- Top of page
Mutation of the non-muscle myosin heavy chain type II-A results in MYH9-related hereditary macrothrombocytopenia (HMTC), including four autosomal dominant platelet disorders: May-Hegglin anomaly (MHA), Sebastian (SBS), Fechtner (FS) and Epstein (EPS) syndrome. Denaturing high-performance liquid chromatography (DHPLC) was optimised for rapid screening of the seven exons harbouring all but one of the previously reported mutations of MYH9. Individuals from 13 families with phenotypes suggestive of MYH9-related HMTC were screened for mutations by DHPLC followed by direct sequencing of samples with aberrant column retention time. Mutations were identified in all 13 families. Six distinct missense heterozygous mutations were found in 10 families, including six families with MHA or SBS (E1841K, D1424N), three families with FS (R702H, R1165C, and D1424Y), and one family with EPS (S96L). A truncating mutation (R1933X) was found in three MHA families. A review of all published mutations suggests that mutation in the C-terminal coiled coil region or truncation of the tailpiece is associated with haematological-only phenotype, while mutation of the head ATPase domain frequently is associated with nephropathy and/or hearing loss. Mutations of other regions have intermediate expression of non-haematological characteristics. Further study is required to confirm these associations and understand the molecular basis for this genotype–phenotype relationship.
The hereditary macrothrombocytopenias (HMTCs) are a diverse group of hereditary disorders characterised by autosomal inheritance (either dominant or recessive) and high penetrance (Jantunen, 1994). Traditionally, these disorders were arbitrarily characterised based on the absence or presence of leucocyte inclusions. In the former group are the autosomal recessive Bernard-Soulier [caused by platelet glycoproteins Ib alpha, Ib beta and IX gene mutations; Online Mendelian Inheritance in Man (OMIM) entry 231200] and autosomal dominant Gray platelet (OMIM 139090) syndromes (Jantunen, 1994). The prototype of the HMTCs with leucocyte inclusions is May-Hegglin anomaly (MHA; OMIM 155100), characterised by the triad of thrombocytopenia, giant platelets and basophilic inclusions in the peripheral blood leucocytes. The triad of MHA also occurs in conjunction with other traits, especially deafness and nephritis. In Fechtner syndrome (FS; OMIM 153640), both deafness and nephritis are present but the leucocyte inclusions are distinct from those of MHA by light and electron microscopy (Heynen et al, 1988; Greinacher & Mueller-Eckhardt, 1990). The presence of deafness and nephritis has suggested that FS is a variant of Alport syndrome (OMIM 104200). The Sebastian platelet syndrome (SBS) shows the same platelet and leucocyte morphology observed in FS, but without the additional Alport-like stigmata (Greinacher & Mueller-Eckhardt, 1990). Macrothrombocytopenia (MTC) with leucocyte inclusions has also been reported in conjunction with renal impairment alone (Brivet et al, 1981; Bepler et al, 1994; Demeter et al, 2001).
The classification of the HMTCs with leucocyte inclusions has recently been clarified by genetic linkage and mutational analyses that led to the identification of the molecular basis for these disorders (Kelley et al, 2000; Seri et al, 2000). Mutation of the non-muscle myosin heavy chain type A, encoded by the MYH9 gene, was found in all three of the HMTCs with leucocyte inclusions (Kelley et al, 2000; Seri et al, 2000). Mutation of MYH9 was also found in the Epstein syndrome (Seri et al, 2002), which is characterised by autosomal dominant inheritance of MTC, hearing loss, and nephropathy without leucocyte inclusions. Thus, these four syndromes are allelic variants that we refer to as MYH9-related MTCs. A single family with predominantly autosomal dominant hearing loss and MYH9 mutation has been reported (Lalwani et al, 2000).
Mutational analysis techniques used to detect MYH9 mutations have included direct DNA sequencing of polymerase chain reaction (PCR) products and PCR-restriction fragment length polymorphism (RFLP) analysis. The limitations of direct DNA sequencing of PCR products are known and include the requirement of purification of PCR products, possible insensitivity of mutation detection and relative expense. The PCR-RFLP analysis is rapid and inexpensive but is able to detect mutations only at known sites. A group of mutation detection techniques, such as single-strand conformational polymorphism analysis, examine DNA fragments for sequence variants on the basis of aberrant electrophoretic migration of single stranded DNA and have yet to be applied to MYH9 mutation analysis. Recently, a new semi-automated method has been introduced for mutation analysis, denaturing high-performance liquid chromatography (DHPLC) (Xiao & Oefner, 2001). Mutation detection with this method relies on the formation and separation of double-strand DNA fragments that contain mismatched bases (heteroduplexes). Heteroduplex DNA is generated by denaturing and reannealing a mixed population of reference wild-type samples and mutant DNA. Due to mismatches in base pairs, the heteroduplex fragments will have melting properties different from those of the homoduplexes. Under conditions of partial heat denaturation (within a linear acetonitrile gradient), heteroduplexes having internal sequence variations display reduced column retention time relative to their homoduplex counterparts, and thus can be resolved.
In this study we have determined the conditions for performing DHPLC on MYH9-related HMTCs and applied this to 13 new families with suspected MYH9-related MTCs. Genotype–phenotype analysis of all reported mutations suggested a domain-specific relationship between the location of the MYH9 mutation and penetrance of the non-haematological characteristics of these disorders.
- Top of page
To rapidly screen for germline MYH9 mutations, we developed a mutation detection schema using DHPLC. Initially, the seven exons containing all but one of the known MYH9 mutations were amplified from an unaffected individual and chromatographed at the optimised melting temperatures (Table III). Initial temperature and Buffer B [WAVE Optimized® TEAA Buffer B (0·1 mol/l Triethylammonium acetate; TEAA, pH 7·0 + 25%v/v Acetonitrile); Transgenomic Inc.] concentration of the elution gradient for DHPLC analysis were determined for each of the amplicons using the Wavemaker software and then adjusted empirically. If the melting profile of the fragment was predicted to be homogeneous, one analysis temperature was used. However, if two different melting domains were detected, the amplicon was analysed under two different sets of conditions. Table III shows the experimental conditions for DHPLC analysis of the commonly mutated exons of human MYH9. A pilot study of previously characterised mutations in exons 25, 38 and 40 was performed with the amplicons and analysis conditions described in our current study. Samples with wild-type sequence eluted as a single peak, whereas samples with heterozygous nucleotide changes were characterised as having a shoulder at the leading edge of the peak or double or multiple peaks (Fig 1). These chromatograms could be resolved at the specified temperatures without mixing sample with wild type as affected individuals are heterozygous for MYH9 mutation.
Figure 1. DHPLC chromatograms corresponding to MYH9 exons 1 and 26 mutations from heterozygous patients compared with a normal sample. The profiles of the mutants (lower panels) show two or more peaks and are easily distinguished from profiles of the wild type (upper panels), which shows a single peak.
Download figure to PowerPoint
Next, we analysed seven exons of MYH9 in an affected individual from each of 13 new families (Table I) by DHPLC. One or more sequence variant was identified in six exons among these 13 families. Mutations were confirmed through direct sequencing (Table IV). All mutations were heterozygous and, except the truncating mutation at R1933, all were missense mutations. Five of the seven mutations were located in the coiled coil domain whereas the remaining two were present in the globular head containing the ATPase domain. Of the latter, the S96L was identified in an individual diagnosed with EPS in Family MHA-26 and the R702C mutation was found in Family MHA-12 with FS. The R702C mutation was not found in the genetically confirmed parents of the single affected individual of Family MHA-12 (data not shown), demonstrating that this was a new mutation in this family.
Table IV. Results of MYH9 mutational analysis in 13 families using DHPLC.
Among mutations in the coiled-coil domain, the E1841K and R1933Ter mutations were identified in five and three families respectively. All eight families were diagnosed with MHA. One family with FS, MHA-24, had the R1165C mutation, but no family had the previously reported R1165L mutation. Aspartic acid at position 1424 is the only residue reported to have three different mutations. We identified a D1424Y substitution in Family MHA-25 with FS and D1424N in Family MHA-21 with MHA.
- Top of page
Non-muscle myosin IIA (NMM-IIA) is a hetero-hexameric complex composed of a pair of heavy chains (NMMHC-IIA) and two pairs of light chains. Each myosin heavy chain contains two characteristic regions: a globular region at the amino terminus that catalyses ATP hydrolysis and binds to actin to generate force and movement, and an alpha-helical carboxy-terminal tail region facilitating assembly of bipolar filaments. A short non-helical sequence at the C-terminus of vertebrate non-muscle myosin, called the tailpiece, is not found in sarcomeric muscle myosin. NMM-IIA subserves essential functions in cell spreading, cell motility, maintenance of cell morphology and cytokinesis (Hartwig & DeSisto, 1991; Maupin et al, 1994; Kolega, 1998). NMMHC-IIA is encoded by the MYH9 gene, which is located on chromosome 22q12–13 and comprises 44 exons, 40 of which contain the coding sequence. Considering the large size and exonic fragmentation of the gene, mutation scanning of MYH9 gene is tedious. DHPLC is an efficient and cost-effective mutational screening technique that has been applied to numerous disease genes. In this study we developed a DHPLC method for the systematic detection of sequence variations in the MYH9 gene. The detection of a mutation in each of 13 families suggests that the DHPLC method is an appropriate technique for mutation screening for individuals with suspected MYH9-related disorders.
Our results on 13 new families in this report and those from prior reports extend to 85 the total number of MYH9 mutations described in the literature (Table V). Mutation of MYH9 occurs at 21 different residues clustered in nine exons (1, 10, 16, 25, 30, 37, 38 and 40). Only one mutation has been reported in each of exons 10 and 37. Eighty-one per cent of mutations occur at only five sites (amino acids 702, 1165, 1424, 1841 and 1933) and all reported germline mutations occur at only about 1% of the 1961 amino acids of NMMHC-IIA despite extensive regions of marked phylogenetic conservation. Among the possible explanations for the limited distribution of MYH9 mutations, several factors argue against a founder effect for this limited mutation pattern: (i) identical MYH9 mutations have been found in diverse populations; (ii) a new, sporadically occurring mutation has been observed at most sites of mutation (such as our family MHA-12); and (iii) absence of a shared haplotype among families with identical mutations (Kelley et al, 2000; Heath et al, 2001).
An alternative explanation for the limited mutational spectrum of MHY9 is a common mutational mechanism operative at restricted sites. Point mutations in 65 of 78 (83%) families with MYH9-related disorders are transitions at sites of CG dinucleotides. Such mutations typically result from deamination of methylated cytosine residues. Arginine, the most frequently mutated amino acid of NMMHC-IIA, is encoded by four codons, all of which contain a CG dinucleotide. Non-sense mutation of arginine in NMMHC-IIA may be analogous to similar mutation in other genes, such as that found in the factor VIII gene (Tuddenham et al, 1991). However, among 121 arginine residues in NMMHC-IIA, only four (R702, R705, R1165 and R1933) have been reported to be mutant and among the 313 CG dinucleotides in the MYH9 coding region, only eight have been reported as mutant. These observations suggest either limited sites of methylation within MYH9 or alternative explanations for the restricted pattern. Limited methylation alone does not adequately explain the observed mutations as one would expect mutation at both sites of a CG dinucleotide, resulting in adjacent CT and GA transitions. Furthermore, the 17% of mutations that are not transitions at CG dinucleotides occur almost exclusively at or near sites of the more common transition mutations of CG dinucleotides. Thus, it is likely that the restricted mutation pattern observed in MYH9 is related to the effects of mutation on NMM-IIA function.
At least two functional scenarios may exist to explain the limited number of mutations observed in humans in NMM-IIA. First, the MYH9-associated disorders may require complete inactivation of MYH9 function and only a limited number of mutations are likely to result in total loss of function by co-inactivation of wild type MYH9. This is supported by the observation that mutant NMMHC-IIA appears to sequester wild-type NMMHC-IIA in leucocyte inclusions (Kunishima et al, 2003), suggesting a dominant negative interaction. A prediction of this scenario is that there would be similar phenotypes (perhaps without leucocyte inclusions) with an autosomal recessive inheritance pattern resulting from biallelic mutation of MYH9. This phenomenon has not been reported to date. A second non-exclusive scenario is that unobserved mutations are incompatible with life. The near ubiquitous expression of MYH9 and its role in essential cellular functions is consistent with this hypothesis. Of note, the MYH10 knock-out mouse is non-viable (Tullio et al, 1997) and an MYH9 knock-out mouse was reported to be embryonic lethal (Conti et al, 2004). However, the marked disruption of ATPase function in some MYH9 mutants (Hu et al, 2002) suggests more subtle differences between observed and expected mutants.
With the accumulation of characterised families, genotype–phenotype patterns are emerging. One consistent correlation was found among the 10 families with R702 mutations, all of which shared the clinical features of MTC, deafness and nephritis. This arginine residue is conserved in non-muscle and smooth muscle myosins in all species whose sequence is known. Molecular modelling of the R702C mutation using chick smooth-muscle X-ray crystallographic coordinates suggested this mutation would disrupt ATPase activity (Seri et al, 2002) and in vitro assay confirms reduced ATPase activity (Hu et al, 2002). Thus, it appears that R702 mutations have a dramatic effect on biochemical function that correlates with a more severe phenotype.
Another pattern is found among the mutations of the C-terminus, E1841K and R1933X. These are among the most common mutations in MYH9 and most commonly associated with only haematological manifestations. MHA or SBS were identified in 15 of 19 (79%) of E1841K families and 13 of 14 (93%) of R1933X (n = 14) families. Two relatively small domains in the C-terminal end of myosin are critical for filament assembly and disassembly, a perquisite to myosin function. The C-terminal 32 residues of the coiled-coil domain has been termed the assembly competence domain (ACD) as proper rod assembly is not possible without the presence of both the charge periodicities along the rod and critical residues at the coiled-coil end (Sohn et al, 1997; Ikebe et al, 2001). Residue E1841 is located 54 amino acids upstream of the ACD, thus, mutation at this site may impair filament formation. In contrast, the non-helical tailpiece plays a role in filament disassembly. Mts1, a 9-kDa S100-family protein, binds to the non-helical tail of non-muscle myosin II and destabilises filaments (Ford et al, 1997; Kriajevska et al, 1998; Murakami et al, 2000). NMM IIA exhibits a dynamic monomer-filament equilibrium that shifts according to changes in cellular functions and motilities. The R1933X mutation and three frame shift mutations (5774delA; 5779delC; 5825delG) in exon 40 truncate the tailpiece containing the mts1 binding site and thus may interfere with filament disassembly. A clear explanation for the close association of mutations near the C-terminus with predominantly haematological manifestations is not readily apparent.
Mutations at R1165 and D1424 have been found in 25 families. The frequency of hearing loss and renal disease, 57% and 35%, respectively, appears to be intermediate between that found in the head mutations and the C-terminal mutations. The variation in phenotype, despite identical gene mutation, may result from genetic modifiers, age or environmental factors. Further study is required to confirm and understand the molecular basis for the genotype–phenotype relationship and to determine the relative contribution of genetic and environmental factors in the manifestations of these disorders.