Mutations in MYLPF Cause a Novel Segmental Amyoplasia that Manifests as Distal Arthrogryposis

Abstract

We identified ten persons in six consanguineous families with distal arthrogryposis (DA) who had congenital contractures, scoliosis, and short stature. Exome sequencing revealed that each affected person was homozygous for one of two different rare variants (c.470G>T [p.Cys157Phe] or c.469T>C [p.Cys157Arg]) affecting the same residue of myosin light chain, phosphorylatable, fast skeletal muscle (MYLPF). In a seventh family, a c.487G>A (p.Gly163Ser) variant in MYLPF arose de novo in a father, who transmitted it to his son. In an eighth family comprised of seven individuals with dominantly inherited DA, a c.98C>T (p.Ala33Val) variant segregated in all four persons tested. Variants in MYLPF underlie both dominant and recessively inherited DA. Mylpf protein models suggest that the residues associated with dominant DA interact with myosin whereas the residues altered in families with recessive DA only indirectly impair this interaction. Pathological and histological exam of a foot amputated from an affected child revealed complete absence of skeletal muscle (i.e., segmental amyoplasia). To investigate the mechanism for this finding, we generated an animal model for partial MYLPF impairment by knocking out zebrafish mylpfa. The mylpfa mutant had reduced trunk contractile force and complete pectoral fin paralysis, demonstrating that mylpf impairment most severely affects limb movement. mylpfa mutant muscle weakness was most pronounced in an appendicular muscle and was explained by reduced myosin activity and fiber degeneration. Collectively, our findings demonstrate that partial loss of MYLPF function can lead to congenital contractures, likely as a result of degeneration of skeletal muscle in the distal limb.

Keywords: Mendelian disease; amyoplasia; congenital contractures; development; distal arthrogryposis; exome sequencing; myosin; skeletal muscle; zebrafish.

Figure 1

Phenotypic Characteristics of Individuals with Recessive or Dominant Distal Arthrogryposis Type 1 due to Variants in MYLPF (A–C) Characteristics shown in family B II-1 with recessive DA1. (A) Camptodactyly of the fingers and radial deviation of the wrists. (B and C) Gross pathology of the right foot illustrating absence of skeletal muscles. (D–F) Characteristics in family H IV-1 with dominant DA1: (D) pursed lips, camptodactyly of the fingers, adducted thumbs, (E) clinodactyly of the fifth digit, and (F) bilateral clubfoot. Table 2 contains a detailed description of the phenotype of each affected individual and Figure S1 provides a pedigree of each family with DA1 due to variants in MYLPF.

Figure 2

Genomic Model of MYLPF MYLPF is composed of seven exons, each of which consists of protein-coding (blue) and non-coding (orange) sequence. The approximate location of each pathogenic variant is indicated by an arrow. The p.Cys157Phe (p.C157F) and p.Cys157Arg (p.C157R) variants are each found in three families (×3) and lead to a recessive phenotype (purple circle). The p.Ala33Val (p.A33V) and p.Gly163Ser (p.G163S) variants lead to a dominant phenotype (red circle).

Figure 3

Zebrafish mylpfa Mutants Have Weakened Myotomes and Paralyzed Fin Muscle (A–D) RNA in situ hybridization showing (A, B) mylpfa and (C, D) mylpfb expression at 20 hpf (A, C) and 52 hpf (B, D). Both genes are expressed exclusively in fast muscle, including somitic muscles, fin muscle (green arrowhead), and posterior hypaxial muscle (red arrowhead). (E) Transcript abundance of mylpfa (blue) and mylpfb (orange) through early larval development, using data provided in the EMBL-EBI expression atlas. (F) Alignment of wild-type (WT) and mutant genomic sequence across the mylpfa^oz30 and mylpfa^oz43 lesions. mylpfa^oz30 is a 20 bp deletion within exon 3 predicted to frameshift the 169 amino acid protein after amino acid 76, and mylpfa^oz43 is a 1 bp deletion within exon 2 predicted to frameshift the protein after amino acid 52. (G) Diagram of wild-type and predicted mutant Mylpfa proteins. Both mutant alleles should truncate the protein within the first EF-hand domain (black boxes) and introduce short stretches of aberrant amino acids after the frameshift (brown). (H) Alignment of wild-type and predicted mutant proteins in the region of frameshift, showing normal sequence (black) and aberrant residues (brown). (I) Superimposed time-lapse images (from Video S1) showing fin motion in a wild-type embryo (yellow crescent arrows) and motionless fins in a mylpfa^oz30 mutant (yellow dots) at 4 dpf. Similar results were obtained using a second mutant allele, mylpfa^oz43 (not shown). (J) Quantification of fin beats per minute averaged over left and right sides, in mylpfa^oz30 mutant (n = 12) or phenotypically wild-type sibling (n = 12) fish at 4 dpf. We have never observed a fin beat in more than 100 mylpfa^oz30 and mylpfa^oz43 mutant fish examined. (K) Trunk muscle contractile force in mylpfa^oz30 and wild-type or heterozygous siblings at 3 dpf. No significant differences are found between wild-type and heterozygous fish. However, homozygous mylpfa mutant fish are significantly weaker than non-mutant siblings at all stimulation frequencies. (L) Representative image of fluorescently labeled actin filaments tracked in the in vitro motility assay, with colored lines showing traces of individual filaments over 50 frames of imaging (2.5 s). Within this time period, actin filaments on wild-type myosin extracts typically move further than do filaments on mylpfa mutant myosin extracts. (M) Actin filament speeds measured using the in vitro motility assay generated by extracted myosin. Myosin from mylpfa homozygous mutant fish propel actin filaments significantly slower than myosin from wild-type siblings. Numbers shown in each bar indicate the experimental N; each experiment uses myosin extracted from two fish (see Material and Methods). Asterisks in (J, K, M) indicate p thresholds for the WT/het pools versus mutant; ^∗p < 0.05, ^∗∗p < 0.001, ^∗∗∗p < 0.0001. Error bars in (J, K, M) represent standard deviation. Statistical comparisons in (J) and (M) use Student’s t test; in (K), thresholds are determined by ANOVA followed by Tukey-Kramer comparisons. Scale bar in (I) is 250 μm.

Figure 4

Embryonic Muscle Degenerates in mylpfa Mutant Zebrafish (A and B) Confocal images showing muscle morphology at 26 hpf in wild-type (A) or mylpfa^−/− (B) embryos. Myonuclei are labeled using Rbfox1l immunolabeling (green), fast-twitch fibers are labeled using F310 immunolabeling (red), and slow-twitch fibers are labeled using Tg(smyhc1:EGFP)i104 (blue). (C and D) Confocal z sections showing muscle morphology of live 6 dpf larvae expressing a fast muscle cell membrane transgene Tg(mylpfa:lyn-Cyan)fb122 (green) and a myonuclear transgene myog:H2B-mRFP (red). These fish also express a slow muscle marker Tg(smyhc1:EGFP)i104 (blue), which is largely lateral to the plane of focus; pink arrowheads point to slow muscle cells within the shown plane. Compared to wild-type controls (C), mylpfa^−/− myofibers have irregular membrane structure (D). (E and F) Confocal projections of the same myotomes from (C) and (D) showing slow muscle fibers (white). (G–J) Confocal projections of pectoral fin and PHM muscles, imaged on 3.25 dpf and again on 4.25 dpf, in fish expressing transgenes Tg(mylpfa:lyn-Cyan)fb122 in fast muscle (green), myog:H2B-mRFP in all myonuclei (red), and Tg(smyhc1:EGFP)i104 in slow muscle (blue). Fin muscle and PHM express Tg(mylpfa:lyn-Cyan)fb122 but not Tg(smyhc1:EGFP)i104, as expected for muscles predominantly comprised of fast-twitch fibers.^, In striking contrast to wild-type (G, H), mylpfa mutant muscle fibers degenerate between 3.25 and 4.25 dpf (I, J). (K–M) Images from a time-lapse of mylpfa mutant PHM degeneration (Video S3). Muscle fibers which initially appear wavy (aqua outline), often become narrow (red arrowheads) before breaking apart (asterisk). A myofiber that appears during imaging is outlined in magenta. Insets in (C), (D), and (G)–(J) show Tg(mylpfa:lyn-Cyan)fb122 in greyscale. Scale bars in (A) for (A) and (B), in (E) for (C)–(F), in (G) for (G)–(J), and in (K) for (K)–(M) are 50 μm.

Figure 5

Mylpf Protein Sequence and Structure Comparisons Identify Key Conserved Residues (A) Model of rabbit Mylpf protein in complex with the neck and head region of myosin heavy chain in rigor. The heavy chain (cyan) and essential light chain (orange) are rendered using a space-filling model and the light chain is shown using a ball and stick model (yellow) except for three residues that align with MYLPF Ala33, Cys157, and Gly163, which are rendered in space filling models; we refer to these by their human numbering. Ala33 (green) is adjacent to a Lys residue in the heavy chain (white). Gly163 (magenta) directly contacts a Phe residue in the heavy chain (dark blue). In contrast, Cys157 (red) is found internal to Mylpf protein. (B) Overview of myosin interaction with thin filaments, color-coded as in (A). Mylpf protein binds to the heavy chain region that bends toward the thin filament (arrows). (C) Alignment of select vertebrate Mylpf proteins and invertebrate Mlc2 proteins highlighting conservation of Ala33, Cys157, and Gly163. The percent identity (% ident.) between human MYLPF and aligned proteins is shown to the right of the alignments. (D–G) Magnified views of myosin heavy and light chain genes showing how Ala33 and Gly163 positions vary between scallop (D), squid (E), chicken (F), and rabbit (G). Cys157 is located internally to the two vertebrate Mylpf structures (F, G). Color coding in (D)–(G) is the same as in (A). (H) Alignment of human regulatory and essential light chain proteins highlighting conservation of Ala33, Gly163, and Cys157. The tissue that express each ortholog is indicated as follows: embryonic skeletal muscle (Emb), fast-twitch skeletal muscle (Fast), slow-twitch skeletal muscle (Slow), cardiac muscle (Card), and non-sarcomeric tissue (NS). The first residue in the shown aligned portions are numbered for each protein. Proteins in (H) are arranged by their similarity to MYLPF.