A model organism pipeline provides insight into the clinical heterogeneity of TARS1 loss-of-function variants

Abstract

Aminoacyl-tRNA synthetases (ARSs) are ubiquitously expressed, essential enzymes that complete the first step of protein translation: ligation of amino acids to cognate tRNAs. Genes encoding ARSs have been implicated in myriad dominant and recessive phenotypes, the latter often affecting multiple tissues but with frequent involvement of the central and peripheral nervous systems, liver, and lungs. Threonyl-tRNA synthetase (TARS1) encodes the enzyme that ligates threonine to tRNA^THR in the cytoplasm. To date, TARS1 variants have been implicated in a recessive brittle hair phenotype. To better understand TARS1-related recessive phenotypes, we engineered three TARS1 missense variants at conserved residues and studied these variants in Saccharomyces cerevisiae and Caenorhabditis elegans models. This revealed two loss-of-function variants, including one hypomorphic allele (R433H). We next used R433H to study the effects of partial loss of TARS1 function in a compound heterozygous mouse model (R432H/null). This model presents with phenotypes reminiscent of patients with TARS1 variants and with distinct lung and skin defects. This study expands the potential clinical heterogeneity of TARS1-related recessive disease, which should guide future clinical and genetic evaluations of patient populations.

Keywords: Mendelian disease; aminoacyl-tRNA synthetases; protein translation; recessive disease; threonine; threonyl-tRNA synthetase.

Figure 1

Engineered TARS1 variants display a loss-of-function effect in S. cerevisiae (A) Conservation analysis of N412, R433, and G541 TARS1 in H. sapiens (NP_001245366.1), Mus musculus (AAH55371.1), C. elegans (NP_001022033), and S. cerevisiae (NP_116578.3). The targeted residues are highlighted in pink, surrounded by flanking sequences from evolutionarily diverse species. (B) A representative image is shown from three replicates of S. cerevisiae haploid strains with THS1 deleted and transformed with a vector with no TARS1 insert (“Empty”), or with one to express wild-type, N412Y, R433H, or G541R TARS1. Yeast samples were spotted on media containing 5-FOA in serial dilutions (undiluted, 1:10, or 1:100) and then grown at 30°C.

Figure 2

R432H tars-1 impairs viability and locomotion and delays development in C. elegans (A) Genotype analysis of offspring from five broods of G540R/+ tars-1 hermaphrodites. A representative genotyping gel image is shown. The observed number of each genotype is shown, along with the expected number from Mendelian segregation of a benign variant. A chi-square test between observed and expected numbers was performed to determine statistical significance (p < 0.0001). (B) Genotype analysis of offspring from four broods of R432H/+ hermaphrodites, exactly as described in (A). (C) Measurements of body length of R432H/R432H tars-1 worms and wild-type tars-1 worms at 6 days after birth. On day 2, n = 3 for R432H/R432H, then n = 18–30 worms for each subsequent day. For wild-type worms, n = 18–30 for each day. (D) Turn count per minute for R432H/R432H worms (n = 28) and wild-type worms (n = 28) at adult stage P9. (E) Number of body waves initiated from either the head or the tail per minute, for R432H/R432H worms (n = 28) and wild-type worms (n = 28) at P9. For (C)–(E), bars indicate the mean value and 95% confidence intervals. Statistical significance was evaluated using an unpaired t-test with Welch’s correction; ∗∗∗∗p < 0.000001; ns = not significant.

Figure 3

Depleted TARS1 function causes reduced viability and a lung phenotype in a mouse model (A) Genotype analysis of Tars1^R432H/R432H and Tars1^{F538Kfs∗4/+} offspring, genotyped upon weaning at 3 weeks of age. The observed and expected number of each genotype is shown. (B) Genotype analysis of 15 deceased pups, identified within 1 day after birth. The observed and expected number of each genotype is shown. For (A) and (B), a chi-square test was used to determine if the difference between the number of observed and expected genotypes was statistically significant. (C) H&E staining of lung sections from three Tars1^R432H/+ P0 pups (top row) and five Tars1^{R432H/F538Kfs∗4} P0 pups (bottom row). All Tars1^R432H/+ pups were alive when identified at P0. The first four Tars1^{R432H/F538Kfs∗4} pups were dead at P0; the fifth was found alive with a gasping, labored breathing pattern. Arrows point to examples of alveoli, which are collapsed in R432H/+ mice. The black scale bar is 100 μm. (D) PAS staining of lung sections from three Tars1^R432H/+ P0 pups (top row) and four Tars1^{R432H/F538Kfs∗4} P0 pups (bottom row), with labeled examples of bronchioles and arterioles. The inset highlights the magenta PAS signal in the bronchioles of Tars1^R432H/+ mice, and the absence of PAS signal in the collapsed bronchioles of Tars1^{R432H/F538Kfs∗4} mice. The black scale bar is 50 μm.

Figure 4

Depleted Tars1 function causes reduced body size in a mouse model (A) Image of four littermates at P11, grouped together for comparison of body size. The genotype of each mouse is provided. (B) The average weekly weights of female Tars1^{R432H/F538Kfs∗4} mice (n = 9) and female Tars1^R432H/+ (n = 11) littermates are shown, until 23 weeks of age. (C) The average weekly weights of male Tars1^{R432H/F538Kfs∗4} mice (n = 6) and male Tars1^R432H/+ (n = 12) littermates are shown, until 23 weeks of age. For (B) and (C), bars represent the mean value and 1 standard deviation. An unpaired t-test was performed for each week to determine if the difference between the two genotypes was statistically significant. ∗∗∗∗p < 0.0001, ∗∗∗p < 0.001, ∗∗p < 0.01, ∗p < 0.05. All values in (C) that are not marked with an asterisk are not significantly different.

Figure 5

Depleted Tars1 function causes skin phenotypes in a mouse model (A) H&E staining of dorsal skin sections from P0 pups. The upper panel shows a representative image of skin from a Tars1^R432H/+ mouse, and the bottom panel shows a representative image of skin from a Tars1^{R432H/F538Kfs∗4} mouse. Black arrows point to hair follicles in each image. (B) Measurements of epidermal thickness on four Tars1^{R432H/F538Kfs∗4} P0 pups and three Tars1^R432H/+ P0 littermates (n = 25 measurements per pup). Bars indicate the mean value and 95% confidence interval. Statistical significance was determined with a one-way ANOVA with Šidák’s multiple comparisons testing, comparing all animals with R432H/+ pup 1. Only p values < 0.05 are shown (differences between R432H/+ pups were not statistically significant).

Figure 6

Depletion of Tars1 function causes hair loss in a mouse model (A) The cumulative percentage of Tars1^{R432H/F538Kfs∗4} mice (pink, n = 14) and Tars1^R432H/+ mice (gray, n = 23) with hair loss on the back of the head or upper back is shown, until 23 weeks of age. (B) Representative images of four individual Tars1^{R432H/F538Kfs∗4} mice with hair loss; white arrows point to the consistent pattern of upper back bald patches. The depicted mice are between 10 weeks and 17 weeks of age. (C) A representative image of hair phenotypes in Tars1^R432H/+ (top) and Tars1^{R432H/F538Kfs∗4} (bottom) animals. Note that extended hair loss stretches from the head to the middle of the back in the Tars1^{R432H/F538Kfs∗4} mouse, at 14 months of age. A Tars1^R432H/+ littermate is shown above, with signs of barbering by the nose and mild age-related hair thinning on the back.