To Charge or Not to Charge: Mechanistic Insights into Neuropathy-Associated tRNA Synthetase Mutations

Abstract

Aminoacyl-tRNA synthetases (ARSs) are ubiquitously expressed, essential enzymes responsible for the first step of protein translation—attaching amino acids to cognate tRNA molecules. Interestingly, ARS gene mutations have been implicated in tissue-specific human diseases, including inherited peripheral neuropathies. To date, five loci encoding an ARS have been implicated in peripheral neuropathy, and alleles at each locus show loss-of-function characteristics. The majority of the phenotypes are autosomal dominant, and each of the implicated enzymes acts as an oligomer, indicating that a dominant-negative effect should be considered. Based on current data, impaired tRNA charging is likely a central component of ARS-related neuropathy. Future efforts should focus on testing this notion and developing strategies for restoring ARS function in the peripheral nerve.

Introduction to Aminoacyl-tRNA Synthetases

The central dogma of molecular biology provides an outline for describing the fundamental processes of information transfer from DNA to mRNA to protein. The last step of this framework—protein translation—requires the covalent attachment of amino acids to cognate tRNA molecules. This essential reaction is required in the cytoplasm and mitochondria of all eukaryotic cells and is performed by a family of aminoacyl-tRNA synthetases (ARSs) [1]. To facilitate tRNA charging, the human nuclear genome harbors 37 ARS loci: 17 encoding a cytoplasmic enzyme, 17 encoding a mitochondrial enzyme, and 3 encoding a bi-functional enzyme—an ARS that functions in both cell compartments via an alternatively employed mitochondrial targeting sequence [2]. ARS enzyme nomenclature includes the single-letter code of the amino acid that the enzyme recognizes, followed by ‘ARS’ and, if relevant, a ‘2’ to indicate a mitochondrial enzyme. For example, glycyl-tRNA synthetase is referred to as ‘GARS’ and mitochondrial tyrosyl-tRNA synthetase is referred to as ‘YARS2’. ARSs catalyze the aminoacylation reaction in two steps:

• Step 1 ARS + AA + ATP     →  ARS(AA-AMP) + PP_i

• Step 2 ARS(AA-AMP) + tRNA → AA-tRNA + AMP + ARS

In the first step, the ARS enzyme binds to the amino acid (AA) and one ATP molecule to form the aminoacyl adenylate intermediate; pyrophosphate (PP_i) is then released. In the second step, the tRNA interacts with the anticodon-binding domain of the ARS and is covalently linked to the amino acid, thus generating free AMP. The charged tRNA is then available for protein translation.

One unexpected finding regarding these ubiquitously expressed, essential enzymes is that mutations in ARS genes cause tissue-specific human diseases. Over the past decade, fourteen ARS loci have been implicated in clinical phenotypes (Table 1), and five of these are associated with inherited peripheral neuropathies. This review considers recent findings on the functional consequences of ARS mutations identified in patients with peripheral neuropathy. A better understanding of the molecular pathology of ARS mutations will improve our knowledge of the function of these enzymes and provide a platform for therapeutic development.

Table 1.

Human diseases associated with ARS mutations

Gene (amino acid)	Locus	Location of Function	Disease Nomenclature	OMIM No.	Clinical Description	References
GARS (glycine)	7p15	Cytoplasm and Mitochondria	CMT2D or dSMA-V	601472 600794	Axonal motor neuropathy with variable sensory involvement, maximal in the upper extremities.	[8]
KARS (lysine)	16q23.1	Cytoplasm and Mitochondria	CMTRIB	613641	A single patient with intermediate CMT disease, developmental delay, self-abusive behavior, dysmorphic features, and vestibular Schwannoma.	[20]
AARS (alanine)	16q22	Cytoplasm	CMT2N	613287	Motor and sensory axonal neuropathy, variable sensorineural deafness.	[16]
HARS (histidine)	5q31.3	Cytoplasm	Not applicable	142810	A single patient with motor and sensory axonal neuropathy. Sensory phenotype is more severe than motor phenotype.	[21]
YARS (tyrosine)	1p35.1	Cytoplasm	DI-CMTC	608323	Motor and sensory neuropathy with moderately reduced conduction velocities.	[15]
AARS2 (alanine)	6p21.1	Mitochondria	Not applicable	614096	Infantile mitochondrial cardiomyopathy.	[4]
DARS2 (asparagine)	1q25.1	Mitochondria	LBSL	611105	Leukoencephalopathy with brain stem and spinal cord involvement and lactate elevation. Spasticity and ataxia with variable cognitive decline.	[35]
EARS2 (glutamine)	16p12.2	Mitochondria	LTBL	614924	Early-onset leukoencephalopathy with thalamus and brain stem involvement and high lactate. Hypotonia, seizures, and visual impairment in more severe cases.	[36]
FARS2 (phenylalanine)	6p25.1	Mitochondria	Alpers syndrome	611592	Early-onset fatal encephalopathy, seizures, developmental delay, and liver dysfunction.	[37]
HARS2 (histidine)	5q31.3	Mitochondria	Perrault Syndrome	233400	Sensorineural hearing loss in males and females, and ovarian dysgenesis in females.	[38]
MARS2 (methionine)	2q33.1	Mitochondria	ARSAL	609728	Autosomal recessive spastic ataxia with leukoencephalopathy and cerebellar atrophy.	[39]
RARS2 (arginine)	6q16.1	Mitochondria	PCH6	611523	Pontocerebellar hypoplasia, cerebral atrophy and variable respiratory chain defects in muscle.	[40]
SARS2 (serine)	19q13.2	Mitochondria	HUPRA Syndrome	613845	Hyperuricemia, pulmonary hypertension, renal failure in infancy and alkalosis.	[41]
YARS2 (tyrosine)	12p11.21	Mitochondria	MLASA	613561	Myopathy, lactic acidosis and sideroblastic anemia. Muscle weakness, exercise intolerance and dysphagia.	[3]

Human Diseases Associated with Aminoacyl-tRNA Synthetase Mutations

Loss-of-function alleles at nine loci encoding a mitochondrial ARS have been implicated in autosomal recessive diseases (Table 1). Given the critical role of these enzymes in mitochondrial translation it is perhaps not surprising that the associated phenotypes are classified as ‘mitochondrial’ diseases. For example, mutations in the mitochondrial tyrosyl-tRNA synthetase (YARS2) gene cause the mitochondrial respiratory chain disorder MLASA Syndrome (Myopathy, Lactic Acidosis, and Sideroblastic Anemia)[3], and mitochondrial alanyl-tRNA synthetase (AARS2) mutations cause infantile cardiomyopathy—a disease characterized by large, hypertrophic heart and fatality within the first 1-2 years of life [4]. Thus, mitochondrial ARS enzymes are critical for cellular respiration, and loss-of-function mutations in the nuclear loci that encode them lead to recessive mitochondrial diseases.

A more puzzling observation is that mutations in genes encoding cytoplasmic ARSs have been identified in patients with inherited peripheral neuropathy and an axonal pathology (Table 1). A diverse group of diseases, peripheral neuropathies are primarily characterized by motor and sensory dysfunction in the distal limbs. These disorders constitute a major public health concern, with prevalence as high as 10% in some aging populations [5]. Peripheral neuropathies can be classified into three main types: 1. Acquired (e.g. alcoholic neuropathy); 2. Syndromic (e.g. diabetic neuropathy); and 3. Primarily inherited.

Inherited motor and sensory neuropathy [also known as Charcot-Marie-Tooth (CMT) disease] encompasses a heterogeneous class of neurological diseases characterized by muscle weakness and sensory loss in the extremities [6]. This axon-length dependent phenotype reflects the requirement of peripheral nerve axons and Schwann cells to function over long distances (e.g., some peripheral nerves extend up to 1 meter in length) and long periods of time (e.g., the decades-long span of a human life). Indeed, the two major subtypes of inherited peripheral neuropathy are those caused by impaired function of Schwann cells (CMT disease type 1) and those caused by impaired axonal function (CMT disease type 2) [7]. The first human genetic diseases associated with ARS mutations were Charcot-Marie-Tooth disease type 2D (CMT2D) and distal spinal muscular atrophy type V (dSMA-V). Both of these axonal neuropathies are inherited in an autosomal dominant fashion, mainly affect the upper extremities, and are caused by mutations in the glycyl-tRNA synthetase (GARS) gene [8]. Sensory deficits in patients with CMT2D distinguish them from patients with dSMA-V. To date, a total of 11 GARS mutations have been identified [8-14].

Mutations in four additional ARS genes have been identified in patients with CMT disease. Two missense mutations (G41R and E196K) and one in-frame deletion of four amino acids (153-156delVKQV) in the tyrosyl-tRNA synthetase (YARS) gene were identified in three families with autosomal dominant intermediate CMT disease type C (DI-CMTC) [15]. DI-CMTC is characterized by axonal and demyelinating pathologies. Three missense mutations (N71Y, R329H, and D893N) in the alanyl-tRNA synthetase (AARS) gene have been identified in five families with autosomal dominant CMT type 2N (CMT2N) [16-19]. CMT2N is an axonal form of CMT disease, and one family with the R329H mutation also contains individuals with sensorineuronal deafness. Interestingly, R329H AARS is a recurrent mutation and is the most common CMT-associated ARS mutation identified to date [18]. Compound heterozygosity for lysyl-tRNA synthetase (KARS) mutations (L133H and Tyr173SerfsX7) was identified in a patient with intermediate CMT disease, developmental delay, self-abusive behavior, dysmorphic features, and vestibular Schwannoma [20]. This phenotype was subsequently termed CMT recessive intermediate type B (CMTRIB). While it was impossible to discern if the neuropathy in this patient is dominant or recessive, the combination of these two mutations in a single individual may provide insight into the molecular pathology of dominant ARS-related neuropathy (see below). Finally, a HARS mutation (R137Q) was recently identified in a patient with sporadic motor and sensory neuropathy [21]. Combined, these observations indicate that tRNA charging enzymes are particularly important for the function of peripheral nerve axons.

Aminoacyl-tRNA Synthetase Mutations are Associated with a Loss-of-Function Effect

Loss-of-function mutations can cause a number of distinct mechanisms including impaired mRNA or protein expression, impaired protein function, or impaired protein localization. The molecular mechanism by which ARS mutations cause autosomal dominant peripheral neuropathy remains elusive. However, functional studies on a subset of disease-associated ARS mutations provide strong evidence for a loss-of-function effect (Table 2). ARS enzyme activity can be assessed via an in vitro aminoacylation reaction. In this assay, purified recombinant ARS (wild-type or mutant) is combined with a radiolabeled amino acid substrate, cognate tRNA, and ATP, and analysis of incorporated versus unincorporated amino acid is determined to assess completion of the reaction. Aminoacylation assays have been conducted for disease-associated ARS enzymes (Table 2) and nine mutations in the AARS, GARS, YARS, and KARS loci have been associated with impaired tRNA charging. However, three GARS mutations (L129P, G240R, and S581L) do not seem to overtly affect enzyme activity. Thus, impaired tRNA charging does not explain the pathogenesis of all disease-associated ARS mutations.

Table 2.

Summary of functional analyses performed on disease-associated ARS mutations

Gene	Variant	Biochemistry	Yeast	Localization	Fly neurons	References
GARS	A57V	NP	NP	NP	NP	[11]
GARS	E71G	Normal	Normal	Normal	LOF	[8,24-26]
GARS	L129P	LOF	LOF	LOF	LOF	[8,24-26]
GARS	G240R	LOF	Normal	LOF	NP	[8,24,25]
GARS	P244L	NP	NP	NP	NP	[13,25,42]
GARS	I280F	NP	NP	NP	NP	[10]
GARS	H418R	NP	LOF	LOF	NP	[14,24,25]
GARS	D500N	Normal	NP	LOF	NP	[9,25]
GARS	G526R	LOF	LOF	LOF	NP	[8,24,25,42]
GARS	S581L	Normal1	NP	LOF	NP	[10,12,25,43]
GARS	G598A	NP	Normal	NP	NP	[10,34]
YARS	G41R	LOF	LOF2	LOF	NP	[15,44]
YARS	153-156ΔVKQV	LOF	NP	NP	NP	[15,44]
YARS	E196K	LOF3	Normal2,4	LOF	NP	[15,44]
KARS	L133H5	LOF	Normal	NP	NP	[20]
KARS	Y173SfsX75	NP	LOF	NP	NP	[20]
AARS	N71Y	LOF	LOF	NP	NP	[16,18]
AARS	R329H	LOF	LOF	NP	NP	[16,18]
AARS	D893N	NP	NP	NP	NP	[19]
HARS	R137Q	NP	LOF	NP	NP	[21]

NP – Not Performed.

LOF – Functional studies consistent with a loss-of-function effect.

¹Also reported as reduced [43].

²Mutation also displays a dominant-negative effect in yeast.

³Also reported as having activity comparable to wild-type YARS [44].

⁴However, reported as only partially complementing the endogenous yeast gene.

⁵These two mutations were identified on different chromosomes in a single patient.

While data from aminoacylation assays have provided insight into the functional consequences of ARS mutations, interpretations from these studies are limited because they are outside the context of a eukaryotic cell. To address this, yeast complementation studies have been employed to determine if ARS mutations impair enzyme function in vivo. Here, the endogenous yeast locus is deleted and replaced with an extra-chromosomal copy of the yeast gene that is mutated to encode the human disease-associated residue. Viability is then evaluated as an indirect measure of ARS function. Of the 13 ARS mutations studied in yeast complementation assays, eight show a loss-of-function effect while five do not (Table 2). Importantly, three ARS mutations (e.g., G240R GARS) do not cause a detectable loss of function in yeast despite data indicating impaired activity in aminoacylation assays. Likely explanations for these observations include differences between the human and yeast enzymes, or residual enzyme activity that is sufficient for yeast cell growth. Indeed, the latter explanation resolves the discrepancy for L133H KARS [20] and underscores a limited resolution of the yeast complementation assay. However, it is important to point out that each ARS mutation that displays a loss of function in yeast assays also shows impaired activity in biochemical assays, indicating the informativeness of a loss-of-function result in yeast. Interestingly, two YARS mutations (G41R and E196K) also impart a dominant-negative effect on yeast cell growth [15]. These data introduced the idea that a loss-of-function ARS protomer could exacerbate ARS functional deficits via interference with the wild-type ARS protomer in the context of a holoenzyme—indeed, all disease-associated ARS enzymes can function as oligomers [8,15,16,20,21].

Yeast has emerged as an informative model system for determining the functional consequences of ARS mutations in vivo. However, these studies are unable to put the mutations in the correct physiological context—a peripheral nerve axon. Protein translation occurs in the cytoplasm and axoplasm of neurons to meet the housekeeping requirements of these highly specialized and polarized cells [22], and impaired protein translation can result in distinct neurological phenotypes [23]. Interestingly, localization studies revealed that wild-type ARS enzymes are present in axons in vitro and in vivo [15,24], suggesting that they function in these structures. In contrast, mutant GARS and YARS proteins do not localize properly to axons [15,24,25]. It has been posited that this defect in ARS localization may deplete tRNA charging (and thus protein translation) within the axon [24].

Examination of biochemical, yeast, and localization studies (Table 2) reveals that all but one of the evaluated human ARS mutations (E71G GARS) are associated with a loss-of-function characteristic—five ARS mutations (A57V, P244L, I280F, and G598A GARS, and D893N AARS) have not been adequately evaluated. It was not until GARS mutations were assessed in the context of a developing neuron in vivo that E71G was found to have a loss-of-function effect. A mosaic genetic screen in Drosophila melanogaster revealed that GARS is required for the proper terminal arborization of axons and dendrites [26]. While wild-type human GARS rescued arborization defects, L129P GARS failed to rescue the phenotype and E71G GARS led to only a partial rescue. From these data, it was concluded that E71G and L129P GARS represent loss-of-function alleles [26]. Thus, each ARS mutation evaluated in functional studies has a loss-of-function effect. Coupled with the observations that the 20 ARS variants (Table 2) occur at highly conserved amino-acid residues and reside in five genes that encode distinct cytoplasmic ARS enzymes [8,15,16,20,21], currently available data suggest that impaired tRNA charging in neurons is a component of ARS-related neuropathy.

How Does a Mutated Aminoacyl-tRNA Synthetase Enzyme Cause an Axonal Pathology?

While impaired tRNA charging likely plays a role in ARS-related axonal peripheral neuropathy, the direct causal link between ARS function and axonal pathology remains unclear. Potential mechanisms range from general protein translation deficiency to undiscovered toxic gain-of-function properties (Figure 1). Reduced tRNA charging may lead to impaired protein translation in the soma of the cell (Figure 1, Point 1a). For proteins transported to the axon, reduced translation in the soma could result in decreased levels of necessary proteins in the axon. Similarly, reduced ARS function in the axon (Figure 1, Point 1b) could lead to impaired local axonal translation. The localization of ARS enzymes to axons in vitro and in vivo supports this notion [15,24]. Indeed, ARS mutations that decrease localization to axons (Figure 1, Point 2) may also result in impaired protein translation within the axon. Defects in mitochondrial protein translation are unlikely to be the sole cause of axonal peripheral neuropathy since mutations have been identified in cytoplasmic ARSs (AARS, YARS, HARS) that have separate mitochondrial counterparts (AARS2, YARS2, HARS2). However, a role for mitochondrial dysfunction cannot be ruled out for disease-associated mutations in bi-functional ARSs (GARS and KARS).

Figure 1.

Potential mechanisms of neuropathy-associated ARS mutations. A neuron is illustrated with the cell body on the left and an axon extending to the right. Based on currently available data, the most likely mode of pathogenesis is impaired tRNA charging in the cytoplasm (1a) or axon (1b) and/or impaired localization of the ARS enzyme to the axon (2), where it may be required for local protein translation. Either of these would be detrimental to the synthesis of key axonal proteins. Loss-of-function effects might also be exacerbated by a dominant-negative effect (3), where the mutant ARS protomer interferes with the function of the wild-type protomer in the context of a heterozygous genotype. One potential side effect of impaired tRNA charging would be an excess of uncharged tRNA molecules in the neuron (4), which would also affect protein translation, albeit in a more general fashion. However, ARS enzymes that are impaired in function or localization may form inappropriate interactions with neuronal (or more specifically, axonal) protein and RNA molecules (5), leading to downstream neurotoxicity. See text for a discussion of each non-mutually exclusive pathogenic mechanism.

The loss-of-function nature of ARS mutations and the dominant CMT phenotypes associated with the majority of mutations raise two possible pathogenic mechanisms: haploinsufficiency or a dominant-negative effect. Haploinsufficiency was addressed using a mouse model of Gars mutations. Mice with a missense mutation (P234KY) in the Gars gene show a severe, lethal, early-onset axonal neuropathy [27]. Mice heterozygous for a gene trap at Gars showed ~50% reduced enzyme activity compared to wild-type mice. Nevertheless, these mice did not have a peripheral neuropathy phenotype. It was concluded that haploinsufficiency was not the cause of the P234KY-mediated axonal neuropathy [27]. Additionally, null alleles have not been identified in patients with autosomal dominant neuropathy, which may also indicate that haploinsufficiency is not the primary cause of ARS-related neuropathy.

A dominant-negative effect (Figure 1, Point 3) is a specific sub-class of loss-of-function where the mutated protein interferes with the function of the wild-type protein in the context of a heterozygous individual [28]. This effect can lead to protein activity levels below 50% and phenotypes more severe than in cases of haploinsufficiency [28,29]. Prerequisites for a dominant-negative effect include: (i) the protein functions as an oligomer; (ii) the mutated allele produces a protein that is expressed and stable; (iii) the mutated protomer has reduced function; (iv) the mutation does not ablate formation of the mutant:wild-type oligomer; and (v) the mutant:wild-type oligomer has reduced function. Indeed, data supporting a dominant-negative effect for ARS mutations are as follows: (1) GARS, YARS, AARS, KARS, and HARS enzymes function as oligomers; (2) All but one disease-associated ARS mutation cause amino-acid changes that are not predicted to affect protein expression; (3) All ARS mutations have loss-of-function characteristics; and (4) certain YARS mutations cause a dominant-negative effect in yeast. Furthermore, compound heterozygosity for one null and one hypomorphic KARS allele causes CMTRIB [20]—the null and hypomorphic KARS alleles may mimic a dominant-negative effect through reducing charging activity below 50%. These observations suggest that restoring ARS function may alleviate the human phenotype. Recent analysis of two mouse strains harboring Gars missense mutations showed that over-expression of human GARS is unable to rescue the neuropathy phenotype [30], suggesting that these mutations convey toxicity independent of wild-type GARS function. However, it is important to consider that neither mouse mutation represents a human disease-associated GARS mutation and that these data may not be relevant for other human ARS genes implicated in peripheral neuropathy (e.g., YARS and AARS). Thus, a dominant-negative effect should still be considered for human neuropathy-associated ARS mutations.

While impaired neuronal protein synthesis appears to be the most likely mechanism of ARS-related neuropathy, there are some interesting alternatives. First, certain ARS enzymes have acquired secondary functions unrelated to aminoacylation [31]. The loss-of-function characteristics of ARS mutations may lead to impaired secondary functions related to neuron or axon function; however, no ARS enzyme has a reported secondary function in neurons. Second, impaired ARS function may lead to toxicity via pathways mediated by uncharged tRNAs (Figure 1, Point 4). Under conditions of amino acid starvation, uncharged tRNAs bind to general control non-repressible 2 kinase (GCN2), which phosphorylates eukaryotic initiation factor 2α (eIF2α) [32]. This leads to a general suppression of protein translation initiation. Thus, impaired charging of specific tRNA molecules could lead to a more global deficit in protein translation—this may explain the ability of mutations in distinct ARS enzymes to cause a similar neuropathic phenotype. Third, ARS mutations may interfere with the fidelity of the enzyme, resulting in mischarged tRNA and incorporation of the wrong amino acid into proteins. This possibility was explored by comparing mice with a neuropathy-associated Gars mutation to a mouse model of ataxia caused by an alanyl-tRNA synthetase (Aars) mutation that impairs the fidelity of the enzyme and allows incorporation of serine in the place of alanine [27,33,34]. The lack of any phenotypic overlap between these two mouse models suggests that mischarging is unlikely to be the mechanism of ARS-related neuropathy. Finally, ARS mutations may cause a toxic gain-of-function unrelated to tRNA charging. For example, the loss-of-function effect may be a prerequisite for inappropriate interactions with neuronal proteins or RNAs (Figure 1, Point 5). While in vitro and in vivo studies have not revealed aggregation or generalized protein misfolding associated with mutant ARS enzymes [15,24,34], more subtle interactions that are toxic to neurons or axons cannot be ruled out.

In summary, it is likely that impaired tRNA charging plays a central role in ARS-mediated peripheral neuropathy. The implication of multiple ARS enzymes in disease onset and the characterization of multiple loss-of-function alleles at each locus strongly support this notion. Importantly, the dominant nature of the human phenotypes may be explained by a dominant-negative effect, suggesting that correcting wild-type ARS function should be explored as a therapeutic modality.

Beating a Dead Enzyme: Future Directions for Characterizing ARS-Related Axonopathy

Despite substantial advances in the field of ARS-related peripheral neuropathy, the molecular pathology of these diseases remains elusive and therapies are currently unavailable. The following studies should help to address these two issues. First, current data suggest that all ARS genes are candidates for peripheral neuropathies. Additional mutation screening in relevant patient populations will be critical for identifying novel disease-associated ARS loci and alleles—as more ARS loci and alleles with loss-of-function characteristics are implicated in neuropathy, a role for impaired tRNA charging in disease pathogenesis will become more likely. Second, it is important to complete the functional analysis of all known neuropathy-associated ARS mutations (Table 2). Combined, these two approaches will help link or unlink loss-of-function characteristics with disease onset. Third, to explore the possibility of a dominant-negative effect, it should be determined if interactions between wild-type and mutant proteins are required for ARS-mediated neurotoxicity. Furthermore, the effect of ARS mutations on enzyme function and localization needs to be studied in the context of a wild-type:mutant oligomer.

Finally, restoration of wild-type ARS function in the peripheral nerve may be a reasonable therapeutic approach for patients with ARS-related neuropathy. The first critical step in assessing this is to determine if wild-type ARS can reverse the neurotoxicity of human ARS mutations in vivo. Subsequently, strategies can be developed for increasing wild-type ARS activity or expression. In summary, the studies proposed above will provide critical knowledge on the molecular pathology of neuropathy-related ARS mutations, a platform for developing therapies for these diseases, and information on neuron- or axon-specific protein translation needs, contributing more broadly to the fields of neurobiology and neurodegenerative disease.