Dispersed disease-causing neomorphic mutations on a single protein promote the same localized conformational opening

Abstract

The question of how dispersed mutations in one protein engender the same gain-of-function phenotype is of great interest. Here we focus on mutations in glycyl-tRNA synthetase (GlyRS) that cause an axonal form of Charcot–Marie–Tooth (CMT) diseases, the most common hereditary peripheral neuropathies. Because the disease phenotype is dominant, and not correlated with defects in the role of GlyRS in protein synthesis, the mutant proteins are considered to be neomorphs that gain new functions from altered protein structure. Given that previous crystal structures showed little conformational difference between dimeric wild-type and CMT-causing mutant GlyRSs, the mutant proteins were investigated in solution by hydrogen-deuterium exchange (monitored by mass spectrometry) and small-angle X-ray scattering to uncover structural changes that could be suppressed by crystal packing interactions. Significantly, each of five spatially dispersed mutations induced the same conformational opening of a consensus area that is mostly buried in the wild-type protein. The identified neomorphic surface is thus a candidate for making CMT-associated pathological interactions, and a target for disease correction. Additional result showed that a helix-turn-helix WHEP domain that was appended to GlyRS in metazoans can regulate the neomorphic structural change, and that the gain of function of the CMT mutants might be due to the loss of function of the WHEP domain as a regulator. Overall, the results demonstrate how spatially dispersed and seemingly unrelated mutations can perpetrate the same localized effect on a protein.

Charcot–Marie–Tooth (CMT) disease, also known as hereditary motor and sensory neuropathy, was named after three physicians who first described the disease in 1886. The disease is characterized by loss of muscle tissue and touch sensation in body extremities, predominantly in the feet and legs but also in the hands and arms (1). Presently incurable, this disease is one of the most common inherited neurological disorders affecting one in 2,500 people (2). Genetically, CMT disease is a heterogeneous group of disorders (3). Among the forty or so genes identified so far whose mutations are linked to the similar clinical presentations of CMT, four are genes encoding aminoacyl-tRNA synthetase, namely glycyl-, tyrosyl-, alanyl- and lysyl-tRNA synthetases (4–7). In most cases, the disease-causing mutation is dominant, and thus implicates a neomorphic (gain of gene function that is different from the normal function) form(s) that engenders the neuropathology.

Aminoacyl-tRNA synthetases are a family of essential enzymes in translation (8). Each member is responsible for charging one specific amino acid onto its cognate tRNAs. The charged tRNAs then use the embedded 3-nucleotide anticodons to decode mRNA and provide the corresponding amino acid building blocks for protein synthesis on the ribosome. Glycyl-tRNA synthetase (GlyRS) was the first tRNA synthetase implicated in CMT (4). So far, eleven different missense mutations of GARS have been reported to cause a dominant axonal form of CMT (CMT type 2D) in patients (4, 9–11). Two separate spontaneous or ENU-induced missense mutations have also been linked to CMT-like phenotypes in mice (12, 13). Interestingly, not all mutations affect the aminoacylation activity of the tRNA synthetase (14, 15). Furthermore, studies in the mouse clearly demonstrated that the CMT-like phenotype was not caused by haploinsufficiency in protein synthesis, but rather by a pathogenic role of the mutant GlyRS itself (12, 16), which remains to be defined at the molecular level.

GlyRS is a class II tRNA synthetase, whose catalytic domain consists of a central antiparallel β sheet flanked with α helices, and three conserved sequence motifs (motifs 1–3) (17). Human GlyRS has three insertions that split the catalytic domain, a metazoan-specific helix-turn-helix WHEP domain, and an anticodon binding domain at the N- and C-terminal side of the catalytic domain, respectively (Fig. 1A). Like most class II tRNA synthetases, GlyRS functions as a dimer for aminoacylation. Interestingly, despite being well-separated in the primary sequence of the three domains of GlyRS, all known CMT-causing mutations are located near the dimer interface of our crystal structure (14). This observation suggested a connection of the dimer interface with the disease-causing mechanism. However, different CMT-causing mutations have different effects on dimer formation: some disrupt, some strengthen and some seem to have no effect on the dimer (14). In addition, crystal structures of two CMT-causing mutant proteins showed little difference from that of the WT protein (17, 18), and suggested that structural differences, if any, between mutant and WT GlyRSs were subtle and could be suppressed by crystal packing forces.

Fig. 1.

Distribution of CMT-causing mutations on GlyRS. (A) All 13 CMT-causing mutations mapped onto the domain structure of GlyRS. The three sequence motifs that are characteristic of the catalytic domain of class II tRNA synthetases are noted as 1, 2, 3, and the three insertions to the catalytic domain as I, II, III. (B) Dimer interface location of two newly identified CMT-associated residues C157 and P244. (C) Close-up view of the location of C157 and P244.

With these considerations, we sought to explore five different CMT-causing mutant proteins in solution by use of hydrogen-deuterium exchange analysis monitored by mass spectrometry (HDX MS). HDX MS determines the solvent exposure of individual peptide segments throughout a protein. By annotating the exposure of the same peptide segments in the WT and mutant proteins, a map of structural differences can be constructed. In addition, we used small-angle X-ray scattering (SAXS) to study protein shape changes in solution. With these approaches, we avoided the problem of crystal packing interactions suppressing conformational differences that existed in solution. Surprisingly, we found that each of the mutations we tested induced a structural opening that was mapped to specific but common regions of the protein.

Results

Dimer Interface Localization of Two Newly Identified CMT-Associated Residues.

Of the 13 CMT-linked mutations on GlyRS (from patients and mice) identified so far (Fig. 1A), two—C201R from ENU-induced mice (corresponding to C157R in the human sequence) and P244L from a Japanese patient—were published recently (10, 13). Before carrying out the conformational studies in solution, we determined that the two newly identified CMT-associated residues P244 and C157, like the other 11, were localized near the dimer interface (Fig. 1B). P244 is located on a hairpin structure (β8–β9) that forms an antiparallel β-sheet across the dimer interface, and that harbors two other CMT-associated residues G240 and P234 (Fig. 1C). We previously reported that P234 and I280—another CMT-linked residue—“kiss” across the dimer interface (14). Remarkably, C157 is located adjacent to I280, and directly across from P234 of the other subunit. The three CMT-linked residues form a triad with a distance of approximately 4 Å between each other (Fig. 1C). Thus, 13 CMT-linked residues are all localized near the dimer interface.

Idiosyncratic Effect of CMT Mutations on Dimer-Monomer Equilibrium.

GlyRS functions as a dimer for aminoacylation. Amino acid substitutions that are located on the dimer interface are likely to affect the dimer-monomer equilibrium. Indeed, using coimmunoprecipitation to detect heterodimer formation in vivo between transfected mutant GlyRS and endogenous WT GlyRS, we previously showed that different mutations have different effects on dimer formation (14). For example, L129P and G240R have a greatly reduced capacity to form heterodimers with WT GlyRS. In contrast, S581L and G526R have an increased capacity for heterodimer formation.

In terms of homodimer formation, the opposite effects of the G240R and G526R mutations were also validated by analytical ultracentrifugation (AUC) analysis. Two complementary views of solution behavior are accessible from AUC analyses. Sedimentation velocity (SV) provides hydrodynamic information about size and shape, whereas sedimentation equilibrium (SE) provides thermodynamic information about the molecule mass, stoichiometry of subunit assembly, and the association constant (19). SV analysis in our previous work showed that a larger population of monomer species was found with G240R relative to WT GlyRS, whereas for G526R, it was the opposite (14, 17).

To obtain quantitative understanding of the effect of CMT-causing mutations on the monomer-dimer equilibrium, we further measured the dissociation constants of WT and mutant GlyRSs by the SE method (Table 1 and Fig. S1). Under our experimental conditions, WT GlyRS had an apparent K_d of 0.56 μM. Consistent with previous observations, L129P (K_d = 41 μM) and G240R (K_d = 11 μM) mutations cause 73- and 20-fold increases in K_d value, whereas S581L (K_d = 0.19 μM) and G526R (K_d = 0.15 μM) decreased the K_d about three- to fourfold. We also determined the K_d of G598A GlyRS, a mutant associated with the most severe phenotype among all CMT 2D patients (9). Interestingly, the G598A mutation had little effect on the monomer-dimer equilibrium (K_d = 0.39 μM). For reasons we do not fully understand, estimates of K_d for dimerization based on size exclusion chromatograph were at least fivefold lower than the numbers calculated from the AUC analysis. Nevertheless, the relative K_d between a mutant and WT GlyRS remain essentially unchanged with both methods. These results indicate that the effect of a CMT-causing mutation on dimer stability is idiosyncratic.

Table 1.

The relative dissociation constants of dimerization for WT and CMT-causing mutant GlyRSs, and the average increase in deuterium incorporation for each mutant relative to WT GlyRS

	WT	L129P	G240R	G598A	S581L	G526R	ΔWHEP
Relative Kd of dimerization	1	73.21	19.64	0.70	0.34	0.27	n.d.
HDX increase*	0%	37%	30%	22%	18%	16%	26%

^*The number is calculated from the deuterium incorporation difference after 1 h exchange for each proteolyzed peptide in a protein and then averaged for all of the peptides.

HDX Analyses Reveal Increased Deuterium Incorporation for All CMT-Associated Mutants.

In spite of their idiosyncratic effect on dimer stability, we wanted to understand if there is a shared conformational change induced by different mutations, given their common dimer interface localization. HDX MS is a powerful analytical tool to study protein dynamics and conformational changes in solution (20, 21). Backbone amide hydrogens in a protein are labile and will exchange with deuteriums when the protein is placed in a D₂O buffer. After each of a series of exchange periods, the exchange reaction is quenched by reducing the pH to approximately 2.5, and the protein is proteolyzed into fragments and subjected to high-resolution mass spectrometry analysis (22). Because exchange of one hydrogen atom for deuterium results in one mass unit increase, deuterium incorporation of each peptide can be monitored by mass spectrometry. The rate and level of exchange are directly related to the local conformation of the peptide: Solvent- accessible peptides exhibit faster and greater deuterium exchange than solvent-inaccessible ones.

We first analyzed the WT protein. According to our K_d measurement for the monomer-dimer equilibrium, GlyRS was predominantly dimeric at the concentration for the HDX experiment (2.6 μM). Consistently, few peptides from the dimerization interface were associated with high deuterium uptake. In contrast, almost all peptides within the disordered regions in the crystal structure of WT GlyRS [e.g., the WHEP domain, Insertion III and the C-terminus (17)] were associated with high deuterium uptake (Fig. S2).

Subsequently, we analyzed the five mutant GlyRSs (L129P, G240R, G526R, S581L, and G598A) that were distinctly different in terms of their dimer stability (Table 1). Remarkably, all five mutations increased the overall deuterium incorporation of the protein (16%–37% over that of WT GlyRS after 1 h exchange), suggesting a conformational opening induced by each mutation. Interestingly, the level of the conformational opening correlates with the dimer stability: The less stable the mutant dimer, the higher the increase in deuterium incorporation (Table 1). However, even mutations that promote dimer association, such as G526R and S581L, increased deuterium uptake, suggesting that the CMT mutation-induced conformational opening is, in part, independent of dimer stability.

Mutations that Either Weaken or Promote Dimer Association Open Up Large Regions.

To identify the specific areas that are opened up by the mutations, we mapped the HDX MS results on the primary sequence and on the crystal structure of GlyRS (Fig. 2 and Fig. 3). For dimer-weakening L129P and G240R mutants, large regions on GlyRS exhibited increased deuterium uptake compared to the WT protein and, not surprisingly, those areas include the dimerization interface (Fig. 3A). [The dimer interaction is mostly provided by three patches: F78-T137 (which includes the entire motif 1), F224-L242, and L252-E291 of the catalytic domain and is, to a lesser degree, contributed by S581-R606 of the anticodon binding domain (17)]. Peptides from these three patches exhibit some of the most pronounced increases in deuterium uptake (e.g., peptides F79-A83, M227-L257, and L258-R288 for L129P, and F79-A83, I232-N253 and L258-E279 for G240R GlyRS). In addition, other areas that are outside of the dimer interface also show large increases in deuterium uptake (e.g., peptides F147-K150 and E515-M531 for both L129P and G240R GlyRS) (Fig. S3 A and B). Apparently, L129P and G240R mutations induced conformational openings beyond those that dissociate the dimer.

Fig. 2.

Changes in deuterium incorporation resulting from CMT-causing mutations or deletion of the WHEP domain. The results are mapped on the primary sequence of GlyRS, with CMT mutation-associated residues highlighted in purple. The percent difference of deuterium incorporation is calculated from the hydrogen-deuterium exchange after 1 h for each mutant relative to WT GlyRS. The uncovered areas for each mutant may result from either lack of sequence coverage for either the mutant or WT GlyRS, or lack of common peptides for direct comparison. (The sequence coverage for WT, L129P, G240R, G526R, S581L, G598A, and ΔWHEP GlyRS themselves were 96%, 87%, 95%, 89%, 98%, 99%, and 96%, respectively.) The eight consensus opened-up areas (hot spots) are labeled. In general, the hot spots are well-covered in at least four of the five CMT mutants we tested, and each mutant peptide within the hot spots has greater than 5% increase in HDX relative to the WT protein. The hot spot areas might slightly increase, if the sequence coverage improves.

Fig. 3.

Changes in deuterium incorporation mapped onto the crystal structure of GlyRS. (A) Mapping of changes in deuterium incorporation caused by different CMT mutations or deletion of the WHEP domain. The monomeric structure is oriented to view the dimerization interface. The color coding is the same as in Fig. 2. (B) Mapping of the consensus areas (or hot spots) that are opened up by all 5 tested CMT-causing mutations. The hot spots are colored in gold and CMT mutation-associated residues in purple. Depicted on the left is a dimeric view of the GlyRS structure.

For G526R and S581L mutations that promote dimer association, each mutant protein has one peptide that is less solvent-exposed than its counterpart in the WT protein (Fig. 2). Peptide I108-E123 (part of motif 1) in G526R and peptide R635-I645 in S581L GlyRS had lower deuterium incorporation and may be responsible for strengthening the dimer interactions of G526R and S581L GlyRS. In spite of these regions of reduced hydrogen-deuterium exchange, overall, both mutants are more solvent-exposed than is the WT protein (Fig. 2 and Fig. S3 C and D). As in the case for L129P and G240R GlyRS, this enhanced exposure includes most of, but is not limited to, the dimerization interface (Fig. 3A). This pattern is also true for G598A GlyRS (Figs. 2 and 3A and Fig. S3E).

Opened-Up Areas in Each Mutant Largely Overlap and Cover Other CMT Mutation Sites.

Most interestingly, the opened-up areas in each mutant protein largely overlap (Fig. 2). Eight consensus opened-up areas can be identified as “hot spots.” To qualify as a hot spot, in general, the area has to be well-covered in all of the mutants we tested, and each mutant peptide within the hot spot must exhibit greater than 5% increase in HDX relative to the WT protein. Those hot spots include peptide A57-A83 adjacent to the WHEP domain and that partially overlaps with the dimer interface (hot spot 1), L129-D161 that bridges motif 1 to Insertion I (hot spot 2), N208-Y320 that covers a large area of the dimer interface and motif 2 (hot spot 3), V366-H378 that is outside the dimer interface (hot spot 4), P518-M531 that is part of motif 3 (hot spot 5), L584-Y604 that is the site nearest to the other subunit in the anticodon binding domain (hot spot 6), and F620-R635 and D654-A663 that are near the C-terminus and distal from the dimer interface (hot spots 7 and 8). With the exception of motif 1, these hot spots are distributed over the entire dimer interface area, and also cover some areas outside the dimer interface.

Two types of structural opening effects can be identified from these mutations. First is a local conformational change near the site of the mutation. Each CMT mutant protein that we tested has increased solvent accessibility in the vicinity of the substitution site (Fig. 2). Second is a distal effect. Interestingly, each mutation also increased the solvent accessibility near the other 12 known CMT-associated sites (Table 2). (Except for H418, all CMT-associated residues are covered by the HDX MS analysis of the WT and at least one mutant GlyRS.) That all CMT-mutation-site-containing peptides in all mutants exhibit increased deuterium uptake suggests that any single CMT-causing mutation can induce a conformational opening that exposes all CMT-associated sites. Consistently, except for H418, D500, and S581, which were poorly covered in the HDX MS analysis, the other ten CMT-associated residues are all within the hot spots of the opened-up areas (Figs. 2 and 3B).

Table 2.

Increase in deuterium incorporation (%) after 1 h exchange at CMT-associated sites of each mutant relative to WT GlyRS

	A57	E71	L129	C157	P234	G240	P244	I280	H418	D500	G526	S581	G598
L129P	7	87	22	25	65	65	65	51	n.c.*	n.c.*	62	n.c.*	17
G240R	6	70	n.c.*	6	58	58	58	56	n.c.*	51	67	n.c.*	26
G526R	20	42	17	n.c.	8	8	8	28	n.c.*	38	26	n.c.*	14
S581L	12	27	27	33	38	38	38	37	n.c.*	n.c.*	38	16	16
G598A	20	30	26	14	56	56	56	37	n.c.*	n.c.*	43	n.c.*	30
ΔWHEP	n.c.*	56	20	n.c.*	55	55	55	37	n.c.*	n.c.*	35	n.c.*	20

^*The residue is uncovered by HDX MS analysis for the mutant.

SAXS Confirms Structure Opening of G526R GlyRS in Solution.

Among the five mutants we analyzed by AUC and HDX MS, particularly interesting are G526R and S581L, which facilitate dimerization and also have overall increased solvent accessibility, including most of the dimerization interface. We imagine that those mutants open up the structure without dissociating the dimer and, as a result, the overall structure should be expanded. To test that hypothesis, and to gain additional insight into the conformational change, we compared one mutant G526R, as an example, with WT GlyRS, by use of SAXS analysis (23).

Scattering curves were measured for both G526R and WT GlyRS (Fig. 4A). The pair distribution function p(r) obtained by an indirect Fourier transformation of the scattering curve showed that G526R GlyRS in solution is an elongated particle with a maximum dimension (D_max) of 185 Å, compared to D_max = 160 Å for WT GlyRS (Fig. 4B). The radius of gyration (R_g) evaluated by Guinier plots of the scattering curve yields R_g = 44 Å (± 0.4) for G526R, and R_g = 41 Å (± 0.9) for WT GlyRS. Both parameters (R_g and D_max) indicate that the mutant is larger in size than the WT protein, consistent with the HDX MS analysis and our hypothesis.

Fig. 4.

SAXS analysis confirms the structure opening of G526R GlyRS and reveals the conformational change of the WHEP domain. (A) Solution scattering data of WT and G526R GlyRS. Small-angle X-ray scattering curves are overlapped with theoretical scattering profiles calculated from ab initio models (black line). The inset shows Guinier plots at the low-angle region (S ∗ R_g < 1.3). (B) Distance distribution P(R) functions of WT and G526R GlyRS. P(R) curves were calculated from SAXS data shown in (A). The main differences between WT and G526R GlyRS are indicated by arrows. (C) SAXS-based ab initio modeling of WT (purple) and G526R (yellow) GlyRS. The crystal structure of the dimeric WT GlyRS was manually docked into the SAXS-based molecular envelope, leaving two extra densities located on each side of the dimer near the N-terminus of the GlyRS structure. Therefore, the extra densities most likely correspond to the disordered WHEP domain in the crystal structure. The extra densities are fit with a model of the WHEP domain from HisRS (PDB∶1X59) and differ between WT and G526R GlyRS, suggesting a conformational change of the WHEP domain induced by the G526R mutation.

We fitted the crystal structure of GlyRS onto the ab initio molecular envelopes of G526R and WT GlyRS generated from their scattering curves (Fig. 4C). Extra unfitted density of each envelope presumably corresponds to the WHEP domain and/or Insertion III that were disordered in the crystal structure (17). Because the extra density is located near the N-terminal residues in the crystal structure, it is most likely contributed from the WHEP domain. Consistently, a model of the WHEP domain with a helix-turn-helix structure fits well into the extra density (Fig. 4C). Interestingly, the density of the molecular envelope of G526R and WT GlyRSs differs at the location of the WHEP domain suggesting a conformational change of the WHEP domain induced by the G526R mutation.

Structure Opening Regulated by the WHEP Domain.

The metazoan-specific WHEP domain is dispensable for aminoacylation (17). Because the CMT phenotype is not correlated with the aminoacylation activity, and because the conformation of the WHEP domain appears primarily affected by one of the CMT mutations G526R (Fig. 4C), it is possible that the WHEP domain plays a role in a CMT-related mechanism and is involved in the structural opening induced by CMT mutations. Thus, we decided to investigate the conformational consequence, if any, of removing the WHEP domain.

Remarkably, HDX MS analysis showed that the deletion of WHEP domain largely opens up GlyRS with average increase in deuterium uptake after 1 h exchange of 26% (Table 1 and Fig. S3F). This level of increase is comparable with that of the CMT-causing mutants we tested. Furthermore, the deletion mutant shares similar areas of structural opening, which include the dimerization interface (Figs. 2 and 3A) and cover all CMT-associated mutation sites (Table 2). The similarity of the HDX MS results for WHEP-deleted GlyRS and for CMT-causing mutants suggests a role for the WHEP domain in suppressing a neomorphic conformational change of GlyRS that may lead to CMT.

Discussion

Several lines of evidence demonstrate that the GlyRS mutation-associated CMT phenotype is not simply caused by a deficiency in aminoacylation (14, 16). Those results raised the possibility that CMT-causing mutations disrupt an unknown, peripheral neuron-specific function of GlyRS. In that connection, tRNA synthetases are known to be multifunctional proteins (24, 25). On the other hand, particularly because of the dominant nature of the GlyRS mutation-associated CMT phenotype, the disease may be linked to a gain-of-function pathogenic role only associated with the mutant GlyRS proteins. In that sense, the CMT-causing mutations in GlyRS resemble mutations in SOD1 that cause a severe motor neuron degenerative disease [amyotrophic lateral sclerosis (ALS)]. Mutant SOD1 proteins have gained toxic properties that promote aggregations in motor neurons, a common pathological feature for both familial and sporadic types of ALS (26, 27). However, overexpression of mutant GlyRS proteins in motor neurons did not promote aggregations (15), and no aggregation of GlyRS or other misfolded proteins were found in CMT- mice expressing a mutant GlyRS (16). Those observations are consistent with the possibility that the gain of function of GlyRS CMT mutant proteins involves specific interaction(s) that lead to pathological consequences.

Also, in contrast to the well-spread ALS-causing mutations throughout the SOD1 structure, we previously observed and herein confirmed that all CMT-causing mutations in GlyRS are located near the dimerization interface. Furthermore, HDX MS analysis elucidated a common area opened up by all tested mutations. These mutations have different effects on aminoacylation activity and on dimer stability, but the same neomorphic conformational opening. This opening may be a unifying feature of all CMT-causing mutations. The opened-up areas partially overlap with the dimerization interface and provide unique surfaces for potential neomorphic interactions specific to the CMT mutants (Fig. 5). Interestingly, not only are the same neomorphic surfaces opened up by different CMT-causing mutations, but all CMT mutation-associated sites are also located within these surfaces. Thus, the mutations may strengthen a neomorphic interaction that contacts the opened-up areas, which would be potential drug targets for treating GlyRS mutation-linked CMT disease.

Fig. 5.

Illustration of the same conformational opening of GlyRS induced by different CMT-causing mutations and of the generation of a common neomorphic surface.

Considering that the physiological concentration of tRNA synthetases is close to our measured K_d for the dimer-monomer equilibrium for GlyRS (Table 1) (28, 29), this synthetase is likely to exist in both dimer and monomer forms in vivo. The monomer form would be especially predominant for L129P and G240R GlyRS and likely to be the form used for mediating neomorphic interaction(s) for those mutants. For other CMT-causing mutations that do not necessarily promote monomer formation, because the opened-up surface overlaps with most of the dimerization interface (Fig. 3), the monomer forms of those mutants may also mediate neomorphic interaction(s) (Fig. 5).

Interestingly, deletion of the WHEP domain from GlyRS induces a conformational change resembling that of the CMT-causing mutations (Figs. 2 and 3). WHEP domains are also found in other human tRNA synthetases and are critically associated with the mechanisms used to expand and regulate a broad functionome of tRNA synthetases (24, 25). For example, the WHEP domain of TrpRS regulates an angiostatic activity embedded in the synthetase (30). Removal of the WHEP domain exposes a binding site for the VE-cadherin receptor that mediates the angiostatic effect of extracellular TrpRS. Possibly, the WHEP domain in GlyRS has a regulatory role like that in TrpRS, and the neomorphic structural opening is associated with an unknown physiological function of GlyRS that is suppressed by the WHEP domain. In this scenario, a CMT-causing mutation may simply disrupt the WHEP domain suppression and thereby give a gain-of-function phenotype. Moreover, the structural opening effect of the CMT mutations could be mediated through changing the WHEP domain conformation that, interestingly, is evident in our SAXS analysis with G526R GlyRS (Fig. 4C). Therefore, whether the neomorphic structure opening is associated with a physiological or a pathological function, the gain of function of the CMT-causing mutants could be due to the loss of function of the WHEP domain as a regulator.

Formation of the same neomorphic structural variant as a result of different CMT-causing mutations suggests that induction of a conformational change of GlyRS can be readily achieved and that the conformation is relatively stable. It is reasonable to speculate that the “wild-type” structure of GlyRS is on a “tipping point” and is separated by a relatively small energy barrier from the neomorphic conformation. Perhaps for this reason and more broadly speaking, this study has demonstrated that spatially dispersed and seemingly unrelated mutations can perpetrate the same localized effect on a protein.

Methods

Sedimentation equilibrium experiments were performed in a Beckman Optima XL-I analytical ultracentrifuge with integrated optical systems and a An-60 Ti rotor. HDX Fourier transform ion cyclotron resonance mass spectrometry experiments were performed as described (31). SAXS data were collected at the 12.3.1 beamline of the Advanced Light Source (Berkeley). Detailed protocols of these experiments and data analyses are described in SI Text.