https://orcid.org/0000-0002-4784-7423
Abstract
The translocase of outer mitochondrial membrane (TOMM) complex is the entry gate for virtually all mitochondrial proteins and is essential to build the mitochondrial proteome. TOMM70 is a receptor that assists mainly in mitochondrial protein import. Here, we report two individuals with de novo variants in the C-terminal region of TOMM70. While both individuals exhibited shared symptoms including hypotonia, hyper-reflexia, ataxia, dystonia and significant white matter abnormalities, there were differences between the two individuals, most prominently the age of symptom onset. Both individuals were undiagnosed despite extensive genetics workups. Individual 1 was found to have a p.Thr607Ile variant while Individual 2 was found to have a p.Ile554Phe variant in TOMM70. To functionally assess both TOMM70 variants, we replaced the Drosophila Tom70 coding region with a Kozak-mini-GAL4 transgene using CRISPR-Cas9. Homozygous mutant animals die as pupae, but lethality is rescued by the mini-GAL4-driven expression of human UAS-TOMM70 cDNA. Both modeled variants lead to significantly less rescue indicating that they are loss-of-function alleles. Similarly, RNAi-mediated knockdown of Tom70 in the developing eye causes roughening and synaptic transmission defect, common findings in neurodegenerative and mitochondrial disorders. These phenotypes were rescued by the reference, but not the variants, of TOMM70. Altogether, our data indicate that de novo loss-of-function variants in TOMM70 result in variable white matter disease and neurological phenotypes in affected individuals.
Introduction
Mitochondria are vital organelles required for energy production, metabolic regulation, signaling and apoptosis (1–4). To perform the majority of these functions, mitochondria rely upon resident proteins located in the outer membrane, inter-membranous space, inner membrane and matrix. In humans, the mitochondrial proteome consists of ~1500 proteins (3,5,6). While the mitochondrial genome encodes only 13 proteins (6), the remaining mitochondrial proteins (~99%) are encoded by the nuclear genome. These proteins typically contain a specific mitochondrial targeting sequence (MTS); ~60% of these proteins contain an N-terminal MTS (7,8), while the remaining 40% contain an internal MTS. Most of these mitochondrial proteins contain large stretches of hydrophobic amino acids, and transporter complexes are required for the import of these proteins into different mitochondrial compartments. Proteins that are translocated to the mitochondria are first translated as pre-protein in the cytoplasm. Cytoplasmic chaperones prevent their misfolding and guide their transport. The coordinated action of the translocase of inner mitochondrial membrane (TIMM) and translocase of outer mitochondrial membrane (TOMM) complexes is required for the import process (9).
The structure of the TOMM complex is highly conserved across species. In humans, the TOMM complex consists of two peripheral receptors—TOMM20 and TOMM70—a central receptor called TOMM22, a central pore-forming protein called TOMM40 and three small TOMM proteins called TOMM5, 6 and 7 (10,11). TOMM70 and TOMM20 help in the transport of the nascent proteins into the mitochondria. TOMM70 binds to the pre-protein-chaperone complex for proteins with an internal MTS (12,13) and facilitates the entry of the pre-proteins into the mitochondria via TOMM40 (14). In human, TOMM70 gene is broadly expressed in many organs including the central nervous system (CNS) (15). The human TOMM70 is 608 amino acid membrane protein integrated in the mitochondrial outer membrane via an N-terminal transmembrane region. A large portion of the protein faces cytoplasm. Structurally, it contains multiple tetratricopeptide (TPR) motifs (16) that facilitate protein–protein interaction and pre-protein recognition (17,18).
Defective mitochondrial protein import has been associated with a spectrum of disorders including cancer, neurodegeneration and metabolic disorders (5,13,19). Previously, a mutation in a TIMM complex member, TIMM8A (OMIM 300356), has been associated with Mohr–Tranebjaerg syndrome (OMIM 304700) (20). In the present study, we report clinical findings of two unrelated children both with de novo missense TOMM70 variants (Table 1; Supplementary Material, Table S1 for a detailed comparison). Here, we assess the nature of these variants in fruit flies. These functional assays show that both are partial loss-of-function variants.
Summary of clinical features of both probands with de novo variants in TOMM70
| Individual 1 | Individual 2 | |
|---|---|---|
| Variant | p.Thr607Ile | p.Ile554Phe |
| Gender | Female | Male |
| Age | 6 | 11 |
| Developmental delay | + | − |
| Dystonia | + | + |
| Hypotonia | + | + |
| Hyper-reflexia | + | + |
| Ankle clonus | + | − |
| Cogwheeling | − | + |
| Tremor | − | + |
| Dysarthria | − | + |
| Ptosis | − | + |
| White matter abnormality | White matter abnormalities with cerebellar atrophy | White matter abnormality with superimposed rarefaction |
| Cerebellar atrophy | + | − |
| Microcephaly | + | − |
| Ataxia | + | + |
| Individual 1 | Individual 2 | |
|---|---|---|
| Variant | p.Thr607Ile | p.Ile554Phe |
| Gender | Female | Male |
| Age | 6 | 11 |
| Developmental delay | + | − |
| Dystonia | + | + |
| Hypotonia | + | + |
| Hyper-reflexia | + | + |
| Ankle clonus | + | − |
| Cogwheeling | − | + |
| Tremor | − | + |
| Dysarthria | − | + |
| Ptosis | − | + |
| White matter abnormality | White matter abnormalities with cerebellar atrophy | White matter abnormality with superimposed rarefaction |
| Cerebellar atrophy | + | − |
| Microcephaly | + | − |
| Ataxia | + | + |
Summary of clinical features of both probands with de novo variants in TOMM70
| Individual 1 | Individual 2 | |
|---|---|---|
| Variant | p.Thr607Ile | p.Ile554Phe |
| Gender | Female | Male |
| Age | 6 | 11 |
| Developmental delay | + | − |
| Dystonia | + | + |
| Hypotonia | + | + |
| Hyper-reflexia | + | + |
| Ankle clonus | + | − |
| Cogwheeling | − | + |
| Tremor | − | + |
| Dysarthria | − | + |
| Ptosis | − | + |
| White matter abnormality | White matter abnormalities with cerebellar atrophy | White matter abnormality with superimposed rarefaction |
| Cerebellar atrophy | + | − |
| Microcephaly | + | − |
| Ataxia | + | + |
| Individual 1 | Individual 2 | |
|---|---|---|
| Variant | p.Thr607Ile | p.Ile554Phe |
| Gender | Female | Male |
| Age | 6 | 11 |
| Developmental delay | + | − |
| Dystonia | + | + |
| Hypotonia | + | + |
| Hyper-reflexia | + | + |
| Ankle clonus | + | − |
| Cogwheeling | − | + |
| Tremor | − | + |
| Dysarthria | − | + |
| Ptosis | − | + |
| White matter abnormality | White matter abnormalities with cerebellar atrophy | White matter abnormality with superimposed rarefaction |
| Cerebellar atrophy | + | − |
| Microcephaly | + | − |
| Ataxia | + | + |
Results
Clinical presentation of the affected individuals
Individual 1 is a 7-year-old female, whose symptoms include severe global developmental delay, mild acquired microcephaly (head circumference − 0.13 SD at birth, −2.48 SD at 6 years), hypotonia, mixed hyperkinetic movement disorder (choreoathetosis, dystonia and ataxia), exaggerated startle response and irritability. Brain MRI performed at the age of 9 months was unremarkable, but subsequent MRIs revealed hypomyelination, cerebellar atrophy and thinning of corpus callosum (Fig. 1A–D). Magnetic resonance spectroscopy was unremarkable.
Brain MRI images of Individuals 1 and 2. (A) Sagittal T1 image of from an MRI of Individual 1 obtained at 9 months of age was reported as normal; however, subsequent imaging (B–D) has revealed significant hypomyelination, diffuse thinning of the corpus callosum, mild cerebral atrophy and moderate cerebellar atrophy. Arrows in (A) and (B) indicate the corpus callosum and cerebellum of Individual 1. (E–G) MRI images from Individual 2 at 4 years and 8 months of age suggest a demyelinating process. There are no earlier scans. (E) On sagittal T1, there are areas of demyelination (arrows) noted in the genu and splenium of the corpus callosum. The body of the corpus callosum is thin; however, there are no other evident structural changes. (F) T2 hyperintensity and (G) FLAIR imaging revealing cystic rarefaction in the deep white matter with relative sparing of the periventricular white matter and the U-fibers. The posterior limb of the internal capsule is also highly affected.
Individual 2 is a 12-year-old male investigated as part of a tiered diagnostic study for individuals with abnormalities of the CNS white matter (21,22) and was identified via GeneMatcher (23). He presented with episodic regression starting at 4 years of age following normal initial development. Regression was primarily associated with alterations in gait, in the context of febrile illness and minor head trauma. He has significant gross motor impairment complicated by proximal weakness and spastic ataxia, as well as hypotonia, cogwheeling and truncal titubation. Fine motor control is remarkable for dysmetria. He also has dysarthria and ptosis, as well as increasing academic difficulties in recent years. Brain imaging revealed diffuse involvement of the brain with T2 hyperintensity and FLAIR imaging revealing cystic rarefaction in the deep white matter with relative sparing of the periventricular white matter (Fig. 1E–G). There were significant abnormalities in the posterior limb of the internal capsule and extending longitudinally through the brainstem and white matter of the spinal cord, as well as through the middle cerebellar peduncles and into the cerebellar white matter including the hilus of the dentate.
Both individuals had extensive metabolic testing that was largely unremarkable. Both had unremarkable serum lactate, plasma amino acids and urine organic acids. Both individuals previously also had extensive genetic testing that was unrevealing, including mitochondrial genome sequencing and deletion testing, nuclear mitochondrial disease gene testing and whole exome sequencing (see Supplementary Material, Table S1 for a detailed comparison).
Next generation sequencing identifies de novo variants in TOMM70
Next generation sequencing approaches in these two independent individuals in the context of different genetic studies identified unique de novo variants in TOMM70 for both individuals. Exome sequencing in Individual 1 identified a de novo missense variant [NM_014820.3:c.1820C > T (p.Thr607Ile)] and in Individual 2, trio genome sequencing identified a separate missense variant confirmed to arise de novo [NM_014820.3:c.1660A > T; (p.Ile554Phe)]. Both variants affect amino acids present in the cytoplasmic C-terminal region of human TOMM70 protein. The p.Ile554Phe variant is in the 10th TPR repeat (Fig. 2A). We used MARRVEL to explore the human and model organism databases (24). Variant pathogenicity prediction software, such as SIFT (25), CADD (26), PolyPhen2 (27) M-CAP (28), RAVEL (29), GERP (30) and MutationTester (31) strongly indicates that these variants are probably pathogenic (Table 2).
Human TOMM70 structure, effect of human transgene overexpression and humanization strategy in Drosophila. (A) Schematic of human TOMM70 protein displaying de novo mutations found in Individuals 1 and 2. TOMM70 is a 608 amino acid long protein consists of 10 tetratricopeptide motifs. Variant amino acids are indicated. Image is based on the Uniprot database using DOG 2.0 software. (B) Table showing the effect of TOMM70 reference and variant transgenes expression using different GAL4 lines in Drosophila. Expression of TOMM70 reference and variant transgene using a ubiquitous Actin-GAL4 line or eye-specific GAL4 lines does not cause any morphological phenotypes. (C) Expression of human TOMM70 reference and variant transgenes using GMR-GAL4 does not affect the stability of the proteins. Size of TOMM70 protein is ~70 kDa and size of actin protein is ~42 kDa. (D) Schematics of generation of Tom70miniGAL4∆ and UAS-TOMM70 transgenic animals. For simplicity, the dominant marker is not included in the Tom70miniGAL4∆ allele schematic.
De novo TOMM70 variants and associated pathogenicity
| Genomic position (hg19) | Variant (de novo) | Amino acid change | CADD score [Kircher et al. (26)] | SIFT [Sim et al. (25)] | REVEL [Ioannidis et al. (29)] | GERP score [Cooper et al. (30)] | PolyPhen2 [Adzhubei et al. (27)] | M-CAP [Jagadeesh et al. (28)] | Mutation tester [Schwarz et al. (31)] |
|---|---|---|---|---|---|---|---|---|---|
| 3:100084415 | NM_014820.3: c.1820C > T | p.T607I (NM_014820.4) | 28.4 (pathogenic) | 0.01 (deleterious) | 0.32 (likely pathogenic) | 5.82 (pathogenic) | 0.987 (probably damaging) | 0.048 (damaging) | 1 (disease causing) |
| 3:100086901 | NM_014820.3: c.1660A > T | p.I554F (NM_014820.4) | 26.3 (pathogenic) | 0 (deleterious) | 0.366 (pathogenic) | 5.88 (pathogenic) | 0.832 (probably damaging) | 0.053 (damaging) | 1 (disease causing) |
| Genomic position (hg19) | Variant (de novo) | Amino acid change | CADD score [Kircher et al. (26)] | SIFT [Sim et al. (25)] | REVEL [Ioannidis et al. (29)] | GERP score [Cooper et al. (30)] | PolyPhen2 [Adzhubei et al. (27)] | M-CAP [Jagadeesh et al. (28)] | Mutation tester [Schwarz et al. (31)] |
|---|---|---|---|---|---|---|---|---|---|
| 3:100084415 | NM_014820.3: c.1820C > T | p.T607I (NM_014820.4) | 28.4 (pathogenic) | 0.01 (deleterious) | 0.32 (likely pathogenic) | 5.82 (pathogenic) | 0.987 (probably damaging) | 0.048 (damaging) | 1 (disease causing) |
| 3:100086901 | NM_014820.3: c.1660A > T | p.I554F (NM_014820.4) | 26.3 (pathogenic) | 0 (deleterious) | 0.366 (pathogenic) | 5.88 (pathogenic) | 0.832 (probably damaging) | 0.053 (damaging) | 1 (disease causing) |
De novo TOMM70 variants and associated pathogenicity
| Genomic position (hg19) | Variant (de novo) | Amino acid change | CADD score [Kircher et al. (26)] | SIFT [Sim et al. (25)] | REVEL [Ioannidis et al. (29)] | GERP score [Cooper et al. (30)] | PolyPhen2 [Adzhubei et al. (27)] | M-CAP [Jagadeesh et al. (28)] | Mutation tester [Schwarz et al. (31)] |
|---|---|---|---|---|---|---|---|---|---|
| 3:100084415 | NM_014820.3: c.1820C > T | p.T607I (NM_014820.4) | 28.4 (pathogenic) | 0.01 (deleterious) | 0.32 (likely pathogenic) | 5.82 (pathogenic) | 0.987 (probably damaging) | 0.048 (damaging) | 1 (disease causing) |
| 3:100086901 | NM_014820.3: c.1660A > T | p.I554F (NM_014820.4) | 26.3 (pathogenic) | 0 (deleterious) | 0.366 (pathogenic) | 5.88 (pathogenic) | 0.832 (probably damaging) | 0.053 (damaging) | 1 (disease causing) |
| Genomic position (hg19) | Variant (de novo) | Amino acid change | CADD score [Kircher et al. (26)] | SIFT [Sim et al. (25)] | REVEL [Ioannidis et al. (29)] | GERP score [Cooper et al. (30)] | PolyPhen2 [Adzhubei et al. (27)] | M-CAP [Jagadeesh et al. (28)] | Mutation tester [Schwarz et al. (31)] |
|---|---|---|---|---|---|---|---|---|---|
| 3:100084415 | NM_014820.3: c.1820C > T | p.T607I (NM_014820.4) | 28.4 (pathogenic) | 0.01 (deleterious) | 0.32 (likely pathogenic) | 5.82 (pathogenic) | 0.987 (probably damaging) | 0.048 (damaging) | 1 (disease causing) |
| 3:100086901 | NM_014820.3: c.1660A > T | p.I554F (NM_014820.4) | 26.3 (pathogenic) | 0 (deleterious) | 0.366 (pathogenic) | 5.88 (pathogenic) | 0.832 (probably damaging) | 0.053 (damaging) | 1 (disease causing) |
The TOMM70 gene is clearly intolerant to loss-of-function variation in the human population. In gnomAD (32), the pLI score is 1, which indicates a high probability of intolerance to loss-of-function. Additionally, the missense ‘z’ score of this gene is 1.92, indicating that the gene is also intolerant to missense variation. The TOMM70 variants reported in this paper are not present in control population database gnomAD (32). In both cases, no alternative diagnosis was found despite an extensive genetic workup. We, therefore, sought to directly test the function of the two TOMM70 variants identified in these individuals. We reasoned that loss-of-function variation in TOMM70 would strongly implicate these variants in one or both of the individuals. Moreover, we considered whether differences in function between the two variants could explain some of the differences in the human phenotypes.
Ectopic expression of reference TOMM70 and the patient variants do not cause a phenotype in Drosophila
In order to assess the pathogenicity of the TOMM70 variants, we used the fruit fly model (33–35). Fruit flies have a single ortholog, Tom70. Because of the high conservation between human TOMM70 and fly Tom70 proteins [DIOPT Score (36) is 14/15, 40% amino acid identity and 60% amino acid similarity], the fly provides an optimal platform for functional assessment of the identified variants. To assess the effect of the TOMM70 variants in adult flies, we performed overexpression assays. We expressed the reference and variant cDNA using a ubiquitous driver, Act-GAL4. These flies are viable as adults and do not exhibit any obvious morphological phenotypes. Using eye-specific GMR-GAL4 and ey-GAL4, we also expressed the reference and the variants of human TOMM70 transgene and did not observe any eye phenotypes (Fig. 2B). We also performed western blots to assess if these mutations affect the stability of TOMM70 proteins and did not observe any changes in the protein levels between reference and variant cDNAs (Fig. 2C). Together these data indicate that expression of TOMM70 is not toxic; based on these assays, the variants do not act as gain-of-function or dominant negative mutations. Hence, the variants might act as loss-of-function mutations.
Expression of reference TOMM70, but not the variants, rescues lethality associated with loss of Tom70 in flies
To determine if the human TOMM70 protein is functional in flies, we performed rescue experiments. Given that there are no available null mutations, we first generated a mini-GAL4 knock-in null mutant allele (Tom70miniGAL4Δ). We used CRISPR-Cas9-mediated homologous recombination approach (37–39) and replaced the coding sequence of the Tom70 gene with a Kozak consensus-mini-GAL4 gene-dominant marker (Fig. 2D). Two guide RNAs targeting the 5′ and 3′ end of the coding sequence of the gene were used to remove the coding sequence and a donor plasmid with a left and right homology arm, and a Kozak sequence followed by the mini-GAL4 gene and a dominant marker (U6-gRNA1) were integrated via homology-directed repair (39). GAL4 acts as a transcriptional coactivator that allows expression of transgenes downstream of an upstream activating sequence (UAS) (40,41). The mini-GAL4 is a smaller version of GAL4 and encodes about half the size of the GAL4 protein (40) with ~40% of the transcriptional activity of the wild-type GAL4 protein. Here, we use this Tom70miniGAL4Δ fly line to determine the loss-of-function phenotype; to assess if the human reference cDNA rescues the fly phenotypes; to assess the nature of the human TOMM70 variants; and finally to determine the expression pattern of fly Tom70 gene.
We used the Tom70miniGAL4Δ to humanize the flies by expressing the human reference cDNA and its variants. Homozygous null mutants (Tom70miniGAL4Δ/Tom70miniGAL4Δ) are early pupal lethal and some escapers are late pupal lethal (Fig. 3A). In order to rule out the possibility of a second site mutation in the Tom70miniGAL4Δ line, we crossed Tom70miniGAL4Δ with a corresponding deficiency that spans Tom70 locus [Df(2 L) Exel6031/Cyo]. These flies [Tom70miniGAL4Δ/Df(2 L) Exel6031] are also pupal lethal (Fig. 3A). Expression of the UAS reference TOMM70 transgene in Tom70miniGAL4Δ/Tom70miniGAL4Δ background rescues the lethality at 29 and 25°C (Fig. 3B and C), but not at 22°C, arguing that the level of expression is important. Higher temperatures drive higher levels of the UAS-mediated transgene expression (42). Hence, the mini-GAL4 may not be able to activate the transgene at the required level at low temperature (39). We then expressed the variants, TOMM70I554F and TOMM70T607I, in Tom70miniGAL4Δ/Tom70miniGAL4Δ mutant background, to assess their rescue potential. At 29 and 25°C, the TOMM70T607I and TOMM70I554F variants rescue the lethality at a significantly reduced level when compared with the reference TOMM70 (Fig. 3A–C), indicating that these two variants are partial loss-of-function alleles.
Human TOMM70 rescues pupal lethality in homozygous Tom70miniGAL4Δ. (A) Tom70miniGAL4Δ homozygous flies and Tom70miniGAL4Δ/Df(2 L) are pupal lethal. The lethality in the homozygous Tom70miniGAL4Δ flies was rescued with human TOMM70 transgene. Both of the variants (p.Thr607Ile and p.Ile554Phe) were able to partially rescue the phenotype. This indicates that both the variants are loss-of-function alleles. (B and C) Graph showing the percentage of progeny flies rescued with the reference and variants of TOMM70 transgene at 29and 25°C, respectively. Error bar is indicative of ±SEM. One-way ANOVA with Tukey’s posthoc test was performed to determine the significance (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).
Drosophila Tom70 has a dynamic expression pattern in many organs including CNS
To determine where Tom70 is expressed, we created Tom70miniGAL4Δ > UAS-mCherry flies and observed that Tom70 is expressed in the larval, pupal and adult stages of Drosophila (Fig. 4A). This mCherry protein typically is localized to the cytoplasm. In the third instar larval stage, the mCherry protein appears broadly expressed. However, upon dissection, expression is mostly confined to fat body, the anterior section of the midgut, and the gastric caeca. We observed no expression in the proventricular region of the gut and negligible expression in the salivary gland (Supplementary Material, Fig. S1). As shown in Figure 4B, Tom70 is expressed in the retina, optic lobes and antennal lobes of the adult brain (Fig. 4B). Labeling with the anti-Elav antibody, a nuclear neuronal marker, reveals that most cells that express Tom70 correspond to neurons (Fig. 4B). Given that many cells do not seem to express Tom70 and given that we observed a more restricted pattern than expected, we performed additional experiments.
Tissue level expression of Tom70. (A) Tom70 is expressed in larval (top), pupal (middle) and adult stages (bottom) of Drosophila. (B) In the adult CNS, Tom70 expression is enriched in optic lobes of the brain and in the Elav-positive neuronal cells. mCherry is indicative of Tom70 expression. Scale bar: 50 μm.
We used the G-TRACE system (43) to determine how broadly Tom70 is expressed. G-TRACE allows us to trace the lineage of a cell and determine the current and historical expression of a gene in cells. The mini-GAL4 protein integrated in the gene (Tom70miniGAL4Δ) activates UAS-RFP.nls, a nuclear localized RFP, and hence marks the cells that currently express the gene (Supplementary Material, Fig. S2A). However, the GAL4 protein also activates the UAS-Flp, and the Flippase excises a transcriptional STOP cassette flanked by FRT sites to drive the expression of UAS-GFP.nls from a constitutively active promoter (Supplementary Material, Fig. S2A). Hence, if the gene is expressed at any time in a cell or a precursor of the cells that are being imaged, the GFP.nls expression marks those cells where the gene was active in the past. As shown in Supplementary Material, Figure S2B, the RFP.nls is sparsely expressed in wing and eye discs in third instar larval stages and in the adult CNS. However, G-TRACE reveals that the precursor of most wing and eye discs cells expressed Tom70 at some point. Broad historical expression was similarly observed in precursors of the third instar larval CNS. In summary, the expression pattern of Tom70 is highly dynamic and not all cells express the protein.
Reference TOMM70, but not the variants, rescues Tom70 RNAi-mediated eye morphology and synaptic transmission defects in flies
Given that Tom70 is expressed in the developing and adult eye (Fig. 4B), we performed eye specific knockdown experiments using RNAi to assess tissue specific phenotypes in adults. Mutations in genes that encode some mitochondrial proteins often result in eye morphology defects, such as glossy or rough eye phenotypes, in Drosophila, even in the absence of an ocular phenotype in the human disorder (44–48). To assess the efficacy of the RNAi constructs, we drove two different RNAi constructs with Actin-GAL4 to ubiquitously knockdown Tom70. This induced pupal lethality at both 25 and 29°C (Supplementary Material, Fig. S3), consistent with the null allele that we generated. We then used the GMR-GAL4, and as shown in Supplementary Material, Figure S3, both Tom70 RNAi lines cause a rough eye phenotype with areas of glossy appearance and black spots in the 3-day-old fly eye. Interestingly, we observed that GMR-GAL4 > Tom70RNAi eye phenotype is progressive, and the eye morphology gets severely affected with time (Supplementary Material, Fig. S4). RNAi#1 (18 112 from VDRC) causes a stronger phenotype at room temperature than RNAi#2 (6756R-1 from NIG). We therefore used RNAi#1 in subsequent experiments. To demonstrate that the phenotype is because of the loss of Tom70, we expressed the reference UAS-TOMM70 transgene and showed that there is a significant rescue of the eye phenotype (Fig. 5A and B). This rescue is not because of a reduction in the expression of RNAi#1 as a UAS-empty vector along with UAS-RNAi#1, exhibits the same phenotype as the UAS-RNAi#1 alone (Fig. 5A). However, expression of the variants of TOMM70 rescue the eye roughening induced by UAS-RNAi#1 to a lesser extent (Fig. 5A and B). These data suggest again that the variants are partial loss-of-function alleles.
RNAi-mediated knockdown of Tom70 causes eye morphology defects and synaptic transmission defects that can be rescued by human TOMM70 reference transgene. (A) GMR-GAL4-mediated knockdown of Tom70 causes rough eye phenotype with black eye patches. The eye morphology phenotype was partially rescued when human TOMM70 reference transgene was co-expressed in the Tom70 RNAi background. Co-expression of the variants of TOMM70 in the Tom70 RNAi background did not significantly rescue the eye phenotype. (B) The graph showing that a significant percentage of flies with TOMM70 reference transgene co-expression rescued the Tom70 RNAi phenotype. However, the co-expression of variants was failed to do so. (C) ERG recording of indicated genotypes and (D and E) the quantification of ERG recording data show that the GMR-GAL4 > Tom70RNAi eye have defect ON and OFF transients that is indicative of synaptic transmission. The synaptic transmission defect because of Tom70 RNAi was rescued by human TOMM70 reference transgene, however, both of the variants of TOMM70 failed to do so. Error bar is indicative of ±SEM. One-way ANOVA with Tukey’s posthoc test was performed to determine the significance (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).
To determine the effect of Tom70 knockdown on photoreceptor function, we performed electroretinogram (ERG) of 5–7-day-old GMR-GAL4 > UAS-RNAi#1 eyes. As shown in Figure 5C, the amplitudes of the depolarization are not changed, arguing that the phototransduction machinery is not severely affected. However, there is an obvious loss in amplitude of the ON and OFF transients (Fig. 5C–E), which may indicate a defect in synaptic transmission when Tom70 levels are reduced. Again, this defect is specific as the amplitude of the ON and OFFs is rescued upon co-expression of reference human TOMM70 transgene (Fig. 5C–E). As shown in Figure 5C–E, the variants do not rescue these defects, again suggesting that the identified variants are partial loss-of-function alleles.
Discussion
In this study, we describe clinical, genomic and functional analyses for two affected individuals from separate families with de novo damaging variants in the mitochondrial outer membrane transport complex component, TOMM70. While both children have some common features including significant white matter abnormalities, hypotonia, hyper-reflexia, dystonia and cognitive deficits, their clinical presentations are not identical. Individual 1 has a relatively stable neurodevelopmental disorder that was noted in infancy. By contrast, individual 2 was developmentally normal, even advanced, before experiencing a subacute neurological regression at 4 years of age. On brain imaging, individual 1 has significant hypomyelination and atrophy of the cerebellum and corpus callosum. Individual 2 has diffuse white matter abnormalities with rarefaction. As both of these variants are partial loss-of-function alleles and there is a strong evidence for intolerance to loss-of-function variation in TOMM70 in humans, we suggest that the two variants are implicated in a phenotypic spectrum of neurological abnormalities including hypotonia, ataxia and white matter abnormalities. The differences between the two cases could relate to the differences in allele severity, environmental factors or other modifiers.
During the submission of this work, a report was published online that documents a single individual with developmental delay, microcephaly, anemia and lactic acidosis (49). This child was found to be a compound heterozygote for TOMM70. The variants in this individual are reported in gnomAD (32). The c.1745C > T (p.A582V) variant was present twice, while the c.794C > T (p.T265M) variant was observed nine times. Neither were found in the homozygous state. These variants do not map to known TOMM70 protein motifs. Additionally, this individual has elevated lactate levels when younger, though a normal level was seen at 3 years of age (49). Abnormal respiratory chain complex activities were documented in lymphoblasts from this patient. Neither lactate nor did respiratory chain complex activity are felt to be sensitive or specific for the diagnosis of mitochondrial disease (50). Only one of the nine previously described mitochondrial outer membrane protein disorders has been demonstrated to result in respiratory chain complex deficiency (51,52). We, therefore, speculate that these are mild loss-of-function mutations, unlike the haploinsufficiency alleles discussed in this work.
In mammalian cells, TOMM70 has been shown to facilitate endoplasmic reticulum (ER) to mitochondrial Ca+2 transport by recruiting IP3R3 at ER–mitochondria contact sites (53). During autophagy, TOMM70 forms a complex with TOMM40 to regulate the phagophore expansion by recruiting Atg2 at the mitochondria-associated ER membrane (54). Hence, TOMM70 has multiple roles in mitochondrial import, quality control and Ca2+ dynamics. During the course of this study, we generated the first GAL4 knock-in null mutant allele of Drosophila Tom70. Interestingly, gene expression analyses using G-TRACE suggest that Tom70 has a dynamic expression pattern including the CNS tissue and a substantial and reproducible number of cells in all tissues tested, such as larval wing disc, eye disc and adult brain, do not express Tom70. Hence, it is possible that the protein is highly stable allowing the perdurance of the protein long after the gene is not expressed anymore. Alternatively, it may be that the limited transcriptional activity of mini-GAL4 results in low level of RFP expression that is not detectable.
Additionally, our data show that Tom70 is an essential gene in Drosophila as homozygous Tom70 null flies are pupal lethal. Moreover, knockdown of Tom70 causes impaired synaptic transmission and degenerative eye phenotype indicating a requirement in neuronal maintenance consistent with the observations that TOMM70 can cause neurological impairments in affected individuals. In functional assays, expression of human reference TOMM70, but not the variants, can compensate the absence or loss of Drosophila Tom70. Additionally, in all assays performed, TOMM70I554F behaves as a stronger loss-of-function allele than TOMM70T607I. The p.Ile554 amino acid is present in the last TPR motif, similar to the corresponding p.Val521 amino acid in fly Tom70 whereas the p.Thr607 amino acid does not belong to a functionally annotated region in both human TOMM70 and Fly Tom70 proteins. Altogether, the functional assays in Drosophila indicate that the human variants affect TOMM70 function and future studies will allow us to determine how loss of Tom70 affects neuronal function. The identification of additional affected individuals should allow for further delineation of the spectrum of neurological phenotypes associated with TOMM70 dysfunction.
Materials and Methods
Next generation sequencing
For the individual 1, Trio exome sequencing raw data from prior clinical sequencing by AmbryGenetics was obtained for research-based reanalysis. Analysis of the raw data by Brigham Genomic Medicine revealed one candidate variant of interest, the de novo missense variant in TOMM70. DNA extracted from new blood samples was obtained from the trio and sent to the HudsonAlpha Clinical Services Lab for Sanger confirmation.
Details of exome sequencing per AmbryGenetics report:
Genomic DNA was isolated from the affected individuals and parents’ whole blood. The samples were prepared using the SeqCap EZ VCRome 2.0 (Roche NimbleGen) or the IDT xGen Exome Research Panel V1.0. Each DNA sample was sheared, adaptor ligated, polymerase chain reaction (PCR)-amplified and incubated with exome baits. Captured DNA was eluted, and PCR was amplified. Final quantified libraries were seeded onto an Illumina flow cell and sequenced using paired-end, 100 cycle chemistry on the Illumina HiSeq 2500.
Genome sequencing for individual 2 and both biological parents was performed as previously described in Helman et al. (22).
Fly stocks and maintenance
The fly stocks used in this study were either generated in house or procured from stock centers namely Bloomington Drosophila Stock Center (BDSC), Vienna Drosophila Resource Center (VDRC) and the National Institute of Genetics (NIG) stock centers. All the flies were reared in a standard flyfood and were cultured in room temperature unless otherwise mentioned. The following stocks were used in this study:
Tom70miniGAL4Δ (Generated in this study).
w[1118]; Df(2 L)Exel6031, P{w[+mC] = XP-U}Exel6031/CyO (BDSC # 7514).
y1 M{vas-int.Dm}ZH-2A w*; PBac{y[+]-attP-3B}VK00033 (BDSC # 24871).
UAS-TOMM70-VK33 (Generated in this study).
UAS-TOMM70I554F-VK33 (Generated in this study).
UAS-TOMM70T607I-VK33 (Generated in this study).
UAS-empty vector-VK37.
UAS-Tom70-RNAi #1 (VDRC stock center #18112).
UAS-Tom70-RNAi #2 (NIG stock center #6756R-1).
UAS-mCherry.
UAS-mCherry.nls.
UAS-GTRACE (43).
Generation of Tom70miniGAL4Δ allele
gRNAs targeting 5′UTR (5′-CTGGGCCATCAATCCGGATTGGG-3′) and end of coding region (5′-GAAAAGAGCTGAATCAGTTTGGG-3′) of Tom70 were cloned in pCDF3 vector (Addgene Accession number #49410) following Port et al. (55). Coding sequence of Tom70 between the gRNA sites is replaced with kozak-miniGAL4-U6gRNA1 cassette using homologous recombination with methods described in Kanca et al. (39). Briefly, the homology donor construct containing gRNA1 target sites to linearize the homology donor construct in vivo, 100 nt of homology to cut region on either side of gene-specific gRNA target sites, miniGAL4 with a kozak sequence to ensure proper translation, and U6gRNA1 dominant marker is synthesized in pUC57-Kan vector backbone by Genewiz (South Plainfield, NJ). Injections were conducted as described in Kanca et al. (39). Resulting flies are crossed as single fly crosses with actin5C-Cas9; actin5C-GF#2-FP; actin5C-mCherr-#1-ry for detecting gRNA1 containing animals. U6gRNA1 positive flies are balanced to obtain Tom70miniGAL4∆ allele. The allele is PCR validated using primers: for 5′ region tom70_ch_for (5′-GATGTCTATCGTTTTTCGATGTCGTG-3′) and kozak_mG4_ch_rev (5′-GTCACAAGCTTGCTCGATGG-3′); for 3′ region kozak_mG4_ch_for (5′-CGATGACGAGGATACCCCC-3′) and tom70_ch_rev (5′-GCTTGTTTTAAGTTTAATCTTACGCTTAAACATAAG-3′).
Generation of UAS-TOMM70 reference and variant stocks
UAS-TOMM70 reference and variant transgenic flies were generated as previously described (56). Briefly, UAS-TOMM70 constructs were generated by LR cloning of Gateway (LR clonase II, Thermo Fisher Scientific) compatible pDONR221-TOMM70 cDNA (GenBank: NM_014820.3) into the pGW-attB-HA destination vector (57). The UAS-TOMM70T607I (NM_014820.3: c.1820C > T) and UAS-TOMM70I554F (NM_014820.3: c.1660A > T) were generated by Q5 site-directed mutagenesis (New England Biolabs) and fully sequenced (Sanger) using the following primers (5′–3′):
To generate I554F point mutant TOMM70_I554F_F: CATGGGA-ACTTTTGAAGTACAAAG TOMM70_I554F_R: GTTTCATAGGCAAAATCAC.
To generate T607I point mutant TOMM70_T607I_F: AAACCACCAATATTATAGAACC TOMM70_T607I_R: TAATCCGTATTTCTTTGCAAC.
Sanger Verification M13_F: GTAAAACGACGGCCAG.
Sanger Verification TOMM70_S1: CGTGCAAAAGCCCATGAGAAG.
Sanger Verification TOMM70_S2: CGAGCAAATGCTCTCATCAAAAGAG.
Sanger Verification M13_R: CAGGAAACAGCTATGAC.
All constructs were inserted into the VK33 (PBac{y[+]-attP}VK00033) docking site by ϕC31 mediated transgenesis (58).
Immunostaining and confocal microscopy
Larval and adult tissues were dissected in 1X cold PBS and proceeded for immunostaining of larval and adult brain (59). In short, the tissues were fixed in 4% paraformaldehyde followed by blocking and incubation in the primary antibody (Rat anti-Elav, 1:500) solution. Alexa488 conjugated secondary antibodies were used to detect the primary antibodies. The images were capture using a confocal microscope (Zeiss 880). Tissues expressing GFP and RFP/mCherry were also imaged using the same confocal microscope.
Protein isolation and western blot
Fly heads were decapitated and homogenized using Cell Lysis Buffer (25 mM Tris-HCl pH 7.5, 100 mM NaCl, 1 mM EDTA, 1% Triton-X 100, 1X liquid protease inhibitor (Gen DEPOT), 0.1 M DTT) (42). The supernatant was collected after centrifugation at 13 000 rpm for 10 min at 4°C. The supernatant was mixed with Laemmli Buffer containing β-mercaptoethanol and heated at 95°C for 10 min. Subsequently, the samples were loaded in 4–20% gradient polyacrylamide gels (Bio-Rad Mini-PROTEAN® TGX™) and the proteins were separated using sodium dodecyl sulfate and polyacrylamide gel followed by a transfer onto polyvinylidene difluoride membrane (Immobilon, Sigma). Further, the membrane was blocked using skimmed milk and treated with the primary antibody for overnight. The following antibodies were used in the present study: mouse anti-TOMM70 (1:1000) (Santa Cruz Biotechnology), mouse anti-Actin (1:20 000) (EMD-Milipore, C4). Horseradish peroxidase-conjugated secondary antibody was used to detect the respective primary antibody.
Imaging of adult fly eye
For the imaging of the eye of an adult fly, the flies were anesthetized and taken on a slide with a bridge cover slip. The head of the fly was placed in an angle so that the whole eye can be viewed. Live images were captured using bright field Stereomicroscope (Leica MZ16) and Image Pro-plus software. The images were further processed and assembled using Adobe Photoshop software.
ERG analysis
ERG recording of the fly eye was performed as previously described by Ansar et al. (60). In short, the flies were anesthetized on CO2 pad and glued on a slide. Two glass electrodes filled with 100 mM NaCl solution was used as the recording and reference electrodes. The reference electrode was placed on the cephalothorax region and the recording electrode was placed on the eye. A light pulse of 1 s was applied to record the ERG traces. Ten flies from each genotype were recorded and analyzed using LabChart Reader software.
The authors wish it to be known that, in their opinion, Debdeep Dutta and Lauren C. Briere should be regarded as joint first authors.
Acknowledgements
We would like to thank the affected individuals and families who participated in this study. We would also like to thank Dr Joshua M. Shulman for use of the ERG rig machine.
Funding
This work is supported by the Undiagnosed Disease Network (U54NS093793 to H.J.B., S.Y. and M.F.W. and (U01HG007690 to L.C.B., D.A.S., F.A.H., C.C., M.A.W. and J.K.). The participation of G.H. and C.S. is in part financed by the Australian National Health and Medical Research Council (NHMRC 1068278). The research conducted at the Murdoch Children’s Research Institute was supported by the Victorian Government’s Operational Infrastructure Support Program. H.J.B. is also supported by NIH grant number R24OD022005 and is an investigator of the Howard Hughes Medical Institute. The confocal microscopy facility at the Neurological Research Institute is a part of Neurovisualization Core of the Intellectual and Developmental Disabilities Research Center (IDDRC) supported by NIH U54HD083092. P.C.M. is supported by CIHR (MFE-164712).
Conflicts of Interest statement
None declared.
