GPT2 mutations in autosomal recessive developmental disability: extending the clinical phenotype and population prevalence estimates

Abstract

The glutamate pyruvate transaminase 2 (GPT2) gene produces a nuclear-encoded mitochondrial enzyme that catalyzes the reversible transfer of an amino group from glutamate to pyruvate, generating alanine and alpha-ketoglutarate. Recessive mutations in GPT2 have been recently identified in a new syndrome involving intellectual and developmental disability (IDD), postnatal microcephaly, and spastic paraplegia. We have identified additional families with recessive GPT2 mutations and expanded the phenotype to include small stature. GPT2 loss-of-function mutations were identified in four families, nine patients total, including: a homozygous mutation in one child [c.775T>C (p.C259R)]; compound heterozygous mutations in two siblings [c.812A>C (p.N271T)/c.1432_1433delGT (p.V478Rfs*73)]; a novel homozygous, putative splicing mutation [c.1035C>T (p.G345=)]; and finally, a recurrent mutation, previously identified in a distinct family [c.1210C>T (p.R404*)]. All patients were diagnosed with IDD. A majority of patients had remarkably small stature throughout development, many < 1st percentile for height and weight. Given the potential biological function of GPT2 in cellular growth, this phenotype is strongly suggestive of a newly identified clinical susceptibility. Further, homozygous GPT2 mutations manifested in at least 2 of 176 families with IDD (approximately 1.1%) in a Pakistani cohort, thereby representing a relatively common cause of recessive IDD in this population, with recurrence of the p.R404* mutation in this population. Based on variants in the ExAC database, we estimated that approximately 1 in 248 individuals are carriers of moderately or severely deleterious variants in GPT2.

Fig 1

Pedigrees and segregation of GPT2 mutations in the current study. a-d Pedigrees and corresponding sequence chromatograms for the compound heterozygous GPT2 mutations identified in Pedigree 1 (a, b) and homozygous mutation identified in Pedigree 2 (c, d). e,f Pedigrees of GPT2 mutations identified in a Pakistani cohort of individuals with autosomal recessive intellectual disability. Note, some pedigrees do not show all siblings to aid in maintaining anonymity of the patient and family. Positions of mutations are according to transcripts NM_133443.3 (a-e) and NM_133443.2 (f) and protein NP_597700.1

Fig 2

Exogenous expression and transaminase enzymatic activity of GPT2 mutants. a Western blot results of lysates from HeLa cells transfected to express C-terminal V5-His-tagged human GPT2 (hGPT2-V5His) or one of the GPT2 mutants P272L (previously identified loss-of-function mutation), C259R, V478Rfs*73, N271T, or Q80E. Lysates from cells transfected with empty vector were used as a control. Blots were probed with an anti-V5 antibody to detect exogenously expressed hGPT2 protein; blotting for tubulin served as a loading control. b,c Results of measurement of transaminase enzymatic activity of wild-type hGPT2 and hGPT2 mutants using lysates from HeLa cells as described in (a). Shown are a time course of hGPT enzymatic activity over a 60-min incubation period with substrate (b) and hGPT enzymatic activity measured in cell lysates after incubating with substrate for 30 min (c). n = 3 (V478Rfs*73, N271T, and Q80E), n = 6 (hGPT2, C259R, and vector control). Data are presented as means ± SEMs. ***p <0.001

Fig 3

Subcellular localization of GPT2 mutants in cultured cells. a-e Structured illumination microscopy (SIM) images of HeLa cells exogenously expressing C-terminal V5-His-tagged human GPT2 (hGPT2-V5His) (a) or one of the hGPT2 mutants C259R (b), V478Rfs*73 (c), N271T (d), or Q80E (e) and immunostained with an anti-V5 antibody to detect hGPT2 (green). Prior to fixation, cells were incubated with MitoTracker Red to label mitochondria (red). Nuclei were labeled with Hoechst dye (blue). Boxed areas are magnified on the right. All images were acquired using the same imaging parameters. Scale bar, 5 μm. f,g Quantification of the percent co-localization of exogenously expressed wild-type or mutant hGPT2 with MitoTracker Red based on images similar to those shown in (a-e). Co-localization was calculated as a percent of total hGPT2 stained area, as indicated by V5 staining (f), and as a percent of total MitoTracker Red stained area (g). n = 11 (hGPT2-V5His), n = 7 (C259R), n = 9 (V478Rfs*73), n = 9 (N271T), n = 13 (Q80E). Data are presented as means ± SEMs. *p ≤0.05, ***p ≤0.001

Fig 4

Structural analysis of two deleterious GPT2 variants. a Bar diagram of the GPT2 protein, with mutations newly characterized in the current study indicated in magenta. Black arrows point to the transaminase catalytic residue, residues involved in binding of the coenzyme pyridoxal 5’-phosphate (PLP), and residues at the homodimer interface. b Stability analysis of wild-type (WT) GPT2 (blue) and the two GPT2 mutants C259R (orange) and N271T (gray), as reflected by Rosetta Energy Units (REU). REU is correlated with energy, which is an indicator of how stable the structure is (lower REU score = higher structure stability). c Visualization of the C259R GPT2 mutant. Wild-type GPT2 is shown in cyan, with C259 in green; mutant GPT2 is shown in magenta, with R259 in orange. di-iii Visualization of the N271T GPT2 mutant and PLP binding. Wild-type GPT2 is shown in blue, with N271 in green (i, iii); mutant GPT2 is shown in magenta, with T271 in cyan (ii, iii); PLP is shown in yellow (i-iii). A hydrogen bond between PLP and N271 is present in wild-type GPT2 (i, dashed yellow line), but is lacking in the N271T mutant (ii)

Fig 5

The GPT2 synonymous mutation c.1035C>T (p.G345=) results in aberrant use of a donor splice site. a Generation of an exon trapping vector for testing of the effects of the c.1035C>T (p.G345=) mutation on splicing. Exon 8 of the human wild-type GPT2 gene, as well as portions of the flanking introns (141 bp of intron 7 and 290 bp of intron 8), was ligated into the multiple cloning site (MCS) of the pET01 vector (mutated base in red). Site-directed mutagenesis was performed to generate a pET01-GPT2 vector with the c.1035C>T mutation in exon 8. Wild-type (WT) and mutant (Mut) vectors were confirmed by Sanger DNA sequencing (bottom chromatograms). The mutated base is underlined. b,c Results of exon trapping assays. RT-PCR were performed with cDNA generated from HEK293T cells transfected with either the wild-type (WT), mutant (Mut), or control (empty pET01) vector. Amplicons were separated by agarose gel electrophoresis (b) and analyzed by Sanger DNA sequencing (c). The c.1035C>T mutation in exon 8 results in a 4-bp deletion due to creation of a new splice donor site (compare bottom two chromatograms)

Fig 6

Pipeline for the classification of GPT2 variants and prediction of allele frequencies. GPT2 variants were characterized as benign, moderate, or severe based on variant annotations provided by ExAC (bottom left table) and use of PolyPhen-2 and SIFT (see Materials and Methods for details). Final variant counts in each category are shown in red boxes. Total allele frequency was computed across all variants in each category (bottom right table). The frequencies of heterozygous alleles and homozygous alleles in each variant category were predicted based on calculations assuming Hardy-Weinberg equilibrium and a recessive mode of inheritance