Human Mutation

Metabolic Unit, Department of Clinical Chemistry, Amsterdam University Medical Centers, Vrije Universiteit Amsterdam, Amsterdam Neuroscience, Amsterdam Gastroenterology & Metabolism, Amsterdam, The Netherlands Department of Child Neurology, Emma Children’s Hospital, Amsterdam University Medical Centers, Vrije Universiteit Amsterdam and Amsterdam Neuroscience, Amsterdam, The Netherlands Department of Functional Genomics, Center for Neurogenomics and Cognitive Research, Amsterdam Neuroscience, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands Department of Pharmacotherapy, College of Pharmacy and Pharmaceutical Sciences, Washington State University, Spokane, Washington Walloon Excellence in Life Sciences and Biotechnology, Brussels, Belgium Laboratory of Biochemistry, de Duve Institute, University of Louvain, Brussels, Belgium Department of Genetic Metabolic Diseases, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam Neuroscience, Amsterdam Gastroenterology & Metabolism, Amsterdam, The Netherlands

Thus far, 32 variants in the D2HGDH gene have been described in patients with D-2-HGA (The Human Gene Mutation Database), of which 21 are missense variants (Ali Pervaiz et al., 2011;Haliloglu et al., 2009;Misra et al., 2005;Struys, Korman et al., 2005;Struys et al., 2006). To date, only three of these variants have been characterized via overexpression studies (Struys, Korman et al., 2005. Truncating variants are usually classified as pathogenic, whereas inferring the pathogenicity of missense variants is more challenging. Bioinformatic prediction tools are important for estimating the potential of missense variants to disturb protein function but do not confirm the pathogenicity, and sometimes may fail to give correct predictions (Frigeni et al., 2017).
Therefore, functional studies via site-directed mutagenesis and enzyme assay remain pivotal in the characterization of variants of unknown significance.
Here, we describe a functional assay to characterize missense variants in the D2HGDH gene, while simultaneously reporting the functional evaluation of 10 novel missense variants in conjunction with 21 previously reported missense variants.

| Variant analysis
The coding region of the D2HGDH gene was analyzed as described . All 10 exons and the adjacent splice sites were amplified and sequenced using an ABI PRISM 3130xl Genetic Analyser (Applied Biosystems, Foster City, CA) with interpretation using Mutation Surveyor (Softgenetics, State College, PA). Nucleotide numbering of the variants reflects cDNA numbering in which +1 corresponds to the A of the ATG translation initiation codon in the reference sequence (NM_152783.3).

| Site-directed mutagenesis
The pCMV5-D2HGDH wild type construct was used for site-directed mutagenesis and overexpression (Achouri et al., 2004). Variants were introduced in the D2HGDH open reading frame using specific primers (Table S1; Betsalel et al., 2012). Successful mutagenesis (as well as the absence of PCR artifacts) was confirmed by full-length sequencing of D2HGDH.

| Overexpression of missense variants in HEK293 cells
Transient expression of constructs was performed in HEK293 cells cultured in Dulbecco's Modified Eagle Medium (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% heat-inactivated fetal bovine serum, 1% penicillin/streptomycin, and 2 mM L-glutamine. Cells were plated in 60 mm dishes and transfected at 70% confluence following 24 hr of culturing. Fugene HD transfection reagent (Promega, Madison, WI) was used following the manufacturer's specifications. A ratio of 10:2 (µl Fugene HD to µg total plasmid DNA) was determined to be optimal.
Every construct was cotransfected with pEGFP-N1 vector at a ratio of 1:50. This enabled the monitoring of transfection efficiency using fluorescent microscopy. Transfections of the variant containing vectors as well as of the pCMV5-D2HGDH wild type and pCMV5 mock vector were performed in triplicate. Untransfected cells (no plasmid, only Fugene HD) also served as controls. Cells were harvested by trypsinization 48 hr posttransfection, divided in aliquots and stored at −80°C until further use.

| Detection of overexpressed D-2-HGDH proteins
To verify successful transfection, SDS PAGE/Western blot analysis was performed as described (Pop et al., 2018). 2.6 | D-2-HGDH enzyme activity of transfectants D-2-HGDH enzyme activity was assessed as described (Wickenhagen et al., 2009), and was optimized with respect to protein concentration, reaction time, and substrate concentration. Cell pellets were resuspended in assay buffer, disrupted by sonication, and homogenates were clarified by centrifugation. The protein content of lysates was determined by Bicinchoninic acid assay (Sigma-Aldrich), and samples were adjusted to 0.14 mg/ml protein. Enzyme assays were carried out at 37°C for 20 min, using 53.7 μΜ [ 2 H 4 ]D-2-HG (total assay volume of 120 µl). The final reaction product [ 2 H 4 ]Lglutamate was quantified by LC-MS/MS. Control assays for Lglutamate dehydrogenase activity were performed by addition of 4.2 μM [ 2 H 4 ]2-KG (total assay volume of 120 µl). Activities of wild type and variants were corrected for the HEK293 endogenous D-2-HGDH activity by subtracting the amount of glutamate in mock transfectants. Pathogenic  13 p.Asp375Tyr 17 2 a Chet 0.0018
D-2-HG excretion and diagnosed with other metabolic diseases. The p.Gly436Val variant (87% residual activity) was detected in a patient with succinic semialdehyde dehydrogenase (SSADH) deficiency (Struys et al., 2006), and the novel variant, p.Leu453Phe (68% residual activity), was found in a patient diagnosed with D-2-HGA Type II. The mild effect of these variants on D-2-HGDH activity and the fact that they are found in heterozygous state in patients diagnosed with other metabolic deficiencies suggests they are likely rare nonpathogenic variants (Table 1). However, we cannot exclude that the mild decrease in D-2-HGDH activity could also contribute to the phenotype of these SSADH or D-2-HGA Type II patients.
p.Ala392Gly (59% residual activity) was detected in the heterozygous state as the sole potentially pathogenic allele in one patient.
Patient-derived fibroblast analyses revealed a complete impairment of D-2-HGDH activity. Yet, the clinical significance of the p.Ala392Gly variant is not fully elucidated, and therefore we cannot exclude that other unidentified variants may exist.
p.Ala446Val displays 67% residual activity. Conservation of Ala446 is poor across D-2-HGDH orthologs ( Figure S2) and the software prediction tools predict the p.Ala446Val variant as benign (  (Table S2)  did not alter protein homodimerization or intracellular localization (Lin et al., 2015). Eventual crystallization and x-ray diffraction analyses of D-2-HGDH will be illuminating in placing the variants we have identified within a 3-dimensional context.
D2HGDH transcriptional regulation studies in colorectal cancer have recently been reported (Han et al., 2018). These investigators identified a putative Hif-1α transcription factor binding site in the D2HGDH promoter, suggesting that Hif-1α controls both the expression of D-2-HGDH as well as D-2-HG levels. Comparable promoter studies are needed in our patients for whom variants with high residual activities and/or a single potentially pathogenic allele, were identified.
The pathogenicity potential of the studied variants was also estimated using PolyPhen-2, SIFT and MutationTaster prediction software tools (Alamut Visual software package, version 2.9; Table   S2a-d). While for 64.5% of the variants the in silico predictions are in agreement with the measured activities, for 35.5% of variants there are dissimilarities within in silico predictions as well as between these and functional data, thus emphasizing the importance of the functional assays and their role in the classification of variants of unknown significance.
In conclusion, we present the functional characterization of 31

ACKNOWLEDGMENTS
The authors are thankful to the clinicians and biochemists involved in the diagnosis of this patient cohort. The authors are also grateful to the colleagues of the Metabolic Unit, Amsterdam for their support.