Congenital vertebral malformation (CVM) refers to abnormal development of the spine structure presenting as congenital scoliosis, kyphosis, or other congenital vertebral defects and may occur simultaneously with other birth defects or as part of an underlying genetic syndrome [1
]. In human embryogenesis, the vertebral column develops at 4–6 weeks of gestation from the paraxial mesoderm (PSM) and is closely related to the spinal cord and other organs originating from mesoderm [3
]. Multiple organ defects, most frequently involving the cardiac system, urogenital system, limbs, and spinal cord, have a higher occurrence in CVM than in the general population [3
]. A genetic component to CVM risk is suspected. However, only 10~20% of the cases are genetically resolved [6
Nicotinamide adenine dinucleotide (NAD) is biologically essential as a key coenzyme in redox reactions participating in cell metabolism, proliferation and inflammation, as well as circadian rhythm [11
]. In mammals, NAD is de novo synthesized from L-tryptophan or recycled via the salvage synthesis pathway from endogenous NAD metabolites [12
]. Recently, genetic perturbations in the de novo NAD synthesis pathway were identified in individuals with birth defects in vertebral, cardiac, renal organs and limbs, namely the VCRL syndrome [MIM: 617660, 617661, 618845] [13
]. Truncating variants and disruptive missense variants in the genes encoding key enzymes of the de novo synthesis pathway, including HAAO
were identified, respectively. Further investigation revealed that the genetic deletion of the Haao
gene, together with deficient dietary NAD precursors during pregnancy, causes VCRL malformations and miscarriages in mice [15
, encoding the final enzyme in the de novo synthesis pathway, NAD synthetase 1, was reported to be a causative gene for congenital NAD Deficiency Disorder [13
]. In a previous study, biallelic variants in NADSYN1
were identified in five individuals from four unrelated families. However, the mutational spectrum of NADSYN1
-associated congenital disorders has not yet been investigated in a large population cohort.
Here we analyzed the genetic variants in NADSYN1 in an exome-sequenced cohort consisting of Chinese patients diagnosed with CVM and other congenital organ defects. We further performed in vitro functional assays to investigate the effects of these variants on protein expression and enzyme activity. These findings identified the involvement of functional NADSYN1 variants in the complex genetic etiology of CVMs.
2. Materials and Methods
2.1. Patient Recruitment and Clinical Evaluation
A total of 424 probands diagnosed with CVMs were consecutively enrolled and collected in the cohort between 2009 and 2018 at the Department of Orthopedic Surgery of Peking Union Medical College Hospital, as a part of the Deciphering disorders Involving Scoliosis and COmorbidities (DISCO) study (http://www.discostudy.org/
(accessed on 10 January 2019)). Detailed phenotypic data was recorded. X-ray, computed tomography (CT), and magnetic resonance imaging (MRI) were also performed. Deformities of limbs, spine and spinal cord of relevant cases were evaluated via X-ray plain films by two independent surgeons. The cardiac anomalies were evaluated via ultrasonic cardiography. Urogenital and gastrointestinal anomalies were evaluated via ultrasonography of the abdomen. Patients diagnosed with clinical features of VACTERL association (vertebral defects, anal atresia, cardiac defects, tracheo-esophageal fistula, renal anomalies, and limb abnormalities), namely, the anal atresia and tracheo-esophageal fistula were evaluated and ruled out from the study. Physical examination was performed to evaluate the conditions of each patient’s parents. The ethical committee at PUMCH approved the study (IRB number: JS-908). Informed consent was obtained from each participant or their guardians.
2.2. Exome Sequencing and Variant Interpretation
Exome sequencing and bioinformatic analysis were conducted. DNA samples were not available for parents. Variants were called, annotated and filtered using the PUMCH developed pipeline (PUMP) as described previously [6
]. Rare variants (MAF < 0.001) were selected for analysis based on 1000 Genomes (October 2013), the Exome Aggregation Consortium (ExAC; http://exac.broadinstitute.org
(accessed on 10 January 2019)), and the Genome Aggregation Database (gnomAD, http://gnomad.broadinstitute.org/
(accessed on 10 January 2019)). In silico prediction tools, including Sorting Intolerant from Tolerant (SIFT) [16
], Polymorphism Phenotyping v2 (Polyphen-2) [17
], Genomic Evolutionary Rate Profiling (GERP++) [18
] and Combined Annotation Dependent Depletion (CADD) [19
] were utilized to predict the deleterious properties of variants. The RefSeq accession numbers of the transcript and corresponding protein isoform of NADSYN1 we used for mutation nomenclature are NM_018161.5 and NP_060631.2, respectively.
2.3. Site-Directed Mutagenesis Plasmid Construction
C-terminal Myc-His-tagged NADSYN1 cDNA (NM_018161.5) in pcDNA3.1+ (Hitrobio Biotechnology, Beijing, China) was generated and used as the template for site-directed mutagenesis following the manufacturer’s instructions for the KOD-NEO-PLUS Kit (TOYOBO, Tokyo, Japan). Primers for site-directed mutagenesis at each mutation site were listed in Supplementary Table S1
. The mutant plasmids were sequenced on both strands to validate the nucleotide mutation.
2.4. Plasmid Transfection and Western Blots
COS-7 cells were acquired from the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Beijing, China). COS-7 cells were cultured in DMEM high glucose (Gibco, Thermo Fisher Scientific, Hudson, NH, USA) supplemented with 10% fetal bovine serum (Gibco, Thermo Fisher Scientific, Hudson, NH, USA) and penicillin-streptomycin solution (50 U/mL, Gibco, USA). COS-7 cells were grown to 60–70% confluency in 6-well plates and then transfected with 2 μg of pcDNA3.1 NADSYN1 WT and all mutant plasmids using Lipofectamine 3000 reagent (Thermo Fisher Scientific, Vilnius, Lithuania). Eight hours after transfection, DMEM high glucose with 20% FBS was added into the transfection mix equivalently. Forty-eight hours post-transfection, after washing with cold PBS, the COS-7 cells were lysed with 80 µL RIPA buffer (Thermo Fisher Scientific, Hudson, NH, USA) containing protease inhibitor cocktail (Roche, Mannheim, Germany).
Then SDS-PAGE and Western blot analysis were conducted on whole cell lysates by standard methods. A rabbit anti-His-tag antibody (1:1000, AE068, ABclonal, Wuhan, China) was used as the primary antibody and goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (ab6721, Abcam, Cambridge, MA, USA) was applied to detect WT and mutant protein expression levels. Then, the cell lysates were re-probed with horseradish peroxidase-conjugated anti-GAPDH antibody (ab9482, Abcam, Cambridge, MA, USA). NADSYN1 and GAPDH expression level were visualized by automatic chemiluminescence imaging system (C300, Azure Biosystems, Dublin, CA, USA). Then quantification of band intensity was performed by ImageJ (National Institutes of Health, Bethesda, MD, USA). Overall expression levels of WT and mutant NADSYN1 were normalized to GAPDH levels, respectively. Statistical comparisons were conducted with the GraphPad Prism software using the one way ANOVA. Each experiment was repeated three times. p-Values < 0.05 (*) were considered statistically significant.
2.5. Expression and Purification of NADSYN1 Protein in Mammalian COS-7 cells
The COS-7 cells were grown to 60–70% confluency on 2 × 15 cm dishes. Forty-eight hours post-transfection, the cells were lysed on ice for 30 min with PBS + 1%TridonX-100 + 1%NP-40 + 1%PMSF (DMSO free), pH = 7.5, then freeze-thawed 5 times, and ultrasonicated for 10 cycles. WT and mutant NADSYN1 protein was purified using HisPur™ Ni-NTA Purification Kit following the manufacturer’s instructions (88227, Thermo Fisher Scientific, Hudson, NH, USA). The equilibration, wash, and elution buffers were 10 mM, 25 mM, and 500 mM imidazole respectively. The resin beds were incubated with protein extract for 30 min on an end-overend rotor at 4 °C. The eluted proteins were quantified by BCA protein assay kit (Sigma-Aldrich, Darmstadt, Germany).
2.6. Enzymatic Assays of NADSYN1 Protein
NADSYN1 protein acts as the final enzyme in NAD+
biosynthesis, therefore the content of NAD+
was measured using an enzymatic assay described in previous studies [13
]. The reaction buffer was prepared by mixing 2 mM ATP, 0.2 mg/mL bovine serum albumin, 5 mM MgCl2
, 56 mM KCl, 50 mM Tris-HCl (pH 8.0), 20 mM glutamine and 1 mM NaAD. Each reaction system contained 20 µL of reaction mix and 0.2 µg of protein and was then incubated for 60 min at 37 °C and terminated at 95 °C for 5 min. After centrifugation at 13,000 rpm for 15 min, the supernatants were collected for NAD detection. NAD assays were performed in 900 µL of 0.1% ethanol, 10 mM sodium pyrophosphate and 20 units of alcohol dehydrogenase (74931, Sigma-Aldrich, Darmstadt, Germany). The absorbance at 340 nm was measured before and after a 30-min room temperature using Multiskan FC Microplate Photometer (Thermo Fisher Scientific, Hudson, NH, USA). Standard NAD (0–100 nmol) was measured under the same conditions. Statistical analysis of NADSYN1 enzymatic activity were conducted in GraphPad Prism software using one way ANOVA method. Each experiment was repeated three times. p
-Values < 0.05 (*) were considered statistically significant.
In the present study, we identified eight potentially disruptive variants in a total number of nine CVM patients, with subsequent functional assessment. All patients presented with CVMs and other defects, either in heart, kidney, limbs, or spinal cord. We further evaluated the pathogenicity of these variants using in silico prediction methods, and tested the protein expression level and enzymatic activity in the COS-7 cell line. These findings provided further evidence of the involvement of NAD deficiency in congenital organ defects.
Nicotinamide adenine dinucleotide (NAD) is essential for mammals and serves endogenously as a key coenzyme in redox reactions and a precursor for cell metabolism, and also a substrate for protein modifications [11
]. Congenital NAD deficiency due to genetic perturbations and dietary deficiency of NAD precursors would impair organogenesis both in humans and mice [14
]. It was reported that genetic loss of non-redundant de novo NAD synthesis genes, KYNU
, was associated with the VCRL syndrome, which is characterized by congenital vertebral malformations and other organ defects. The vertebral phenotype of humans was reproduced in mouse models with genetic deletion of Kynu
, complicated with dietary deficiency of NAD precursors in pregnant mice. These data suggested that congenital NAD Deficiency Disorder may be caused by a gene–environment interaction.
encodes the final enzyme of the NAD de novo synthesis pathway [23
]. The role of NADSYN1
during organ development has not yet been elucidated. Notably, all five individuals presented in a previous study did not survive more than three months postnatally [13
, like many non-redundant genes in the NAD synthesis pathway, appears tolerant to heterozygous missense mutations in humans. Here, we present nine individuals carrying heterozygous deleterious variants of NADSYN1
with multiple organ defects who survived postnatally with relatively good quality of life (admitted to the hospital from age 1 to 22 years). In addition to CVM, all patients in this study presented with other organ defects involving either the heart, limbs, kidney, liver or spinal cord. Notably, two of the patients presented with multiple hepatic polycysts, which has not been previously reported in NAD Deficiency Disorder cases. Considering that congenital NAD deficiency disorder is caused by a gene–environment interaction, we propose that functional NADSYN1
variants are involved in CVMs with variable penetrance and expressivity.
The functional NADSYN1 protein contains glutaminase and synthetase domains [21
] and catalyzes the ATP-dependent formation of NAD+ from nicotinic acid adenine dinucleotide (NaAD+) at the synthetase domain using the ammonia generated from the glutaminase domain [23
]. It has been previously reported that variants in both the glutaminase (p. Cys49Arg) domain and synthetase domain (p. Ala573Thr) will cause a significant decrease in NAD synthesis activities of encoded NADSYN1
]. To validate the proposed hypothesis, we performed in vitro assays to evaluate the consequences of NADSYN1
variants located in different domains. One frameshift variant, c.861delT(p. Arg288GlufsTer14), led to complete loss of the NAD synthetase domain in the translated protein (Figure 2
A,B). Another possible explanation is that the non-sense mediated decay (NMD) is activated in COS-7 cells to mediate the degeneration of mutant mRNA transcripts from variant c.861delT(p. Arg288GlufsTer14) and c.1216C > T(p.Arg406Ter). Two missense variants in the NAD synthetase domain, c.1511G > A(p. Arg504Gln) and c.1762G > A(p. Glu588Lys), are structurally close to the P2 loop, the active site for NAD synthetase [21
], significantly decreasing the enzymatic activity (Figure 1
A). Consistent with previous results, the missense variant in the glutaminase domain, c.232G > A(p. Val78Ile), demonstrated significantly decreased enzymatic activity [13
]. However, given the fact that congenital Deficiency Disorder is known to be caused by biallelic variants in NADSYN1
, we still consider these variants as uncertain significance.
As a multifactorial disease with a prevalence of ~1/1000 live births [1
], genetic perturbations, gene–environment interactions, and epigenetic changes are suspected to play a role in the etiology of CVMs [4
]. Previously, we have reported the mutational burden and potential oligogenic models underpinning CVM development, as well as compound heterozygous inheritance model in TBX6
-associated congenital scoliosis (TACS) [9
]. The combined effect of deleterious variants in multiple genes might synergistically lead to the development of the malformations, which gives insight into the complex disease-causing model of CVMs. Apart from genetic perturbations, hypoxia, insufficient folate acid supply, and teratogenic drug use during pregnancy are also considered to be potential CVM risk factors [8
]. Considering the complexity of NAD-involved biological processes, the exact mechanism underpinning NAD-mediated organogenesis defects remains to be elucidated. As anticipated, defects in organs originating from the mesoderm in the cardiac system and urogenital system, as well as limbs, often co-occur with somitogenesis defects in other congenital disorders [3
], highlighting the vital role of NADSYN1
during embryogenesis of the mesoderm.
In summary, functional variants in NADSYN1 were involved in the complex genetic etiology of CVMs. Further investigation of congenital NAD Deficiency Disorder will help to understand the pathogenesis of syndromic congenital defects involving CVMs.