WES analysis of the five patients
Five patients with severe EOPE were recruited in this study; their clinical information and data are shown in Table 1. The incidence of EOPE was about 0.5–0.8%, so the incidence of severe EOPE is much lower than EOPE. We speculated that genetic factors may play an important role in patients with severe EOPE; therefore, WES was performed for each patient. The quality of sequencing met the requirements of the bioinformatics analysis, as shown in the WES report (Supplementary Table 1).
Next, we performed a filtering process on the WES data (Supplementary Table 2). First, we retained missense, nonsense, frameshift, and splice site mutations, then filtered out the variants with allele frequencies above 1% in the 1000G, ESP6500, ExAC, and gnomAD databases. We then used our inhouse database to further filter out duplicate mutations, and thus obtained variants for each patient. For example, we identified 419 single nucleotide variants and 36 insertion-deletion mutations in Patient No. 1 (P1, Supplementary Table 2). All variants were further filtered according to the list of 40 selected candidate genes that are known to be associated with severe preeclampsia [9]. In this way, we were able to narrow down the scope of the target. We identified a rare variant of the GOT1 gene c.44C>G:p.P15R in P1; the mutation was confirmed by Sanger sequencing (Figure 1A).
In silico analysis of the GOT1 variant
The Pro15 site is highly conserved across species from human to zebrafish (Figure 1B), suggesting that Pro15 plays an important role in the function of the GOT1 protein. After searching exome and genome sequencing databases, we found that the allele frequency of c.44C>G was 0 in all databases (Table 2), indicating that the mutation is extremely rare. Furthermore, several online mutational prediction tools predicted P15R to be a pathogenic mutation (Table 2). As the variant is heterozygous, we wanted to know whether the heterozygous variant influences disease tolerance. Constraint Metrics Z score for missense variation analysis [22] found that GOT1 was predicted to be intolerant to variation (z = 2.17, http://exac.broadinstitute.org/gene/ENSG00000120053). Thus, the above bioinformatics analysis suggests that the rare variant c.44C>G:p.P15R of the GOT1 gene may be associated with severe EOPE.
Molecular modeling of the mutated protein
Fortunately, we found that the structure of GOT1 protein had been resolved (PDB 3ii0, Figure 2A). By structure analysis, we found that there was an ‘open pocket’ at the surface of GOT1 protein, which was positive charged (Figure 2C). The 15th residue proline was located at the edge of this pocket (Figure 2C). In our study, this residue mutated to arginine (Figure 2B). Arginine is a basic amino acid with a big side chain (Figure 2D). We proposed that this point mutation would change the electric environment of enzymatic center, meanwhile, affect the binding affinity of substrate or product even more the whole process of enzyme-catalyzed reactions. Computer modeling had confirmed our proposal that P15R mutation changed the electricity and enhanced the positive charge of this area (Figure 2D).
Detailed clinical information for P1
Patient No. 1 (P1) is a 35-year-old woman, body weight 62.5 kg, with a BMI of 22.9 kg/m2 prior to pregnancy. Six years ago, she delivered a baby girl by cesarean section because the umbilical cord was wrapped around the fetus’s neck. During the same period, she was diagnosed with severe PE (gestational age of onset is unknown), and her blood pressure returned to normal postpartum. She refused to provide a personal medical history or a family history. The patient’s blood pressure rose to 140/90 mmHg at 25 weeks of gestation, and an ultrasound examination showed that the fetus was small. She began to experience limb edema at 26 weeks of gestation but was not treated at this time. At 28 weeks of gestation, her blood pressure rose to 197/114 mmHg and urinary protein appeared as 3+, but she made no complaints. When the patient visited our hospital at that time, her body weight was 80 kg, and an ultrasound test showed that umbilical cord blood flow was S/D 4.8 and fetal growth was restricted. Her urinary protein went up to 4+, and 24 h urine protein was 7431.2 mg. The highest level of serum creatine was 85.6 μmol/L, blood urea nitrogen was 10.56 mmol/L, uric acid was 472.9 mmol/L, with no detected anemia or thrombocytopenia. The lowest level of albumin was 22.6 g/L. As for liver damage, the highest level of alanine aminotransferase was 52.5 U/L, aspartate aminotransferase was 41.6 U/L, and lactate dehydrogenase was 332 U/L. The highest level of d-dimer was 10.87 mg/L. Echocardiography showed no abnormalities. Both antinuclear antibodies, antibody spectrum, and cardiolipin antibody were all negative. After administering drugs and inducing labor, P1 gave birth to a stillborn child. She recovered well and was released from the hospital 10 days later.