The samples used in this study were peripheral blood obtained from proband and their parents. The protocols for this study were evaluated and approved by the Ethics Committee of Xiangyang Central Hospital. Informed consent forms for WES analysis were signed by the parents of the proband, specifically for the purpose of this research.

The HEK293T and HeLa cell lines were purchased from the China Center for Type Culture Collection (CCTCC, China). The cells were cultured with the high-glucose Dulbecco’s Modified Eagle Medium (DMEM; Gibco, USA) containing 10% fetal bovine serum (Gibco, USA) and 1% penicillin-streptomycin (Gibco, USA), in a constant temperature incubator at 37 °C and 5% CO₂ with saturated humidity.

WES was performed by inputting 150–300 ng of genomic DNA from each sample, which was extracted from peripheral blood samples of the proband and her parents according to the manufacturer’s instructions (MagPure Buffy Coat DNA Midi KF Kit, MAGEN). Exome capture using the MGIEasy Exome Capture V4 Probe (MGI) was followed by paired-end read sequencing (2 × 100 bp read length) on the MGISEQ-2000 platform with an average depth of ≥100-fold. WES and related data analysis were performed by Hangzhou Bosheng Medical Laboratory.

Bioinformatics analysis was performed to investigate the effect of these variations on the MED25 gene at two levels of regulation, the mRNA transcription level and the protein translation level. PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/index.shtml) was applied for the conservation analysis of genetic mutations. HSF (https://hsf.genomnis.com/login), RDDC^SC (https://rddc.tsinghua-gd.org/search-middle?to=SplitToolModel), and SpliceAI (https://spliceailookup.broadinstitute.org/) algorithms are used to predict whether mutations affect the splicing of MED25 mRNA.

First, the InterPro (http://www.ebi.ac.uk/interpro/search/sequence/#opennewwindow), UniProt (https://www.uniprot.org/), and NCBI (https://www.ncbi.nlm.nih.gov/) databases were utilized to predict the MED25 protein domain. Subsequently, Consurf (https://consurf.tau.ac.il/consurf_index.php) was employed to analyze the conservation of the mutation site. The effects of mutations on the secondary structure of the MED25 protein were assessed using the SOPMA database. Additionally, Alphafold3 (https://golgi.sandbox.google.com/) and Missense3D (http://missense3d.bc.ic.ac.uk/∼missense3d/) were used to predict the impact of mutations on the tertiary structure of the MED25 protein. Finally, a comprehensive analysis of the effects of mutations on the stability of the MED25 protein was performed using multiple databases, including DUET (http://biosig.unimelb.edu.au/duet/stability), DynaMut2 (https://biosig.lab.uq.edu.au/dynamut2/submit_prediction_mm), I-Mutant2.0 (https://folding.biofold.org/i-mutant/i-mutant2.0.html), SAAFEC-SEQ (http://compbio.clemson.edu/SAAFEC-SEQ/#started), and MUpro (https://mupro.proteomics.ics.uci.edu/).

To study the effect of variations on MED25 gene splicing, we used reverse transcription combined with nested PCR to analyze the subjects’ periphery MED25 ectopic transcript. For the in vitro splicing study, the minigene pcMINI-N/C-wt/mut and pcDNA3.1-wt/mut were constructed. Based on the mutation site, we devised two vector construction strategies to assess any abnormalities in the splicing patterns. For the mutation c.670C > G p. R224G, which is situated at position −19 of Exon 6, we inserted Exon 5 - Intron 5 - Exon 6 - Intron 6 - Exon seven into the pcDNA3.1 or pcMINI-N vectors. In contrast, for the mutation c.1965 + 1dup, located at the +1 position of Intron 16, we incorporated a segment of Intron 15 - Exon 16 - Intron 16 - Exon 17 into the pcMINI-C vector.

We used genomic DNA as the template for nested PCR. The nested PCR products of the second round were used as templates to amplify wild-type and mutant DNA fragments. Vector and fragments were digested, recovered, ligated, and transformed into colonies for PCR identification and Sanger sequencing.

The successfully constructed wild-type/mutant recombinant minigene were transfected into Hela or HEK293T cells. 48 h later, total RNA was extracted. After reverse transcription synthesis of cDNA, we used primers on both sides of minigene for PCR amplification. Agarose gel was used to detect the size of products.

In the investigation of the c.1965 + 1dup mutation in the MED25 gene, we utilized the father’s blood sample as a control to assess whether the mutation present in the mother’s sample influenced mRNA splicing. Total RNA was extracted from the whole blood samples of the proband’s parents and subsequently reverse transcribed. The cDNAs from both the father and mother served as templates for PCR amplification. The PCR products were analyzed using agarose gel electrophoresis and Sanger sequencing.

The proband is a girl who was admitted to Xiangyang Central Hospital in Hubei Province at the age of 3 years and 10 months due to cognitive, language, and intellectual disabilities, as well as microcephaly. Neither of the proband’s parents had relevant clinical manifestations.

Now, at the age of 6.5 years, the proband exhibits persistent growth retardation according to WHO growth standards, with a height of 1.0 m (−2.3 SD) and a weight of 15 kg (−1.8 SD). In terms of motor function, the patient exhibits significant limitations and is unable to independently perform actions such as running, jumping, or climbing stairs. Comparative assessment results indicate that since the initial evaluation, there has been moderate development in both motor skills and cognitive function. The patient’s language comprehension remains intact; however, verbal expression has been completely lost. Notably, epileptic seizures began at the age of 3.5, presenting as high-frequency paroxysmal activity. During the most recent 2-month observation period, three episodes have been recorded.

The proband’s blood biochemistry and liver and kidney function tests revealed no abnormalities; however, the MRI and electroencephalogram indicated abnormalities. The brain MRI results showed scattered punctate lesions that appeared isointense on T1-weighted images and hyperintense on T2-weighted images in the subcortical white matter of both cerebral hemispheres, with high signal intensity on FLAIR sequences. Additionally, there was bilateral hippocampal volume loss with widening of the temporal horns, and the cavum septum pellucidum was present (Figure 1C). The EEG examination results indicated an abnormal electroencephalogram III (awake/sleep), with typical epileptic discharges detected in the central parietal and temporal brain regions and the anterior and posterior heads of the bilateral brain during the intermittent period, accompanied by tachycardia. A large number of low to mid-amplitude spine slow waves and multiple spine slow waves were observed in the bilateral central parietal and temporal regions, particularly in the left central parietal and temporal regions. A small amount of low to mid-amplitude spine slow waves and multiple spine slow waves were also seen in the bilateral posterior and anterior heads.

Using WES analysis, compound heterozygous variants of c.670C > G (p.R224G) and c.1965 + 1dup in the MED25 gene (NM_030973.4) on chromosome 19 of the proband were detected, which have never been described before. These compound heterozygous variants were inherited from her father and mother respectively (Figure 1A). The analysis results indicated the proband’s mother is a heterozygous carrier of the MED25 gene mutation c.1965 + 1dup, and the father is a carrier of the mutation c.670C > G (p.R224G) (Figure 1B). Moreover, the MED25 gene is associated with Basel-Vanagaite-Smirin-Yosef syndrome, which is an autosomal recessive inheritance. Representative clinical features of BVSY include microcephaly and global developmental delay, which are consistent with the clinical phenotype of the proband. However, according to American Association for Human Genetics and Genomics (ACMG) guidelines, the MED25 c.670C > G (p.R224G) variant was classified as a variant of uncertain significance (VUS), as it meets the PM2 criterion (the variant is absent in the gnomAD population database). The MED25 c.1965 + 1dup variant was classified as a variant of uncertain significance (VUS) based on: PVS1 (a splice-site variant likely leading to aberrant splicing and potential functional impact on the gene) +PM2 (absent in the gnomAD population database).

The WES analysis results indicated the c.1965 + 1dup variant of the MED25 gene is a splice site mutation, which might lead to abnormal gene splicing and affect gene function. However, the impact of the variant c.670C > G (p.R224G) on gene splicing is unclear. Therefore, the in vitro minigene splice assays were firstly utilized to analyze whether the two variants detected in this study affect the alternative splicing of MED25 gene, thereby further affecting gene function.

For each mutation site, we designed two sets of minigene vectors and validated them across different cell lines to ensure the objectivity and accuracy of our detection methods. In vitro MSA results indicated that the variant c.670C > G (p.R224G) did not disrupt the normal splicing of the MED25 gene (Supplementary Figure S3). In contrast, the variant c.1965 + 1dup affected normal splicing, resulting in two abnormal transcripts: retention of 1 bp in intron 16; skipping of exon 16 (Figure 2). The band intensity was scanned and analyzed using ImageJ to calculate the proportion of each transcript variant. The results showed that in the pcMINI-C vector, band b accounted for 53%, while band c made up 47%. In the pcDNA3.1 vector, band b represented 38% and 40%, whereas band c constituted 62% and 60%, respectively (Supplementary Figure S4). Both vectors demonstrated a higher proportion of the aberrantly spliced c band after mutation. Notably, the retention of 1 bp in intron 16 could lead to alterations in the subsequent reading frame, potentially resulting in a premature stop codon (PTC) in the 3′UTR region. Meanwhile, the skipping of exon 16 did not alter the reading frame but might result in the internal deletion of 73 amino acids in the protein.

To further confirm that the variant c.1965 + 1dup can cause abnormal splicing of MED25 mRNA in vivo, we conducted an in vivo splicing experiment. Using a sample from the proband’s father as a control, we assessed whether the mutation in the sample of proband’s mother affected mRNA splicing. PCR amplification was performed using the cDNAs from both the father and mother as templates. The results of the electrophoresis test indicated that both the father and mother exhibited a single band (Figure 3A). The gel electrophoresis bands were subsequently subjected to Sanger sequencing, revealing that the father’s band corresponded to the expected size of the normal splice variant, measuring 393 bp, while the mother’s sequencing results displayed a set of peaks. Consequently, the mother’s PCR product was purified and subjected to TA cloning followed by Sanger sequencing. The results indicated the presence of two splice variants: the normal splice variant and a variant with a retention of 1 bp on the left side of intron 16 (Figures 3B,C).

Here, we demonstrated through both in vivo and in vitro splicing experiments that the variant c.1965 + 1dup can disrupt the normal splicing of MED25 mRNA, while, the variant c.670C > G (p.R224G) does not impact gene splicing.

Although the results of the in vitro minigene splice assays indicated that the variant c.670C > G (p.R224G) has no effect on the splicing of the MED25 gene, the NCBI prediction results showed that the mutation at position 224 is located in the Med25_VWA functional domain (Figure 4A), and the amino acid sequence is highly conserved. Therefore, this missense mutation is likely to affect the structure and function of MED25 protein.

In order to clarify whether the variant c.670C > G (p.R224G) has an impact on the MED25 protein structure, we constructed the secondary and tertiary structures of the mutated MED25 protein through bioinformatics methods and analyzed them. SOPMA database analysis results showed that the proportions of secondary structures such as Alpha helix (Hh), Extended strand (Ee), and Beta turn (Tt) change after mutation. Among them, Hh accounted for 19.68% before mutation and increased to 21.95% after mutation. The proportions of Ee and Tt decreased after the mutation, from 14.59% to 6.83% before the mutation to 13.65% and 5.49% respectively (Figure 4B). This suggests that mutation could affect protein secondary structure.

Consurf was utilized for conservation analysis, and the analysis results showed that the wild-type amino acid at position 224 is highly conserved in terms of exposure and functionality of the protein (Figure 4C), indicating that this mutation is likely to affect the tertiary structure of the protein. So we analyzed the effect of mutation on the tertiary structure of MED25 protein subsequently. Through Alphafold 3 modeling and PyMOL mapping analysis, it was found that the mutation did not affect the structure of the MED25 main chain. However, when ARG224 is mutated into GLY224, the hydrogen bond formed between the side chain and the original PHE195 position is broken (Figures 4D,E), thus affecting the structure of the MED25 side chain. Using the Adaptive Poisson-Boltzmann Solver (APBS) in ChimeraX for mapping analysis, the results indicated that the mutation also has a certain impact on the surface electrostatic potential of the MED25 protein molecule (Figure 4F). Subsequent Missense3D prediction results suggested that the c.670C > G (p.R224G) mutation could lead to cavity alteration, which would have a destructive effect on the tertiary structure of MED25 (Figure 4G).

Finally, we analyzed the impact of the 224-position mutation site on the stability of MED25 protein through multiple software such as DUET, MUpro, DynaMut2, SAAFEC-SEQ and I-Mutant2.0 SEQ. The negative ΔΔG values across all algorithms suggest that the mutation reduces MED25 protein stability. Take mCSM or DUET for instance, a ΔΔG below −0.5 kcal/mol is indicative of high instability. As shown in Figure 4H, the mutation can significantly affect the stability of MED25 protein. It can be seen that the missense mutation c.670C > G (p.R224G) not only affects the structure of MED25, but also has a significant impact on protein stability.