Identification of Variants in Two Families With Congenital Cataract by Whole Exome Sequencing

ABSTRACT

Background

Congenital cataract (CC) is an ophthalmic disorder diagnosed at birth or during early infancy, characterized by severe visual impairment and potential blindness. This study aimed to evaluate the pathogenicity of two variants: GJA8 c.201C>G (p.Asp67Glu) and CRYBB2 c.325A>T (p.Ile109Phe).

Methods

The patient's lenses were examined using slit‐lamp photography and B‐ultrasound. Whole‐exome sequencing (WES) and sanger sequencing were performed to identify the variants. Multiple sequence alignment was conducted using CLUSTALW to assess the evolutionary conservation of the affected residues. Structural modeling and conformational analyzes of wild‐type (WT) and mutant‐type (MT) proteins were conducted using AlphaFold2, PyMOL, and computational structural dynamics approaches, including root mean square fluctuation (RMSF) and normal mode analysis (NMA).

Results

Two likely pathogenic missense variants were identified in two unrelated families: c.201C>G (p.Asp67Glu) in exon 2 of GJA8, which encodes Connexin 50 (Cx50) protein, was paternally inherited in family 1, whereas c.325A>T (p.Ile109Phe) in exon 4 of CRYBB2 occurred de novo in family 2. The c.201C>G (p.Asp67Glu) variant mapped to the first extracellular loop (EL1) of Cx50. Structural modeling revealed disruption of the Asp67‐Arg185 ionic interaction (Cα‐OD distance: 3.4 → 4.9 Å) alongside a novel hydrogen bond with Asn63. Reduced RMSF in the mutant indicated localized rigidification of EL1, potentially impairing channel gating and correlating with severe cataract in affected family members. The co‐segregation of the variant with the affected phenotype within the family 1 was confirmed by sanger sequencing combined with phenotypic analysis. The c.325A>T variant (p.Ile109Phe) in CRYBB2 increased protein flexibility, and the bulky aromatic phenylalanine side chain perturbed the hydrophobic core of Greek key motif III, causing global dynamic instability. The CRYBB2 c.325A>T variant was further confirmed to be de novo by sanger sequencing in conjunction with phenotypic analysis in family 2. Furthermore, multiple sequence alignment demonstrated that both residues are highly evolutionarily conserved.

Conclusions

This study expands the pathogenic spectrum of congenital cataract. Our findings provide novel insights into the molecular diagnosis of CC and have important implications for genetic counseling.

We describe likely pathogenic missense variants causing congenital cataract and investigate these variants using high‐throughput gene sequencing, Sanger sequencing, bioinformatics, and computational biology tools. Our findings provide novel insights into the molecular diagnosis of congenital cataract and have important implications for genetic counseling.

1. Introduction

Congenital cataract is characterized by lens opacification, with a global incidence of approximately 1.91–4.24 per 10,000 live births, accounting for roughly 14% of childhood blindness worldwide and representing one of the leading causes of preventable visual impairment in children (Li et al. 2020). Although classification by age of onset, etiology, or locus of opacification is possible, morphological subtyping remains the clinical standard, encompassing anterior (e.g., polar, pyramidal, subcapsular), central (e.g., nuclear, sutural, lamellar, cerulean, Christmas tree, pulverulent, coronary), and posterior variants (e.g., lenticonus, subcapsular, polar, oil‐droplet, persistent hyperplastic primary vitreous, Mittendorf dot), as well as total cataracts (Taylan Şekeroğlu and Utine 2021). The etiology of CC is heterogeneous, encompassing genetic factors, prenatal infections, metabolic disorders, and environmental factors, among which genetic predisposition constitutes the predominant causative mechanism (Tătaru et al. 2020). CC is predominantly inherited in an autosomal dominant (AD) manner, accounting for approximately 75% of hereditary CC, followed by autosomal recessive (AR) and X‐linked recessive (XR) patterns (Ding et al. 2020). The website (https://cat‐map.wustl.edu/, accessed on June 1, 2026) recorded 558 genes that may be associated with cataract. To date, more than 100 genes have been implicated in CC pathogenesis, functionally categorized primarily into lens crystallin genes (accounting for ~50%) and membrane protein genes, including connexins and aquaporins (~25%) (Shi et al. 2022). The intricate genetic and phenotypic heterogeneity of CC poses considerable challenges for establishing reliable genotype–phenotype correlations in clinical practice (Patel et al. 2019).

Untreated congenital cataract may result in irreversible loss of visual, hence, prompt genetic diagnosis and timely surgical management are essential to preserve the visual function of pediatric patients (Al‐Bakri et al. 2022; Lloyd et al. 2025). Over the past decade, WES has emerged as a powerful tool for identifying pathogenic genes in monogenic diseases (Bamshad et al. 2011). Moreover, whole‐exome sequencing, by virtue of its cost‐effective targeting of coding regions and canonical splice sites, has emerged as a first‐tier diagnostic tool for such heterogeneous disorders. Despite considerable research into congenital cataract, the molecular mechanisms underlying its onset and progression in humans remain incompletely understood (Shanbagh et al. 2023). In the present study, WES identified a likely pathogenic GJA8 variant, c.201C>G (p.Asp67Glu), inherited from the father, and a de novo CRYBB2 variant, c.325A>T (p.Ile109Phe). We further investigated the potential molecular mechanisms by which these two distinct mutations contribute to cataract pathogenesis.

2. Materials and Methods

2.1. Patient Consent and Ethics Statement

Written informed consent was obtained from each subject or their guardians for participation in this study. The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of the First People's Hospital of Changde City (protocol code 2024‐088‐01) on May 24, 2024.

2.2. Whole Exome Sequencing and Bioinformatic Analysis

Genomic DNA was isolated from peripheral blood samples of the proband using the Qiagen DNA Blood Midi/Mini kit (Qiagen GmbH, Hilden, Germany) following the manufacturer's protocol. The extracted DNA was subsequently sheared to generate fragments of approximately 200 bp in length, followed by end‐repair and the addition of a single adenine nucleotide to the 3′ termini. Following PCR amplification, the DNA libraries underwent hybridization‐based capture with the Nano WES Human Exome V1 capture kit (Berry Genomics) as per the manufacturer's protocol. Paired‐end sequencing (2 × 150 bp) was performed on the NovaSeq 6000 platform (Illumina, San Diego, CA, USA). Base calling and generation of raw sequence data were accomplished by processing the raw image files with CASAVA v1.82. Sequence reads were mapped to the human reference genome (GRCh37/hg19) using the Burrows‐Wheeler Aligner (BWA), and PCR duplicates were subsequently eliminated with Picard v1.57 (http://picard.sourceforge.net/). Variant annotation and interpretation were carried out with ANNOVAR and the Enliven Variants Annotation Interpretation System (v2.10.0), both licensed to Berry Genomics. In this workflow, Enliven was utilized for variant calling, whereas Verita Trekker was applied for annotation and automated classification of pathogenicity based on the established American College of Medical Genetics and Genomics (ACMG) guidelines in conjunction with the Genome Analysis Toolkit (GATK; https://software.broadinstitute.org/gatk/) (Wang et al. 2010; Huang et al. 2025). Variants exceeding an MAF threshold of 1% under a recessive, or 0.1% under a dominant, inheritance model were discarded. All identified variants were annotated and filtered against population frequency databases, including the Exome Aggregation Consortium (ExAC http://exac.broadinstitute.org/) and the Genome Aggregation Database (gnomAD http://gnomad.broadinstitute.org/). Pathogenicity assessment was conducted utilizing multiple in silico prediction algorithms: MutationTaster (http://www.mutationtaster.org/), SIFT (http://sift.jcvi.org), PolyPhen‐2 (http://genetics.bwh.harvard.edu/pph2), Combined Annotation Dependent Depletion (CADD https://cadd.gs.washington.edu), PROVEAN (https://provean.jcvi.org/), FATHMM (https://fathmm.biocompute.org.uk/) and Rare Exome Variant Ensemble Learner (REVEL https://sites.google.com/site/revelgenomics/). Disease‐gene and phenotype associations were curated by cross‐referencing the Online Mendelian Inheritance in Man (OMIM http://www.omim.org), ClinVar (http://www.ncbi.nlm.nih.gov/clinvar), and Human Gene Mutation Database (HGMD http://www.hgmd.org). Clinical whole exome sequencing encompasses the entire human exome, while our bioinformatics analysis and interpretation were restricted to genes with an established disease association in the OMIM database (phenotype mapping key of 3). In total, 5133 genes were analyzed, a number that varies as the OMIM database was updated. Furthermore, we analyzed 64 genes with a clearly established cataract association. The gene list was provided in Table S1. Variants interpretation and classification into five categories‐pathogenic (P), likely pathogenic (LP), variant of uncertain significance (VUS), likely benign (LB), and benign (B) were performed in strict accordance with the American College of Medical Genetics and Genomics guidelines (Richards et al. 2015).

2.3. Mutation Validation by Sanger Sequencing

For segregation analysis, peripheral blood samples were drawn from I‐2, II‐1, II‐2, II‐3 and III‐2 of family 1, as well as from I‐1 and I‐2 of family 2. Primers were designed using Primer 5 based on the DNA fragments that needed to be sequenced. The sequences of the GJA8 primers were GJA8‐Forward: 5′‐GTGCACATTGACCGTTCTGG‐3′ and GJA8‐Reverse: 5′‐CCTCCAGCCGGAACTTCTTA−3′. The sequences of the CRYBB2 primers were CRYBB2‐Forward: 5′‐GCCTTGGTCATCTGACATGC‐3′ and CRYBB2‐Reverse: 5′‐TTCAAGCAATTCTTCTGCCTC‐3′. PCR products were purified and sequenced using an ABI 3730xl DNA Analyzer with the BigDye Terminator Cycle Sequencing Kit (Applied Biosystems, Foster, CA, USA).

2.4. Conservation Analysis, Protein Structural Modeling, Structural Dynamics, and Mutational Impact Analysis

Cx50 and CRYBB2 protein sequences across multiple species were retrieved from the Ensembl database (https://asia.ensembl.org/index.html). Multiple sequence alignment was subsequently performed using the online CLUSTALW tool (https://www.genome.jp/tools‐bin/clustalw), with the resulting alignments visualized in BioEdit. For structural modeling, the AlphaFold2 protein structure database (https://alphafold.ebi.ac.uk/) was queried to generate three‐dimensional models of wild‐type and mutant Cx50 and CRYBB2 proteins. Final structural visualization and analysis were accomplished using PyMOL v2.6.0 (open‐source).

To investigate the structural effects of the identified variants, point mutations were introduced into the WT structures to generate mutant models. Energy minimization and structural optimization were subsequently performed using the CHARMM36m force field implemented in OpenMM (v8.4) to relieve steric clashes and optimize local conformations. Structural dynamics and conformational stability analyzes were further conducted using molecular dynamics related structural evaluation approaches. Protein structural features, residue flexibility, and local conformational alterations surrounding the mutation sites were analyzed using BioPython and MDAnalysis (v2.10.0). The mutational effects on protein stability and intermolecular interactions were assessed by comparing WT and mutant structural models, with particular attention to hydrogen bonding patterns, side‐chain orientation, and local secondary structure alterations.

2.5. Mutational Hotspots of GJA8 and CRYBB2

ClinVar records for GJA8 and CRYBB2 were retrieved through the NCBI E‐utilities on 15 August 2026. Records were restricted to missense variants on NM_005267.5 and NM_000496.3 with a non‐conflicting aggregate germline classification of pathogenic/likely pathogenic or benign/likely benign and with assertion criteria provided. Variants of uncertain significance, conflicting records, records lacking assertion criteria, non‐missense variants, variants on other transcripts, and records that could not be mapped to the corresponding protein were excluded. Protein positions were derived from ClinVar protein‐change annotations. Coding‐exon boundaries were mapped from NCBI RefSeq, and protein features were obtained from UniProtKB/Swiss‐Prot entries P48165 (Cx50) and P43320 (CRYBB2). Distinct substitutions affecting the same codon were retained and stacked at the same protein position. The study index variants were displayed separately and were not included in the ClinVar counts. P/LP and B/LB variants were plotted on separate tracks together with coding‐exon and protein‐feature tracks. Region‐specific P/LP density was calculated as the number of qualifying P/LP records divided by the region length and expressed per 100 amino acids.

3. Result

3.1. Clinical Characteristics

In family 1, a female proband was diagnosed at birth with congenital cataract accompanied by visual blurring. At the age of 14 years, she was admitted to the Departments of Ophthalmology and Medical Genetics at the First People's Hospital of Changde City, where she underwent surgical treatment and received genetic counseling. The pedigree of this family 1 is illustrated in Figure 1A. Slit‐lamp photograph of the proband (III‐1) revealed diffuse, homogeneous white pupillary opacification consistent with a dense mature total cataract. B‐scan ultrasonography demonstrated dense echogenicity within the lens, confirming marked lens opacification, while the retina remained flat and completely attached. Postoperatively, the proband (III‐1) achieved a best‐corrected visual acuity (BCVA) of 0.02 in both the right eye (OD) and the left eye (OS) after surgery (Table 1). Intraocular pressure (IOP) (OD, 13 mmgh; OS, 14 mmgh) was within normal limits, with mild nystagmus and bilateral amblyopia also noted (Figure 2A–E). The proband's father (II‐1) had suffered from bilateral congenital cataracts since childhood and did not undergo bilateral cataract extraction until the age of 19, without intraocular lens (IOL) implantation. No psychiatric phenotypes were observed. The proband's mother (II‐2) exhibited normal visual acuity and IOP (OD, 12 mmgh; OS, 16 mmgh); however, she was diagnosed with Grade III schizophrenia at age 24, with a Positive and Negative Syndrome Scale (PANSS) score of 67 and a Social Disability Screening Schedule (SDSS) score of 9, and had been receiving long‐term antipsychotic medication. The proband's elder sister (III‐2), sharing the same parents, was also born with bilateral congenital cataracts and underwent bilateral cataract extraction without IOL implantation at age 1. She currently presented with bilateral nystagmus, amblyopia, and strabismus, with normal IOP (OD, 10 mmgh; OS, 14 mmgh). At age 17, she was diagnosed with Grade III schizophrenia (PANSS score, 48; SDSS score, 7) and had been receiving continuous psychiatric medication. The proband's paternal grandmother (I‐2) was noted to have poor bilateral vision since birth and was diagnosed with bilateral congenital cataracts at age 12. She underwent cataract extraction with IOL implantation in the left eye at age 19 and in the right eye at age 45, accompanied by bilateral nystagmus (Table 1). Additionally, the proband's paternal aunt (II‐5) and two male cousins (III‐4 and III‐5) were affected with congenital cataracts but had not yet undergone genetic testing or genetic counseling. Conversely, another paternal aunt (II‐3) was unaffected and tested negative for pathogenic variants associated with congenital cataracts (Figure 3A).

FIGURE 1.

Pedigrees structure of two families with congenital cataract. (A) For the Family 1 genetic pedigree diagram. (B) For the Family 2 genetic pedigree diagram. The arrow points to the proband. The black solid square and circle represent the affected male patient and female patient with congenital cataract, respectively. The gray solid square and circle represent the affected male patient and female patient with schizophrenia, respectively. Unfilled squares and circles indicate normal males and females, respectively.

TABLE 1.

Demographic, clinical, and ophthalmological characteristics of affected individuals from two unrelated families in this study.

Family ID	Family 1	Family 2
Diagnosis	ADCC	ADCC
Cataract	BL	BL
No. of patients	7	1
Patient ID	I‐2, II‐1, II‐5, III‐1, III‐2, III‐4, III‐5	II‐1
Proband underwent WES	III‐1	II‐1
Age of onset	Infancy	Infancy
Current vision status	0.02 (opaque lens)	NA
Nystagmus	+	+
Photophobia	+	+
Achromatopsia	+	NA
B‐scan	III‐1	II‐1
SD‐OCT scan	Normal	NA
Fundus examination	Normal	NA
Additional features	II‐2 and III‐2 Grade III schizophrenia	Unilateral hearing impairment

Abbreviations: +, present; −, absent; ADCC, autosomal dominant congenital cataract; BL, bilateral; NA, not available; WES, whole‐exome sequencing.

FIGURE 2.

Ophthalmic examinations of two probands. (A) Slit‐lamp photographs of the proband (III‐1) from Family 1. (B) Optical coherence tomography (OCT) analysis of the proband (III‐1) from Family 1. (C, D) Fundus images of the proband (III‐1) from Family 1. (E) B‐scan ultrasonography images of the proband (III‐1) from Family 1. (F) B‐scan ultrasonography images of the proband (II‐1) from Family 2. OD, right eye; OS, left eye.

FIGURE 3.

GJA8 and CRYBB2 variant analysis of the pedigree. (A) The sanger sequencing of the family 1, red arrow indicates mutation site. (B) The sanger sequencing of the family 2, red arrow indicates mutation site. (C) The evolutionary conservation Cx50 p.Asp67 across multiple species clusters, red arrow indicates mutation site. (D) The evolutionary conservation CRYBB2 p.Ile109 across multiple species clusters, red arrow indicates mutation site.

In family 2, a 4‐month‐old female infant (II‐1) presented with photophobia and poor visual behavior in family 2 (Figure 1B). She was diagnosed with congenital cataracts accompanied by mild nystagmus and unilateral hearing loss, and underwent bilateral phacoemulsification at age 4 months (Figure 2F, Table 1). Her parents (I‐1 and I‐2) exhibited no ocular disease‐related phenotypes and were negative for the pathogenic variant (Figure 3B).

3.2. Variant Identification, Bioinformatic Analysis and Mutation Validation by Sanger Sequencing

In family 1, we performed WES on the proband (III‐1), which generated 18.18 GB data covering 99.97% target region (99.73% achieving a sequencing depth of ≥ 20×). Following exclusion of pathogenic variants in other known congenital cataract‐associated genes, we identified a heterozygous missense variant (NM_005267.4: c.201C>G, p.Asp67Glu) in GJA8 at chromosome 1:147380283 in family 1. According to the ACMG guidelines, this variant was classified as likely pathogenic based on the following criteria: (1) The variant was absent from large population databases, including ExAC, the 1000 Genomes Project, gnomAD, and the NHLBI Exome Sequencing Project (ESP), among East Asian populations (PM2_Supporting); (2) Multiple in silico predictive algorithms, including SIFT, PolyPhen‐2, MutationTaster, CADD, and REVEL (score = 0.864), consistently indicated a damaging effect on the gene or its product (PP3_Moderate) (Table 2); (3) Sanger sequencing confirmed the presence of the heterozygous missense mutation (GJA8:c.201C>G, p.Asp67Glu) in all affected family members (I‐2, II‐1, III‐1 and III‐2), whereas unaffected individuals (e.g., II‐2) harbored the wild‐type sequence, demonstrating co‐segregation of the variant with the disease phenotype in this pedigree (PP1_Supporting) (Figure 3A); (4) The Cx50 p.Asp67Glu variant is located within a highly conserved and significant region of EL1 (Figure 6A), where it disrupts the stable salt bridge between Asp67 and Arg185, of note, benign variants also being present in this EL1 region (PM1_Supporting); (5) This variant had previously been documented in the ClinVar database (accession VCV000952446.9) in a case also presenting with congenital cataract (PS4_Supporting).

TABLE 2.

Description and pathogenic predictions from different bioinformatics tools of disease‐causing variants identified in this study.

Family ID	Gene physical location (hg19)	c.notation	p.notation	gnomAD All AF	Mutation taster	SIFT	Polyphen‐2	PROVEAN	FATHMM	CADD PHRED	REVEL	ACMG/AMP classification
Family 1	chr1:147380283	c.201C>G	p.Asp67Glu	0	Deleterious	Deleterious	Damaging	Deleterious	Damaging	17.5	0.864	Likely Pathogenic
Family 2	chr22:25625421	c.325A>T	p.Ile109Phe	0	Deleterious	Deleterious	Benign	Deleterious	Damaging	25	0.816	Likely Pathogenic

Note: gnomAD All AF represents the allele frequencies among all the populations (v.2.1.1). Reference transcripts; GJA8  = NM_005267.5, CRYBB2  = NM_000496.3.

Abbreviation: AF, allele frequency.

FIGURE 6.

Graphical representation of genes indicating their mutational hotspots. (A) Schematic diagram of GJA8 with all the identified mutations shown in their respective exons. (B) Schematic diagram of CRYBB2 with the frequently found mutations in their respective exons. The disease‐causing variants identified in this study are written in red and their location in the protein domain is also marked red.

In family 2, WES was performed on the proband (II‐1), generating 8.83 GB of data with 99.96% coverage of the target region, of which 99.58% achieved a sequencing depth of ≥ 20×. Sequence analysis revealed a heterozygous missense variant (NM_000496.4: c.325A>T, p.Ile109Phe) in CRYBB2 at chromosome 22:25625421 in family 2. According to the ACMG guidelines, this variant was classified as likely pathogenic based on the following criteria: (1) The variant was absent in large population databases, including ExAC, 1000 Genomes Project, gnomAD, and ESP, specifically within East Asian populations (PM2_Supporting); (2) Multiple computational prediction algorithms, including SIFT, Polyphen‐2, Mutation Taster, CADD, and REVEL (score = 0.816), independently predicted a deleterious impact of this variant on gene function or gene product (PP3_Moderate) (Table 2); (3) Sanger sequencing confirmed the presence of this heterozygous missense variant (CRYBB2:c.325A>T, p.Ile109Phe) in the proband (II‐1), whereas both parents harbored the wild‐type allele, indicating that this variant occurred de novo (PM6) (Figure 3B); (4) This variant had previously been documented in the ClinVar database (accession VCV002442008.1) in two cases also presenting with congenital cataract (PS4_Supporting).

3.3. Conservation Analysis and Protein Structure Analysis

The conservation analysis revealed that c.201C (amino acid Asp67) of the human GJA8 gene is highly conserved among various species, including Homo sapiens , Pan troglodytes , Mus musculus , Macaca mulatta , Rattus norvegicus , Canis lupus familiaris , Bos indicus , Gallus gallus , Xenopus tropicalis , and Danio rerio (Figure 3C). The Cx50 p.Asp67Glu mutation is situated within the first extracellular loop, a region essential for connexon docking and channel gating (Figure 6A). In the wild‐type structure, Asp67 establishes a stable salt bridge with Arg185, with an interatomic distance of approximately 3.4 Å (Figure 4A). By contrast, the glutamate side chain in the variant induces an additional methylene group (‐CH₂‐) relative to aspartate, resulting in excessive spatial extension. This conformational change elongates the Cα–OD distance to Arg185 to approximately 4.9 Å, thereby disrupting the native salt bridge (Figure 4B). Molecular dynamics simulations further revealed that the amino acid substitution markedly reduced the root mean square fluctuation values of the overall protein and residues surrounding the mutation site compared with the wild‐type, leading to localized rigidification and decreased flexibility of EL1(Figure 5A,B). Such structural alterations may compromise the channel gating function of GJA8. Nevertheless, the Asp67Glu variant did not significantly alter the low‐frequency mode eigenvalues, normal mode analysis suggesting that the global folding stability of the Cx50 protein remains preserved (Figure 5C). Additionally, this variant is located within the mutational hotspot region of GJA8 (Figure 6A).

FIGURE 4.

Protein structural modeling analysis and schematic representation of Cx50 and CRYBB2 protein. (A) Three‐dimensional model and enlarged image of residues of Asp67 in wild‐type Cx50. (B) Three‐dimensional model and enlarged image of residues of Glu67 in mutant Cx50. (C) Three‐dimensional model and enlarged image of residues of Ile109 in wild‐type CRYBB2. (D) Three‐dimensional model and enlarged image of residues of Phe109 in mutant CRYBB2. MT, mutant type; WT, wild type.

FIGURE 5.

Structural dynamics and mutational impact analysis of Cx50 and CRYBB2 proteins. (A–C) Analysis of the Cx50 protein. (A) Root mean square fluctuation (RMSF) profiles comparing wild‐type (WT) and mutant‐type (MT) structures, showing residue‐level flexibility differences. (B) Atomic fluctuation (RMSF) analysis of Cx50 WT and mutant structures, showing differences in residue‐level flexibility and dynamic behavior. (C) Eigenvalue spectrum derived from normal mode analysis, reflecting differences in structural stiffness between WT and MT. (D–F) Analysis of the CRYBB2 protein. (D) RMSF comparison between WT and MT, highlighting local flexibility changes. (E) Atomic fluctuation (RMSF) analysis of CRYBB2 WT and mutant structures, showing differences in residue‐level flexibility and dynamic behavior. (F) Eigenvalue spectrum comparison, indicating alterations in global motion dynamics caused by the mutation.

Cross‐species sequence alignment revealed that the c.325A nucleotide position, specifying isoleucine at residue 109 of human CRYBB2 is evolutionarily invariant across a broad phylogenetic spectrum, from mammals ( Homo sapiens , Pan troglodytes , Mus musculus , Macaca mulatta , Rattus norvegicus , Canis lupus familiaris , and Bos indicus ) to non‐mammalian vertebrates ( Gallus gallus , Xenopus tropicalis , and Danio rerio ) (Figure 3D). The CRYBB2 p.Ile109Phe mutation is located within the Greek key III (GKIII) domain (Figure 6B). Residue 109 serves as a critical component of the hydrophobic core in the GKIII domain. In the wild‐type structure, Ile109 adopts a characteristic branched aliphatic “Y”‐shaped conformation that fits precisely into the hydrophobic cavity, forming a tight hydrophobic interaction with Ile124. However, substitution of isoleucine with phenylalanine introduces a rigid hexagonal aromatic ring side chain. This bulky moiety significantly reduces the spatial distance to Ile124 and induces severe steric clash, thereby weakening the native hydrophobic interaction and compromising the integrity of the hydrophobic core (Figure 4C,D). MD simulations and NMA demonstrated that the Ile109Phe substitution markedly elevated the RMSF values of the overall protein and residues surrounding the mutation site compared with the wild‐type, resulting in enhanced global flexibility and disruption of local structural rigidity (Figure 5D–F). Notably, this missense variant resides within the mutational hotspots of CRYBB2, a region recurrently implicated in congenital cataract pathogenesis (Figure 6B).

4. Discussion

Through integrating genetic variant identification, evolutionary conservation analysis, protein structure prediction, structural dynamics and mutational hotspots analysis, the paternally inherited GJA8 variant c.201C>G (p.Asp67Glu) and the de novo CRYBB2 variant c.325A>T (p.Ile109Phe) are likely pathogenic, and we further discussed potential molecular mechanisms. The GJA8 gene mapped to chromosome 1q21.2 and comprising three exons, encodes connexin 50, a 433‐amino‐acid protein that assembles into gap junction channels on the membranes of lens fiber cells, mediating intercellular communication and maintaining lens homeostasis (Shi et al. 2022). Mutations in GJA8 have been extensively documented in association with congenital cataracts in humans, with more than 146 disease‐causing variants cataloged in inherited cataract pedigrees (Cat‐Map; https://cat‐map.wustl.edu/) (Shiels 2024). The majority of reported pathogenic GJA8 variants are inherited in an autosomal dominant (AD) pattern or de novo heterozygous missense changes, although rare cases of homozygous and compound heterozygous variants have also been described (Shiels 2024; Merepa et al. 2025). Previous studies have established associations between GJA8 variants and developmental eye disorders, highlighting the presence of mutational hotspots and distinct genotype–phenotype correlations (Merepa et al. 2025; Huang et al. 2026). Furthermore, GJA8 mutations have been implicated in a broader spectrum of ocular and systemic conditions, including acorea‐microphthalmia‐cataract syndrome, congenital aphakia and long‐segment Hirschsprung's disease (Dong et al. 2024; Lucas et al. 2025; Ma et al. 2024). The Cx50 protein comprises four transmembrane domains (TM1‐TM4), two extracellular loops (EL1 and EL2), one cytoplasmic loop (CL), and N‐/C‐terminal domains (Merepa et al. 2025; Dong et al. 2024). The extracellular loops of Cx50 represent critical structural determinants for gap junction channel function, with EL1 and EL2 mediating hemichannel docking via conserved disulfide bonds, hydrogen bonds, and hydrophobic interactions (Nielsen et al. 2012). Among the three connexins expressed in the lens (Cx43, Cx46, and Cx50), Cx50 plays a particularly indispensable role in lens morphogenesis. Notably, during early postnatal development, Cx50‐mediated intercellular communication accounts for approximately 63% of coupling in lens epithelial cells (White et al. 2007). The majority of reported pathogenic GJA8 variants are clustered within EL1, EL2, and the C‐terminal domain (Shi et al. 2022). The Cx50 p.Asp47His mutation alters charge polarity, resulting in impaired connexin channel function (Yan et al. 2019). The Cx50 p.Glu201Lys mutant protein was aberrantly localized in the cytoplasm, which impaired its proper localization at the plasma membrane, disrupting channel stability (Su et al. 2013). The Cx50 p.Ser258Phe variant induces conformational changes in the C‐terminal domain that lead to aberrant recognition and degradation by the autophagy‐lysosome pathway (Li et al. 2025). Notably, a missense mutation at the same codon as identified in the present study, GJA8 c.200A>G (p.Asp67Gly), has previously been reported to cause congenital cataract (Reis et al. 2013). These variants demonstrate that perturbations in extracellular domains and C‐terminal domain converge on channel destabilization, as same to the p.Asp67Glu (White et al. 2007). Previous studies have demonstrated that approximately 40% of disease‐causing GJA8 mutations manifest as nuclear cataracts (Guo et al. 2022). Nuclear cataracts localize to the embryonic and fetal nuclear regions, which are composed of primary and early secondary lens fibers and are not renewed postnatally, whereas total cataracts involve diffuse opacification of the entire lens, indicating more widespread disruption of lens homeostasis (Hejtmancik 2008).

In family 1 of this study, all affected members consistently presented with bilateral severe total cataract, distinct from the typical nuclear pattern. The Cx50 p.Asp67Glu mutation is located within a highly conserved region of EL1, where it disrupts the stable salt bridge between Asp67 and Arg185, thereby perturbing local domain architecture. This conformational alteration most likely interferes with disulfide bond pairing within the EL1 structure, leading to functional loss or impaired trafficking of the assembled gap junction channels, and ultimately resulting in the severe total cataract phenotype rather than the classic nuclear cataract.

The β‐crystallins family comprises seven major lens proteins, designated βB1, βB2, βB3, βA1/A3, βA2, and βA4 (Bloemendal et al. 2004). The CRYBB2 localized to chromosome 22q11.2 and contains six exons, encodes a 205 amino acid protein containing four Greek key motifs. This protein constitutes approximately 35% of the total lens crystallin and serves as a critical component for maintaining lens transparency and normal optical function. Mutations in CRYBB2 can lead to aberrant protein folding, aggregation, or degradation, thereby disrupting normal lens development (Tătaru et al. 2020; Qi et al. 2023; Shiels and Hejtmancik 2007). CRYBB2 expression in rodents is restricted to the adult lens; in humans it is present throughout all developmental stages, attributable to its capacity for domain and subunit swapping, which enables it to associate with other β‐crystallins family members to solubilize and stabilize them (Zhang et al. 2003). CRYBB2 assembles into homodimers via intermolecular interactions between β‐strands, thereby providing the structural foundation for lens transparency (Hussain et al. 2023). This intricate folding renders the Greek key motifs highly conserved, whereas the majority of pathogenic CRYBB2 mutations cluster within the Greek key III and IV domains, where disease‐causing variants typically disrupt protein stability and structural integrity (Yu et al. 2021). For instance, the CRYBB2 c.62T>A (p.Ile21Asn) variant has been shown to induce abnormal protein aggregation, activate stress response pathways, and trigger both excessive cellular proliferation and apoptosis (Chen and Zhu 2021). The CRYBB2 p.Gly149Val mutation severely disrupts oligomeric interactions and promotes the formation of abnormal aggregates (Wu et al. 2023). The p.Ala2Val variant exhibits poor thermal stability due to its impaired ability to form tetramers at high concentrations (Xu et al. 2012).

In a sporadic congenital cataract patient enrolled in this study, we identified a heterozygous missense mutation in CRYBB2, c.325A>T (p.Ile109Phe), located within the GKIII domain encoded by exon 4 (Figure 6B). The CRYBB2 p.Ile109Phe mutation substitutes a highly conserved isoleucine in the hydrophobic core of the GKIII motif with phenylalanine, which possesses a bulky aromatic side chain. This non‐conservative replacement disrupts the internal hydrophobicity of the Greek key motif, while the steric hindrance imposed by the bulky aromatic ring of phenylalanine distorts the local β‐sheet, markedly increasing local conformational instability, ultimately leading to congenital cataract through protein aggregation. This likely pathogenic mechanism is consistent with previously reported βB2‐crystallin mutations, including p.Trp59Cys, p.Trp151Cys, p.Ser175Gly, p.His181Gln, p.Arg188His, and p.Val187Glu (Chen et al. 2013; Rocha et al. 2021; Zhao et al. 2017).

In the present study, we identified heterozygous missense variants in GJA8 and CRYBB2 in two unrelated families, and investigated the underlying two pathogenic mechanisms of congenital cataract: disruption of membrane channel protein function and misfolding of lens structural proteins through conservation analysis and structural modeling. Moreover, we acknowledge that these predictions require more experimental validation.

5. Conclusion

In conclusion, we have identified a likely pathogenic variant c.201C>G of Cx50 (p.Asp67Glu) and a de novo c.325A>T likely pathogenic variant of CRYBB2 (p.Ile109Phe) associated with congenital cataracts in two Chinese families. Our finding can broaden the mutation spectrum of GJA8 and CRYBB2, enriching the phenotype–genotype correlation of congenital cataract.

Author Contributions

Yazhou Huang: conceptualization, funding acquisition, writing – original draft, investigation, data curation. Xinwei Leng: data curation, investigation, writing – original draft, conceptualization. Jibo Zhang: investigation, data curation. Xingxin Fang: data curation, investigation. Ting He: data curation, investigation. Lihua Zhou: data curation. Zhenkai Wu: writing – original draft. Yuankun Wang: writing – original draft, writing – review and editing. Dan Peng: conceptualization, writing – review and editing, funding acquisition.

Funding

This work was supported by the Hunan Provincial Natural Science Foundation of China (2024JJ7014, 2025JJ70643) and research fund of The First People's Hospital of Changde City (2024ZC05).

Consent

Consent for publication was obtained from parents of all the enrolled patients.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: A list of cataract‐related genes in the analysis of the WES data.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.