A Novel p.Tyr129His Variant in SlX1 Leads to Dominant,Delayed-onset Hearing Loss with Possible Association with Congenital Anosmia＊

PANG Xiu Hong, ZHENG Xiao Yong, LIN Yun, ZHENG Hao, XU Jun,LIU Dong, JIN Chun Yan, ZHANG Lu Ping, ZHANG Yu Ting, CHU Jiu Sheng,#,CHAI Yong Chuan,#, and YANG Tao,#

SIX1(OMIM 601205), as a member of the SIX homeobox transcription factor family, belongs to homologs of the Drosophila ‘sine oculis’ gene that expresses primarily in the developing visual system of the fly. The SIX family includes six members(SIX1–SIX6), which share a homologous DNA-binding homeodomain (HD) and a highly conserved protein–protein interacting SIX domain (SD)[1]. These genes are involved in vertebrate and insect development and maintenance of the differentiated state of tissues[1,2,3]. Branchio-oto-renal (BOR) and branchio-otic (BO) syndrome have been reported to be associated with dominant mutations inSIX1, both are characterized by hearing loss and branchial anomalies, and the former also has the characteristics of renal malformations[4]. In addition,SIX1mutations may also lead to dominant nonsyndromic deafness DFNA23 with variable audiogram profile[5].

For embryonic development of ear, kidney, and other organs,SIX1,SIX4(OMIM 606342), andSIX6(OMIM 606326) are absolutely necessary[4]. In animal models, it has been shown thatSix1andSix4control formation and early neurogenesis of the olfactory placode[6]. The transcriptional regulation ofSix1andSix4had multiple olfactory-specific genes, and the olfactory placode development could not be initiated by the double knock-out mice[7]. However,no olfactory-related disease has been associated with mutations in theSIXgene family in humans.

Since SIX1 does not have intrinsic activation domains, co-transcription factors are needed to facilitate the transcriptional activation, such as the EYA family[8]. The four EYA-related proteins(EYA1–EYA4) are mammalian homologs of the Drosophila eyes absent (eya) gene, with a highly conserved C-terminal Eya domain of -270 amino acids interacting with the SIX1 SD domain[1]. The crystal structure of human SIX1–EYA2 complex (PDB:4EGC, https://www.rcsb.org/structure/4EGC) is available as a model for the study of structural,electrical potential, and functional changes resulting fromSIX1andEYAmutations[9]. Furthermore, Site Directed Mutator (SDM), an online computational tool that predicts effects of mutations on protein stability, has successfully revealed the structural and functional basis of a number of missense mutations inSIX1that destabilize the SIX1–EYA2 complex structure[9].

In this study, we recruited a dominant Chinese Han family with delayed-onset, progressive hearing loss. Interestingly, four of the seven affected family members were found with additional anosmia. Our study identified p.Tyr129His, a novel variant inSIX1, as the probable cause of hearing loss and proposed that anosmia maybe a new phenotype associated withSIX1mutations with incomplete penetrance .

Thirteen adult members, seven affected and six unaffected, of the Chinese Han Family P-D26(Figure 1A) were recruited through the Department of Otolaryngology—Head and Neck Surgery, Taizhou People’s Hospital. All seven patients were given a thorough medical history investigation and subsequent clinical examination, with special attention paid to audiological, renal, branchial,olfactory, cardiac, ophthalmologic, skeletal, mental,and dermatologic abnormalities. Hearing loss was evaluated through otoscope, pure-tone test,distortion product otoacoustic emission, immittance,and auditory brainstem response. Possible abnormalities of middle and inner ear were explored by high-resolution computed tomography and magnetic resonance imaging. Renal abnormalities were excluded by ultrasonic test. Olfactory function was evaluated using smelling agents, such as alcohol,essence, vinegar, camphor oil, kerosene, and water as control. Informed consent was signed by all patients to participate in this study. No identifying information was shown in the manuscript. This study was approved by the Ethics Committee of Taizhou People’s Hospital and was in compliance with the Declaration of Helsinki.

Figure 1. Characterization of Family P-D26. (A) Pedigree showing genotype–phenotype co-segregation for the variant p.Tyr129His in SIX1. Proband III-2 is pointed by the arrow. Black and gray on the left and right halves indicate hearing loss and anosmia, respectively. (B) Audiograms showing moderate-to-severe sensorineural hearing loss with the sloping audiometric feature in the affected family members.

DNA samples were extracted from the venous whole blood by the Blood DNA Extraction Kit(Tiangen Biological Technology Co., Ltd, China).Targeted next-generation sequencing (NGS) was used to sequence a total of 406, 20, and 3 known causative genes for deafness, olfactory disorders,and gustory disorders, respectively. Data analysis and bioinformatics processing were performed as previously described[10]. Minor allele frequencies(MAFs) of the candidate variants were extracted from public databases Exome Variant Server,gnomAD, and 1,000 genomes. Through Sanger sequencing in all subjects and 400 Chinese Han controls, candidate pathogenic variants were confirmed. Possible deleterious effect was predicted by computational tools PolyPhen-2, Mutation Taster,and PROVEAN (cut-off score≤ 1.3)[10].

The mutant complex was built by PyMOL version 1.8.X (www.pymol.org) using the wild-type crystal structure of human SIX1–EYA2 complex (PDB: 4EGC)as the template to predict the structural and functional changes of SIX1 resulting from the p.Tyr129His. Protein stability of the mutant SIX1–EYA2 complex was predicted by SDM. The adaptive Poisson–Boltzmann solver (APBS) method was applied to calculate the electrostatic potential[11]. The wild-type and mutant SIX1–EYA2 complexes were subjected to conversion of partial charges and atomic radiiviaPDB2PQR[12]. Structure and electrostatic potential energy of wild and mutant complexes was visualized by Visual Molecular Dynamics (VMD) version 1.9.3.

Family P-D26 has seven family members with bilateral, delayed-onset, moderate-to-severe sensorineural hearing loss segregated in autosomal dominant inheritance mode (Figure 1A and 1B). The age of onset varied from approximately 3 (III-2 and III-4) to 10 (II-2, II-4, II-5, II-7, and II-8) yr old. All patients had a sloping audiometric feature toward higher frequencies, started with moderate hearing loss and progressed with age. Interestingly, four(III-2, II-4, II-5, and II-7) of the seven patients also had congenital olfactory dysfunction (Figure 1A).Renal, branchial, cardiac, ophthalmologic, skeletal,mental, intestinal, dermatologic, and inner ear abnormalities were excluded through a series of clinical examinations.

Three heterozygous non-synonymous variants with MAFs less than 0.01 including c.2408C>T and c.4315C>A inTECTAand c.385T>C inSIX1were identified by targeted NGS of 406 recognized deafness-causative genes (Supplementary Table S1 available in www.besjournal.com) in proband III-2. In all members, Sanger sequencing showed that within this family, only the c.385T>C (p.Tyr129His) inSIX1(NM_005982.4) segregates with hearing loss phenotype (Figure 1A and Figure 2A). The p.Tyr129His inSIX1was not seen in databases Exome Variant Server, gnomAD, 1,000 genomes, and 400 Chinese Han controls. It is predicted as deleterious by computational tools PolyPhen-2, Mutation Taster,and PROVEAN (Supplementary Table S2 available in www.besjournal.com). The p.Tyr129 residue was evolutionarily conserved (Figure 2B). In proband III-2,targeted NGS of 20 and 3 known causative genes for olfactory and gustatory disorders, respectively, did not identify any non-synonymous variants with MAF less than 0.01.

The mutated SIX1-EYA2 complex model was built based on the human wild crystal structure (PDB:4EGC) as the template to predict the possible pathogenic effects of the p.Tyr129His on the SIX1 protein structure. The electrostatic potential calculations were done by PDB2PQR and APBS and visualized by VMD (Figure 3). Compared with the wild-type SIX1 protein, the p.Tyr129His mutant protein showed a significant bulge outward in the surface model and the electrostatic potential energy changes from positive (red) to neutral (white)(Figure 3). Compared with the wild-type SIX1–EYA2 complex, the stability of the mutant protein complex was predicted to be reduced (predicted ddg: –0.29)by SDM.

Figure 2. The novel p.Tyr129His variant in SIX1 identified in Family P-D26. (A) Chromatograms showing the heterozygous missense c.385T>C (p.Tyr129His) variant. (B) Multi-species conservative analysis of SIX1 showing the highly conserved Tyr129 residue (pointed by the arrow) in human, mouse, cattle, chicken,and the drosophila.

In this study, a p.Tyr129His variant inSIX1was identified as the probable cause for hearing loss in Family P-D26. Evidence supporting its pathogenic role includes as follows: (1) The dominant, delayedonset, audiometrically sloping hearing loss(Figure 1B) in Family P-D26 resembled the hearing phenotype, which is previously reported for the majority of the DFNA23 patients[5]; (2) the p.Tyr129His variant segregated with the hearing phenotype within the family (Figure 1A); (3) it was predicted to be pathogenic by multiple computational tools and was not seen in public variant databases of the general populations and 400 ethnically matched controls.

The p.Tyr129His variant substituted a highly conserved tyrosine residue (Figure 2B) for histidine.It was located in the HD of SIX1 and was the fourth amino acid of the ETSY tetrapeptide, which is essential for sub-classification of the six family proteins[13](Figure 2; Supplementary Figure S1 available in www.besjournal.com). Mutations in the HD domain, such as p.Tyr129Cys, p.Gln125Lys, and p.delGln133, have been shown to disrupt the EYA1–SIX1–DNA complexes[4,5]. Since p.Tyr129His changed the same residue as p.Tyr129Cys, we postulated that these two variants may have a similar pathogenic mechanism. According to our protein structural simulation, the p.Tyr129His variant may result in structural and electrostatic potential changes of the SIX1–EYA2 protein complex and decrease in affinity and stability between SIX1 and EYA2 (Figure 3).

Of numerousSIX1missense mutations identified to date, most mutations in the SD domain are associated with BO syndrome, whereas most mutations in the HD domain are associated with non-syndromic deafness DFNA23 (https://www.uniprot.org/uniprot/Q15475) (Supplementary Figure S1). Our result for the p.Tyr129His variant apparently strengthened this apparent genotype–phenotype correlation.

Figure 3. Simulation of structure and electrostatic potential energy in wild-type. (A) and mutant. (B) SIX1-EYA2 complex. Upper (enlarged) and lower boxes highlight the structural alteration from the wild Tyr129 to the mutant His129 residue. Note the mutant His129 residue bulged outward in comparison with the wild-type Tyr129 in the surface mode, and the electrostatic potential energy shifted from positive (red) to neutral (white).

As a surprising result of the current study, we observed congenital anosmia in four of the seven affected family members in addition to hearing loss(Figure 1A). To our knowledge, association of anosmia withSIX1mutations has not been previously reported in humans. Studies in the mouse models, however, suggest that the Six–Eya regulatory pathway controls early differentiation and survival of the cranial neurons[14]. Both Six1 and Eya2 are highly expressed in the olfactory pit. Loss of Six1 leads to defective olfactory epithelium neurogenesis[6]. Therefore, it was conceivable that anosmia was a novel phenotype associated withSIX1mutations with incomplete penetrance, which is quite frequent in otherSIX1-associated BOR phenotypes[4,15]. However, we do acknowledge that sequencing of only known causative genes for deafness and anosmia may generate a bias, and the anosmia phenotype in the four family members may result from an independent, as-yet-unknown genetic cause.

To conclude in this study, a novel p.Tyr129His variant inSIX1had been identified as the probable cause of the dominant, delayed-onset hearing loss in Family P-D26. In addition, anosmia may be a new phenotype associated withSIX1mutations with incomplete penetrance, which remains to be investigated by replication in other studies.

Author and ContributionsYANG Tao, PANG Xiu Hong, CHAI Yong Chuan, and CHU Jiu Sheng were involved in the conception and design of work, data acquisition and interpretation, manuscript drafting and revision. ZHENG Xiao Yong, ZHENG Hao, and LIU Dong were involved in samples collection. LIN Yun,ZHANG Yu Ting, and ZHANG Lu Ping were involved in physical examination. LIN Yun, XU Jun, and JIN Chun Yan were involved in experiment, data acquisition,and analysis. All authors read and approved the final manuscript.

AcknowledgmentsWe thank all family members who enrolled in this study.

&These authors contributed equally to this work.

#Correspondence should be addressed to YANG Tao,Professor, PhD, Tel: 86-21-38452508, E-mail: yangtfxl@sina.com; CHAI Yong Chuan, Assistant Director Physician,PhD, Tel: 86-21-38452508, E-mail: cycperfect@163.com;CHU Jiu Sheng, Director Physician, Master, Tel: 86-523-86361471, E-mail: taizhoucjs@163.com

Biographical notes of the first authors: PANG Xiu Hong, female, born in 1977, PhD, Assistant Director Physician, majoring in molecular biology of deafness;ZHENG Xiao Yong, male, born in 1977, Master, Assistant Director Physician, majoring in molecular biology of hereditary skin diseases; LIN Yun, female, born in 1989,Master, Technician, majoring in molecular biology of deafness.

Received: June 15, 2020;

Accepted: November 6, 2020

Supplementary Table S1. Previously reported causative genes for deafness, olfactory and gustatory disorders that were screened in the current study

Supplementary Table S2. Pathogenic prediction by computational tools

Supplementary Figure S1. Schematic diagram of the SIX1 structure with locations of the previously and currently reported pathogenic mutations. Mutations in gray and black indicates those associated with BO syndrome (BOS3) and non-syndromic deafness DFNA23, respectively. The p.P249L mutation is suspected to be associated with BOR with uncertainty.