Novel Sequence Variants in the TMC1 Gene in Pakistani Families With Autosomal Recessive Hearing Impairment

Abstract

Though many hearing impairment genes have been identified, only a few of these genes have been screened in population studies. For this study, 168 Pakistani families with autosomal recessive hearing impairment not due to mutations in the GJB2 (Cx26) gene underwent a genome scan. Two-point and multipoint parametric linkage analyses were carried out. Twelve families had two-point or multipoint LOD scores of 1.4 or greater within the transmembrane cochlear expressed gene 1 (TMC1) region and were subjected to further screening with direct DNA sequencing. Five novel putatively functional non-synonymous sequence variants, c.830A>G (p.Y277C), c.1114G>A (p.V372M), c.1334G>A (p.R445H), c.2004T>G (p.S668R) and c.2035G>A (p.E679K), were found to segregate within seven families, but were not observed in 234 Pakistani control chromosomes. The variants c.830A>G (p.Y277C), c.1114G>A (p.V372M) and c.1334G>A (p.R445H) occurred at highly conserved regions and were predicted to lie within hydrophobic transmembrane domains, while non-synonymous variants c.2004T>G (p.S668R) and c.2035G>A (p.E679K) occurred in extracellular regions that were not highly conserved. There is evidence that the c.2004T>G (p.S668R) variant may have occurred at a phosphorylation site. One family has the known splice site mutation c.536 -8T>A. The prevalence of non-syndromic hearing impairment due to TMC1 in this Pakistani population is 4.4% (95%CI: 1.9, 8.6%). The TMC1 protein might have an important function in K⁺ channels of inner hair cells, which would be consistent with the hypothetical structure of protein domains in which sequence variants were identified.

INTRODUCTION

Due to the highly complex structure and function of the human inner ear, it is not surprising that sensorineural hearing impairment (HI) is genetically heterogeneous. Currently, 37 non-syndromic hearing impairment (NSHI) genes have been identified, of which 21 are associated with autosomal recessive (AR) NSHI (Van Camp G, Smith RJH. Hereditary Hearing Loss Homepage. URL: http://webhost.ua.ac.be/hhh/ Accessed 1 March 2005). However, of the ARNSHI genes, only a few [e.g. GJB2 (MIM# 121011), CDH23 (MIM# 605516) and WFS1 (MIM# 606201)] have been well characterized in terms of prevalence and sequence variants found in different populations. Additional studies of genes that are involved in NSHI within various populations are necessary to better understand their prevalence, spectrum of sequence variants and public health significance. In this study, TMC1 (MIM# 606706) was studied within the Pakistani population.

A total of 168 Pakistani families with ARNSHI that were negative for GJB2 mutations underwent a 10cM genome scan. Twelve families had two-point or multipoint LOD scores of ≥1.4 within the 9q21.13 region where TMC1 is physically mapped. These families were selected for sequencing of the TMC1 gene. Thereafter the prevalence of TMC1 variants in this population was estimated. Also, by locating the residues at predicted transmembrane domains and studying evolutionary conservation in multiple sequence alignment, the possible effects on the TMC1 protein product of both novel and previously reported sequence variants were examined.

MATERIALS AND METHODS

Ascertainment of study subjects

The study was approved by the Quaid-I-Azam University Institutional Review Board and by the Institutional Review Board for Human Subject Research for Baylor College of Medicine and Affiliated Hospitals. Informed consent was obtained from all family members who participated in the study.

For this study, 192 unrelated Pakistani families with at least two ARNSHI individuals were ascertained from various regions of Pakistan. Medical and family history and information on pedigree structure was obtained from multiple family members. Pure tone audiometry at 250–8000 Hz was performed for selected subjects. All hearing-impaired family members underwent physical examination. Within these families, no clinical features that would indicate that the HI was part of a syndrome were observed. In addition, no gross vestibular involvement was noted. The HI phenotype was prelingual, severe to profound, and was not known to be caused by inflammatory middle ear disease or specific environmental factors. Additionally, 117 unrelated hearing individuals without a family history of hearing impairment were ascertained from Pakistan.

Genome scan

DNA was isolated from venous blood samples following a standard protocol (Grimberg et al., 1989), quantified by spectrophotometry at optical density 260, and stored at −20°C. Of the 192 families, twelve families were positive for GJB2 (GenBank accession # NM_004004.3) mutations. Twelve other families were not included for genome scan due to an insufficient number of DNA samples in order to carry out a linkage analysis. DNA samples from 126 families were diluted to 40 ηg/μl and sent to the Center for Inherited Disease Research (CIDR; URL: http://www.cidr.jhmi.edu/) for genome scan, while diluted DNA samples from 42 families were sent to the National Heart, Lung and Blood Institute (NHLBI) Mammalian Genotyping Service (Center for Medical Genetics, Marshfield, WI, USA; URL: http://research.marshfieldclinic.org/genetics/) for genotyping. From 2002 to 2004, samples were sent in six batches, with an average of 395 short tandem repeat (STR) markers spaced at ~10cM apart for each genome scan that was done on all 22 autosomes and the X and Y chromosomes.

Linkage analysis

Linkage analyses were performed under a fully penetrant autosomal recessive model with a disease allele frequency of 0.001. The MLINK program of the FASTLINK computer package was utilized for two-point linkage analysis (Cottingham et al., 1993), while multipoint analysis was performed using ALLEGRO (Gudbjartsson et al., 2002) with map distances from the Marshfield genetic map (Broman et al., 1998). Some of the families were too large to analyze in their entirety using ALLEGRO and were therefore broken into two or more branches for the analysis, then the LOD scores from each branch of the family were summed. Marker allele frequencies were estimated from the data by means of both observed and reconstructed genotypes of founders from each pedigree and the other pedigrees in the same genome scan.

Sequencing TMC1

Polymerase chain reaction (PCR) primers were designed for 24 exons plus 1000 base pairs of the promoter region of TMC1 (GenBank accession # NM_138691.2) via Primer3 software (Rozen and Skaletsky, 2000). DNA from one unaffected and two affected individuals from each family were diluted to 5ηg/μl, amplified, and purified with ExoSAP-IT^® (USB Corp., Cleveland, Ohio, USA; URL: http://www.usbweb.com/). Sequencing with the appropriate primers was performed with the BigDye® Terminator v3.1 Cycle Sequencing Kit together with an Applied Biosystems 3700 DNA Analyzer (Applera Corp., Foster City, CA; URL: http://www.appliedbiosystems.com/). Sequence variants were identified via Sequencher^™ Version 4.1.4 software (Gene Codes Corp., Ann Arbor, MI, USA; URL: http://www.genecodes.com/sequencher/). When a sequence variant was found, DNA samples from the rest of the family were sequenced for the exon in which the variant was identified. Similarly, 117 control individuals from Pakistan were screened for the same exon.

Protein sequence analysis

To determine the evolutionary conservation of identified substitutions, the ExPASy (Expert Protein Analysis System; URL: http://us.expasy.org/) proteomics server of the Swiss Institute of Bioinformatics (SIB) was used to look for homologues of the TMC1 protein. ExPASy uses the NCBI BLASTP 1.5.4-Paracel program (Altschul et al., 1997) to search the ExPASy/UniProt database. To perform the BLASTP search, the default settings were used, except that the threshold for expected random matches (E) was set to a more conservative value of 10⁻⁴ in order to minimize false positive results. Of the 81 query matches (score range 58–1196), one match per species was chosen for further alignment. Nine proteins, including the human TMC1 sequence, were then submitted for multiple sequence alignment via ClustalW (Thompson et al., 1994) at the European Bioinformatics Institute (EBI; URL: http://www.ebi.ac.uk/clustalw/) using default settings.

Membrane-spanning domains of the TMC1 protein were predicted through the TMHMM Server v.2.0 (Krogh et al., 2001). An independent study demonstrated that TMHMM v.2.0 had 92% true positive predictions, which were higher than all other transmembrane protein prediction programs (Moller et al., 2001). To check for motifs or patterns within the protein sequence, the PROSITE database of SIB and the European Molecular Biology Laboratory (EMBL) - EBI was scanned (Sigrist et al., 2002). The web server PolyPhen, also from EMBL, was used to assess the functional effect of identified substitutions (Ramensky et al., 2002).

RESULTS

Of 168 families that underwent two-point and/or multipoint linkage analysis, twelve families had two-point or multipoint LOD scores of 1.4 or greater between markers flanking the TMC1 gene. Table 1 shows the LOD scores per family, along with ethno-linguistic information and sequence variants that were found to segregate within families. These sequence variants were found in the homozygous state in the hearing-impaired individuals but not among unaffected family members.

Table 1.

Families screened for TMC1 mutations^*

Family	Place of origin†	Language	2-point LOD	Multipoint LOD	Sequence variant
4008	DG Khan, Punjab	Sairiki	3.1	4.0	c.2035G>A (p.E679K)
4027	Mian Wali, NWFP - Punjab border	Punjabi	2.5; 2.6 at chr.11	1.1; 2.0 at chr.11	None
4033	Sarghoda, Punjab	Punjabi	2.2; 1.2 at chr.2	2.9; 2.1 at chr.2	c.1334G>A (p.R445H)
4049	Bahwalpur, Punjab	Sairiki	4.7	4.8	c.830A>G (p.Y277C)
4070	Kashmir, AJK	Kashmiri	2.4	3.0	c.2004T>G (p.S668R)
4090	Chistian, Punjab	Punjabi	1.4; 2.1 at chr.10	1.8; −0.6 at chr.10	c.536 -8T>A‡
4119	Sadiqabad, Punjab	Sairiki	2.4	2.6	c.1114G>A (p.V372M)
4138	Kotli Kshmir, AJK	Kashmiri	2.6	3.3	c.2004T>G (p.S668R)
4156	Larkhana, Sind	Sindi	2.2; 2.3 at chr.8	3.1; 0 at chr.8	None
4160	Dhandi, Sind	Sindi	3.0	3.6	c.1114G>A (p.V372M)
4165	Taj Ghar, Punjab	Sairiki	1.5; 2.0 at chr.17	2.7; 2.8 at chr.17	None
4173	Patoki, Punjab	Punjabi	1.6; 1.7 at chr.1	1.6; 1.9 at chr.1	None

^*TMC1 (GenBank accession # NM_138691.2).

^†NWFP = Northwestern Frontier Province; AJK = Azad Jammun and Kashmir.

^‡The traditional nomenclature for c.536 -8T>A was IVS10 -8T>A.

Eight families were positive for putatively functional sequence variants. Of the identified variants, c.830A>G (p.Y277C), which occurred at a highly conserved residue, was deemed to be damaging because of a change from a hydroxyl to a sulfhydryl side chain in the amino acid that was located within the second transmembrane (TM) domain (Table 2). The c.1114G>A (p.V372M) and c.1334G>A (p.R445H) variants were also at highly conserved TM residues, but the amino acid was changed to another of the same subclass. Residues for variants c.2004T>G (p.S668R) and c.2035G>A (p.E679K) were less conserved and occurred within the extracellular region. These last four sequence changes were considered benign by the Polyphen website, though scanning through PROSITE showed that the serine residue at position 668 may belong to a protein kinase C phosphorylation site. The nucleotide change c.2004T>C removes this putative phosphorylation site. Also c.1114G>A (p.V372M) and c.2004T>G (p.S668R) were each found to segregate with the NSHI status in two families. One family has the known splice site mutation c.536 -8T>A (Kurima et al., 2002). All the reported sequence variants were not found in control subjects, who were linguistically matched for the Punjabi- and Sairiki-speaking families. Due to the limited samples from the remote Sindi and Kashmiri regions, most of the controls for the c.2004T>G variant are Punjabi, while for the c.1114G>A variant most of the controls speak Sairiki.

Table 2.

TMC1 sequence variants found to segregate within Pakistani families

Exon	Nucleotide change	Amino acid change	Domain*	Evolutionary conservation†	Predicted functional effect ‡	Allele frequencies among control chromosomes
11	c.536 -8T>A	Splice acceptor	N/A	N/A	Exon 11 skipped	0 / 234
13	c.830A>G	p.Y277C	TM2	Identical	Probably damaging	0 / 234
15	c.1114G>A	p.V372M	TM3	Conserved	Benign	0 / 234
16	c.1334G>A	p.R445H	TM4	Identical	Benign	0 / 234
21	c.2004T>G	p.S668R	EC	Non-conserved	Benign; PKC phosphorylation site§	0 / 234
21	c.2035G>A	p.E679K	EC	Non-conserved	Benign	0 / 234

^*TMHMM v.2.0 predicted six membrane-spanning domains for the TMC1 protein (Krogh et al., 2001). TM = transmembrane; EC = extracellular; N/A = not applicable.

^†Human TMC1 sequence compared with TMC1-like protein sequences of 8 species which were identified via the NCBI BLASTP (Altschul et al., 1997) program and the ExPASy/UniProt databases. Species include 1 mammal, 1 amphibian, 2 bony fishes, 3 insects and 1 nematode. Comparison done via ClustalW (Thompson et al., 1994). The serine residue at position 668 and glutamic acid at 679 are conserved in M. musculus Tmc1 and F. rubripes Tmc2-related proteins. Additionally, S668 is also conserved in A. gambiae TmcA-like protein and D. melanogaster CG3280-PA.

^‡Possible effect of amino acid substitutions on the protein structure and function were predicted via the web server PolyPhen (Ramensky et al., 2002).

^§The p.S668R sequence change was found to occur at a possible protein kinase C (PKC) phosphorylation site according to the PROSITE database (Sigrist et al., 2002).

Several polymorphisms were also observed in the twelve families (Table 3). None of these variants segregated with hearing impairment status within the families. Two of the variants were non-synonymous substitutions, but were also considered benign polymorphisms because they did not segregate with HI status. These substitutions also occurred at non-conserved non-transmembrane residues.

Table 3.

TMC1 polymorphisms that do not segregate within families

Exon	Nucleotide Change
1	g.74G>A (-467 from the ATG)
3	g.[236 -67G>A (+) 322A>G (-219 from the ATG)]
4	g.346 -11delT
6	c.45C>T (synonymous)
8	c.237 -44G>A
8	c.241G>A (p.E81K) *
10	c.[535 +101A>G (+) 535 +106_113delCAAACAAA]
14	c.1029 +85T>C
16	c.1404 +32A>G
17	c.1457T>C (p.M486T) †
22	c.2208 +22A>G

^*p.E81K does not segregate with the HI status in three families. Additionally p.E81K is at the intracellular glutamic acid (E)-rich region and is non-conserved.

^†p.M486T was found in the heterozygous state in two hearing individuals from two different families but not among the hearing-impaired. The residue is predicted to be intracellular and is non-conserved.

Aside from c.536 -8T>A, other TMC1 sequence variants that were identified in previous research were not observed in the families in this study, but were also analyzed in terms of evolutionary conservation and occurrence at transmembrane domains and known motifs (Table 4).

Table 4.

Previously published sequence variants in the TMC1 gene^*

Exon	Nucleotide Change	Protein Change	Domain	Evolutionary Conservation	Predicted Functional Effect
Intron 3–5	IVS3_IVS5 del27kb†	Transcription initiation site removed	N/A	N/A	Exons 4 and 5 deleted
7	c. 100C>T	p.R34X	IC	Non-conserved	Truncated protein; Occurs at glutamic-acid rich N-terminus region; cAMP-and cGMP-dependent protein kinase phosphorylation site
8	c.295delA	Frameshift	N/A	N/A	Truncated protein
13	c.884 +1G>A‡	Splice donor	N/A	N/A	Exon 13 skipped
15	c. 1165C>T	p.R389X	EC	Non-conserved	Truncated protein
17	c. 1534C>T	p.R512X	IC	Non-conserved	Truncated protein
19	c. 1714G>A§	p.D572N	IC	Non-conserved	Benign; Casein kinase II phosphorylation site
20	c. 1960A>G	p.M654V	TM5	Non-conserved	Benign

^*TMC1 sequence changes were reported by Kurima et al. (2002) except for p.R389X at exon 15 (Meyer et al., 2005). Membrane-spanning domains, evolutionary conservation and functional effect were determined as described in Table 2.

^†Because the article did not specify the exact position at which this deletion occurred, it was not possible to rename this sequence variant according to the current nomenclature recommendations.

^‡The traditional nomenclature for c.884 +1G>A was IVS13 +1G>A.

^§c.1714G>A (p.D572N) was observed in a North American family with dominant hearing impairment.

DISCUSSION

Putatively functional TMC1 variants were observed in eight of the 168 Pakistani families that underwent a genome scan. Twelve additional families in the study did not undergo a genome scan but are known to have functional variants in the GJB2 gene (Santos et al., 2005). Thus, when GJB2-affected families were included, the prevalence of TMC1 mutations in this Pakistani population was 4.4% (95%CI: 1.9, 8.6). This is a lower prevalence rate than the previously reported prevalence of 5.4 + 3.0% among consanguineous Indian and Pakistani families (Kurima et al., 2002). However, the difference in these prevalence rates is not statistically significant.

Among 12 families in which the TMC1 gene was sequenced, a potentially functional variant was not identified in four families (Table 1). Three of these families (4027, 4165 and 4173) had LOD scores of suggestive linkage (1.9–2.8) in other chromosomal regions, which may contain the gene that causes the HI in these families. In family 4156, the two-point LOD score was relatively high for chromosome 2, but the multipoint LOD score at chromosome 2 was zero, while a multipoint LOD score of 3.1 was obtained in the TMC1 region. For this family, it is possible that the functional variant(s) are within the intronic regions of TMC1, or in another gene that lies close to the TMC1 gene.

The predicted structure of TMC1 bears similarity to that of the α-subunit of voltage-dependent K⁺ channels, which has six α-helical TM segments and intracellular N and C termini (Hanlon and Wallace, 2002). The first four TM domains of the K⁺ channel α-subunit act as voltage sensors for activation gating (Li-Smerin et al., 2000), whereas the intervening segment between TM5 and TM6 appears to confer channel selectivity (Hanlon and Wallace, 2002). Three highly conserved TMC1 sequence variants in this study – c.830A>G (p.Y277C), c.1114G>A (p.V372M) and c.1334G>A (p.R445H) – lie within the central portion of hydrophobic TM segments. Meanwhile, the TMC1 variants c.2004T>G (p.S668R) and c.2035G>A (p.E679K) were found between TM5 and TM6 among a cluster of less conserved residues. Though these amino acid residues were non-conserved when human TMC1 was aligned with eight proteins from other species, alignment with other human and murine TMC proteins shows that these residues are conserved for members of the TMC subfamily A, TMC1, TMC2 and TMC3 (Keresztes et al., 2003; Kurima et al., 2003). Additionally, the serine residue at position 668 is a putative protein kinase C phosphorylation site. Therefore, even though the Polyphen server labeled four of five novel substitutions as benign, we believe that all five variants are functional. It should be noted that the Polyphen server successfully predicts only 82% of mutations in the SwissProt protein database (Ramensky et al., 2002).

The deafness (dn) mouse is the homologous model for recessive TMC1 mutations, while the Beethoven (Bth) mouse is identified with dominant TMC1 mutations that cause postlingual hearing impairment (Kurima et al., 2002; Vreugde et al., 2002). In the mouse Tmc1 expression has been found to localize to cochlear hair cells (Kurima et al., 2002). In microscopic sections of dn mouse cochlea, Tmc1 mutations caused the greatest damage to inner hair cells (IHC), with progressive degeneration that begins at postnatal day 13, which is the time of onset of hearing function in mice (Vreugde et al., 2002; Pujol et al., 1983). In functional studies, the cochlear microphonic, which is a measure of the electric potential that is generated by the organ of Corti after acoustic stimulation, was never detected in dn mice from postnatal day 12 onwards (Steel and Bock, 1980; Bock and Steel, 1983). These studies further support the findings of bilateral profound prelingual hearing impairment due to recessive homozygous TMC1 mutations, and implicate the IHC as the main site of hearing dysfunction.

It was suggested that TMC1 might be an ion channel or transporter which mediates K⁺ homeostasis in the inner ear (Keresztes et al., 2003). Because K⁺ transport across hair cell membranes requires very rapid response, the use of second messenger or cascade mechanisms has been deemed highly unlikely, and the gating-spring model for mechanoelectrical transduction is largely favored (Corey and Hudspeth, 1983). This model proposed that mechanical forces brought about by bending of stereocilia and tension on the tip links directly activate ion channels. If it is true that TMC1 is an ion channel that is mainly localized in the IHC, then it might be involved in the most basic auditory process of hair-cell transduction.

In conclusion, five novel sequence variants of the TMC1 gene for hearing impairment were identified in this study. These variants were found in TMC1 regions that might be critical to function of K⁺ channels in hair cells. In the highly consanguineous Pakistani population, the prevalence of putatively functional TMC1 variants was found to be low. However, the true impact of genetic defects in TMC1 cannot be known until more studies are done in other populations.