Article Text


Pathogenic mutations but not polymorphisms in congenital and childhood onset autosomal recessive deafness disrupt the proteolytic activity of TMPRSS3
  1. Y J Lee,
  2. D Park,
  3. S Y Kim,
  4. W J Park
  1. Department of Life Science and National Research Laboratory of Proteolysis, Kwangju Institute of Science and Technology (K-JIST), Kwangju, Korea
  1. Correspondence to:
 Dr W J Park, Department of Life Science, Kwangju Institute of Science and Technology (K-JIST), 1 Oryong-dong, Puk-gu, Kwangju, 500-712 Korea; 

Statistics from

Congenital hearing loss occurs in approximately 1 in 1000 live births and 60% of these cases are hereditary.1,2 Non-syndromic autosomal recessive deafness accounts for about 70% of congenital hereditary hearing loss cases. To date, at least 33 genetic loci have been mapped for non-syndromic deafness, and the causative genes for 17 of these loci have been identified.3 Mutations in the TMPRSS3 gene, which encodes a transmembrane serine protease, were originally identified in Pakistani DFNB8 and Palestinian DFNB10 families, and later in two Tunisian families and several white patients.4–9

TMPRSS3 belongs to a family of transmembrane serine proteases, which includes TMPRSS1-5.5,10 Like the other members of this family, TMPRSS3 contains a short amino terminus, a transmembrane domain and a large extracellular or lumenal carboxyl terminal segment characterised by a stem region containing LDLRA (low density lipoprotein receptor class A), SRCR (scavenger receptor cysteine rich) domains, and a catalytic domain. In addition to a deleterious β-satellite repeat insertion and splice acceptor site substitution,6 six pathogenic missense mutations have been isolated.7 Since these mutations occur not only in the catalytic domain, but also in the LDLRA and SRCR domains involved in interactions with extracellular molecules, the molecular mechanism for pathogenesis is unclear. In this study, we determine whether these missense mutations affect the proteolytic activity of TMPRSS3.

Key points

  • Mutations in the TMPRSS3 gene, which encodes a transmembrane serine protease, are responsible for non-syndromic autosomal recessive deafness, DFNB8 and 10. Pathogenic mutations were found in the LDLRA (low density lipoprotein receptor class A) and SRCR (scavenger receptor cysteine rich) domains as well as the serine protease domain.

  • We have assayed the proteolytic activity of the TMPRSS3 proteins containing pathogenic mutations or polymorphisms using a yeast based protease assay.

  • All the six tested pathogenic missense mutations disrupted the proteolytic activity of TMPRSS3, while two non-pathogenic polymorphisms did not affect the activity.

  • The disruption of proteolytic activity of TMPRSS is tightly correlated with the pathogenesis of deafness.


Generation of wild type and mutant TMPRSS3 constructs

The lumenal region of TMPRSS3 (accession number AB038157) was amplified by PCR and fused to the truncated STE13. The PCR primers used in this amplification (5′-ATG CCT CGA GTT CGA CTG CTC AGG GAA GTA C-3′ and 5′-ATG CGC GGC CGC TCA GGT TTT TAG GTC TCT CTC-3′) was also used for generation of mutated TMPRSS3. Missense mutations and polymorphisms were introduced by PCR based mutagenesis.11 Primer sets containing the nucleotide alterations are as follows: D103G (5′-CAA AGA CGG GGA GGG CGA GTA CCG CTG TG-3′ and 5′-CAC AGC GGT ACT CGC CCT CCC CGT CTT TG-3′), R109W (5′-CCG CTG TGT CTG GGT GGG TGG-3′ and 5′-CCA CCC ACC CAG ACA CAG CGG-3′), C194F (5′-AGG GAG GGA TTT GCC TCT GGC-3′ and 5′-GCC AGA GGC AAA TCC CTC CCT C-3′), W251C (5′-CAT CAC GCC CCT GTG CAT CAT CAC TGC TG-3′ and 5′-CAG CAG TGA TGA TGC ACA GGG GCG TGA TG-3′), P404L (5′-GAC AGC GGG GGG CTC CTG GTG TGT CAA G-3′ and 5′-CTT GAC ACA CCA GGA GCC CCC CGC TGT C-3′), C407R (5′-GCC CCT GGT GCG TCA AGA GAG G-3′ and 5′-CCT CTC TTG ACG CAC CAG GGG C-3′), G111S (5′-CGC TGT GTC CGG GTG AGT GGT CAG AAT GCC GTG-3′ and 5′-CAC GGC ATT CTG ACC ACT CAC CCG GAC ACA GCG-3′), I253V (5′-ACG CCC CTG TGG ATC GTC ACT GCT GCA CAC TGT-3′ and 5′-ACA GTG TGC AGC AGT GAC GAT CCA CAG GGG CGT-3′), D173N (5′-GAG TTT GTG TCC ATC AAT CAC CTC TTG CCA GAT-3′ and 5′-ATC TGG CAA GAG GTG ATT GAT GGA CAC AAA CTC-3′), and A426T (5′-TTT GGC ATC GGC TGC ACA GAG GTG AAC AAG CCT-3′ and 5′-AGG CTT GTT CAC CTC TGT GCA GCC GAT GCC AAA-3′).

Yeast based protease assay

Proteolytic activity of TMPRSS3 and its variants were assayed as described previously.12 Briefly, a yeast strain KSY01 (MATα, leu2 ura3 his3 trp1 lys2 suc2-Δ9 kex2::HIS3) was cotransformed with the wild type and mutant TMPRSS3 expression vectors and the substrate vector. The Leu+/Trp+ transformants were selected on minimal media containing 2% glucose but lacking Leu and Trp, and then replica plated onto YPD media containing 2% sucrose and 10 μg/ml antimycin A. Colonies usually appeared in seven days at 30°C.


In this study, we determined whether the missense mutations affect the proteolytic activity of TMPRSS3 by using a yeast based protease assay, designated sGASP (secretory Genetic Assay for Site specific Proteolysis), which is effective for the study of secretory proteases.12 The principle of sGASP is illustrated in fig 1A and B. Using this method, proteolysis can be monitored simply by the growth of yeast cells on selective plates. Assays were performed with the substrate sequence VNLNSSRQSR_IVGGE and the extracellular domain of TMPRSS3 expressed as a fusion protein with STE13 (fig 1C). The structure of TMPRSS3, locations of pathogenic missense mutations, and polymorphisms are depicted in fig 2A and the results of assays are shown in fig 2B. The substrate sequence alone was not cleaved (data not shown), but cleavage was evident upon coexpression with wild type TMPRSS3, as shown by the growth of transformants on sucrose plates (panel I). The mutation in the potential activation cleavage site, R216A, greatly diminished the proteolytic activity, and the mutations in the serine residue of the catalytic triad, S401A, completely abolished the activity, as evaluated by the failure of yeast cells transformed with these mutants to grow on sucrose plates (panels II and III). Pathogenic missense mutations (D103G, R109W, C194F, W251C, P404L, and C407R), were all defective in protease activity (panels IV-IX). In contrast, the two polymorphisms G111S and I253V did not affect the proteolytic activity of TMPRSS3 (panels X and XI). Our data indicate that the pathogenesis is strongly correlated with the defective proteolytic activity of TMPRSS3.

Figure 1

Principle of sGASP. In a yeast strain lacking invertase activity (suc2), a fusion protein is expressed in which invertase is linked to the truncated lumenal domain of an integral Golgi membrane protein, STE13, by a short substrate sequence containing linker. In the absence of cleavage of the substrate sequence, the invertase moiety remains anchored to the Golgi membrane (A). However, upon cleavage of the substrate sequence by a specific protease, invertase is released into the periplasmic space where it degrades sucrose to glucose and fructose. As a result, transformants are able to grow on selective plates containing sucrose as the sole carbon source (B). The structures of the STE13-substrate-invertase fusion protein and STE13-TMPRSS3 are shown in (C). Targeting proteases to Golgi apparently augments the protease-substrate interaction and hence the proteolysis.

Figure 2

Protease assays for pathogenic mutations and polymorphic alterations. (A) Schematic representation of TMPRSS3 with pathogenic mutations and polymorphisms. (B) Transformants expressing both the STE13-substrate-invertase fusion protein and STE13-TMPRSS3 (wild type or variants) were plated on non-selective (glucose) and selective (sucrose) plates. I, wild type TMPRSS3; II, R216A; III, S401A; IV, D103G; V, R109W; VI, C194F; VII, W251C; VIII, P404L; IX, C407R; X, G111S; XI, I253V; XII, D173N; XIII, A426T. The numbers in parentheses are plating efficiencies, which represent relative proteolytic activity. They are calculated by dividing the number of colonies on sucrose plates by the number of colonies on glucose plates.

We additionally examined two reported polymorphisms, D173N and A426T, which could not be clearly categorised as non-pathogenic.9 Our results showed that the D173N mutant possesses full proteolytic activity, while that of A426T is significantly diminished (panels XII and XIII). We suggest that D173N is a non-pathogenic polymorphism, while A426T could be pathogenic in certain circumstances (for example, when the critical substrates of TMPRSS3 become slightly less cleavable owing to gene alterations, the A426T mutation in TMPRSS3 could worsen the situation and eventually lead to pathogenesis). It is interesting to note that D173 is not conserved, while A426 is highly conserved among TMPRSS proteases.

Recently, the epithelial amiloride sensitive sodium channel (ENaC) was suggested to be a potential substrate of TMPRSS3.13 It was suggested that TMPRSS3 proteolytically activates ENaC, which might control important signalling pathways in the inner ear. Consistent with this report, our data indicate that disruption of the proteolytic activity of TMPRSS3 is tightly correlated with the pathogenesis of hearing loss. It remains to be seen how the mutations in the LDLRA and SRCR domains affect the proteolytic activity of TMPRSS3. It may be possible that these domains are necessary for proper folding or assembly of the catalytic domain or protease substrate recognition and binding.


The first two authors contributed equally to this work. This work was supported by grants from the National Research Laboratory programme (M1-0104-00-0007) of the Ministry of Science and Technology and the BK21 programme of the Ministry of Education, Korea.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.