Non-USH2A mutations in USH2 patients
We have systematically analyzed the two known minor genes involved in Usher syndrome type 2, DFNB31 and GPR98, for mutations in a cohort of 31 patients not linked to USH2A. PDZD7, an Usher syndrome type 2 (USH2) related gene, was analyzed when indicated. We found that mutations in GPR98 contribute significantly to USH2. We report 17 mutations in 10 individuals, doubling the number of GPR98 mutations reported to date. In contrast to mutations in usherin, the mutational spectrum of GPR98 predominantly results in a truncated protein product. This is true even when the mutation affects splicing, and we have incorporated a splicing reporter minigene assay to show this, where appropriate. Only two mutations were found which we believe to be genuine missense changes. Discrepancy in the mutational spectrum between GPR98 and USH2A is discussed. Only two patients were found with mutations in DFNB31, showing that mutations of this gene contribute to only a very small extent to USH2. Close examination of the clinical details, where available, for patients in whom no mutation was found in USH2A, GPR98, or DFNB31, showed that most of them had atypical features. In effect, these three genes account for the vast majority of USH2 patients and their analysis provide a robust pathway for routine molecular diagnosis.
Mutations in USH2 Patients
CHU Montpellier, Laboratoire de G
eculaire, Montpellier, France;
INSERM U827, Montpellier, France;
Univ, Montpellier I,
CHU Montpellier, Centre National de R
erence maladies rares “Affections Sensorielles G
etiques”, Montpellier, France;
opital Sud, Service de G
edicale, Rennes, France;
CHU Montpellier, Service de G
edicale, Montpellier, France;
opital Purpan, Service de G
edicale, Toulouse, France;
Centre de R
erence pour les Affections Rares en G
Ophtalmologique (CARGO), H
opitaux Universitaires de Strasbourg, Strasbourg, France;
CHU Edouard Herriot, Service d’ORL, Lyon, France;
edicale, Chambery, France;
CHU de Nantes, Service de G
edicale, Nantes, France;
CHU Amiens, Service de
etique Clinique, Amiens, France;
Clinical and Molecular Genetics, Institute of Child Health, University College London, United Kingdom
Communicated by Andreas Gal
Received 16 September 2011; accepted revised manuscript 28 November 2011.
Published online 2 December 2011 in Wiley Online Library (www.wiley.com/humanmutation).DOI: 10.1002/humu.22004
ABSTRACT: We have systematically analyzed the two
known minor genes involved in Usher syndrome type 2,
DFNB31 and GPR98, for mutations in a cohort of 31 pa-
tients not linked to USH2A. PDZD7, an Usher syndrome
type 2 (USH2) related gene, was analyzed when indicated.
We found that mutations in GPR98 contribute signiﬁ-
cantly to USH2. We report 17 mutations in 10 individu-
als, doubling t he number of GPR98 mutations reported to
date. In contrast to mutations in usherin, the mutational
spectrum of GPR98 predominantly results in a truncated
protein product. This is true even when the mutation af-
fects splicing, and we have incorporated a splicing reporter
minigene assay to show this, where appropriate. Only two
mutations were found which we believe to be genuine mis-
sense changes. Discrepancy in the mutational spectrum
between GPR98 and USH2A is discussed. Only two pa-
tients were found with mutations in DFNB31, showing
that mutations of this gene contribute to only a very small
extent to USH2. Close examination of the clinical details,
where available, for patients in whom no mutation was
found in USH2A, GPR98,orDFNB31, showed that most
of them had atypical features. In effect, these three genes
account for the vast majority of USH2 patients and their
analysis provide a robust pathway for routine molecular
Hum Mutat 33:504–510, 2012.
2011 Wiley Periodicals, Inc.
KEY WORDS: Usher syndrome; GPR98; DFNB31; func-
tional analysis; molecular analysis
Additional Supporting Information may be found in the online version of this article.
Correspondence to: Anne-Franc¸oise Roux, Laboratoire de G
CHU Montpellier, INSERM U827, IURC, 641 Avenue du Doyen Gaston Giraud, F-34093
Montpellier Cedex 5, France. E-mail: firstname.lastname@example.org
Usher syndrome is the most common form of hereditary syn-
drome associating sensorineural hearing loss (HL) and retinitis pig-
mentosa (RP). The prevalence has long been estimated to be 1 of
17,000 to 1 of 33,000 but recent studies suggest a much higher fre-
quency of 1 of 6,000 [Kimberling et al., 2010]. This group of recessive
disorders is divided into three clinical subtypes deﬁned according to
the degree of HL and the presence or not of vestibular dysfunction.
To date, nine causative genes have been identiﬁed (for review see
Saihan et al., ).
Usher syndrome type 2 (USH2) is the most frequent form and
accounts for more than half of the cases (for review see [Petit,
2001]). Patients present with congenital moderate to severe HL
and normal vestibular function [Abadie et al., 2011]. RP usually
develops from the second decade on. Currently, three genes are
known to be responsible for USH2: USH2A (MIM# 608400), GPR98
(MIM# 602851), and DFNB31/WHRN (MIM# 607928). Although
USH2A has been extensively studied in terms of mutational screen-
ing over the last 6 years, data on GPR98 and DFNB31 remain scarce
(see LOVD-USHbases: https://grenada.lumc.nl/LOVD2/Usher_
montpellier/). USH2A is by far the most frequently implicated gene
and pathogenic mutations are identiﬁed in approximately 75–80%
of cases [Baux et al., 2007]. The prevalence of mutations in each
of the GPR98 and DFNB31 genesisasyetunknown.GPR98 (also
known as VLGR1) was ﬁrst described as implicated in USH2 in
2004 [Weston et al., 2004] with the identiﬁcation of four causative
mutations in ﬁve unrelated female patients. However, only a limited
number of USH2-related deleterious mutations (n
= 17) is known to
date (see USHbases), possibly because molecular testing of this gene,
which contains 90 exons remains challeng ing and cannot be offered
routinely. The DFNB31 gene was ﬁrst identiﬁed as being respon-
sible for nonsyndromic deafness in two s eparate families [Mburu
et al., 2003]. Its involvement in Usher syndrome was demonstrated
by Ebermann et al. , in a family presenting with typical USH2
syndrome. Because both identiﬁed mutations altered the N termi-
nal portion of the whirlin long isoform, the authors suggest that
two PDZ(Post-synaptic density protien 95, Drosophila disk-large
2011 WILEY PERIODICALS, INC.
tumor supressor, Zona occludens-1 protien) domains, PDZ1 and
PDZ2, involved with binding to other USH2 proteins (usherin and
G protein-coupled receptor 98) are not only cr itical for auditory but
also for retinal function.
The three USH2 proteins belong to an Usher interactome. PDZ
domains of whirlin interact with the PDZ-binding motif of both
USH2 transmembrane (TM) proteins (G protein-coupled receptor
98 and usherin). These proteins are major components of the ankle-
link of the hair bundle of the hair cells cochlea [Michalski et al., 2007]
and play an anchoring role in the connecting cilium with the inner
segment of the photoreceptors [Maerker et al., 2008; van Wijk et al.,
About 20% of the patients referred to our laboratory are not
linked to USH2A [(Baux et al., 2007] and unpublished results),
which represents a signiﬁcant proportion of unresolved cases. To
improve our diagnostic service we have tested the involvement of
GPR98 and DFNB31 genes and established the mutational spec-
trum of these two genes. We show that similarly to USH2A,the
whole GPR98 gene needs to be screened. We also conﬁrm that,
integrity of the N-terminal region (PDZ1 and PDZ2 domains) of
the whirlin long isoform is crucial. We performed functional analy-
sis to determine the outcome of the predicted aberrant pre-mRNA
Finally, because PDZD7 has been described as a contributor to
digenism in Usher syndrome and as a modiﬁer of the retinal phe-
notype [Ebermann et al., 2010], its involvement was tested where
Patients and Methods
Patients were recruited from all over France. All patients were
referred with a clinical diagnosis of Usher type 2 based on the degree
of HL and RP. In most cases, audiograms and electroretinograms
When possible, blood samples of relatives were collected and used
for linkage and segregation analysis.
This study was approved by the local Ethical Committee and
consent to genetic testing was obtained from adult probands or
parents in the case of minors.
Haplotype Analyses of the USH2 Genes
Microsatellites markers were developed at the three different
USH2 loci: USH2A (USH2A gene), USH2C (GPR98 gene), and
USH2D (DFNB31 gene) (see Supp. Fig. S1). Most of the markers
correspond to Genethon markers. However, some additional intra-
genic markers were identiﬁed in the laboratory at the USH2A locus
in introns 13, 52, and 70 of the gene, which proved to be more
informative. Their primer sequences are listed in Supp. Figure S1.
PCR Sequencing of the Genes
Primers were designed to amplify all the coding exons and their
ﬂanking intronic sequences of the GPR98 (90), DFNB31 (12), and
PDZD7 (16) genes. They are listed in Supp. Table S1. Most of them
ampliﬁed with a uniform PCR condition on a single 96-well plate
(adapted from [Roux et al., 2006]). The same primers were used
for sequencing, except in a few cases, for which new design of in-
ternal primers improved the sequence quality. Sequencing puriﬁca-
tion was carried out with Agencourt CleanSeq on a Biomek 3000
Laboratory Automation Workstation (Beckman Coulter, Villepinte,
France), following the manufacturer’s recommendations or by us-
ing sephadex (Sephadex G-50 S, GE Healthcare, V
Sequences were run on an Applied Biosystems 3130xl Genetic Ana-
lyzer (Applied Biosystems, Courtaboeuf, France) and analyzed with
the ABI Prism SeqScape V2.5 software.
Nomenclature of the variants follows the HGVS (http://www.
hgvs.org/rec.html) recommendations with nucleotide
sponding to the A of the ATG initiation codon. Genotypes are given
according to HGVS nomenclature.
Large Rearrangement Identiﬁcation
Detection of large genomic rearrangements was performed on
a custom designed CGH-microarray chip (12 × 135 k), as already
described [Roux et al., 2011].
In Silico Analysis
The pathogenic effect of missense variants was predicted as pre-
viously described [Baux et al., 2007] following a multistep anal-
ysis [Roux et al., 2011] and is available in Usher Syndrome Mis-
sense Analysis ( USMA) (https://neuro-2.iurc.montp.inserm.fr/cgi-
bin/USMA/USMA.fcgi). In addition, alignments of orthologs are
accessible in USMA.
Prediction of a splice effect was tested using the Human Splicing
Finder (HSF) tool which gives a position Weight matrices-derived
scores (HSF) for potential splice sites, and also integrates the results
from the MaxEnt program [Desmet et al., 2009; Yeo and Burge,
Splicing Variant Analysis
The potential impact upon splicing of missense and intronic un-
known variants was functionally established as previously described
edard-Mereuze et al., 2010] and using the pSPL3 exon-
trapping vector [Bottillo et al., 2007]. Brieﬂy, patients’ genomic
DNA was used as a template to generate wild ty pe and mutated in-
serts with the Phusion
High ﬁdelity DNA polymerase (Finzymes,
Espoo, Finlande). Ampliﬁed products were cloned between the NheI
and XhoI restriction sites of pSPL3 vector. Ligation was obtained
Dry-Dow n PCR cloning kit according to manu-
facturer’s instructions (Clontech, Saint-Germain-en-Laye, France).
Constructs were transiently transfected into HeLa cells and total
RNA was extracted 48 hr after transfection; Reverse Transcription-
PCR(RT-PCR) and vizualization of splicing alterations were per-
formed as already published [Le Gu
edard-Mereuze et al., 2010].
GenBank reference sequences: GPR98: NM_32119.3; DFNB31:
NM_015404.2; PDZD7: NM_001195263.1.
UNIPROT: G-protein coupled receptor 98: Q8WXG9; Whirlin:
Among the 31 patients analyzed, a likely pathogenic genotype
was identiﬁed in 12 of them, 10 in GPR98 and 2 in DFNB31
HUMAN MUTATION, Vol. 33, No. 3, 504–510, 2012 505
Table 1. Pathogenic
Genotypes and Clinical Findings in USH2 Patients
Patient Degree of HL Onset of RP (years) Ophthalmologic ﬁndings Genotyp e
U229 Moderate 17 Typical RP c.[14365C
U347 Severe 20 ND c.[7770delC(;)17204
U361 Severe 47 Rod-cone dystrophy c.[3945dupA];[17933A
U380 Moderate 35 Typical RP c.[10935_10938del(;)9042G
U416 Moderate 13 Typical RP c.[17668_17669del(;)17020-?_17856
U457 Severe 30 ND c.[13320dupC(;)16940delT]
Severe 40 Typical RP c.[5671A
U787 Moderate 35 Typical RP c.[17062C
Moderate 20 Typical RP c.[2984_2988del(;)2984_2988del]
U919 Moderate 35 ND c.[7001T
Moderate 30 Typical RP c.[680dupG];[737delC]
Moderate 30 Typical RP c.[737delC];[737delC]
HL, hear ing loss; RP, retinitis pigmentosa; ND, not documented.
Sequencing was performed in the three USH2 genes for all the listed patients with the exception of U734, U831, U797, and U922. For these four patients preliminary haplotypes
analyses were performed for the three USH2 loci.
Exclusion of USH2A and USH2D loci.
Exclusion of USH2A and USH2C loci.
Homozygosity identiﬁed for the targeted locus.
GenBank accession numbers:GPR98 : NM_32119.3; DFNB31 : NM_015404.2.
Clinical ﬁndings are indicated for each patient. Patients were mainly Caucasians; U831 and U922 were of Turkish and Portuguese origin, respectively.
Table 2. List of Likely Pathogenic Variants Identiﬁed in the
Exon/intron Nucleotide exchange Predicted translation effect Classiﬁcation Prediction of splice effect Allelic frequency
16 c.2984_2988del p.(Leu995fs) Pathogenic No
Intron 19 c.3635-2A
20 c.3945dupA p.(Gln1316fs) Pathogenic No
) Pathogenic No
) Pathogenic No
33 c.7770delC p.(Glu2591fs) Pathogenic No
52 c.10935_10938del p.(Ser3646fs) Pathogenic No
Intron 65 c.13232-3C
66 c.13320dupC p.(Ser4441fs) Pathogenic No
T p.(Arg4789Trp) UV3 No 0/176
78 c.16940delT p.(Val5647fs) Pathogenic No
) Pathogenic No
Intron 79 c.17204
7del p.(?) UV4 Yes
82 c.17668_17669del p.(Met5890fs)
G p.(His5978Arg) UV3 No 0/170
?dup p.(Ile5674_Gln5952dup) UV4 ?
2 c.680dupG p.(Tyr228fs) Pathogenic No
2 c.737delC p.(Pro246fs)
For missense variants, allelic frequency was estimated in control chromosomes.
Splicing alteration as demonstrated by further ex vivo splicing assays.
Already described (see USHbases).
In 4 of these 12 families because consanguinity was documented
(U831, U922) or as several affected siblings were available (U734,
U797), haplotype analyses had previously been performed at the
three USH2 loci to prioritize which gene to sequence (Table 1).
Thus, GPR98 sequencing was prioritized in family U734 as USH2A
and USH2D loci had been excluded in the three affected siblings,
while DFNB31 sequencing was performed directly for family U797
after exclusion in the three affected siblings of USH2A and USH2C
loci. In the two consanguineous families U831 and U922, GPR98 and
DFNB31 weresequenced due to homozygosity revealedby haplotype
analyses at the USH2C and USH2D locus, respectively.
Ten variants were considered “a priori” deleterious in GPR98
as they predict a premature termination codon (PTC) leading to a
truncated G protein-coupled receptor 98 (three nonsenses, ﬁvesmall
deletions, and two small duplications [Tables 1 and 2]). Three sub-
stitutions predicting so-called missense alterations were classiﬁed
as likely pathogenic after the multistep analysis already described
[Roux et al., 2011]; however, one of them lying at the last nucleotide
of exon 41 (c.9042G
C) was eventually classiﬁed as a splicing alter-
ation (see below).
506 HUMAN MUTATION, Vol. 33, No. 3, 504–510, 2012
Figure 1. Ex vivo analysis of four splicing identiﬁed variants in
. Schematic minigene constructs of the region of interest (A) and gel
electrophoresis of RT-PCR products obtained from the assay (B) for (1–1) c.9042G>C in exon 41; (1–2) c.17204+4_17204+7del in intron 79; (1–3)
c.13232-3C>G in intron 65 and for (1–4) c.3635-2A>G in intron 19. Splicing patterns obtained are speciﬁed on the right-hand of each agarose gel.
Underlined nucleotides represent the alteration; the deletion of the four bases is indicated with hyphens (1–2). MaxEnt scores are displayed for
reference sites wild type (Wt), mutated (Mut), and for other used and identiﬁed by sequencing of the RT-PCR of the constructs. For 1 –3, HSF
score is indicated for the predicted de novo acceptor 3
ss. In Wt constructs 1–1, 1–3, 1–4, traces of exon skipping were detected by sequencing.
)Alternative splice sites used in Wt constructs: (1–2) a cryptic acceptor splice site leads to the retention of 12 intronic nucleotides; (1–4) the use
of an exonic cryptic donor site results in the exclusion of the last 697 bases of exon 20.
Three additional putative splicing alterations were found in in-
trons and likely to alter the strength of consensus splice sites. Finally,
screening for large genomic rearrangements [Roux et al., 2011] was
performed in cases where only a single deleterious mutation had
been identiﬁed after sequencing (U416, U787), and one large du-
plication encompassing several exons was detected (U416).
Because the presence of large rearrangement was excluded in pa-
tient U787, further PDZD7 gene investigation was performed in
order to i dentify the second alteration, considering the fact that
this gene has been described as a contributor to digenism in Usher
syndrome [Ebermann et al., 2010]. Sequencing of all PDZD7 ex-
ons did not reveal any mutation in combination with the GPR98
Prediction of a splice effect was tested for all the identiﬁed
variants (Table 2). For four of them, all within GPR98, a signif-
icant alteration in the score of the wild-type splice sites strength
was highlighted by the two prediction tools (Fig. 1 and Supp.
Two of the variants tested destroy the 5
C is located at the last position of exon 41 and affects
1 of the site, whereas the deletion of the four bases AGTA
7del) abolishes the
5G positions of the
splice site (ss) of exon 79 by replacing it with
four base deletion arose in an AGTA–AGTA repeat motif. The two
other base changes c.3235-2A
G and c.13232-3C
sites affecting variants but with uncertain splicing outcomes: the
results concerning their precise effects are different between both
programs. For c.3235-2A
G, the mutation d rastically reduces the
strength of the 3
ss and reinforces the strength of a downstream 3
ss according to MaxEnt (score of 3.21
6.12), while for c.13232-
G, HSF (http://www.umd.be/HSF/) but not MaxEnt indicates
the potential use of the de novo 3
ss (77.36) created by the mutation
(Supp. Table S2).
HUMAN MUTATION, Vol. 33, No. 3, 504–510, 2012 507
In order to conﬁrm these splicing predictions, minigene assays
were carried out for the four variants and the results are presented
in Figure 1.
In accordance with predictions, the presence of each variant re-
sulted in an aberrant splicing pattern, and in all cases to the skipping
of the corresponding exon (Fig. 1-1 to 1-4). For the two constructs
ss, the use of either the de novo (c.13232-3C
the exonic cryptic (c.3235-2A
G) acceptor sites was not detected
(Fig. 1-3; 1-4).
In Silico Studies of the Missense Variants
Two missense variants suspected to be disease causing have been
T (p.(Arg4789Trp)) has been shown to lie
in trans to c.3635-2A
G and has not been found in 176 control
chromosomes. USMA analysis [Roux et al., 2011] shows a global
conservation in 19 of 20 orthologs (95%). The single di-
vergent protein sequence belongs to Anolis carolinensis (EN-
SACAP00000013363), which strongly differs from all others in this
G (p.(His5978Arg)) variation has been identiﬁed
in two unrelated patients, U361 and U919, in trans to c.3945dupA
(p.(Gln1316fs)) and to c.7001T
)), respectively. It
has not been found in 170 control chromosomes. USMA analysis
shows that the residue His is found in 16 of 17 ortholog sequences
(including Homo sapiens). The presence of a divergent residue Glu
is noted in Bos Taurus and includes a totally divergent alignment
in this region. The mutated residue Arg at position 5978 lies in the
ﬁrst extracellular loop of the 7-TM domain of G-protein coupled
receptor 98, the third TM region being predicted to begin at position
5980. Alignment of 237 sequences of 7-TM proteins shows that His
is the major residue at this position (97 sequences, 40.93%). Mutant
residue Arg is found only once (0.42%) in this alignment.
Two patients were identiﬁed who carried an unambiguous
pathogenic DFNB31 genotype. The two identiﬁed alterations
were frameshift mutations predicting a truncated whirlin protein
(Tables 1 and 2).
In addition, among all patients, 87 (68 in GPR98 and 19 in
DFNB31) variants were identiﬁed during this study, which could
be neutral. Fifty-eight additional changes were classiﬁed as UV1 or
UV2 (unclassiﬁed variants likely to be nonpathogenic) in GPR98
(55) and DFNB31 (3). They are all listed in USHbases.
Patients with no Mutation
After exhaustive analysis of USH2A, GPR98,andDFNB31 or
exclusion of their involvement by linkage, 19 patients remained with
no identiﬁed mutation. The coding exons of PDZD7 (considered to
be a potential candidate gene for Usher syndrome) were sequenced
but no pathogenic mutations were identiﬁed.
A further assessment of the phenotype was performed (Supp.
Table S3) and se veral atypical ﬁndings were highlighted in most
of these USH2 geno-negative patients, althoug h lack of sufﬁcient
clinical detail was inconclusive for two patients (U767 and U603).
Atypical audiological and/or ophthalmologic impairment was noted
for 12 patients. For two patients, the clinical presentation may in-
dicate the possibility of another syndrome because of additional
clinical ﬁndings such as renal impairment (U583) or the presence
of dysmorphic features (U607).
We have conﬁrmed in this study that mutations in GPR98 account
for a small but signiﬁcant proportion of mutations in USH2, that is,
6.4%. On the other hand, mutations in DFNB31 account for a very
small proportion (1.3%). This is to our knowledge the largest study
offering exhaustive molecular analysis of USH2 patients.
Mutations in GPR98 are of all types, including large gene rear-
rangements, and are spread throughout the gene, which additionally
contains numerous polymorphisms, but in almost all cases lead to
a truncated protein product. As the gene contains 90 exons, this
represents a considerable diagnostic challenge.
Although two patients were found in whom USH2 was caused by
mutations in DFNB31, only two mutant alleles were identiﬁed as one
patient was homozygous for c.737delC and the other patient was a
compound heterozygote for the same allele plus c.680dupG. It is,
spectrum in DFNB31.
Analysis and Outcome of the Likely Pathogenic Mutations
(UV3 and 4) Identiﬁed in this Study
Many of the predicted mutations in GPR98 could be deduced with
some conﬁdence as they predicted premature chain termination due
to small frameshifts or single changes, but several alterations needed
more care to analyze the predicted effect. This includes the large
rearrangement, the s plicing, and the missense alterations.
The large rearrangement consists of a duplication of exons 79
to 83, identiﬁed by means of a CGH array, and predicts an in-
frame duplication of 837 coding nucleotides, resulting in a protein
enlarged by 279 amino acids, from position 5,674 to 5,952. This
protein would contain a duplication of the entire G protein-coupled
receptors (GPCR) Proteolysis Site (GPS), the ﬁrst TMdomain, and
part of the second. It is very likely that the extracellular structure of
the resulting protein would be strongly affected.
Ex vivo splicing assays have shown that, in fact, the four tested
alterations result in exon skipping. They all predict an out of frame
sequence and a PTC with the exception of c.3635-2A
G, which re-
sults in an in-frame deletion, predicting a protein lacking 248 amino
acids (positions 1213 to 1460). This deletion removed the region be-
tween Calx-beta domains 9 and 10. It is difﬁcult to speculate on the
resulting changes, but the local structure is likely to be altered.
Unfortunately, in silico analysis of the two missense
p.(Arg4789Trp) and p.(His5978Arg) is not complete due to the lack
of appropriate three-dimensional models. Both missense show per-
fect alignments with ortholog sequences with the exception of one
species in each case (Anolis carolinensis in the ﬁrst and Bos Taurus
in the second case). Because the surrounding sequences diverge in
both species, the misalignment could be the result of an error in the
translation. In addition, p.(His5978Arg) arises in a well-conserved
position in the 7-TM domain. The fact that both missense lie in
trans to deleterious changes, are absent from a control population,
and are conserved among orthologs is in favor of their involvement
in the disease.
508 HUMAN MUTATION, Vol. 33, No. 3, 504–510, 2012
Figure 2. Distribution of the mutations identiﬁed in
(B). The exon numbering is indicated in blue squares. Identiﬁed
protein domains are represented with different pictograms. Pathogenic mutations and UV3–4 are designated at cDNA and protein level: the newly
identiﬁed variants are in bold and above the schematic proteins, the known variants are indicated below.
Mutation Spectrum of GPR98 and DFNB31 in Usher
A total of 19 likely pathogenic alterations have been identiﬁed in
the cohort described here, 17 in GPR98 and 2 in DFNB31. Seventeen
of them are newly described (Fig. 2). They raise the total number to
33 and 4 Usher-linked mutations in GPR98 and DFNB31, respec-
tively. All GPR98 mutations described previously and in this work
are schematically shown, distributed along the G protein-coupled
receptor 98 in Figure 2. Mutations are spread along the whole se-
quence but a cluster emerges in the terminal end. All the mutations
are predicted to lead to a PTC resulting in a truncated protein, with
the exception of four missenses (p.(His3399Pro), p.(Arg4789Trp),
p.(His5978Arg), and p.(Thr6044Cys)). It can be noted that as in the
majority of other Usher genes, the spectrum of pathological muta-
tions in GPR98 encompasses all possible type of DNA/RNA/protein
alterations, including at the DNA level: substitutions, small and large
deletions and duplications, alterations of splicing of the pre-mRNA
due to exonic or intronic mutations, creation of PTC and missense.
Again, as in the other genes (with maybe the exception of USH2A
c.2299delG), no major recurrent mutations have been identiﬁed.
These observations on ∼36 kb of DNA sequence, spread among
the 600 kb of the gene, show again the complexity of the evolution
of the DNA molecule, which modiﬁes itself using various mecha-
nisms occurring by chance which can sometimes lead to devastating
Only two mutations, located in the cytoplasmic domain
p.(Ala6216fs)and p.(Thr6244fs), would predict a G protein-coupled
receptor 98 with a functional extracellular domain and a native TM
domain. Because the PDZ binding motif, located in the COOH ex-
tremity, is then lacking in both cases, the resulting mutated proteins
are likely nonfunctional as, if present, are unable to be part of the
The mutational spectrum of GPR98 somewhat differs from that
found in USH2A, as pathogenic missense accounts here for 12%
versus 37%. This discrepancy is surprising considering that the two
proteins share general common properties, such as: (i) a large ex-
5,000 residues) part, (ii) the presence of numerous
repeated domains (35 ﬁbronectin type III domains in usherin, and
35 calx-beta core domains in G protein-coupled receptor 98), (iii)
the localization in the cells, and (iv) of course the phenotype the y
can induce, but can be explained by several factors. The ﬁrst one
is the bias introduced by the number of analyzed patients in each
gene (more than 150 in USH2A,31inGPR98), but even consid-
ering this, USH2A clearly shows a tendency to be more prone to
generate pathogenic missense variants. Another clue is the global
organization of the amino acids between the two proteins: 88%
of the residues in usherin belong to identiﬁed domains, while this
proportion falls down to 35–60% in G protein-coupled receptor
98 depending on the considered size of the calx-beta repeats. The
40–65% of remaining sequence may tolerate numbers of missense
Based on data supporting the hy pothesis that PDZD7 is part
of the Usher interactome, and because Usher patients have been
found carrying one PDZD7 mutation in association with one GPR98
mutation or a pathogenic USH2A genotype [Ebermann et al., 2010],
it was important to screen this gene as a potential Usher candidate
on its own. Implication of the PDZD7 gene has also been excluded
in the 19 patients with no identiﬁed mutation.
Among these patients, ﬁve remained with typical audiological
and ophthalmologic ﬁndings (U277, U496, U585, U654, U391).
These patients will be further screened for the USH1 genes, as some
of them can be associated with variable phenotypes and we have
indeed identiﬁed a pathogenic CDH23 genotype in a patient pre-
senting with clinical USH2 (severe HL and Usher type RP) pheno-
type (unpublished results). Implication of an as yet uncharacterized
HUMAN MUTATION, Vol. 33, No. 3, 504–510, 2012 509
Usher gene is not to be excluded but is likely to either constitute
a small minority or be the result of genotype/phenoty pe variation.
For example, mutation in the Myo15 gene (whose human ortholog
is MYO15A) is known to be involved in mainly profound deafness,
but has recently been shown to cause an Usher-like phenotype in
rats [Held et al., 2011].
Next-generation sequencing (NGS) of whole genome, exome, or
targeted reg ions is becoming more cost effective [Choi et al., 2009;
Lim et al., 2011]. More mutations will be found, categorized and
put into the databases and indeed NGS will have a role in solving
these atypical Usher cases efﬁciently. Sequence capture of Usher
exome is currently being developed and could be implemented in
“routine” diagnostics for any Usher patient in the future. As well,
whole exome s equencing will make a contribution to identify as yet
uncharacterized Usher genes.
Overall, a total of 176 USH2 patients has been investigated in our
laboratory, in either USH2A [Baux et al., 2007] and unpublished
data, or in the other USH2 genes (this study). The USH2 mutation
detection rate has now been raised to 89.2% (157/176) in clinically
diagnosed USH2 patients, which is quite similar to that of USH1
patients [Roux et al., 2011]. Interestingly, most patients remaining
with an uncharacterized pathogenic genotype, present with a ques-
tionable clinical Usher diagnosis. These observations could only be
done a posteriori and in these terms, we believe that our work is very
helpful to clinicians and diagnostic laboratories in their practice.
This work has been supported in par t by le Minist
ere de la Recherche “PHRC
National 2004, PROM 7802”. T he authors are grateful to “UNADEV” and
etinite” foundations for their support.
Abadie C, Blanchet C, Baux D, Larrieu L, Besnard T, Ravel P, Biboulet R, Hamel C,
Malcolm S, Mondain M, Claustres M, Roux AF. 2011. Audiological ﬁndings in
100 USH2 patients. Clin Genet
Baux D, Larrieu L, Blanchet C, Hamel C, Ben Salah S, Vielle A, Gilbert-Dussardier B,
Holder M, Calvas P, Philip N, Edery P, Bonneau D, Claustres M, Malcolm S, Roux
AF. 2007. Molecular and in silico analyses of the full-length isoform of usherin
identify new pathogenic alleles in Usher type II patients. Hum Mutat 28:781–
Bottillo I, De Luca A, Schirinzi A, Guida V, Torrente I, Calvieri S, Gervasini C, Larizza
L, Pizzuti A, Dallapiccola B. 2007. Functional analysis of splicing mutations in
exon 7 of NF1 gene. BMC Med Genet 8:4.
Choi M, Scholl UI, Ji W, Liu T, Tikhonova IR, Zumbo P, Nayir A, Bakkaloglu A, Ozen S,
Sanjad S, Nelson-Williams C, Farhi A, Mane S, Lifton RP. 2009. Genetic diagnosis
by whole exome capture and massively par allel DNA sequencing. Proc Natl Acad
Sci U S A 106:19096–19101.
Desmet FO, Hamroun D, Lalande M, Collod-Beroud G, Claustres M, Beroud C. 2009.
Human Splicing Finder: an online bioinformatics tool to predict splicing signals.
Nucleic Acids Res 37:e67.
Ebermann I, Phillips JB, Liebau MC, Koenekoop RK, Schermer B, Lopez I, Schafer
E, Roux AF, Daﬁnger C, Bernd A, Zrenner E, Claustres M, Blanco B, Nurnberg
G, Nurnberg P, Ruland R, Westerﬁeld M, Benzing T, Bolz HJ. 2010. PDZD7 is a
modiﬁer of retinal disease and a contributor to digenic Usher syndrome. J Clin
Ebermann I, Scholl HP, Charbel Issa P, Becirovic E, Lamprecht J, Jurklies B, Millan
JM, Aller E, Mitter D, Bolz H. 2007. A novel gene for Usher syndrome type 2:
mutations in the long isoform of whirlin are associated with retinitis pig mentosa
and sensorineural hearing loss. Hum Genet 121:203–211.
HJ, Wedekind D. 2011. A mutation in Myo15 leads to Usher-like symptoms in
LEW/Ztm-ci2 rats. PLoS One 6:e15669.
Kimberling WJ, Hildebrand MS, Shearer AE, Jensen ML, Halder JA, Trzupek K, Cohn
ES, Weleber RG, Stone EM, Smith RJ. 2010. Frequency of Usher syndrome in
two pediatric populations: implications for genetic screening of deaf and hard of
hearing children. Genet Med 12:512–516.
edard-Mereuze S, Vach
ere V, Larrieu L, Abadie C, Janecke A,
Claustres M, Roux AF, Tuffery-Giraud S. 2010. Ex vivo splicing assays of mutations
at non-canonical positions of splice sites in USHER genes. Hum Mutat31:347–355.
Lim BC, Lee S, Shin JY, Kim JI, Hwang H, Kim KJ, Hwang YS, Seo JS, Chae JH.
2011. Genetic diagnosis of Duchenne and Becker muscular dystrophy using next-
generation sequencing technology: comprehensive mutational search in a single
platform. J Med Genet 48:731–736.
Maerker T, van Wijk E, Overlack N, Kersten FF, McGee J, Goldmann T, Sehn E,
Roepman R, Walsh EJ, Kremer H, Wolfrum U. 2008. A novelUsher proteinnetwork
at the periciliary reloading point between molecular transport machineries in
vertebrate photoreceptor cells. Hum Mol Genet 17:71–86.
Mburu P, Mustapha M, Varela A, Weil D, El-Amraoui A, Holme RH, Rump A, Hardisty
RE, Blanchard S, Coimbra RS, Perfettini I, Parkinson N, Mallon AM, Glenister
P, Rogers MJ, Paige AJ, Moir L, Clay J, Rosenthal A, Liu XZ, Blanco G, Steel KP,
Petit C, Brown SD. 2003. Defects in whirlin, a PDZ domain molecule involved
in stereocilia elongation, cause deafness in the whirler mouse and families with
DFNB31. Nat Genet 29:29.
Michalski N, Michel V, Bahloul A, Lefevre G, Bar ral J, Yagi H, Chardenoux S, Weil D,
Martin P, Hardelin JP, Sato M, Petit C. 2007. Molecular characterization of the
J Neurosci 27:6478–6488.
Petit C. 2001. Usher syndrome: from genetics to pathogenesis. Annu Rev Genomics
Hum Genet 2:271–297.
Roux AF, Faugere V, Le Guedard S, Pallares-Ruiz N, Vielle A, Chambert S, Marlin
S, Hamel C, Gilbert B, Malcolm S, Claustres M. 2006. Survey of the frequency
of USH1 gene mutations in a cohort of Usher patients shows the importance of
cadherin 23 and protocadherin 15 genes and establishes a detection rate of above
90%. J Med Genet 43:763–768.
Mondain M, Gilbert-Dussardier B, Edery P, Lacombe D, Bonneau D, Holder-
Espinasse M, Ambrosetti U, Journel H, David A, Lina-Granade G, Malcolm S,
Claustres M. 2011. Four year follow-up of diagnostic service in USH1 patients.
Invest Ophthalmol Vis Sci 52:4063–4071.
Saihan Z, Webster AR, Luxon L, Bitner-Glindzicz M. 2009. Update on Usher syndrome.
Curr Opin Neurol 22:19–27.
van Wijk E, van der Zwaag B, Peters T, Zimmermann U, Te Brinke H, Kersten FF,
Marker T, Aller E, Hoefsloot LH, Cremers CW, Cremers FP, Wolfrum U, Knipper
M, Roepman R, Kremer H. 2006. The DFNB31 gene product whirlin connects
to the Usher protein network in the cochlea and retina by direct association with
USH2A and VLGR1. Hum Mol Genet 15:751–765.
Weston MD, Luijendijk MW, Humphrey KD, Moller C, Kimberling WJ. 2004. Mu-
tations in the VLGR1 gene implicate G-protein signaling in the pathogenesis of
Usher syndrome Type II. Am J Hum Genet 74:357–366.
Yeo G, Burge CB. 2004. Maximum entropy modeling of short sequence motifs with
applications to RNA splicing signals. J Comput Biol 11:377–394.
510 HUMAN MUTATION, Vol. 33, No. 3, 504–510, 2012