2001 Oxford University Press Human Molecular Genetics, 2001, Vol. 10, No.

25 29452951
Mutations in GJA1 (connexin 43) are associated with
non-syndromic autosomal recessive deafness
Xue Zhong Liu
1,2,
*, Xia Juan Xia
1
, Joe Adams
3
, Zheng Yi Chen
4
, Katherine O. Welch
5
,
Mustafa Tekin
1
, Xiao Mei Ouyang
2
, Arther Kristiansen
3
, Arti Pandya
1
, Thomas Balkany
2
,
Kathleen S. Arnos
5
and Walter E. Nance
1
1
Department of Human Genetics, Medical College of Virginia of Virginia Commonwealth University, Richmond,
VA 23298-0033, USA,
2
Department of Otolaryngology, University of Miami, Miami, FL 33101, USA,
3
Department of
Otolaryngology, Harvard Medical School and Massachusetts Eye and Ear Infirmary and
4
Neurology Department,
Massachusetts General Hospital and Neurobiology Department, Harvard Medical School, Boston, MA 02114, USA
and
5
Department of Biology, Gallaudet University, 800 Florida Avenue NE, Washington,
DC 20002, USA
Received August 21, 2001; Revised and Accepted September 24, 2001 DDBJ/EMBL/GenBank accession no. X17027459
Mutations in four members of the connexin gene
family have been shown to underlie distinct genetic
forms of deafness, including GJB2 [connexin 26
(Cx26)], GJB3 (Cx31), GJB6 (Cx30) and GJB1 (Cx32).
We have found that alterations in a fifth member of
this family, GJA1 (Cx43), appear to cause a common
form of deafness in African Americans. We identified
two different GJA1 mutations in four of 26 African
American probands. Three were homozygous for a
LeuPhe substitution in the absolutely conserved
codon 11, whereas the other was homozygous for a
ValAla transversion at the highly conserved codon
24. Neither mutation was detected in DNA from 100
control subjects without deafness. Cx43 is
expressed in the cochlea, as is demonstrated by PCR
amplification from human fetal cochlear cDNA and
by RTPCR of mouse cochlear tissues. Immuno-
histochemical staining of mouse cochlear preparations
showed immunostaining for Cx43 in non-sensory
epithelial cells and in fibrocytes of the spiral ligament
and the spiral limbus. To our knowledge this is the
first connexin gene to be associated with non-
syndromic deafness. Cx43 must also play a critical
role in the physiology of hearing, presumably by
participating in the recycling of potassium to the
cochlear endolymph.
INTRODUCTION
Hearing loss is an important cause of human morbidity that
eventually affects at least 15% of the population. The
incidence of profound prelingual deafness is approximately
one per 1000 at birth and includes many known genetic and
environmental causes (1). Dramatic progress has been made in
the localization of 71 loci that can cause dominant or recessive
forms of non-syndromic deafness, and, of these, 22 have been
cloned (G.Van Camp and R.J.H.Smith, Hereditary Hearing
Loss Homepage. World Wide Web address: http:dnalab-
www.uia.ac.be/dnalab/hhh/). Mutations in GJB2 [connexin 26
(Cx26)] can give rise to recessive non-syndromic deafness or,
more rarely, dominantly transmitted deafness (2). Mutations
involving this locus are the most common cause of deafness in
many populations. One particular mutation, 30delG or 35delG,
is especially common, accounting for two-thirds of all patho-
logic Cx26 mutations in most populations. Cx31 (GJB3) is also
expressed in the cochlea and is the cause of deafness in several
Chinese pedigrees (3,4). Likewise, mutations in GJB6 (Cx30)
are the apparent cause of deafness in a few small families
(http:dnalab-www.uia.ac.be/dnalab/hhh/) (5), whereas muta-
tions in GJB1 (Cx32) lead to an X-linked form of Charcot-
Marie-Tooth disease (CMTX) associated with hearing loss
(http:dnalab-www.uia.ac.be/dnalab/hhh/) (6).
Connexins are the proteins which line the intercellular chan-
nels or gap junctions that connect adjacent cells and facilitate
the exchange of ions, secondary messengers and small
molecules (7). The numbers assigned to the various connexins
refer to their approximate molecular weight. In addition, the
human connexins can be classified into sub groups: , , and ,
based on similarities at the nucleotide and amino acid level.
Six connexin molecules assemble to form a half-channel or
connexon, which docks with its counterpart in an adjacent cell
to form a complete intercellular channel or gap junction. In the
sensory epithelia of the inner ear, it is thought that an important
function of these gap junctions is to facilitate the recycling of
potassium ions from the hair cells back into the cochlear endo-
lymph during auditory transduction (8). The involvement of
several members of the connexin gene family in deafness
suggests that others should be considered as candidates for
non-syndromic deafness. To pursue this possibility, we initi-
ated a study to determine whether there are mutations in GJA1,
*To whom correspondence should be addressed at: Department of Otolaryngology (D-48), University of Miami, 1666 NW 12th Avenue, Miami, FL 33136, USA.
Tel: +1 305 243 4923; Fax: +1 305 243 4925; Email: xliu@med.miami.edu
2946 Human Molecular Genetics, 2001, Vol. 10, No. 25
encoding Cx43, which lead to deafness and to document the
expression of Cx43 in the inner ear.
RESULTS AND DISCUSSION
In total, 26 deaf African American probands were screened for
Cx43 mutations, including five from multiplex and 21 from
simplex families. In addition, screening was conducted on
several control groups including 40 African American subjects
with normal hearing. Four of the deaf probands who were born
to non-consanguineous parents in three simplex (NSDF 056,
062 and 205) and one multiplex family (NSDF 158) were
found to have pathological mutations in GJA1 (see below).
Clinical and audiological evaluation showed that all four
probands had a profound bilateral non-syndromic
sensorineural loss that appeared to have been congenital with
no evidence for progression, which is the most common type
of audiograms in non-syndromic recessive deafness (9). Phys-
ical examination of probands in families NSDF 056, 062 and
205 showed no dysmorphic features of the face, jaw, palate, or
external ears with no history of vertigo or evidence for vestib-
ular dysfunction. No other clinical abnormalities were noted.
Likewise, the proband in family NSDF 158 had no reported
abnormalities. In family NSDF 158 (Fig. 1A), there is a history
of a brother who, by report, is also deaf but not available for
testing.
The DNA samples from probands were screened for Cx43
mutations by a single-strand conformation polymorphism
(SSCP)-sequencing approach after initial testing for mutations
in Cx26, Cx30 and Cx31 with negative results (data not
shown). GJA1 contains a single, uninterrupted ORF of 1148 nt
(10). Eight pairs of overlapping primers covering the entire
coding region of GJA1 were used for the SSCP analysis (see
Materials and Methods). Variant SSCP patterns for the PCR of
primer 1FR were observed in four African American patients
(data not shown). The four affected subjects were all
homozygous for one or the other of two variants, whereas the
remaining 22 black probands and 40 black control subjects
with normal hearing all had normal SSCP patterns. Sequence
analysis of the SSCP variants demonstrated that all three deaf
individuals from families NSDF 056, 062 and 205 were
homozygous for a 30CT transversion which leads to
LeuPhe substitution at codon 11 (L11F) (Fig. 1B). In family
NSDF 205 (Fig. 1A), a hearing sister (II.2) and the mother (I.2)
were found to be heterozygous for the L11F mutation. The
other mutant L11F allele was apparently inherited from the
unaffected father, who was not available for testing. In family
NSDF 158, the affected individual (II.1, Fig. 1B) was
homozygous for a 71TC change which creates a ValAla
substitution at codon 24 (V24A).
To exclude the possibility that these mutations are simply
polymorphisms, we have studied samples from 100 unrelated
Figure 1. Mutation analysis of four families. (A) Pedigrees of family NSDF 205 and family NSDF 158. (B) Direct sequence analysis of control and patients for
GJA1. Top, the sequences of individual subjects at position 11. Patient II.1 from family NSDF 205 shows a homozygous CT substitution, as were the probands
of families NSDF 056 and 062 (data not shown). Both mother (I.2) and sister (II.2) with normal hearing are heterozygous for the mutation. The sequence of normal
control is also shown. Bottom, the sequences of individual subjects at position 71. Patient II.1 from family NSDF 158 shows a homozygous CT substitution. The
sequence of the normal control is also shown.
Human Molecular Genetics, 2001, Vol. 10, No. 25 2947
control subjects (including 40 African Americans and 60 from
other populations) for both mutations. Neither of the two muta-
tions was detected in the control panel. The mutations are
believed to be pathological, first, because of their location and
conservation (see below) and, secondly, because the changes
were not observed in a series of normal controls. To search for
the mutations in other racial or population groups, we screened
510 deaf probands from other ethnic backgrounds obtained in
part from the National DNA Repository for Genetic Deafness
at the Medical College of Virginia (11). None carried the two
mutations found in our patients.
Like other connexins, Cx43 consists of four transmembrane
domains linked by one cytoplasmic and two extracellular
loops, with cytoplasmic C- and N-terminal ends (Fig. 2). GJA1
is located on human chromosome 6q21q23.2 (10). Currently,
none of the known loci for non-syndromic deafness maps to
this chromosomal region. The only human disease previously
attributed to GJA1 was a recessive form of lateralization
defects (heterotaxy) in which the substitutions of phosphorat-
able Ser or Thr residues in the cytoplamic tail domain were
reported (12). However, intensive efforts to confirm these find-
ings in a large number of patients have failed to reveal similar
findings (13). To date, mutations in six connexin genes are
known to cause a variety of clinical abnormalities (14). In
addition to Cx43, four other connexins are associated with
hearing impairment (http:dnalab-www.uia.ac.be/dnalab/hhh/).
The altered Leu residue in families NSDF 056, 062 and 205
lies in the cytoplasmic N-terminal domain (NT, Fig. 2) in a
stretch of 14 residues that exhibit a high degree of evolutionary
conservation across species in all Cx43 genes studied to date as
well as in other members of the connexin family (Fig. 3). This
domain plays a role in the voltage-gating of gap junction chan-
nels and in the insertion of connexins into membranes (15,16).
In GJB2 (Cx26), more than 50 different mutations have been
identified in patients with non-syndromic deafness, including
missense, nonsense and small deletions or insertions across all
protein domains. More than five of these Cx26 and two Cx30
mutations lie within the N-terminal domain, and a number of
mutations in this region occur in Cx32 in patients with CMTX
as well as in Cx31 in patients with erythrokeratodermia vari-
abilis. Moreover, mutations in residues close to the Leu at
codon 11 altered in the families reported here, namely T8 and
G12 in Cx26, as well as T5 and G11 in Cx30, have been
reported previously in patients with hearing impairment
(http:dnalab-www.uia.ac.be/dnalab/hhh/) (5). Mutations in the
N-terminal domain of Cx32 or Cx26 have been shown to either
reverse the gating polarity or cause defective trafficking of
proteins (15,16). Interestingly, a substitution of the Gly residue
12 to Ser (G12S) in Cx32, has been shown to cause defective
trafficking with no measurable conductance (16). Therefore, it
is highly likely that the Leu11Phe substitution we observed
could alter the structure of the N-terminus, in a way that might
block the assembly, insertion or cellular processing of the
Cx43 subunit. The identification of pathological mutations
associated with hearing loss in the homologous region of
several other connexins supports the view that this domain is of
functional importance for the Cx43 protein.
The Val24Ala substitution in family NSDF 158 occurs in the
first putative transmembrane domain (M1) of Cx43, which is
critical for the normal formation of the pore of gap junction
Figure 2. Schematic representation of Cx43 and its mutations. Closed circles, two recessive mutations found in the four families studied; M
14
, transmembrane
domains; E1 and E2, extracellular domains; CL, cytoplasmic linking domain; NT, N-terminal domain; CT, C-terminal domain.
2948 Human Molecular Genetics, 2001, Vol. 10, No. 25
channels (13,17). Moreover, this Val is located at the NT/M1
boundary (Fig. 2) and is conserved across connexins from
many species, whereas others contain a structurally conserva-
tive isoleucine substitution at this position (Fig. 3). Amino
acids at this transmembrane border form part of a charged
complex that is a component of the connexin voltage sensor
(15). Non-conservative connexin mutations affecting this Val
and adjacent residues are pathologic (14). Interestingly, two
substitutions involving this residue, Val23Ala and Val23Glu in
Cx32, have been implicated in CMTD (18). A mutation
involving the adjacent Trp24 in Cx26 has also been reported
previously in patients with deafness (http:dnalab-
www.uia.ac.be/dnalab/hhh/). The mutation in the neighboring
S26 residue in Cx32 reduces channel permeability by
decreasing the pore size (19). Since the Val24Ala substitution
that we observed introduces a less hydrophobic residue at the
membrane cytoplasm boundary (Fig. 2), functionally significant
changes in the structure of M1 might be expected.
To support the causal relation between Cx43 mutations and
deafness in our patients, we characterized the expression of
Cx43 by RTPCR of mouse cochlear tissues and by PCR
amplification from human fetal cochlear cDNA (data not
shown). We detected GJA1 expression in mouse cochlear
tissues, and were able to amplify Cx43 in human fetal cochlear
cDNA from the Morton cochlear library.
Using two independent affinity purified antibodies to Cx43,
we performed immunohistochemical localization of the Cx43
molecule in sections of mouse cochlea. The pattern of staining
for Cx43 was dependent upon fixation conditions of the tissue.
In tissue fixed with formalin\glutaraldehyde, staining was
strongest and most consistently present in non-sensory epithe-
lial cells (closed arrows, Fig. 4A), in the spiral ligament and in
the spiral limbus (open arrows, Fig. 4A). In the ligament, type
I fibrocytes were much more darkly stained than adjacent type
III fibrocytes. Other positive staining cells included the inter-
dental cells of the spiral limbus and outer sulcus and the root
cells beneath the spiral prominence. The pattern of staining
was essentially the same with both antibodies, with the Zymed
antibody usually providing superior results. A similar distribu-
tion of positive staining cells was also found in rat and guinea
pig cochlea (data not shown). In tissue fixed in ethanol\acetic
acid, staining was limited to the apices of supporting cells of
the organ of Corti (Fig. 4B and C). Previous work on the
distribution patterns of Cx26, Cx30 and Cx31, including
ultrastructural confirmation of the locations of gap junctions
within the cochlea, led to the conclusion that gap junctional
systems provided the pathways by which potassium ions are
Figure 3. Sequence alignment of human GJA1 with other gap junction proteins. The arrows indicate the range of putative domains. Asterisk indicates the position
of L11F and V24A, identified in four families. Identical residues are highlighted with a dark background. L11 is conserved throughout all connexins.
Human Molecular Genetics, 2001, Vol. 10, No. 25 2949
recirculated from the organ of Corti to the stria vascularis (8).
According to this model, the potassium ions that enter hair
cells in response to acoustic activation of the organ of Corti are
expelled basolaterally by hair cells and are accumulated by
supporting cells. From the organ of Corti the ions moved laterally
through gap junctions to the spiral ligament, where they
expelled from the terminal cells of the epithelial gap junctional
system and are accumulated by type II fibrocytes, from which
they are transported to the stria vascularis via the gap
junctional system of the connective tissue cells of the spiral
ligament. Cx26 and Cx30 are present in both the epithelial and
connective tissue gap junctional systems and mutations in
Cx26 and Cx30 genes can result in sensorineural hearing loss
(2,5). On the other hand, Cx31 is present primarily in type II
fibrocytes (20) and mutations in that gene can also result in
hearing loss (3,4). Apparently failure of gap junctions that are
either distributed throughout the cochlea or are limited to key
cells can lead to hearing loss. The present results indicate that
Cx43 is either present in a highly restricted group of supporting
cells (Fig. 4B and C) or is found throughout non-sensory
epithelial cells and in type I fibrocytes. Apparently this issue
will have to be resolved by using in situ hybridization. In either
case, it appears that the contribution of Cx43 to the function of
cochlear gap junctions is critical for normal hearing. These
immunohistochemical data are consistent with the idea that the
Cx43 mutation is the underlying cause of deafness in our four
families.
To gain further insight into the ontogeny of GJA1 expres-
sion, we have characterized the expression profiles of mouse
cochlea at three developmental stages by Genechip analysis
(21). Cx43 was expressed in the mouse cochlea at P1, P2 and
P32, and at significantly higher levels in P32 (>2-fold greater
than at P1 or P2). Cx43 was also expressed during early devel-
opment of the mouse utricle (Z.-Y.Chen, unpublished data).
Cx43, Cx30 and Cx26 all increased their expression from P1 to
P32, with the highest increase for Cx30 (>5-fold) and the
smallest for Cx26 (marginally). In the developing utricle, in
contrast to Cx43, both Cx26 and Cx30 are up-regulated post-
natally. The observed variation in the expression of the three
connexins may indicate that they have distinct but overlapping
functional roles in the inner ear.
The present data demonstrate that Cx43 is responsible for a
recessive form of non-syndromic deafness. The identification
of mutations in four of 26 African American families suggests
that Cx43 mutations could be a common cause of deafness in
this group, and like Cx26, its prevalence in the United States
could be influenced by the mating structure of the deaf popula-
tion (22). Additional data will be required to determine the
frequency of GJA1 mutations as a cause for deafness in other
populations. Since the connexins function by forming hetero-
meric or homomeric connexons, it is interesting to notice that
all pairs of the mutations identified in deaf patients to date have
involved changes in the same connexin subunit. It seems
possible that mutations involving different but interacting
connexins may exist in some deaf individuals. Our findings of
mutations in Cx43 in four deaf families and information on its
localization and the expression profile should aid identification
of epistatic interactions of this type.
MATERIALS AND METHODS
Clinical evaluations of patients
Probands were ascertained through the Genetics Clinic at
Virginia Commonwealth University in Richmond and the
Genetics Program at Gallaudet University in Washington, DC.
Informed consent was obtained from all participants or from
parents of patients younger than 18 years of age. In total, 26
African American probands with recessive or sporadic non-
syndromic deafness were included in this study. All probands
Figure 4. Immunostaining of mouse cochlea for Cx43 using glutaraldehyde/
formalin fixative (A) or ethanol/acetic acid (B and C). Filled arrows in (A)
indicate staining of non-sensory epithelial cells. Open arrows indicate stained
type I fibrocytes in the spiral ligament. The box in (A) encloses the region of
the organ of Corti that is shown at higher magnification (in a different speci-
men) in (B) and (C). The closed arrow in (B) indicates staining at the apical
surfaces of Deiters cells. The open arrow indicates staining along the apical
border of the inner pillar cell as seen in a radial section. The arrow in (C) indi-
cates the surface view of the region indicated by the filled arrow in (B) as seen
in a tangential section. The dark curved lines in (C) are immunostained mar-
gins of the apical portion of Deiters cells. The calibration bar in (C) applies to
(B) and (C) and is 10 m in length.
2950 Human Molecular Genetics, 2001, Vol. 10, No. 25
had profound prelingual deafness. The clinical and family
history was obtained on each proband and complete physical
examinations were performed on three of the four probands by
one of the investigators. The hearing of all affected individuals
in the present series was examined using pure tone audiometry.
Air conduction thresholds were measured at 250 Hz, 500 Hz,
1 kHz, 2 kHz, 4 kHz, 6 kHz and 8 kHz. DNA samples were
obtained from peripheral blood in the study families and
controls.
SSCP and sequence analysis of Cx43
GJA1 has a single exon with a coding region of 1148 bp. Eight
pairs of overlapping primers were designed from the complete
coding and flanking UTR sequence of Cx43 (GenBank acces-
sion no. XM027459) using the Primer 3 program (http://www-
genome.wi.mit.edu/genome-software/other/primer3.html).
The amount of overlap between each of the primer pairs ranged
from 34 to 65 bp. Primer sequences for the SSCP screen of
Cx43 will be provided by request. The coding region of the
Cx43 gene was amplified from genomic DNA by PCR using
the above primers. The amplification conditions were: 95C
for 5 min, then 30 cycles of 95C for 1 min, 60C for 1 min and
72C for 1 min, with a final extension for 5 min at 72C.
For mutation analysis, the PCR products were initially run
on a 1 mm thick 8% non-denaturing polyacrylamide gel (acryl-
amide:N,N-methylene bisacrylamide, 49:1) at 4C. SSCPs
were detected using silver staining as previously described by
Liu et al. (23). Direct sequencing of PCR products from
patients with SSCP variants was then performed on both
strands using the fluorescent dideoxy terminator method and
an ABI 377 DNA sequencer.
RTPCR
We performed reverse transcription reactions with the Super-
script reverse transciptase (Gibco BRL, Boston, MA) follow-
ing the manufacturers protocol. All soft tissues from cochleas
of young male CBA/CaJ mice were collected. Total RNA was
isolated using TRIzol reagent (Gibco BRL) in accordance with
the manufacturers directions. PCR conditions were: 93C for
2 min, followed by 39 cycles of 93C for 1 min, 56C for 1 min
and 72C for 1 min. The final primer extension was at 72C for
7 min. The primers used for RTPCR were: forward 5-TGCG-
GTCTACACCTGCAAGA-3; reverse 5-ACCAAGGACACC-
ACCAGCAT-3.
Immunohistochemical analysis
CBA/CaJ mice were deeply anesthetized and perfused intra-
cardially with phosphate buffered saline (PBS) and then 10%
formalin in PBS. The stapes was removed, the round window
pierced and formalin gently perfused through the cochlear
scalae. In some cases the fixative also included 0.1% glutar-
aldehyde. In other cases, no aldehyde fixative was used.
Instead, following exsanguination with PBS, a solution of 50%
ethanol and 5% acetic acid was perfused through the cochlear
scalae. In all cases, the head was immersed in fixative for 2 h.
Then a block of tissue including the temporal bones was
removed from the skull and placed in a large volume of EDTA
(pH 7.0) and gently agitated for 1014 days until the bone was
decalcified. The tissue was then dehydrated and embedded in
paraffin.
Serial 8 m sections were mounted on slides. Selected slides
were dewaxed in xylene, hydrated and immunostained for
Cx43. Primary antibodies included two well characterized
affinity-purified rabbit antibodies (Zymed Labs, San Francisco,
CA and a gift antiserum Petunia from Dr David Paul).
Following 3060 min in 5% normal horse serum to block non-
specific IgG binding, serial dilutions of the primary antibodies
were performed and the tissue sections bathed in primary anti-
body overnight at room temperature. A biotinylated donkey
anti-rabbit secondary antibody (Jackson ImmunoResearch
Labs, West Grove, PA) was applied for 1 h, followed by an
avidinbiotinHRP complex (Standard ABC kit; Vector Labora-
tories, Burlingame, CA) for 1 h, biotinylated tyramine/H
2
O
2
for
10 min (24), ABC solution for 30 min and then diamino-
benzidine/H
2
O
2
for 2 min. Copious washing in PBS followed
each step.
Following the final step, sections were dehydrated and
coverslipped in Permount. Inner ears of rat and guinea pig
were also immunostained using the same protocol.
ACKNOWLEDGEMENTS
We would like to thank the families for their contribution to
this study. We thank Dr C.C.Morton for providing human fetal
cochlear cDNA and Wanda Hunt for her technical help. This
work was supported by NIH grants DC 05575 and DC 04530
to X.Z.L., DC 02530 and DC 04293 to W.E.N., DC 03929 to
J.A. and DC 04546 to Z.Y.C.
REFERENCES
1. Gorlin, R.J., Toriello, H.V. and Cohen, M.M. (1995) Hereditary Hearing
Loss and Its Syndromes. Oxford University Press, Oxford, UK.
2. Kelsell, D.P., Dunlop, J, Stevens, H.P., Lench, N.J., Liang, J.N., Parry, G.,
Mueller, R. F. and Leigh, I.M. (1997) Connexin 26 mutations in hereditary
non-syndromic sensorineural deafness. Nature, 387, 8083.
3. Xia, J.H., Liu, C.Y., Tang, B.S., Pan, Q., Huang, L., Dai, H.P.,
Zhang, B.R., Xie, W., Hu, D.X., Zheng, D. et al. (1998) Mutations in the
gene encoding gap junction protein -3 associated with autosomal
dominant hearing impairment. Nat. Genet., 20, 370373.
4. Liu, X.Z., Xia, X.J., Xu, L.R., Pandya, A., Liang, C.Y., Blanton, S.H.,
Brown, S.D., Steel, K.P. and Nance, W.E. (2000) Mutations in connexin
31 underlie recessive as well as dominant non-syndromic hearing loss.
Hum. Mol. Genet., 9, 6367.
5. Grifa, A., Wagner, C.A., DAmbrosio, L., Melchionda, S., Bernardi, F.,
Lopez-Bigas, N., Rabionet, R., Arbones, M., Monica, M.D., Estivill, X.
et al. (1999) Mutations in GJB6 cause nonsyndromic autosomal dominant
deafness at DFNA3 locus. Nat. Genet., 23, 1618.
6. Bergoffen, J., Scherer, S.S., Wang, S., Scott, M.O., Bone, L.J., Paul, D.L.,
Chen, K., Lensch, M.W., Chance, P.F. and Fischbeck, K.H. (1993)
Connexin mutations in X-linked Charcot-Marie-Tooth disease. Science,
262, 20392042.
7. Goodenough, D.A., Goliger, J.A. and Paul, D.L. (1996) Connexins,
connexons, and intercellular communication. Annu. Rev. Biochem., 65,
475502.
8. Kikuchi, T., Kimura, R.S., Paul, D.L. and Adams, J.C. (1995) Gap
junctions in the rat cochlea: immunohistochemical and ultrastructural
analysis. Anat. Embryol., 191, 101118.
9. Liu, X.Z. and Xu, L.R. (1994) Non-syndromic hearing loss: an analysis of
audiogram. Ann. Otol. Rhinol. Laryngol., 10, 428432.
10. Fishman, G.I., Spray, C. and Leinwand, L.A. (1990) Molecular
characterization and functional expression of the human cardiac gap
junction channel. J. Cell Biol., 111, 589598.
11. Nance, W., Pandya, A., Liu, X.-Z. and Arnos, K. (2001) Creation of a DNA
repository to identify new genes for human deafness. Genet. Med., 3, 209.
Human Molecular Genetics, 2001, Vol. 10, No. 25 2951
12. Britz-Cunningham, S.H., Shah, M.M., Zuppan, C.W. and Fletcher, W.H.
(1995) Mutations of the Connexin43 gap-junction gene in patients with
heart malformations and defects of laterality. New Engl. J. Med., 18,
13231329.
13. Toth, T., Hajdu, J., Marton, T., Nagy, B. and Papp, Z. (1998) Connexin 43
gene mutations and heterotaxy. Circulation, 9, 117118.
14. Krutovskikh, V. and Yamasaki, H. (2000) Connexin gene mutations in
human genetic diseases. Mutat. Res., 462, 197207.
15. Verselis, V.K., Ginter, C.S. and Bargiello, T.A. (1994) Opposite voltage
gating polarities of two closely related connexins. Nature, 368, 348351.
16. Deschenes, S.M., Walcott, J.L., Wexler, T.L., Scherer, S.S. and
Fischbeck, K.H. (1997) Altered trafficking of mutant connexin 32.
J. Neurosci., 17, 90779084.
17. Zhou, X.W., Pfahnl, A., Werner, R., Hudder, A., Llanes, A., Luebke, A.
and Dah, G. (1997) Identification of a pore lining segment in gap junction
hemichannels. Biophys. J., 72, 19461953.
18. Nelis, E., Haites, N. and Van Broeckhoven, C. (1999) Mutations in the
peripheral myelin genes and associated genes in inherited peripheral
neuropathies. Hum. Mutat., 13, 1128.
19. Oh, S., Ri, Y., Bennett, M.V., Trexler, E.B., Verselis, V.K. and
Bargiello, T.A. (1997) Changes in permeability caused by connexin 32
mutations underlie X-linked Charcot-Marie-Tooth disease. Neuron, 19,
927938.
20. Xia, A.P., Ikeda, K., Katori, Y., Oshima, T., Kikuchi, T. and Takasaka, T.
(2000) Expression of connexin 31 in the developing mouse cochlea.
Neuroreport, 1, 24492453.
21. Chen, Z.Y. and Corey, D.P. (2001) An inner ear gene expression database.
J. Assoc. Res. Otolaryngol., in press.
22. Nance, W.E., Liu, X.Z. and Pandya, A. (2000) Relation between choice of
a partner and high frequency of connexin-26 deafness. Lancet, 356, 500512.
23. Liu, X.Z., Walsh, J., Mburu, P., Kendrick-Jones, J., Cope, J.T.V., Steel, K.
and Brown, S.D.M. (1997) Mutations in myosin VIIA gene caused non-
syndromic recessive deafness. Nat. Genet., 16, 188190.
24. Adams, J.C. (1992) Biotin amplification of biotin and horseradish
peroxidase signals in histochemical stains. J. Histochem. Cytochem., 40,
14571463.
2952 Human Molecular Genetics, 2001, Vol. 10, No. 25