 1997 Oxford University Press
Human Molecular Genetics, 1997, Vol. 6, No. 1
17–26
ARTICLE
Mutations in the gene encoding the
inwardly-rectifying renal potassium channel, ROMK,
cause the antenatal variant of Bartter syndrome:
evidence for genetic heterogeneity
International Collaborative Study Group for Bartter-like Syndromes, consisting of:
Group 1: Lothar Károlyi1,2, Martin Konrad1, Arnold Köckerling1, Andreas Ziegler3,
Dorthe K. Zimmermann1, Bernd Roth4, Christian Wieg5, Karl-Heinz Grzeschik2,
Manuela C. Koch2, Hannsjörg W. Seyberth1, Group 2: Rosa Vargas6, Lionel Forestier6,
Genevieve Jean6, Michele Deschaux6, Gian Franco Rizzoni7, Patrick Niaudet6,
Corinne Antignac6, Group 3: Delphine Feldmann8, Frederique Lorridon8,
Emmanuel Cougoureux8, France Laroze8, Jean-Luc Alessandri10, Louis David11,
Pascal Saunier12, Georges Deschenes9, Group 4: Friedhelm Hildebrandt13,
Martin Vollmer13, Willem Proesmans14, Matthias Brandis13, Group 5:
Lambertus P. W. J. van den Heuvel15, Henny H. Lemmink15, Willy Nillesen16,
Leo A. H. Monnens15, Nine V. A. M. Knoers16, Group 6: Lisa M. Guay-Woodford17,*,
Christopher J. Wright18, Gilbert Madrigal19 and Steven C. Hebert20
1Department
of Pediatrics, 2Institute of Human Genetics, 3Institute of Medical Biometry, Philipps University, 35033
Marburg, Germany, 4University Children’s Hospital, 50924 Cologne, Germany, 5Children’s Hospital, 63411 Hanau,
Germany, 6INSERM U423, Necker Hospital, University Paris 5, France, 7Bambino Gesu Children’s Hospital, Rome,
Italy, Departments of 8Biochemistry and 9Pediatrics, Armand-Trousseau Hospital, Paris, France, 10Felix Guyon
Hospital, Saint Denis, France, 11Edouard Herriot Hospital, Lyon, France,12Fontainebleau Hospital, Fontainebleau,
France, 13Department of Pediatrics, University of Freiburg, Germany, 14Department of Pediatrics, University of
Leuven, Belgium, Departments of 15Pediatrics and 16Human Genetics, University of Nijmegen, The Netherlands,
17Departments of Medicine and Pediatrics and 18Department of Cell Biology, University of Alabama at Birmingham,
624 Zeigler Research Building, 703 South 19th Street, Birmingham, AL 35294, USA, 19Hospital Nacional de Ninos,
San Jose, Costa Rica and 20Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA
Received September 10, 1996; Revised and Accepted October 14, 1996
Inherited renal tubular disorders associated with hypokalemic alkalosis (Bartter-like syndromes) can be
subdivided into at least three clinical phenotypes: (i) the hypocalciuric-hypomagnesemic Gitelman variant; (ii) the
classic variant; and (iii) the antenatal hypercalciuric variant (also termed hyperprostaglandin E syndrome).
Mutations in the Na-Cl cotransporter (NCCT) underlie the pathogenesis of the Gitelman variant and mutations in
the Na-K-2Cl cotransporter (NKCC2) have recently been identified in the antenatal hypercalciuric variant. We now
describe mutations in the gene encoding the inwardly-rectifying potassium channel, ROMK, in eight kindreds with
the antenatal variant of Bartter syndrome. These findings indicate that antenatal Bartter syndrome is genetically
heterogeneous and provide new insights into the molecular pathogenesis of Bartter-like syndromes.
*To whom correspondence should be addressed
18
Human Molecular Genetics, 1997, Vol. 6, No. 1
INTRODUCTION
Bartter syndrome encompasses a set of primary renal tubular
disorders associated with hypokalemic metabolic alkalosis; often
increased urinary prostaglandin excretion; hyperreninemic
hyperaldosteronism with normal blood pressure; and hyperplasia
of the juxtaglomerular apparatus (1). Within this relatively rare
set of disorders, familial cases occur commonly and inheritance
best fits with autosomal recessive transmission (2,3). This set of
disorders can be subdivided into at least three clinical phenotypes:
(i) the hypocalciuric-hypomagnesemic variant described by
Gitelman et al.; (ii) the classic syndrome originally described by
Bartter et al.; and (iii) the antenatal hypercalciuric variant
associated with severe systemic manifestations (3–6).
Among these disorders, Gitelman syndrome is characterized by a
relatively mild course and late age of onset. Profound hypokalemia
and hypomagnesemia lead to the major manifestations of fatigue,
muscle weakness and recurrent episodes of tetany (7,8).
Hypocalciuria and mildly impaired urinary concentrating ability are
constant features of this sub-type (9). Recent studies demonstrate
that this disorder results from presumptive loss of function mutations
in the gene encoding the thiazide-sensitive sodium-chloride (Na-Cl)
cotransporter (NCCT) of the distal nephron (10). Genetic
homogeneity in Gitelman syndrome has been confirmed in
subsequent studies (11,12).
In comparison, patients with classic Bartter syndrome (CBS)
generally present during early childhood. These patients fulfill
the core criteria of Bartter’s original description without
associated tetanic episodes or profound hypercalciuria (3,6,13).
Typically, urinary calcium excretion is normal or slightly elevated
and urinary concentrating capacity is close to normal (9,14).
While much attention has been focused on defective chloride
transport in the medullary thick ascending limb of the loop of
Henle (mTAL) as the proximate abnormality in CBS (15–18), the
underlying defect in the classic variant remains to be elucidated.
Finally, the antenatal variant of Bartter syndrome (ABS) is a
life-threatening disorder in which both the renal tubular
hypokalemic alkalosis as well as profound systemic symptoms
are manifest (5,19,20). The abnormalities begin in utero with
marked fetal polyuria that leads to polyhydramnios between 24
and 30 weeks gestation and typically, premature delivery
(5,19–21). The amniotic fluid contains high chloride levels but
normal concentrations of sodium, potassium, calcium and
prostaglandin E2 (22). Affected neonates have severe salt wasting
and hyposthenuria, moderate hypokalemic metabolic alkalosis,
hyperprostaglandinuria and failure to thrive. An essential
manifestation is marked hypercalciuria and as a secondary
consequence affected infants develop nephrocalcinosis and
osteopenia (23–26). The fever, vomiting and occasional
diarrhoea associated with this disorder has been attributed to the
stimulation of renal and systemic prostaglandin E2 activity in
affected infants and these symptoms are effectively treated with
inhibitors of prostaglandin synthesis. Based on these clinical
features, the term hyperprostaglandin E syndrome was coined to
describe this antenatal variant of Bartter syndrome (27).
Table 1. Clinical characteristics of index patients with antenatal Bartter syndromea
Families
AB.Paris1
AB.Paris2
AB.Paris3
AB.Paris5
AB.Paris6
AB.Paris8
AB.Paris9
AB.Marb1
AB.Marb2
AB.Marb3
AB.Marb4
AB.Marb5
AB.CR1
AB.CR2
Sporadic cases
AB.MarbS1
AB.MarbS2
AB.NijS1
AB.FreiS1
AB.ParisS1
Poly
hydramnios
Gestational
age (weeks)
Iso-/hypos
thenuriab
Hyper –
calciuriac
Nephro –
calcinosisd
Alkalosise
Hypokalemiaf and/or
elevated FEK+g
+
+
+
+
+
+
+
+
+
+
+
+
+
+
32
35
33
35
28
32
24
33
31
32
32
30
35
34
+
+
+
+
+
+
+
+
+
+
+
+
nd
nd
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
28
34
31
34
+
+
+
+
+
+
+
+
+
+
–
+
+
+
+
+
+
+
+
+
+
35
+
+
+
+
+
aWithout indomethacin therapy; burine osmolality ≤300 mosmol/kg with plasma osmolality >290 mosmol/kg; crenal calcium excretion >0.6 mg/mg creatinine; drenal
sonographic evidence of nephrocalcinosis; eserum bicarbonate >26 mmol/l; fserum potassium <3.3 mmol/l; gfractional excretion of potassium >15%.
Sporadic ABS patients are indicated by the suffix, S1.
nd, not done.
19
Human
Genetics,
1997,
6, No.
NucleicMolecular
Acids Research,
1994,
Vol. Vol.
22, No.
1 1
19
Figure 1. Pedigrees and haplotypes of antenatal Bartter syndrome families not linked to the chromosome 15NKCC2 interval. For each child, the paternal chromosome
is represented on the left and the maternal chromosome on the right. The thin lines indicate uninformative markers.
Recent studies have demonstrated mutations in the gene
encoding the mTAL luminal bumetanide-sensitive Na-K-2Cl
cotransporter (NKCC2) in a cohort of five ABS kindreds (28).
These genetic data indicate that the antenatal variant of Bartter
syndrome results from a primary defect in mTAL chloride
transport. This hypothesis is consistent with clinical studies
demonstrating that ABS patients have an impaired diuretic
response to furosemide (29). However, in applying haplotype
analyses with polymorphic chromosome 15 markers tightly
linked to NKCC2 in our cohort of 14 antenatal Bartter syndrome
families, we excluded NKCC2 as the disease-susceptibility locus
in six families.
Transepithelial chloride transport in the mTAL involves an
interplay between the apical Na-K-2Cl cotransporter (NKCC2),
the luminal, ATP-regulated potassium channel, ROMK (KCNJ1),
the basolateral chloride channel (ClC), a basolateral K-Cl
cotransporter, and the Na-K-ATPase (reviewed in 30,31).
Therefore, defective chloride transport may result from loss of
function mutations in any of the genes encoding these channels
or cotransporters. We now report the identification of mutations
in the ROMK gene (KCNJ1) in three kindreds and five sporadic
cases with ABS. These findings indicate that the antenatal variant
of Bartter syndrome is genetically heterogeneous.
RESULTS
Antenatal Bartter syndrome kindreds
We have identified a cohort of 14 families with antenatal Bartter
syndrome and five sporadic cases for study (Table 1). In nine of
these families, the affected children were the products of
consanguineous unions. All index cases had a history of
polyhydramnios and premature delivery. Postnatally, patients
developed hypokalemic alkalosis and all had hyposthenuria.
Prior to the initiation of indomethacin therapy, all affected infants
had failed to thrive with growth parameters less than the third
percentile for age. Urinary excretion rates of prostaglandin E2
(PGE2) and/or its major metabolite, prostaglandin E-M (PGE-M)
were markedly elevated (100% above upper limit of normal) in
all patients in whom these measurements were performed.
Hypercalciuria with associated sonographic evidence of
nephrocalcinosis was demonstrated in all 14 kindreds. Isolated
hypercalciuria was detected in one sporadic case who was treated
in the immediate postnatal period with indomethacin. Renal
biopsies obtained for three unrelated hypercalciuric infants
(index cases from families AB.Paris2, AB.Marb5 and
AB.MarbS2) revealed medullary calcifications with associated
tubular atrophy, interstitial inflammation and fibrosis.
Genotype analysis
Initially, we sought to determine whether the NKCC2 gene was the
disease-susceptibility locus in our ABS families. Markers tightly
linked to the NKCC2 locus on chromosome 15 were typed in these
kindreds. Lod scores for this set of markers were not conclusive
(data not shown), which raised the possibility of genetic heterogeneity in our ABS cohort. Haplotypes constructed for these
pedigrees demonstrate that six ABS families (AB.Paris2, AB.Paris5,
AB.Paris6, AB.Marb3, AB.Marb4 and AB.Marb5) are not linked
to the NKCC2 locus either because two affected siblings do not
share the same haplotypes or an affected child from a consanguineous union is not homozygous for markers of the NKCC2
gene interval (Fig. 1). Of note, there were no clinical or laboratory
features that distinguish the group linked to NKCC2 and the group
unlinked to this locus (Table 1).
In those families where the NKCC2 gene is excluded as the
disease-susceptibility locus, we evaluated whether ABS is linked
to the gene encoding ROMK, the inwardly-rectifying, ATP-regulated, potassium channel. Fluorescence in situ hybridization has
mapped the ROMK gene, KCNJ1, to chromosome 11q24–25
(32,33). Therefore, we typed these families for a set of five
microsatellite markers that spanned a 14 cM candidate interval on
20
Human Molecular Genetics, 1997, Vol. 6, No. 1
Figure 2. Pedigrees and haplotypes of Bartter syndrome families with apparent linkage to chromosome 11 markers that flank theKCNJ1 gene. In the consanguineous
families, the regions of homozygosity are framed with a solid line. In those families in which consanguinity cannot be confirmed, regions of homozygosity are framed
with a dotted line.
chromosome 11q24–q25. The haplotype data for these six
families suggested linkage between ABS and the KCNJ1 gene
(Fig. 2). In two of the consanguineous families, (AB.Paris2 and
AB.Paris6), the affected children are homozygous for the 9 cM
interval between D11S912 and D11S1320. Since a homozygous
mutation has been found in family AB.Marb3 (see below), the
region of homozygosity defined by this family allows us to
position the KCNJ1 gene in the genetic interval flanked by
D11S912 and D11S910. Given the haplotype data, we initiated a
search for mutations in the KCNJ1 gene in these six families and
the five sporadic cases.
Mutation analysis of the human KCNJ1 gene in
antenatal Bartter syndrome kindreds
Previous studies by Shuck et al. have demonstrated that the gene
encoding human ROMK produces five distinct transcripts by
differential splicing of its five exons (34). All five transcripts
share a common 3′ exon (exon 5) that encodes the majority of the
channel protein. The largest product, ROMK-1, is initiated at a
start codon in exon 4, contains 391 amino acid residues, and is the
most efficiently translated isoform (34). Based on the sequence
of ROMK-1, sets of primer pairs were used to amplify the coding
sequence for exon 4 and exon 5 from genomic DNA of our test
families (Table 2). The amplified products were analyzed by
single-strand conformation polymorphism analysis (SSCA).
Aberrant SSCA patterns were detected for exon 5 in three
consanguineous families (AB.Paris2, AB.Marb3 and AB.Marb4)
and the five sporadic cases. The affected children in the
consanguineous families were homozygous for these aberrations.
None of these aberrant SSCA patterns has been observed in
examination of at least 50 normal unrelated Caucasian controls
(100 test alleles).
In total, 11 mutations were identified and four of these
mutations were homozygous (Table 3; Fig. 3). Sequence analysis
reveals that these mutations are distributed throughout the
ROMK-1 protein (Fig. 4). Nine mutations are missense, causing
substitutions of residues which are highly conserved in wild-type
human, rat and Xenopus laevis. One mutation introduces a stop
codon that deletes the terminal 63 residues of the channel protein
and one 4 bp deletion causes a frameshift that results in a
truncated, 351 amino acid protein. All of these mutations were
inferred to alter the ROMK protein structure.
Mutations in the KCNJ1 gene have not been detected in
families AB.Paris5, AB.Paris6 or AB.Marb5. It is possible that
these families have KCNJ1 gene mutations involving the
exon–intron splice junctions of exon 4 or exon 5 or in an as yet
unexamined exon. In addition, we cannot exclude the possibility
that another gene may be the disease-susceptibility locus in these
families.
DISCUSSION
Antenatal Bartter syndrome is a phenotypically distinct variant of
the Bartter-like syndromes characterized by polyhydramnios
with preterm delivery, severe salt wasting, hyposthenuria, and
hypercalciuria (Table 1). Our identification of 11 independent
mutations in the gene KCNJ1 that segregate specifically with
ABS and result in either non-conservative amino acid substitutions in important functional domains or premature truncation of
the channel protein indicate that defects in KCNJ1 cause the
antenatal variant of Bartter syndrome. These findings demonstrate that ABS is genetically heterogeneous, involving mutations
in KCNJ1 as well as NKCC2 (28).
21
Human
Genetics,
1997,
6, No.
NucleicMolecular
Acids Research,
1994,
Vol. Vol.
22, No.
1 1
21
Table 2. PCR primers for SSCA analysis of the human KCNJ1 gene exons 4 and 5 (5′→3′)
Primer name
Forward primer
Reverse primer
hROMKex4
GCATAGAAAGACCAACAA
ACTTACCAACGTGTCAAA
hROMKex5.5
TTGATGTGTTTTCAGATCA
GGAGGTCTTTGTGAATGTA
hROMKex5.4
AAGTGGAGATACAAAATGACC
GCAGAAAAATGGCAGTGG
hROMKex5.3
TATGGATTCAGGTGTGTG
AAAATAATGGTCTCTCCT
hROMKex5.2
CTTATTGGCAGTCACATTTA
TATTTCCCTTCCTTTGTCTT
hROMKex5.1
CTACCGTTTTGCTCCCATA
ACTTTGCTTTACTCCTGTTGAA
Based on the sequence previously reported by Shuck et al. (34), a single primer pair was designed to amplify exon 4 and a set of nested primer pairs were designed
to amplify the coding region of exon 5.
Table 3. Mutations in the human ROMK gene (KCNJ1) in antenatal Bartter syndrome
Ethnic origin
Mutation
Nucleotide
Consequence
Mutation #
North Africa
A198T
G→A at 1153
Ala→Thr at 198
1
AB.Marb3
India
G167E
G→A at 1062
Gly→Glu at 167
2
AB.Marb4
Turkey
D108H
G→C at 883
Asp→His at 108
3
Germany/Croatia
P110L
C→T at 890
Pro→Leu at 110
4
V72E
T→A at 776
Val→Glu at 72
5
Families
AB.Paris2
Sporadic cases
AB.MarbS1
AB.MarbS2
Germany
V315G
T→G at 1505
Val→Gly at 315
6
1557del AAAG
4 bp deletion from 1557
frameshift→Stop 36 aa downstream
7
AB.NijS1
Netherlands
W99C
G→T at 858
Trp→Cys at 99
8
AB.FreiS1
Belgium
A198T
G→A at 1153
Ala→Thr at 198
1
R338X
C→T at 1573
Arg→Stop at 338
9
V122E
T→A at 926
Val→Glu at 122
10
D74Y
G→T at 781
Asp→Tyr at 74
11
AB.ParisS1
France
Sense strand sequences are shown. Amino acids are numbered from the first Met in exon 4. The mutation # refers to the mutation position in Figure 4. Nucleotides
are numbered as in Shuck et al. (34).
The ROMK gene is a disease-susceptibility gene in
antenatal Bartter syndrome
KCNJ1 encodes the luminal, ATP-sensitive, potassium channel in
the distal nephron (35–37). This channel belongs to a family of
inwardly rectifying potassium channels (Kir family; ROMK is
Kir1.1) and is characterized by little-to-no voltage dependence,
inward rectification, exquisite pH-sensitivity and modulation by
ATP (31,38–40). The ROMK channel is involved in potassium
secretion in the TAL and the distal nephron. In the mTAL, this
secretory function recycles potassium across the apical membrane, thus ensuring that adequate luminal potassium is available
for the efficient functioning of the Na-K-2Cl cotransporter (41).
Inhibition of luminal potassium channels reduces NaCl reabsorption in this nephron segment (42). In human kidney, differential
splicing produces five distinct transcripts of ROMK. While the
relative distribution of these isoforms in distal nephron segments
has not been established, exon 5 is common to all of these
isoforms and encodes the majority of the channel protein (33,34).
The detection of 11 independent mutations involving this
common exon predicts that ROMK activity is disrupted in our
cohort of ABS patients.
We have identified mutations involving the amphipathic M0
segment; the M1 and M2 transmembrane segments; the M1–H5
linker segment that comprises the channel pore; the putative
ATP-binding domain; and the C-terminus of the ROMK channel
protein (Fig. 4). Two non-conservative missense mutations
within the M0 region of the N-terminus have been identified,
V72G (mutation #5 in Table 3, Fig. 4) and D74Y (mutation #11).
This region of ROMK has been implicated in internal pH
regulation of ROMK (43). Since this channel is exquisitely
sensitive to reductions in cytosolic pH, these mutations in M0
could significantly alter ROMK channel function. The homozygous missense mutation, W99C (mutation #8), that involves
the substitution of a polar cystine for Trp-99 in the M1 domain is
likely to alter the configuration of the channel pore and thus,
would be predicted to disrupt channel function. The missense
mutations in AB.Marb4 (mutation #3), AB.MarbS1 (mutation
#4) and AB.ParisS1 (mutation #10) all cause non-conservative
amino acid changes in the M1–H5 linker region. Experimental
studies have demonstrated that this segment is important in
external pH regulation of the inwardly-rectifying potassium
channels, HIR (44) and ROMK. Mutation of the glycosylation
site, Asn-117, in this region of ROMK has been shown to
22
Human Molecular Genetics, 1997, Vol. 6, No. 1
23
Human
Genetics,
1997,
6, No.
NucleicMolecular
Acids Research,
1994,
Vol. Vol.
22, No.
1 1
23
Figure 4. Structural model of the ROMK channel protein and the position of mutations identified in antenatal Bartter syndrome patients. The H5 region contains the
putative channel pore and is located between the M1 and M2 membrane-spanning domains. The M0 segment is an amphipathic region. The protein kinase A (PKA:
open circles with ‘P’) and kinase C (PKC: closed circles with ‘P’) phosphorylation sites are indicated. The single possible N-linked glycosylation (N-glycosyl) site
is shown. Each symbol represents a single amino acid. Mutations are indicated by numbered arrows and each number corresponds to a mutation listed in Table 3.
Structural model adapted with permission (37).
dramatically decrease channel function (45). In addition, this
region is conserved in different classes of potassium channels and
in voltage-gated channels, single amino acid substitutions within
this segment result in loss of channel function or alterations in
single channel conductance and ion selectivity (37). Within the
M2 transmembrane segment, prior in situ mutagenesis studies
have indicated that residue Gly-167 is oriented so that it faces the
channel pore (46). Therefore, the homozygous mutation, G167E
(mutation #2), that substitutes a glutamic acid with its large polar
R group for Gly-167 would be predicted to alter the pore
conformation and thus, be deleterious for channel function. The
homozygous missense mutation, A198T (mutation #1), causes a
non-conservative substitution within a 63 amino acid segment
that lies just 3′ to the M2 domain and is involved in ROMK
channel regulation. This regulatory region contains important
phosphorylation sites [two protein kinase C (PKC) and one
protein kinase A (PKA) sites] and a Mg2+-ATP-binding motif
(Walker A site) that have been shown to be important in the
non-hydrolytic Mg2+-ATP regulation of ROMK (37). Both PKA
(35,47) and PKC (37,48) are important regulators of ROMK
function and structural alterations due to non-conservative amino
acid substitutions within this region may disrupt channel
regulation. Of note, the same A198T mutation was found in a
non-related compound heterozygote (AB.FreiS1). That this
Figure 3. Mutational analysis with the KCNJ1 gene in antenatal Bartter syndrome patients. Aberrant band patterns were identified by SSCA and evaluated by DNA
sequence analysis. Patients are identified as in Table 1. A representative sample of eight mutations in six patients are shown. Each composite panel shows the sequence
data from a control individual and the affected individual on the left and a representative autoradiograph of the SSCA pattern on the right. The sequence of the sense
strand is shown unless otherwise indicated. (A) In index case AB.Paris2, the first base of codon 198 is mutated from G to A causing the A198T substitution. (B) In
index case AB.Marb3, the second base of codon 167 is mutated from G to A resulting in the G167E change. (C) In index case AB.Marb4, a G to C change in the first
base of codon 108 cause the substitution, D108H (anti-sense strand data shown). (D) In the compound heterozygote, AB.MarbS2, the first mutant allele has a T to
G mutation in codon 315 resulting in the V315G substitution. In the second mutant allele, a 4 bp deletion produces a frameshift in codon 334, which causes the premature
termination of the protein (anti-sense strand data shown). (E) In the compound heterozygote, AB.FreiS1, the first mutant allele has the same G to A change in codon
198 as AB.Paris2, resulting in the A198T substitution (anti-sense strand data shown). In the second mutant allele, a C to T change in the first base of codon 338
introduces a stop codon. (F) In AB.NijS1, a homozygous G to T mutation in the last base of codon 99 causes the W99C change.
24
Human Molecular Genetics, 1997, Vol. 6, No. 1
specific point mutation occurs in two unrelated families and is not
evident in 100 test chromosomes provides further supporting
evidence that this mutation is pathogenic and not simply a
polymorphism. Finally, we have identified three mutant alleles
that alter the C-terminus of the ROMK protein. Since the
C-terminus has been shown to be important in determining pore
properties (49), mutations in this region could affect channel
function. In one mutant allele, a 4 bp deletion (mutation #7)
causes a frameshift mutation at codon 334 and results in the
introduction of a termination codon 36 amino acids downstream.
In another mutant allele, a stop codon is substituted at R338
(mutation #9), thereby deleting the terminal 53 residues of the
carboxy tail. These two mutations could alter phosphorylation at
the tyrosine kinase site at Y337, which is conserved in rat and
human ROMK. In the third mutant allele, a glycine for Val-315
substitution (mutation #6) lies adjacent to the cyclic AMP-dependent PKA phosphorylation site in the carboxy tail which has
been shown to be critical for channel function (47). In all affected
individuals, these mutations would be predicted to cause loss of
function in the channel protein.
Proposed pathogenesis of antenatal variant of Bartter
syndrome
In the mTAL, the functional coupling of ROMK and the luminal
Na-K-2Cl cotransporter (NKCC2) is crucial for NaCl reabsorption. Therefore, loss of function in ROMK as well as NKCC2 (28)
would be predicted to disrupt electrogenic chloride reabsorption
in the mTAL.
A primary defect in mTAL chloride transport is consistent with
the complex constellation of features evident in ABS patients.
Several lines of evidence support this proposed pathogenic
model. Chronic inhibition of salt reabsorption in the mTAL by
long-term furosemide treatment mimics the clinical and biochemical findings in ABS, i.e. hypokalemic hypochloremic
alkalosis, hypo- or isosthenuria, and impaired renal conservation
of both calcium and magnesium (5,50,51). Furthermore, ABS
patients have impaired diuretic, saluretic and hormonal responses
to furosemide (29). In contrast, furosemide sensitivity is maintained in Gitelman syndrome (52), which is caused by mutations
in the thiazide-sensitive Na-Cl cotransporter of the distal
convoluted tubule (10).
In patients with either KCNJ1 or NKCC2 mutations, defective
mTAL chloride transport results in increased NaCl delivery to the
distal nephron with consequent salt wasting, loss of concentrating
capacity and volume contraction. In addition, impaired electrogenic chloride transport in mTAL also inhibits the voltage-driven,
paracellular reabsorption of calcium and magnesium (53), thus
leading to hypercalciuria and in some patients, hypermagnesuria.
In ABS, stimulation of renal and systemic PGE2 formation is
likely to be a secondary phenomenon, at least in those patients
with KCNJ1 or NKCC2 mutations. The elevated PGE2 levels play
a central role in disease pathogenesis (16,27). Experimental data
demonstrate that elevated PGE2 activity stimulates the reninangiotensin-aldosterone axis (54), specifically inhibits ROMK
channel activity (48,55) and impedes NaCl transport in the mTAL
(56,57), as well as vasopressin-induced water reabsortion in the
collecting duct (58). Therefore, stimulation of PGE2 activity
exacerbates the primary defect in mTAL chloride transport,
contributes to the aldosterone-mediated hypokalemic metabolic
alkalosis and promotes hyposthenuria. As would be predicted,
suppression of PGE2 formation with cyclo-oxygenase inhibitors,
such as indomethacin, attenuates the salt wasting and hypokalemic alkalosis and converts hyposthenuria to isosthenuria (29).
In summary, our identification of molecular defects in the
KCNJ1 gene, coupled with the previously reported mutations in
NKCC2 (28), indicate that antenatal Bartter syndrome is genetically heterogeneous. Together, these data support the hypothesis
that ABS involves a primary mTAL transport defect (29). While
we cannot exclude the possibility that there may be additional
gene defects that cause ABS, these data provide new insights into
the complex, interrelated mechanisms involved in the pathogenesis of antenatal Bartter syndrome. In addition, our findings
expand the genetic basis for distinguishing Bartter-like syndromes from one another. While specific disorders within the
spectrum of Bartter-like syndromes can usually be differentiated
by applying rigorous clinical criteria, confounding presentations
are not unusual. Further identification of the basic genetic defects
in Bartter-like syndromes will provide potential molecular tools
for distinguishing specific disorders within this often confusing
spectrum, and for dissecting the complicated pathophysiology
involved in each disorder.
MATERIALS AND METHODS
Bartter syndrome families
Index cases from families, AB.Paris1, AB.Marb2, AB.Marb3 and
AB.Marb4, have been previously reported (5,19,24,29). The
index cases of the other 10 families and the five sporadic cases
were identified either in Neonatal Intensive Care Units or
Pediatric Nephrology Clinics at the collaborating institutions.
These studies were approved by local Ethics Committees and
Informed Consent was obtained from the parents.
Genotype analysis and haplotype construction
Using standard methods, genomic DNA was extracted from whole
blood lymphocytes of all family members studied. Markers tightly
linked to the NKCC2 locus on chromosome 15 were typed in all of
our families. The markers, D15S132 and D15S209 have been
previously reported (28). The markers, D15S132–D15S143–
D15S123 are all located on the YAC 956E3 (CEPH library) (59).
PCR analysis using a primer set derived from the human NKCC2
cDNA sequence amplified a 1200 bp fragment from genomic DNA
and from YAC 956E3, confirming that this YAC contains the
NKCC2 gene. Additional tightly linked markers (D15982, D15S126
and D15S121) were selected from the Genethon database (accessible at http://www.genethon.fr/genethon_en.html ). Highly polymorphic markers from the 14 cM interval on chromosome
11q24–25 that contains the KCNJ1 gene were selected from the
Genethon database (60). The following marker order for the 14 cM
interval between D11S934 and D11S1320 has been established by
radiation hybrid mapping (Genome Data Base; accessible at
http://gdbwww.gdb.org
):
centromere-D11S934–D11S912–
D11S874–D11S910–D11S1320-telomere.
Microsatellite polymorphisms were amplified by the polymerase chain reaction using forward primers labeled at the 5′-end
either with fluorescent dye or γ-32P. PCR reactions were
performed in a 15 µl volume containing 50 ng of genomic DNA,
1.5 mM MgCl2, 5 mM Tris (pH 8.3), 50 mM KCl, 10 pmol of each
primer and 0.5 U of Taq polymerase. After an initial denaturation
step at 94C for 5 min, PCR was conducted for 30 cycles with
25
Human
Genetics,
1997,
6, No.
NucleicMolecular
Acids Research,
1994,
Vol. Vol.
22, No.
1 1
denaturation at 94C for 45 s, annealing at 55C for 45 s and
extension at 72C for 45 s. The reaction was completed with a
final elongation step at 72C for 10 min. Amplified products were
separated on 6% polyacrylamide gels run under denaturing
conditions. The gels were analyzed either using the Genscan 672
software (Ver 1.2) or autoradiography. Linkage analysis was
performed using the LINKAGE programs (61). Haplotypes were
constructed from the genotype data. The most likely haplotypes
were inferred by minimizing the number of crossover events in
each sibship.
SSCA and DNA sequencing
Aberrant band patterns for the KCNJ1 gene were sought using
single-strand conformation polymorphism analysis (SSCA) (62).
Based on the sequence for the human KCNJ1 gene previously
described (34), an overlapping set of primer pairs were designed
(Table 2) and used to amplify the coding sequence for exons 4 and
5 from genomic DNA of our test families as well as 50 unrelated
controls. PCR was performed as described above except that the
products were labeled by inclusion of 1 µCi of [α-32P]dCTP or
[α-33P]dCTP in the reaction mixture. Amplified products were
separated either at room temperature or at 4C by electrophoresis
at 40 W constant power for 6–8 h on 0.5× MDE gels (AT
Biochem) prepared in 0.6× TBE [54 mM Tris borate (pH 7.8)].
Following autoradiography, conformational variants were excised from the gel and the DNA was isolated by centrifugation
through a Micropure 0.22 filter (Amicon, Inc.). The isolated
product was then re-amplified with the same primer pair and
subjected to direct DNA sequence analysis using an ABI 373A
automated DNA sequencer (Applied Biosystems). Alternatively,
non-radioactive PCR was performed and gels were silver-stained
as previously described (63). Direct sequencing was performed
after reamplification of the remaining PCR product using a direct
blotting electrophoresis sequencing system (GATC 1500, GATC,
Konstanz, Germany). In all cases, DNA sequences were confirmed by sequencing both strands.
ACKNOWLEDGEMENTS
The authors are indebted to the participating families for their
cooperation. LK is a recipient of a fellowship grant from the
Deutsche Forschungsgemeinschaft (Ka 1252/1-1). Group 1 was
supported in part by grants from the Deutsche Forschungsgemeinschaft (HWS: Se 263/11-1) and from the Kempkes Stiftung
Marburg. Group 4 was supported by the Zentrum Klinische
Forschung, A1 Freiburg. Financial support for Group 5 was
provided by the Dutch Kidney Foundation (Grant C96-1540).
Funding was also provided by grants from the University of
Alabama Health Services Foundation and the Alabama Kidney
Foundation (LG-W). The Collaborative Study Group is supported by a Concerted Action ‘Identification of the gene defects
in Bartter syndrome and Gitelman syndrome’ funded by the
European Economic Community (EEC) within BIOMED II
program PL950260. This study was performed in cooperation
with the International Study of Genetic Renal Disease (ISGRD).
25
REFERENCES
1. Bartter, F.C., Pronove, P., Gill, J., and MacCardle, R. (1962) Hyperplasia of
the juxtaglomerular complex with hyperaldosteronism and hypokalemic
alkalosis. Am. J. Med. 33, 811–828.
2. Hogewind, B., Van Brummelen, P., and Veltkamp, J. (1981) Bartter’s
syndrome: an autosomal recessive disorder? Study of four patients in one
generation of the same pedigree and their relatives. Acta Med. Scand. 209,
463–67.
3. Schwartz, I., and Alon, U. (1996) Bartter syndrome revisited. J. Nephrol. 9,
81–87.
4. Gitelman, H., Graham, J., and Welt, L. (1966) A new familial disorder
characterized by hypokalemia and hypomagnesemia. Trans. Assoc. Am.
Physicians 79, 221–235.
5. Seyberth, H.W., Rascher, W., Schweer, H., Kuehl, P.C., Mehls, O., and
Schaerer, K. (1985) Congenital hypokalemia with hypercalciuria in preterm
infants: a hyperprostaglandinuric tubular syndrome different from Bartter’s
syndrome. J. Pediatr. 107, 694–701.
6. Clive, D. (1995) Bartter’s syndrome: the unsolved puzzle. Am. J. Kidney Dis.
25, 813–823.
7. Rodriguez-Soriano, J., Vallo, A., and Garcia-Fuentes, M. (1987) Hypomagnesemia of hereditary renal origin. Pediatr. Nephrol. 1, 465–72.
8. Gitelman, H. (1992) Hypokalemia, hypomagnesemia, and alkalosis: A rose is
a rose - or is it? J. Pediatr. 129, 79–80.
9. Bettinelli, A., Bianchetti, M., Girardin, E., Caringella, A., Cecconi, M.,
Appiani, A.C., Pavanello, L., Gastaldi, R., Isimbaldi, C., and Lama, G. (1992)
Use of calcium excretion values to distinguish two forms of primary renal
tubular Hypokalemia Alkalosis: Bartter and Gitelman syndromes. J. Pediatr.
120, 38–43.
10. Simon, D., Nelson-Williams, C., Bia, M., Ellison, D., Karet, F., Molina, A.,
Vaara, I., Iwata, F., Cushner, H., M, M.K., Gainza, F., Gitelman, H., and
Lifton, R. (1996) Gitelman’s variant of Bartter syndrome, inherited
hypokalemic alkalosis, is caused by mutations in the thiazide-sensitive Na-Cl
cotransporter. Nature Genet. 12, 24–30.
11. Lemmink, H., Heuvel, L.v.d., Dijk, H.v., Merkx, G., Smilde, T., Taschner, P.,
Monnens, L., Hebert, S., and Knoers, N. (1996) Linkage of Gitelman
syndrome to the thiazide-sensitive sodium-chloride cotransporter gene with
identification of mutations in Dutch families. Pediatr. Nephrol. 10, 403–407.
12. Karolyi, L., Ziegler, A., Pollak, M., Fischbach, M., Grzeschik, K.-H., Koch,
M., and Seyberth, H. (1996) Gitelman’s syndrome is genetically distinct from
other forms of Bartter syndrome. Pediatr. Nephrol. 10, 551–554.
13. Gladziwa, U., Schwarz, R., Gitter, A., Bijman, J., Seyberth, H., Beck, F., Ritz,
E. and Gross, P. (1995) Chronic hypokalaemia of adults: Gitelman’s
syndrome is frequent but classical Bartter’s syndrome is rare. Nephrol. Dial.
Transplant. 10, 1607–1613.
14. Geven, W., Willems, J., Schroder, C. and Monnens, L. (1994) Study of the
pathophysiology of Bartter/Gitelman’s syndrome. Magnesium Bull. 16,
29–36.
15. Fujita, T., Sakaguchi, H., Shibagaki, M., Fukui, T., Nomura, M., and
Sekiguchi, S. (1977) The pathogenesis of Bartter’s syndrome. Functional and
histologic studies. Am. J. Med. 63, 467–74.
16. Gill, J., and Bartter, F.C. (1978) Evidence for a prostaglandin independent
defect in chloride reabsorption in the loop of Henle as a proximal cause of
Bartter’s syndrome. Am. J. Med. 65, 766–772.
17. Bartter, F.C. (1980) On the pathogenesis of Bartter syndrome. Miner.
Electroyte Metab. 3, 61–65.
18. Stein, J.H. (1985) The pathogenetic spectrum of Bartter’s syndrome. Kidney
Int. 28, 85–93.
19. Deschenes, G., Burguet, A., Guyot, C., Hubert, P., Garabedian, M., Dechaux,
M., Loirat, C., and Broyer, M. (1993) Forme antenatale du syndrome de
Bartter. Ann. Pediatr. 40, 95–101.
20. Proesmans, W., Devlieger, H., Assche, A.V., Eggermont, E., Vandenberghe,
K., Lemmens, F., Sieprath, P. and Lijnen, P. (1985) Bartter syndrome in two
siblings - antenatal and neonatal observations. Int. J. Pediatr. Nephrol. 6,
63–70.
21. Ohlsson, A., Sieck, U., Cumming, W., Akhtar, M., and Serenius, F. (1984)
Bartter syndrome associated with hydraamnios, prematurity, hypercalciuria
and nephrocalcinosis. Acta Pediatr. Scand. 73, 868–874.
22. Proesmans, W., Massa, G., Vanderberghe, K. and Assche, A.V. (1987)
Prenatal diagnosis in Bartter syndrome. Lancet I, 394.
23. Fanconi, A., Schachenmann, G., Nussli, R., and Prader, A. (1971) Chronic
hypokalemia with growth retardation, normotensive hyperrenin-hyperaldosteronism (Bartter’s syndrome) and hypercalciuria. Helv. Pediatr. Acta. 2,
144–63.
26
Human Molecular Genetics, 1997, Vol. 6, No. 1
24. Leonhardt, A., Timmermanns, G., Roth, B., and Seyberth, H. (1992) Calcium
homeostasis and hypercalciuria in hyperprostaglandin E syndrome. J.
Pediatr. 120, 546–54.
25. McCredie, D.A., Rotemberg, E., and Williams, A.L. (1974) Hypercalciuria in
potassium losing nephropathy. A variant of Bartter’s syndrome. Aust. Pediatr.
J. 10, 286–295.
26. Shoemaker, L., Welch, T., Bergstrom, W., Abrams, S., Yergey, A., and Vieira,
N. (1993) Calcium kinetics in the hyperprostaglandin E syndrome. Pediatr.
Res. 33, 92–96.
27. Seyberth, H., Koniger, S., Rascher, W., Kuhl, P., and Schweer, H. (1987) Role
of prostaglandins in hyperprostaglandin E syndrome and in selected renal
tubular disorders. Pediatr. Nephrol. 1, 491–497.
28. Simon, D., Karet, F., Hamdan, J., DiPetro, A., Sanjad, S., and Lifton, R.
(1996) Bartter’s syndrome, hypokalemic alkalosis with hypercalciuria, is
caused by mutations in the Na-K-2Cl cotransporter NKCC2. Nature Genet.
13, 183–188.
29. Kockerling, A., Reinalter, S., and Seyberth, H. (1996) Impaired response to
furosemide in hyperprostaglandin E syndrome: Evidence for a tubular defect
in the loop of Henle. J. Pediatr. 129, 519–528.
30. Giebisch, G., and Wang, W. (1996) Potassium transport: from clearance to
channels and pumps. Kidney Int. 49, 1624–1631.
31. Lang, F., and Rehwald, W. (1992) Potassium channels in renal epithelial
transport regulation. Physiol. Rev. 72, 1–32.
32. Krishnan, S., Desai, T., Ward, D., and Haddad, G. (1995) Isolation and
chromosomal localization of a human ATP-regulated potassium channel.
Hum. Genet. 96, 155–160.
33. Yano, H., Philipson, L., Kugler, J., Tokuyama, Y., Davis, E., Beau, M.l.,
Nelson, D., Bell, G., and Takeda, J. (1994) Alternative splicing of human
inwardly rectifying K+ channel ROMK1 mRNA. Mol. Pharmacol. 45,
854–860.
34. Shuck, M., Bock, J., Benjamin, C., Tsai, T.-D., Lee, K., Slightom, J., and
Bienkowski, M. (1994) Cloning and characterization of multiple forms of the
human kidney ROM-K potassium channel. J. Biol. Chem. 269,
24261–24270.
35. McNicholas, C., Yang, Y., Giebisch, G., and Hebert, S. (1996) Molecular site
for nucleotide binding on an ATP-sensitive renal K+ channel (ROMK2). Am.
J. Physiol. 271, F275–F280.
36. Boim, M., Ho, K., Shuck, M., Bienkowski, M., Block, J., Slightom, J., Yang,
Y., Brenner, B., and Hebert, S. (1995) ROMK inwardly rectifying ATP-sensitive K+ channel. II. cloning and distribution of alternative forms. Am. J.
Physiol. 268, F1132–F1140.
37. Ho, K., Nichols, C., Lederer, W., Lytton, J., Vassilev, P., Kanazirska, M., and
Hebert, S. (1993) Cloning and expression of an inwardly rectifying
ATP-regulated potassium channel. Nature 362, 31–38.
38. McNicholas, C., Wang, W., Ho, K., Hebert, S., and Giebisch, G. (1994)
Regulation of ROMK1 K+ channel activity involves phosphorylation
processes. Proc. Natl. Acad. Sci. USA 91, 8077–8081.
39. Tsai, T., Shuck, M., Thompson, D., Bienkowski, M., and Lee, K. (1995)
Intracellular H+ inhibits a cloned rat kidney outer medulla K+ channel
expressed in Xenopus oocytes. Am. J. Physiol. 268, C1173–C1178.
40. Wang, W., Sackin, H., and Giebisch, G. (1992) Renal potassium channels and
their regulation. Annu. Rev. Physiol. 54, 81–96.
41. Hebert, S., and Andreoli, T. (1984) Control of NaCl transport in the thick
ascending limb. Am. J. Physiol. 246, F745-F756.
42. Hebert, S., Friedman, P., and Andreoli, T. (1984) The effects of antidiuretic
hormone on cellular conductive pathways in mouse medullary thick
ascending limbs of Henle. I. ADH increases transcellular conductance
pathways. J. Membrane Biol. 80, 201–219.
43. Doi, T., Fakler, B., Schultz, J., Schulte, U., Brandle, U., Weidemann, S.,
Zenner, H.-P., Lang, F., and Ruppersberg, J. (1996) External K+ and
intracellular pH allosterically regulate renal Kir 1.1 channels. J. Biol. Chem.
271, 17261–17266.
44. Coulter, K., Perier, F., Radeke, C., and Vandenberg, C. (1995) Identification
and molecular characterization of a pH-sensing domain for the inward
rectifier potassium channel, HIR. Nephron 15, 1157–1168.
45. Schwalbe, R., Wang, Z., Wible, B., and Brown, A. (1995) Potassium channel
structure and function as reported by a single glycosylation sequon. J. Biol.
Chem. 270, 15336–15340.
46. Choe, S., Stevens, C., and Sullivan, J. (1995) Three distinct structural
environments of a transmembrane domain in the inwardly rectifying
potassium channel ROMK1 defined by perturbation. Proc. Natl. Acad.
Sci.USA 92, 12046–12049.
47. Xu, Z.-C., Yang, Y., and Hebert, S. (1996) Phosphorylation of the
ATP-sensitive, inwardly rectifying K+ channel, ROMK, by cyclic AMP-dependent protein kinase. J. Biol. Chem. 271, 9313–9319.
48. Macica, C., Yang, Y., and Hebert, S. (1996) Arachidonic acid inhibits activity
of the cloned renal K channel, ROMK1. Am. J. Physiol. 271, F588–F594.
49. Taglialatela, M., Wible, B., Caporaso, R., and Brown, A. (1994) Specifications of pore properties by the carboxyl terminus of inwardly rectifying K+
channels. Science 264, 844–847.
50. Hufnagle, K., Khan, S., Penn, D., Cacciarelli, A., and Williams, P. (1982)
Renal calcifications: a complication of long-term furosemide therapy in
pre-term infants. Pediatrics 70, 360–363.
51. Venkataraman, Han, B., Tsang, R., and Daugerthy, C. (1983) Secondary
hyperparathyroidism and bone disease in infants receiving long-term
furosemide RX. J. Pediatr. Adolesc. Med. [Am. J. Dis. Child.]. 137,
1157–1161.
52. Sutton, R., Mavichak, V., Halabe, A., and Wilkins, G. (1992) Bartter’s
syndrome: Evidence suggesting a distal tubular defect in a hypocalciuric
variant of the syndrome. Miner. Electrolyte Metab. 18, 43–51.
53. Greger, R., and Heidland, A. (1991) Action and clinical use of diuretics. In
Cameron, S., Davison, A.M., Grunfeld, J.P., eds, Textbook of Clinical
Nephrology. London: Oxford University Press. pp 197–223.
54. Schlondorff, D. and Ardaillou, R. (1986) Prostaglandins and other arachicondic acid metabolites in the kidney. Kidney Int. 29, 108–119.
55. Wang, W.-H. (1994) Two types of K+ channel in thick ascending limb of rat
kidney. Am. J. Physiol. 267, F599–F605.
56. Stokes, J. (1979) Effect of prostaglandin E2 on chloride transport across the
rabbit thick ascending limb of Henle. J. Clin. Invest. 64, 495–502.
57. Culpepper, R., and Andreoli, T. (1983) Interactions among prostaglandin E2,
antidiuretic hormone and cyclic adenosine monophosphate in modulating Cl
absorption in single mouse medullary thick ascending limbs of Henle. J. Clin.
Invest. 71, 1588–1591.
58. Nakao, A., Allen, M., Sonnenburg, W., and Smith, W. (1989) Regulation of
cAMP metabolism by PGE2 in cortical and medullary thick ascending limbs
of Henle’s loop. Am. J. Physiol. 256, C652–C657.
59. Fougerousse, F., Broux, O., Richard, I., Allamand, V., Pereira de Souza, A.,
Bourg, N., Brenguier, L., Devaud, C., Pasturaud, P., Roudaut, C., Chiannilkulchai, B., Hillaire, D., Bui, H., Chumakov, I., Weissenbach, J., Cherif, D.,
Cohen, D., and Beckmann, J.S. (1994) Mapping of a chromosome 15 region
involved in limb girdle muscular dystrophy. Hum. Mol. Genet. 3, 285–293.
60. Dib, C., Faure, S., Fizames, C., Samson, D., Drout, N., Vignal, A., Millasseau,
P., Marc, S., Hazan, J., Seboun, E., Lathrop, M., Gyapay, G., Morissette, J.,
and Weissenbach, J. (1996) A comprehensive genetic map of the human
genome based on 5,264 microsatellites. Nature 380, 152–154.
61. Lathrop, G.M., and Lalouel, J.M. ( l984) Easy calculations of lod scores in
genetic risks on small computers. Am. J. Hum. Genet. 36, 460–465.
62. Orita, M., Suzuki, Y.H., Sakiya, T., and Hayaski, K. (1989) Rapid and
sensitive detection of point mutations and DNA polymorphisms using the
polymerase chain reaction. Genomics 5, 874–879.
63. Hiort, O., Wedtke, A., Zollmer, A. and Sinneckes, G. (1994) Detection of
point mutations in the androgen receptor gene using non-isotopic single
strand conformation polymporphism analysis. Hum. Mol. Genet. 3,
1163–1166.