Detection of mutations in KLHL3 and CUL3 in families with FHHt (familial hyperkalaemic hypertension or Gordon's syndrome)

The study of families with rare inherited forms of hypo- and hyper-tension has been one of the most successful strategies to probe the molecular pathophysiology of blood pressure control and has revealed dysregulation of distal nephron Na+ reabsorption to be a common mechanism. FHHt (familial hyperkalaemic hypertension; also known as Gordon's syndrome) is a salt-dependent form of hypertension caused by mutations in the regulators of the thiazide-sensitive Na+–Cl− co-transporter NCC [also known as SLC12A3 (solute carrier family 12 member 3)] and is effectively treated by thiazide diuretics and/or dietary salt restriction. Variation in at least four genes can cause FHHt, including WNK1 [With No lysine (=K) 1] and WNK4, KLHL3 (kelch-like family member 3), and CUL3 (cullin 3). In the present study we have identified novel disease-causing variants in CUL3 and KLHL3 segregating in 63% of the pedigrees with previously unexplained FHHt, confirming the importance of these recently described FHHt genes. We have demonstrated conclusively, in two unrelated affected individuals, that rare intronic variants in CUL3 cause the skipping of exon 9 as has been proposed previously. KLHL3 variants all occur in kelch-repeat domains and so probably disrupt WNK complex binding. We have found no evidence of any plausible disease-causing variants within SLC4A8 (an alternative thiazide-sensitive sodium transporter) in this population. The results of the present study support the existing evidence that the CUL3 and KLHL3 gene products are physiologically important regulators of thiazide-sensitive distal nephron NaCl reabsorption, and hence potentially interesting novel anti-hypertensive drug targets. As a third of our non-WNK FHHt families do not have plausible CUL3 or KLHL3 variants, there are probably additional, as yet undiscovered, regulators of the thiazide-sensitive pathways.

Correspondence: Dr Mark Glover (email mark.glover@nottingham.ac.uk). 49 % of heart disease are attributable to blood pressure [1][2][3]. The current limitations in anti-hypertensive therapeutics are perhaps not surprising since for most affected individuals the molecular mechanisms driving their hypertension remain undefined.
Although rare, Mendelian forms of hypo-and hypertension represent experiments of Nature that have informed our understanding of the physiology of the distal nephron. Remarkably, given the variety of physiological systems that affect arterial pressure, all of these Mendelian syndromes for which the molecular mechanism is understood converge around a common theme: distal nephron sodium wasting in hypotensive syndromes and excessive sodium reabsorption in hypertensive conditions [4].
Although the amiloride-sensitive ENaC (epithelial sodium channel) has classically dominated research interests, NaCl reabsorption via the thiazide-sensitive Na + -Cl − co-transporter NCC [also known as SLC12A3 (solute carrier family 12 member 3)] is at least as important [5]. Thiazide diuretics are potent antihypertensive agents [6] and mimic the effects of loss-of-function mutations of NCC observed in the hypotensive monogenic syndrome of Gitelman [7]. Moreover, the heritable condition of FHHt (familial hyperkalaemic hypertension) results from increased sodium reabsorption via NCC and is effectively ameliorated by thiazide diuretics and/or dietary sodium restriction [8].
The current model for the regulation of NCC is complex and involves a scaffold of at least 12 interacting proteins centred on a WNK signalling cascade, with intermediary STE20 (sterile 20) kinases [SPAK (STE20/SPS1-related proline/alanine-rich kinase) and OSR1 (oxidative stress-responsive kinase-1)] activated by WNKs which in turn activate NCC [17][18][19]. CUL3 and KLHL3 are both components of the cullin/Ring E3 ligase ubiquitination pathway and at least some variants of KLHL3 appear to affect NCC via the control of WNK1 ubiquitination [15,20].
We have identified previously three FHHt pedigrees carrying WNK4 mutations (D564H, E562K and Q565E) [21], but none carrying WNK1 mutations. To assess whether our remaining pedigrees with FHHt and without WNK1/4 mutations had either CUL3 or KLHL3 mutations, we undertook NGS (nextgeneration sequencing) of these genes and also screened an alternative thiazide-sensitive sodium transporter (SLC4A8) hypothesized to be an additional candidate [14].

Study population
The present study was carried out in accordance with the Declaration of Helsinki (2013) of the World Medical Association. Study participants with an FHHt phenotype were identified through tertiary specialist hypertension clinics in the U.K. and Australia. Diagnosis of FHHt was confirmed by the authors. All affected patients were Caucasian and shared a phenotype of persistent hyperkalaemia (plasma potassium >5.0 mmol/l in blood collected without stasis) and hypertension (>140/90 mmHg for adults) following exclusion of the relevant co-morbidities and pharmacotherapies. Detailed phenotypes of the affected individuals are given in Supplementary Figure S1 (at http://www.clinsci.org/ cs/126/cs1260721add.htm). All non-affected individuals demonstrated plasma potassium <5 mM/l. The disparity in ages prevented comparison of age-related blood pressure between affected and non-affected individuals. DNA was extracted using a standard method from venous blood acquired following informed consent (Princess Alexandra Hospital Human Research Ethics Committee ID EC00 167 in Australia and National Research Ethics Committee reference 12/EM/0317 in the U.K.).

DNA analysis
CUL3, KLHL3 and SLC4A8 genes were sequenced in the affected proband of each family using NGS. PCR amplicons covering all coding exons and exon/intron boundaries were prepared from genomic DNA (Fluidigm Access Array TM ; the amplicons used are listed in Supplementary Table S1 at http://www.clinsci.org/cs/126/cs1260721add.htm) and sequenced on the Illumina HiSeq platform. Reads were aligned to the human reference sequence hg19 using the Burrows-Wheeler Aligner, and the Genome Analysis Toolkit was used for base recalibration, local realignment and variant calling, following published best practice guidelines, and as described previously [22]. Variants were filtered for rarity and protein consequence: variants altering the protein-coding sequence [missense and nonsense SNPs (single nucleotide polymorphisms), insertions or deletions, or intronic variants at the exon/intron boundary] that were absent from public databases [dbSNP, 1000 Genomes and the NHLBI ESP (National Heart, Lung, and Blood Institute Exome Sequencing Project) Exome Variant Server] were considered candidates. All candidates detected by NGS were confirmed in the proband and assessed for segregation in the pedigree using Sanger sequencing. Variants are reported using Human Genome Variation Society standard nomenclature (http://www.hgvs.org/ mutnomen/). The reference sequences used for each gene and protein are listed in Supplementary

RNA studies
The functional effects of putative splice variants were confirmed using RNA studies. Peripheral blood mononuclear cell RNA was isolated from 5 ml of whole blood using a PAXgene blood RNA kit (Qiagen) according to the manufacturer's instructions. The RNA was then transcribed using a Promega AMV reverse transcriptase kit (catalogue number A3500) according to manufacturer's instructions using either random primers (RT1) or a CUL3-specific primer (5 -TTATGCTACATATGTGTATAC-TTTGC-3 ; RT2). The resulting cDNA was then PCR-amplified using exon-specific primers to amplify exons 8-10 of the CUL3 transcript (forward, 5 -TCAACCTCAACTCCAGGTCTCC-3 and reverse, 5 -TGTTGCCTGAATTCATCCATCG-3 ). The PCR products were run on a 2 % agarose gel to visualize them, excised, cleaned using a Promega PCR clean-up kit and Sanger-sequenced on a Beckman CEQ 6800 sequencer. The expected PCR product sizes were 338 bp and 167 bp for the exon 8-10 and del9 transcripts respectively.

Paralogue mapping
For each gene we first identified paralogues using pre-defined Ensembl protein families (http://www.ensembl.org; release 70), and constructed a multiple sequence alignment using M-Coffee [23]. Reported Mendelian disease-causing variants (nonsynonymous SNPs causing a single non-terminal amino acid change) in paralogues of the FHHt genes were identified using the Human Gene Mutation Database Professional version (http://www.hgmd.cf.ac.uk; release 2012.3), and mapped to the equivalent residue of the FHHt gene in the multiple sequence alignment.

Exon-directed array and identity by descent analysis
Representative affected individuals in pedigrees 6, 7 and 8 were genotyped using the Illumina Infinium HumanExome BeadChip array. Pair-wise IBD (identity by descent) analysis was undertaken using PLINK version 1.0.7 [24] on the basis of a subset of 27 402 informative autosomal SNPs with a minor allele frequency >5 %. A proportion of IBD (PI_HAT) <0.05 was considered to indicate no excess of sharing (i.e. unrelated individuals).

RESULTS
Genetic analysis of 25 affected individuals from 16 families with FHHt who had already been screened and found negative for WNK1/4 mutations was performed. A total of 95 % of the targeted bases were sequenced adequately for variant calling. The sequencing depth and coverage achieved by gene and exon are shown in Supplementary Figure S2 (at http://www.clinsci. org/cs/126/cs1260721add.htm).
Affected individuals (n = 16) from ten of these 16 families were found to have CUL3 or KLHL3 variants not reported in the general population (Table 1 and Supplementary Figure S3 at http://www.clinsci.org/cs/126/cs1260721add.htm). We found no evidence of rare variants in SLC4A8 which segregated with disease phenotype.
As shown in Table 1 and Supplementary Figure S2, affected individuals from eight pedigrees carried variants that have been associated previously with FHHt, two in CUL3 and six in KLHL3. Affected individuals from two pedigrees carried variants unreported previously in CUL3 (c.1207-12T>A and c.1377 + 1G>T). In addition an affected individual from pedigree 5 was homozygous for a previously reported heterozygous KLHL3 variant (c.1499G>T; p.G500V) [15]. In keeping with previous observations, CUL3 mutations were intronic and probably affected splicing of exon 9, whereas KLHL3 mutations were non-synonymous exonic SNPs (Supplementary Figures S4 and S5 at http://www.clinsci.org/cs/126/cs1260721add.htm).
An affected individual from pedigree 1 had an alternative G>T variant at the same position in CUL3 as one from pedigree 2 (c.1377 + 1G>C; the original proband reported by Gordon et al. [9]). Both had a severe hyperkalaemic phenotype apparent during   childhood despite coming from different families and living on opposite sides of the globe. Although variants surrounding this exon 9/intron 9 acceptor splice site have been predicted to affect splicing of exon 9 [14], the present study has provided the first evidence of this effect in FHHt patients. Specifically, RNA from peripheral blood monocytes of the index case in pedigrees 1 and 2 contains exon 9-deficient transcripts from the mutated CUL3 allele (Figure 1). Pedigrees 6, 7 and 8 all carry the same KLHL3 p.L387P mutation that segregates completely with an FHHt phenotype, raising the question whether these families have a common founder. IBD analysis (Illumina Infinium HumanExome Bead-Chip) revealed that these pedigrees were no more related than by chance (PI_HAT = 0.0440), indicating that the mutation has probably arisen independently in each lineage. Although the KLHL3 R528H mutation has also been reported in three pedigrees [15], it was not established whether they shared a common founder. Hence in our pedigrees, KLHL3 p.L378P is the most commonly identified FHHt-causing KLHL3 mutation with robust evidence of independent founder mutations.
To assess the pathogenicity of the KLHL3 variants associated with FHHt, we used a Paralogue Annotation ap-proach [25]. KLHL3 is one of a family of evolutionarily related cytoskeletal BTB/kelch repeat proteins, variation in several of which cause Mendelian disease. Using multiple sequence alignment to identify structurally and functionally equivalent residues across the protein family, we observed that one of the KLHL3 variants reported previously to be associated with FHHt [14,15] (KLHL3 p.R384W) co-locates with a reported disease-causing variant in another member of the protein family KBTBD13 p.R248S {where KBTBD13 is kelch repeat and BTB [BR-C (Broad Complex), ttk (tramtrack) and bab (bric a brac)] (POZ) domain-containing 13}, which is associated with nemalin myopathy [26]. This suggests that the variants lie at a functionally important site conserved across the protein family that is intolerant of sequence variation. Similarly, two of the KLHL3 FHHt variants in our patients (L387P and A494T) are very close to the location of known diseasecausing variants in GAN (gigaxonin) [27], suggesting that these too are probably functionally important sites. GAN p.G368 and p.G474 (at which substitutions are associated with giant axonal neuropathy [27]) are equivalent to KLHL3 p.G388 and p.G496, and are adjacent to rare variants found in our FHHt pedigrees.

DISCUSSION
In the present study we have identified disease-causing variants in CUL3 and KLHL3 in 63 % of our pedigrees with FHHt who had been screened and found to be negative for WNK1/4 mutations, confirming recent reports of association between CUL3 and KLHL3 variants and FHHt [14,15]. In the case of CUL3 mutation at position c.1377 + 1 we report a second variant allele associated with a similar thiazide-responsive FHHt phenotype, strengthening further the case for a functional role of aberrant CUL3 function on sodium reabsorption in the distal nephron. We have also demonstrated that the predicted exon 9 splicing effect produced by c.1377 + 1G>T and c.1377 + 1G>A is, in fact, observed.
We have found that KLHL3 p.L387P associated with FHHt in three unrelated pedigrees, making this the most commonly occurring single FHHt mutation not only within our FHHt consortium, which includes three FHHt pedigrees carrying different WNK4 mutations (D564H, E562K and Q565E) [21], but also among all KLHL3 mutations reported to date [14,15]. That KLHL3 variants in our pedigrees are restricted to kelch repeats, and that other FHHt-associated KLHL3 variants cluster in these domains provides further support for disruption of WNK complex binding as reported previously [20].
Accepting the limitations of bioinformatics tools to predict pathogenicity, we did not find evidence of probable diseasecausing variants within an alternative thiazide-sensitive sodium bicarbonate exchanger, SLC4A8, hypothesized as an alternative genetic candidate for FHHt [14]. A third of our pedigrees with non-WNK FHHt therefore remain without a genetic diagnosis, which is somewhat greater than that reported in other pedigree collections [14,15]. This highlights the genetic heterogeneity of the FHHt phenotype and the likelihood that additional, as yet undiscovered, regulators of thiazide-sensitive pathways exist. It is also worth emphasizing that we set out to identify KLHL3 and CUL3 variants in subjects with a clinical diagnosis of FHHt on the basis of measurements routinely recorded in the clinic. Similar data are recorded for unaffected relatives, but because of the large disparity in ages it is often impossible to provide a comparison of age-related blood pressure between affected and non-affected individuals. Nevertheless, all non-affected individuals were normokalaemic with a plasma potassium <5 mmol/l, and we are confident that we have correctly assigned affected compared with non-affected status within our pedigrees.
Further detailed laboratory and clinical studies are required to establish whether the effects of the reported heterogeneity of variant KLHL3 on WNK1 immunoprecipitation and ubiquitination translate into differential effects on thiazide-sensitive distal nephron sodium trafficking and phenotype within FHHt [20]. For instance, do patients with KLHL3 A340V and A494T Gordon's syndrome have the same CUL3/KLHL3/WNK/SPAK/ NCC pathway abnormalities as those with KLHL3 L387P?
In conclusion we have identified disease-causing variants in CUL3 and KLHL3 in patients with FHHt screened previously and found to be negative for WNK1 and WNK4 mutations, but did not find evidence of such variants in the alternative candidate SLC4A8. Approximately one-third of our non-WNK patients with FHHt remain without a molecular diagnosis raising the possibility that there may be additional regulators of thiazide-sensitive distal nephron sodium trafficking which remain to be discovered.

CLINICAL PERSPECTIVES
r The present study was performed to acertain whether pedigress with FHHt, but without mutation in WNK1/WNK4, contained mutation in CUL3, KLHL3 or SLC4A8.
r The present study confirms recent findings of CUL3 and KLHL3 mutations in FHHt and identifies novel diseasecausing variants. This strengthens the argument that these gene products are physiologically important regulators of distal nephron NaCl reabsorption via thiazide-sensitive pathways, and hence are potentially interesting novel anti-hypertensive drug targets.
r As only 63 % of our non-WNK FHHt families were found to contain plausible CUL3 or KLHL3 variants, there are probably additional, as yet undiscovered, regulators of thiazide-sensitive pathways.    Figure S3 Pedigree structure of the ten kindreds with FHHt and mutations in CUL3 and KLHL3 detailed in Table 1 of the main text Affected, unaffected and phenotype-undetermined subjects are denoted by black, white and grey symbols respectively. Alongside each pedigree, Sanger sequence traces are shown showing the wild-type sequence and the corresponding CUL3 or KLHL3 variant segregating with the affected phenotype within each family. * The affected individual whose sequence is shown; c, the following co-ordinate is of coding DNA. Variants are described using standard Human Genome Variation Society nomenclature. Reference sequences for each gene are listed in Table S1. Pedigree 1: the affected individual carries a previously unreported variant disrupting a consensus splice site in CUL3. The lower chromatogram of CUL3 shows that in the affected individual the wild-type guanine at position c.1377 + 1 (the first base of intron 9 at the border with exon 9) is mutated to thymine in a heterozygous manner compared with the wild-type chromatogram shown above. The position of this variant is shown in more detail in Figure S5 and full genetic co-ordinates are detailed in Table 1 of the main text. Pedigree 2: the affected individual carries a previously reported variant co-locating with the variant observed in pedigree 1, which also disrupts the consensus splice site in CUL3. The lower chromatogram shows that in the affected individual the wild-type guanine at position c.1377 + 1 (the first base of intron 9 on the border with exon 9) is mutated to cytosine in a heterozygous manner compared with the wild-type chromatogram above. The position of this variant is shown in more detail in Figure S5 and full genetic co-ordinates are detailed in Table 1  The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/) which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.

Figure S3
Continued chromatogram above. The position of this variant is shown in more detail in Figure S5 and full genetic co-ordinates are detailed in Table 1 of the main text. Pedigree 4: the affected individual carries a previously reported variant disrupting a consensus splice site at the border of exon 9 in CUL3. The lower chromatogram of CUL3 shows that in the affected individual the wild-type guanine, one base into intron 8 from the exon 9 border position c.1207, is mutated to adenine in a heterozygous manner compared with the wild-type chromatogram above. The position of this variant is shown in more detail in Figure S5 and full genetic co-ordinates are detailed in Table 1 of the main text. Pedigree 5: the affected individual carries a previously reported FHHt-associated missense variant in KLHL3. The lower chromatogram of KLHL3 shows that in the affected individual the wild-type guanine is mutated to thymine in a homozygous manner compared with the wild-type chromatogram. This leads to exon 13 of KLHL3 encoding the amino acid mutation G500V. The position of this mutation is shown in more detail in Figure S4 and full genetic co-ordinates are detailed in Table 1 of the main text. Pedigree 6: the affected individual carries a previously reported FHHt-associated missense variant in KLHL3. The lower chromatogram of KLHL3 shows that in the affected individual the wild-type thymine is mutated to cytosine in a heterozygous manner compared with the wild-type chromatogram. This leads to exon 10 of KLHL3 encoding the amino acid change L387P . The position of this mutation is shown in more detail in Figure S4 and full genetic co-ordinates are detailed in Table 1 of the main text. Pedigree 7: the affected individual carries a previously reported FHHt-associated missense variant in KLHL3.
The lower chromatogram of KLHL3 shows that in the affected individual the wild-type thymine is mutated to cytosine in a heterozygous manner compared with the wild-type chromatogram. This leads to exon 10 of KLHL3 encoding the amino acid mutation L387P . The position of this mutation is shown in more detail in Figure S4 and full genetic co-ordinates are detailed in Table 1 of the main text. Pedigree 8: the affected individual carries a previously reported FHHt-associated missense variant in KLHL3. The lower chromatogram of KLHL3 shows that in the affected individual the wild-type thymine is mutated to cytosine in a heterozygous manner compared with the wild-type chromatogram. This leads to exon 10 of KLHL3 encoding the amino acid mutation L387P . The position of this mutation is shown in more detail in Figure S4 and full genetic co-ordinates are detailed in Table 1 of the main text. Pedigree 9: the affected individual carries a previously reported FHHt-associated missense variant in KLHL3. The lower chromatogram of KLHL3 shows that in the affected individual the wild-type cytosine is mutated to thymine in a heterozygous manner compared with the wild-type chromatogram. This leads to exon 9 of KLHL3 encoding the amino acid mutation A340V. The position of this mutation is shown in more detail in Figure S4 and full genetic co-ordinates are detailed in Table 1 of the main text. Pedigree 10: the affected individual carries a previously reported FHHt-associated missense variant in KLHL3. The lower chromatogram of KLHL3 shows that in the affected individual the wild-type cytosine is mutated to thymine in a heterozygous manner compared with the wild-type chromatogram. This leads to exon 13 of KLHL3 encoding the amino acid mutation A494T. The position of this mutation is shown in more detail in Figure S4 and full genetic co-ordinates are detailed in Table 1 of the main text. www.clinsci.org