Human Hypertension Caused by Mutations in WNK Kinases Frederick H. Wilson, Sandra Disse-Nicodème, Keith A. Choate, Kazuhiko Ishikawa, Carol Nelson-Williams, Isabelle Desitter, Murat Gunel, David V. Milford, Graham W. Lipkin, Jean-Michel Achard, Morgan P. Feely, Bertrand Dussol, Yvon Berland, Robert J. Unwin, Haim Mayan, David B. Simon, Zvi Farfel, Xavier Jeunemaitre, and Richard P. Lifton

Determination of WNK1 and WNK4 structures.

Publicly available genomic sequence, cDNA sequences and ESTs from human, mouse and rat were used to predict genomic and mRNA structures of hWNK1 and hWNK4. A comparison of the genomic sequence of the BAC clone RPCI11-388A16 with ESTs assembled from public databases reveals that exons 1-12 of hWNK1 are contained on the clone. The remaining 16 exons of hWNK1 are located on an overlapping BAC clone RP11-359B12 (GenBank # AC004803). EST database searches, GENSCAN exon predictions, and PCR amplification from kidney cDNA provide no evidence for transcripts or exons within the deleted interval. Exons 9, 11 and 12 are contained in some, but not all transcripts, indicating alternative splicing. A BLAST search of the assembled hWNK1 cDNA sequence identified a partially sequenced BAC clone, RP11-506G7 (GenBank # AC016889), containing significant sequence similarity to the kinase domain of hWNK1. A comparison of ESTs with the BAC genomic sequence identified exons later found to comprise the 3' end of the gene. Human ESTs, exon predictions by GENSCAN, and homology both with hWNK1 and with mouse genomic sequence containing the WNK4 ortholog (mouse BAC RP23-286N22; GenBank # AC025424) were used to design primers for PCR amplification from human kidney cDNA. Amplified products were sequenced to confirm the intron-exon structure of the gene.

Search for deletions in intron 1 of hWNK1.

Other deletions in the segment shared between the K22 and K4 deletions were sought in controls by quantitative PCR. Primers amplifying a product within this segment (nucleotides 62653-62856, GenBank # AC004765) and a product of similar size from an unlinked gene (KCC4, GenBank # AF105365) were used in the same reactions to direct PCR using DNA from individuals of K22, 40 unrelated unaffected subjects and affected members of other PHAII kindreds as template. The ratio of WNK1:control amplification was quantitated; the mean of the ratio for each subject was determined from at least 4 independent measurements. Individuals with heterozygous WNK1 deletions showed markedly lower ratios than their wild-type relatives, with no overlap between the two distributions. None of the 40 unrelated subjects or members of other PHAII kindreds studied showed ratios in the range of patients with deletions, indicating that deletions of this segment must be rare in the population. Southern blotting was performed by hybridizing 3 probes across this interval to genomic DNA of 20 control individuals digested with enzyme EcoRV. No fragments other than those predicted by the wild-type genomic sequence were detected. A smaller deletion lying in the proximal portion of intron 1 and which does not overlap with PHAII deletions was identified. This deletion removes 7983 nucleotides from intron 1 (nucleotides 27074-35056 in GenBank # AC004765). The deletion endpoints lie in a 5 bp segment of sequence identity. Its allele frequency was estimated at 10% by PCR genotyping of 70 unrelated unaffected subjects.

Preparation and characterization of anti-WNK1 and anti-WNK4 antibodies. The peptides SQPGGSLAQAPTTSSQQ (residues 1017-1033 of WNK1) and MGQMRRPPGRNLRR (residues 657-670 of WNK4) were coupled to keyhole limpet hemocyanin. Rabbits were immunized by successive injections, and serum was harvested. Antibodies were affinity purified using the immunizing peptide linked to an iodoacetyl cross-linked agarose column. Other antibodies used included Guinea pig polyclonal anti-mouse sodium-chloride cotransporter (NCCT, gift of David Ellison), rat monoclonal anti-ZO-1 (R-4036) (gift of James Anderson), mouse monoclonal anti-GST (B-14) (Santa Cruz Biotechnology), goat polyclonal anti-aquaporin-2 (C-17) (Santa Cruz Biotechnology), and goat polyclonal anti-vinculin (C-20) (Santa Cruz Biotechnology).

cDNAs encoding portions of WNK1 and WNK4 containing the immunizing peptides were separately cloned into pGEX4T-1 (Pharmacia) and transformed into E. coli to produce GST fusion proteins. Lysates expressing GST-WNK constructs were prepared by inducing log-phase cultures with 0.1 mM IPTG for 3 hours. Bacteria were pelleted by centrifugation and resuspended in 2 sample buffer and boiled. Samples were subjected to 12% PAGE, transferred to nitrocellulose membranes, and immunoblotting was performed with indicated sera. The proteins identified in Supplemental Fig. 2A and 2C with anti-WNK1 and anti-WNK4 were also detected by immunoblotting with anti-GST (data not shown). In addition, whole kidney lysates were prepared by Dounce homogenization from 8-week-old mice. Debris and nuclei were pelleted by centrifugation. The resulting samples were subjected to 5% PAGE, transferred to nitrocellulose membranes and immunoblotted with indicated sera, demonstrating staining of proteins of appropriate size.

Measurement of WNK1 transcript levels in leukocytes.

Levels of gene transcripts were evaluated by quantitative RT-PCR using RNA extracted from leukocytes of members of K4 and control subjects. RNA was purified by guanidinium thiocyanate-phenol-chloroform followed by treatment with RNAase-free DNAase. First strand cDNA was synthesized followed by amplification with specific primers from exons 7 and 8 of hWNK1, GAPDH, or the 18S ribosomal RNA locus. Products were quantitated using an ABI PRISM 7700 instrument. The threshold cycle number (C_T) at which the fluorescence reaches 10 the standard deviation of the baseline measured with no template was determined in triplicate. C_T is linearly related to the log of initial mRNA copy number. WNK1 level in each subject was evaluated as the WNK1:GAPDH ratio, 2^-CT, where C_T is defined as the difference between the mean C_T value from WNK1 and GAPDH. Similar estimates of WNK1 transcript levels were obtained by comparison to 18S ribosomal RNA.

Supplemental Figure 1. Comparison of hWNK1 and hWNK4. The amino acid sequence of hWNK4 is shown, with the corresponding sequence of hWNK1 shown below. Residues within conserved segments that are identical in WNK1 are indicated by dots. Amino acids that are missing are indicated by dashes. The positions of introns (In) in the genomic sequence of each gene are indicated. The kinase domain is highlighted in green (15), the putative coil domains in red, and the amino acids that are mutated in WNK4 in PHAII in blue.

Medium version | Full size version

Supplemental Figure 2. Characterization of anti-WNK1 and anti-WNK4 antibodies. (A) Western blot of bacteria expressing a 30 kD GST control protein (lane 1), a 39 kD GST-WNK1 fusion protein (lane 2), and a 41 kD GST-WNK4 fusion protein (lane 3) immunoblotted with anti-WNK1 antibody. The antibody specifically recognizes the GST-WNK1 fusion protein in lane 2. (B) Western blot of mouse kidney probed with anti-WNK1 antibody. The immune serum recognizes a protein of approximately 220 kD (lane 1) that is not seen with pre-immune serum (lane 2). (C) Western blot of bacteria expressing a GST control protein (lane 1), a GST-WNK1 fusion protein (lane 2), and a GST-WNK4 fusion protein (lane 3) immunoblotted with anti-WNK4 antibody. The antibody specifically recognizes the 41 kD GST-WNK4 fusion protein in lane 3. (D) Western blot of mouse kidney probed with anti-WNK4 antibody. The immune serum (lane 1) recognizes a predominant protein of approximately 135 kD and a second protein of approximately 155 kD; neither protein is labeled by pre-immune serum (lane 2).

Medium version | Full size version

Supplemental Figure 3. Localization of WNK1 in medullary collecting duct. Frozen mouse kidney sections were stained with antibodies and analyzed by fluorescence microscopy. (A) Transverse section of medullary collecting duct showing co-staining with anti-WNK1 and anti-AQP2. (B) Same view as (A) showing only anti-WNK1 channel, and demonstrating cytoplasmic distribution of the protein. White bar represents 10 m. We have not detected WNK1 or WNK4 in other nephron segments either on morphologic grounds (proximal tubule) or by costaining with markers of other nephron segments (thick ascending limb of Henle identified using anti-Tamm-Horsfall protein antibody) or by triple staining with antibodies to AQP2, NCCT and either WNK1 or WNK4.

Medium version | Full size version

Supplemental Figure 4. Normal salt, K⁺ and H⁺ handling in the distal nephron. A schematic diagram of cells of the distal convoluted tubule (DCT) and cortical collecting duct (CCD) is shown. In the DCT, salt is reabsorbed by the combined action of the electroneutral apical thiazide-sensitive Na-Cl cotransporter and the basolateral Na⁺/K⁺ ATPase and K-Cl cotransporter. In the CCD, apical Na⁺ entry in the principal cell is mediated by the electrogenic epithelial sodium channel which provides the electrical driving force for apical secretion of K⁺ from the principal cell, H⁺ from the -intercalated cell and the reabsorption of Cl^- via the paracellular pathway through the tight junction.

Medium version | Full size version

Supplemental Table 1. PCR primers for STS analysis.
STS Name	Location	Forward Primer	Location	Reverse Primer
STS42K	41762	GCATTTTCCCATTTAAGATCTTGAG	41982	CACCATACAGCAACATAACCGTG
STS45K	45765	TTTTCAGTTTTCAAGCTGCTGTTTG	46197	AACACCACGCAGAGCAAGAACAC
STS60K	60614	TAGAAATCTAGAAATGTCAGTAACAG	60782	ATGAAAATTTCCAGATAGGCCTCTC
All primers are indicated 5'-3'. Nucleotide position of 5' base with respect to GenBank # AC004765 is shown.

Supplemental Table 2. PCR primers for amplification across WNK1 deletions.
Kindred	Location	Forward Primer	Location	Reverse Primer
K22	35763	AATGAATAAGTTCAGGGAAGGTTGG	77605	TTTGAATGTGAAAGCAGCTAACAGG
K4	43794	CCCTCTTGAGTATAATCCAACCAG	68264	AGAAATACCACGAGTGTCAAGGAG
All primers are indicated 5'-3'. Nucleotide position of 5' base with respect to GenBank # AC004765 is shown.