J Am Soc Nephrol 11: S129–S134, 2000 Scnn1 Sodium Channel Gene Family in Genetically Engineered Mice EDITH HUMMLER* and FRIEDRICH BEERMANN† *Institut de Pharmacologie et de Toxicologie, Universite´de Lausanne, Lausanne, and †Swiss Institute for Experimental Cancer Research, Epalinges, Switzerland. Abstract. The amiloride-sensitive epithelial sodium channel is have been obtained by introducing specific mutations or using the limiting step in salt absorption. In mice, this channel is transgenic rescue experiments. In this report, these approaches composed of three subunits (␣, ␤, and ␥), which are encoded are summarized and a current gene-targeting strategy to obtain by different genes (Scnn1a, Scnn1b, and Scnn1c, respectively). conditional inactivation of the channel is illustrated. This latter The functions of these genes were recently investigated in approach will be indispensable for the investigation of channel transgenic (knockout) experiments, and the absence of any function in a wide variety of organ systems. subunit led to perinatal lethality. More defined phenotypes The highly amiloride-sensitive epithelial sodium channel epithelial potential differences in ␣ENaCϪ/Ϫ (Scnn1atm1/ (ENaC) is a membrane constituent of many salt-reabsorbing Scnn1atm1) mice revealed that ENaC activity was completely epithelia that facilitates Naϩ movement across the tight epi- abolished (Table 1). This suggests that channels composed of ␤␥ thelia that line the distal renal tubule, the distal colon, the ducts subunits alone do not confer sufficient activity to compensate for of salivary and sweat glands, and the lung (1). ENaC is a loss of the ␣ subunit in the lung. In human subjects, mutations in heterotetramer composed of three homologous subunits, i.e., ␣, all three Scnn1 genes do result in ENaC hypofunction and are ␤, and ␥; each subunit contains intracellular amino and car- expected to induce a salt-wasting syndrome similar to pseudohy- boxyl termini, two membrane-spanning domains, and a large poaldosteronism type 1 (PHA-1) (9). In contrast to the ␣ENaC extracellular loop (2–4). cDNA encoding the three homolo- deficiency in mice, PHA-1 variants in human subjects might gous subunits of this channel have been characterized in sev- retain sufficient residual ENaC activity to rescue or attenuate the eral species, including humans and mice (5–7). Here we sum- lung phenotype (10,11). In an in vitro expression system, none of marize transgenic experiments addressing channel function in the mutations causing PHA-1 has been shown to be a null muta- mice and present the three different approaches that we have tion and those tested still confer ENaC-mediated sodium transport used, i.e., classic knockout experiments, transgenic rescue ex- activity (albeit diminished) (11). However, pulmonary ENaC dys- periments, and the introduction of a specific mutation in the function and excess airway liquid have been demonstrated in ␤-subunit. Finally, we discuss an additional tool for analyzing patients with PHA-1 (12). Species-specific differences, such as the channel in vivo, namely conditional gene inactivation. anatomic immaturity of newborn mouse lungs (13) or the pres- ence of additional ENaC subunits in human subjects (e.g., Role of the Channel Subunit Genes ␦ENaC) (14), might also explain the phenotypic differences. Gene targeting is defined as the introduction of site-specific Scnn1b-deficient (15) and Scnn1c-deficient (16) mice lack a modifications into the genome. It allows the in vivo analysis of pulmonary phenotype but exhibit defects in renal function that diverse aspects of gene function. The classic gene-targeting cause perinatal death. In both knockout types, hyponatremia and (knockout) approach leads to inactivation of a gene in all tissues hyperkalemia indicated a defect in renal collecting duct Naϩ of the body, from the onset of development through the entire channel activity and Kϩ secretion. In addition, Scnn1b- and lifespan. Phenotypes of targeted mouse mutants cannot always be Scnn1c-deficient mice exhibited small but significant increases in predicted from the presumed function of the given gene product wet/dry lung weight ratios, although this was not the primary and/or the pattern of expression of the gene. For example, con- cause of their death. We suggest that, in contrast to the ␤␥ channel stitutive inactivation of the ␣ subunit of ENaC revealed an im- subunits in Scnn1a-deficient mice, ␣␤ channels (Scnn1c-defi- portant role for the channel in lung liquid clearance after birth (8). cient) or ␣␥ channels (Scnn1b-deficient) provide low but suffi- ␣ENaC knockout mice died within 40 h after birth, as a result of cient levels of Naϩ absorptive capacity in lung to account for the respiratory failure. Measurements of the amiloride-sensitive trans- transient slower liquid clearance a few hours after birth. We suggest that, for neonatal survival, low levels of ENaC activity are sufficient in lung but not in kidney. Correspondence to Dr. Edith Hummler, Institut de Pharmacologie et de Toxi- cologie, Rue du Bugnon 27, CH-1005 Lausanne, Switzerland. Phone: 41-21- 692-5357; Fax: 41-21-692-5355; E-mail: [email protected] Introduction of Specific Mutations into Scnn1 1046-6673/1111-0129 Genes Journal of the American Society of Nephrology Direct evidence that ENaC dysfunction is involved in patho- Copyright © 2000 by the American Society of Nephrology logic processes has come from the molecular analysis of two S130 Journal of the American Society of Nephrology J Am Soc Nephrol 11: S129–S134, 2000 Table 1. Transgenic models targeting ENaC functiona ENaC Activity Allele and Genotype (%)b Organ Phenotype Reference Null alleles Scnn1atm1/Scnn1atm1 0 Lung (kidney) Perinatal lethal 8 ␤ENaCϪ/Ϫ ND Kidney Perinatal lethal 15 Scnn1ctm1/Scnn1ctm1 Ͻ15 Kidney Perinatal lethal 16 Mutant alleles Scnn1bneo/Scnn1bneo ϳ20 Kidney PHA-1 22 Scnn1bLid/Scnn1bLid Ͼ100 Kidney Liddle’s syndrome 23 Transgenic rescue Scnn1atm1/Scnn1atm1 Tgr␣ENaC ϳ15 Kidney PHA-1 29 a According to the International Committee on Standardized Genetic Nomenclature, the actual gene symbol for the epithelial sodium channel (ENaC) is Scnn1 (sodium channel, non-voltage-gated 1). Genes encoding the three subunits are Scnn1a,-b, and -c (41). The ␤ENaC null allele was not indicated in the naming of the different alleles, because it has been produced elsewhere. PHA-1, pseudohypoaldosteronism type 1; ND, not determined. b The values for resting ENaC activity were estimated from in vitro experiments. human genetic diseases, i.e., Liddle’s syndrome (aldosteron- ism) and PHA-1 (9,17). In patients with Liddle’s syndrome, mutations in the Scnn1b and Scnn1c genes are responsible for a monogenic form of salt-sensitive hypertension (for review, see reference (10). These mutations are clustered in the last exons of the Scnn1b and Scnn1c genes, and, in the resulting protein, PPxY motifs are affected or deleted. It has been shown that these PPxY motifs act as binding sites for a protein called Nedd4 (18). Normally, ENaC is ubiquitinated and degraded, and the number of channels at the cell surface is tightly controlled. In patients with Liddle’s syndrome, lack of this protein-protein interaction has been proposed as a cause for the regulatory defect. Interference of the mutant ENaC proteins Figure 1. Strategy for generating two new alleles at the Scnn1b with the clathrin-mediated endocytosis pathway has also been epithelial sodium channel (␤ENaC) gene locus. A schematic repre- discussed (19). As a consequence of the mutations, increased sentation of the Scnn1b targeting vector, the 3Ј portion of the wild- sodium channel activity is obtained through increased channel type allele (including exons 12 and 13), and the two newly generated numbers and greater open probability. This renders the channel alleles is presented. Arrow, premature termination codon introduced constitutively active (20). In patients with PHA-1 (a salt- by the targeting vector. Black triangles, loxP sites. Black lines, ho- wasting syndrome), mutations have been identified in all three mologous sequences included in the targeting vector. Gray boxes, exons. HSV-Tk, herpes simplex virus thymidine kinase; neo, neomy- Scnn1 genes. In the variant proteins, mutations are well spread cin selection marker. throughout the functional domains, including the gating do- main, which renders the channel hypoactive (for review, see references (10 and 21). The introduction of a Liddle mutation into mice should in appearance, growth rate, and fertility (22). However, the permit elucidation of the causal relationship between dysfunc- introduction of this mutation resulted in large decreases in tion of the Scnn1 genes, dietary salt intake, and hypertension. Scnn1b mRNA and protein levels in all organs tested and We recently introduced the Liddle mutation R566STOP into reduced in vivo ENaC activity in the colon. This was unex- the mouse Scnn1b gene locus and thus established two mouse pected, because this mutation should not affect RNA levels and models, i.e., a model for salt-dependent PHA and a model for instead should lead to increased ENaC density and activity salt-induced Liddle’s syndrome (22,23). First, to reproduce (24). The only difference between our mutated mouse Scnn1b Liddle’s syndrome in mice, we introduced the stop codon in locus and that in human subjects is the presence of the selection the last exon of the mouse Scnn1b gene. This mutation corre- marker neo controlled by the phosphoglycerate kinase (PGK) sponds to the R566STOP mutation in human subjects (17) promoter, which are needed to select for stable transfection in (Figure 1). In the targeting vector, the neomycin selection embryonic stem cell culture (Figure 1). It is possible that the marker (neo), flanked by two loxP sites, followed this muta- insertion of an additional selection marker (neo) into 3Ј un- tion. Phenotypically, mice homozygous mutant for this new translated sequences resulted in destabilization and/or degra- allele (Scnn1bneo) were indistinguishable from wild-type mice dation of the Scnn1b mRNA transcripts, thus interfering with J Am Soc Nephrol 11: S129–S134, 2000 ENaC Genes S131 the expression of the corresponding gene locus (25,26).