Further evidence for functional recovery of AQP2 mutations associated with nephrogenic diabetes insipidus

Abstract

Aquaporin-2 (AQP2) is a homotetrameric water channel responsible for the final water reuptake in the kidney. Disease-causing AQP2 mutations induce nephrogenic diabetes insipidus (NDI), a condition that challenges the bodily water balance by producing large urinary volumes. In this study, we characterize three new AQP2 mutations identified in our lab from NDI patients (A120D, A130V, T179N) along the previously reported A47V variant. Using Xenopus oocytes, we compared the key functional and biochemical features of these mutations against classical recessive (R187C) and dominant (R254Q) forms, and once again found clear functional recovery features (increased protein stability and function) for all mutations under study. This behaviour, attributed to heteromerization to wt-AQP2, challenge the classical model to NDI which often depicts recessive mutations as ill-structured proteins unable to oligomerize. Consequently, we propose a revised model to the cell pathophysiology of AQP2-related NDI which accounts for the functional recovery of recessive AQP2 mutations.

Keywords: aquaporin-2; functional recovery; nephrogenic diabetes insipidus; recessive mutations.

FIGURE 1

Family pedigrees and location of four mutations in study. (a) Pedigree of three families with congenital nephrogenic diabetes insipidus probands showing homozygosity for A120D, compound heterozygosity for A130V/A47V, and heterozygosity for T179N. (b) Positioning of the four mutations in study within the reported structure of AQP2 (top and side views). Sections in red represent alpha helices segments participating in interunit association.

FIGURE 2

Expression of AQP2 variants in oocytes. Functionality determinations (a) along corresponding Western Blot (b) from a typical experiment (oocyte batch). Oocytes were injected with equivalent amounts of mRNAs for each AQP2 variants (0.5 ng) and incubated for 24 h prior to testing. (a) Specific P_f values are in % ± SD of wt‐AQP2 with n = 6 to 8 oocytes for each variant. Activity levels of all variants were similar to Ctrl (see statistics in supporting information). (b) Typical double bands (29 kDa (*) and 31 kDa (**) labeling for non‐glycosylated and high‐mannose forms, respectively, with 25 kDa MW marker indicated) are found for A47V, A130V and T179N. Note that although faint, the double bands are also found for A120D in overexposed blot.

FIGURE 3

Functional analysis of T179N‐AQP2 in wt/mutant conditions. Oocytes were injected with 0.5 ng mRNA coding for WT, T179N, R187C and R254Q either in absence (−) or presence (+) of same amount of wt‐AQP2 and incubated for 24 h prior to water permeability measurements. Activities are presented in % ± SD of wt‐AQP2, with n ≥ 6 for each condition. (a) Total activity in single (−) and dual expressions (+). (b) Specific activity for single (Ctrl value subtracted from all (−) expressions), and double (wt‐AQP2 value subtracted from all (+) expressions) conditions. (c) Specific activity of each mutant resulting from coexpression along wt‐AQP2, calculated by subtracting single (−) from dual (+) values determined in (b). As expected, coexpressing with wt‐AQP2 does not elicit any specific activity on R187C (p = 0.1176, see supporting information) in opposition to T179N which display a 78 ± 7% increase in activity, similar to WT (p = 0.0105). On the other hand, dominant R254Q induces a 69 ± 13% reduction in activity of wt‐AQP2.

FIGURE 4

Quantization of protein abundance for AQP2 variants in wt/mutant conditions. The conditions tested in Figure 3 for functionality were also used to evaluate the protein stability for each mutant variant, this time using equimolar amount of GFP‐wt‐AQP2 in (+) condition in order to isolate the untagged signal of mutant forms for densitometry analysis (GFP‐wt‐AQP2 in bracket). Western blots in panel (a) represent the specific labeling for each mutant either in absence (−) or presence (+) of GFP‐wt‐AQP2. Here again, both non‐glycosylated (*) and high‐mannose bands (**) are shown. (b) Densitometry analysis for each condition displayed in (a) combining both 29 kDa (*) and 31 kDa (**) bands and presented in % values against WT.

FIGURE 5

Functional recovery of four rec‐AQP2 mutations. Similarly to procedure in Figures 3 and 4, oocytes were injected with 0.5 ng mRNA coding for A47V, A120D, A130V and T179N either in absence (−) or presence (+) of wt‐AQP2 (equimolar GFP‐wt‐AQP2 in Western blots) and incubated for 24 h prior testing for activity (water permeability in a) and protein content (densitometry analysis of Western blot in b). (a) Recovery evaluation for each variant calculated by subtracting specific activities form single to dual (wt + mutant) expression conditions (similar to Figure 3c). Data represent % ± SD of wt‐AQP2 with n = 3–5 assays (see supporting information for p values). (b) Fold increase in protein contents (29 + 31 kDa bands) for each mutant (+ over − values, similar to Figure 4b) with n = 3–6 assays. (a) Plot correlating increased protein stability to functional recovery using data from panel (a) and (b), to which are added values for wt‐AQP2 (□) and R187C (∆).

FIGURE 6

Functional recovery distribution in NDI related AQP2 mutations. The functional recovery capacities for 19 AQP2 mutations identified from NDI case‐studies was performed using the same strategy as previously, i.e. subtracting mutant activities in single expression (− condition) to that found in coexpression (+ condition, see Figure 3). As shown, mutations located in the core structure of the channel (A19V to G211R) all display clear functional recovery properties, with notable exception of N68S and R187C which are part of, or directly adjacent to, the double NPA water selectivity filter (water pore). The two last mutations studied (R254L and R254Q), located in the C‐ter regulatory segment of the channel, display dominant negative effect typical of dom mutations through inhibition of the wt‐AQP2 subunit(s) activity. Values are mean ± SD in % of wt‐AQP2, n = 3–5 assays (seesupporting information for p values).

FIGURE 7

Molecular model describing AQP2‐depended NDI. In normal condition (wt/wt), AQP2 proteins are synthesized and assembled (tetramer) in ER before Golgi maturation and final storage in endosomes (Endo). AQP2 particles can therefore be recruited to the plasma membrane (PM) through vasopressin signaling (phosphorylation of C‐ter segment) and recycled on demand. Defective rec forms found within the core structure of the protein fail to produce functional tetramers in sufficient amount to support adequate level of activity, hence inducing NDI phenotype in homozygous rec/rec conditions (mean activity = 11% of wt/wt condition). Yet, aside a few exceptions (N68S, R187C), the same variants in heterozygous condition (wt/rec) can associate with wt counterparts to create functional wt/rec heteromers, increasing overall activity (87% of wt/wt condition) so to prevent NDI phenotype. Although dominant (dom) mutations also multimerize with wt, the wt/dom heteromers are mostly found to be sequestered (Golgi, endosomes), lowering global activity and causing NDI even in heterozygotes (25% of wt/wt condition).