Renal defects associated with improper polarization of the CRB and DLG polarity complexes in MALS-3 knockout mice

Abstract

Kidney development and physiology require polarization of epithelia that line renal tubules. Genetic studies show that polarization of invertebrate epithelia requires the crumbs, partition-defective-3, and discs large complexes. These evolutionarily conserved protein complexes occur in mammalian kidney; however, their role in renal development remains poorly defined. Here, we find that mice lacking the small PDZ protein mammalian LIN-7c (MALS-3) have hypomorphic, cystic, and fibrotic kidneys. Proteomic analysis defines MALS-3 as the only known core component of both the crumbs and discs large cell polarity complexes. MALS-3 mediates stable assembly of the crumbs tight junction complex and the discs large basolateral complex, and these complexes are disrupted in renal epithelia from MALS-3 knockout mice. Interestingly, MALS-3 controls apico-basal polarity preferentially in epithelia derived from metanephric mesenchyme, and defects in kidney architecture owe solely to MALS expression in these epithelia. These studies demonstrate that defects in epithelial cell polarization can cause cystic and fibrotic renal disease.

Figure 1.

MALS-3^−/− mice display renal defects. (A) Dissected urinary tracts from P0 pups show significantly smaller kidneys in MALS-3 knockout (−/−) mice as compared with heterozygote (+/−) littermates. A, adrenal gland; B, bladder; K, kidney; T, testis; U, ureter. (B) Kidneys from adult MALS-3^−/− mice are hypoplasic, dysplasic, and cystic; arrows mark cysts. (C) Quantification of body and organ weight of adult MALS-3^−/− mice and littermates. Kidneys from MALS-3^−/− are significantly smaller (57.9% ± 1.8; P < 0.01). (D) MALS-3^−/− mice backcrossed to 129/Sv or C57BL/6 for 8 generations show reduced kidney size (64.8% ± 3.2; P < 0.01 or 53.4% ± 8.5; P < 0.01, respectively). Mice lacking both MALS-1 and MALS-2 (MALS 1/2 KO) have normal kidneys. For C and D the number of animals is within the parentheses. (E) Hematoxylin and eosin stained kidney section from a 6-wk-old MALS-3^+/− mouse shows normal arrangement of densely packed tubules (left). Kidney from a MALS-3^−/− littermate shows marked tubulointerstitive changes, including dilatation of tubular lumina, tubular dedifferentiation, and fibrosis (right). Asterisks mark tubules with epithelia undergoing dedifferentiation, and arrows show dedifferentiated tubules. (F) Trichrome stained kidney section from a 6-wk-old MALS-3^+/− mouse displays little or no collagen deposition (left). In contrast, a kidney from a MALS-3^−/− littermate stains bright blue (arrows), revealing extensive fibrosis (right). (G) Immunolocalization of the Na⁺/K⁺ ATPase shows basolateral membrane localization in renal tubules of MALS-3^+/− mice (left). Polarized expression of the Na⁺/K⁺ ATPase is lost (arrows) in renal tubules undergoing simplification in MALS-3^−/− mice.

Figure 2.

MALS-3 localizes at the basolateral membrane and tight junction of renal epithelia. (A) Schematic depicting the arrangement of tubular segments in the kidney. ATL, ascending thin limb; CD, collecting duct; DCT, distal convoluted tubule; DTL, descending thin limb; G, glomerulus; LH, loop of Henle; PT, proximal tubule; TAL, thick ascending limb. (B) Low magnification fluorescent images from the cortex of kidney sections from a 6-wk-old mouse reveal that MALS-3 (red) primarily localizes to the basolateral membrane of tubular structures. MALS-3 staining is absent from MALS-3^−/−. AQP2 (green) specifically labels collecting ducts. (C) High magnification of proximal tubules shows a prominent brush border (phalloidin, green). DAPI (blue) stains nuclei. In proximal tubules, MALS-3 (red) is at both the basolateral membranes (arrowheads) and the tight junctions (arrows). (D) Cartoon summarizing MALS-3 localization in proximal tubules. Brush border and tight junction (TJ) are shown in green and yellow, respectively. (E and F) Low and high magnification images of the inner medulla show basolateral localization (arrows) of MALS-3 in collecting ducts (AQP2 positive tubules; green). (G) Cartoon depicting AQP2 (green) and MALS-3 (red) localizations in collecting ducts. Bars are 50 μm in B and E; 10 μm in C and F.

Figure 3.

MALS-3 interacts with both apical and basolateral polarity complexes in the kidney. (A) Silver staining of kidney extracts immunoprecipitated with anti-MALS-3 antibody shows a series of proteins from MALS-3^+/− homogenates that are absent from MALS-3^−/− precipitates. (B) Western blotting of extracts immunoprecipitated with anti-MALS-3 antibody confirms specific association of CASK, DLG, PALS, PATJ, and CRB-3 with MALS-3. (C) Kidney homogenates from MALS-3^+/− and MALS-3^−/− mice immunoblotted for numerous proteins associated with MALS-3 and with epithelial cell junctions. (D) Quantification of protein levels in C. The number of blots quantified for each protein is in parentheses. In A–D, proteins of the CRB, PAR-3, and DLG complexes are shown in red, blue, and green, respectively. Asterisks denote P values < 0.05.

Figure 4.

The L27 domain of MALS mediates cooperative formation of a stable MALS/CASK/DLG ternary complex. (A) Purification of a recombinant binary complex containing the single L27 domain of DLG and the tandem L27 domains of CASK shows degradation (asterisk, lane 1) and further degrades over 24 h at 4°C (asterisk, lane 2). Co-expression of MALS L27 domain stabilizes the MALS/CASK/DLG complex (lane 3), which does not degrade after 3 d at 35°C. (B) Comparison of the urea-induced denaturation profile of the L27_DLG/L27NL27C_CASK/L27_MALS complex with those of the (L27_MALS/L27C_CASK)₂ and (L27N_CASK/L27_DLG)₂ complexes. The ellipticities of each spectrum at 222 nm were used to construct the denaturation curves. The MALS/CASK/DLG ternary complex is more stable than the two binary L27 domain complexes. (C) The tandem L27 domains of CASK assemble MALS and DLG into a stable 1:1:1 MALS/CASK/DLG ternary complex with a molecular mass of ∼38 kD over a wide concentration range. Cartoon inset shows stoichiometry and molecular oligomerization of the MALS/CASK/DLG complex.

Figure 5.

Disruption of the PALS/PATJ/CRB-3 polarity complex in renal epithelia lacking MALS-3. In 6-wk-old MALS-3^+/− mice, PALS (A, top), PATJ (B, top), and CRB-3 (C, top) localize to the tight junction (arrows) below the apical brush border (phalloidin, green) of proximal tubule epithelia. In MALS-3^−/− mice, PALS (A, bottom) and PATJ (B, bottom) are no longer detected in proximal tubules. Most proximal tubules in MALS-3^−/− mice also lack CRB-3 (C, bottom), but some show a sub-apical punctate pattern, reminiscent of apical endosomes (arrows). (D) Cartoon summarizing the localization of the PALS/PATJ/CRB-3 complex (red) in wild-type proximal tubules (left) and the disruption of the complex in proximal tubule epithelia lacking MALS-3 (right). Bar = 10 μm.

Figure 6.

DLG mislocalizes to the tight junction in renal epithelia lacking MALS-3. (A, top) At the medullary boundary of a MALS^+/− kidney, DLG strongly labels the basolateral membranes of tubules (arrows) and faintly stains the tight junction of proximal tubule segments (arrowheads). Proximal tubule brush border is stained with phalloidin (green). (A, bottom) Basolateral localization of DLG is lost in renal epithelia lacking MALS-3 (arrows and arrowheads, as above) and replaced by staining at the tight junction. (B, top) Higher magnification of A shows basolateral localization of DLG (arrows). (B, bottom) Epithelia lacking MALS-3 show specific loss of DLG at the basolateral membrane and accumulation of DLG at the tight junction. (C, top) High magnification images of proximal tubules in the cortex show basolateral (arrowheads) and tight junction staining for DLG. (C, bottom) Proximal tubules of MALS-3^−/− mice have a discrete loss of DLG from the basolateral membrane (arrowheads). (D) Cartoon summarizing the localization of DLG (red) in control renal epithelia and epithelia lacking MALS-3. Bar = 50 μm in A and 10 μm in C.

Figure 7.

Cilia show no changes in structure or localization for CRB-3. (A) Cilia were visualized using anti-acetylated tubulin antibody. No obvious changes in cilia morphology were observed between control and MALS-3^−/− kidneys. (B) CRB-3 (red) localizes to cilia in both control and MALS-3^−/− kidneys. (C) Alignment of the intracellular portion of the translated protein sequences of canonical CRB-3 with three predicted protein sequences from the EST database. Underlined is the PDZ ligand in CRB-3 that is required for PALS/PATJ association; italicized are differences in the alternatively spliced C termini.

Figure 8.

Increased apoptosis in developing kidney of MALS-3^−/−. (A) Kidneys from E14.5 embryos were stained with anti-phospho-histone3 antibody (red) to visualize cells in mitosis and counterstained with cytokeratin (green) and DAPI (blue). No changes in cell proliferation were observed between MALS-3^+/− and MALS-3^−/− kidneys. (B) Quantification of the number of mitotic cells. (C) Lysotracker (red) was used to label apoptotic cells and counterstained with cytokeratin (green) and DAPI (blue) in embryonic kidneys (E14.5). MALS-3^−/− kidneys showed a dramatic increase in the number of apoptotic cells. (D) Quantification of Lysotracker fluorescence intensity from kidneys. Bar = 100 μm.

Figure 9.

Defective polarity cues in renal mesenchyme lacking MALS-3. (A) Cartoon shows nephron segments and their embryonic origins. Abbreviations are as in Fig. 2 A. (B) Cartoon shows the expression of cre recombinase (cre) driven by Pax3 promoter (Pax3-cre; blue) or HoxB7 promoter (HoxB7-cre; orange). MM, metanephric mesenchyme; UB, ureteric bud; W, Wolfian duct. (C) Western blotting of adult kidney homogenates from flox/flox mice shows differential loss of MALS-3 in mice expressing cre from the HoxB7 (Hox-cre) and Pax3 (Pax-cre) promoters. (D) Elimination of MALS-3 from the metanephric mesenchyme fully reproduces the MALS-3^−/− kidney deficiencies. MALS-3 flox/flox; Pax3-Pro-cre (MMKO) mice have kidneys that are reduced in magnitude to similar extent as MALS-3^−/− mice (54.7% ± 13.3; P < 0.01). In contrast, kidneys from MALS-3 flox/flox; HoxB7-Pro-cre (UBKO) mice have kidneys similar in size (99.7% ± 4.6) to littermate controls. (E) Trichome-stained kidney section from a 6-wk-old UBKO mouse shows normal renal anatomy (left), whereas kidney from a MMKO mouse (right) shares anatomical abnormalities with MALS-3^−/− mice, including tubular dilatation and fibrosis (arrow). (F) Immunohistochemistry from the cortex of kidneys from 6-wk-old UBKO and MMKO mice stained with phalloidin (green) to identify proximal tubule brush borders and DAPI (blue) for nuclei. PALS, PATJ, and CRB-3 all properly localize to the tight junction (arrows) of epithelia in proximal tubules of a kidney from UBKO mouse (F, top). In contrast, PALS, PATJ, and CRB-3 are lost from tight junction in proximal tubules from a MMKO mouse (F, bottom). Asterisk marks apical localization of CRB-3 in collecting duct.

Figure 10.

Increased apoptosis associated with disrupted polarity complexes in developing kidney of MMKO. (A) Apoptotic cells were visualized in embryonic kidneys (E14.5) with Lysotracker (red). Cytokeratin (green) and DAPI (blue) label epithelia and nuclei, respectively. As compared with control littermates, MMKO kidneys showed a dramatic increase in the number of apoptotic cells. Bar = 100 μm. (B) Kidney homogenates from flox/flox and MMKO mice immunoblotted for proteins of the CRB-3 and DLG polarity complexes show reduced expression of both complexes in MMKO embryonic kidneys.