Analysis of uromodulin polymerization provides new insights into the mechanisms regulating ZP domain-mediated protein assembly

Abstract

Uromodulin is the most abundant protein secreted in urine, in which it is found as a high-molecular-weight polymer. Polymerization occurs via its zona pellucida (ZP) domain, a conserved module shared by many extracellular eukaryotic proteins that are able to assemble into matrices. In this work, we identified two motifs in uromodulin, mapping in the linker region of the ZP domain and in between protein cleavage and glycosylphosphatidylinositol (GPI)-anchoring sites, which regulate its polymerization. Indeed, mutations in either module led to premature intracellular polymerization of a soluble uromodulin isoform, demonstrating the inhibitory role of these motifs for ZP domain-mediated protein assembly. Proteolytic cleavage separating the external motif from the mature monomer is necessary to release the inhibitory function and allow protein polymerization. Moreover, we report absent or abnormal assembly into filaments of GPI-anchored uromodulin mutated in either the internal or the external motif. This effect is due to altered processing on the plasma membrane, demonstrating that the presence of the two modules has not only an inhibitory function but also can positively regulate protein polymerization. Our data expand previous knowledge on the control of ZP domain function and suggest a common mechanism regulating polymerization of ZP domain proteins.

Figure 1.

Identification of putative EHP and IHP motifs in uromodulin sequence. (A) A schematic representation of uromodulin is shown. The leader peptide is shown as a black box, EGF-like domains are displayed as numbered white boxes, and the ZP domain is shown in blue and divided in ZP-N and ZP-C subdomains (Boja et al., 2003). Urinary consensus cleavage site (CCS) is shown in pink and GPI-anchoring site in green. Glycosylation sites are represented as Y. The sequence of human uromodulin C terminus is shown below. Highlights show the C-terminal end of the ZP domain (blue), the consensus cleavage site (pink), the predicted EHP motif (red) and the GPI-anchoring site (green). The alignment (ClustalW) of C-terminal sequences of human (Hs, gi 56550049), murine (Mm, gi 22128623), rat (Rn, gi 8394509), canine (Cf, gi 50950249), and bovine (Bt, gi 27806359) uromodulin is shown. Each residue is highlighted in color, ranging from red for highly hydrophobic residues to blue for the most hydrophilic residues. The predicted secondary structure for human uromodulin obtained by using five different programs is shown. β-Strands are depicted as green arrows; α-helices as gray cylinders. (B) Uromodulin domain structure is represented as above. Alignment of sequences of the ZP domain linker region from human (Hs, gi 56550049), murine (Mm, gi 22128623), rat (Rn, gi 8394509), canine (Cf, gi 50950249), and bovine (Bt, gi 27806359) uromodulin is given. The two putative IHP motifs are boxed in yellow. Hydrophobicity of each residue is shown as described above. The prediction of secondary structure for human uromodulin obtained by using five different programs is shown: β-strands are represented by green arrows.

Figure 2.

Soluble uromodulin mutated in the EHP motif assembles into intracellular filaments. (A) Schematic representation of the C-terminus of truncated and soluble (indicated by _s) uromodulin EHP mutants. The putative EHP motif is highlighted in red. (B) Western blot detection of uromodulin in the medium (M) and in the soluble (S) and unsoluble (P) fraction of cell lysates from stably transfected MDCK cells. Uromodulin is found as a 100- and an 80-kDa isoform, corresponding to the mature and ER precursor form, respectively. Mutations in the EHP motif with the exception of the mutations affecting TRK residues (TRK/AAA_s and ΔTRK_s) dramatically impair protein secretion. Tubulin is shown as a loading control. (C) Immunofluorescence analysis of permeabilized MDCK cells stably expressing WT_s, m1-589 or TRK/AAA_s isoforms and stained for uromodulin and calnexin (ER marker). Mutations in the EHP sequence lead to the formation of intracellular filaments of mutant uromodulin (magnified in the insets). In the merged picture, the uromodulin signal is shown in red and the calnexin signal in green. Colocalization with calnexin shows the presence of mutant protein in the ER. However, uromodulin intracellular filaments are calnexin negative. Bar, 40 μm.

Figure 3.

Polymerization defect in GPI-anchored EHP mutants. (A) Domain structure of uromodulin protein. A schematic representation of the C terminus of all GPI-anchored EHP mutants is given. (B) Western blot detection of wild-type or mutant uromodulin in the medium (M) and in the soluble (S) and unsoluble (P) fractions of cell lysates from stably transfected MDCK cells. Uromodulin is found as a 100-kDa mature form and an 80-kDa precursor. Wild-type as well as mutant isoforms are fully glycosylated and secreted in the culturing medium, although to a different extent. Tubulin is shown as a loading control. (C) Immunofluorescence analysis showing wild-type and mutant (VLN/AAA, GP/AA, DQ/AA) uromodulin on the cell surface of stably transfected MDCK cells fixed with PFA. All isoforms are trafficked to the plasma membrane, but only wild-type uromodulin and the mutant not affecting the EHP motif (DQ/AA) can properly polymerize. On the contrary, mutations in the EHP motif impair protein polymerization. Bar, 40 μm.

Figure 4.

Mutations in the IHP lead to polymerization dysfunction. (A) Domain structure of uromodulin protein. Amino acidic sequence in the linker region 425-466 is shown. Putative IHP motifs 430DMKVSLK436 and 456FTVRMAL462 are highlighted in red and gray, respectively. Point mutations in both motifs are shown in red. (B) Immunofluorescence analysis of MDCK cells stably expressing the soluble isoforms of uromodulin IHP mutants (indicated by _s). Permeabilized cells were incubated with anti-uromodulin and anti-calnexin antibodies. In the merged picture, uromodulin and calnexin signals are shown in red and green, respectively. In cells expressing uromodulin that carries mutations in the first putative IHP (430DMKVSLK436), intracellular filaments of mutant protein similar to the ones observed for EHP mutants are visible (magnified in the insets). Intracellular assembly cannot be observed for mutants in the other putative IHP motif (456FTVRMAL462). As for EHP mutants, intracellular filaments formed by IHP mutants do not colocalize with calnexin. Bar, 40 μm. (C) Western-blot detection of soluble and GPI-anchored uromodulin mutants in the putative IHP motifs. Culturing medium (M), soluble (S), and unsoluble (P) fractions of cell lysates from stably transfected MDCK cells are loaded. All mutants are secreted in the culturing medium with the exception of isoforms carrying V458R mutation that is fully retained in the ER and shows only the precursor form of the protein. Tubulin is shown as a loading control. (D) Immunofluorescence analysis of unpermeabilized MDCK cells stably expressing wild-type or GPI-anchored IHP mutants. Wild-type uromodulin and L462S mutant assemble into polymers. Mutant V458R is barely trafficked to the membrane and does not show any polymerized protein. Mutants D430L and L435S form abnormally short and less organized polymers on the plasma membrane. Bar, 40 μm.

Figure 5.

The EHP motif is lost upon uromodulin assembly into filaments. Immunofluorescence analysis on MDCK cells stably expressing N- and C-terminally tagged uromodulin. A schematic representation of each tagged isoform is shown above the respective immunofluorescence panel. Uromodulin domain structure is depicted as before. EHP motif is indicated as a red box, the position of 6xHis and HA tags as green and red triangles, respectively. Immunofluorescence analysis was carried out on unpermeabilized PFA-fixed cells that were stained for the 6xHis tag and uromodulin. In the merged picture, uromodulin signal is in red whereas the 6xHis tag one is in green. Uromodulin polymers are positive for both antibodies when the 6xHis tag is present at the N terminus (N-His). However, they are negative for the anti-His antibody when the tag is localized in the C terminus, either before (C-His1) or after (C-His2) the EHP motif. These data suggest that the EHP sequence is lost in polymeric uromodulin. Bar, 10 μm.

Figure 6.

Release of the EHP motif in polymeric uromodulin. (A) Western blot detection of wild-type and RFRS/AAAA uromodulin isoforms that are secreted in the medium of stably transfected MDCK cells. Where indicated (+), samples were deglycosylated with PNGase F before loading. The short uromodulin band is lost in RFRS/AAAA mutant. (B) Immunofluorescence analysis of unpermeabilized MDCK cells stably expressing either wild-type uromodulin or uromodulin mutated in the CCS (RFRS/AAAA). Mutation of the CCS abolishes the formation of uromodulin polymers on the plasma membrane. Bar, 20 μm. (C) Western blot detection of N- and C-terminally tagged wild-type uromodulin isoforms that are immunoprecipitated from the medium of stably transfected MDCK cells. Samples were deglycosylated with PNGase F before loading. Although N-His uromodulin shows the presence of two bands with an anti-His and an anti-uromodulin antibody, for C-His1 and -2 only the upper band is positive for both antibodies. The lower band is negative for anti-His antibody, suggesting that in both cases the 6xHis tag is lost in the short uromodulin isoform that is incorporated into polymers.

Figure 7.

EHP/IHP mutations affect uromodulin cleavage. Western-blot detection of wild-type and EHP or IHP mutant uromodulin isoforms that are secreted in the medium of stably transfected MDCK cells. Samples were deglycosylated with PNGase F before loading. (A) EHP mutants that do not form polymers (*) (see Figure 3C) are abnormally processed as compared with wild-type and mutants that are able to polymerize. (B) IHP mutants (D430L and L435S) showing altered polymerization (*) (see Figure 4C) are abnormally processed.

Figure 8.

Conservation of IHP and EHP motifs in different ZP domain proteins. Schematic representation of the secondary structure prediction (obtained by PSI-Pred analysis) of the ZP domain and the C terminus of different human ZP domain proteins. Predicted β-strands and α-helices are indicated as dark and light gray arrows and striped cylinders, respectively. For each protein the position of known or putative cleavage site is indicated with a striped box. The position of predicted EHP and IHP motifs is marked by dashed lines and their sequence is shown below. Consensus sequences are also indicated. Uppercase letters within consensus sequence specify the corresponding amino acid in one-letter code. Lowercase letter indicate the following grouping sets: l, aliphatic (I, L, V); s, small (A, C, D, G, N, P, S, T, V); h, hydrophobic (F, H, I, L, M, V, W, Y); p, polar (C, D, E, H, K, N, Q, R, S, T); and t, turnlike (A, C, D, E, G, H, K, N, Q, R, S, T). Despite the apparent lack of significant sequence conservation among different EHP and IHP motifs, their secondary structure is remarkably conserved in ZP domain proteins.