The Drosophila Claudin Kune-kune is required for septate junction organization and tracheal tube size control

Abstract

The vertebrate tight junction is a critical claudin-based cell-cell junction that functions to prevent free paracellular diffusion between epithelial cells. In Drosophila, this barrier is provided by the septate junction, which, despite being ultrastructurally distinct from the vertebrate tight junction, also contains the claudin-family proteins Megatrachea and Sinuous. Here we identify a third Drosophila claudin, Kune-kune, that localizes to septate junctions and is required for junction organization and paracellular barrier function, but not for apical-basal polarity. In the tracheal system, septate junctions have a barrier-independent function that promotes lumenal secretion of Vermiform and Serpentine, extracellular matrix modifier proteins that are required to restrict tube length. As with Sinuous and Megatrachea, loss of Kune-kune prevents this secretion and results in overly elongated tubes. Embryos lacking all three characterized claudins have tracheal phenotypes similar to any single mutant, indicating that these claudins act in the same pathway controlling tracheal tube length. However, we find that there are distinct requirements for these claudins in epithelial septate junction formation. Megatrachea is predominantly required for correct localization of septate junction components, while Sinuous is predominantly required for maintaining normal levels of septate junction proteins. Kune-kune is required for both localization and levels. Double- and triple-mutant combinations of Sinuous and Megatrachea with Kune-kune resemble the Kune-kune single mutant, suggesting that Kune-kune has a more central role in septate junction formation than either Sinuous or Megatrachea.

Figure 1.—

kune-kune encodes a Drosophila claudin. (A) A schematic representation of the kune-kune (kune) locus. The kune^C309 allele contains a PiggyBac transposable element inserted within the 5′-UTR. (B) The Kune protein is predicted to be 264 amino acids with four transmembrane domains (red) and a C-terminal PDZ binding motif (blue). Residues conserved throughout claudin family members are shown in black. (C) The topology of Kune is similar to other claudins, with intracellular N and C termini and a large extracellular loop followed by two smaller loops. Kune is predicted to have an N terminus that is shorter than the Drosophila claudins Mega and Sinu and a C terminus that is longer. (D) Sequence comparisons show that Kune is more closely related to Sinu and Mega than to other predicted Drosophila claudins. The numbers on branch points in the figure indicate the frequency (%) with which claudins or groups of claudins clustered together in a “bootstrap” analysis where 1000 CLUSTLW alignments were performed on subsets of the dataset using the MacVector program (MacVector, Inc.). This approach more robustly determines relationships of poorly conserved proteins than does single-pass phylogenetic alignment approaches. Distance along the x-axis is arbitrary and does not reflect sequence similarity. Human claudin 3 (hClaudin 3) was used as an outgroup.

Figure 2.—

Kune localizes to septate junctions in primary epithelia. (A, B) Kune localizes to primary epithelia, including the trachea (TR), epidermis (EP), foregut (FG), salivary gland (SG), and hindgut (HG). A and B are different focal planes of the same embryo. (C–F″) Kune colocalizes with Cor at septate junctions (SJ) and does not colocalize with DE-cadherin at the adherens junction (F). Images of the trachea (C–F), hindgut (C′–F′), and salivary gland (C″–F″) are shown. (G–J″) Kune localization to the SJ is dependent on the SJ components, Mega, Sinu, Cor, and Atpα, since Kune is reduced and mislocalized to more basal positions in these mutants. Scale bars: 20 μm for A and B in B, 5 μm for C–J″ in J″.

Figure 3.—

Kune is required for septate junction organization and barrier function. (A, B, A′, B′) kune^C309 embryos show a complete absence of Kune by immunohistochemistry and result in reduced levels and mislocalization of Cor. (C, C′) Expression of kune-RNAi also causes Cor mislocalization. Arrow in C indicates the presence of some residual Kune, suggesting that RNAi knockdown is incomplete. (D, D′) Expression of the kune ORF using the da-Gal4 driver rescues Cor localization. (E–L′) The SJ proteins Dlg (E–F, E′–F′), Atpα (G–H, G′–H′), Mega (I–J, I'–J'), and Sinu (K–L, K′–L′) are reduced and/or mislocalized in kune epithelia. Dashed lines indicate the basal surface. (M–P) A fluorescent 10-kDa dye injected into the body cavity of WT embryos is excluded from the lumen of the trachea (M) and salivary gland (O), indicated by dashed lines. The dye readily leaks into the lumens of kune embryos (N, P), indicating a loss of the paracellular barrier. Scale bars: 5 μm for A–L′ in L′, 5 μm for M and N in P, 10 μm for O and P in P.

Figure 4.—

Kune is required for establishing the blood–brain barrier. (A–B) Kune is expressed in glial cells and is enriched in the glia at the midline (arrow in A). (C–D) A 10-kDa fluorescent dye is excluded from the ventral nerve cord of 20- to 21-hr-old (st 17) WT embryos (C), indicating a functional blood–brain barrier. In contrast, dye penetrates into the nerve cord of kune embryos (D). All images are ventral views. Scale bars: 10 μm for A and B in B and 10 μm for C and D in D.

Figure 5.—

Kune is required for tracheal tube size control. (A–B) The dorsal trunk of kune homozygous embryos (B) is overly elongated, resulting in tubes that follow a tortuous path. (C) This phenotype is identical in kune/Df(2R)BSC696 embryos. (D) Embryos expressing kune-RNAi using the da-Gal4 driver also display overly elongated trachea. (E–H) Lumenal Verm (E and F) and Serp (G and H) are absent in kune tracheal tubes. (I) Lumenal Verm is also absent in embryos expressing kune-RNAi using the da-Gal4 driver. (J) Expression of UAS-kune with da-Gal4 largely rescues Verm secretion of kune mutants. (K and L) WT trachea contain a lumenal, fibrilar chitin cable that is separated from the apical surface of tracheal cells, marked by Crb (arrow in K). In kune embryos, the chitin matrix is disorganized and the gap between the cable and the apical surface is missing (L). (M) Measurement of the posterior dorsal trunk (DT) between transverse connectives (TC) 5 and 10 indicate that the length of kune DTs is longer than WT. The posterior DT was measured because tracheal elongation of SJ mutants is most prominent at the posterior end. (**) P < 0.0005. Scale bar: 10 μm for A–L in L.

Figure 6.—

Kune-related claudins function in the same pathway of tracheal tube size control, but have unique roles in SJ organization. (A–C) Mutation to the Kune-related claudins, mega (A), sinu (B), and kune (C) all cause similar tracheal tube elongation defects. (D–F) Staining for Cor in the hindgut reveals that SJ organization is different in each of the three Kune-related claudins. mega embryos show complete mislocalization of Cor, but appear to have WT levels (D). Cor is reduced in sinu embryos, but retains an enrichment of Cor at the SJ (E, bracket). Cor is reduced and completely mislocalized in kune hindguts (F). (G–I) In contrast to mega (G) and sinu (H) salivary glands, kune salivary glands display a dramatic disruption in Cor levels and localization (I). (J–L) The trachea of mega; kune (J), kune; sinu (K), and mega; kune; sinu (L) embryos do not appear to be any worse than the single mutants (A–C). (M–R) Cor levels and localization in the double and triple mutants appear similar to kune single mutants (F and I), suggesting that the kune phenotype is the most severe. See Figure 2, D′ and D″, for WT comparison. Dashed lines indicate basal surface. Scale bar: 10 μm for A–C and J–L in R, 5 μm for D–I and M–R in R.

Figure 7.—

Kune is not required for epithelial apical-basal polarity. (A–D) Apical localization of Crb is normal in the ventral epidermis of WT (A), yrt(z) (B), kune (C), and Atpα (D) animals at stage 12. (E and F) Embryos containing mutations in kune and yrt show normal localization of Crb (E). In contrast, Crb is completely mislocalized in yrt, Atpα double mutants (F), indicating a loss of apical-basal polarity. Scale bar: 10 μm for A–F in F.