Pasiflora proteins are novel core components of the septate junction

Epithelial sheets play essential roles as selective barriers insulating the body from the environment and establishing distinct chemical compartments within it. In invertebrate epithelia, septate junctions (SJs) consist of large multi-protein complexes that localize at the apicolateral membrane and mediate barrier function. Here, we report the identification of two novel SJ components, Pasiflora1 and Pasiflora2, through a genome-wide glial RNAi screen in Drosophila. Pasiflora mutants show permeable blood-brain and tracheal barriers, overelongated tracheal tubes and mislocalization of SJ proteins. Consistent with the observed phenotypes, the genes are co-expressed in embryonic epithelia and glia and are required cell-autonomously to exert their function. Pasiflora1 and Pasiflora2 belong to a previously uncharacterized family of tetraspan membrane proteins conserved across the protostome-deuterostome divide. Both proteins localize at SJs and their apicolateral membrane accumulation depends on other complex components. In fluorescence recovery after photobleaching experiments we demonstrate that pasiflora proteins are core SJ components as they are required for complex formation and exhibit restricted mobility within the membrane of wild-type epithelial cells, but rapid diffusion in cells with disrupted SJs. Taken together, our results show that Pasiflora1 and Pasiflora2 are novel integral components of the SJ and implicate a new family of tetraspan proteins in the function of these ancient and crucial cell junctions.


Production of antibodies
For each protein two 15-16 amino acids-long peptides were synthetically generated and their mixture was injected in rabbit and guinea pig (for Pasiflora1) and hen (for Pasiflora2) (Eurogentec, Seraing, Belgium). The epitopes were: for Pasiflora1: SPLFETDIRSSMPVA, IIWSDNVRTGSYAVA, and for Pasiflora2: NLHSKMSRSTRSVRI, STANSLAGSRPTTPHS. The sera, as well as affinity-purified antibodies were tested by immunostainings in wt embryos in various concentrations (including 1:2).

Analysis of FRAP data
Image Registration and analysis. Embryo movements are unavoidable and pose severe challenges for the analysis of time-lapse recordings. We used a home-written Definiens script to correct for lateral drift and non-linear distortions of the raw confocal images due to changes of cellular shape. In brief, for a confocal stack of n images with index 1..n, a built-in image registration algorithm was first applied to three reference images with rounded indexes n/6, n/2 and 5n/6, respectively (the middle image was used as reference image for registration). The remaining images were then registered with respect to the reference image closest in index number. Given the strong embryo movements and drift that we observed, this strategy ensured a more robust alignment compared to a registration procedure based on only one reference image for the whole stack.
A second Definiens script was then used to automatically extract the fluorescence intensity trajectories of the photobleached membrane regions. To detect the photobleached region we applied to registered images a 2D-Gaussian filter with a kernel size of 5x5x3 pixels, followed by an edge 3D filter. This filter is sensitive to signal variations between successive time-lapse images, and is thus ideal to detect the photobleached region that exhibits a strong decrease in fluorescence intensity just after the photobleaching step. The average fluorescence intensity in the identified region can then be extracted for each time point, and normalized with Development | Supplementary Material respect to its maximal and minimal values at the time points before and immediately after the photobleaching step, respectively. FRAP data analysis. In a first approximation, the diffusion in the thin photobleached membrane can be modelled by one-dimensional free diffusion. The experimental data were fitted to the empirical formula given in equation (1), which agrees within 5% with the solution of the diffusion equation in one dimension for recovery into an interval of zero intensity (Ellenberg et al., 1997;Ellenberg and Lippincott-Schwartz, 1999) with I(t) = intensity as a function of time; t0= time right after photobleaching; I(final) = final intensity reached after complete recovery; τD = characteristic time of diffusion.
The fitting procedure was performed using Origin 8.5. We kept t0 constant, and extracted I(final) and τD from the fitted curves. Mobile fractions were calculated as ratios of fluorescence intensity in the bleached area after recovery of the signal to fluorescence intensity before photobleaching.
Another common approach used to analyze FRAP recovery curves is the calculation of half-time (t1/2) as the time required for the bleached fluorescence to recover to half of its maximum recovery value (Yguerabide et al., 1982;Oshima and Fehon, 2011). We extracted t1/2 from exponential fits of the recovery curves and found for Nrg-GFP t1/2= 0.