Supplementary Figure 5: rLCx overexpression constructs for subcellular localization.

(a) Schematic representation of all UAS constructs generated in this study (left panel). Quantification of Dpt mRNA (at 8h after infection, left) or rLCx mRNA (fold induction, right) in flies overexpressing rLCx constructs (driver: c564-Gal4) and challenged with heat-killed Ecc15 (WT control: c564-Gal4 crossed to w¹¹¹⁸). Unfortunately, any large truncation of the rLC cytosolic tail will leave a short cytosolic stump coupled to a peptidoglycan-binding PGRP ectodomain, strongly resembling PGRP-LF. It is likely that overexpression of such constructs will still regulate IMD activation, albeit by receptor competition mechanistically resembling regulation through PGRP-LF rather than through rLC-specific (or PHD-domain specific) mechanisms. An adequate but time-intensive approach would be to replace wild-type rLC with deletion mutants at its genomic locus.

(b) Confocal imaging of overexpressed (driver: c564-Gal4) FYVE-GFP (left, single z slice through cell body) and PLCδ-PH-GFP (right, single z slices close to cell surface and through cell body) shows localization of FYVE marker to intracellular vesicles and of PLCδ-PH marker to plasma membranes. Scale bars, 20 μm.

(c) Confocal imaging of overexpressed GFP-rLCx in fixed and DAPI-stained hemocytes from hml-Gal4>UAS-GFP-rLCx larvae shows GFP localizes to plasma membrane (solid arrowheads), to cell-cell contact points (open arrowheads), and to filopodia (arrow). Image is an average intensity z-projection of 10 consecutive z-slices. Scale bar 10 μm.

(d) Quantification of immunoblot in Fig. 5f shows that GFP-rLCx accumulates when the endocytic machinery is impaired.

(e) and (f) Localization of overexpressed GFP-rLCx (driver: c564-Gal4) in tsg101 knock-down (e) and Rab5 knock-down (f) fat bodies. Sum projection of 3 consecutive apical slices; scale bar 10 μm (compare to Fig. 5g).

(g) Quantification of fat body cell size for indicated genotypes shows that only Rab5 knock-down significantly affects cell size. To correct for this, vesicle numbers were normalized to respective cell area in Fig. 5j.

(h) Quantification of Dpt mRNA expression 8h after infection with heat-killed Ecc15 in controls (c564-Gal4) or in flies overexpressing rLCx (c564-Gal4 > UAS-rLCx) in the presence or absence of RNAi against indicated endocytic components. WT control, w;c564-Gal4 x w¹¹¹⁸. Fab1 is a PI3P-5 kinase, Vps28 and Tsg101 are components of the ESCRT machinery, UbcD10 is the unique Drosophila E1 ligase involved in protein ubiquitination. * P < 0.05, **P < 0.01, *** P < 0.001 (one-way ANOVA, Dunnett’s post hoc test (g), two-way ANOVA, Bonferroni post hoc test (h). Data are pooled from three (a, h, except for Rab5 RNAi which caused significant developmental lethality) or n (indicated above graph (g)) independent experiments and represent mean + s.e.m. (a, g, h).