Figure 2: Swelling in P deaths results from E. coli infection.

(a) Comparison of P and p corpses of worms fed with E. coli expressing RFP. Scale bar, 40 μm. (b) Proportion of P (n=70) and p (n=75) corpses with widespread RFP, small intra-pharyngeal RFP inclusions or no detectable RFP foci. (Trials: 2). (c) Representative TEM of a swollen pharynx from a live, 8-day-old adult. L, lumen; Ph, pharynx; B, bacteria. The elongated appearance of some bacilli suggests that proliferation occurs within pharyngeal tissue. For TEMs of pharynxes at different stages of infection, see Supplementary Fig. 5. Scale bar, 5 μm. (d) Early stages of E. coli-RFP invasion of pharynx in transgenic C. elegans expressing GFP in selected cell types within the pharynx. Representative images of the muscle (left) and marginal cell marker (right) are shown with nascent intra-pharyngeal infection near the grinder (arrowheads). E. coli were noted previously in senescent marginal cells³¹. For gland cell marker image, see Supplementary Fig. 4c. Scale bar, 10 μm. (e) Blocking E. coli proliferation prevents P death. Proportion of P corpses when worms are grown with or without proliferating E. coli. Data are mean±s.e.m. (Carbenicillin trials: 4; ultraviolet trials: 2, for sample size, see Supplementary Table 3). (f) Carbenicillin-treated bacteria remove mortality rate deceleration in mid life. Control wild-type mortality rate (black) as in Fig. 1g. For carbenicillin-treated log mortality rate (green), the slopes of regression are not significantly different when segmented at any point.