Phagocytic Receptor CED-1 Initiates a Signaling Pathway for Degrading Engulfed Apoptotic Cells

Apoptotic cells in animals are engulfed by phagocytic cells and subsequently degraded inside phagosomes. To study the mechanisms controlling the degradation of apoptotic cells, we developed time-lapse imaging protocols in developing Caenorhabditis elegans embryos and established the temporal order of multiple events during engulfment and phagosome maturation. These include sequential enrichment on phagocytic membranes of phagocytic receptor cell death abnormal 1 (CED-1), large GTPase dynamin (DYN-1), phosphatidylinositol 3-phosphate (PI(3)P), and the small GTPase RAB-7, as well as the incorporation of endosomes and lysosomes to phagosomes. Two parallel genetic pathways are known to control the engulfment of apoptotic cells in C. elegans. We found that null mutations in each pathway not only delay or block engulfment, but also delay the degradation of engulfed apoptotic cells. One of the pathways, composed of CED-1, the adaptor protein CED-6, and DYN-1, controls the rate of enrichment of PI(3)P and RAB-7 on phagosomal surfaces and the formation of phagolysosomes. We further identified an essential role of RAB-7 in promoting the recruitment and fusion of lysosomes to phagosomes. We propose that RAB-7 functions as a downstream effector of the CED-1 pathway to mediate phagolysosome formation. Our work suggests that phagocytic receptors, which were thought to act specifically in initiating engulfment, also control phagosome maturation through the sequential activation of multiple effectors such as dynamin, PI(3)P, and Rab GTPases.


Introduction
An engulfing cell recognizes an apoptotic cell through phagocytic receptor(s) and extends thin pseudopods around it to generate a phagocytic cup. The fusion of growing pseudopods and scission of a vacuole containing the apoptotic cell from the plasmalemma generate a phagosome inside the host cell ( Figure 1A) [1]. The swift engulfment (phagocytosis) and subsequent degradation of apoptotic cells inside phagosomes eliminate dying cells before they release any potentially harmful contents, and actively prevent tissue injury, inflammatory responses, and autoimmune diseases [2]. Inefficient degradation of the components of engulfed apoptotic cells, such as nuclear DNA, similar to inefficient engulfment, results in severe inflammatory and autoimmune responses [3,4].
Despite its importance, the molecular mechanisms driving the degradation of apoptotic cells remain largely unknown. Our current knowledge of phagosome maturation, the process that results in the complete degradation of phagosomal contents, is primarily obtained through the studies of phagosomes containing opsonized foreign pathogens or undigestible latex beads. Early during phagosomal maturation, signaling molecules such as Rab5 and PI(3)P are observed to be enriched on nascent phagosomes and act to recruit multiple downstream factors [5]. Early and late endosomes and lysosomes continue to fuse with phagosomes, and provide necessary lipids and proteins for phagosome maturation. Phagosome lumen is gradually acidified. The final state, known as a phagolysosome, is characterized by its abundance of acid hydrolases and an acidic pH environment [5,6]. The maturation of phagosomes containing apoptotic cells, by contrast, is less characterized, and may display distinct mechanistic features. Macrophages that internalize foreign pathogens secrete proinflammatory cytokines that induce inflammatory responses, whereas those engulfing apoptotic cells secrete anti-inflammatory cytokines, indicating that the same type of phagocytes elicit different responses to different phagocytic targets [2,7].
Rab GTPases are important regulators of many vesicle trafficking events. Rab7 has been implicated in the endocytic pathway, lysosomes biogenesis, and phagosomal maturation [5]. In both mammalian and Dictyostelium cells that ingest latex beads, the overexpression of a dominant-negative form of Rab7 (Rab7(DN)) blocks the phagosomal acquisition of lysosomal contents [8,9]. It remains to be elucidated, however, whether the defect in phagolysosome formation is a secondary consequence of the defects in lysosomal biogenesis and/or functions. In addition, although the enrichment of Rab7 was observed on phagosomes containing apoptotic cells [10], whether Rab7 is required for the processing of engulfed apoptotic cells and how Rab7 is activated by upstream signals remain unknown.
During the development of Caenorhabditis elegans hermaphrodites, 131 somatic cells and 300-500 germ cells undergo programmed cell death [11]. These apoptotic cells (referred to as cell corpses) are distinguishable from living cells as highly refractile discs under a Nomarski differential interference contrast (DIC) microscope [12]. In C. elegans, cell corpses are removed by their neighboring cells [11]. Genetic screens have identified two parallel and partially redundant pathways that regulate the engulfment of cell corpses ( Figure 1B) [1,11,13,14]. In one pathway, the phagocytic receptor cell death abnormal 1 (CED-1) is an engulfing cell-specific, single-pass transmembrane protein that recognizes neighboring cell corpses, clusters to the growing phagocytic cups, and initiates pseudopod extension [15]. The large GTPase DYN-1 (dynamin) acts downstream of CED-1 and CED-6, a candidate adaptor for CED-1, to control the delivery of intracellular vesicles to both phagocytic cups and phagosomes [1]. In the other pathway, CED-5/Dock180, CED-2/CrkII, and CED-12/ELMO1 act together to regulate CED-10/Rac1 GTPase and promote cell surface extension during engulfment [11]. CED-10 may also mediate certain activities for the CED-1 pathway [16].
Previously, transmission electron microscopy (TEM) analyses indicated that those cell corpses that persisted for hours in ced-1, -6, -7, -2, -5, -10, and -12 mutants were mostly unengulfed. It was thus proposed that these ced genes specifically control the engulfment of cell corpses [15,17,18]. However, none of the corresponding ced mutants completely block engulfment, since even in the strongest single engulfment mutants, such as ced-1 or ced-5 mutants, only approximately 25% of embryonic apoptotic cells persist for longer than 5 h [1]. Previous TEM analyses, which only characterized the terminal mutant phenotypes hours after the apoptosis events, were unable to determine whether the rest of cell corpses are engulfed, or to detect defects in the kinetics of engulfment and/or phagosome maturation.
Recently, we developed a time-lapse imaging protocol and used it to characterize the function of DYN-1 in developing embryos [1]. We found that in engulfing cells, DYN-1 was recruited to the surface of phagocytic cups and phagosomes in a CED-1-, CED-6-, and CED-7-dependent manner. DYN-1 promoted the recruitment and fusion of endosomes to phagocytic membranes, events that provide lipid and protein material to support pseudopod extension and phagosome maturation [1]. These results suggest that the engulfment and degradation of apoptotic cells may share common regulatory mechanisms. Like dyn-1 mutants, ced-1 null mutant embryos display defects in the incorporation of early endosomes to phagocytic cups and phagosomes [1], suggesting that CED-1, and perhaps other pathway components, may also control phagosome maturation in addition to engulfment. To test the above hypothesis and to further identify the mechanism controlling the degradation of cell corpses, we developed multiple markers and further expanded the scope of our time-lapse imaging assays to monitor the processes of engulfment and phagosome maturation as well as the cellular events occurring during these processes in living embryos. We identified novel functions of both engulfment pathways in regulating the degradation of engulfed cell corpses. We further elucidate that the signaling pathway led by CED-1 promotes phagolysosome formation through activating RAB-7, whose function is essential for promoting the recruitment and fusion of lysosomes to phagosomes.

PI(3)P and RAB-7 Are Sequentially Enriched on Phagosomal Surfaces
We examined the maturation process of phagosomes containing apoptotic cells in C. elegans. Previously, we reported the transient recruitment of DYN-1 and the incorporation of early endosomes, which are labeled with HGRS-1, the worm homolog of mammalian endosomal protein Hrs, to the surface of phagosomes [1]. Here, we examined two additional molecular events on phagosomal surfaces: the accumulation of PI(3)P [19,20] and the recruitment of small GTPase Rab7 [8]. We monitored both events in developing embryos using modified versions of an established protocol (Materials and Methods) [1]. In all time-lapse experiments described below, we chose to follow the engulfment and degradation of three particular apoptotic somatic cells, C1, C2, and C3, which are located near each other on the ventral side of an embryo and die almost simultaneously during embryogenesis, at approximately 330 min past the first cell division (the first cleavage) ( Figure 1C(l) and 1D(a)). ABplaapppa, ABpraapppa, and ABplaapppp, three ventral hypodermal cells, engulf C1, C2, and C3, respectively ( Figure  1C(f) and 1D(a)), during their extension to the ventral midline [1]. In addition, all green fluorescent protein (GFP) or monomeric red fluorescence protein (mRFP1) [21] tagged reporters were expressed specifically in engulfing cells under the control of P ced-1 , the ced-1 promoter [15].
The FYVE domain of C. elegans EEA-1, in a tandem repeat, specifically associates with PI(3)P [22]. The FYVE-FY-VE::mRFP1 reporter was detected primarily in cytoplasm as bright puncta, consistent with its endosomal localization ( Figure 1C) [1]. The 2xFYVE markers were also evenly distributed in nuclei ( Figure 1D(d)). In engulfing cells, we observed bright mRFP1 circles surrounding cell corpses

Author Summary
Cells undergoing programmed cell death, or apoptosis, within an animal are swiftly engulfed by phagocytes and degraded inside phagosomes, vesicles in which the apoptotic cell is bounded by the engulfing cell's membrane. Little is known about how the degradation process is triggered and controlled. We studied the degradation of apoptotic cells during the development of the nematode Caenorhabditis elegans. Aided by a newly developed livecell imaging technique, we identified multiple cellular events occurring on phagosomal surfaces and tracked the initiation signal to CED-1, a phagocytic receptor known to recognize apoptotic cells and to initiate their engulfment. CED-1 activates DYN-1, a large GTPase, which further activates downstream events, leading intracellular organelles such as endosomes and lysosomes to deliver to phagosomes various molecules essential for the degradation of apoptotic cells. As well as establishing a temporal order of events that lead to the degradation of apoptotic cells, the results suggest that phagocytic receptors, in addition to initiating phagocytosis, promote phagosome maturation through the sequential activation of multiple effector molecules.
C. elegans rab-7 encodes RAB-7, a likely ortholog of human Rab7 ( Figure S1A and Text S1). In wild-type animals expressing P ced-1 gfp::rab-7, a functional GFP::RAB-7 was detected in the cytoplasm, and a portion of it is enriched on cytoplasmic puncta ( Figures S1B and S3D). This punctate localization pattern is consistent with previous reports indicating that RAB-7 is localized to late endosomes and lysosomes [24,25]. In both embryos and adult hermaphrodite gonads, we observed robust GFP::RAB-7 signals around cell corpses ( Figure S1C). Quantitative measurements of timelapse images (Materials and Methods) indicated that the intensity of the GFP signal on phagosomal surfaces could reach as high as 2.5 times that in the cytosol of the same engulfing cell ( Figure 1D(o)). In a time-lapse recording of embryos that coexpress GFP::RAB-7 and 2xFYVE::mRFP1, RAB-7 was recruited to the surface of a phagosome approximately 3 min after PI(3)P appeared on the same phagosome ( Figure 1D and 1E). After the completion of engulfment, yet prior to the recruitment of RAB-7, a nascent phagosome appeared as a dark hole surrounded by evenly distributed GFP::RAB-7 in the engulfing cell cytoplasm ( Figure 1D(h) and 1D(i)). Like 2xFYVE::mRFP1, GFP::RAB-7 persisted on the surface of a phagosome until the phagosome disappeared ( Figure 1D and 1E, and Video S1).
(E) Summary of the temporal order of each reporter detected on the surfaces of phagocytic cups and phagosomes. Data represent means obtained from time-lapse recording of multiple C1, C2, and C3 cell corpses (see text for actual numbers). ''0 min'' represents when budding pseudopods are first detected by CED-1::GFP. 2xFYVE, HGRS-1, RAB-7, and CTNS-1 reporters localize to phagosome surface until the cell corpse is fully degraded. The conversion of pale to deep pink colors of HGRS-1::mRFP1 and CTNS-1::mRFP1 indicates that starting from an initial weak signal, the signal intensity gradually increases. doi:10.1371/journal.pbio.0060061.g001  Previously, whether C. elegans RAB-7 is involved in cellcorpse removal was unknown. To determine the function of RAB-7 in the degradation of cell corpses suggested by its specific phagosomal localization, we characterized rab-7(ok511), a deletion allele obtained from the C. elegans Gene Knockout Consortium (http://www.wormbase.org). We confirmed the presence of a genomic deletion that eliminates the first three exons and a 265-bp upstream sequence of rab-7 in this allele ( Figure 3A). Since no additional in-frame ATG codon exists flanking or downstream of the site of deletion, rab-7(ok511) is likely a null allele.
Homozygous rab-7(ok511) (m þ z À ; where m represents the maternal gene product and z the zygotic gene product) embryos produced by rab-7(ok511)/þ mothers develop normally and do not contain excess cell corpses ( Figure 3C). However, after reaching the adult stage, their gonads contain a large number of germ-cell corpses ( Figure 3B and 3C). In contrast, due to the swift removal of cell corpses, the gonads of wild-type hermaphrodites at the same age contain very few cell corpses despite the continuous occurrence of apoptosis events in the germline ( Figure 3B and 3C).
The homozygous rab-7 (m À z À ) progeny of rab-7(ok511) (m þ z À ) animals display 98% embryonic lethality and arrest at various embryonic stages ( Figure 3D), indicating an essential function for maternal RAB-7 during embryonic development. We further observed various developmental defects in rab-7 (m À z À ) embryos, including defects in the consumption of yolk and failure in hypodermal cell closure ( Figures 3D(c) and S2, Text S1, and Video S2). During embryogenesis, 113 cells undergo apoptosis [12]. However, at late 4-fold stage, the final stage before hatching, no cell corpses can be found in wildtype embryos due to their prompt removal ( Figure 3C and 3D(a)). By contrast, rab-7 (m À z À ) embryos that arrest at 4-fold stage contain many persistent cell corpses ( Figure 3C and 3D(d)), again suggesting defects in their removal.
To verify that the persistent cell-corpse-like objects are indeed apoptotic corpses, the ced-3(n717) mutation, which blocks most, if not all, programmed cell deaths [26,27] was introduced into the rab-7(ok511) background. In rab-7(ok511); ced-3(n717) double mutants, no cell-corpse-like objects were observed in the m þ z À adult hermaphrodite gonads or m À z À embryos ( Figure 3C), confirming that the cell-corpse-like objects accumulating in rab-7 mutant animals are apoptotic cells. rab-7 Mutant Animals Are Defective in the Removal of Cell Corpses, Not the Execution of Programmed Cell Death Accumulation of cell corpses may be caused by excessive apoptosis or defects in the removal of cell corpses. To distinguish between these two possibilities, we monitored the generation and duration of germ-cell corpses induced by cray irradiation in rab-7 (m þ z À ) adults (Materials and Methods). We chose a 2-h recording period starting at 3 h after irradiation, since in the wild-type germline, the number of cell corpses reaches its peak 4 h after irradiation [28]. Within this period, we observed similar numbers of germ cells undergoing apoptosis in wild-type and rab-7(ok511) animals (average ten and nine apoptotic cells, respectively). This result suggests that rab-7(ok511) mutants are no more susceptible to apoptosis stimuli than wild-type animals. On the other hand, we observed that in rab-7 mutants, all germcell corpses analyzed persisted for a significantly longer period of time than those in wild-type animals ( Figure 3E), indicating that rab-7 mutant gonads are defective in the prompt removal of germ-cell corpses.

Loss of rab-7 Function Results in Specific Defects in the Degradation of Engulfed Cell Corpses
To further distinguish whether the persistent germ-cell corpses observed in rab-7 mutants are results of defects in engulfment or degradation, we analyzed germ-cell corpses and their surroundings using TEM (Materials and Methods). Apoptotic germ cells cellularize from the germline syncytium and are quickly engulfed by gonadal sheath cells, a single layer of somatic cells that wrap around germline syncytium [27]. We analyzed serial sections that span the entire diameter of each cell corpse to determine whether a cell corpse is totally inside a sheath cell. In rab-7(ok511) mutants, all germ-cell corpses observed in three different gonad arms were engulfed by gonadal sheath cells ( Figure 3F). The accumulation of engulfed germ-cell corpses suggests that the degradation, and not the engulfment, of cell corpses is blocked. In addition, in rab-7 mutants, 90% of excessive germ-cell corpses are labeled with 2xFYVE::GFP from engulfing cells, respectively ( Figure  2B and 2C), again indicating that they persist inside phagosomes.
Under TEM, in 59% of corpses in rab-7 mutants, the nuclear envelope and plasma membrane were readily distinguishable, and the nucleolus appeared as a distinct patch with high electron density ( Figure 3F(c) and 3F(e)). These well-preserved cellular characteristics suggest a defect in the initiation of cell-corpse degradation. In the rest of cell corpses, the electron density of osmium tetroxide-stained sections was greatly enhanced, and it was impossible to distinguish nucleoli or any intracellular membranous struc-fold stage (d). rab-7(ok511) (m À z À ) embryos were synchronized and were scored 12 h after being laid. The percentage of each class of phenotype is labeled. Arrows indicate cell corpses. Arrowheads indicate granules. Scale bar indicates 10 lm. (E) DIC time-lapse recording of germ-cell death induced by c-ray irradiation and their subsequent removal in the gonad of wild-type (a-e) and rab-7(ok511) (f-j) adult hermaphrodites. Scale bar indicate 5 lm. (k) Histogram indicating distribution of the durations of germ-cell corpses. n, number of cell corpses scored. Two animals were used for each genotype. (F) TEM images of a rab-7(ok511) (m þ z À ) gonad. (e), (c), and (g) Indicate a cross section (e) and its enlarged regions that contain germ-cell corpses (c) and (g). Traces of membranes corresponding to (e), (c) and (g) are shown in (a), (b) and (f), respectively. Further enlarged images of framed regions in (c) and (g) are shown in (d) and (h), respectively. White arrows indicate germ-cell membrane; black arrows indicate sheath cell membrane; white arrowheads indicate cell corpse membrane; white open arrowheads indicate the nuclear envelope of the cell corpse. Scale bars indicate 5 lm in (e); 1 lm in (c) and (g); and 250 nm in (d) and (h). DCC, dark cell corpse. (i) Percentages of engulfed cell corpses in the gonads of wild-type and rab-7(ok511) at 48 h post-L4. A superscript letter a ( a ) indicates data from [1]. doi:10.1371/journal.pbio.0060061.g003 tures ( Figure 3F(e) and 3F(g)), suggesting that a certain degree of phagosome maturation had occurred.
In 1.5-and 2-fold stage rab-7 (m À z À ) embryos, 78% of cell corpses are labeled with the PI(3)P marker (Figure 2A and 2C), indicating that they are inside phagosomes. Thus, somatic cell corpses in rab-7 embryos, like germ-cell corpses, appear to be engulfed efficiently yet remain undegraded inside engulfing cells. We refer to the accumulation of engulfed cell corpses as a cell-corpse degradation defective (Ded) phenotype.
The rab-7 (m þ z À ) Adults Are Relatively Normal in Endocytosis and Lysosomal Functions To examine whether the defects in the degradation of germ-cell corpses are an indirect consequence of defects in endocytosis, lysosome biogenesis and function, and/or other vesicle transport events that might be affected by the loss of rab-7 function, we analyzed a number of relevant aspects in rab-7(ok511) (m þ z À ) adult hermaphrodites, which display a strong Ded phenotype, but develop normally and appear normal in general morphology (see above). To examine possible defects in exocytosis and endocytosis, we first characterized the delivery of yolk to oocytes. In adults, yolk, a lipoprotein complex, is secreted from intestinal cells and transported to the gonad where it is internalized by oocytes through receptor-mediated endocytosis [29]. By monitoring YP170 (a yolk component)::GFP [29], we observed that just as in wild type [1], 100% of fertilized embryos inside rab-7(ok511) (m þ z À ) adult hermaphrodite gonads contain internalized YP170::GFP ( Figure 4A). This result indicates that in these adults, the exocytosis and the subsequent endocytosis of yolk into oocytes are both normal. The accumulation of germ-cell corpses thus is unlikely to be a secondary consequence of a general endocytosis and/or exocytosis defect.
We compared the number and morphology of CTNS-1::GFP(þ) particles in several tissues in wild-type and rab-7 (m þ z À ) adult hermaphrodites and did not observe any obvious difference between these strains ( Figure 4B). Recently, it was reported that in rab-7(ok511) (m þ z À ) adult hermaphrodites, the acidification of lysosomal compartments in coelomocytes, special scavenger cells that are highly active in endocytosis, was normal [32]. These lines of evidence suggest that the biogenesis and functions of lysosomes in rab-7(ok511) (m þ z À ) adult animals are relatively normal, probably owing to the residual wild-type maternal gene activity. By contrast, in rab-7(ok511) (m À z À ) embryos that lack both maternal and zygotic gene activity, we detected severe defects in lysosomal functions ( Figure S2 and Text S1). We thus conclude that the accumulation of engulfed cell corpses observed in the gonads of rab-7(ok511) (m þ z À ) adult animals is unlikely a secondary consequence of general defects in lysosome biogenesis or function.
We observed that in rab-7(ok511) (m þ z À ) hermaphrodites, coelomocyte vacuoles, which represent endocytic compartments [33], appeared the same size as in wild type, although they accumulated in a larger number ( Figure 4C). This is in sharp contrast to the phenotypes observed in typical lysosome biogenesis mutants, such as cup-5 or ppk-3 mutants, on which coelomocyte vacuoles are enlarged [25,33]. On the other hand, mutations in cup-5 or ppk-3 resulted in no or very weak Ded phenotypes ( Figure 4D). In addition, in cad-1 mutants, in which the activity of aspartyl protease cathepsin D, a lysosomal hydrolase, was severely reduced [34], no Ded phenotype was observed ( Figure 4D). These results imply that in addition to its function in regulating lysosome biogenesis, RAB-7 mediates an independent function related to phagosome maturation.
Rab proteins are known to cycle between the GTP-bound, active state and the GDP-bound, inactive state [36]. A T22N mutation of mammalian Rab7 ( Figure S1A) results in a great decrease of its in vitro GTP-binding activity [37]; on the other hand, the Q67L mutation ( Figure S1A) decreases its GTP hydrolysis activity, and thus locks it in a GTP-bound and active state [38,39]. We analyzed these two mutations for their effects on RAB-7 function. The transgene P ced-1 gfp::rab-7(T23N) did not rescue the Ded phenotype in rab-7(ok511) mutants ( Figure 5A), indicating that GTP binding is essential for RAB-7 function. On the other hand, P ced-1 gfp::rab-7(Q68L) displayed a rescuing activity similar to (in gonads and embryos) or stronger than (in L1 heads) that of P ced-1 gfp::rab-7 ( Figure 5A), suggesting that like many other small GTPases, GTP RAB-7 is an active form in cell-corpse degradation.
In engulfing cells, GFP::RAB-7(Q68L), like GFP::RAB-7, was enriched on the surface of phagosomes (unpublished data). In contrast, GFP::RAB-7(T23N) failed to accumulate to phagosomes ( Figure 5C). These results indicate that only the GTPbound, active form of RAB-7 is recruited to phagosomal surfaces and establish a correlation between the localization and function of RAB-7 on phagosomal surfaces.

The Loss of rab-7 Function Does Not Affect the Presentation of PI(3)P on Phagosomal Surfaces or the Incorporation of Early Endosomes to Phagosomes
We observed that in both embryos and gonads, the percentage of cell corpses labeled by 2xFYVE::GFP circles were similar between wild type and rab-7 mutant (Figure 2), indicating that the enrichment of PI(3)P on phagosomal surfaces is normal in rab-7 mutant. In addition, in embryos that in engulfing cells expressed HGRS-1::GFP, a marker for early endosomes [1], both the percentage of phagosomes labeled with HGRS-1::GFP (39% and 42%, respectively) and the intensity of GFP on phagosomal surfaces were similar between wild type and rab-7(ok511) (Figure 2A and 2C) [1]. This result indicates that the function of RAB-7 is not essential for the delivery of early endosomes to phagosomes. The numbers of DIC(þ) cell corpse in rab-7(ok511) (m À z À ) animals at different stages that express different transgenes. Data are presented as mean 6 SD and are labeled above the bars. NA, not applicable due to 100% embryonic lethality. At least ten animals were scored for each datum. (B) Epifluorescence (a) and DIC (b) images of a 4-fold stage rab-7(ok511) (m À z À ) embryo that expresses P egl-1 gfp::rab-7. Arrows indicate dying cells that express GFP::RAB-7. Scale bar indicates 10 lm. The Incorporation of Lysosomes to Phagosomes Is Severely Defective in rab-7 Mutants In wild-type embryos, 64% of cell corpses are labeled with CTNS-1::GFP on their surfaces ( Figure 6A), indicating that lysosomes are incorporated into phagosomes. We monitored the delivery of CTNS-1::GFP signal to phagosomal surfaces in both embryos that expressed P ced-1 ctns-1::gfp and embryos that coexpressed P ced-1 gfp::rab-7 and P ced-1 ctns-1::mrfp1. The dynamic CTNS-1::GFP recruitment to phagosomes occurred in two distinct phases. During phase I, which started after the formation of nascent phagosomes and lasted approximately 20 min, a weak and punctate CTNS-1::GFP signal was observed on the surface of phagosomes ( Figure 6B(g-i), 6C(a-d), and 6C(o)). During the following phase II, CTNS- (B) Time-lapse images of epifluorescence signals on the C2 phagosome (arrows) in a wild-type embryo coexpressing P ced-1 ctns-1::mrfp1 and P ced-1 gfp::rab-7. Scale bar indicates 5 lm. ''0 min'' represents when GFP::RAB-7 is first recruited to C2. Arrowheads indicate tubular structures extended from phagosomes. (C) Time-lapse images of CTNS-1::GFP signal on the C1 phagosome (arrows) in wild-type and rab-7(ok511) (m À z À ) embryos. ''0 min'' represents when the nascent C1 phagosome is formed. Scale bars indicate 5 lm. (o) and (p) Quantitative data measured from images (a-n), including CTNS-1::GFP intensity ratio (phagosome/cytosol) and phagosome volume, are plotted over time. The two phases of CTNS-1 localization pattern are labeled as I and II. (D) Histogram indicating the distribution of the durations of phagosomes in wild-type and rab-7(ok511) (m À z À ) embryos measured using CTNS-1::GFP as a reporter. n, number of phagosomes scored. doi:10.1371/journal.pbio.0060061.g006 1::GFP rapidly accumulated on a phagosome, and gradually evolved from multiple punctate spots to a smooth, continuous circle ( Figure 6B(i-l) and 6C(d-g)). Following the rapid enrichment of CTNS::GFP signal, the volume of a phagosome quickly decreased ( Figure 6B, 6C(o), and 6C(p)), indicating degradation of phagosomal contents. A phagosome remained CTNS-1::GFP(þ) until it was completely degraded ( Figure 6B and 6C). The accumulation of the CTNS-1::GFP signal is indicative of the recruitment of lysosomes to phagosomal surfaces, whereas the subsequent transition from punctate to continuous circles may represent the fusion of lysosomes to phagosomal membranes. The recruitment of RAB-7 to phagosomes took place earlier than the phase II CTNS-1 enrichment (Figures 1E and 6B), with an average interval of 13.4 6 5.8 min (n ¼ 9).
In rab-7(ok511) embryos, as in wild type, CTNS-1::GFP is localized to cytoplasmic puncta ( Figures 6A). However, we observed severe defects in the incorporation of lysosomes into phagosomes. Most (65%) of all CTNS-1::GFP circles observed around cell corpses in 1.5-fold rab-7(ok511) embryos remain punctate ( Figure 6A(e)). Time-lapse recording results indicated that the localization of CTNS-1::GFP to phagosomes was mostly arrested at phase I: the signal intensity on phagosomes remained low ( Figure 6C(o)), suggesting a defect in the recruitment of lysosomal particles; furthermore, the signal observed on phagosomal surfaces remained in a punctate pattern for an extended period of time (.100 min) ( Figure 6C(h-n)), suggesting a likely defect in lysosomes/ phagosome fusion.
To quantify the degradation defect caused by the rab-7(ok511) mutation, we monitored the duration of CTNS-1::GFP-labeled phagosomes. The volume of phagosome C1 in wild-type embryo decreased 5-fold within 40 min of its presence, yet remained unchanged over 2 h in a rab-7(ok511) embryo ( Figure 6C(p)), indicating a lack of digestion of phagosomal contents. In wild-type embryos, phagosomes disappeared within a period of 30-70 min after their formation; in rab-7(ok511) embryos, ten phagosomes monitored all lasted much longer than 90 min ( Figure 6D). These results indicate that the defects in the incorporation of lysosomes to phagosomes caused by rab-7 deletion are closely associated with the blockage of the degradation of phagosomal contents.

RAB-7 Promotes the Extension of Tubular Structures for the Recruitment of Lysosomes to Phagosomal Surfaces, and the Fusion of Lysosomal Particles to Phagosomes
We further monitored individual CTNS-1::GFP(þ) puncta over time and observed that in wild-type embryos, the CTNS-1(þ) puncta were delivered to the surfaces of phagosomes through either direct encounters with phagosomes or attachment to membrane tubules originating from phagosomes. Tubules extending from phagosomal surfaces were frequently observed; these tubules were labeled with RAB-7 and CTNS-1 ( Figures 6B and 7A), but not with PI(3)P ( Figure 1D). As shown in Figure 7A and Video S3, within 6 min, a CTNS-1(þ) punctum was seen to attach to such a tubule, and was carried to the phagosome by the retracting tubule. In rab-7(ok511) embryos, in contrast, we did not observe any tubular structures extending from phagosomes (Figure 7 and Video S4). The lack of extended tubules presumably leads to the decreased targeting of lysosomes to phagosomes, which subsequently results in insufficient delivery of lysosomal enzymes. Indeed, on the surface of 43% of phagosomes that we monitored in rab-7(ok511) embryos, no CTNS-1(þ) particles were ever observed to be attached during the entire 100-min observation period ( Figure 7B and Video S4). These observations suggest that RAB-7 is essential for the extension of membrane tubular structures from phagosomes, a process necessary for the efficient recruitment of lysosomes to the surface of phagosomes.
In wild-type embryos, after attaching to the phagosomal surface, lysosomal particles rapidly lost their distinct shape and became part of the phagosomal surface, a process likely suggesting the fusion of lysosomal particles to phagosomal membranes ( Figures 6C(a-g) and 7A). In rab-7(ok511) embryos, however, in some cases, even after the attachment to phagosomal surfaces, the particles maintained their distinct shape for various amounts of time, sometimes greater than 50 min ( Figure 7A(i-p) and Video S4), indicating an apparent lack of fusion. The defects in the attachment and possible fusion of individual puncta to phagosomes apparently correlate with the persistent punctate state and the low signal intensity of CTNS-1::GFP on phagosomal surfaces ( Figure 6C(h-n)).

Mutations in ced-1, dyn-1, and ced-5 Delay the Degradation of Engulfed Cell Corpses
Previously, due to the lack of markers to label phagosomes and time-lapse imaging techniques to follow phagosome maturation, the ced genes acting in the two parallel pathways led by ced-1 and ced-5 were only implicated in the engulfment of cell corpses (Introduction). The diffused cytoplasmic localization pattern of GFP::RAB-7, in addition to its association to cytoplasmic puncta, allows us to follow the extension and closure of pseudopods during engulfment ( Figure S4A and S4B). In addition, a phagosome can be distinguished by the enriched GFP::RAB-7 signal on its surface ( Figure 8A(b)), or even before being labeled by GFP::RAB-7, recognized as a dark, GFP(À) hole inside a GFP::RAB-7(þ) engulfing cell ( Figure 8A(a)). To further study whether the engulfment and degradation of cell corpses employ common mechanisms, we examined the kinetics of the engulfment and degradation in ced-1, dyn-1, and ced-5 mutant embryos using time-lapse imaging. ced-1 and ced-5 mutant embryos are viable [13]. Although dyn-1(n4039) embryos undergo developmental arrest, the arrest point is at 4-fold stage (.620 min after first cleavage) [1]. Thus our time-lapse recording, performed between 300-420 min after first cleavage, was not affected by the arrest.
We monitored the engulfment of cell corpses C3 in ced-1(e1735), dyn-1(n4039), and ced-5(n1812) null mutant embryos [1,15,40] that express P ced-1 gfp::rab-7. We demonstrated that these three mutants displayed distinct and severe defects in the initiation and/or the extension of pseudopods around cell corpses ( Figure S4B-S4E and Text S1), consistent with the well-characterized functions of these proteins in initiation of engulfment. Interestingly, in only 19%-25% of single-mutant embryos, C3 remained unengulfed for longer than 100 min; in the other embryos, C3 were eventually engulfed ( Figure  S4E). In addition, in 1.5-fold stage ced-1 and ced-5 mutant embryos, approximately 20%-30% of DIC(þ) cell corpses appeared as either dark holes or GFP-labeled phagosomes inside engulfing cells ( Figure S4F and S4G), indicating that these cell corpses were engulfed, but not yet degraded.
The partial defects in engulfment allowed us to monitor the degradation of those cell corpses that were engulfed in ced-1, dyn-1, and ced-5 embryos. In wild-type embryos, the duration of a phagosome, which was the total time period that it existed as a dark hole (nascent phagosome) and then as a GFP::RAB-7(þ) circle, was 30-70 min ( Figure 8A(a-e) and 8B). In ced-1, dyn-1, and ced-5 mutants, however, the duration varied within a wider range, with a distinct population of phagosomes that lasted longer than 90-100 min, the time limit of our time-lapse recording due to the vigorous body movement of embryos occurring after reaching approximately 440 min after first cleavage ( Figure 8B). The volume of the phagosomes belonging to this population did not reduce throughout the observation period ( Figure 8A(v)), indicating a severe defect in the degradation of phagosomal content. This defect is further confirmed using CED-1C (the intracellular domain of CED-1)::GFP, another reporter that is evenly distributed in the cytoplasm of engulfing cells and allows the recognition of phagosomes as GFP(À) holes ( Figures 8C and S5). The above observations confirm the defects of dyn-1 mutants in phagosome maturation, which was observed before using different assays [1], and further reveal novel functions for ced-1 and ced-5 in the degradation of apoptotic cells, a step downstream of engulfment.

ced-1, dyn-1, and ced-5 Mutants Are Defective in the Recruitment of Lysosomes to Phagosomes
Since the formation of phagolysosomes is an essential step in the digestion of phagosomal contents (Figure 6), we first analyzed the incorporation of lysosomes to maturing phag-osomes containing C1, C2, and C3 in ced-1, dyn-1, and ced-5 embryos by monitoring CTNS-1::GFP. We observed that although the initial localization of weak CTNS-1::GFP(þ) signal on phagosomal surfaces in phase I was relatively normal; the enrichment of CTNS-1::GFP signal expected for phase II was delayed or absent on the majority of phagosomes in all three mutants (Figure 9). In many cases, the weak CTNS-1::GFP signal on phagosomal surfaces was unstable ( Figure  9A). The defects of lysosomes incorporation in ced-1 and dyn-1 mutant embryos are more severe than that in ced-5 embryos ( Figure 9D(a)), correlating with the levels of degradation defects displayed by the corresponding mutants ( Figure 8B). Although the recruitment of CTNS-1(þ) particles was severely delayed, once recruitment was complete, in most cases, the corresponding phagosome would disappear within 50 min ( Figure 9B and 9D(b)). For instance, we observed different fates of two phagosomes in one dyn-1(n4039) embryo: the weak, initial CTNS-1 signal labeling the phagosome containing C3 failed to increase over time, and C3 remained undegraded after 65 min ( Figure 9A), whereas the phagosome containing C1 was eventually degraded after a delayed enrichment of CTNS-1 signal (at 31 min) on its surfaces ( Figure 9B). These results suggest that the inefficient incorporation of lysosomes into phagosomes is one of the major causes for the cell-corpse degradation defects observed in ced-1, dyn-1, and ced-5 mutants.
The Accumulation of RAB-7 on the Surface of Phagosomes Is Delayed or Blocked in ced-1 and dyn-1, but Not ced-5, Mutants Since RAB-7 activity is essential for the incorporation of lysosomes to phagosomes, we monitored the dynamic :GFP-labeled lysosome particle (white arrowheads) being delivered to C3 phagosome (arrows) in wild-type (a-h) and rab-7(ok511) (m À z À ) (i-p) embryos that express P ced-1 ctns-1::gfp. Open arrowheads indicate tubular structures. Scale bars represent 5 lm. (B) Quantitative data deduced from multiple time-lapse recording experiments similar to (A) that score the defects in the recruitment and fusion of lysosome to phagosomes in rab-7(ok511) (m À z À ) embryos. doi:10.1371/journal.pbio.0060061.g007 association of RAB-7 with phagosomes in ced-1, dyn-1, and ced-5 mutants. In both wild-type and ced-5 embryos, greater than 70% of phagosomes are labeled with a GFP::RAB-7 circle within 10 min after the completion of engulfment ( Figure  8D). By contrast, in ced-1 and dyn-1 embryos, the recruitment of RAB-7 to phagosomal surfaces is delayed or blocked. The majority of phagosomes in ced-1 (82%) and dyn-1 (79%) mutants remained RAB-7(À) (detected as dark holes) for greater than 10 min ( Figure 8A(f-t) and 8D, and Videos S5-S7). In particular, 35% and 17% of phagosomes in ced-1 and dyn-1 mutant embryos, respectively, remained RAB-7(À) for greater than 60 min ( Figure 8A(k-t) and 8D); all 11 such (B and C) Histogram distribution of phagosome durations in embryos using GFP::RAB-7 (B) and CED-1C::GFP (C) as reporters. Duration is measured from the formation of a nascent phagosome (indicated by a dark hole) until its degradation. n, number of phagosomes C1, C2, and C3 measured using timelapse recording. (D) Histogram distribution of the time that it took to recruit RAB-7 to phagosomes, measured from the formation of a nascent phagosome (dark hole) until the GFP::RAB-7 signal intensity on the phagosomal surface is at least 1.2 fold as high as that in the host cell cytosol. n, number of phagosomes scored. doi:10.1371/journal.pbio.0060061.g008 phagosomes failed to reduce in size throughout the recording period (for instance, Figure 8A(p-t)). These phenotypes indicate that (1) the recruitment of RAB-7 to phagosomal surfaces is essential for phagosome maturation, and (2) the recruitment is inefficient in ced-1 and dyn-1 mutants but relatively normal in ced-5 mutants.
We examined the localization of RAB-7(Q68L) in ced-1(e1735) mutant embryos and found that the recruitment of RAB-7(Q68L) to phagosome surface is also delayed. However, the defect is less severe: whereas greater than 30% of phagosomes took more than 60 min to recruit RAB-7 after engulfment completes, all phagosomes recruit RAB-7(Q68L) within 50 min, with the majority within 30 min ( Figure S6). These results suggest that enhancing GTP binding appears to alleviate the defect of RAB-7 recruitment caused by the loss of ced-1 function.

Mutations in ced-1, ced-6, and dyn-1 Delay the Presentation of PI(3)P on Phagosomal Surfaces
To identify additional events important for phagosomal maturation regulated by the CED-1 pathway, we monitored the presentation of PI(3)P on phagosomes, which occurs prior to the recruitment of RAB-7 in wild-type ( Figure 1D), ced-1(e1735), ced-5(n1812), ced-6(n2095) (a strong loss-of-function allele) [41], and dyn-1(n4039) embryos using the 2xFYVE::GFP or ::mRFP1 markers. In time-lapse experiments, nascent phagosomes not yet labeled with 2xFYVE markers were distinguished as dark holes inside engulfing cells ( Figure  10A(b)). We observed the delays of PI(3)P presentation to different extents in different mutant embryos. In wild-type embryos, all phagosomes were labeled with 2xFYVE::GFP within 10 min after the completion of engulfment, among which 87% were labeled within 6 min ( Figures 1C and 10C). The ced-5(n1812) mutation resulted in a slight delay of the appearance of PI(3)P: on 62% of phagosomes, PI(3)P was detected within 11-20 min after engulfment (Figures 10C and  S7C, and Video S8). ced-6(n2095) embryos displayed an intermediate delay of PI(3)P presentation (Figures 10C and  S7A). The ced-1 and dyn-1 embryos displayed the most severe defects-only 14% and 8% of phagosomes, respectively, were labeled with PI(3)P within 10 min of engulfment ( Figure 10A-10C, and Videos S9 and S10). In addition, PI(3)P was absent from 21% and 42% of phagosomes, respectively, during a period that lasted greater than 60 min, suggesting that the presentation of PI(3)P on the surface of these phagosomes was blocked (Figures 10C and S7B, and Video S10). These observations suggest that the severe delay and/or blockage of PI(3)P presentation in ced-1 and dyn-1 mutants might affect multiple downstream maturation events.

C. elegans as a Genetic System for Studying the Mechanisms Controlling the Degradation of Apoptotic Cells
Previously, due to the lack of cellular markers, whether a cell corpse was engulfed could not be easily distinguishable in C. elegans embryos. We have developed a series of live-cell imaging-based assays and have provided the first real-time observation of the entire cell-corpse removal process in the context of a developing embryo. Using these assays, we also established the temporal sequence of multiple phagosome maturation events. These in vivo assays have revealed that C. elegans RAB-7 plays essential roles in the formation of phagolysosomes. Our study indicates that the degradation of apoptotic cells in C. elegans shares certain mechanistic features with that of opsonized foreign objects in mammalian macrophages.
Unlike TEM, which only detects the terminal phenotypes, the time-lapse recording system allows us to observe previously unknown defects in the dynamic process of phagosome maturation. We thus identified distinct roles of four C. elegans proteins, CED-1, CED-6, DYN-1, and CED-5, in the degradation of engulfed apoptotic cells ( Figure 11). CED-1, CED-6, and CED-5 were only known for their functions in engulfment (see Introduction). Whether the mammalian homologs of these proteins are similarly involved in the maturation of phagosomes remains to be examined. Our findings suggest that C. elegans is a powerful model system for identifying additional genes and new mechanisms regulating phagosome maturation.
Mammalian macrophages elicit pro-or anti-inflammatory responses after engulfing pathogens, which are considered ''foreign,'' or apoptotic cells, which are considered ''self,'' respectively (see Introduction). The mechanisms behind these distinct responses remain largely unknown, and phagosomal surfaces are possible sites where these immune responses could be initiated. By studying the removal of apoptotic cells in C. elegans, we may find clues to those unknown mechanisms. For example, we observed that PI(3)P was enriched on C. elegans nascent phagosomes and persisted there until the complete digestion of phagosomal content. On the contrary, PI(3)P was reported to be transiently present on the surface of mammalian phagosomes containing opsonized objects, disappearing well before the completion of phagosome maturation [20]. These observations may reflect differences in the maturation of phagosomes with different contents. The Specific and Direct Functions of C. elegans RAB-7 for the Degradation of Apoptotic Cells C. elegans rab-7 was implicated in controlling cell survival based on the observation that rab-7 (RNA interference [RNAi]) resulted in excessive germ-cell corpses in the adult hermaphrodite gonads [42]. Here, we report multiple lines of evidence demonstrating that the excessive germ-cell corpses observed in rab-7 null mutants are unlikely caused by excessive cell death; rather, they are the result of defects in the degradation of engulfed apoptotic cells (Figure 3).
Mammalian Rab7 is involved in endocytic pathway and lysosome biogenesis [37,[43][44][45]. Consistently, C. elegans RAB-7 is essential for endosomal trafficking and yolk consumption ( Figure 11C) [29,32]. Our research provides several lines of evidence to suggest that the Ded phenotype of rab-7 mutants is not merely a secondary consequence of the defects in the aforementioned functions, rather, RAB-7 plays a direct and independent role in phagosome maturation by acting on the surface of phagosomes. In adult rab-7 (m þ z À ) mutant hermaphrodites, whereas intracellular trafficking and lysosome biogenesis are relatively normal due to the activity of rab-7 maternal product, the degradation of engulfed cell corpses is severely impaired. In addition, mutations in ced-1 and dyn-1 cause strong and correlated defects in the recruitment of RAB-7 to phagosomal surfaces and in the formation of phagolysosomes, further supporting an independent role of RAB-7 on phagosomal surfaces.
Studies of the mammalian phagosomes containing infectious bacteria indicate that the functions and regulation of host-cell Rab7 during phagosome maturation are complex and diverse, and depend on the ingested bacteria species [46][47][48]. We found that regarding the degradation of apoptotic cells, which are distinct from latex beads, which are not degradable, and from infectious bacteria, which are foreign objects, RAB-7 displays both conserved and unique features. Mammalian Rab7 and its effector RILP were implicated in the extension of tubular structures from phagosomal surfaces, which attract late endosomes and lysosomes to phagosomes containing latex beads, based on the effects of the overexpressed dominant-negative forms of Rab7 and RILP [8]. We observed a similar extension of tubular structures from phagosomes and the recruitment of CTNS-1(þ) lysosomal particles along those tubules in C. elegans (Figure 7). Furthermore, by analyzing rab-7 null mutant embryos, we discovered that one of the functions of endogenous RAB-7 is to promote the extension of these tubules and the recruitment of lysosomes to phagosomes. We have also identified another function of RAB-7 in promoting the fusion of lysosomes to phagosomal surfaces. This fusion activity is consistent with the reported activity of Ypt7, the yeast counterpart of Rab7, in promoting the physical contact of vacuoles during homotypic vacuole fusion (reviewed in [49]), and indicates that RAB-7 is capable of mediating the fusion of organelles of different origins. In addition, dominantnegative Rab7 inhibits acidification of mammalian phagosomes that contain latex beads [8]. In contrast, we found that the null mutation of rab-7 did not significantly affect the acidification of phagosomes containing apoptotic cells ( Figure S8 and Text S1). Phagosomes containing apoptotic cells thus provide a new experimental system for further studying the mechanisms behind the functions and regulation of Rab7.

How Is RAB-7 Recruited to the Phagosomal Surface-A Novel Signaling Pathway That Initiates the Formation of Phagolysosomes
In engulfing cells, RAB-7 exists in the cytoplasm both in a diffused form and in association with intracellular organelles, in particular lysosomes. The initial phagosomal appearance of RAB-7 occurs prior to the phagosomal acquirement of lysosomes, suggesting that free, cytosolic RAB-7 molecules are likely the primary source for phagosome-associated RAB-7. The GTP-bound form of RAB-7 is the active form recruited to and stably associated with phagosome, perhaps through a particular protein or protein complex that that mediates this association.
We have identified the novel functions of at least two genes, ced-1 and dyn-1, for the timely recruitment of RAB-7 to phagosomes. This finding, together with the finding that RAB-7 is essential for the recruitment and fusion of lysosomes into phagosomes, indicate that the CED-1 pathway controls phagolysosome formation through regulating RAB-7 recruitment and function. Previously, we have reported that the phagocytic receptor CED-1 acts upstream of DYN-1 in the signaling pathway for engulfment [1]. We have also established that the transient localization of DYN-1 on the surface of phagocytic cups and phagosomes are both dependent on CED-1 function [1]. Thus we propose that CED-1 acts upstream of DYN-1 to control RAB-7 recruitment ( Figure  11B).
The molecular mechanism for the recruitment of RAB-7 remains to be investigated. The observation that the recruitment of a GTP-bound, constitutively active form of RAB-7 to phagosomal surface was less affected than that of RAB-7 by the ced-1 null mutation ( Figure S6) may indicate that the CED-1 pathway may induce the formation of GTP-RAB-7 around phagosomal surfaces and the subsequent RAB-7/phagosome association. Alternatively, the CED-1 pathway may act in a more indirect manner, by recruiting or activating early signaling molecules that will further recruit RAB-7 ( Figure  11A). We observed that PI(3)P was dramatically enriched on phagosomal surfaces prior to RAB-7. Furthermore, in dyn-1, ced-6, and ced-1 mutants, the enrichment of PI(3)P on phagosomal surfaces was delayed. These observations suggest that PI(3)P might be related to the recruitment of RAB-7.

The CED-1 Pathway Controls Multiple Important Events for Phagosome Maturation
The enrichment of RAB-7 on phagosomes is not the only phagosome maturation event regulated by the CED-1 signaling pathway. For instance, the recruitment of early endosomes to phagosomes, which is independent of RAB-7 ( Figure  2), is dependent on CED-1, CED-6, and DYN-1 [1]. In addition, we observed that the rate of enrichment of PI(3)P on phagosomal surfaces is under the control of ced-1, ced-6, and dyn-1. PI(3)P attracts FYVE or PX domain-containing effectors, some of which play diverse roles during phagosome maturation, to phagosomal surfaces (reviewed in [50]). CED-1::GFP was observed to transiently cluster on nascent phagosomal surfaces for merely 9 min, in a time period that barely overlaps with that of PI(3)P and RAB-7 ( Figure 1E). However, DYN-1 remains on phagosomal surfaces for a longer period (18 min). CED-1 thus may establish an active state of phagosomal surfaces through the activation of DYN-1. The rapid accumulation of PI(3)P on the surfaces of nascent phagosomes is likely to reflect the local activation of a Class III PI(3) kinase [20], whose activity and/or localization on phagosomal surfaces may be controlled by the CED-1 pathway.
We thus propose that CED-1, CED-6, and DYN-1 act in a signaling pathway to control both the engulfment and the degradation of engulfed apoptotic cells, through regulating different downstream effectors ( Figure 11B). On nascent phagosomes, or perhaps even before the phagocytic cup is enclosed, this pathway may actively prime the phagosomal membrane for the reception of multiple signaling molecules and sequential incorporation of intracellular vesicles ( Figure  11A).
Although extensive studies suggest that nascent phagosomes possess multiple signaling molecules on their surfaces, the identity of the ultimate initial signal that triggers phagosome maturation remains unclear. An activated phagocytic receptor may well be the source of this signal. When in complex with extracellular ligands, phagocytic receptors were known to undergo certain conformational changes that lead to the activation of their intracellular domains [51], which in turn may act on the cytoplasmic side of phagocytic cups and phagosomes to recruit cytosolic factors. Our report that CED-1 initiates phagosomal maturation through activating downstream effectors suggests that phagocytic receptors for apoptotic cells in different organisms, including mEGF10, the human CED-1 homolog [52], and perhaps even phagocytic receptors for foreign pathogens, may act to regulate the downstream phagosome maturation events. Furthermore, by activating different phagocytic receptors, phagocytic targets that are ''self'' or ''foreign'' might thus initiate differential phagosome maturation processes.

CED-5 Controls Phagosome Maturation through a Distinct Pathway
ced-5 null mutants are defective in the recruitment of lysosomes to phagosomes. However, unlike ced-1 or dyn-1 mutations, null mutation in ced-5 does not seem to affect the recruitment of RAB-7 on phagosomal surfaces, nor does it cause any severe defect in the presentation of PI(3)P on phagosomal surfaces. These results imply that CED-5 acts in a distinct pathway to control phagolysosome formation ( Figure  11B), just like in engulfment ( Figure 1B). During engulfment, the pathway led by CED-5 was shown to regulate cytoskeletal reorganization (reviewed in [11]). Recently, it was reported that cytoskeletal reorganization also plays an active role in phagosome maturation in mammalian cells [10,53]. It would be interesting to know whether CED-5 and other members of its pathway contribute to phagosome maturation through remodeling the cytoskeleton.
Determining the gene structure of rab-7 in the rab-7(ok511) allele. rab-7 genomic DNA was PCR amplified using VC308 lysate as a template. The fragment amplified from the ok511 allele, which was 742 bp shorter than wild type, was sequenced, and the exact boundaries of the deletion, tgtgatttttc... aattcctccaa, as reported by C. elegans Gene Knockout Consortium, was confirmed.
Transmission electron microscopy. Adult hermaphrodites 48 h past the mid-L4 stage were used for TEM analysis of germ-cell corpses and the neighboring gonadal sheath cells as previously described [1]. Serial 50-nm thin sections that cover the entire length of each cell corpse were analyzed to determine whether a germ-cell corpse is entirely inside the sheath cell.
Quantifying the number of cell corpses. Both somatic cell corpses in embryos and germ-cell corpses in the adult hermaphrodite gonads were scored under the Nomarski DIC microscope by their highly refractile appearance as previously established [1]. Embryonic stages were determined according to [1]. In brief, 1.5-and 2-fold stages are embryos whose body lengths are 1.5 or two times the length of an egg, respectively. Late 4-fold stages embryos are embryos whose bodies make three turns and contain fully developed pharyngeal grinders. Wild-type embryos at 1.5-, 2-, and early and late 4-fold stages correspond to embryos at approximately 420 min, 445 min, 660 min, and approximately 780 min past the first cleavage, respectively [1].
Irradiation and time-lapse recording of the duration of germ-cell corpses. Time-lapse analysis was performed on a DeltaVision system with Olympus IX70 microscope (Applied Precision). Adult hermaphrodites 24 h post-L4 stage were irradiated with a 137 Cs source (Gammacell 1000, 8.33Gy/min) at 180Gy. 2.5 h after irradiation, samples were transferred onto NGM plates containing 1 mM Aldicarb; after 20 min, samples were mounted on a slide immersed in 0.8 mM Aldicarb solution and observed under DIC optics at 20 8C. Twenty serial Z-sections, at 1.0 lm/section, of the gonadal region were recorded every 3 min. Animals were closely monitored for viability during recording. The total number and the period of duration of cell corpses that occur during the first 2 h were scored.
Fluorescence microscopy and time-lapse recording. An Olympus IX70-Applied Precision DeltaVision microscope equipped with a Photometris Coolsnap digital camera and Applied Precision Softworx software were used to capture and deconvolve fluorescence images. To score the number of phagosomes labeled with a certain fluorescence marker, serial Z-section images of an entire embryo were recorded at 0.5-lm intervals, in 40 sections, and images were deconvolved before scoring. To monitor the process of engulfment and degradation of cell corpses C1, C2, and C3, as well as the subcellular localization of multiple cellular markers, recording of the ventral surface of embryos started at 310-320 min after first cleavage and lasted for 2 h in 1-or 2-min intervals. For each time point, eight serial Z-sections at a 0.5-lm interval were recorded. Signs such as embryo elongation and movement were closely monitored to ensure the embryo being recorded developed normally.
Quantitative image analyses. Fluorescence signal intensity was measured and images were analyzed using the ImageJ software. To calculate the ratio of signal intensity on the surface of phagosomes compared to that in the engulfing cell cytosol, the phagosomal surface and cytoplasm were individually defined using the ''Free Hand'' tool, and the modal value of each selected area was used in the ratio calculation. The enrichment of fluorescence signal on the phagosomal surface was considered significant when the ratio was greater than 1.2. To measure the volume (V) of phagosomes, we assumed they were spheres. The radius (R) of each phagosome was measured using the ''Measuring'' tool provided by ImageJ, and V is calculated as 1.333 p R 3 . Under the described image-capture conditions, the pixel distance is 0.133 lm/pixel. To measure the size of yolk droplets, the threshold tool was adjusted to mark all droplets labeled with YP170::GFP in an embryo. The size of every droplet larger than 0.13 lm 2 was analyzed and collected to generate the distribution histogram. Figure S1. Characterization of C. elegans RAB-7 (A) Sequence alignment between C. elegans and human Rab7. Solid triangles indicate the residues mutated in RAB-7(T23N) and RAB-    (B) Time-lapse images of the engulfing process in embryos that express P ced-1 gfp::rab-7. ''0 min'' represents when ABplaapppp started to fuse with ABpraapppp at the ventral midline (arrowheads). Scale bars indicate 2 lm. Arrows indicate C3. In the wild-type embryo, pseudopods extension started and completed before ABplaapppp fused with ABpraapppp. In ced-1(e1735) mutant, pseudopods extension did not start until fusion occurred. In ced-5(n1812) mutant, pseudopods extension started before cell fusion occurred, but did not complete until fusion was complete due to a slower pseudopod extension process.    Figure 8D). Found at doi:10.1371/journal.pbio.0060061.sg006 (319 KB TIF). Figure S7. Examples of the Delayed Appearance of PI(3)P on Phagosomal Surfaces Observed in ced-6, dyn-1, and ced-5 Mutants Time-lapse image series of three embryos in different mutant backgrounds, either coexpressing P ced-1 ced-1::gfp and P ced-1 2xFY-VE::mrfp1 (A) or expressing P ced-1 2xFYVE::gfp (B-C), starting at approximately 330 min after first cleavage. Anterior is to the top. Ventral faces readers. Scale bars indicate 5 lm. ''0 min'' represents the time point that nascent phagosomes (white arrows) first form. The inset inside each panel is a 2-fold enlargement of the region around the white arrow, with the shape of the engulfing cell highlighted. (A) PI(3)P is not presented on the surface of the C1 phagosome (white arrows) until 18 min after engulfment is completed in a ced-6(n2095) embryo. (B) PI(3)P is not presented on the surface of the C2 phagosome (white arrows) throughout the recording period of 95 min in a dyn-1(n4039) embryo. A yellow arrow indicated a C3 phagosome on which the appearance of PI(3)P is relatively normal. (C) PI(3)P appears on the surface of the C1 phagosome (white arrows) 10 min after the completion of engulfment in a ced-5(n1812) embryo. A yellow arrow indicates a C3 phagosome on which the timing of the PI(3)P appearance was normal. Found at doi:10.1371/journal.pbio.0060061.sg007 (7.5 MB TIF). Figure S8. Phagosome Acidification in rab-7(ok511) Mutant Is Normal (A) The numbers of germ-cell corpses detected by DIC and 2xFYVE::GFP at 28 h post-L4 stage in wild-type and rab-7(ok511) (m þ z À ) adult hermaphrodites that express P ced-1 2xfyve::gfp, in the absence or presence of irradiation (IR) treatment (Materials and Methods). Data are presented as mean 6 SD. n, number of animals scored. (B) Fluorescence and DIC images of irradiated gonads of wild-type and rab-7(ok511) (m þ z À ) adult hermaphrodites that are stained with Lysosensor Yellow/Blue (Materials and Methods). Arrows indicate cell corpses. Arrowheads indicate normal germ cells. Scale bar indicates 10 lm. (C) Percentage of cell corpses that display different staining intensities in the gonads of wild-type and rab-7(ok511) (m þ z À ) adult hermaphrodites. n, number of cell corpses examined. At least three different gonadal arms were scored for each genotype. Found at doi:10.1371/journal.pbio.0060061.sg008 (2.2 MB TIF).
Video S1. The Dynamic Process of the Enrichment of GFP::RAB-7 on Phagosomal Surface and the Degradation of a Cell Corpse Epifluorescence images of GFP::RAB-7 in a wild-type embryo expressing P ced-1 gfp::rab-7 are assembled in a time-lapse movie. The interval between two consecutive frames is 2 min. The white arrow indicates C2. At the ''0 min'' time point, C2 was partially internalized into the engulfing cell. At ''2 min,'' C2 is completely engulfed and a nascent phagosome first formed. Anterior is to the top. Ventral faces the reader. Found at doi:10.1371/journal.pbio.0060061.sv001 (6.1 MB MOV).
Video S2. The Development Process of a rab-7(ok511) (m À z À ) Embryo Showing Multiple Phenotypes DIC images were recorded starting at the first cleavage. The first image of this movie is at 400 min after first cleavage. The interval between two consecutive frames is 2 min. Anterior is to the left. Judging by morphology at 400 min after first cleavage, the embryonic development before that stage was delayed. The movie also shows that this embryo failed to continue elongation after reaching 2-fold of body length. In addition, hypodermal cells failed to fully enclose at the head region and as a result, the head tissue leaked out of the soma. Found at doi:10.1371/journal.pbio.0060061.sv002 (8.5 MB MOV).
Video S3. In a Wild-Type Embryo, CTNS-1(þ) Particle First Docks onto the Tubular Structure Extended from Phagosome and Subsequently Fuse into Phagosome through Retraction of the Tubule The movie was captured with 1 min between each frame. The whole process, including both the recruitment and fusion, is less than 8 min. Found at doi:10.1371/journal.pbio.0060061.sv003 (2.4 MB MOV).
Video S4. In a rab-7(ok511) (m À z À ) Embryo, the Recruitment and Fusion of CTNS-1-Labeled Lysosomal Particles to Phagosomal Surface Are Both Defective This movie records the journey of a single CTNS-1(þ) particle (arrow) towards a phagosome in a rab-7 embryo expressing P ced-1 ctns-1::gfp. Images captured in 2-min intervals are shown. It took the CTNS-1(þ) particle 16 min to be recruited to phagosomal surface, which is significantly longer than what was observed in wild-type embryos, indicating a defect in recruitment. In addition, this CTNS-1(þ) particle failed to fuse to phagosomal membrane after attachment, in a period of 52 min, indicating a fusion defect. Found at doi:10.1371/journal.pbio.0060061.sv004 (9.7 MB MOV).
Video S5. The Enrichment of RAB-7 on Phagosomal Surfaces Is Moderately Delayed in a ced-1(e1735) Embryo This embryo expresses P ced-1 gfp::rab-7. Arrows indicate phagosome C2. ''0 min'' is the time point when cell corpse C2 is half engulfed. Timelapse images are shown in 2-min intervals. RAB-7 enrichment is not detected on the surface of C2 until the 20 min point. The volume of phagosome C2 does not decrease throughout the recording period,