• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. Apr 2008; 74(8): 2275–2287.
Published online Feb 15, 2008. doi:  10.1128/AEM.02646-07
PMCID: PMC2293164

Cell Invasion and Matricide during Photorhabdus luminescens Transmission by Heterorhabditis bacteriophora Nematodes[down-pointing small open triangle]


Many animals and plants have symbiotic relationships with beneficial bacteria. Experimentally tractable models are necessary to understand the processes involved in the selective transmission of symbiotic bacteria. One such model is the transmission of the insect-pathogenic bacterial symbionts Photorhabdus spp. by Heterorhabditis bacteriophora infective juvenile (IJ)-stage nematodes. By observing egg-laying behavior and IJ development, it was determined that IJs develop exclusively via intrauterine hatching and matricide (i.e., endotokia matricida). By transiently exposing nematodes to fluorescently labeled symbionts, it was determined that symbionts infect the maternal intestine as a biofilm and then invade and breach the rectal gland epithelium, becoming available to the IJ offspring developing in the pseudocoelom. Cell- and stage-specific infection occurs again in the pre-IJ pharyngeal intestinal valve cells, which helps symbionts to persist as IJs develop and move to a new host. Synchronous with nematode development are changes in symbiont and host behavior (e.g., adherence versus invasion). Thus, Photorhabdus symbionts are maternally transmitted by an elaborate infectious process involving multiple selective steps in order to achieve symbiont-specific transmission.

In most animals, healthy intestines are colonized by commensal and beneficial bacteria (8, 34). A fundamental question is how beneficial interactions are established and maintained while pathogenic interactions are resisted. To better understand symbiont discrimination in the animal intestine, transmission of the insect pathogen Photorhabdus luminescens by nematode hosts was investigated. Photorhabdus spp. are associated with entomopathogenic nematodes, such as Heterorhabditis bacteriophora (6). Knowledge about the relationship between these two taxa and about the insecticidal toxins produced by the symbionts may prove to be useful for the control of insect pests. An emerging human pathogen, Photorhabdus asymbiotica, was recently shown to be vectored by Heterorhabditis entomopathogenic nematodes (13). Colonization of developmentally arrested infective juvenile (IJ)-stage H. bacteriophora by P. luminescens is essential for both partners to infect insects and reproduce in nature (7, 20). Germfree IJs infect insects, but they do not cause insect mortality or reproduce efficiently (20). H. bacteriophora IJs selectively vector P. luminescens or Photorhabdus temperata bacteria in their gut mucosa before regurgitating the symbionts into an insect host (7, 14). Since both partners of the symbiosis are required for insect pathogenesis and insects are thought to be the preferred niche for both partners, there is likely strong selective pressure on symbiont-specific transmission.

The life cycle of H. bacteriophora is initiated by the environmentally resistant IJs usually inhabiting soil. After sensing an insect host, the IJs enter the insect hemocoel through natural openings, such as the mouth or anus, or by using a buccal tooth to slice through the cuticle and then regurgitate their intestinal symbionts (7). Following symbiont release, insect mortality occurs rapidly (usually <48 h), and the IJs exit diapause and resume development, a process known as recovery. Symbiotic bacteria that proliferate in the hemocoel preserve the insect cadaver by producing broad-spectrum antibiotics and/or signals or nutrients essential for nematode growth and reproduction (5, 33).

H. bacteriophora reproduces by either laying eggs or developing eggs internally inside the maternal body cavity. Internal development of larvae ultimately causes matricide, a process termed endotokia matricida, and the larvae develop into IJs (24). Nematodes that develop with symbiotic bacteria inside infected insects or on agar media undergo two or three generations before the IJs are generated en masse, and most of the IJs vector symbiotic bacteria to new insect hosts. It is reported here that symbiotic bacteria are maternally transmitted to IJs by the following elaborate sequence of events: (i) adherence to the maternal posterior intestine, (ii) growth within the intestinal lumen, (iii) invasion of the rectal gland cells (RGCs), (iv) release into the maternal body cavity, (v) adherence to the pharyngeal intestinal valve cells (PIVCs), (vi) invasion of the PIVCs, and (vii) colonization of the IJ intestinal lumen.


Media and culture conditions.

The sources and descriptions of strains used in this study are shown in Table Table1.1. Photorhabdus spp. were grown in PP3salt-2% proteose peptone no. 3 (Difco, Detroit, MI) containing 0.5% NaCl (Sigma-Aldrich, St. Louis, MO), and agar (1.5%), gentamicin (0.75 μg/ml), streptomycin (40 μg/ml), and kanamycin (3.75 μg/ml) were added when they were required. Escherichia coli was grown in lysogeny broth (3) modified so that it contained 5 g/liter NaCl, and agar (1.5%), gentamicin (5 μg/ml), ampicillin (50 μg/ml), and diaminopimelic acid (300 μg/ml) were added when they were required. P. luminescens subsp. laumondii was isolated from H. bacteriophora TT01 kindly provided by Ann Burnell (NIU-Maynooth, Ireland) by placing IJs that were surface sterilized for 5 min in 1% commercial bleach onto PP3salt.

Strains and plasmids used

Propagation of H. bacteriophora.

Nematodes were propagated on 8 g of nutrient broth with 10 μg/ml cholesterol (NA+chol) or on 8 g of nutrient broth with 12 ml of corn oil (Mazola) per liter containing a lawn of Photorhabdus spp. pregrown for 18 to 24 h at 28°C in a culture dish (60 by 15 mm) or on one side of a divided petri dish (100 by 15 mm). After growth for 10 to 14 days at 28°C, the IJs were washed off the lid of the culture dish with sterile saline (0.85% NaCl), or 12 ml of sterile saline was added to the empty side of the split petri dish. When IJs form on lawns, they have dispersive behavior and become trapped on condensation on a culture dish lid or in the saline opposite the culture in a split-well dish. Germfree IJs were first generated by growing the H. bacteriophora nematodes on P. temperata subsp. temperata isolated from Heterorhabditis megidis as described previously (21). However, better worm yields were obtained with P. temperata strain NC1 TRN16, a transmission-defective mutant that is unable to colonize the IJs and normally is associated with H. bacteriophora nematodes. An inbred strain of H. bacteriophora TT01, M31e, had been self-fertilized for 13 generations prior to use in this study (9). H. bacteriophora was maintained as a germfree stock and added to P. luminescens subsp. laumondii TT01, P. temperata strain NC1, or one of the strains labeled with Tn7-green fluorescence protein (GFP) as described previously (19). The bacteria were frozen in 4.5% dimethyl sulfoxide in PP3salt and stored at −80°C, and the IJs were cryopreserved and stored in liquid nitrogen as described previously (31).

Analysis of worm development.

Greater wax moth (Galleria mellonella) larvae were obtained from Nature's Way (Ross, OH). Larvae were infected by placing 20 larvae in sterile a 100-mm Pyrex culture dish containing a 9.0-cm Whatman no. 1 (Florham Park, NJ) filter, after which ~1,000 monoxenic IJs were added in 0.7 ml of sterile saline. Beginning 48 h postinfection, larvae were removed every 24 h and disrupted in Ringer's solution (100 mM NaCl, 1.8 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 5 mM HEPES; pH 6.9) in a petri dish (100 by 15 mm), in which all worms, laid eggs, and external IJs were counted. For analysis of nematode development on lawns of symbiotic bacteria, ~25 IJs were placed on NA+chol preseeded with P. luminescens TT01 as described above. Beginning 48 h after IJ addition, worms were washed off the lawns every 24 h with 1.8 ml of Ringer's solution, centrifuged for 1 min at 2,000 rpm, and washed three times in Ringer's solution, after which all worms, laid eggs, and external IJs were counted. When the number of worms exceeded the number which could be counted accurately (>300 worms), values were extrapolated from counts for three aliquots containing 60 to 300 worms each. Worms were counted in triplicate in three independent experiments.

Assay for retention of intestinal symbionts by IJs after exit from diapause.

To determine if IJs vertically transmit symbiotic bacteria directly to offspring, IJs containing GFP-labeled bacteria were placed on lawns of unlabeled bacteria. Two to twelve hours after addition, nematodes were removed, and the presence of symbiotic bacteria in the nematodes was determined by fluorescent and differential interference contrast or Nomarski microscopy using a Leica DM5000 compound microscope (Leica Microsystems, Wetzlar, Germany) equipped with an X-cite 120 fluorescence illuminator (EXFO, Quebec, Canada), a Spot Pursuit charge-coupled device camera (Diagnostic Instruments, Sterling Heights, MI), and a GFP filter set (Leica). To determine if starvation stress triggered vertical transmission, IJs in the process of regurgitating their intestinal symbionts were removed and starved for 12 h in Ringer's solution before they were placed on lawns of unlabeled bacteria and imaged as described above.

Test for the presence of intestinal bacteria.

H. bacteriophora IJs and Caenorhabditis elegans dauer larvae were each added to lawns of GFP-labeled P. temperata NC1, an H. bacteriophora symbiont that does not kill C. elegans on nematode growth media (data not shown), and observed 24 h after addition. To test if nonsymbiotic bacteria are permitted in the H. bacteriophora intestine, IJs were incubated with TRN16, a P. temperata transmission-defective mutant, for 24 h before transfer for 4 h to lawns of GFP-labeled Escherichia coli OP50 and then directly observed. To determine if P. temperata persists in the H. bacteriophora intestine in the presence of nonsymbiotic bacteria, H. bacteriophora IJs were added to lawns of DSRedexpress-labeled NC1 symbionts and incubated for 24 h before transfer to GFP-labeled OP50 for 24 h and then observed.

Detection of symbiotic bacteria that have infected nematodes.

To determine when nematodes are infected by symbiotic bacteria, nematodes were pulsed with GFP-labeled symbionts, which was followed by a 4-h chase with unlabeled symbionts. Most transient intestinal bacteria were defecated from the intestine during the chase, so only GFP-labeled symbionts that had established persistent infections were visible. Unless specified otherwise, these experiments were performed by adding ca. 25 germfree IJs to lawns of GFP-labeled symbionts and incubating the preparations for 12 to 144 h at 28°C. At least every 12 h during this period, at least 15 worms were picked from the bacterial lawns into a drop of Ringer's solution on a sterile PP3salt plate and then transferred to a lawn of unlabeled bacteria for 4 h before imaging. At least three independent experiments were performed. To determine the growth of adherent bacteria on the maternal intestine, IJs were added to lawns of GFP-labeled bacteria, incubated for 12 h, removed, washed three times in Ringer's solution, and placed on lawns of unlabeled bacteria as described above. Nematodes were removed after 4, 12, or 24 h, and the growth of the initially adherent GFP-labeled bacteria was analyzed by counting fluorescent cells by epifluorescent microscopy. To determine if GFP-labeled bacteria adhered to the male nematodes, first-generation (F1) adult males were removed, washed, chased with unlabeled bacteria, and analyzed to determine the presence of fluorescent bacteria as described above. To determine if adherence occurred in a maternal intestine containing a biofilm of unlabeled bacteria, germfree IJs were placed on unlabeled bacteria for 30 h, washed, and then placed on GFP-labeled bacteria and incubated for 4 h, after which they were washed in Ringer's solution, chased with unlabeled bacteria for 4 h, and then imaged as described above. To determine when the window of transmission was closed (i.e., the time at which new symbionts could no longer infect the intestine), germfree IJs were propagated using either unlabeled symbionts or TRN16 that was unable to colonize the maternal intestine and 24, 36, 48, and 60 h after IJ addition were placed on GFP-labeled bacteria for 4 h and then chased with unlabeled bacteria as described above.

Transmission electron microscopy.

For electron microscopy, live animals were fixed, rinsed, and stained using microwave irradiation to help with penetration of the animal's thick cuticle. A Pelco Biowave oven (Ted Pella, Inc., Redding, CA) was tuned to very low power (70 W), and the animals were kept in a plastic culture plate containing liquid fixative that was placed on top of the Pelco Coldspot device inside the oven chamber, which helped minimize sample heating during extended irradiation. A temperature probe was used, and the restriction temperature used was 39°C, although this temperature was rarely reached. Multiple processing steps were done in the chamber, using irradiation to aid each rinse step or staining. Samples were fixed first in aldehydes and then in osmium tetroxide plus potassium ferrocyanide, stained en bloc in uranyl acetate, and finally embedded in 3% agarose prior to dehydration and embedding in Embed812 plastic resin (18). The following conditions and temperature regimen were used. (i) Samples in cacodylate buffer containing 3.5% glutaraldehyde and 1.5% paraformaldehyde were microwaved twice (5 min on and 3 min off) and then kept at room temperature for 60 min. The temperature ranged from 15 to 39°C during irradiation. The 0.1 M cacodylate buffer (pH 7.2) included 2 mM CaCl2 and 50 mM NaCl. (ii) The samples were rinsed three times in 0.2 M cacodylate buffer, microwaved for 1 min, and then kept at room temperature for 10 min. (iii) The samples were placed in 1% OsO4-0.5% KFe(CN)6 in 0.1 M cacodylate buffer, microwaved twice (5 min on and 3 in off), and then kept at room temperature for 15 min. The temperature ranged from 15 to 35°C during irradiation. (iv) The samples were rinsed three times in 0.2 M cacodylate buffer, microwaved for 1 min, and then kept at room temperature for 7 min. (v) The samples were rinsed three times in 0.2 M sodium acetate buffer (pH 5.2), microwaved for 1 min, and then kept at room temperature for 7 min. (vi) The samples were placed in 0.5% uranium acetate in 0.2 M sodium acetate buffer, microwaved twice (5 min on and 3 min off), and then kept at room temperature for 15 min. The temperature ranged from 15 to 33°C during irradiation. (vii) The samples were rinsed three times in 0.2 M sodium acetate buffer (pH 5.2), microwaved for 1 min, and then kept at room temperature for 7 min. (viii) The samples were rinsed three times in 0.2 M cacodylate buffer, microwaved for 1 min, and then kept at room temperature for 7 min. (ix) Samples were embedded in parallel in 3% type VII agarose and then kept at 4°C overnight. (x) The samples were cut into small agar cubes and transferred to snap cap vials in buffer. (xi) The samples were dehydrated at room temperature using the following conditions: 70% ethanol for 10 min, 80% ethanol for 10 min, 90% ethanol for10 min, three treatments with 100% ethanol (10 min each), and three treatments with 100% propylene oxide (PO) (10 min each). (xii) The samples were infiltrated at room temperature on a rotator using the following conditions: 2 parts PO to 1 part resin for 2 h, 1 part PO to 2 parts resin for 2 h, and four changes of 100% resin over 1 day. (xiii) Samples were arranged in a flat embedding mold and cured at 60°C for 65 h. Thin sections were obtained using a Power Tome XL ultramicrotome (RMC, Boekeler Instruments, Tucson, AZ), stained with 2% uranyl acetate in 50% ethanol for 10 min and with lead citrate (Reynold's formulation) for 15 min, and viewed with a JEOL100 CXII transmission electron microscope (Japan Electron Optics Laboratories Ltd., Tokyo, Japan) located at the Michigan State University Center for Advanced Microscopy.


IJs develop inside maternal nematode body cavities and not from laid eggs.

To determine if P. luminescens transmission to the IJ vector directly involves maternal nematodes, it is important to determine if IJs develop inside maternal nematodes, outside nematodes from laid eggs, or both. Development of the IJs involves three developmental pathways and behaviors: reproduction by egg laying versus endotokia matricida (i.e., intrauterine egg hatching); development of offspring into IJs (i.e., alternative developmentally arrested third-stage larval nematodes that vector P. luminescens to insect hosts) versus vegetative noninfective hermaphrodites, females, or males; and persistence of IJ diapause leading to dispersal behavior versus exit from diapause (i.e., recovery) and resumption of development (Fig. (Fig.1A).1A). H. bacteriophora offspring arising from endotokia matricida develop predominantly into IJs (24). Conversely, hermaphroditic nematodes presumed to be pre-IJs develop from laid eggs (25), but whether such progeny can also develop into IJs has not been determined previously.

FIG. 1.
IJs develop inside maternal body cavities and not from laid eggs. (A) H. bacteriophora makes three key developmental or behavioral choices related to IJ formation: (i) egg-laying behavior versus intrauterine egg hatching (endotokia matricida); (ii) development ...

To determine if IJs develop from laid eggs and/or via endotokia matricida, nematode development was monitored both inside infected insect larvae (G. mellonella) and on agar-based media seeded with symbiotic bacteria by determining the numbers of laid eggs, IJs, and total worms (laid eggs plus IJs plus other stages) present 2 to 11 days after IJs were added (Fig. (Fig.1B1B and and1C).1C). Most IJs that were added to the symbiont lawns or that infected insects recovered and resumed development. Two days after IJ addition, 13 recovered IJs (standard error of the mean [SEM], 7.9 IJs) and 13 recovered IJs (SEM, 6.7 IJs) were present (recovered IJs were distinguished from nonrecovered IJs on the basis of feeding behavior and morphology) inside infected insects and on symbiont lawns, respectively (Fig. (Fig.1B1B and and1C).1C). Two days later (4 days after IJ addition), the animals had molted twice and developed into adults that were laying eggs, and 1,650 laid eggs (SEM, 1,315 eggs) and 1,852 laid eggs (SEM, 1,979 eggs) were present in insects and on symbiont lawns, respectively (Fig. (Fig.1B1B to to1D).1D). First- and second-stage larvae (L1 and L2, respectively) derived from hatched eggs were also observed (Fig. (Fig.1D1D and data not shown). The offspring derived from laid eggs developed into vegetative (non-IJ) stages, based on morphology, the fact that no IJs were present on the symbiont lawns prior to day 8, and the fact that IJs accounted for a minority of the total worms (17 and 30% on days 6 and 7, respectively) inside insects (Fig. (Fig.1B1B and and1C).1C). In addition, no IJs appeared on the symbiont lawns before day 8 following removal of parental (P0) nematodes on day 6 (data not shown). After the initial egg-laying period 3 to 5 days after IJ addition, reproduction occurred exclusively by intrauterine egg hatching (endotokia matricida) because few or no eggs were observed either on symbiont lawns or inside insects (Fig. (Fig.1B1B and and1C).1C). Offspring arising from endotokia matricida developed into IJs (which were distinguished from vegetative stages by morphology and behavior). IJs arising from endotokia matricida were first detected 6 days after addition of IJs to insects, but most IJs recovered and resumed development, as shown by the increase in the total worm numbers from day 5 (7,403 worms; SEM, 8,745 worms) to day 6 (15,984 worms; SEM, 21,511 worms), whereas only a minority were IJs (3,567 worms; SEM, 5,173 worms) that remained in diapause (Fig. (Fig.1B).1B). F1 IJs also developed via endotokia matricida on symbiont lawns (data not shown), but all of them recovered and resumed development after they left the maternal body cavity (and thus were not scored as IJs), as shown by the complete absence of IJs before day 8 on symbiont lawns (Fig. (Fig.1C).1C). These data suggest that during the first generation of reproduction and development both in insects and on symbiont lawns maternal nematodes first laid eggs that developed into vegetative (non-IJ) stages and then reproduced via endotokia matricida, where offspring developed into IJs but most IJs recovered and resumed development.

In contrast to P0 nematodes, F1 nematodes reproduced only via endotokia matricida; essentially no laid eggs were observed after day 5 and during days 7 to 11, when IJs developed en masse (Fig. 1 B, ,1C,1C, and and1E).1E). On average, 405,000 IJs (SEM, 77,800 IJs) and 28,500 IJs (SEM, 5,600 IJs) were observed on day 11 in insects and on day 10 on symbiont lawns, respectively (Fig. (Fig.1B1B and and1C).1C). Egg laying occurred too early and the numbers of eggs were insufficient to account for the large numbers of IJs that developed. These data suggest that IJs develop exclusively in maternal nematodes reproducing via endotokia matricida.

Intestinal symbionts are completely released when IJs exit diapause.

Because F1 IJs developed inside the body cavities of P0 nematodes that were originally IJs harboring bacteria, some symbionts might have been retained by the IJ-derived P0 nematodes and directly transmitted to IJ progeny. To determine if IJs retained symbionts after recovery, IJs harboring GFP-labeled symbionts were added to lawns of unlabeled bacteria, and assayed to determine whether they retained GFP-labeled symbionts. Prior to recovery, each IJ contained approximately 130 symbiont bacterial cells that were predominantly in the anterior intestine (see Fig. S1A in the supplemental material). During recovery on lawns of symbiotic bacteria, IJs regurgitated the intestinal symbionts, as previously reported for IJs immersed in insect hemolymph (7). At 4 h after IJ addition to symbiont lawns, IJs regurgitating symbiotic bacteria were observed (data not shown), and ~75% of the IJs contained residual GFP-labeled symbiont cells in the posterior intestine (Fig. S1B), possibly due to the progressive release of the symbionts from the anterior intestine to the posterior intestine. At 8 h, no GFP-labeled symbionts were found in any IJs that had resumed development (see Fig. S1C in the supplemental material).

To determine if stress could induce direct symbiont transmission, IJs which were in the process of recovery and which still contained residual intestinal symbionts (at 4 h) were removed, starved, and assayed to determine whether they retained GFP-labeled intestinal symbionts. No fluorescent bacteria were retained by the starved IJs (data not shown).

Feeding H. bacteriophora contains viable P. luminescens in the intestinal lumen.

H. bacteriophora is a bacteriovore like C. elegans and utilizes bacteria as its primary food source. C. elegans feeds on bacteria by efficiently grinding them using a chitin grinder located in the basal bulb of the pharynx and by producing lytic proteins and enzymes. Recovered H. bacteriophora IJs develop by feeding on symbiotic bacteria and macromolecular components provided by the bacteria. Because intestinal symbionts are completely released during IJ recovery, we sought to determine if the conditions in the intestine of fourth-stage H. bacteriophora larvae are conducive for symbiont survival or growth compared with the conditions in C. elegans. To test this, H. bacteriophora and C. elegans were propagated on GFP-labeled P. temperata NC1, since this strain does not kill C. elegans when it is grown on nematode growth media and is symbiotically associated with H. bacteriophora (data not shown). Intact P. temperata cells were present only in the intestines of H. bacteriophora (see Fig. S2A and S2B in the supplemental material). To determine if nonsymbiotic bacteria are also permitted in the H. bacteriophora intestine, H. bacteriophora IJs were recovered on TRN16, a transmission-defective mutant of P. temperata that is unable to colonize maternal nematodes (see below), and transferred to GFP-labeled E. coli OP50. Labeled OP50 cells were found to survive or grow in the H. bacteriophora intestine to an extent similar to or greater than that of symbiont bacteria (see Fig. S4C in the supplemental material). When H. bacteriophora nematodes were allowed to recover for 24 h on lawns of DSRedexpress-labeled P. temperata symbionts and then transferred for 24 h to lawns of GFP-labeled OP50, DSRedexpress-labeled symbionts were found to persist in the posterior intestine (see below), while GFP-labeled OP50 cells were found throughout the intestine (see Fig. S4D in the supplemental material). Compared to the C. elegans intestine, the H. bacteriophora intestine appears to be more permissive for both symbionts and nonsymbionts, and such cells are available to infect the intestine and other tissues.

Symbiont infection of the posterior maternal intestine.

To determine if and when symbiont transmission is initiated in maternal nematodes, a series of pulse-chase experiments were performed. Germfree (i.e., axenic) IJs were transiently exposed to GFP-labeled symbionts and then chased with unlabeled symbionts (for 4 h) before an assay to determine the presence of GFP-labeled symbionts by fluorescence microscopy was performed. IJs exposed to GFP-labeled symbionts for less than 6 h were not colonized by GFP-labeled bacteria (data not shown). However, symbionts were detected in ~25% of the worms following 8 h of exposure of IJs to GFP-labeled symbionts (data not shown). At 12 h, all IJs that exited diapause contained 1 to 3 GFP-labeled symbiont cells (mean, 2.2 cells; SEM, 0.7 cells) adhering only to the two most posterior intestinal epithelial cells (left and right intestinal epithelial cells [INT9L and INT9R, respectivel]) (Fig. (Fig.2A).2A). More GFP-labeled symbiont cells (mean, 12.2 cells; SEM, 2.3 cells) were found adhering to the posterior intestinal cells following 24 h of exposure (Fig. (Fig.2B),2B), and still more were found following 36 h of exposure (Fig. (Fig.2C2C).

FIG. 2.
Adherence of P. luminescens to maternal and adult male nematode intestines. Transient GFP-labeled cells were chased from the intestine so that only labeled symbiont cells that had established persistent infections were visible. (A) Single GFP-labeled ...

To determine if and when second-generation progeny acquired symbionts, maternal nematodes were placed onto lawns of GFP-labeled symbionts and allowed to lay eggs for 5 h, after which the adults were removed. At 12 h after eggs were laid, most L1 and L2 nematodes contained symbionts in the posterior intestine (data not shown). Like recovered IJs, the L1 and L2 nematodes initially contained one to three symbiont cells in the posterior intestine. Symbiont cells were also found adhering to the adult male posterior intestine in an anatomical location corresponding to that of the INT9 cells, as seen in hermaphrodites (Fig. (Fig.2D),2D), but these cells usually did not persist into late adulthood (data not shown).

To determine if adherent symbiont cells prevent additional symbiont cells from adhering, nematodes containing a biofilm of unlabeled symbionts were exposed to labeled symbionts (for 4 h), chased with unlabeled symbionts (for 4 h), and observed to determine the presence of adherent GFP-labeled symbionts. Adherence of GFP-labeled bacteria was still possible in maternal nematodes containing a mature community of attached symbionts 36 to 42 h after IJ addition (Fig. (Fig.2E).2E). However, at 42 to 48 h after IJ addition, no GFP-labeled bacteria colonized the maternal intestine (Fig. (Fig.2F).2F). Thus, symbiont adherence to the maternal intestine can occur 8 to 42 h after IJ addition to symbiont lawns.

Growth of adherent bacteria on the maternal intestinal epithelium.

Since adherence of new bacteria occurred during the entire time that bacteria were found to be attached to the posterior maternal intestine, it is not clear to what extent the symbionts grow while they are attached to the maternal intestine. To determine the growth of adherent bacteria, IJs were pulse-chased as described above (12-h pulse and 4-h chase), which resulted in one to four adherent symbiont cells (see Fig. S3A in the supplemental material). The equivalent worms 20 h later contained 16 to 25 adherent GFP-labeled symbiont cells, suggesting that there were three or four doublings of adherent bacteria during this time (see Fig. S3B in the supplemental material). These data suggest that the mass of symbiont bacteria present as a biofilm on the maternal posterior intestine was a result of both growth and new adherence.

Invasion of maternal RGCs.

Because the symbiont biofilm persists for up to 36 h before it disappears, the fate of the adherent bacteria was determined during the abrupt transition. At 42 h after addition of IJs to GFP-labeled symbionts, a mass of ~50 cells was found to be attached to the posterior intestinal lumen (Fig. (Fig.2C).2C). At 42 to 48 h after addition of IJs, GFP-labeled cells appeared to migrate, cluster around the rectum, and invade the cytoplasm of RGCs (Fig. (Fig.3A).3A). At 48 h after addition of IJs, GFP-labeled cells formerly attached to the posterior intestine began breaching the gland epithelium and invading the RGCs (Fig. (Fig.3B).3B). Actively invading symbiont cells were recognized by the appearance of a symbiont-containing vacuole in contact with the luminal (apical) face of the INT9 cells (Fig. (Fig.3B).3B). Active invasion is also shown in Fig. Fig.3C3C and and3D,3D, in which in one focal plane symbionts are visible both adhering to the intestinal epithelium and invading the RGCs (Fig. (Fig.3C)3C) and in another focal plane above the intestinal lumen a symbiont-containing vacuole is visible (Fig. (Fig.3D).3D). The likely route for invasion is through the apical surface of the RGCs, which are exposed to the intestinal lumen near the rectum (Fig. (Fig.3A3A to to3C).3C). Most or all of the adherent symbionts appeared to invade the RGCs, because when a few GFP-labeled symbionts were loaded onto an almost mature unlabeled biofilm (36 h after IJ addition), the labeled cells invaded the RGCs (data not shown). By invading the RGCs, the symbionts partially breached the glandular epithelium, but never the intestinal epithelium (see below).

FIG. 3.
Invasion of maternal RGCs in maternal nematodes 48 h after IJs were added to GFP-labeled lawns and then chased with unlabeled symbionts for 4 h. The intestinal lumen (i), rectum (r), and RGC vacuole (v) are indicated. (A) Adherent GFP-labeled symbionts, ...

Intracellular symbiont behavior.

To determine the growth and behavior of the intracellular symbionts, the RGCs were observed after symbiont invasion. Recently invaded maternal RGCs (60 h after IJ addition) contained one to three symbiont-containing vacuoles per cell (Fig. (Fig.4A).4A). Twenty-six hours later (96 h after IJ addition), the symbiont-containing vacuoles had multiplied and there were 12 to 30 vacuoles per cell (Fig. (Fig.4B).4B). The cloverleaf appearance of three RGCs, each filled with symbiont-containing vacuoles and tethered to the rectum, is evident in Fig. Fig.4B.4B. The intracellular symbionts also appeared to multiply during this time (compare Fig. Fig.4A4A and and4B4B).

FIG. 4.
Behavior of intracellular symbionts. (A) Recently invaded RGCs each contained one to three vacuoles following 54 h of exposure of IJs to labeled symbionts. (B) Multiplication of symbiont-containing vacuoles was evident 38 h later, and each RGC contained ...

Since symbiont-containing vacuoles multiply inside the RGCs, the possibility that the intracellular symbionts induce vacuole multiplication is plausible. To test this possibility, maternal nematodes were propagated using P. temperata TRN16, which was unable to infect and invade the maternal intestine. Only a few (<3) enlarged vacuoles were observed in the RGCs without intracellular symbionts 96 h after addition of IJs to TRN16 (Fig. (Fig.4D).4D). Thus, intracellular symbionts influence the size and number of RGC vacuoles.

Symbiont invasion and growth inside RGCs preceded and occurred during egg-laying behavior of recovered IJs (P0 nematodes) (80 to 96 h, equivalent to day 4 in Fig. Fig.1C).1C). However, after 96 h after IJ addition, egg laying ceased and new IJs developed via intrauterine hatching (endotokia matricida). During this time (96 to 108 h), pre-IJs (L2s developing into IJs) began consuming maternal protoplasm and developing inside the maternal body cavity (Fig. (Fig.4C).4C). At 100 h after IJ addition, while the maternal intestine remained intact, the RGCs lysed, and the symbiont-containing vacuoles were liberated into the maternal body cavity but not into the intestinal lumen (Fig. (Fig.4C).4C). The symbiont-containing vacuoles lysed shortly after release (data not shown), which made their bacterial contents available to the pre-IJs present there. The intracellular symbionts were the primary or sole inoculum for the IJs because maternal nematodes propagated on unlabeled bacteria until shortly after RGC invasion (>48 h after IJ addition) and transferred to GFP-labeled symbiont lawns did not transmit any GFP-labeled symbionts to the IJs (data not shown). These data suggest that both biofilm formation on the maternal intestine and subsequent RGC invasion are required for symbionts to breach the glandular epithelium and infect the pre-IJs developing in the maternal pseudocoelom. Otherwise, the intestinal epithelium is well protected against transit of live bacteria to the pseudocoelom.

Ultrastructure of symbiont-containing vacuoles.

To better understand the behavior of intracellular symbionts in RGCs, the ultrastructure of symbiont-containing vacuoles was analyzed using transmission electron microscopy. Symbiont-containing vacuoles with a granular appearance were observed in RGCs (Fig. (Fig.5A).5A). The nearby INT9 cells were easily recognized and never showed any sign of invasion (not shown). The symbiont-containing vacuoles occurred in chains and may have been dividing or expanding, as shown by granular connections between the globular intracellular compartments (Fig. (Fig.5A5A and and5B).5B). Several bacteria were visible inside each RGC vacuole, and a vacuolar membrane was always visible (Fig. (Fig.5B5B and and5C).5C). Numerous membranous blebs were apparent inside the symbiont-containing vacuoles (Fig. (Fig.5C)5C) and were often directly attached to the intracellular symbionts (Fig. (Fig.5D).5D). The inner and outer membranes of the intracellular bacteria, as well as unusual compartments with an unknown function, were clearly visible (Fig. (Fig.5D5D).

FIG. 5.
Ultrastructure of symbiont-containing vacuoles. (A) Transmission electron micrograph of a cross-section of a maternal nematode following 96 h of exposure to symbiont lawns. Symbiont-containing vacuoles are indicated by thick arrows. A few bacterial cells ...

Symbiont infection of IJs developing inside the maternal body cavity.

To monitor the symbiont infection of IJs in maternal nematodes undergoing endotokia matricida, GFP-labeled bacteria were observed in the intestines of pre-IJs developing inside the body cavity. Symbiont infection of the pre-IJs occurred after symbiont release from the RGCs and was apparent before the maternal intestine was disrupted by the pre-IJs developing inside the body cavity (Fig. (Fig.6A).6A). Large numbers of free GFP-labeled bacteria were usually observed in the maternal pseudocoelom (evident in the body cavity in Fig. Fig.6A),6A), suggesting that symbiont replication occurs in the pseudocoelom. Infection of the pre-IJs was initiated inside the maternal body cavity ca. 120 h after IJ addition, and a single symbiont cell usually adhered to the PIVCs located between the pharynx and the intestine (Fig. (Fig.6B).6B). After adherence, the symbiont cell appeared to invade the PIVCs and multiply (Fig. (Fig.6C).6C). Definitive ultrastructural evidence, such as the presence of a vacuolar membrane, was not observed for the intracellular symbionts in PIVCs, and it was also not clear whether the initial adherent cells replicated before invasion or while they were inside the PIVCs. However, 16 h later, multiple cells were again found in the intestinal lumen of IJs (Fig. (Fig.6D);6D); 24 to 72 h later, bacteria were growing in the anterior intestine (Fig. (Fig.6E);6E); and about 7 days later, bacteria were observed throughout the intestine. This full colonization of the IJ intestine completed the transmission cycle.

FIG. 6.
Colonization of pre-IJ second-stage juveniles. (A) Fluorescent micrograph of GFP-labeled symbiont cells (arrows) in the body cavities of two nematodes undergoing endotokia matricida and inside pre-IJs (arrows) (the image was not pseudocolored or overlaid ...

Ultrastructure of IJ intestinal symbionts.

To infect a new insect host, IJs must vector viable symbionts to the insect hemocoel. To better understand how the symbionts are maintained in a semidormant state, sometimes for several months between insect hosts, we compared the ultrastructure of symbiotic IJs and the ultrastructure of germfree IJs (see Fig. S4 in the supplemental material). The IJs were enclosed by a specialized outer (“dauer”) cuticle protecting the nematode from the environment (see Fig. S4A in the supplemental material); the animal's mouthparts were closed (1, 11). The symbionts were present the intestinal lumen in a presumed semidormant state, replicated little, and apparently were not digested by the host. The intestinal lumen contained microvilli that sometimes were in close contact with bacteria (see Fig. S4B and S4E in the supplemental material). No differences in the density of microvilli were observed when germfree and symbiotic nematodes were compared, although an electron-dense matrix was more uniform and diffuse in the intestinal lumen of germfree IJs than in the intestinal lumen of symbiotic IJs (compare Fig. S4C to Fig. S4B and S4D in the supplemental material). A direct physical connection between the symbiotic bacteria and the IJ intestinal epithelium was usually not apparent (see Fig. S4E in the supplemental material). In some IJs, the bacteria appeared to be surrounded by a thin electron-lucent zone distinct from a thick electron-dense matrix in the intestinal lumen (see Fig. S4F in the supplemental material), suggesting that these bacteria were enclosed in a protective acellular matrix or biofilm.


A simple model can explain how mutualistic associations (i.e., associations beneficial to both partners) are established and maintained to the exclusion of nonsymbiotic or pathogenic associations. Visualization of labeled bacteria inside the transparent nematode body revealed an infectious developmental program that ensures symbiont-specific transmission to the IJ nematode. Notably, transmission involved (i) IJ development only inside the maternal nematode body cavities, (ii) symbiont infection of both maternal and IJ offspring intestines, and (iii) temporal and spatial specific host and symbiont cell behaviors.

The majority of the information concerning host-bacterium interactions has been obtained by studying pathogens that cause acute disease, even though many organisms are associated with commensal or mutualistic bacteria, some of which are close relatives of pathogens (29). Recent findings for the model mutualism between the Hawaiian bob-tailed squid (Euprymna scolopes) and the bacterium Vibrio fischeri have revealed a selective gauntlet employed by the squid to ensure symbiont-specific colonization of the light organ (32). This association has revealed the role of virulence-like factors and innate immunity in establishing a symbiotic relationship (16, 26). Similarly, virulence-like genes and innate immunity are involved in the colonization of legumes by symbiotic rhizobia (12). The nematode C. elegans has been used extensively to study the host-bacterium interactions of a variety of pathogens (17, 36). Here we investigated mutualistic host-bacterium interactions with another rhabditid nematode, where transmission of P. luminescens is essential for the insect-pathogenic lifestyle of the two organisms (21). Due to the obligate nature of the association for both the nematode and the symbiont, it is not surprising that the elaborate selective process described in this paper is employed for symbiont transmission.

However, the process of transmission of Photorhabdus to H. bacteriophora IJ nematodes was notably sophisticated. Transmission of a related insect pathogen, Xenorhabdus nematophila, by Steinernema carpocapsae IJs, although far from simple, seems to involve selective binding, growth, and survival of symbionts only in the IJ intestine (15). In the P. luminescens-H. bacteriophora symbiosis, the symbionts are maternally acquired and maintained both extracellularly and intracellularly during most of the nematode life span. Furthermore, symbiont acquisition occurs shortly after IJ recovery and after hatching from laid eggs in vegetative stages destined to produce IJs in subsequent generations via endotokia matricida.

Since symbiont transmission is essential to both the nematode and the bacterium in nature (where each organism depends on the other to infect insects), it is interesting that IJs completely release their intestinal symbionts during IJ recovery. One possible advantage of symbiont release is to avoid Mueller's ratchet, a phenomenon where deleterious mutations tend to accumulate in some vertically (e.g., strictly transovarian) transmitted symbionts (30). In addition, release of the intestinal symbionts into insect hemolymph might select for insect virulence or at least survival in hemolymph that contains efficient humoral and cellular innate immune effectors (28). This selection might also eliminate cheaters which do not contribute to insect virulence. Because a single symbiont cell (and sometimes pairs of cells) was observed adhering to the pre-IJ PIVCs, symbiont transmission is highly clonal. Clonal symbiont transmission might be an effective mechanism to eliminate cheaters (cells that are transmitted but rely on other cells for insect pathogenicity), because in single IJ infections symbiont clones are directly selected for insect virulence.

Symbiont transmission is initiated soon after expulsion from IJs or after hatching from eggs and proceeds through a series of infectious steps (Fig. (Fig.7).7). Symbiont infection of the maternal and IJ intestines is a developmental process, where symbiont cells adhere to, invade, and exit specific nematode cells at specific times in nematode development (Fig. (Fig.7).7). For example, symbiont adherence to nematode cells occurs on only two posterior INT9 intestinal cells in a ~40-h window in the maternal intestine and only to the PIVCs in the pre-IJ anterior intestine (Fig. (Fig.7).7). Similarly, invasion of nematode cells occurs only in specific cells at specific stages of nematode development. Like metazoan development, symbiont transmission likely involves surface components to order cells in space and signals that affect cell behavior. For the symbiont, expression of 1 of 10 predicted fimbrial genetic loci is required for the symbiont to adhere to the maternal intestine (B. Kaufmann-Daszczuk and T. A. Ciche, unpublished data).

FIG. 7.
Model of the transmission cycle. Symbionts that have colonized the maternal intestine (top panel) or pre-IJs or IJs (bottom panel) are shown in the context of select nematode cells in the same orientation on the upper left, where the anterior (A) is on ...

The signals mediating the changes in symbiont behavior are currently not known. The global second messenger, cyclic diguanylate, is a key signaling molecule for social behavior and virulence in many pathogenic bacteria (10, 37). It is surprising that P. luminescens, a highly virulent insect pathogen as well as a mutualistic symbiont, appears to lack proteins containing GGDEF, EAL, and PilZ domains involved in the synthesis, degradation, and sensing of cyclic diguanylate, respectively (22). Other proteins and/or signals might regulate P. luminescens behavior while this bacterium is infecting nematode and insect hosts. It is clear from these studies that the P. luminescens behaviors exhibited during infection of the nematode are also behaviors important to other pathogens. Because P. luminescens is an insect pathogen and a mutualistic symbiont, the mechanisms underlying these infectious behaviors can be directly compared in the two hosts.

H. bacteriophora has adapted what in C. elegans is a stress response behavior and developmental choice, namely, egg laying and dauer formation, for symbiont transmission. Because endotokia matricida occurs in every maternal hermaphrodite and female nematode and results in the development of symbiont-containing IJs, it is clear that endotokia matricida is an adaptation for symbiont transmission. Furthermore, giant nematodes also undergo endotokia matricida. Among the largest nematodes that we observed was a nematode retrieved from an infected G. mellonella larva, which was 6,000 μm long and contained 472 IJs. In C. elegans, egg-laying behavior is modulated by environmental conditions, such as food availability (35), and may result in the preferential development of offspring to the environmentally resistant dauer stage (4). However, the total absence of egg laying is rare in C. elegans and has been the subject of mutant screens to identify genes involved in behavior and tissue development (23). In contrast, H. bacteriophora nematodes always reproduced by endotokia matricida even after an initial period of laying by maternal P0 nemotodes, despite high food (symbiont) availability and low worm densities. Therefore, egg laying might be an elective behavior in H. bacteriophora or may involve different sensitivities to environmental cues, such as dauer pheromone.

It is striking that the majority of offspring that develop via endotokia matricida develop into IJs (analogous to the dauer stage) in H. bacteriophora. In C. elegans, a transforming growth factor β pathway (Daf-1, -3, -4, -7, -8, and -14) regulates both egg laying and dauer formation, and mutants defective in this egg-laying pathway develop into dauer larvae (35). Egg laying and IJ formation might be regulated in H. bacteriophora by a homologous pathway or by other stress response pathways (e.g., insulin- and mitogen-activated protein kinase pathways). The host biology related to symbiont transmission will likely become amenable to genetic analysis soon, as RNA interference in H. bacteriophora has recently been developed (9), and an H. bacteriophora genome project at the National Human Genome Research Institute is currently in progress (R. Wilson, personal communication).

Supplementary Material

[Supplemental material]


We acknowledge Alicia Pastor at the Center for Advanced Microscopy at Michigan State University for her exceptional patience and expertise in assisting with transmission electron microscopy. We acknowledge John Breznak, David Brian Butvill, Elissa Hallem, and Paul Sternberg for advice and insightful comments on the manuscript.

This work was supported in part by the Center for Microbial Pathogenesis at Michigan State University and by grant NIH RR 12596 to D.H.H.


[down-pointing small open triangle]Published ahead of print on 15 February 2008.

Supplemental material for this article may be found at http://aem.asm.org/.


1. Albert, P. S., and D. L. Riddle. 1983. Developmental alterations in sensory neuroanatomy of the Caenorhabditis elegans dauer larva. J. Comp. Neurol. 219:461-481. [PubMed]
2. Bao, Y., D. P. Lies, H. Fu, and G. P. Roberts. 1991. An improved Tn7-based system for the single-copy insertion of cloned genes into chromosomes of Gram-negative bacteria. Gene 109:167-168. [PubMed]
3. Bertani, G. 1951. Studies on lysogenesis. I. The mode of phage liberation by lysogenic Escherichia coli. J. Bacteriol. 62:293-300. [PMC free article] [PubMed]
4. Chen, J., and E. P. Caswell-Chen. 2003. Why Caenorhabditis elegans adults sacrifice their bodies to progeny. Nematology 5:641-645.
5. Ciche, T. A., S. B. Bintrim, A. R. Horswill, and J. C. Ensign. 2001. A phosphopantetheinyl transferase homolog is essential for Photorhabdus luminescens to support growth and reproduction of the entomopathogenic nematode Heterorhabditis bacteriophora. J. Bacteriol. 183:3117-3126. [PMC free article] [PubMed]
6. Ciche, T. A., C. Darby, R.-U. Ehlers, S. Forst, and H. Goodrich-Blair. 2006. Dangerous liaisons: the symbiosis of entomopathogenic nematodes and bacteria. Biol. Control 38:22-46.
7. Ciche, T. A., and J. C. Ensign. 2003. For the insect pathogen Photorhabdus luminescens, which end of a nematode is out? Appl. Environ. Microbiol. 69:1890-1897. [PMC free article] [PubMed]
8. Ciche, T. A., and S. K. Goffredi. 2007. General methods to investigate microbial symbioses, p. 394-419. In C. A. Reddy, T. J. Beveridge, J. A. Breznak, G. M. Arzluf, T. M. Schmidt, and L. R. Synder (ed.), Methods for general and molecular microbiology, 3rd ed. ASM Press, Washington, DC.
9. Ciche, T. A., and P. W. Sternberg. 2007. Postembryonic RNAi in Heterorhabditis bacteriophora: a nematode insect parasite and host for insect pathogenic symbionts. BMC Dev. Biol. 7:101. [PMC free article] [PubMed]
10. Cotter, P. A., and S. Stibitz. 2007. c-di-GMP-mediated regulation of virulence and biofilm formation. Curr. Opin. Microbiol. 10:17-23. [PubMed]
11. Endo, B. Y., and W. R. Nickle. 1991. Ultrastructure of the intestinal epithelium lumen and associated bacteria in Heterorhabditis bacteriophora. J. Helminthol. Soc. Wash. 58:202-212.
12. Gage, D. J. 2004. Infection and invasion of roots by symbiotic, nitrogen-fixing rhizobia during nodulation of temperate legumes. Microbiol. Mol. Biol. Rev. 68:280-300. [PMC free article] [PubMed]
13. Gerrard, J. G., S. A. Joyce, D. J. Clarke, R. H. ffrench-Constant, G. R. Nimmo, D. F. Looke, E. J. Feil, L. Pearce, and N. R. Waterfield. 2006. Nematode symbiont for Photorhabdus asymbiotica. Emerg. Infect. Dis. 12:1562-1564. [PMC free article] [PubMed]
14. Gerritsen, L. J. M., and P. H. Smits. 1997. The influence of Photorhabdus luminescens strains and form variants on the reproduction and bacterial retention of Heterorhabditis megidis. Fund. Appl. Nematol. 20:317-322.
15. Goodrich-Blair, H. 2007. They've got a ticket to ride: Xenorhabdus nematophila-Steinernema carpocapsae symbiosis. Curr. Opin. Microbiol. 10:225-230. [PubMed]
16. Goodson, M. S., M. Kojadinovic, J. V. Troll, T. E. Scheetz, T. L. Casavant, M. B. Soares, and M. J. McFall-Ngai. 2005. Identifying components of the NF-κB pathway in the beneficial Euprymna scolopes-Vibrio fischeri light organ symbiosis. Appl. Environ. Microbiol. 71:6934-6946. [PMC free article] [PubMed]
17. Gravato-Nobre, M. J., and J. Hodgkin. 2005. Caenorhabditis elegans as a model for innate immunity to pathogens. Cell. Microbiol. 7:741-751. [PubMed]
18. Hall, D. H. 1995. Electron microscopy and three-dimensional image reconstruction. Methods Cell Biol. 48:395-436. [PubMed]
19. Hallem, E. A., M. Rengarajan, T. A. Ciche, and P. W. Sternberg. 2007. Nematodes, bacteria, and flies: a tripartite model for nematode parasitism. Curr. Biol. 17:898-904. [PubMed]
20. Han, R. C., and R. U. Ehlers. 1998. Cultivation of axenic Heterorhabditis spp. dauer juveniles and their response to non-specific Photorhabdus luminescens food signals. Nematologica 44:425-435.
21. Han, R. C., and R. U. Ehlers. 2000. Pathogenicity, development, and reproduction of Heterorhabditis bacteriophora and Steinernema carpocapsae under axenic in vivo conditions. J. Invert. Pathol. 75:55-58. [PubMed]
22. Heermann, R., and T. M. Fuchs. 2008. Comparative analysis of the Photorhabdus luminescens and the Yersinia enterocolitica genomes: uncovering candidate genes involved in insect pathogenicity. BMC Genomics 9:40. [PMC free article] [PubMed]
23. Horvitz, H. R., and J. E. Sulston. 1980. Isolation and genetic characterization of cell-lineage mutants of the nematode Caenorhabditis elegans. Genetics 96:435-454. [PMC free article] [PubMed]
24. Johnigk, S.-A., and R.-U. Ehlers. 1999. Endotokia matricida in hermaphrodites of Heterorhabditis spp. and the effect of the food supply. Nematology 1:717-726.
25. Johnigk, S.-A., and R.-U. Ehlers. 1999. Juvenile development and life cycle of Heterorhabditis bacteriophora and H. indica (Nematoda: Heterorhabditidae). Nematology 1:251-260.
26. Koropatnick, T. A., J. T. Engle, M. A. Apicella, E. V. Stabb, W. E. Goldman, and M. J. McFall-Ngai. 2004. Microbial factor-mediated development in a host-bacterial mutualism. Science 306:1186-1188. [PubMed]
27. Lambertsen, L., C. Sternberg, and S. Molin. 2004. Mini-Tn7 transposons for site-specific tagging of bacteria with fluorescent proteins. Environ. Microbiol. 6:726-732. [PubMed]
28. Lemaitre, B., and J. Hoffmann. 2007. The host defense of Drosophila melanogaster. Annu. Rev. Immunol. 25:697-743. [PubMed]
29. Merrell, D. S., and S. Falkow. 2004. Frontal and stealth attack strategies in microbial pathogenesis. Nature 430:250-256. [PubMed]
30. Moran, N. A. 1996. Accelerated evolution and Muller's rachet in endosymbiotic bacteria. Proc. Natl. Acad. Sci. USA 93:2873-2878. [PMC free article] [PubMed]
31. Nugent, M. J., S. A. O'Leary, and A. M. Burnell. 1996. Optimised procedures for the cryopreservation of different species of Heterorhabditis. Fund. Appl. Nematol. 19:1-6.
32. Nyholm, S. V., and M. J. McFall-Ngai. 2004. The winnowing: establishing the squid-vibrio symbiosis. Nat. Rev. Microbiol. 2:632-642. [PubMed]
33. Richardson, W. H., T. M. Schmidt, and K. H. Nealson. 1988. Identification of an anthraquinone pigment and a hydroxystilbene antibiotic from Xenorhabdus luminescens. Appl. Environ. Microbiol. 54:1602-1605. [PMC free article] [PubMed]
34. Ruby, E., B. Henderson, and M. McFall-Ngai. 2004. Microbiology. We get by with a little help from our (little) friends. Science 303:1305-1307. [PubMed]
35. Schafer, W. R. 14 December 2005, posting date. Egg-laying. In The C. elegans Research Community (ed.), WormBook. http://www.wormbook.org/chapters/www_egglaying/egglaying.html.
36. Sifri, C. D., J. Begun, and F. M. Ausubel. 2005. The worm has turned—microbial virulence modeled in Caenorhabditis elegans. Trends Microbiol. 13:119-127. [PubMed]
37. Tamayo, R., J. T. Pratt, and A. Camilli. 2007. Roles of cyclic diguanylate in the regulation of bacterial pathogenesis. Annu. Rev. Microbiol. 61:131-148. [PMC free article] [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...