For the Insect Pathogen Photorhabdus luminescens, Which End of a Nematode Is Out?

Transcription

1 APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Apr. 2003, p Vol. 69, No /03/$ DOI: /AEM Copyright 2003, American Society for Microbiology. All Rights Reserved. For the Insect Pathogen Photorhabdus luminescens, Which End of a Nematode Is Out? Todd A. Ciche and Jerald C. Ensign* Department of Bacteriology, University of Wisconsin Madison, Madison, Wisconsin Received 25 July 2002/Accepted 17 December 2002 The nematode Heterorhabditis bacteriophora is the vector for transmitting the entomopathogenic bacterium Photorhabdus luminescens between insect larvae. The dauer juvenile (DJ) stage nematode selectively retains P. luminescens in its intestine until it releases the bacteria into the hemocoel of an insect host. We report the results of studying the transmission of the bacteria by its nematode vector. Cells of P. luminescens labeled with green fluorescent protein preferentially colonized a region of the DJ intestine immediately behind the basal bulb, extending for various distances toward the anus. Incubation of DJ nematodes in vitro in insect hemolymph induced regurgitation of the bacteria. Following a 30-min lag, the bacteria migrated in a gradual and staggered movement toward and ultimately exited the mouth. This regurgitation reaction was induced by a low-molecular-weight, heat- and protease-stable, anionic component present in arthropod hemolymph and in supernatants from insect cell cultures. Nematodes anesthetized with levamisole or treated with the antihelmenthic agent ivermectin did not release their bacteria into hemolymph. The ability to visualize P. luminescens in the DJ nematode intestine provides the first clues to the mechanism of release of the bacteria during infection of insect larvae. This and the partial characterization of a component of hemolymph triggering release of the bacteria render this fascinating example of both a mutualistic symbiosis and disease transmission amenable to future genetic and molecular study. The mutualistic association of the enteric entomopathogenic bacterium Photorhabdus luminescens and its nematode vector Heterorhabditis bacteriophora is a potentially useful model for studies of vector-borne disease and use as biological control agents for arthropod vectors of disease and agricultural insect pests. The two partners, acting in concert, are voracious pathogens to a wide variety of insect larvae (24, 28; for a recent review, see reference 15). The dauer juvenile (DJ) stage of the nematode carries in its intestine a monoculture of symbiotic P. luminescens (13, 20, 21, 25). The nematode locates a susceptible insect victim, enters the hemocoel, and releases its charge of bacteria, and death of the insect follows, usually within 24 h (15, 24, 28). Insect mortality is primarily attributable to the extreme virulence of P. luminescens, where the 50% lethal dose for some insects is less than 30 cells when injected into the hemocoel (24, 25, 29). The bacteria produce potent orally active insecticidal protein toxins that are being developed as biological control agents (5, 6, 18). During growth inside the insect cadaver, the DJ develops into a reproductively mature adult hermaphrodite. The nematode offspring feed on the P. luminescens cells and are highly specific in requiring their specific strain of P. luminescens for growth and reproduction (1, 12, 16). Following several rounds of reproduction, lasting approximately 2 weeks and in apparent response to impending nutrient limitation, several thousand DJ stage nematodes, which are selectively colonized by P. * Corresponding author. Mailing address: Department of Bacteriology, University of Wisconsin, Madison, WI Phone: (608) Fax: (608) jcensign@facstaff.wisc.edu. Present address: Hopkins Marine Station of Stanford University, Pacific Grove, CA luminescens (20, 21), disperse from the cadaver in search of another insect victim. Little is known concerning the mechanisms for specific colonization of the DJ intestine by P. luminescens, and its subsequent release during insect infection. A possible clue is the presence of three fimbrial homologs located 54 bp 5 of ngra, a gene required by the bacteria to support the growth and reproduction of H. bacteriophora nematodes (9). Nematodes grown on an equal mixture of the wild type and the ngra mutant retain only the wild-type bacteria. Whether these fimbrial genes located close to ngra directly affect retention of the bacteria in DJ nematodes is not yet known. Vivas and Goodrich- Blair (36) recently reported that a stationary-phase sigma factor homolog, rpos, is required for Xenorhabdus nematophilus, a bacterium symbiotically associated with the nematode Steinernema carpocapsae, to colonize a specialized intestinal vesicle of the DJ stage nematodes. In earlier work, Poinar and colleagues (27, 29) isolated symbiotic bacteria from surface-sterilized DJ nematodes by adding them to a hanging drop of hemolymph. They also reported phase-contrast microscope observations indicating that the bacteria were localized in the ventricular portion of H. bacteriophora DJ intestines and that the nematodes, upon ingesting hemolymph, passed the bacteria through the intestine, excreting them from the anus (29). To study nematode transmission of P. luminescens to insect hosts, we labeled the bacterium by transposing a green fluorescent protein (GFP) gene, contained in a mini-tn5 transposon, into the bacterium s DNA (9). Here we report the results of studies of the colonization and persistence of the GFPlabeled cells in the DJ intestine, behavior of the bacterial cells during release from the DJ nematodes when incubated in insect hemolymph, and properties of factors involved in eliciting release of the bacteria. 1890

2 VOL. 69, 2003 REGURGITATION OF P. LUMINESCENS 1891 TABLE 1. Strains and plasmids used in this study. Strain or plasmid Relevant characteristic(s) Source or reference Nematode strain H. bacteriophora NC1 Host for P. luminescens NP1/1 24 Bacterial strains P. luminescens NP1/1 Primary phase; isolated from H. bacteriophora (identical to ATCC , strain NC-19) NC1/2 Secondary phase; isolated from NP1/1 9 NC1/1GFP NC1/1::mini-Tn5rsGFP This study Meg/1 Primary phase; isolated from H. megidis 30 E. coli DH5 Cloning strain Gibco BRL Plasmid vectors psu19 Source of p15a ori; Cm r for mini-tn5rsgfp 2 pbc SK( ) Cloning vector Stratagene pub193 Mini-Tn5 delivery vector; Cm r Sm r Sp r Suc s This study pqb163 Source of rsgfp Q-Biogene pbc SK( )GFP pbc SK containing XbaI-BamHI fragment of rsgfp from pqb163 This study pub193ssgfp Mini-Tn5rsGFP delivery vector; SalI-SacI fragment of GFP63 inserted into pub193 This study MATERIALS AND METHODS Microbial methods. The strains and plasmids used in this study are described in Table 1. Microbial media were obtained from Difco (Detroit, Mich.) and chemicals were from Sigma (St. Louis, Mo.), unless otherwise indicated. P. luminescens was grown at 29 C in the dark in 2% (wt/vol) Proteose Peptone 3 (PP3), with 1.5% agar, 7.5% (wt/vol) sucrose, 20 g of streptomycin per ml, 20 g of spectinomycin per ml, or 4 g of chloramphenicol per ml added when required. Escherichia coli was grown at 37 in Luria-Bertani medium, with 1.5% agar, 20 g of streptomycin per ml, 20 g of spectinomycin per ml, or 20 or 120 g of chloramphenicol per ml added when required. GFP labeling of bacteria. P. luminescens was labeled with GFP by transposon mutagenesis of a mini-tn5 containing a constitutively expressed GFP gene randomly incorporated into its DNA. The mini-tn5rsgfp was constructed by cloning a XbaI-BamHI fragment containing an enhanced red-shifted GFP (rsgfp) gene and ribosome binding site from pqbi63 (Quantum Biotechnologies Inc.) into an XbaI-BamHI-digested pbc SK( ) vector (Stratagene). A SalI-SacI fragment was cloned into a SalI-SacI-digested pub193 vector. Plasmid pub193 is identical to pub394 except that a psu19 plasmid (2), instead of psu39 plasmid, was inserted between the inverted repeats of the mini-tn5 (9). The resulting plasmid, pub193rsgfp, was transformed into P. luminescens, and the mini- Tn5rsGFP was inserted into the chromosome as described previously (9). These bacteria displayed bright and stable fluorescence, even in the absence of antibiotic selection for the transposon. Observation and counting of bacteria in nematodes. Nematodes were propagated on lipid agar seeded with wild-type, Meg/1, or GFP-labeled bacteria as described previously (9). Nematodes were observed for GFP-labeled P. luminescens by using an epifluorescence microscope that was equipped with a standard fluorescein isothiocyanate filter set (Nikon). The numbers of bacteria retained in the DJ nematode intestines were determined by plating dilutions of surfacesterilized and homogenized nematodes as described previously (9) or by counting the GFP-labeled cells by epifluorescence microscopy. Release of bacteria from nematodes. The release of P. luminescens from DJ nematodes was observed by mixing a 0.01-ml sample containing more than 30 DJ nematodes retaining GFP-labeled P. luminescens with an equal volume of test material (see below) on a microscope slide. The mixture was covered with a glass slide, and the edges were sealed with petroleum jelly. The nematodes were observed with epifluorescence microscopy for the presence of bacteria and their migration through the pharynx and exiting the mouth. Insect hemolymph was obtained by bleeding third- or fourth-instar larvae of Manduca sexta by making a lesion in an anterior forelimb. The hemolymph was collected on ice, 0.1% dithiothreitol was added to inhibit melanization of hemolymph, and the hemolymph was then stored at 20 C. The rate of release from individual nematodes was determined by counting the numbers of fluorescent bacteria exiting individual nematodes. The rate of release of GFP-labeled cells by a population of nematodes was determined by incubating 750 DJ nematodes in 1.0 ml of M. sexta hemolymph or in Grace s insect tissue culture medium (Sigma). At various times during incubation at 30 C, 0.l-ml samples were removed and centrifuged at 1,400 g for 1 min to sediment nematodes, serial dilutions of the supernatant were plated on 2.0% PP3, and colony counts were made following 72 h of incubation. We attempted to determine if nematode muscle contraction, pharyngeal pumping, or bacterial viability is required for release of the bacteria. Hemolymph obtained from M. sexta was amended with levamisole (0.5 mm) or ivermectin (100 g/ml, with 1% dimethyl sulfoxide [DMSO]), which paralyzes nematodes or prevents pharyngeal pumping, respectively, or with kanamycin (30 g/ml), which is lethal for P. luminescens. Release of bacteria was visualized by using epifluorescence microscopy. Characterization of release activity. To further characterize factors involved in release of bacteria, nematodes containing GFP-labeled bacteria were incubated in M. sexta hemolymph that was modified by either heating at 80 C for 15 min, centrifugation through a 10-kDa-cutoff Microcon microconcentrator (Millipore Corp., Bedford, Mass.), addition of EDTA at 25 mm (Sigma), treatment with pronase (20 g/ml) at 50 C for 2 h, treatment with Chelex resin (10 mg/ml) (Bio-Rad Laboratories, Hercules, Calif.), dialysis for 24 h at 4 C with 5- or 10-kDa-cutoff cellulose ester membranes (Spectrum Laboratories, Inc., Rancho Dominquez, Calif.), or melanization at room temperature for 3 h. Hemolymph from diverse arthropods was tested for the ability to induce DJ release of bacteria. Freshwater crabs, feeder crickets (Archeta domestica), and tropical cockroaches (Blaberus sp.) were obtained from a local pet store (Pet World, Madison, Wis.), and the hemolymph was collected from a lesion at the base of a posterior limb of anesthetized animals. Tenebrio molitor and Galleria mellonella larvae were obtained from Ja-Da Bait Co. (Antigo, Wis.), and hemolymph was collected as described above for M. sexta. Ants (Lasius alienus [Foerster], provided by H. Goodrich-Blair, University of Wisconsin) were anesthetized and compressed by pressure in a 1-ml syringe, resulting in release of a few drops of hemolymph. The following supernatants of insect cell culture fluids were also tested: TN 368 (22) and TNHI5 (10, 17) from Trichoplusia ni, SF21 (35) from Spodoptera frugiperda grown in TC 100 medium containing 10% heat-inactivated fetal bovine serum (FBS), DL-1 (31) from Drosophila melanogaster grown in Schneider s medium plus 10% FBS (spent tissue culture medium provided by Paul Friesen, Department of Biochemistry, University of Wisconsin-Madison), and 4a-7B (26), a hemocyte-like cell line from Anopheles gambiae grown in Schneider s medium plus 10% FBS (provided by L. Moita and F. C. Kafatos, European Molecular Biology Laboratory, Heidelberg, Germany). Also tested were sterile Grace s insect cell culture medium, Schneider s medium with 10% FBS, distilled water, 2% PP3 (Difco), culture supernatants of P. luminescens NC1 grown for 48 h at 29 C in 2% PP3, and fresh human blood donated by T.A.C. The incubations were for 60 min at 30 C, after which the numbers of released bacteria were determined by epifluorescence microscopy. Release of bacteria was also studied by using nematodes grown on wild-type P. luminescens. In these experiments, 35 to 50 nematodes were incubated for 60 min in the test media, after which 0.01-ml samples were plated onto nutrient agar (Difco) and the colonies arising from released bacteria were counted after 72 h of incubation.

3 1892 CICHE AND ENSIGN APPL. ENVIRON. MICROBIOL. FIG. 1. Location of GFP-labeled P. luminescens cells in the intestines of H. bacteriophora DJ nematodes. (A) Differential interference contrast micrograph showing cells (arrow) located anterior to the nematode basal bulb. (B) Epifluorescent micrograph of same nematode as in panel A, showing GFP-labeled P. luminescens cells present throughout the DJ nematode lower intestine. (C) Epifluorescent micrograph taken with higher magnification, showing bacteria in the DJ intestine. (D and E) Epifluorescent micrographs of several nematodes freshly harvested (D) and after 30 days of incubation in saline (E). a, anterior orientation of nematode; b, posterior orientation of nematode. RESULTS Colonization of DJ nematodes by bacteria. Differential interference contrast microscopy of DJ nematodes (Fig. 1A) shows the esophagus, basal bulb, and location of bacterial cells in the intestine (the intestinal bacteria are usually obscured by the nematode body). However, epifluorescence microscopy of the DJ nematodes containing GFP-labeled bacteria clearly shows the bacteria to be located in the anterior region of the intestine, immediately posterior to the basal bulb and extending throughout the intestine (Fig. 1B [same nematode as shown in Fig. 1A]), and under higher magnification individual bacteria were easily distinguished (Fig. 1C). DJ nematodes retained variable numbers of bacteria in their intestine, with many nematodes retaining no GFP-labeled bacteria. Epifluorescence microscopic examination revealed that 26% (standard deviation [SD], 10%; n 3) of the DJ nematodes grown on GFP-labeled P. luminescens retained bacteria; the others contained no bacteria. Viable plate counts of bacteria released from individual DJ nematodes by mechanical disruption confirmed that the numbers of GFP-bacterial cells per nematode varied from 0 to 300, with the majority of the nematodes containing none. The counts obtained by epifluorescence microscopy were as follows (mean SD): 74% 10% of the DJ nematodes contained no visible bacteria, 3% 2% of the DJ nematodes contained 1 to 25 cells, 7% 3% of the DJ nematodes contained 26 to 50 cells, 5% 4% of the DJ nematodes contained 51 to 100 cells, 3% 2% of the DJ nematodes contained 101 to 150 cells, 3% 2% of the DJ nematodes contained 151 to 200 cells, and 3% 4% of the DJ nematodes contained more than 200 cells. The numbers of bacteria released from disrupted DJ nematodes propagated on wild-type bacteria also varied from 0 to 300, but only 5% did not retain any viable bacteria (data not shown). The numbers of bacteria retained by GFP-propagated nematodes that visibly retained the bacteria were approximately the same (mean of 29 cells/dj nematode retaining bacteria; SD, 16; n 2) as found in nematodes grown on wild-type bacteria (46 cells/dj nematode; SD, 9; n 4). In addition to a fourfold reduction in the numbers of DJ nematodes that retained bacteria, GFP-labeled bacteria were unable to compete with wild-type bacteria for colonization of the DJ nematodes. DJ nematodes propagated on an equal mixture of GFP-labeled and unlabeled P. luminescens retained only the unlabeled bacteria (data not shown). Three independently generated GFP-labeled mutants displayed these same colonization characteristics. Reinfection of DJ nematodes with colonization-competent GFP-labeled mutants, cultivated from DJ intestines, did not increase the infection rate in DJ nematodes (data not shown). The virulence of DJ nematodes containing GFP-labeled P. luminescens for M. sexta larvae was the same as that for the wild-type cells, and the resulting DJ nematodes, which emerged from the insect cadaver after several generations of growth on the bacteria, retained GFP-labeled bacteria to the same extent as agar-propagated nematodes. Numbers and locations of bacteria in freshly harvested and aged DJ nematodes. The DJ nematodes incubated in saline at 25 C for 30 days contained 80 cells/dj nematode (SD, 23; n 3), which is not significantly different from the number contained in freshly harvested DJ nematodes (mean, 29 cells/dj nematode retaining bacteria; SD, 16; n 2). The location of bacteria in the intestines of aged DJ nematodes (Fig. 1E) also resembled that of freshly harvested DJ nematodes (Fig. 1D).

4 VOL. 69, 2003 REGURGITATION OF P. LUMINESCENS 1893 FIG. 2. Association of GFP-labeled P. luminescens with non-dj nematodes. (A, C, and D) Digestion of GFP-labeled bacteria by adult nematodes indicated by a bright and diffuse fluorescence in the intestine (i). Intact bacteria (arrows) are found to be associated with the vulva (v) opening (A and C) and rectum (r) (C and D). Eggs (e) are indicated in panel C. (B) Four nematodes undergoing endotokia matricida containing variable amounts of bacteria in their body cavities (arrows). Two nematodes are brightly fluorescent due to large numbers of bacteria in their body cavities ( 1,000 bacteria/nematode). Two other nematodes are weakly fluorescent due to a small population of bacteria ( 200 bacteria/nematode). Bacteria are also present in the remnants of the pharynx (p) and rectum (r). A DJ nematode outside the adult nematodes is also indicated. The DJ nematodes containing few ( 25) bacteria were equally abundant in those DJ nematodes incubated for 30 days and in those that were freshly harvested (data not shown). Thus, it seems that the bacteria have a limited capacity to multiply or spread throughout the intestine during incubation and aging of the DJ nematodes. In some 30-day-old or deceased nematodes, a slight swelling of the nematode intestine could be seen (Fig. 1E), and P. luminescens was present in the entire body cavity of many deceased DJ nematodes (data not shown), suggesting that the processes involved in localization of the bacteria in healthy nematodes do not persist into death. Association with non-dj stage nematodes. Nematode stages other than the DJ stage actively digest P. luminescens cells as a major food source. A diffuse green fluorescence was observed in the intestines of adult nematodes feeding on the GFPlabeled cells resulting from bacterial cell lysis (Fig. 2A). Few intact bacteria were observed in the intestine, but they were sometimes observed at the surface of the vulva opening (Fig. 2A and C) and in the rectums of adult nematodes (Fig. 2C and D). Conversely, the DJ nematodes appeared not to digest their intestinal bacteria, since fluorescence was limited to intact cells in the DJ intestine and the numbers of bacteria did not decrease over at least a 1-month incubation period. Many DJ nematodes develop inside the body cavity of adult female nematodes, a phenomenon termed endotokia matricida (23). We observed a large variation in numbers of GFP-labeled P. luminescens cells in the body cavities of endotokia matricida nematodes (Fig. 2B). Two of the hermaphroditic nematodes were brightly fluorescent, being replete with bacteria and having approximately 100 bacteria per DJ nematode. One nematode contained 10 bacteria and another contained just 1 bacterium per DJ nematode. It is plausible that DJ nematodes developing in hermaphrodites with few bacteria will be less efficiently colonized with bacteria than DJ nematodes developing in bodies with copious amounts of the bacteria. Release of bacteria from nematodes. The DJ nematodes encountering an insect larva penetrate the insect cuticle or gut by using a buccal tooth and enter the hemocoel (3). A second outer cuticle, retained from the J2 developmental stage, maintains the DJ nematodes in a nonfeeding and environmentally resistant state when they are between insect victims and is involved in their survival in soil, in water, and on plant surfaces for long periods of time (7). The cuticle was shed almost immediately after immersion in hemolymph, accompanied by increased movement of the nematodes, and shortly thereafter P. luminescens cells began to migrate toward the mouth. The cuticle was also shed during surface sterilization of the DJ nematodes with 0.05% sodium hypochlorite for 5 min (7), but the intestinal bacteria were not released or damaged (data not shown). These nematodes did not lose insect virulence during incubation in saline for several weeks. No difference was observed in the timing or rate of release of bacteria from the normal and exsheathed DJ nematodes when they were incubated in hemolymph (data not shown). Differential interference contrast (Fig. 3A) and epifluorescence (Fig. 3B) microscopy of the same nematode after 60 min of incubation in hemolymph show GFP-labeled cells outside the nematode and in the hemolymph. The fluorescent micrographs clearly reveal that the bacteria migrated from the intestine into the pharynx, toward and ultimately exiting the nematode mouth, a process suggestive of regurgitation (Fig. 3B). During the initial 20-min incubation, the bacteria remained clustered in the anterior region of the intestine, with no bacteria visible in the pharynx (Fig. 3C). After 30 min of incubation, individual bacterial cells began to migrate from the intestine into the pharynx (Fig. 3D), and at 40 min (Fig. 3E) and 60 min (Fig. 3F), they were observed to migrate progressively through the pharynx toward the nematode mouth. Higher-magnification micrographs of different DJ nematodes releasing bacteria at 90 min show more detail of the bacteria leaving the anterior region of the intestine (Fig. 3G) and poised to exit the mouth (Fig. 3H). The bacteria moved in a punctuated and gradual manner, with pulses of movement separated by pauses of 5 to 20 min (see a time-lapsed movie of the release process at taciche). No movement of bacteria toward the posterior region of the nematode intestine or out the anus was observed. The pulsatile migration of the bacteria through the pharynx and toward the mouth and the observation that the bacteria were nonmotile upon release indicate that the bacteria were not using flagellar motility to exit the nematode. Rate of bacterial release from DJ nematodes. During incubation in hemolymph, the DJ nematodes began to release their intestinal bacteria after a 30-min lag, where no movement of bacteria was seen, and continued to release the bacteria at a gradual rate for more than 300 min thereafter (Fig. 4). During the 30-min lag period, nematode movement decreased and rapid pumping of a vesicle inside of the excretory pore was observed. Following 30 min of incubation, the bacteria began

5 1894 CICHE AND ENSIGN APPL. ENVIRON. MICROBIOL. FIG. 3. Release of GFP-labeled P. luminescens by DJ nematodes incubated in M. sexta hemolymph. In all panels the anteriors of the nematodes are orientated to the left of the field of view. (A) Phase-contrast micrograph showing bacterial cells that were released from a DJ nematode. (B) Epifluorescent micrograph of the same nematode clearly shows bacteria in the intestine and pharynx. (C) A DJ nematode incubated for 20 min, prior to releasing GFP-labeled P. luminescens. (D, E, and F) Nematodes incubated for 30, 40, and 60 min, respectively. Note progressive movement of bacteria toward the nematode mouth. (G and H) A nematode observed under higher magnification, showing individual bacterial cells leaving a cluster of bacteria in the posterior region of the intestine after 90 min of incubation (G) and showing bacteria poised to exit the mouth after 90 min of incubation (H). to be released in hemolymph but not in Grace s insect cell culture medium. The average rate of release was one bacterial cell every 2 min for 90 min, followed by a lower rate of release. Characterization of the release activating factor(s). We attempted to identify the components of hemolymph that initiate the release of bacteria. The bacterial release factor was present in the hemolymph or insect cell culture supernatants obtained from at least eight orders of the phylum Arthropoda (Table 2). Approximately 30 to 50% of the nematodes were observed to FIG. 4. Numbers of GFP-labeled P. luminescens organisms released by DJ nematodes incubated in M. sexta hemolymph or Grace s insect cell culture medium. CFU counts were made, and the numbers of bacteria released were normalized to those DJ nematodes that epifluorescence microscopy showed to be actively releasing bacteria at each experimental point. Error bars indicate standard deviations. emit the bacteria during the 1-h incubation in hemolymph of M. sexta, D. melanogaster, T. molitor, G. mellonella, crab, ant, tropical cockroach, and cricket. Supernatant fluids obtained following growth of several insect cell cultures induced release in between 17 and 31% of the nematodes. Significantly fewer nematodes released their bacteria when incubated in uninoculated Grace s or Schneider s tissue culture medium (with 10% FBS), fresh human blood, water, or P. luminescens culture supernatants. The release-inducing activity was not affected by heat, pronase digestion, Chelex treatment, EDTA addition, or melanization; was lost following dialysis; and was present in the filtrate but not the retentate of a 10-kDa-cutoff Centricon membrane. The mechanism of release does not appear to be intrinsic to the bacteria but depends on the nematode activity. Nematodes anesthetized with levamisole or treated with ivermectin, a compound that inhibits pharyngeal pumping, regurgitated few bacteria (Table 2). Although bacteria were observed in the pharyngeal region in 12 to 16% of the nematodes and were counted as positive for release, the pharyngeal bacteria were not released outside the nematodes. Furthermore, incubation of nematodes in hemolymph amended with the bactericide kanamycin did not affect release of bacteria. The release experiments were repeated and extended by counting the numbers of wild-type bacteria released from the DJ nematodes. These results verified that virtually no bacteria were released during 1 h of incubation in buffer, nutrient broth, growth liquor of P. luminescens, or Grace s medium but were released in untreated, heated, and melanized M. sexta hemolymph (Table 3). The release activity was removed from he-

6 VOL. 69, 2003 REGURGITATION OF P. LUMINESCENS 1895 TABLE 2. Induction of DJ nematode release of GFP-labeled P. luminescens Sample molymph by dialysis and when passed through a 10-kDa-cutoff Centricon membrane, was not removed by passage through a Sepharose Q column, but was completely removed by the cation exchanger Sepharose S. Cell-free fluids obtained from a culture of D. melanogaster DL-1 cells effected release of bacteria, and this activity was also removed by Sepharose S treatment. DISCUSSION The labeling of P. luminescens with GFP allowed the bacteria to be visualized in living nematodes. This facilitated the observations that variable numbers of the bacteria preferentially colonized a region of the DJ intestine where they persisted until they were released into insect hemolymph. In contrast to a previous report that release is through the anus (29), our observations clearly showed the bacteria to be regurgitated as individual cells, exiting the mouth, in a gradual and pulsitile manner following exposure to hemolymph. It is remarkable that the nematodes growing inside an insect Treatment Release activity a Mean % SD (%) n M. sexta larva hemolymph None Incubated at 68 C for 15 min Pronase (20 g/ml, 50 C,2h) EDTA (25 mm) Chelex (10 mg/ml) Centricon 10-kDa filtrate Centricon 10-kDa retentate 10 1 Melanized Dialyzed (5 kda) 0 1 Dialyzed (10 kda) 0 1 Diluted 1/ Levamisole (0.5 mm) Ivermectin (100 g/ml, 1% DMSO) DMSO (1%, vol/vol) 33 1 Kanamycin (30 g/ml) Other arthropod hemolymph G. mellonella larvae T. molitor larvae D. melanogaster larvae 30 1 Freshwater crab A. domestica (feeder cricket) Blaberus sp. (tropical cockroach) L. alienus Foerster (Ant) Insect cell culture supernatants A. gambiae 4a-3B T. ni TN T. ni TNH S. frugiperda SF D. melonogaster DL Nonarthropod fluids Fresh human blood Culture of P. luminescens NC1/ Peptone broth Grace s insect cell culture medium Schneider s insect cell culture medium 10% FBS Water a Percentage of DJ nematodes containing GFP-labeled P. luminescens cells that released bacteria 60 min after incubation. The mean percentage, SD, and number of independent experiments (n) where at least 30 DJ nematodes containing GFP-labeled bacteria were observed are shown. cadaver, in the presence of a dense mixed culture of bacteria upon which they are apparently feeding indiscriminately, retain only their specific P. luminescens symbiont in the DJ stage (20, 21). With few exceptions, a species of Heterorhabditis will grow only on that strain of P. luminescens, or a closely related strain, from which it was originally isolated (20, 21). One notable exception is the ability of H. bacteriophora to grow on the P. luminescens symbiont of the H. megidis nematode. The resulting DJ nematodes, however, are axenic, retaining no bacteria (19). A surprising observation was that 5% of DJ stage nematodes propagated on a pure culture of wild-type bacteria and 74% propagated on GFP-labeled bacteria were axenic. The decreased colonization ability of GFP-labeled P. luminescens was not likely the result of insertion of the GFP gene into a region of the P. luminescens genome directly associated with adhesion to the nematode intestine, since the same result was obtained with three independent mini-tn5rsgfp-transposon insertions. Apparently some unknown consequence of the high level of GFP expression is responsible for the colonization

7 1896 CICHE AND ENSIGN APPL. ENVIRON. MICROBIOL. TABLE 3. Effects of different incubation conditions on DJ nematode release of unlabeled P. luminescens Incubation conditions Release activity a Mean CFU Buffer 5 7 2% nutrient broth 6 8 Growth liquor (72 h) 4 5 Grace s medium 4 5 M. sexta hemolymph No treatment Centrifuged, filter sterilized Incubated at 80 C for 30 min Melanized Dialyzed (10 kda) 18 9 Centricon 10-kDa retentate 68 5 Centricon 10-kDa filtrate Passed through anion-exchange column (Sepharose Q) Passed through cation-exchange column (Sepharose S) 11 8 D. melanogaster tissue culture fluid (DL-1) No treatment Passed through Sepharose S a Thirty monoxenic DJ nematodes were incubated for 60 min, and samples were plated on nutrient agar. Mean CFU and SD are shown. Experiments were performed twice. defect. It is significant that labeling of the cells with GFP did not affect the location of the bacteria in the nematode intestine (Fig. 1). Also, conditions effecting release of the bacteria were the same in nematodes harboring normal and GFP-labeled cells (Tables 2 and 3). The nature of the nematode and bacterial factors involved in transmission are unknown. An adhesin-receptor interaction might allow the bacteria to adhere to specific receptors in the anterior region of the DJ intestine. Putative fimbrial genes were identified 54 bp 5 of ngra, a gene required for P. luminescens to support nematode growth and reproduction (9), and might be involved in the selective retention of the bacteria in the nematode gut mucosa. In addition, the nematode innate defense reaction or reactive oxygen species produced by the nematode might select for the symbiotic bacteria while eliminating others. DJ nematodes develop from eggs deposited either outside or inside (endotokia matricida) the body cavities of adult hermaphroditic nematodes. It is likely that those DJ nematodes developing inside the maternal nematodes acquire their charge of P. luminescens that had infected the body cavity. The variable numbers of bacteria, which we observed to be present in the nematode bodies undergoing endotokia matricida, might partially explain the variability in the numbers of bacteria observed to colonize the DJ intestines. DJ nematodes developing in adult nematodes infected with few or many P. luminescens cells would correspondingly retain few or many of the bacteria. How the bacteria infect pregnant hermaphroditic nematodes is unknown; possible routes of infection are through the intestine by pharyngeal, rectal, or intestinal bacteria that penetrate the intestinal mucosa, perhaps after being ruptured by the developing DJ nematodes, or through the vulva by bacteria associated with the vulva opening. These observations are probably relevant to the complex and selective process of horizontal transmission of symbiotic bacteria by SD the nematodes, which plays an important role in efficiency of insect pathogenicity. It is interesting that the bacteria are regurgitated from the DJ nematodes by a pulsitile and gradual process. Perhaps there is some advantage for the slow continual release of a few bacteria in overwhelming the insect defense mechanisms. A rationale for such a slow release is not obvious, since the 50% lethal dose of P. luminescens when injected into larvae of several different insect genera is less than 30 cells (24, 25). Perhaps the nematodes and P. luminescens have chosen to use guerrilla tactics of stealth and ambush for attacking their insect victims rather than a strategy of mass attack. A possible rationale for the gradual release would be to allow the insect host s innate defense mechanisms to destroy potential competitive bacteria that might have entered during nematode penetration while P. luminescens evades the innate immune defenses and begin to reproduce. In the larvae of the greater wax moth, G. mellonella, H. bacteriophora nematodes evade the innate immune system while the P. luminescens cells are engulfed by hemocytes and remain hidden in fat bodies (11). During this time the innate immune system functions normally. After 5 h, the bacteria emerge along with damaged hemocytes and rapidly kill the insect host. Other microorganisms carried into the hemocoel might have been destroyed at that time, ensuring that the insect cadaver is initially devoid of saprophytic microorganisms, which might have had an adverse affect on nematode growth and subsequent colonization of their intestine. Recently, Silva et al. (32) reported that GFP-labeled P. luminescens subsp. akhustii evades phagocytosis and suppresses the innate immune response of M. sexta. The data presented here are the first published report of the process by which an entomopathogenic nematode regurgitates its symbiotic bacteria, a process critical for efficient pathogenesis of insect larvae. The GFP-labeled bacteria should prove to be a valuable tool for future studies of the selective colonization and persistence of P. luminescens in the DJ intestine, the mechanism of regurgitation of the bacteria into insect hemolymph, and subsequent virulence of P. luminescens. Our discovery that a low-molecular-weight, heat- and protease-resistant factor found only in hemolymph or insect culture fluids elicits regurgitation opens the way for identification of the factor and elucidation of the mechanism involved. It appears that the target of the factor(s) is the nematode and not the bacterial partner, since anesthetized nematodes did not regurgitate the bacteria. The nematode genera Heterorhabditis and Caenorhabditis share some important properties: both are bacteriovores, show hermaphroditic reproduction, and are phylogenetically related (4). The genetic tools developed for use in C. elegans (such as RNA interference [14]), information from the complete genome sequence (8), and insights about the innate immune system (33) are valuable resources for future studies of the entomopathogenic bacterium-nematode interactions. Our observation that the bacteria are released from nematodes during incubation in D. melanogaster hemolymph and cell culture supernatants opens the possibility for future studies using a genetically tractable insect host with a compete genome sequence available and which is a model for host-pathogen interactions with an innate immune system (34). Thus, powerful genetics tools are available that might soon be applied to aid our un-

8 VOL. 69, 2003 REGURGITATION OF P. LUMINESCENS 1897 derstanding of tripartite interactions between pathogen, vector, and host that are fundamental to vector-borne disease. ACKNOWLEDGMENTS We thank the members of the Heidi Goodrich-Blair laboratory for excellent comments, suggestions, and assistance, and we thank Paul Friesen and Luis Moita, who kindly provided us with the insect cell cultures. This research was funded by an S. C. Johnson Wax Distinguished Scientist Fellowship to T.A.C. and grants from the College of Agriculture and Life Sciences of the University of Wisconsin Madison to J.C.E. REFERENCES 1. Akhurst, R. J., R. G. Mourant, L. Baud, and N. E. Boemare Phenotypic and DNA relatedness study between nematode-symbiotic and clinical strains of the genus Photorhabditis (Enterobacteriaceae). Int. J. Syst. Bacteriol. 43: Bartolome, B., Y. Jubete, E. Martinez, and F. de la Cruz Construction and properties of a family of pacyc 184-derived cloning vectors compatible with pbr322 and its derivatives. Gene 102: Bedding, R. A., and A. S. Molyneux Penetration of insect cuticle by infective juveniles of Heterorhabditis spp. (Heterorhabditidae: Nematoda). Nematologica 28: Blaxter, M. L., P. De Ley, J. R. Garey, L. X. Liu, P. Scheldeman, A. Vierstraete, J. R. Vanfleteren, L. Y. Mackey, M. Dorris, L. M. Frisse, J. T. Vida, and W. K. Thomas A molecular evolutionary framework for the phylum Nematoda. Nature 392: Bowen, D. J., and J. C. Ensign Purification and characterization of a high-molecular-weight insecticidal protein complex produced by the entomopathogenic bacterium Photorhabdus luminescens. Appl. Environ. Microbiol. 64: Bowen, D., T. A. Rocheleau, M. Blackburn, O. Andreev, E. Golubeva, R. Bhartia, and R. H. French-Constant Insecticidal toxins from the bacterium Photorhabdus luminescens. Science 280: Cambell, L. R., and R. Gaugler Role of the sheath in desiccation tolerance of two entomopathogenic nematodes. Nematologica 37: C. elegans Sequencing Consortium Genome sequence of the nematode C. elegans: a platform for investigating biology. Science 282: Ciche, T. A., S. B. Bintrim, A. H. Horswill, and J. C. Ensign A phosphopantetheinyl transferase homolog is essential for Photorhabdus luminescens to support growth and reproduction of the entomopathogenic nematode Heterorhabditis bacteriophora. J. Bacteriol. 183: Davis, T. R., K. M. Trotter, R. R. Granados, and H. A. Wood Baculovirus expression of alkaline phosphatase as a reporter gene for evaluation of production, glycosylation and secretion. Bio/Technology 10: Dunphy, G. B., and J. M. Webster Virulence mechanisms of Heterorhabditis heliothidis and its bacterial associate, Xenorhabdus luminescens, in non-immune larvae of the greater wax moth, Galleria mellonella. Int. J. Parasitol. 18: Ehlers, R.-U., S. Stoessel, and U. Whyss The influence of phase variants of Xenorhabdus spp. and Escherichia coli (Enterobacteriaceae) on the propagation of entomopathogenic nematodes of the genera Steinernema and Heterorhabditis. Rev. Nematol. 13: Endo, Y., and W. R. Nickle Ultrastructure of the intestinal epithelium, lumen, and associated bacteria in Heterorhabditis bacteriophora. J. Helminthol. Soc. Wash. 58: Fire, A., S. Xu, M. K. Montgomery, S. A. Kostas, S. E. Driver, and C. C. Mello Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391: Forst, S., B. Dowds, N. Boemare, and E. Stackebrandt Xenorhabditis spp. and Photorhabditis spp.: bugs that kill bugs. Annu. Rev. Microbiol. 60: Gerritsen, L. J. M., and P. H. Smits Variation in pathogenicity of recombinations of Heterorhabditis and Xenorhabditis luminescens strains. Fundam. Appl. Nematol. 16: Granados, R. R., L. Guoxun, A. C. G. Derksen, and K. A. McKenna A new insect cell line from Trichoplusia ni (BTI-Tn 5B1 4) susceptible to Trichoplusia ni single enveloped nuclear polyhedrosis virus. J. Invert. Pathol. 64: Guo, L., R. O. Fatig, I. I. I., G. L. Orr, B. W. Schafer, J. A. Strickland, K. Sukhapinda, A. T. Woodsworth, and J. K. Petell Photorhabdus luminescens W-14 insecticidal activity consists of at least two similar but distinct proteins: purification and characterization of toxin A and toxin B. J. Biol. Chem. 274: Han, R. C., and R.-U Ehlers Cultivation of axenic Heterorhabditis spp. dauer juveniles and their response to non-specific Photorhabdus luminescens food signals. Nematologica 44: Han, R., W. M. Wouts, and L. Li Development of Heterorhabditis spp. strains as characteristics of possible Xenorhabditus subspecies. Rev. Nematol. 13: Han, R., W. M. Wouts, and L. Li Development and virulence of Heterorhabditis spp. strains associated with different Xenorhabditus luminescens isolates. J. Invert. Pathol. 58: Hink, W. F Established insect cell line from cabbage looper, Trichoplusia ni. Nature 225: Johnigk, S.-A., and R.-U. Ehlers Endotokia matricida in hermaphrodites of Heterorhabditis spp. and the effect of the food supply. Nematology 1: Khan, A., and W. M. Brooks A chromogenic bioluminescent bacterium associated with the entomophilic nematode Chromonema heliothidis. J. Invert. Pathol. 29: Milstead, J. E Heterorhabditis bacteriophora as a vector for introducing its associated bacterium into the hemocoel of Galleria mellonella larvae. J. Invert. Pathol. 33: Müller, H.-M, G. Dimopoulos, C. Blass, and F. C. Kafatos A hemocyte-like cell line established from the malaria vector Anopheles gambiae expresses six prophenoloxidase genes. J. Biol. Chem. 274: Poinar, G. O., Jr The presence of Achromobacter nematophilus in the infective stage Neoplectana sp. (Steinernematidae: Nematoda). Nematologica 12: Poinar, G. O., Jr Description and biology of a new insect parasitic rhabditoid Heterorhabditis bacteriophora. Nematologica 21: Poinar, G. O., Jr., G. M. Thomas, and R. Hess Characteristics of the specific bacterium associated with Heterorhabditis bacteriophora (Heterorhabditidae: Rhabditida). Nematologica 23: Poinar, G. O., Jr., T. Jackson, and M. Klein Heterorhabditis megidis sp. n. (Heterorhabditidae: Rhabditida), parasitic on the Japanese beetle, Popillia japonica (Scararabaeidae: Coleoptera) Ohio. Proc. Helminthol. Soc. Wash. 54: Schneider, I Cell lines derived from late embryonic stages of Drosophila melanogaster. J. Embryol. Exp. Morphol. 27: Silva, C. P., N. R. Waterfield, P. J. Daborn, P. Dean, T. Chilver, C. P. Y. Au, S. Sharma, U. Potter, S. E. Reynolds, and R. H. French-Constant Bacterial infection of a model insect: Photorhabdus luminescens and Manduca sexta. Cell. Microbiol. 4: Tan, M.-W., S. Mahajan-Miklos, and F. M. Ausubel Killing of Caenorhabditis elegans by Pseudomonas aeruginosa used to model mammalian bacterial pathogenesis. Proc. Natl. Acad. Sci. USA 96: Tzou, P., E. De Gregorio, and B. Lemaitre How Drosophila combats microbial infection: a model to study innate immunity and host-pathogen interactions. Curr. Opin. Microbiol. 5: Vaughn, J. L., R. H. Goodwin, G. L. Thomkins, and D. McCawley The establishment of two cell lines from the insect Spodoptera frugiperda (Lepidoptera; Noctuidae). In Vitro 13: Vivas, E. I., and H. Goodrich-Blair Xenorhabdus nematophilus as a model for host-bacterium interactions: rpos is necessary for mutualism with nematodes. J. Bacteriol. 183: