Nematode Maintenance

S. carpocapsae strain LBIN-1 was maintained under laboratory conditions by infecting G. mellonella larvae and collecting the infective juveniles (IJs) from White traps, basically as described by Kaya & Stock (1997). If required, IJs were stored at 12 °C until needed.

Nematode Identification

Identification of the strain, S. carpocapsae LBIN-1, was done by morphological examination of the IJs, and was based on the taxonomic description of entomopathogenic nematodes genera by Adams & Nguyen (2002). Corroboration of the genus and identification of the species were based on comparisons of DNA sequences to reported reference sequences. Genomic DNA was extracted as described by Stock (2009), and its integrity was examined by agarose gel electrophoresis.

DNA from IJs was used as a template to amplify and sequence the 18S (partial), ITS1, 5.8S, ITS2, and 28S (partial) fragments from the rDNA operon, using the following primers: direct 5′TTGATTACGTCCCTGCCCTTT-3′ and reverse 5′-TTTCACTCGCCGTTACTAAGG-3′ (Vrain et al. 1992). The reaction mixture and amplification conditions by Vrain et al. (1992) were followed, except that the annealing temperature was 52 °C for 30 sec. All amplifications were performed in a Perkin Elmer GeneAmp PCR System 2400. Amplicons were sequenced by the Sanger technique, using a 3730 XL DNA Analyzer (Applied Biosystems). Sequences were analyzed by NCBI-Blastn and NCBI-Blastx (Altschul et. al. 1990,  http://blast.ncbi.nlm.nih.gov/Blast.cgi).

Identification of Bacteria

Both the symbiotic and the associated bacteria of S. carpocapsae LBIN-1 were identified by 16S rDNA Operon subunit DNA sequence comparison. The symbiotic bacterium was isolated from surface sterilized IJs, as described by Kaya & Stock (1997), except that a 0.7% thiomersal (Merthiolate™) solution was used, along with the streptomycin solution (6.64 mg/ml) to eliminate any bacterial contamination on the IJ's surface. Treated IJs were inoculated onto nutrient agar medium to corroborate surface-sterilization, and then the IJs were homogenized with a small pestle. Serial dilutions of homogenates were inoculated on NBTA plates (Akhurst 1980) to recover the symbiotic bacterium. The associated bacterium was recovered from the hemolymph of infected G. mellonella larvae, by extracting samples with a syringe. Serial dilutions were inoculated into nutrient agar, and bright red colonies were transferred to Serratia medium (7 gr K₂HPO₄, 3 gr KH₂PO₄, 1 gr casamino acids, 1 gr (NH₄)₂SO₄, 1 gr yeast extract, 0.1 gr MgSO₄-7H₂O, 15 gr bacteriologic agar, water to 1 liter, and 5 ml 50% dextrose) (ATCC medium 1399). The isolated bacterium was registered as LBSe-17.

Once both bacterial species had been isolated, single colonies of each were inoculated into nutrient broth and incubated at 28 °C, shaken at 150 rpm for 24 h, or until a minimum OD₆₀₀ = 1 was obtained. Genomic DNA from both bacterial species was extracted by the procedure described earlier (Chen & Kuo 1993), and its integrity was checked by agarose gel electrophoresis. These DNA samples were used as the template to amplify and sequence the 16S subunit from the rDNA operon with universal primers: direct EuF 5′-AGAGTTTGATCATGGCTCAG-3′; reverse EuR 5′-TACCTTGTTACGACTTCACC-3′ (Heddi et al. 1999). The reaction mixture and the amplification conditions were basically as described by Heddi et al. (1999) and performed in a Perkin Elmer GeneAmp PCR System 2400. Amplicons were sequenced by the Sanger technique, using a 3730 XL DNAAnalyzer (Applied Biosystems). Sequences were analyzed by NCBI-Blastn and at the Bio Informatic Bacterial Identification (BIBI) website ( http://blast.ncbi.nlm.nih.gov/Blast.cgi;  http://umr5558sud-strl.univ-lyonl.fr/lebibi/lebibi.cgi).

Fig. 1.

A. Galleria mellonella larvae infected with Steinernema carpocapsae LBIN-1 strain. B. Bright red colonies of Serratia marcescens LBSe-17 growing on nutrient agar plates after being inoculated with hemolymph of infected larvae.

Pathogenicity of the Serratia marcescens LBSe-17 Strain

To test the ability of the associated bacterium, S. marcescens LBSe-17, to cause septicemia of G. mellonella larvae, Koch's postulates were followed. The S. marcescens LBSe-17 strain, previously isolated from cadavers, was inoculated by dipping a dental needle (30 gauge, 19 mm long) into a bacterial suspension, and introducing it into the hemocoels of 24 G. mellonella last instar larvae by puncturing the area between the prologs (see Supplementary Fig. 1A). Control larvae were punctured in the same way with sterile needles. Treated larvae were placed into 24-well microtiter plates and incubated at 28 °C and 80% RH for 24 h, when mortality was recorded. Next the larvae were surface-sterilized with 0.3% NaClO and a hemolymph sample was withdrawn with a syringe from each cadaver and inoculated into Serratia medium. Three replicates of 24 larvae each were performed.

Phoretic Ability of S. carpocapsae LBIN-1 IJs to Carry S. marcescens LBSe-17

The ability of the S. carpocapsae LBIN-1 strain to carry the S. marcescens LBSe-17 strain into G. mellonella larvae was tested by exposing the larvae only to the bacterium and also to a mixture of the bacterium and the nematode. Filter paper circles were impregnated with 80 µl of S. marcescens LBSe-17 concentrate (OD₆₀₀≈30) and placed at the bottom of each well of two 24-well microtiter plates. Approximately 40 active S. carpocapsae LBIN-1 IJs were added to each well in these plates. A third plate with only water was used as a control; one G. mellonella last-instar larva was placed into each well (24 larvae per treatment). Three replicates per treatment were conducted. Plates were incubated at 28 °C and 80% RH for 48 h, then mortality was recorded and hemolymph samples from cadavers and living larvae were inoculated on Serratia medium after the larvae had been surface-sterilized with 0.3% NaClO.

Location of S. marcescens LBSe-17 When Carried by S. carpocapsae LBIN-1 IJs

Whether the S. marcescens LBSe-17 strain was carried within the S. carpocapsae LBIN-1 IJs or on their surfaces during the infection process was determined by surface-sterilizing the IJs. IJs were surface sterilized with three 10 min washes of a 0.7% thiomersal (Merthiolate™) and streptomycin (6.64 mg/ml) solution and then rinsed with double-distilled water (ddH₂O). Sterility of the treated IJs suspension was tested by inoculating an aliquot on Serratia medium. Treated and untreated IJs, as well as a mixture of untreated IJs with a concentrated suspension of the S. marcescens LBSe-17 strain were used to infect 24 G. mellonella last instar larvae in 24-well microtiter plates. There were 3 treatments and 3 replicates per treatment. Plates were incubated at 28 °C and 80% RH for 48 h, then mortality was recorded, and hemolymph samples drawn from cadavers were inoculated onto Serratia medium after larvae had been surface-sterilized with 0.3% NaClO.

Artificial Induction of a Bacterium-Nematode Association

The possibility of artificially inducing an association between a bacterial strain and a nematode strain was tested by mixing the S. marcescens LBSe-17 strain with another nematode strain. For this purpose, IJs of Steinernema sp. LBIN-2 strain, native to central Mexico and with no history of bacterial association (except its own symbiont), were mixed with a concentrated suspension of S. marcescens LBSe-17. A mixture of S. marcescens LBSe-17 and S. carpocapsae LBIN-1 was also included in the experiment (see above). Infection and incubation of G. mellonella larvae as well as inoculations of hemolymph samples were carried out as described above.

Possible Specific Association between S. carpocapsae LBIN-1 and S. marcescens LBSe-17

Once the artificially induced association between the bacterial strain and a nematode strain was tested (see above), the specificity of such an association was compared to the natural association LBIN-1/LBSe-17 by maintaining both associations during 5 cycles of larval infections. IJs from larval cadavers were used to infect a set of new larvae, and in each cycle 5 cadavers were homogenized and diluted to be inoculated on Serratia medium; colony forming units (CFU) were quantified in each cycle. Additionally, 5 cadavers from the continuous maintenance of the S. carpocapsae LBIN-1 laboratory colony at the 29th infection cycle were used for CFU quantification.

Effect of S. marcescens LBSe-17 on the Nematode

The effect of the S. marcescens LBSe-17 strain on nematode growth was tested by quantifying IJs from 2 types of G. malonella larval cadavers: a) larvae showing from just a slight red coloration to unnoticeable S. marcescens LBSe-17 strain growth (see Supplementary Fig. 3A); and b) larvae showing an abundance of S. marcescens LBSe-17 strain growth (visible red tissue throughout the body) (see Supplementary Fig. 3B). IJ quantification was made from 20 larvae from each of these 2 types. Additionally, IJs were inoculated on NBTA plates with growing X. nematophilus and on Serratia medium with growing LBSe-17 strain. Nematode growth and reproduction were monitored on both treatments.

Nematode Identification

On 18 Jun 2003, a strain of an entomopathogenic nematode was isolated from a white grubinfested corn field located at Mazatan, state of Chiapas, in southern Mexico. When maintenance of the isolate was carried out by infecting G. mellonella larvae, a reddish coloration of the cadavers was observed, suggesting a Heterorhabditis infection (Fig. 1A). However, examination of the morphology of these nematodes indicated that it was a Steinernema species. When the genomic DNA from the nematode was used as a template to amplify and sequence the 18S (partial), ITS1, 5.8S, ITS2, and 28S (partial) fragments from the rDNA operon, 100% identity was observed with the reported AF121049.1 sequence, which corresponds to the species, S. carpocapsae. This corroborated the morphological observations and confirmed the species identification. This strain was registered as LBIN-1 at the CINVESTAV's Collection of Entomopathogenic Microorganisms log.

Identification of the Associated Bacterial Species

The reddish coloration of larval cadavers, sometime only observed at the spiracles, mouth and anus, was also detected in the hemolymph (Fig. 1A). Microscopic observations of infected hemolymph showed a concentrated suspension of homogeneously short bacilli. When samples of hemolymph were inoculated on nutrient agar plates, mainly bright-red bacterial colonies appeared (Fig. 1B). DNA extracted from this bacterial species was used as a template to amplify the 16S ribosomal subunit, and amplicons were sequenced. Sequence analysis at the Bio Informatic Bacterial Identification (BIBI) website showed a positive identification within the species Serratia marcescens, with 99% identity with the reported FJ360759.1 sequence of S. marcescens; which positively identified the bacterium associated with the S. carpocapsae LBIN-1 strain. The S. marcescens strain was registered as LBSe-17 in CINVESTAV's Entomopathogenic Microorganisms Collection log.

Pathogenicity of the S. marcescens LBSe-17 Strain

The high proliferation of the S. marcescens strain LBSe-17 in the insect cadavers infected by the nematode indicated a potential pathogenic effect of the bacterial strain. Once Koch's postulates were completed, all of the injected larvae (see Supplementary Fig. 1A) were killed in less than 24 h, and they displayed the typical reddish coloration (see Supplementary Fig. 1B). Diluted hemolymph from these cadavers plated on Serratia medium resulted in a large number of bright-red colonies (see Supplementary Fig. 1C), indicating that the S. marcescens LBSe-17 strain caused a lethal septicemia of G. mellonella larvae when inoculated into their hemocoels.

Presence of the Symbiotic Xenorhabdus bacteria

The high proliferation of the S. marcescens LBSe-17 strain in the insect cadavers infected by the nematode might suggest the loss of the natural symbiotic Xenorhabdus bacteria. When homogenates of surface sterilized nematodes and hemolymph samples drawn from the cadavers were inoculated on NBTA medium, typical dark purple colonies of Xenorhabdus were observed; and from them genomic DNA was extracted. Once the 16S subunit of the rDNA operon was amplified and sequenced, the analysis at the Bio Informatic Bacterial Identification (BIBI) website showed a 100% identity with the AY278674.1 strain of Xenorhabdus nematophila. This result indicates that despite the apparent proliferation of the S. marcescens LBSe-17 in the insect cadavers infected by the nematode, they still carried their symbiotic X. nematophila bacteria; which grew in the insect cadavers along with the S. marcescens strain LBSe-17.

Phoretic Ability of S. carpocapsae Carrying S. marcescens

To corroborate if S. marcescens LBSe-17 is infective to G. mellonella larvae only if S. carpocapsae LBIN-1 carries it into the insect, larvae were exposed to a mixture of IJs and this bacterial species as well as to the bacterial species alone. Forty eight h after inoculation, 100% mortality was observed in the treatment with both the IJs and these bacteria, while treatment with the bacteria alone showed 0% mortality. Once larvae were surface-sterilized, and hemolymph was withdrawn and inoculated onto agar plates, abundant growth of S. marcescens was observed from larvae exposed to both the IJs and these bacteria, whereas no bacterial growth was observed from the hemolymph of larvae exposed only to S. marcescens.

Location of S. marcescens When Carried by S. carpocapsae IJs

Only media inoculated with the unsterilized IJ treatments showed growth of S. marcescens, and this demonstrated that these bacteria had been removed from the sterilized IJ's surface. When G. mellonella larvae were infected with sterilized IJs (see Supplementary Fig. 2A), or unsterilized IJs (see Supplementary Fig. 2C), or with a mixture of the bacteria and the IJs, all 3 treatments showed 100% larval mortality. However, when hemolymph samples from the cadavers were inoculated on Serratia medium (see Supplementary Figs. 2B and 2D), bright red colonies failed to grow only from samples obtained from larvae that had been infected with sterilized IJs (see Supplementary Fig. 2C). In the bacteria plus nematode mixture, the bright red colony growth was much more abundant. Also, red-colored cadavers were observed only on the unsterilized treatments (see Supplementary Fig. 2C). These observations indicated that the S. marcescens LBSe-17 strain is carried by the S. carpocapsae LBIN-1 IJs on their body surfaces during the infection process.

Fig. 2.

Trends of Serratia marcescens LBSe-17 CFU counts from consecutive infection cycles on Galleria mellonella larvae involving 2 Steinernema carpocapsae strains, LBIN-1 and LBIN-2. S. marcescens LBSe-17 was found naturally associated with LBIN-1, and was artificially associated with LBIN-2 in the laboratory. A. Colony forming units (CFU) in larval homogenates infected by the natural association, LBIN-1/LBSe-17, throughout five experimental infection cycles. (triangles, continuous line), and the artificially induced association with LBIN-2/LBSe-17 (squares, dotted line). B. Colony forming units (CFU) in larval homogenates as in A, but with a purported 29th infection cycle (from colony maintenance) added to the natural association LBIN-1/LBSe-17.

Artificial Induction of a Bacteria-Nematode Association

Once the association between S. marcescens LBSe-17 and S. carpocapsae LBIN-1 was established, questions about the singularity of this association motivated an experiment to induce a similar association. For this purpose, the strain LBIN-2 of Steinernema sp., originally isolated from Chapingo, State of Mexico (central Mexico), which had no bacterial association history but with its own symbiotic bacterial species, was mixed with the S. marcescens LBSe-17 strain. G. mellonella larvae were exposed to this mixture, as well as to a mixture of S. marcescens LBSe-17 and S. carpocapsae LBIN-1, and 100% mortality ensued in both treatments. Samples of hemolymph and larval homogenates from cadavers of both treatments showed abundant growth of bright red colonies when inoculated on Serratia medium. This result indicated that this type of association may be induced on other nematodes and that the LBSe-17/LBIN-1 association may be fortuitous.

Possible Specific Association between S. carpocapsae LBIN-1 and S. marcescens LBSe-17

From the time samples were received from southern Mexico in our laboratory, to the end of this work (∼4 years), S. carpocapsae LBIN-1 strain was passed through G. mellonella larvae a total of 29 times. During this period, the association between S. carpocapsae LBIN-1 and S. marcescens LBSe-17 was consistently observed. To test this presumed stability, infections from the experiment described above were followed through 5 consecutive infection cycles on G. mellonella larvae (Fig. 2A). All treatments showed total larval mortality; however, substantial differences were observed in the CFU counts for each treatment. The artificially induced association LBSe-17/ LBIN-2 showed a drastic decline in the CFU count after each infection cycle, decreasing from an initial average count of 127,100 colonies per sample at the first infection cycle, to only one colony per sample at the fifth infection cycle. Meanwhile, the natural association LBSe-17/LBIN-1 showed a moderate decline from an initial average count of 634,900 colonies per sample at the first infection cycle, to 78,800 colonies at the fifth infection cycle. Figure 2A shows the relationship between the CFU count and the infection cycles as well as the logarithmic fit for each treatment. When the CFU quantification was made on samples from larval cadavers of the 29th infection cycle (made for maintenance of the S. carpocapsae LBIN-1 colony in the laboratory), this count was added to the previous data as a purported extension of the experiment to a hypothetical 29th cycle (Fig. 2B). This result supports the hypothesis of a possible specific, stable relationship between S. marcescens LBSe-17 and S. carpocapsae LBIN-1.

Effect of S. marcescens LBSe-17 on the Steinernema carpocapsae Nematodes

Because the growth of the nematode in the larval cadaver requires the presence of its bacterial symbiont, the possibility that the simultaneous growth of X. nematophila and S. marcescens in the infected larvae could have some effect on the nematode required investigation. The number of IJs varied from 0 to 1,300 per larva when S. marcescens growth was abundant, but the number IJs varied from 12,162 to 110,093 per larva when larvae showed just slight red to unnoticeable red coloration, i.e., visually undetectable S. marcescens growth. Additionally, when LBIN-1 IJs were inoculated on S. marcescens and X. nematophila axenic agar plate cultures, abundant growth of the nematode was observed only on the X. nematophila plates (see Supplementary Fig. 4A). S. marcescens plates not only showed a complete absence of nematode growth but also 100% mortality (see Supplementary Fig. 4B).