Entomopathogenic nematode-associated microbiota: from monoxenic paradigm to pathobiome

Abstract

Background: The holistic view of bacterial symbiosis, incorporating both host and microbial environment, constitutes a major conceptual shift in studies deciphering host-microbe interactions. Interactions between Steinernema entomopathogenic nematodes and their bacterial symbionts, Xenorhabdus, have long been considered monoxenic two partner associations responsible for the killing of the insects and therefore widely used in insect pest biocontrol. We investigated this "monoxenic paradigm" by profiling the microbiota of infective juveniles (IJs), the soil-dwelling form responsible for transmitting Steinernema-Xenorhabdus between insect hosts in the parasitic lifecycle.

Results: Multigenic metabarcoding (16S and rpoB markers) showed that the bacterial community associated with laboratory-reared IJs from Steinernema carpocapsae, S. feltiae, S. glaseri and S. weiseri species consisted of several Proteobacteria. The association with Xenorhabdus was never monoxenic. We showed that the laboratory-reared IJs of S. carpocapsae bore a bacterial community composed of the core symbiont (Xenorhabdus nematophila) together with a frequently associated microbiota (FAM) consisting of about a dozen of Proteobacteria (Pseudomonas, Stenotrophomonas, Alcaligenes, Achromobacter, Pseudochrobactrum, Ochrobactrum, Brevundimonas, Deftia, etc.). We validated this set of bacteria by metabarcoding analysis on freshly sampled IJs from natural conditions. We isolated diverse bacterial taxa, validating the profile of the Steinernema FAM. We explored the functions of the FAM members potentially involved in the parasitic lifecycle of Steinernema. Two species, Pseudomonas protegens and P. chlororaphis, displayed entomopathogenic properties suggestive of a role in Steinernema virulence and membership of the Steinernema pathobiome.

Conclusions: Our study validates a shift from monoxenic paradigm to pathobiome view in the case of the Steinernema ecology. The microbial communities of low complexity associated with EPNs will permit future microbiota manipulation experiments to decipher overall microbiota functioning in the infectious process triggered by EPN in insects and, more generally, in EPN ecology.

Keywords: Entomopathogenic nematode; Insect disease; Microbiota; Multigenic metabarcoding; Pathobiome; Pseudomonas; Xenorhabdus.

Fig. 1

Ex situ isolation, storage and laboratory multiplication of Steinernema entomopathogenic nematodes. a Ex situ Galleria trap methodology. Soil was sampled at a depth of 0–20 cm and transferred to 1 L plastic containers. Five last-instar Galleria mellonella larvae were placed in each container (ex situ Galleria trap). The containers were stored in the dark at 18 °C. After 7 days, dead G. mellonella were placed in a White trap consisting of an overturned 35-mm Petri dish covered with a piece of material (in blue) and placed in a larger Petri dish containing Ringer solution. After the production of a few generations of nematodes in G. mellonella cadavers, the IJs left the cadaver and migrated into the Ringer solution via the wet material (black dashes in Ringer solution). b IJ storage and laboratory Galleria trap method. IJs emerging from Steinernema and Heterorhabditis species were stored in 250-mL flasks containing 80 mL of Ringer solution supplemented with 0.1% formaldehyde at 9 °C, and flasks containing 80 mL of Ringer solution supplemented with 0.01% formaldehyde at 15 °C, respectively. Every 6 months, stocks were multiplied by adding 50–100 IJs to Galleria larvae placed on filter paper in a Petri dish (laboratory Galleria trap). Once the insects had died, their cadavers were placed in a White trap, as described above. Multiplication batches were labelled with the infestation date

Fig. 2

Bacterial communities associated with IJs of S. carpocapsae SK27, Galleria mellonella larvae and control samples. The IJ microbiota was investigated by metabarcoding with the V3V4 region of the 16S rRNA gene. a Venn diagram of the OTU detected in of S. carpocapsae SK27 batch SK27_23_08_16 samples, G. mellonella samples and control samples (Kitome_QE, Kitome_MN, Tap water and Ringer). OTUs (frequency cut-off > 80% and abundance cut-off > 0.01%) were assigned to genus level. Non-core set embeds OTUs that are not shared by all the control, the Galleria or the Steinernema replicates. b Heatmap showing the composition of the microbiota of S. carpocapsae SK27 batch SK27_23_08_16 samples. Each column represents a technical replicate. The 30 most abundant OTUs across the samples at the genus level of affiliation (Top30 Genus) are listed on the left. The percentage relative abundance is indicated by the gradient of blue hues.

Fig. 3

Bacterial communities associated with IJs from different species of Steinernema. a Simplified phylogeny of Steinernema. Phylogenetic relationships between 16 Steinernema strains based on a maximum likelihood (ML) analysis of partial internal transcribed spacer (ITS) regions (~ 850 bp). Heterorhabditis bacteriophora was used as outgroup. Branch support values (estimated by the aLRT [SH-like] method) are shown at the nodes (percentages of 100 replicates). The branch length scale bar below the phylogenetic tree indicates the number of nucleotide substitutions per site. The accession numbers of the sequences are indicated after the names of the nematodes. The sequences of strains used for metabarcoding studies are shown in colour. b Principal co-ordinates analysis (PCoA) based on Bray-Curtis distances for the IJ microbiota. Heterorhabditis and Steinernema samples are indicated by circles and triangles, respectively. Colours indicate the different Steinernema species. Each point represents a technical replicate. The proportion of the variance explained by each axis is shown. The five samples are significantly different (Permanova rpoB, Df = 4, R² = 0.73, p value = 10⁻⁴). c Heatmap showing the microbiota composition of IJ samples. Each column represents an IJ species. The 30 most abundant OTUs across the samples at the genus level of affiliation (Top30 Genus) are listed on the left. The percentage relative abundance is indicated by the gradient of blue hues. For b and c, IJ microbiota were described by metabarcoding with the 435 bp rpoB region. The EPN strains used here belong to the following batches: S. carpocapsae SK27_23_08_16 and B10_27_04_16; S. weiseri 583_09_06_15, TCH02_11_08_16 (t1), S. glaseri SK39_09_06_15; S. feltiae FRA200_09_06_15 and FRA200_12_08_15 and H. bacteriophora TT01_22_06_16 and TT01_15_03_16. Three to six technical replicates per batch were performed

Fig. 4

Bacterial communities associated with different samples from the species S. carpocapsae. Principal co-ordinate analysis (PCoA) based on Bray-Curtis distances for IJ microbiota obtained by metabarcoding with the 435-bp rpoB region. Each point represents an individual sample replicate. The different strains, multiplication batches and origins of S. carpocapsae are indicated by colours and symbols. The proportion of variance explained by each axis is shown. a Comparison of six S. carpocapsae strains. The six samples are significantly different (Permanova, Df = 5, R² = 0.87, p value = 10⁻⁴). b Comparison of four multiplication batches of S. carpocapsae SK27. The four samples are significantly different (Permanova, Df = 3, R² = 0.75, p value = 10⁻⁴). c Comparison of two laboratory origins of S. carpocapsae DD136 and S. carpocapsae All (Permanova: DD136 versus All, Df = 1, R² = 0.12, p value = 10⁻⁴; DGIMI versus USDA, Df = 1, R² = 0.25, p value = 10⁻⁴)

Fig. 5

Occurrence (x axis) and frequency of high abundance (y axis) of OTUs associated with laboratory-reared S. carpocapsae IJs. Abundant OTUs with read numbers accounting for more than 0.1% of the total reads for the sample were plotted. The darker the dot, the more OTUs with these characteristics are found at the position concerned. Among the OTUs with high occurrence rates (present in more than 70% of the samples), the yellow and orange areas delimit two groups. Orange group: frequency of high abundance > 90% (core symbiont). Yellow group: frequency of high abundance < 90% (frequently associated microbiota or FAM). OTU identity is indicated to the right of the graph (bootstrap confidence ≥ 0.9). *OTU present in the control samples

Fig. 6

Distance phylogenetic tree of 62 bacterial isolates from S. carpocapsae IJs. The phylogenetic tree of taxa isolated from S. carpocapsae IJs was constructed with the 16S rRNA gene sequences (1377 nucleotides), with the Kimura two-parameter model [43] and the neighbour-joining method [44] included in SeaView 4.7 software. Bootstrap values (percentages of 1000 replicates) of more than 90% are shown at the nodes. Twelve type strains (in bold) of the Xenorhabdus, Pseudomonas, Stenotrophomonas, Alcaligenes, Ochrobactrum, Pseudochrobactrum, Achromobacter and Brevundimonas genera were added. The bar represents 1% sequence divergence

Fig. 7

Survival curves of the insect Spodoptera littoralis after the direct injection of IJ-associated Pseudomonas protegens and Pseudomonas chlororhaphis. IJ-associated Pseudomonas: P. protegens strains PPSg_SG6 APO, PpSw_SW4, PpSw_TCH07 2-2, PpSc_PP-SC-10 and P. chlororaphis strain PcSg_SK39 ApoA. Rhizosperic strains: P. protegens CHAO^T, P. chlororaphis CFBP2132^T. Positive controls: X. nematophila strain XnSc_F1 (bold purple curve). Negative control (bold red curve): E. coli CIP7624. We injected 10² or 10³ bacterial cells in the exponential growth phase and injected each of 20 last-instar larvae