Introduction
Steinernema carpocapsae (Weiser Reference Weiser1955) Wouts, Mráček, Gerdin and Bedding, 1982 is a cosmopolitan species of entomopathogenic nematode (EPN) (Půža et al. Reference Půža, Machado and Malan2025). Its presence has been documented in the Balkans (Grubišić et al. Reference Grubišić, Drmić, Balaško, Bažok, Škvorc and Čačij2025; Nježić Reference Nježić2017; Shishinova et al. Reference Shishinova, Budurova and Gradinarov1998) and other European countries (Laznik et al. Reference Laznik, Tóth, Lakatos and Trdan2008; Torrini et al. Reference Torrini, Landi, Benvenuti, De Luca, Fanelli, Troccoli, Tarasco, Bazzoffi and Roversi2014; Weiser Reference Weiser1955). Morphological analysis of infective juveniles and adults remains a fundamental approach in nematode taxonomy. However, due to morphological similarities among species and overlapping morphometrics, molecular techniques such as DNA sequencing and phylogenetic analyses have become indispensable tools for distinguishing S. carpocapsae from closely related species. Steinernema carpocapsae shows considerable intraspecific variability in morphometrics, as reviewed by Mráček et al. (Reference Mráček, Půža and Nermut2014), who remeasured morphometric characters of several previously described populations, including the US population redescribed by Poinar (Reference Poinar1967) and those from early descriptions of later synonymised populations of S. carpocapsae. There are several molecularly confirmed descriptions of S. carpocapsae populations from Italy (Torrini et al. Reference Torrini, Landi, Benvenuti, De Luca, Fanelli, Troccoli, Tarasco, Bazzoffi and Roversi2014), Norway (Haukeland et al. Reference Haukeland, Klingen and Brurberg2006), Türkiye (Ertürk et al. Reference Ertürk, Ustaoğlu and Akyazı2014), Bosnia and Herzegovina (Đekanović Reference Đekanović2024), and Iran (Ebrahimi et al. Reference Ebrahimi, TanhaMaafi and Sharifi2019), which provide additional data on intraspecific variability in the morphometrics of S. carpocapsae.
Steinernema carpocapsae forms a symbiotic relationship with Xenorhabdus nematophila, a Gram-negative bacterium that plays a crucial role in insect pathogenicity. The nematode carries bacterial cells in a specialised vesicle in the upper intestine and, after infecting a host, releases them into the insect’s haemocoel. The bacterium rapidly multiplies and produces toxins and enzymes that suppress the immune system and degrade the host’s tissue. This mutualistic interaction ensures the nutrition and reproduction of the nematode and leads to the death of the host within 24 to 48 hours. The efficiency of this nematode-bacterium complex makes S. carpocapsae a valuable tool in biological pest control and one of the most successful and frequently used commercialised EPN species in insect pest management (EPPO 2021).
In addition to the typical bacterial symbionts, recent studies have revealed the existence of other bacteria associated with EPN, termed the ‘second bacterial circle’ (Ogier et al. Reference Ogier, Akhurst, Boemare and Gaudriault2023). This refers to the diverse collection of environmental and opportunistic bacteria that coexist in the insect cadaver after infection by the nematode-bacteria complex. While the primary symbiotic bacteria are responsible for rapidly killing the host and suppressing competing microbes, the role of second bacterial circle species in the infection process remains largely unknown. These secondary bacteria are usually not vertically transmitted and can originate from the soil (Ogier et al. Reference Ogier, Pagès, Frayssinet and Gaudriault2020), the gut flora of the insect (Dillon et al. Reference Dillon, Vennard, Buckling and Charney2005; Gomes et al. Reference Gomes, de Almeida and Cônsoli2023; Mason et al. Reference Mason, Stepien, Blum, Holt, Labbe, Rush and Raffa2011), and may enter the host carried by the nematode in the intercuticular space (Gouge and Snyder Reference Gouge and Snyder2006) or attached to the cuticle of the nematode, as demonstrated for some plant-parasitic nematodes (Topalović et al. Reference Topalović, Elhady, Hallmann and Richert-Pöggeler2019). There are reports of an entomopathogenic effect of several bacterial species from the second bacterial circle associated with steinernematids and S. carpocapsae, including Pseudomonas spp. (Mahar et al. Reference Mahar, Darban, Gowen, Hague, Jan, Munir and Mahar2004; Ogier et al. Reference Ogier, Pagès, Frayssinet and Gaudriault2020; Ruiu et al. Reference Ruiu, Marche, Mura and Tarasco2022; Shah et al. Reference Shah, Maistrou, van Loon and Dicke2025; Sugiyama and Hasegawa Reference Sugiyama and Hasegawa2025; Zwyssig et al. Reference Zwyssig, Spescha, Patt, Belosevic, Machado, Regaiolo, Keel and Maurhofer2024), Serratia spp. (Guo et al. Reference Guo, Zou, Chen, Liu, Ye, Huang and Wu2025; Kim et al. Reference Kim, Kim, Seo, Shrestha, Kim, Nalini and Yi2009; Ortega-Estrada Reference Ortega-Estrada, Del Rincon-Castro, Basurto-Rios, Toledo and Ibarra2012; Ruiu et al. Reference Ruiu, Li and Tarasco2024; Sugiyama and Hasegawa Reference Sugiyama and Hasegawa2025), Stenotrophomonas sp. (Li et al. Reference Li, Pu, Han, Li, An, Song, Fan, Chen and Tong2023; Ogier et al. Reference Ogier, Akhurst, Boemare and Gaudriault2023; Soleimani et al. Reference Soleimani, Jamshidi, Khakvar and Hasheminia2022), Enterococcus faecalis (Mason et al. Reference Mason, Stepien, Blum, Holt, Labbe, Rush and Raffa2011), and several others (Ogier et al. Reference Ogier, Akhurst, Boemare and Gaudriault2023). Bacterial entomopathogenicity is mostly demonstrated by direct injection into the haemocoel and rarely by oral ingestion. Some bacteria associated with entomopathogenic nematodes, when ingested, may directly contribute to killing the host by degrading gut tissues (Guo et al. Reference Guo, Zou, Chen, Liu, Ye, Huang and Wu2025; Lin et al. Reference Lin, Zhou, Ali, Fu, Gao, Jin, Fang and Wang2024), or by degrading gut tissue, allowing other bacteria to enter the haemocoel and become pathogenic (Mason et al. Reference Mason, Stepien, Blum, Holt, Labbe, Rush and Raffa2011).
Phenotypic traits can be influenced by genetic and environmental differences. Environmental influences are often neglected or diminished. Environmental conditions can affect EPN morphometrics in S. carpocapsae (Nguyen and Smart Reference Nguyen and Smart1995); for example, food depletion in infected cadavers over time and the absence of crucial ingredients in artificial food can shorten the body and tail length of infective juveniles. It has been reported that oral administration of Serratia marcescens can influence changes in the microbial composition of the environment and affect the nutrient cycle in the cadaver (Guo et al. Reference Guo, Zou, Chen, Liu, Ye, Huang and Wu2025), which could potentially impact nematode reproduction and development (Ogier et al. Reference Ogier, Pagès, Frayssinet and Gaudriault2020), which may also affect the EPN morphometrics. The presence of another bacterium frequently associated with EPN, Alcaligenes faecalis (Ogier et al. Reference Ogier, Pagès, Frayssinet and Gaudriault2020), has been connected with changes in the morphology of Heterorhabditis bacteriophora (Poinar Reference Poinar1988) and S. carpocapsae (Ogier et al. Reference Ogier, Pagès, Frayssinet and Gaudriault2020). Apart from naturally caused differences, some variability can have its origin in methodological inconsistencies in descriptions of populations.
This study aims to provide a characterisation of Serbian populations of S. carpocapsae by combining traditional taxonomic methods with genetic analyses. Furthermore, this work aims to isolate and molecularly characterise several of their bacterial symbionts and cultivable associates and to test their pathogenicity towards Galleria mellonella. By combining these approaches, we seek to enhance the accuracy of species identification by pointing out some possible causes of morphological variability that could provide an improved description of this nematode-bacterium complex, contributing to a deeper understanding of the diversity and ecology within this important nematode species.
Materials and methods
Soil sampling and isolation of entomopathogenic nematodes
The extensive sampling was part of a broader ongoing survey on EPN in the country. Seventy-six soil samples from several locations were collected from Mt Avala, near Belgrade, Serbia. The samples were taken in a way to equally represent different ecosystems and habitats present. Thirty soil samples were collected from the protected area Avala, dominated by natural habitats (mainly deciduous mixed forests of oak species, beech with the presence of maple, ash, elm, lime, and hazel trees), and the remaining samples were collected from lower surrounding areas dominated by meadows, ruderal meadow and border meadow-forest habitats, and agroecosystems (orchards, vineyards, and arable fields). Each sampling site has been photographed to assess vegetation, and physical and chemical analyses of collected soil on positive sites have been performed. Soil was taken from the centre and five points on the edge of an approximately 5 m diameter circle, at a depth of 5–10 cm in the soil profile. Approximately 1,000 ml of soil was collected per sample and placed in plastic containers. Five last-instar larvae of the greater wax moth, Galleria mellonella (Lepidoptera: Pyralidae), were added to each container as EPN traps (Bedding and Akhurst Reference Bedding and Akhurst1975). Three larvae were placed in pre-punctured 1.5 ml Eppendorf tubes to allow direct contact with the soil while retaining the larvae inside the tubes, and the remaining two larvae were allowed to move freely within the soil substrate. The containers were kept in the dark at room temperature (20 ± 3°C) (Glazer and Lewis Reference Glazer, Lewis, Navon and Ascher2000). Larval mortality was checked over a 7- to 10-day period. Dead larvae showing symptoms of EPN infection were collected and transferred to White traps (White Reference White1927) for the isolation of infective juveniles (IJs). Emerging IJs were collected, and their pathogenicity was confirmed by reinoculation of fresh G. mellonella larvae. IJs emerging from reinoculated larvae were stored in water in cell culture flasks (Greiner Bio-One) and kept in a refrigerator at 7°C.
Molecular identification and phylogenetic analysis of EPNs
For molecular identification, genomic DNA was extracted from freshly collected infective juveniles using the DNeasy Blood and Tissue Kit (Qiagen, Germany) following the manufacturer’s protocol. The internal transcribed spacer (ITS) region of rDNA was amplified using the primers TW81 (5′-GTTTCCGTAGGTGAACCTGC-3′) and AB28 (5′-ATATGCTTAAGTTCAGCGGGT-3′) (Spiridonov et al. Reference Spiridonov, Reid, Podrucka, Subbotin and Moens2004). The resulting PCR products were purified and sent for sequencing to Macrogen (the Netherlands). The quality of the sequence chromatograms was assessed using FinchTV software (Geospiza, Inc 2004), and manual corrections were made where necessary. Multiple sequence alignment was performed using the ClustalW algorithm (Larkin et al. Reference Larkin, Blackshields, Brown, Chenna, McGettigan, McWilliam, Valentin, Wallace, Wilm, Lopez, Thompson, Gibson and Higgins2007), and final sequence editing was carried out in the MEGA 12 package (Tamura et al. Reference Tamura, Stecher and Kumar2021). The sequences were then compared with existing sequences of S. carpocapsae in the NCBI GenBank according to Torrini et al. (Reference Torrini, Landi, Benvenuti, De Luca, Fanelli, Troccoli, Tarasco, Bazzoffi and Roversi2014), with the addition of sequences from populations for which morphometric data were available and some sequences of other newly described valid species of the ‘carpocapsae’ group according to Půža and Machado (Reference Půža and Machado2024). The phylogenetic tree was constructed using the maximum likelihood (ML), neighbour-joining (NJ), and minimum evolution (ME) methods (Felsenstein Reference Felsenstein1985), with branch support evaluated by the bootstrap test with 1,000 replicates.
Morphological analysis
Morphological analyses were conducted on infective juveniles, males, and females of the first and second generations. First-generation males and females were collected from dissected Galleria larvae on the third day, and second-generation adults were collected on the fifth day after the first symptoms of infection, using 100 IJs per larva. Infective juveniles used for morphological analysis were collected on the fifth day after the first IJ emerged from cadavers.
Specimens were heat-killed and fixed in 60°C triethanolamine formalin (TAF), gradually transferred into glycerin (Seinhorst Reference Seinhorst1959), and mounted on permanent glass slides with appropriately thick supports. Morphological analysis was performed using a Leica DM LS2 light microscope and a microQ camera (ToupTek Photonics). Digital images were captured and processed using ToupView software, which was also used for the morphometric analysis of taxonomic characters in 25 individuals. Measurements included total body length (TBL), maximum body width (MBW), distance from the anterior end to the excretory pore (EP) and nerve ring (NR), and oesophagus length (EL). For males, additional measurements were taken for spicule length (SpL), measured as a curved line, and gubernaculum length (GuL), and relevant ratios were calculated.
Isolation and identification of symbiotic bacteria
Symbiotic and associated bacteria were isolated from the haemolymph of G. mellonella larvae infected with the isolated IJ EPNs after one life cycle in G. mellonella, following isolation from the original soil sample for both S. carpocapsae populations. In parallel, isolation from uninfected larvae was performed. Baiting from original soil samples was repeated to get infected G. mellonella larvae directly with IJ present in the soil. Insect cadavers were surface-sterilised by sequential immersion in 1% sodium hypochlorite (NaOCl) and 70% ethanol, followed by rinsing in sterile distilled water. Haemolymph was aseptically collected five days after infection (Ruiu et al. Reference Ruiu, Marche, Mura and Tarasco2022). The samples were streaked onto Luria–Bertani agar (LA) and tryptic soy agar (TSA) plates and incubated at 30°C for 48 h. Pure bacterial colonies were selected and subjected to molecular identification. Genomic DNA was extracted from pure bacterial cultures, and the 16S rRNA and rpoB genes were amplified by PCR using gene-specific primers. The 16S rRNA gene was amplified using the forward primer UNI16SF (5′-GAGAGTTTGATCCTGGC-3′) and reverse primer UNI16SR (5′-AGGAGGTGATCCAGCCG-3′) (Kim et al. Reference Kim, Kim, Seo, Shrestha, Kim, Nalini and Yi2009). PCR was performed with an initial denaturation step at 94°C for 5 min, followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 50°C for 30 s, and extension at 72°C for 1 min 30 s, with a final extension at 72°C for 7 min. The rpoB gene was amplified using the forward primer rpoB1206 (5′-ATCGAAACGCCTGAAGGTCCAAACAT-3′) and reverse primer rpoBR3202 (5′-ACACCCTTGTTACCGTGACGACC-3′) (Kim et al. Reference Kim, Kim, Seo, Shrestha, Kim, Nalini and Yi2009). The PCR conditions consisted of an initial denaturation at 95°C for 5 min, followed by 30 cycles of denaturation at 95°C for 40 s, annealing at 54°C for 40 s, and extension at 72°C for 1 min 30 s, with a final extension at 72°C for 7 min. The PCR products were purified using a PCR purification kit (Thermo Fisher Scientific, Waltham, MA, USA) and submitted to Eurofins Genomics (Cologne, Germany) for sequencing. The obtained sequences were compared with reference sequences available in the National Center for Biotechnology Information (NCBI) database using the Basic Local Alignment Search Tool (BLAST) to identify the bacterial isolates. The nucleotide sequences obtained in this study were deposited in the NCBI GenBank (Table S6). Phylogenetic analysis based on the obtained 16S rRNA gene sequences was performed using MEGA11 software. A phylogenetic tree was constructed using the Neighbour-Joining (NJ) method with default settings. For each bacterial species, a 16S rRNA gene sequence of a reference strain retrieved from the NCBI database was included in the analysis. Micrococcus luteus was used as the outgroup to root the tree.
Pathogenicity bioassays of the bacterial isolates
The pathogenicity of the bacterial strains was evaluated using last-instar larvae of G. mellonella weighing approximately 400 mg. The experiment was conducted in six-well culture plates with one larva per well, each representing one replicate (six larvae weighed to be 2–2.5 g). The pathogenicity of the bacterial isolates Xenorhabdus nematophila AV27-4, X. nematophila AV10-2, Enterococcus casseliflavus AV10-3, Stenotrophomonas maltophilia AV27-3, and Serratia marcescens AV10-1 was assessed using five concentrations of bacterial suspensions (102, 103, 104, 106, and 108 CFU/ml). Overnight bacterial cultures were centrifuged at 4,000 rpm for 30 min, and the resulting cell pellets were resuspended in sterile distilled water to obtain suspensions at the required concentrations. Each concentration was tested in four replicates (24 larvae in total at each concentration), and the entire experiment was independently repeated twice (240 larvae in total per bacterial isolate). Two application methods were used: direct injection of 20 μl of the bacterial suspension into the last larval proleg (Salgado-Morales et al. Reference Salgado-Morales, Martínez-Ocampo, Obregón-Barboza, Vilchis-Martínez, Jiménez-Pérez and Dantán-González2019), and ingestion of food mixed with 0.1 ml of 108 CFU/ml per gram of food (Kalia et al. Reference Kalia, Sharma, Shapiro-Ilan and Ganguly2014). As controls, larvae were injected with the same volume of sterile water or fed uncontaminated food (24 larvae with each independent repetition). The pathogenicity of Bacillus cereus, B. paranthracis, and Priestia megaterium was tested only at the maximal concentration using the injection method. Each treatment was conducted under controlled laboratory conditions at 27°C in a FOC Connect incubator (VELP Scientifica). Larval mortality was recorded at 24, 48, 72, and 96 hours after the start of the experiment. Dead larvae were counted, and insect mortality was recorded for each treatment.
Mortality data were analysed using probit regression to estimate median lethal concentrations (LC₅₀ and LC₉₀) (Ruiu et al. Reference Ruiu, Marche, Mura and Tarasco2022). When control mortality exceeded 5%, observed mortality in treated groups was corrected using Abbott’s formula prior to analysis. Probit analysis was performed in R Statistical Software (v4.6.0, R Core Team 2021) using a generalised linear model (GLM) with a binomial error distribution and probit link function, with log₁₀-transformed bacterial concentrations as the explanatory variable. LC₅₀ and LC₉₀ values were estimated from the fitted model, and 95% confidence intervals were calculated using the delta method based on the variance–covariance matrix of model parameters. Model fit was evaluated using a Pearson chi-square goodness-of-fit test, and the slope with standard error was used to describe the relationship between bacterial concentration and larval mortality. As no significant differences in mortality were observed between the experimental runs, the data were analysed together.
Results
Nematode isolates and molecular identification
Four soil samples (5%) tested positive for Steinernema spp., of which two isolates (AV10 and AV27) were identified as S. carpocapsae and the remaining two (AV24 and AV35) as S. thesami (Gorgadze, 1988) Gorgadze & Lordkipanidze, 2004. Both species are the first records in Serbia. Three isolates were found in natural forest habitats, representing 10% of positive records (AV24, AV27, and AV35), and one (AV10) was found from more or less disturbed habitats due to human activities (representing 2% of positive samples). Sample AV10 was collected from a border ruderal meadow habitat with Poaceae species and wild Rubus caesius, in the vicinity of mixed forest at coordinates 44° 40′ 20.13800″ N, 20° 30′ 57.97600″ E at an altitude of 224 m a.s.l. Sample AV10 was isolated from silty clay loam approximately 2.5 km from locations where the other three EPNs were found. The rest of the EPN-positive samples were collected at coordinates 44° 41′ 41.68400″ N, 20° 30′ 40.91900″ E at 440 m a.s.l. (AV27), 44° 41′39.8″N 20°30′55.7″E at 422 m a.s.l. (AV24) and 44° 41′47.2″N 20°30′58.7″E at 436 m a.s.l. (AV35), and isolated from silt loam soil under the same mixed deciduous forest habitat. After amplification and sequencing of the ITS region, the newly obtained DNA sequences were submitted to GenBank under accession numbers PZ618321 (AV10), PZ618322 (AV27), QB070757 (AV24), and QB070758 (AV35). AV10 and AV27 sequences were a 100% match with NCBI-deposited sequences of the corresponding ribosomal DNA region of S. carpocapsae and AV24 and AV35 had 99.74% and 99.87% shared sequence identity with sequence KC287223.1 belonging to S. thesami. The map with sampling locations and positive sampling sites, including results of physical and chemical soil analysis, is provided in Supplementary material (Figure S1 and Tables S1 and S2).
Molecular phylogenetic analysis of Steinernema carpocapsae isolates
For the phylogenetic analysis of the ITS region, the sequences belonging to S. carpocapsae isolates obtained in this study were aligned with ITS sequences of S. carpocapsae from previously published work, including some species of the ‘carpocapsae’ group, and S. feltiae sequences used as an outgroup. Phylogenetic trees generated by Neighbour-Joining, Maximum Parsimony, and Minimum Evolution methods showed no conflict in branching order and no significant difference in branch support levels; therefore, only the ML tree is presented (Figure 1). The ML tree supported several clades corresponding to species of the ‘carpocapsae’ group and clearly distinguishable, highly supported S. carpocapsae clades. Both sequences from Serbian populations clustered with all S. carpocapsae sequences. The Steinernema carpocapsae clade included two moderately supported subclades: one comprising all European populations and a single non-European sequence from Iran, and the second formed by the remaining non-European populations.
Molecular phylogenetic analysis of Steinernema species based on ITS rDNA sequences. Evolutionary history was inferred using the maximum likelihood method and the Tamura-Nei model. The tree is drawn to scale, with branch lengths measured as the number of substitutions per site (scale bar = 0.050). The analysis included 36 nucleotide sequences. Evolutionary analyses were conducted in MEGA 12. Bootstrap support values (1,000 replicates) are shown next to the branches. Steinernema feltiae was used as the outgroup.
Figure 1. Long description
The phylogenetic tree begins with a root on the left, branching into two main lineages. The bottom lineage contains the outgroup S. feltiae K M 0 1 6 4 1 9 dot 1 and L N 6 1 1 1 3 6 dot 1 with a bootstrap value of 99. The upper lineage contains the remaining 34 sequences.
Within the upper lineage, a large clade of S. carpocapsae and Steinernema species dominates the top section. Two specific sequences are highlighted in blue boxes: S. carpocapsae P Z 6 1 8 3 2 2 Serbia A V 2 7 and S. carpocapsae P Z 6 1 8 3 2 1 Serbia A V 1 0. These are nested within a group including isolates from Switzerland, Iran, Portugal, Italy, Slovenia, Russia, Bosnia and Herzegovina, Belgium, Turkey, Croatia, Jordan, Israel, India, China, U S A, and Mexico.
A separate lower sub-clade within the upper lineage includes S. nepalense H Q 1 9 0 0 4 4 dot 1, Steinernema spp. G U 1 3 0 1 8 2 dot 1, S. surkhense H Q 1 9 0 0 4 2 dot 1, S. siamkayai J Q 2 6 7 9 9 2 dot 1, S. minutum G U 6 4 7 1 5 6 dot 1, and S. tami A Y 1 7 1 2 8 0 dot 1. S. scapterisci A F 3 3 1 9 1 5 dot 1 sits at the base of this upper lineage.
Bootstrap values ranging from 73 to 99 are positioned at major nodes. A scale bar at the bottom left indicates 0 dot 0 5 0 substitutions per site.
Morphometric analysis of Steinernema carpocapsae isolates
Morphological analysis was performed only on Steinernema carpocapsae isolates. Morphological examination of infective juveniles and adults from both S. carpocapsae populations revealed characters typical of the species as described by Poinar. No qualitative morphological differences were observed between the two populations and previous descriptions. Morphometric measurements and body ratios of IJs and males of the first and second generations are presented in Table 1. All measurements of both Serbian populations of IJ fell within the ranges reported by Poinar (Reference Poinar1967). Infective juveniles of the population AV10 showed mean values for all measured morphometric characters mainly typical for the species. Means of all morphometric characters and ratios of IJ of AV10 were very similar to the corresponding averages calculated from the means of each character for all populations presented in Supplementary material (Table S3). Infective juveniles of population AV27 were more distinct, with a low position of the excretory pore (43.4) and nerve ring (87), a long oesophagus (132.2), and a short tail length (50). These atypical values influenced the ratios derived from them, with the most prominent being an E ratio of 87.7 (due to the lower position of EP and shorter tail) and ratio b of 4.1 (due to the long oesophagus and relatively short body length).
Morphometric measurements and body ratios of infective juveniles, first- and second-generation males of Steinernema carpocapsae populations (AV10 and AV27). All measurements are in μm: mean ± standard deviation, with min and max values in brackets
Table 1. Long description
The table presents data for n equals 25 for all groups. Measurements are provided as mean plus or minus standard deviation with min-max ranges in brackets.
Header Structure:
* Infective juveniles: A V 2 7 and A V 1 0.
* 1st generation males: A V 2 7 and A V 1 0.
* 2nd generation males: A V 2 7 and A V 1 0.
Key Morphometric Rows:
* T B L (Total Body Length): Infective juveniles range from 464 to 618; 1st generation males are significantly larger (1190 to 2219); 2nd generation males range from 724 to 1187.
* M B W (Maximum Body Width): Increases from approximately 25 in juveniles to a peak of 144 in 1st generation A V 1 0 males, then decreases in 2nd generation males (50 to 60).
* E P (Excretory Pore): Values range from 32.3 to 55.3 in juveniles, 43 to 79 in 1st generation, and 40 to 64 in 2nd generation.
* N R (Nerve Ring): Ranges from 72.6 to 98.7 in juveniles, increasing to 86 to 120 in 1st generation, and 80 to 111 in 2nd generation.
* E L (Oesophagus Length): Ranges from 113 to 143.2 in juveniles, 136 to 184 in 1st generation, and 124 to 170 in 2nd generation.
* T L (Tail Length): Only recorded for Infective juveniles, ranging from 31.2 to 59.
* A B D (Anal Body Diameter): Recorded only for males, ranging from 35 to 52 in 1st generation and 22 to 35 in 2nd generation.
* S p L (Spicule Length): Recorded only for males, ranging from 61 to 73 in 1st generation and 42 to 63 in 2nd generation.
* G u L (Gubernaculum Length): Recorded only for males, ranging from 39 to 52 in 1st generation and 25 to 44 in 2nd generation.
Ratios:
* E percent: Only for juveniles (62.8 to 125.3).
* D percent: Ranges from 24.2 to 42.1 in juveniles, 28 to 56 in 1st generation, and 29 to 47 in 2nd generation.
* a, b, and c ratios: Only provided for Infective juveniles.
* S W: Recorded for males, ranging from 1.2 to 2.0 in 1st generation and 1.45 to 2.61 in 2nd generation.
* G S percent: Recorded for males, ranging from 54 to 75 in 1st generation and 49 to 87 in 2nd generation.
TBL, total body length; MBW, maximum body width; EP, excretory pore; EL, oesophagus length; TL, tail length; NR, nerve ring; ABD, anal body diameter; SpL, spicule length (measured as a curve); GuL, gubernaculum length.
Ratios: E(%) = EP/TL × 100; a = TBL/MBW; b = TBL/EL; c = TBL/TL; D(%) = EP/EL × 100; SW = SpL/ABD; GS(%) = GuL/SpL × 100.
Average values of measured-characters mean calculated from the corresponding means of all populations presented in Table S3 were almost exact matches (with less than 1 μm difference) with the mean values given for DD136 by Poinar (Reference Poinar1967). Conversely, and unexpectedly, average ratios calculated from the corresponding means of all populations presented differed from ratios of DD136, especially values of E and D of 73.5 vs. 60 and 32.4 vs. 26%, respectively. When ratios were calculated from the means of character measurements given by Poinar (although mathematically incorrect), they were 71.6, 31.7, 22.3, 4.7, and 10.5 for E, D, a, b, and c, respectively (see Table S3, data in brackets).
The morphometric measurements of first-generation males of both examined populations also fell within the ranges given by Poinar (Reference Poinar1967), except for total body length, maximal body diametre, and oesophageal length, which exceeded the maximal values, especially in the large males of population AV10. Atypical mean values were recorded for TBL (1850), MBW (144), EL (173), and from EL-derived small ratio D (38%) within population AV10. In both populations atypically small mean values of gubernaculum lengths were recorded (43 and 45) that influenced small ratio GS of 65 and 67% in AV10 and AV27, respectively. Second-generation males were generally smaller and more slender than first-generation males, as expected. However, we could not find similar descriptions for comparison with the results of this study.
Additionally, the comparison of means for some morphometric characters for several populations of S. carpocapsae is presented in Table 2. The comparison of means for the complete morphometrics of the same populations described in the literature available to us (including those from this study) is provided in the Supplementary material, in Table S3 for infective juveniles and in Table S4 for first-generation males.
Comparative morphometrics of various Steinernema carpocapsae strains
Table 2. Long description
The table contains 8 columns: Country, Strain, Reference, I J T B L, I J D, I J E (%), M 1 G S (%), and M 1 S W (%).Data rows include:* USA, DD 136 (Poinar 1967): 558.0, 26.0, 60.0, 72.0, 151.0.* USA, DD 136 (Mráček et al 2014): 525.0, 33.0, 71.0, 82.0, 155.0.* USA, All (Mráček et al 2014): 594.0, 32.0, 73.0, 79.0, 143–156.* Russia, Agriotos (Mráček et al 2014): 566.0, 33.0, 77.0, 82.0, 150.0.* Switzerland, Dok83 (Mráček et al 2014): 626.0, 35.0, 79.0, 77.0, 161.0.* Czechia, Pieridarum (Mráček et al 2014): 556.0, 38.0, 81.0, 60.0, 173.0.* Iran, I R Moghan 1 (Ebrahimi et al 2019): 538.5, 27.3, 65.1, 68.0, 187.0.* Türkiye, 52–118 (Ertürk et al 2014): 557.8, 29.1, 71.1, 67.0, 143.0.* Norway, (Haukeland et al 2006): 562.7, 30.8, 72.6, 70.0, 210.0.* Italy, Caorle (Torrini et al 2014): 559.0, 30.7, 69.1, 66.0, 150.0.* Bosnia and Herzegovina, N 0 2 7 / 22 (ekanovi 2024): 584.0, 39.0, 76.0, 70.0, 187.0.* Bosnia and Herzegovina, N n 0 0 9 / 22 (ekanovi 2024): 551.0, 35.5, 72.0.* Serbia, A V 27 (This work): 539.3, 32.8, 87.7, 65.0, 170.0.* Serbia, A V 10 (This work): 586.2, 31.3, 74.9, 67.0, 150.0.Summary statistics:* Average: 564.5, 32.4, 73.5, 71.2, 165.6.* Median: 558.5, 32.4, 72.8, 70.0, 158.0.* S D: 25.0, 3.6, 6.5, 6.6, 19.5.* C V: 4.4, 11.0, 8.9, 9.3, 11.8.
Data rows include:
* USA, DD 136 (Poinar 1967): 558.0, 26.0, 60.0, 72.0, 151.0.
* USA, DD 136 (Mráček et al. 2014): 525.0, 33.0, 71.0, 82.0, 155.0.
* USA, All (Mráček et al. 2014): 594.0, 32.0, 73.0, 79.0, 143–156.
* Russia, Agriotos (Mráček et al. 2014): 566.0, 33.0, 77.0, 82.0, 150.0.
* Switzerland, Dok83 (Mráček et al. 2014): 626.0, 35.0, 79.0, 77.0, 161.0.
* Czechia, Pieridarum (Mráček et al. 2014): 556.0, 38.0, 81.0, 60.0, 173.0.
* Iran, IRMoghan1 (Ebrahimi et al. 2019): 538.5, 27.3, 65.1, 68.0, 187.0.
* Türkiye, 52–118 (Ertürk et al. 2014): 557.8, 29.1, 71.1, 67.0, 143.0.
* Norway, (Haukeland et al. 2006): 562.7, 30.8, 72.6, 70.0, 210.0.
* Italy, Caorle (Torrini et al. 2014): 559.0, 30.7, 69.1, 66.0, 150.0.
* BandH, N027/22 (Đekanović 2024): 584.0, 39.0, 76.0, 70.0, 187.0.
* BandH, Nn009/22 (Đekanović 2024): 551.0, 35.5, 72.0.
* Serbia, AV27 (This work): 539.3, 32.8, 87.7, 65.0, 170.0.
* Serbia, AV10 (This work): 586.2, 31.3, 74.9, 67.0, 150.0.
Summary statistics at the bottom:
* Average: 564.5, 32.4, 73.5, 71.2, 165.6.
* Median: 558.5, 32.4, 72.8, 70.0, 158.0.
* S D: 25.0, 3.6, 6.5, 6.6, 19.5.
* C V: 4.4, 11.0, 8.9, 9.3, 11.8.
Isolated bacteria
Bacteria isolated under the media and conditions used from cadavers infected with IJs of the studied isolates of S. carpocapsae are presented in Table S5. Xenorhabdus nematophila and Enterococcus casseliflavus were isolated from cadavers infected with IJ of both populations studied. Priestia megaterium, Bacillus paranthracis, and Bacillus cereus were found only in cadavers infected with the AV27 isolate and Serratia marcescens was isolated only from AV10-infected cadavers. In repeated baiting, we failed to get infected Galleria mellonella larvae in AV10 original soil sample, and from AV27 newly infected cadavers, only Xenorhabdus nematophila, Bacillus cereus, and Enterococcus casseliflavus were isolated. Phylogenetic analysis (Figure 2), together with results from the BLAST analysis (Table S6), confirmed bacterial identification at the species level. Two bacterial strains were isolated in repeated experiments from uninfected larvae: Staphylococcus saprophyticus (QB104198) and Enterococcus mundtii (QB104199).
Phylogenetic relationships of strains isolated from Steinernema carpocapsae AV10 and S. carpocapsae AV27. The branching pattern, rooted using Micrococcus luteus NCTC2665 as the outgroup, was generated by the neighbour-joining method, and distances were calculated using the Kimura two-parameter model. Bootstrap values are shown for each node with 50% or greater support, based on 1,000 replicates. The isolate names and most similar MegaBLAST search results are shown, with the GenBank identifier presented next to the reference strain label.
Figure 2. Long description
A neighbour-joining phylogenetic tree rooted with Micrococcus luteus N C T C 2665 as the outgroup at the bottom. Moving upward, the tree branches into several distinct clades with bootstrap values indicated at the nodes.
* The first major clade above the outgroup contains Xenorhabdus nematophila strains, including A V 102, A V 274, and A V 27 Z 1, grouped with reference strain A T C C 19061 with a bootstrap value of 100.
* The next clade upward contains Serratia marcescens A V 101 grouped with strain E L P 1 dot 10 with a bootstrap value of 100.
* Above this is a clade for Stenotrophomonas maltophilia A V 273 and strain A C Y C a dot 6 E with a bootstrap value of 100.
* The next section contains a cluster of Enterococcus casseliflavus strains, including A V 103 and A V 27 Z 2, grouped with strain A S E 4 with a bootstrap value of 100.
* Near the top, a clade contains Priestia megaterium A V 271 grouped with N B R C 15308 equals A T C C 14581 with a bootstrap value of 100.
* The uppermost clade consists of Bacillus species, including Bacillus cereus A V 27 Z 3 and A T C C 14579 (bootstrap 84), and Bacillus paranthracis A V 272 and strain B t C 4 (bootstrap 100).
Isolates from the current study are highlighted within blue rectangular boxes.
Pathogenicity of symbiotic and associated bacteria
Oral ingestion of food contaminated with bacterial suspensions of any tested bacterial species did not cause mortality in last-instar larvae of G. mellonella. In contrast, injection of bacterial suspensions into the larval body usually resulted in a strong positive mortality response. G. mellonella larval mortality increased with both bacterial concentration and exposure time for all tested bacterial species, except P. megaterium, which did not appear virulent to the tested insect (Figure 3).
Mortality of Galleria mellonella larvae during the 96-hour pathogenicity bioassay after injection with different concentrations (102–108 CFU/ml) of bacterial suspensions (Enterococcus casseliflavus, Serratia marcescens, Stenotrophomonas maltophilia) (a), and comparison of insect mortality with Bacillus cereus, Bacillus paranthracis, Priestia megaterium, and Xenorhabdus nematophila strains at the maximal dose (b).
Figure 3. Long description
Two-panel multi-graph layout.
Panel (a) contains three line graphs. Each graph has an x-axis for Time (h) from 0 to 96 and a y-axis for Mortality (%) from 0 to 100.
1. Serratia marcescens (top-left): Concentrations 10 super 4, 10 super 6, and 10 super 8 C F U / ml show a rapid increase to 100% mortality within 24 to 48 hours. 10 super 3 and 10 super 2 reach approximately 90% mortality. The control remains below 20%.
2. Stenotrophomonas maltophilia (top-right): 10 super 8 reaches 100% mortality by 24 hours. 10 super 6 reaches 100% by 72 hours. Lower concentrations (10 super 2 to 10 super 4) show a gradual increase to 60-75% mortality.
3. Enterococcus casseliflavus (bottom-center): Shows a dose-dependent response where only 10 super 8 reaches approximately 80% mortality by 96 hours. Lower doses range between 20% and 50% mortality.
Panel (b) is a single large line graph comparing different strains at the maximal dose.
- Xenorhabdus nematophila A V 10, Serratia marcescens, Stenotrophomonas maltophilia, Bacillus cereus, and Bacillus paranthracis all show a steep vertical increase reaching 100% mortality by 24 hours.
- Xenorhabdus nematophila A V 27 reaches 100% mortality by 72 hours.
- Enterococcus casseliflavus reaches approximately 80% mortality by 96 hours.
- Priestia megaterium and the Control show the lowest mortality, staying below 25% throughout the 96-hour period.
The highest pathogenicity was observed for S. marcescens, which caused rapid and nearly complete mortality across most tested concentrations. At 48 hours post-treatment, concentrations from 103 to 108 CFU/ml resulted in mortality levels exceeding 95%. Even the lowest tested concentration (102 CFU/ml) caused high mortality, exceeding 85%. A similar pathogenicity pattern at the highest concentration was recorded for both Bacillus isolates tested. Stenotrophomonas maltophilia exhibited high pathogenicity, with estimated LC50 and LC90 values of 2.66 × 103 and 7.54 × 105 CFU/ml/larva at 48 hours post-infection, respectively (Table 3). High concentrations (106 and 108 CFU/ml) caused rapid increases in mortality, reaching nearly complete larval death by 48 hours. For E. casseliflavus, mortality increased gradually over time but remained generally lower compared with the other tested bacteria, with estimated LC50 and LC90 values of 4.94 × 106 and 5.09 × 1010 CFU/ml, respectively. At the highest concentration (108 CFU/ml), mortality reached approximately 85% by 96 hours, whereas lower concentrations resulted in moderate to low mortality.
Pathogenicity of four bacterial strains isolated from Steinernema carpocapsae-infected cadavers towards Galleria mellonella last-instar larvae 48 hours after injection of 20 μl of bacterial suspension at five concentrations, expressed as lethal concentration (LC) values (CFU/ml)
Table 3. Long description
The table contains seven columns: Bacterial strain, n, L C sub 50 (C F U forward slash ml) (95% C L s), L C sub 90 (C F U forward slash ml) (95% C L s), Slope (plus or minus S E), chi-squared, and d f. All strains have an n value of 288.
* X dot nematophila A V 27-4: L C sub 50 is 3.81 times 10 super 4 (range 1.11 times 10 super 4 to 1.27 times 10 super 5); L C sub 90 is 2.72 times 10 super 8 (range 1.86 times 10 super 7 to 3.97 times 10 super 9); Slope is 0.332 plus or minus 0.044; chi-squared is 9.13; d f is 3.
* X dot nematophila A V 10-2: L C sub 50 is 5.94 times 10 super 4 (range 3.20 times 10 super 4 to 1.10 times 10 super 5); L C sub 90 is 6.25 times 10 super 5 (range 2.49 times 10 super 5 to 1.57 times 10 super 6); Slope is 1.253 plus or minus 0.1; chi-squared is 0.01; d f is 2.
* E dot casseliflavus: L C sub 50 is 4.94 times 10 super 6 (range 9.44 times 10 super 5 to 2.59 times 10 super 7); L C sub 90 is 5.09 times 10 super 10 (range 1.01 times 10 super 9 to 2.57 times 10 super 12); Slope is 0.319 plus or minus 0.050; chi-squared is 1.18; d f is 2.
* S dot maltophilia: L C sub 50 is 2.66 times 10 super 3 (range 1.01 times 10 super 3 to 6.97 times 10 super 3); L C sub 90 is 7.54 times 10 super 5 (range 1.59 times 10 super 5 to 3.55 times 10 super 6); Slope is 0.522 plus or minus 0.06; chi-squared is 8.26; d f is 2.
A clear dose-dependent response was observed for both X. nematophila isolates, with larval mortality increasing progressively with both bacterial concentration and time (Figure 4). The highest concentration (108 CFU/ml) resulted in complete mortality at 72 hours for X. nematophila AV27, while for X. nematophila AV10 after 24 hours. Intermediate mortality of 50% was observed at 104 CFU/ml for both strains after 48 and 72 hours for AV27 and AV10, respectively. At 48 hours post-infection, the estimated LC50 and LC90 values for X. nematophila AV10 and AV27 were 5.94 × 104 and 6.25 × 105, and 3.81 × 104 and 2.72 × 108 CFU/ml, respectively.
Mortality of Galleria mellonella larvae during the 96-hour pathogenicity bioassay after injection with different concentrations (102–108 CFU/ml) of Xenorhabdus nematophila AV27 and X. nematophila AV10 bacterial suspensions.
Figure 4. Long description
Two side-by-side line graphs. Both share a Y axis labeled Mortality percentage from 0 to 100 and an X axis labeled Time h from 0 to 96.
Left panel titled Xenorhabdus nematophila A V 27.
At 24 hours, mortality for 10 super 8 C F U per ml jumps to 80 percent, while 10 super 6 reaches 60 percent. Other concentrations remain below 25 percent.
By 96 hours, mortality follows a dose-dependent trend. 10 super 8 reaches nearly 100 percent. 10 super 6 reaches 80 percent. 10 super 4 and 10 super 3 converge around 55 percent. 10 super 2 is at 30 percent, and the Control is lowest at approximately 15 percent.
Right panel titled Xenorhabdus nematophila A V 10.
At 24 hours, only 10 super 8 shows a sharp increase to nearly 100 percent mortality.
At 48 hours, 10 super 6 rises sharply to meet 10 super 8 at nearly 100 percent.
By 96 hours, 10 super 8 and 10 super 6 remain at 100 percent. 10 super 4 reaches 55 percent. 10 super 3 reaches 40 percent. The Control is at 25 percent. Notably, 10 super 2 shows the lowest mortality, remaining near 5 percent throughout the duration.
Discussion
In recent decades, research on EPN has increased significantly in several areas, with many new descriptions of both new and known species from different countries around the globe. Apart from the typical bacterial symbionts, recent studies have revealed the existence of other bacteria associated with insect cadavers after infection by IJs of EPN. Their influence on the biology and ecology of EPN remains largely unknown. In this work, we have isolated four isolates of EPN belonging to two different species. As expected, a higher percentage of positive samples have been found in natural forests compared to arable land and human-disturbed habitats sampled in this work, and these results agree with Tian et al. (Reference Tian, Blanco-Pérez, Zhu, Zheng, Chen, Sui, Wu, Li and Zhu2026). Due to differences in habitat, soil characteristics, and the relatively long distance between the sites, we considered two isolates of S. carpocapsae to be different populations and S. thesami isolates to be members of the same population of this species. Phylogenetic analysis of ITS sequences demonstrates that Serbian populations AV10 and AV27 belong to S. carpocapsae and show high genetic similarity to other European and global isolates. This finding is consistent with previous studies reporting relatively low intraspecific variation in ITS sequences of S. carpocapsae (e.g., Torrini et al. Reference Torrini, Landi, Benvenuti, De Luca, Fanelli, Troccoli, Tarasco, Bazzoffi and Roversi2014) and clear separation, with a similar tree branching order, of S. carpocapsae from closely related species (Spiridonov and Subbotin Reference Spiridonov, Subbotin, Hunt and Nguyen2016). However, the ITS region provides limited resolution to show intraspecific genetic differences, for which higher-resolution molecular approaches would be required. Morphometric values of IJ of AV10 were comparable with those of DD136 (Poinar Reference Poinar1967) and matched well with the average values for S. carpocapsae populations described. Within the infective juveniles of population AV27, the atypical values mentioned in the results section are comparable with some complement extremes already reported in the literature for several populations (Ertürk et al. Reference Ertürk, Ustaoğlu and Akyazı2014; Ebrahimi et al. Reference Ebrahimi, TanhaMaafi and Sharifi2019; Đekanović Reference Đekanović2024). Similarity in measured values of taxonomic characters, and dissimilarity of ratios derived from the same character values given by Poinar (Reference Poinar1967), with the average character values for all described populations mentioned, suggests that the ratios given by Poinar, and cited subsequently, may be erroneous. The timing of the extraction of juveniles of S. carpocapsae after their first emergence from infected cadavers for morphological studies has rarely been addressed in the literature. IJ body size can decrease with the time of extraction, and the length of the tail and EP distance within IJ can be significantly variable (Nguyen and Smart Reference Nguyen and Smart1995). Depending on the time of extraction, the mean tail length (based on 20 measurements) is reported to be between 45 and 55 μm and EP distance between 35 and 38 (range 28–45), within the same (Pieridarum) population of S. carpocapsae (Nguyen and Smart Reference Nguyen and Smart1995). Both morphometric measurements influence ratio E, which has been reported to be below 70 in only three populations (including the redescription by Poinar discussed previously) out of 14 descriptions, and above 80 or close to it in three. Considering that the timing of the IJ extraction may be different in published descriptions, part of the intraspecific variability could be attributed to methodological error rather than natural differences among populations.
Contrary to infective juveniles, males of population AV10 appeared more distinct than average males of described populations, with mean values of MBW and EL being much higher than in reported descriptions. Both populations had a lower-positioned EP, average spicule length, and shorter gubernacula than average, which resulted in a smaller GS ratio than average compared with all populations and as given in Poinar’s redescription. All the other atypical and other measurements presented for the studied populations are comparable with some of the previous reports.
The main issue in male morphometrics lies in the inconsistent method of measuring spicule length (which is often not addressed) and the consequent artificial variability of calculated ratios. In four populations where the spicules were measured as a curve, the mean GS ratio is below 70%, with maximum spicule lengths of 74 μm recorded. In contrast, in four populations remeasured by Mráček et al. (Reference Mráček, Půža and Nermut2014), where spicule length was measured along the cord, mean GS ratios were in the range of 77–82, well above the average calculated values of means for all populations and DD136 given by Poinar (Reference Poinar1967), which are around 72%. In the case of Serbian populations, the ratio of along-the-cord measurement to spicule-curve length was around 0.8. If this factor is multiplied by the values of mean spicule lengths measured along the curve (from descriptions where the method of spicule measurement is explicitly stated), values in the range from 55 to 61 for spicule length are calculated. When gubernaculum lengths are divided by these values, GS ratios are above 77, which is consistent with the observations of Mráček et al. (Reference Mráček, Půža and Nermut2014).
Bearing in mind the above discussion, and excluding genetically and environmentally influenced differences, the variability of taxonomic characters and the reported differences in morphometric values of both IJs and males could be partially influenced by artificial variability caused by methodological inconsistencies in the published descriptions. Overall, the results indicate significant differences in pathogenicity among the tested bacterial species towards G. mellonella. No mortality was observed after larvae were fed bacteria-contaminated food, suggesting either an inability of the bacteria to cross the gut barrier and enter the haemocoel or an insufficient inoculum. Pseudomonas protegens can apparently be pathogenic when orally administered to several insect species by forced feeding with a bacterial suspension (Ruiu et al. Reference Ruiu, Marche, Mura and Tarasco2022; Zwyssig et al. Reference Zwyssig, Spescha, Patt, Belosevic, Machado, Regaiolo, Keel and Maurhofer2024), and in Zwyssig et al. (Reference Zwyssig, Spescha, Patt, Belosevic, Machado, Regaiolo, Keel and Maurhofer2024), exposure to IJ most likely enhanced crossing of the gut barrier. There are reports of oral entomopathogenicity of some human pathogenic strains of Bacillus cereus to second-stage larvae of Galleria mellonella, where larvae were naturally fed, with or without Cry1C toxin, and much higher mortality was observed in the mixed treatment (Fedhila et al. Reference Fedhila, Buisson, Dussurget, Serror, Glomski, Liehl, Lereclus and Nielsen-LeRoux2010). The referenced work used much higher bacterial inocula than those used in this study and targeted earlier larval stages. This suggests stronger immune responses in older larval stages to pathogenic bacteria and indicates the need for additional factors to complement pathogenicity. Similarly, B. thuringiensis toxins were required to lyse midgut epithelial cells and assist Enterococcus faecalis, a frequent member of insect microbiota, in becoming pathogenic by entering the haemocoel (Mason et al. Reference Mason, Stepien, Blum, Holt, Labbe, Rush and Raffa2011).
Ramos et al. (Reference Ramos, Farinacio, Angel and Rojas2025) showed over 90% relative growth inhibition but no mortality of naturally fed fall armyworm treated with several species of Enterococcus, including E. casseliflavus, which is consistent with our results. We have shown that E. casseliflavus can cause significant mortality when injected into G. mellonella larvae, especially at the highest concentration, and we could not find similar results or direct evidence of its pathogenicity in the published literature. Serratia marcescens used in our study exhibited an entomopathogenic effect only when directly injected into the haemocoel of G. mellonella larvae. Kim et al. (Reference Kim, Kim, Seo, Shrestha, Kim, Nalini and Yi2009) obtained similar results, where Serratia sp. did not show pathogenicity when orally administered to Spodoptera exigua larvae but caused 100% mortality when injected into the haemocoel. Likewise, S. marcescens, when injected, killed 100% of Zaphobus morio and Periplaneta fuliginosa larvae (Sugiyama and Hasegawa Reference Sugiyama and Hasegawa2025). Interestingly, Serratia sp. pathogenicity had a synergistic effect on the mortality of S. exigua when orally administered together with B. thuringiensis (Kim et al. Reference Kim, Kim, Seo, Shrestha, Kim, Nalini and Yi2009). However, Lin et al. (Reference Lin, Zhou, Ali, Fu, Gao, Jin, Fang and Wang2024) reported that serralysin, a vital toxic protein from S. marcescens, caused disruption of the gut tissue of Mythimna separata, resulting in larval death when S. marcescens was orally administered. Similarly, Guo et al. (Reference Guo, Zou, Chen, Liu, Ye, Huang and Wu2025) demonstrated the pathogenicity of S. marcescens to immature stages of S. frugiperda; larvae naturally fed on maize leaves immersed in bacterial suspension showed changes in gut tissues. The bacterial isolates used by Kim et al. (Reference Kim, Kim, Seo, Shrestha, Kim, Nalini and Yi2009) and Sugiyama and Hasegawa (Reference Sugiyama and Hasegawa2025) were associated with insects and EPN, while the S. marcescens isolates used by Lin et al. (Reference Lin, Zhou, Ali, Fu, Gao, Jin, Fang and Wang2024) and Guo et al. (Reference Guo, Zou, Chen, Liu, Ye, Huang and Wu2025) were endophytic bacteria from sugarcane and a soil isolate, respectively. This may suggest that, apart from other possible reasons that may contribute to oral pathogenicity, plant-associated members of Serratia spp. might have the ability to cross the gut barrier by secreting toxins, a capability that insect- and EPN-associated members may lack.
In our study, both B. cereus and B. paranthracis, injected at a concentration of 108 CFU/ml, caused total mortality of G. mellonella larvae. The entomopathogenicity of B. cereus has been confirmed (Loulou et al. Reference Loulou, Mastore, Caramella, Bhat, Brivio, Machado and Kellel2023), with comparable results at lower doses than those used in the present study. Regarding B. paranthracis, we have found only indirect evidence of its entomopathogenicity in the literature (de Sousa Reference de Sousa2021). However, because B. paranthracis is a recently separated member of the B. cereus sensu lato group, older studies may have identified it simply as B. cereus using 16S methods. Therefore, some historical ‘B. cereus’ cadaver isolates could actually include B. paranthracis.
Our results showed pathogenicity of S. maltophilia to G. mellonella, but only when injected into the haemocoel. Our findings are consistent with Li et al. (Reference Li, Pu, Han, Li, An, Song, Fan, Chen and Tong2023), but not with Soleimani et al. (Reference Soleimani, Jamshidi, Khakvar and Hasheminia2022), who sprayed larvae with bacterial suspension and confirmed pathogenicity, or with Ogier et al. (Reference Ogier, Pagès, Frayssinet and Gaudriault2020), who did not confirm pathogenicity using washed cells (105–106 CFU) even when injecting S. litoralis larvae. Published differences may be due to variation in virulence among bacterial isolates and susceptibility of insect hosts, and it appears that insect mortality may be influenced or initiated by S. maltophilia metabolites.
The pathogenicity of X. nematophila documented here is comparable with published work of Ogier et al. (Reference Ogier, Pagès, Frayssinet and Gaudriault2020), who reported 50% mortality as LT50 of 27 hours with 102–103 CFU/larva used. Chapuis et al. (Reference Chapuis, Pagès, Emelianoff, Givaudan and Ferdy2011) reported similar mortality 28–30 hours after a higher dose of 2,000 bacterial cells was injected into Galleria mellonella larvae and much higher mortality above 95% was observed at 46 hours after injection. Chapuis et al. (Reference Chapuis, Pagès, Emelianoff, Givaudan and Ferdy2011) observed very fast changes in mortality after 24 hours ranging from 0% at 24 hours to 40%–70% in just the next 8 hours. With the 10 times lower or higher dose tested (200 CFU/larva in concentration of 104 CFU/ml and 20,000 CFU/larva at 106 CFU/ml), we recorded similar mortality at 48 h, suggesting the need for intermediary concentrations to be tested for better mortality resolution and more frequent observations of mortality. Differences in recorded mortality between the tested strains used in this work could be explained by their differences in virulence (Hasan et al. Reference Hasan, Ahmed, Mollah, Lee and Kim2019).
Many bacteria of the second bacterial circle likely require nematodes and/or their core symbionts, other members of the microbiota, or their metabolites to express virulence and pathogenicity. As mentioned previously, some of the bacterial members associated with EPN-infected cadavers may originate from the soil or insect gut and are most likely vectored into the insect haemocoel by the nematode (Gouge and Snyder Reference Gouge and Snyder2006). We isolated only two bacteria from the gut of three healthy larvae that differed from those used in this study. This result suggests other than gut origin of tested bacteria. However, we did not isolate the whole microbiome, and this statement should be interpreted with caution. Within the insect haemocoel, many EPN-associated bacteria may cause insect death, and it is reasonable to assume that they can influence the environment inside the cadaver in different manners. Some EPN bacterial associates, such as Serratia marcescens, have been reported to alter the microbial composition of the environment inside the cadaver and to affect the nutrient cycle in the cadaver (Guo et al. Reference Guo, Zou, Chen, Liu, Ye, Huang and Wu2025).
The environmental conditions inside EPN-infected cadavers, which can be difficult to demonstrate, can affect morphometrics in several steinernematids and H. bacteriophora (Nguyen and Smart Reference Nguyen and Smart1995), and the presence of some atypical bacterial species can alter the morphology of H. bacteriophora (Poinar Reference Poinar1988) and S. carpocapsae (Ogier et al. Reference Ogier, Pagès, Frayssinet and Gaudriault2020). Some of the isolated bacterial members differed between the two studied populations, and it could be hypothesised that, in addition to other possible causes, differences in microbial composition inside the cadaver may be one of the environmental factors influencing EPN morphometrics and may partially explain the differences between the populations studied.
In conclusion, we have demonstrated the pathogenicity of some bacteria isolated from S. carpocapsae-infected cadavers to Galleria mellonella. Apart from genetically influenced differences among S. carpocapsae populations, we have shown that some variability may result from methodological inconsistencies in population descriptions, which artificially introduce variation. Among environmentally caused differences, we hypothesise a potential influence of different bacteria on the morphometrics of the populations studied. It would be interesting to directly test this hypothesis in future experiments where the bacteria would be injected with IJ of EPN into experimental hosts. Research on the origin of the bacteria found is planned.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0022149X26102521.
Acknowledgements
We would like to thank Aleksa Novaković, an undergraduate student at the Faculty of Agriculture, for his help in collecting the soil samples, Dr Ivana Jovičić from the Institute of Pesticides and Environmental Protection, Belgrade, for providing us with additional batches of Galleria mellonella larvae when needed, and Dr Jelena Bogosavljević Lipovac, from the Faculty of Agriculture, for physical and chemical soil analysis performed. We acknowledge the Ministry of Science, Technological Development and Innovation of the Republic of Serbia for the scholarship granted to Angelina Popović for her PhD studies.
Financial support
This research was funded by the Ministry of Science, Technological Development and Innovation of the Republic of Serbia, grant numbers 451-03-34/2026-03/200116 (NG), 451-03-12/2026-03/3499 (AP), 451-03-33/2026-03/200178 (AM), and 451-03-34/2026-03/200178 (JL).
Competing interests
The authors declare no conflicts of interest.
Open access
