Introduction
The Japanese beetle, Popillia japonica (Coleoptera: Scarabaeidae), stands as a major insect pest native to Japan that reached mainland Europe in 2014 causing great damage in the Italian region (Glazer et al. Reference Glazer, Santoiemma, Battisti, De Luca, Fanelli, Troccoli, Tarasco, Sacchi, Bianchi, Gilloki, Gherardo and Mori2022; Gotta et al. Reference Gotta, Ciampitti, Cavagna, Bosio, Gilloli, Alma, Battisti, Mori, Mazza, Torrini, Paoli, Santoiemma, Simonetto, Lessio, Sperandio, Giacometto, Bianchi, Roversi and Marianelli2023). The larval stage exerts detrimental effects on grasslands and pastures, whereas adults feed on the foliage of agricultural crops and ornamental plants (Potter and Held Reference Potter and Held2002). Early investigations on this pest identified naturally occurring parasitism of P. japonica larva by entomopathogenic nematodes (EPNs) leading to the isolation of Steinenerma glaseri (Rhabditida: Steinernematidae) (Poinar Reference Poinar, Gaugler and Kaya1990). Indeed, EPNs are obligate insect parasites that belong mainly to the families Steinernematidae and Heterorhabditidae (order Rhabditida) (Alatorre-Rosas and Kaya Reference Alatorre-Rosas and Kaya1990; Peters Reference Peters1996). These nematodes form symbiotic associations with gram-negative bacteria of the genera Xenorhabdus and Photorhabdus respectively (Bird and Akhurst Reference Bird and Akhurst1983; Griffin et al. Reference Griffin, Boemare, Lewis, Grewal, Ehlers and Shapiro-Ilan2005). Infective juveniles (IJs) penetrate the host via natural openings or by penetrating the cuticle and releasing the bacteria while developing to the adult stage (Snyder et al. Reference Snyder, Stock, Kim, Flores-Lara and Forst2007). The nematode-bacteria complex kills the host within 24 to 48 hours through septicaemia or toxaemia (Dowds and Peters Reference Dowds, Peters and Gaugler2002; Forst and Clarke Reference Forst, Clarke and Gaugler2002). Adult nematodes feed on bacteria-degraded insect tissues, generating subsequent generations of IJs that emerge from the depleted cadaver to seek out new hosts (Poinar Reference Poinar, Gaugler and Kaya1990).
Our understanding of EPNs has progressively expanded enabling their use as effective biological control agents. EPNs offer a distinctive advantage because of their ability to find and infect soil-dwelling and cryptic environment pests, mirroring their natural habitat characteristics. The application of EPNs yielded favourable outcomes in managing soil pests such as Diabrotica spp. (Coleoptera: Chrysomelidae) and Sphenophorus spp. (Coleoptera: Curculionidae) (Jaffuel et al. Reference Jaffuel, Sbaiti and Turlings2020; Leite et al. Reference Leite, Chacon-Orozco, Shapiro-Ilan, Baldo and Cardoso2023). Nevertheless, field applications of EPNs against P. japonica larvae have yielded inconsistent control over the years (Helmberger et al. Reference Helmberger, Thaler, Shields and Wickings2018). Georgis and Gaugler (Reference Georgis and Gaugler1991) suggested that factors like unsuitable nematode strains or adverse environmental conditions impeded an effective control of P. japonica during field trials. This led to focus on isolating new strains, seeking native EPNs better adapted to the local environment, and climatic, and edaphic soil conditions (Aryal et al. Reference Aryal, Nielsen, Sumaya, De Faveri, Wilson and Riegler2022a; Lacey et al. Reference Lacey, Grzywacz, Shapiro-Ilan, Frutos, Brownbridge and Goettel2015). EPNs exhibit a ubiquitous global distribution, thriving in diverse landscapes ranging from agricultural fields to urban gardens (Aryal et al. Reference Aryal, Nielsen, Sumaya, De Faveri, Wilson and Riegler2022a; Campos-Herrera et al. Reference Campos-Herrera, Escuer, Labrador, Robertson, Barrios and Gutiérrez2007; Julià et al. Reference Julià, Morton and Garcia-Del-Pino2023; Londoño-Caicedo et al. Reference Londoño-Caicedo, Uribe-Londoño, Buitrago-Bitar, Cortés and Muñoz-Flórez2023; San-Blas et al. Reference San-Blas, Campos-Herrera, Dolinski, Monteiro, Andaló, Leite, Rodríguez, Morales-Montero, Sáenz-Aponte, Cedano, López-Nunez, Del Valle, Doucet, Lax, Navarro, Báez, Llumiquinga, Ruiz-Vage, Guerra-Moreno and Stock2019). This broad ecological adaptability positions EPNs as versatile and cosmopolitan agents with potential applicability in a wide array of environments. Besides, nematode strains can display large differences regarding their biological traits; thus, the selection of the best EPNs for one pest target entails the search for new isolates (Gulzar et al. Reference Gulzar, Usman, Wakil, Gulcu, Hazir, Karagoz, Hazir and Shapiro-Ilan2020; Campos-Herrera et al. Reference Campos-Herrera, Vicente-Díez, Galeano, Chelka, Del Mar González-Trujillo, Puelles, Labarga, Pou, Calvo and Belda2021). Evaluation of novel isolates must go beyond taxonomic classification and extend to defining their traits such as fitness, infection potential, and environmental stress, as explored in previous studies on EPN strains (Morton and García-del-Pino Reference Morton and García-del-Pino2009; Khashaba et al. Reference Khashaba, Moghaieb, Abd-El-Azim and Ibrahim2020; Aryal et al. Reference Aryal, Nielsen, Sumaya, Wilson and Riegler2022b). Virulence characterisation is fundamental to grasp the potentiality of these newfound isolates and serves to optimise the forthcoming assessments directed at various insect pests.
In the early 1970s, P. japonica reached the Azores Islands from America, firmly establishing itself in the archipelago (Simões et al. Reference Simões, Laumond and Bonifassi1993). To assess the potential for biological control, a nematode sampling was undertaken across the islands with the isolation of several strains of Heterorhabditis bacteriophora (Rhabditida: Heterorhabditidae) and Steinernema carpocapsae (Rhabditida: Steinernematidae) (Rosa et al. Reference Rosa, Bonifassi, Amaral, Lacey, Simões and Laumond2000). Although these strains demonstrated notable virulence against P. japonica larvae only in controlled experiments, they proved ineffective under field conditions (Simões et al. Reference Simões, Laumond and Bonifassi1993). In 2015, the designation of P. japonica as a priority pest in Europe renewed interest in exploring biological control strategies (EPPO 2014). Subsequent investigations confirmed the superior virulence of Heterorhabditis spp. compared to Steinernema spp. against the larval stage of P. japonica under laboratory conditions and semi-field experiments (Renkema and Parent Reference Renkema and Parent2021; Torrini et al. Reference Torrini, Paoli, Mazza, Simoncini, Benvenuti, Strangi, Tarasco, Barzanti, Bosio, Cutino, Roversi and Marianelli2020). The continuous search for new EPN strains adapted to the local environment is crucial for the efficacy of biocontrol measures against insect pests. In this sense, a recent soil survey across the Azorean archipelago was conducted to isolate novel H. bacteriophora strains. The present study focused on the isolation and molecular identification of Azorean native strains of H. bacteriophora, and the evaluation of the virulence and reproductive potential of the strains using Galleria mellonella L. (Lepidoptera: Galleridae) as model organism, with the goal of determining their potential use as biological control agents.
Material and methods
Soil sampling
A total of 164 soil samples were collected from 82 sites in all the islands of the Azorean Archipelago (Portugal) except Santa Maria Island. Ten sites were sampled on each island, with an additional two points on São Miguel Island from spring to autumn 2022. The collection sites were selected across each island in areas infested with P. japonica and two soil samples were collected in each site. Each soil sample weighed approximately 1 kg resulting from three subsamples taken with a garden shovel at a depth of 10–25 cm. Samples were placed in a polyethylene bag to prevent water loss and transported to the laboratory under cool conditions. Before EPN isolation, samples were sieved to remove stones, organic matter, and residues. Landscape features, geographical coordinates, and altitude were all recorded at each point using a GPS. The Azorean Archipelago lies between 0 to 2300 m above sea level, although all collection sites were obtained below 950 m. The prevailing conditions across the sampling area were a mean temperature of 13–24 °C and a mean rainfall of 800–1500 mm, recorded by the official stations. Due to the volcanic origin, the prevailing soil is Androsols with a pH of 5.4–6.
Isolation of entomopathogenic nematodes
EPNs were isolated from soil samples using the insect baiting technique described by Bedding and Akhurst (Reference Bedding and Akhurst1975). Each soil sample was processed by filling two plastic jars (500 mL, 12 cm in diameter) with soil. Four G. mellonella larvae were then placed on the soil surface in each jar. The plastic jars were incubated in darkness at 23 ± 1 °C, sealed, and inverted daily. Moisture levels were adjusted by adding water as needed during the baiting period. After 7 days of incubation, insect larvae were removed from the jars. If no EPNs were detected initially, the assay was repeated. Dead larvae exhibiting signs of nematode infection were individually transferred onto modified White traps (Kaya and Stock Reference Kaya, Stock and Lacey1997) and maintained at 23°C ± 1 °C. The collected IJs underwent several cycles of exposure to fresh G. mellonella larvae to confirm pathogenicity and establish the new culture.
Nematodes and bacteria culture
Once EPNs were isolated, they were maintained as laboratory cultures according to the procedures of Woodring and Kaya (Reference Woodring and Kaya1988). EPNs were reared in vivo using larvae of G. mellonella as previously described and collected in a White trap (White Reference White1927). Nematodes were stored in Ringer solution at 10 °C for 7–14 days. Before any experiment, IJs were acclimated at room temperature, checked for vitality, and adjusted to the required test dose.
The symbiotic bacteria were isolated from nematode-infected G. mellonella, following the protocol described in Rosa et al. (Reference Rosa, Cabral and Simões2002). Larvae were disinfected on the surface, and the haemolymph was streaked onto NBTA (nutrient agar supplemented with 5 g yeast extract, 25 mg bromothymol blue, and 40 mg triphenyl tetrazolium chloride per liter). After 48 hours’ incubation at 27 °C in the dark, the cultures were transferred to 10 °C and subcultured weekly.
DNA extraction
The whole body of IJs was used to extract the DNA from both nematode and symbiont bacteria. Several thousand IJs were frozen at -80 °C and homogenized with a pestle for 30 seconds to break the cuticle. This was followed by a 3-hour incubation at 55 °C with protease K and lysis buffer while maintaining constant agitation at 700 rpm, and a subtle agitation every 30 minutes. After incubation, the DNA was obtained with the extraction kit Zymo Quick-DNA MiniPrep (Zymo Research Corporation) following the manufacturer’s instructions. This process was repeated with three biological replicates for each isolate. The complete genomic DNA was sequenced using universal primers by the Next-Generation Sequencing (NGS) whole-genome sequencing technique via the Illumina platform and based on the mechanism of sequencing by synthesis. The quality of reads was assessed using FastQC and trimmed from adaptors and low-quality nucleotides (Andrews Reference Andrews2010). Then, Burrows-Wheeler Aligner software was used to filter and map the reads to the corresponding Heterorhabditis genome of reference (GCA_000223415) and the Photorhabdus genome of reference (GCF_001083805) (Li and Durbin Reference Li and Durbin2009).
Phylogenetic analysis of nematode and bacteria
From the mapped reads to the genome, manual curation was done to extract the fragment containing the complete 18s - ITS - 28s region (1290 bp) from each nematode isolate. To obtain the reference sequences for the nematode phylogenetic analysis, the fragments of our isolates were submitted to Blastn for similarity and Heterorhabditis sequences were obtained altogether with Oscheius tipulae (Rhabditida: Rhabditidae) as an outgroup. Sequences were aligned using the Multiple Sequence Alignment of the ClustalW software. The phylogenetic tree was calculated with the Tamura-Nei + G model, using maximum likelihood with the MegaX software (Kumar et al. Reference Kumar, Stecher, Li, Knyaz and Tamura2018; Tamura Reference Tamura1992). The robustness of the phylogenetic tree was evaluated by a bootstrap analysis of 10,000 replicates.
For the bacterial isolates, the 16s rRNA region (1545 bp) and the GyrB gene (partial cds 1136 bp) were manually selected from each isolate. Reference sequences of the genera Photorhabdus were obtained from NCBI Reference Sequences following the phylogeny proposed by Tailliez et al. (Reference Tailliez, Laroui, Ginibre, Paule, Pagès and Boemare2010). Escherichia coli was used as an outgroup. Sequences for the two regions were aligned separately using the ClustalW software and then concatenated to be modelled as one. The phylogenetic tree was calculated using the maximum likelihood method with the Kimura 2-parameter + G+I model and evaluated by a bootstrap of 10,000 replicates (Kimura Reference Kimura1980).
Bioassays using G. mellonella larvae
To determine the virulence of the six strains, three different bioassays were performed using the last instar larvae of G. mellonella. All larvae proceeded from our laboratory culture and only healthy insects weighing 240 ± 20 mg were used in the experiments. Different batches of insects were used for each experimental replicate. Throughout the experiments, insect larvae were deemed dead if they exhibited: (i) a lack of movement and unresponsiveness to probing with a dissecting needle, (ii) loss of turgor, or (iii) displayed symptoms characteristic of EPN infection, such as brick-red coloration resulting from Photorhabdus proliferation (Kazimierczak et al. Reference Kazimierczak, Lis, Skrzypek and Kreft2018). Besides, bioassays were performed alongside the native H. bacteriophora strain Az29 as a reference strain due to its well-described virulence (Rosa et al. Reference Rosa, Cabral and Simões2002), thus serving as a positive control in experimental setups.
Determination of the dose–mortality relationship
The determination of the lethal dose (LD50), expressed in terms of IJ cm-2, involved individual exposures of insects subjected to different doses of EPNs in 24-well plates. Each well, lined with two filter papers, housed one G. mellonella larva and the corresponding dose of IJs, administered in 100 μL of sterile tap water. Eight doses were used for each EPN strain corresponding to 5, 10, 20, 40, 80, 160, 320, and 640 IJs per insect larvae (equivalent to 3, 5, 11, 21 44, 88, 168, and 337 IJ cm-2). Employing a randomised block design, plates were sealed with Parafilm and incubated in darkness at 23 ± 1 °C. Insect mortality was checked at 24, 48, 72, and 96 hours of exposure, based on examination of the insects´ condition as previously described. The assay consisted of four G. mellonella larvae per nematode strain and dose and was replicated twice. Negative control using sterile tap water instead of IJs was conducted with the same number of replicates and insect larvae.
Determination of the lethal time at fixed dose
The lethal time (LT50) was determined for each EPN strain by administering a fixed dose of IJs to G. mellonella. The assay was conducted in 24-well plates filled with two filter papers where insect larvae were individually exposed to 200 IJs suspended in 100 μL solution. Plates were sealed with Parafilm and placed in the dark at 23 ± 1 °C. Twelve hours after the application of IJs, we performed larval mortality checks every 2 hours until >80% of insect larval mortality was achieved in all experimental treatments. The experiment was conducted twice, with each repetition involving 24 G. mellonella larvae per EPN strain. Negative controls, comprising 24 G. mellonella larvae, were subjected to the same manipulation but received sterile tap water instead of IJs.
Short exposure assay
To test the ability of each EPN strain to infect in a short exposure time, we assess the G. mellonella larvae mortality after three exposure periods time: 20, 60, and 180 minutes. The infection was performed as described previously, individualising insect larvae in 24-well plates filled with two filter papers. Each well received a dose of 211 IJ cm-2 suspended in 100 μL of sterile tap water. After the designated exposure period, each insect larva was carefully removed from the test well and rinsed with diluted 70% ethanol and subsequently with sterilised distilled water to remove any attached nematodes. Rinsed larvae were then incubated at 23 ± 1 °C for an additional 48 hours to monitor the mortality induced by nematodes. Negative controls, comprising the same number of G. mellonella larvae per exposure time were subjected to the same manipulation but without exposure to IJs. A total of six insect larvae per EPN strain and exposure time was used and repeated twice.
Reproductive potential
An assessment to determine the reproductive potential of the EPN strains was conducted using G. mellonella larvae following the established infection methodology and dose for rearing EPNs (Bedding and Akhurst Reference Bedding and Akhurst1975; Woodring and Kaya Reference Woodring and Kaya1988). Dead insect larvae were transferred individually to small White traps made with 5-mm Petri dishes. Emerging nematodes from the insect larvae cadaver were collected daily and stored in Ringer solution at 10 °C. After 1 week without further collection, the total number of EPN offspring was counted through serial dilution. There were six G. mellonella larvae replicates per nematode strain and the experiment was done twice.
Statistical analysis
Survival data were fitted to a four-parameter logistic curve to determine LT50 and LD50 values. The estimated LD50 values at a particular time point and the LT50 at a fixed dose were compared between each pair of nematode strains using the extra sum-of-squares F test. For the short exposition and reproduction potential tests, the Kruskal-Wallis test was applied to detect significative differences among EPN strains, when the positive Dunn test was used to determine pairwise differences. Statistical analyses were conducted using GraphPad Prism 10 for Mac and RStudio, and the significance level was set at 0.05 (R Core Team 2017).
Results
Soil sampling
EPNs were recovered from 18 of the 164 sites sampled in the islands, accounting for 10.98% of positive samples (Fig 1). Among the isolated EPNs, we selected strains of Heterorhabditis sp. because of their significant role in the biological control of P. japonica. This particular nematode species was isolated from six different sites (3.7% of the total) across the islands of São Miguel, São Jorge, and Faial (Table 1). Besides, positive samples from the Islands of Terceira, Flores, and Graciosa contained unidentified EPNs. EPNs were found at various locations, ranging from sea level to 400 m of altitude, and spread across diverse landscapes. Three strains were isolated from pastures, the most common landscape type of the islands, characterised by loamy soils and grasslands. Two strains (Az170 and Az180) were obtained from a coastal landscape dominated by sandy soil and little vegetation. In São Jorge, a single positive sample was obtained from an area outside an urban zone composed of shrubs. All isolated strains of H. bacteriophora were maintained in continued culture at our facilities at 10 °C (Fig. S1).
Geographical distribution of the sampling sites (x) in the Azores Islands with the identification of the positive samples for (▲) Steinernema sp. or (●) H. bacteriphora.
Localisation of the isolated strains of Heterorhabditis sp. with GPS coordinates, altitude (in meters), and landscape
Phylogenetical identification of EPNs
The six new isolates were phylogenetically identified with the regions 18s-ITS-28s for nematodes and with the concatenated 16s rRNA and GyrB gene for symbiotic bacteria. All nematode fragments presented over 99% similarity to H. bacteriophora, with a close phylogenetical relationship to Heterorhabditis georgiana (Rhabditida: Heterorhabditidae) (Fig. 2). Overall, we can firmly affirm that the new strains belong to H. bacteriophora with strong bootstrap support. Sequences for the isolated H. bacteriophora strains can be found in the following accession numbers of GenBank (PP738128, PP738129, PP738130, PP738131, PP738132, PP738133). Concerning the symbiont bacteria, the six isolates had over 97% similarity with P. luminescens. The phylogenetic analysis separated Az171 and Az186 bacteria with reliability (bootstrap value = 45) from the main node that includes Az180, Az172, and Az179 (Fig. 3). Az170 was clearly separated from the other isolates. Besides the minor differences pointed out by the analysis among bacterial strains, the phylogenetic tree located robustly all samples within the branch of Photorhabdus luminescens, specifically the subspecies laumondii. Sequences for the isolated Photorhabdus can be found in the following accession numbers of GenBank for 16s rRNA (PP738150, PP738151, PP738152, PP738153, PP738154, PP738155) and for GyrB gene (PP746494, PP746495, PP746496, PP746497, PP746498, PP746499).
Phylogenetic tree using maximum likelihood method of EPN isolates and other Heterorhabditis spp. based on the analysis of the region 18s-ITS-28s, with O. tipulae as outgroup. Bootstrap values are indicated in each node after 10,000 replicates. GenBank accession numbers are given in front of each entry. Scale bar represents the number of substitutions per site.
Phylogenetic tree using maximum likelihood method of bacteria isolates and other Photorhabdus spp. based on the analysis of the 16s rRNA and GyrB genes, with E. coli as outgroup. Bootstrap values are indicated in each node after 10,000 replicates. GenBank accession numbers are given in front of each entry. Scale bar represents the number of substitutions per site.
Determination of the dose–mortality relationship
The bioassays included the comparison of the new strains alongside the Azorean reference Az29. The determination of LD50 revealed notable differences at the 72- and 96-hour intervals, serving as the initial metric to determine the potentiality of EPNs (Table 2). Az29 exhibited the minor LD50 at 48 hours with only 17.96 IJs and maintained the dose effect consistently over time. Among the new strains, Az171 showed the best performance against G. mellonella, which achieved LD50 values from 25 to 6.9 IJs cm-2. Az172 and Az186 followed closely, showing no significant differences with Az171. Remarkably, these three strains exhibited an enhanced dose response effect after 72 hours, with marginal progress noted at 96 hours. In contrast, the strains Az179, Az170, and Az180 needed a higher dose to kill G. mellonella larvae, although these strains improved greatly the dose response when the experiment was extended to 96 hours. Especially, Az179 was the most favoured, presenting no differences with the top virulent strains at 96 hours. Therefore, Az170 and Az180 were the least virulent strains regarding the dose response.
LD50 and 95% confidence interval measured as IJs cm-2 for the three time points of the new strains alongside the reference Az29; the assignment of letter groups (a-c) is based on statistically significant differences among the strains
Determination of the lethal time
The virulence of strains was also measured by the time of death of G. mellonella while strains exhibited significant differences among them, no mortality was registered in negative control larvae (Table 3). The reference strain Az29 exhibited the highest virulence of the experiment with an LT50 of 26.7 hours. Regarding the performance of the new strains, we split them into two main groups. One was composed by Az172, Az186, Az171, and Az170, which presented an intermediate LT50 from 35 to 39 hours. Az170 was the lowest among them, differing significantly from Az172 and Az186. The second group comprised strains with slower virulence, specifically Az179 and Az180, exhibiting LT50 values exceeding 40 hours. Particularly, there were statistically significant differences in the virulence between these strains manifesting the notably delayed virulence of Az180 compared to Az179.
LT50 and 95% confidence interval measured in hours of the new strains alongside the reference Az29; ahe assignment of letter groups (a-f) is based on statistically significant differences among the strains
Short exposure assay
In the exposure assay, we observed a positive correlation between exposure time and G. mellonella mortality. All H. bacteriophora strains caused a larval mortality below 30% when exposition lasted only 20 minutes, pointing to no differences among them. In contrast, from 60 minutes of exposure to 180 minutes, strains presented significant differences regarding their virulence (Fig 4). Mortality of G. mellonella increased rapidly to 50% with the 60-minute treatment of Az29, Az172, and Az186. The highest mortality rate was achieved by Az186, reaching 92% mortality at 180 minutes of exposure. Moreover, Az29, Az172, and Az171 reached 75% G. mellonella mortality at 180 minutes, whereas Az170 only killed 58% of larvae. Az180 and Az179 exhibited the lowest virulence in these experiments, reaching a maximum of 35% mortality even at the longest exposure time. As was observed in the LD50 test, the infectivity performance of Az179 was dependent on a long exposition time to achieve a higher mortality rate on larvae. Besides, no mortality was registered in control larvae during the experiment.
Mean mortality percentage of G. mellonella larvae obtained by the EPN strains after 60 minutes of exposure (A) and 180 minutes (B). The assignment of letter groups (a-c) is based on statistically significant differences among the strains.
Reproductive potential
The reproductive potential of EPNs differed greatly between larvae, highlighting the influence of various factors on offspring production, even in model insects like G. mellonella (Fig. 5). However, we observed significantly higher reproduction of the Az179 strain producing a mean of over 400,000 IJs per larva (equivalent to 1701 IJs/mg of larvae). In contrast, Az172 and Az186 presented the lowest production, with only 480 and 778 IJs/mg, respectively. Strains Az29, Az171, Az170, and Az180 exhibited intermediate reproductive potential with a mean offspring of 1050 IJs/mg. Particularly, Az180 showed a wider range of offspring production from 364 to 2364 IJs/mg.
Violin plot showing the reproductive potential expressed as the total number of IJs produced per larvae of the new strains alongside the reference Az29. The assignment of letter groups (a-b) is based on statistically significant differences among the strains.
Discussion
The isolation of new EPN strains enlarges the chances of efficient biological control methods while helping to preserve the native biodiversity and ensure adaptability to regional conditions. In our survey across eight islands of the Archipelago of Azores, the total recovered EPNs represented 10.9% of the samples, which aligns with the average reported nematode recovery rates from cultivated and natural spots, including 11.8% in Oregon (Liu and Berry Reference Liu and Berry1995), 5% in Spain (Campos-Herrera et al. Reference Campos-Herrera, Escuer, Labrador, Robertson, Barrios and Gutiérrez2007; Julià et al. Reference Julià, Morton and Garcia-Del-Pino2023), 6.6% in Italy (Tarasco et al. Reference Tarasco, Clausi, Rappazzo, Panzavolta, Curto, Sorino, Oreste, Longo, Leone, Tiberi, Vinciguerra and Triggiani2015), or 6.8% in Hawaiian Islands (Hara et al. Reference Hara, Gaugler, Kaya and Le-Beck1991). From the total of our isolates, 7.3% were unidentified steinernematid species, consistent with the trend of higher steinernematid recovery compared to heterorhabditids (Hominick Reference Hominick and Gaugler2002). The observed prevalence of H. bacteriophora in Azorean soils was similar to the previous occurrence study that showed a range from 1.2 to 4.6 positive samples depending on the island (Rosa et al. Reference Rosa, Bonifassi, Amaral, Lacey, Simões and Laumond2000). Because of the archipelago’s nature, H. bacteriophora was found across islands, as it is commonly found in maritime environments (Emelianoff et al. Reference Emelianoff, Le Brun, Pagès, Stock, Tailliez, Moulia and Sicard2008). In agreement with the previous survey, São Miguel, Faial, and São Jorge exhibited the highest occurrence of H. bacteriophora predominantly in pastures and coastal landscapes at lower altitudes.
The molecular identification revealed all novel strains as H. bacteriophora as it was suggested by the reddish colour during the infection of G. mellonella larvae. In particular, the tight relation between H. bacteriophora and H. georgiana observed in the analysis was already pointed out in other phylogenetic studies producing often unreliable nodes because of their proximity (Aryal et al. Reference Aryal, Nielsen, Sumaya, De Faveri, Wilson and Riegler2022a). For bacterial identification, phylogeny performed only with 16s rRNA gene often produces inconclusive results. Thus, we selected the GyrB gene to improve the analysis as suggested for the phylogenetical classification of Photorhabdus species (Shapiro-Ilan et al. Reference Shapiro-Ilan, Mbata, Nguyen, Peat, Blackburn and Adams2009; Tailliez et al. Reference Tailliez, Laroui, Ginibre, Paule, Pagès and Boemare2010). With this procedure, the new isolates were robustly identified as Photorhabdus luminiscens subsp. laumondii. Interestingly, H. bacteriophora Az29 and Az186 were isolated from practically the same area but 30 years apart. However, Az29 harbours the bacterium P. temperata subsp. thracensis, distinct from the symbiont associated with Az186.
Recently isolated EPN strains can exhibit novel biological traits such as increased virulence, broad host range, or abundant offspring generation. Thus, it is crucial to evaluate through bioassays their abilities first under laboratory conditions. Virulence of the novel strains was evaluated with the tests of LD50, LT50, and short exposure, altogether with the assessment of the reproductive potential. The isolates were able to infect G. mellonella at high rates with the bioassays, although differences in their performance were observed. Therefore, we classified the strains from high to low following the measured parameters (Table 4). Az29 exhibited the highest virulence in the current experiments because it occurred when the EPN was first isolated (Rosa and Simões Reference Rosa and Simões2003). The high virulence of Az29 was followed by the new strains Az172, Az186, and Az171. Similar LD50 values were obtained by the Egyptian isolates of H. bacteriophora, whereas lower LD50 (2 to 6 IJs/larvae of G. mellonella) were achieved by the Thai EPNs isolates (Noosidum et al. Reference Noosidum, Hodson, Lewis and Chandrapatva2010; Saleh et al. Reference Saleh, Metwally, El-Behery, Moawad and Mahmoud2016). The LD50 results obtained with Az179 suggest that providing more exposure time or extended contact with IJs could reach high virulence. In contrast, when the dose was limited Az170 strain exhibited the worst performance. The strains displaying the highest virulence as determined by the LT50 were Az29, Az172, and Az186 consistently across all bioassays. The obtained LT50 values with H. bacteriophora strains were comparable to the obtained with Steinernema feltiae (Rhabditida: Steinernematidae) and S. carpocapsae strains against G. mellonella ranging from 12 to 48 hours (Campos-Herrera et al. Reference Campos-Herrera, Escuer, Labrador, Robertson, Barrios and Gutiérrez2007; Caroli et al. Reference Caroli, Glazer and Gaugler1996). The capacity of EPNs to infect in short exposure periods is essential for efficient biological control of pests, with Heterorhabitis species showing the ability to infect diverse insects in short time experiments (Corrie et al. Reference Corrie, Grewal and Taylor2004; van Niekerk and Malan Reference van Niekerk and Malan2012). Notably, Az186 induced higher mortalities within short exposure times and Az170 demonstrated superior performance in these scenarios achieving almost 60% of mortality. Nevertheless, Az179 and Az180 suffered the most when exposure time was limited. The low general virulence exposed by the novel strain Az180 suggests the potential unsuitability of G. mellonella as its host.
Classification of the new H. bacteriophora strains performance from high to low regarding their virulence with the LD50, LT50, and short exposure (Expo.) test and the reproductive potential
Considering the weight of the G. mellonella larvae used, our strains produced between 1600 and 420 IJs per mg, similar to values obtained with strains of H. bacteriophora and S. carpocapsae (Campos-Herrera et al. Reference Campos-Herrera, Escuer, Labrador, Robertson, Barrios and Gutiérrez2007; Nouh Reference Nouh2016; Wang and Bedding Reference Wang and Bedding1996). The reproductive potential differs greatly among larvae, although our results suggest a correlation between IJs production and virulence. Because the most virulent strains Az186 and Az172 exhibited the lowest reproduction while the less virulent Az179 and Az180 achieved a production almost three times higher. This observation could hint at a specialised adaptation of the less virulent EPNs to increase their persistence with an enhanced reproductive capacity. Although this association has not been firmly elucidated, it is established that the reproductive potential of EPNs is liable to insect size and the number of IJs that enter the host (Bastidas et al. Reference Bastidas, Portillo and San-Blas2014; Caroli et al. Reference Caroli, Glazer and Gaugler1996).
These isolates have the potential to become biological control agents for P. japonica as EPNs isolated from areas already infested are proposed to exhibit heightened virulence against the pest (Lacey et al. Reference Lacey, Grzywacz, Shapiro-Ilan, Frutos, Brownbridge and Goettel2015; Glazer et al. Reference Glazer, Santoiemma, Battisti, De Luca, Fanelli, Troccoli, Tarasco, Sacchi, Bianchi, Gilloki, Gherardo and Mori2022). Besides, these strains could display superior adaptability to the environmental and soil conditions for field applications. Native strains of H. bacteriophora from Italy exhibited an increased virulence in field trials proving the importance of using regional isolates for the integrated pest management strategies of P. japonica (Torrini et al. Reference Torrini, Paoli, Mazza, Simoncini, Benvenuti, Strangi, Tarasco, Barzanti, Bosio, Cutino, Roversi and Marianelli2020). After this first assessment with G. mellonella, Azorean native strains showing high virulence traits will be tested against the pest aligning with the ongoing efforts to develop effective management strategies for P. japonica.
In conclusion, we isolated six native strains of H. bacteriophora with diverse virulence performances. With the bioassays, we revealed the potentiality of Az172, Az186, and Az171 whose performance is close to the well-known virulent Az29. Meanwhile, Az179 can be an interesting strain to use if an extended period of exposure can be assured because their reproductive potential could favour an enhanced persistence on field applications. Concerning Az180, the observed potentiality was rather low though future bioassays against other insects would clarify if the low virulence is related to the host suitability or a trait of the strain. Further experiments will be carried out to elucidate the efficacy of these strains against P. japonica as well as the behaviour and persistence of these EPNs under field conditions. Nevertheless, this preliminary evaluation serves as a foundation for directing further investigations aimed at assessing the potential of the novel Azorean strains for the control of P. japonica or other insect pests.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0022149X24000336.
Acknowledgements
We thank the Serviços de Desenvolvimento Agrário dos Açores from the different islands for their cooperation in fieldwork and sampling. Besides, we thank the colleagues from the UAç-CBA laboratory that carried out the sampling as well: Jorge Frias, Mario Teixeira, Marina Oliveira, and Ángel Peñalver.
Author contribution
Conceptualisation and supervision: N.S. and A.G.; Investigation: R.B., H.M., and A.G.; Formal analysis: H.M.; Writing – original draft: R.B. and A.G.; Writing – review and editing: H.M., D.T., and N.S; Funding acquisition: N.S., D.T., and A.G. All authors read and approved the manuscript.
Financial support
This work was supported by the International Research Project-Integrated Pest Management of the invasive Japanese Beetle, Popillia japonica (IPM-Popillia; grant Nr. H2020-EU.3.2.1.1/ID: 861852); the Fundação Ciência e Tecnologia (FCT) with the project AGWAS-EPNs (2022.06153.PTDC). R.B. was hired by the IPM-Popillia project as a research assistant. A.G. received a postdoctoral grant from the M. Universidades—Margarita Salas. H.M. received the support of the FCT Pluriannual Program for the Funding of R&D Units (UIDB/05292/2020 and UIDP/05292/2020) granted to CBA.
Competing interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Data availability
All sequences used for the phylogenetical identification can be found in the GenBank NCBI Database; see the Results section for the corresponding accession numbers.
Open access
