Introduction

Entomopathogenic nematodes (EPNs), which belong to the genus Steinernema (Nematoda: Steinernematidae), are associated with bacteria from the genus Xenorhabdus (Enterobacteriacae) (Goodrich-Blair and Clarke 2007). The life cycle of insecticidal nematodes appears simple. The non-feeding third-stage juvenile of the nematode is the infective juvenile (IJ). It searches for suitable insect hosts and, when a host has been located, the IJs penetrate the hemocoel by mechanical and enzymatic means (Abu Hatab et al. 1995). Shortly after entry, the nematode releases the symbiotic bacterium. The nematode and its bacterial partner work together to overcome the immune system of susceptible insects (Goodrich-Blair and Clarke 2007), quickly causing host death. The nematodes feed upon the rapidly multiplying bacteria, mature, mate, and produce two or more generations within the insect cadaver before emerging as IJs in search of fresh hosts. The attributes of nematodes as biological insecticides include ease of mass culture, ease of application, high lethality to diverse insect pests and safety (Ehlers 2003; Grewal et al. 2005). Consequently, nematodes are produced commercially on four continents, and are being successfully applied against important agricultural pests, such as thrips and root and trunk borers (Grewal et al. 2005).

The natural habitat of entomopathogenic steinernematid nematodes is the soil. Steinernematids are well adapted for survival under the various moisture, temperature, texture and chemical conditions of different soils. Consequently, they are found on all continents except Antarctica, and at nearly all latitudes and altitudes (Hominick 2002). The sensitivity of nematodes to extreme environmental conditions prevents them from reaching their full potential as biocontrol agents. The intolerance of the infective stages to desiccation (Glazer 1992, 2002), extreme temperatures (Ehlers et al. 2005; Shapiro-Ilan et al. 2006) and solar radiation (Gaugler and Boush 1978) has contributed to generally erratic results in the field (Georgis et al. 2006). The sensitivity of EPNs to environmental conditions has discouraged their application onto exposed surfaces, such as foliage. When they have been applied in these settings, results have been inconsistent (Gaugler 1981; Kaya 1985). Successful applications have been restricted to habitats with favorable microenvironments, such as soil and cryptic habitats (i.e., locations sheltered from desiccation and solar radiation) (Georgis et al. 2006).

In order to overcome these limitations, several physical and biological approaches have been suggested. There have been several attempts to develop alternative application formulations that would provide protection for the nematodes (Grewal 2002). Alginate capsules (Kaya et al. 1987; Kaya and Nelsen 1985), baits (Capinera and Hibbard 1987; MacVean et al. 1982) and dehydrated preparations (Connick et al. 1993) have been evaluated. However, the efficacy of these formulations has been limited and inconsistent, varying with the prevailing environmental conditions (Georgis et al. 2006; Shapiro-Ilan et al. 2006).

The genetic improvement of EPNs has been suggested and utilized by various experts over the years (Burnell 2002; Segal and Glazer 1998). Genetic improvement of EPNs for heat and cold tolerance (Ehlers et al. 2005; Grewal et al. 1996), nematicide resistance (Glazer et al. 1997) or host-seeking behavior (Gaugler et al. 1989a) has been achieved mainly by repeated selection under defined regimes. Cross-hybridization (Shapiro et al. 1997) and gene transfer (Hashmi et al. 1995) have also been shown to enhance heat tolerance.

In the present study, we genetically improved a heterogeneous population of the EPN Steinernema feltiae Filipjev for desiccation tolerance (both rapid and gradual) and host-seeking behavior. Both traits were chosen to enhance the nematode’s ability to overcome environmental conditions and improve its efficacy as a biocontrol agent. In a recent survey conducted in Israel, several populations of S. feltiae were isolated and their desiccation tolerances and downward dispersal abilities in sand columns were characterized (Salame et al. 2010). We used these populations and the information obtained previously about their environmental tolerances and host-seeking abilities as the basis for the establishment of a heterogeneous study population and genetic selection.

Materials and methods

Establishment of a heterogeneous foundation population (HFP) of S. feltiae

The strains of S. feltiae used to establish the HFP were isolated in different locations in Israel, as reported recently by Salame et al. (2010). These diverse sites are described in Table 1. To produce the heterogeneous foundation, the strains were cross-bred following the round-robin design scheme used by Gaugler et al. (1989b), in the order described in Table 2. Females from one strain and males from another strain were combined on agar media, pre-seeded with the symbiotic bacterium Xenorhabdus nematophilus, in multi-well plates (12 wells per plate; Corning, Corning, NY, USA) (Iraki et al. 2000). These wells were filled with 0.5 ml of an agar medium containing the following components (per liter of the medium): 12 g agar (Difco, Detroit, MI, USA), 10 g trypcase soy (API BioMerieux, Marcy l’Etoile, France), 5 g yeast (Merck, Darmstadt, Germany), 5 g nutrient broth (Difco), 1 g NaCl, 0.5 g MgSO4 × 6 H2O, 0.2 g CaCl2 and 0.5% sunflower oil. The medium was adjusted to pH 7.0 with NaOH. Each well was inoculated with 60 μl of a 24-h X. nematophilus culture (that originated from the female nematode strain) in YS broth (Dye 1968). Since S. feltiae reproduces amphimictically, it is easy to identify 4th stage juveniles that will develop into females or males on 5-cm-diam culture plates. In each well, 25–30 individual females from one population were combined with a similar number of males from another population (Table 2). The plates were incubated at 24°C in the dark for 10–12 days. The new, crossed populations were mixed and allowed to mate freely within G. mellonella larvae to produce the foundation strain (Gaugler et al. 1989b).

Table 1 Locations and characteristics of the sites at which new population of Steinernema feltiae were isolated. These populations were employed to establish the heterogeneous foundation population used in the present study
Full size table
Table 2 Crosses between different Steinernema feltiae strains for establisment of the heterogeneous foundation population for further genetic selection
Full size table

Further propagation of the HFP, as well as selected populations, was done on last-instar G. mellonella larvae, using the procedure described by Kaya and Stock (1997). For genetic selection, we used IJs that had been stored for 2–3 weeks in a water suspension kept at 8°C.

Genetic selection

In general the selection procedure was done by exposure of IJs from the HFP to the different selection regimes (see below). The effects of the treatments were recorded and the surviving (in the stress assays) or selected (in the sand-columns assay) IJs were propagated for further selection. Following a number of selection cycles (at least 20), the new, selected population was subjected to a series of bioassays (see below) for determination of fitness loss.

Selection for tolerance of rapid desiccation

The IJs from the HFP were concentrated by vacuum filtration onto ten 5-cm-diam discs of filter paper (Whatmann No. 1) at a density of 10,000 nematodes per disc. The discs were then exposed to ambient conditions (22–25°C; 55–65% r.h.) for 100 min before being re-hydrated in distilled water. Twenty-four hours later, the viability of the nematodes was determined by observing their motility and their responses to probing under a stereomicroscope. Six hundred individuals were examined. Preliminary tests indicated that these conditions led to a 3–7% survival rate among IJs of the HFP. We separated the surviving IJs from the dead ones by allowing them to migrate through a 30-mesh plastic sieve to a water suspension. They were then re-exposed to 30 G. mellonella larvae for propagation on moist 5-cm-diam filter paper in three plastic petri dishes (ten insects per dish) and incubated for 3 days at 25°C, in the dark. The dead insects were transferred to ‘White traps’ and further incubated for 12 days. The emerging IJs which were collected in ‘White traps’ were re-subjected to the desiccation regime. Each exposure cycle consisted of six replicates. Following a dramatic increase in nematode viability after the fourth selection cycle, the exposure time was extended to 140 min.

Selection for tolerance of slow desiccation

Infective juveniles from the HFP were concentrated by vacuum filtration onto 5-cm-diam discs of filter paper, as described above. The discs were then transferred to desiccators that were kept at 25°C and 97% r.h. The humidity level was maintained through the use of a saturated potassium sulfate solution (Solomon et al. 1999). Following 72 h of exposure to these conditions, the nematodes entered an anhydrobiotic state (Solomon et al. 1999) caused by the slow removal of water from their bodies. After this period, the IJs were further incubated in the desiccators at 85% r.h. (potassium chloride saturated solution; Solomon et al. 1999) for an additional 72 h. The nematodes were then re-hydrated by direct immersion in distilled water. Their viability was determined after 24 h and the survivors were re-exposed to G. mellonella larvae for further propagation as described above. Each exposure cycle consisted of six replicates.

Selection for enhanced host-seeking ability

The IJs from the HFP were allowed to move through the soil and locate a target host in a sand column assay (Salame et al. 2010). Plastic pipes (5 cm diam, 21 cm in height) were filled with moistened sandy soil (8% moisture w/w). The IJs were applied to the tops of the columns in 1 ml water, at a concentration of 4000 IJs ml−1. Four last-instars of G. mellonella were placed at the bottom of the column. The columns were incubated in the dark at 25°C for 72 h. After the incubation period, the soil was pushed out of the column and divided into three layers: upper, middle and lower (see Fig. 1). The nematodes were extracted from each of the soil samples overnight using Baermann funnels and the number of nematodes recovered from each layer was recorded. The dead insect larvae from each column were washed in water and incubated further for the propagation of the nematodes. Each cycle included six columns. After nine cycles the number of IJs applied to the top of the column was reduced to 2000 and the incubation time was shortened to 48 h. That is due to a substantial increase in the number of nematodes found in the bottom layer of the column.

Fig. 1
figure 1

Changes in survival of infective juveniles (IJs) of the heterogeneous foundation population (HFP) of Steinernema feltiae, which was subjected to selection for rapid-desiccation tolerance. The initial four selection cycles, the IJs were exposed to ambient conditions (22–25°C; 55–65% r.h.) for 100 min before being re-hydrated in distilled water (black columns). Due to rapid increase in viability the selection was intensified at the fifth cycle to 140 min exposure (gray columns). Y error bars represent the standard errors of the means and different letters above the different columns indicate significantly different means (P < 0.05) (capital letter for the initial selection)

Full size image

A control treatment, which consisted of non-selected IJs from the HFP, was included in all assays and selection cycles. In each assay and selection cycle, there were five replicates of the control treatment.

Bioassays for fitness determination

The progeny of the 15th selection cycle for tolerance of rapid desiccation and the 20th cycle of the other two selections were subjected to a series of bioassays to detect any loss of fitness due to the selection regimes.

Infectivity

Infectivity was evaluated in two separate bioassays: an invasion-rate assay and a dose-response assay. Both assays are described in detail by Glazer and Lewis (1998). Briefly, in the invasion rate assay, 500 IJs from the selected populations were exposed to five last-instar G. mellonella larvae in a 5-cm-diam petri dish lined with moist filter paper. After 24 h of exposure at 25°C, the insects were washed in tap water to remove any nematodes from the surface of their bodies, and then incubated further in nematode-free dishes for 48 h. The insects were then dissected under a stereomicroscope and the numbers of invading nematodes were determined. Each treatment consisted of five replicates (1dish = 1 replicate) and the experiment was repeated twice.

In the dose-response assay, five instars of the mealworm Tenebrio molitor, at the 5th–6th stage, were exposed to 0, 125, 250, 500 or 1000 IJs from the selected populations in 5-cm-diam petri dishes lined with moist filter paper. The dishes were incubated at 25°C for 96 h, after which insect viability was determined. As in the infectivity assay, each treatment consisted of five replicates (1 dish = 1 replicate) and the experiment was repeated twice.

Heat tolerance

Infective juveniles from the selected populations were suspended in 15 ml distilled water, in 25-ml conical flasks at a concentration of 5000 IJs per flask. The flasks were shaken gently at 40 rpm for 6 h in a water bath kept at 37°C. Three 1-ml samples were then withdrawn from each flask and diluted in 10 ml water at room temperature (22–25°C) in petri dishes. Nematode viability was recorded for each population by observing nematode motility and response to probing under a stereomicroscope 24 h later.

Reproductive potential

Ten last-instars of G. mellonella larvae, with an average weight of 0.7 ± 0.06 g, were exposed to 500 IJs of the selected populations in 5-cm-diam petri dishes lined with moist filter paper. Following 72 h of incubation at 25°C, the dead insects were placed on White traps and incubated further for 14 days. The IJs that emerged from the cadavers were collected daily after the 8th day. The total IJs harvested from each treatment/population were then counted.

All of the bioassays described above included a control treatment, which consisted of non-selected IJs from the HFP. The treatments in each of the assays consisted of five replicates. Each assay was repeated twice.

Statistical analysis

The survival data presented in percent were arcsin transformed and analyzed by ANOVA. Differences among treatments were estimated by contrasts using Tukey’s HSD test (P ≤ 0.05). All of the statistical analyses were conducted using JMP software (SAS Institute 2009). In the dose-response assay, LD50 values were calculated for each nematode population using a probit analysis (SAS Institute 2009).

Results

Rapid desiccation

The change in the viability of the HFP following repeated exposure of the IJs to a rapid desiccation regime is presented in Figure 1. A rapid increase in the survival rate was observed in the initial selection process. Within three to four selection cycles under mild conditions (100 min exposure), IJ viability increased fivefold, from 5 ± 1.3% to 25 ± 2.7% (Fig. 1). Following this finding, the selection regime was intensified to 140 min of exposure. Under such conditions, nematode tolerance of rapid desiccation was further increased and reached 80–90% after the 10th selection cycle. After an additional five selection cycles, their viability level had increased by an additional 10% (Fig. 1).

Slow desiccation

A rapid increase in the survival of IJs exposed to the slow desiccation regime was detected in the first three selection cycles (Fig. 2). This was followed by a constant survival rate for the next 13 selection cycles, ranging between 40% and 47%. There was a further increase in the survival rate over the last four selection cycles, for a final survival rate of 81% (±2.4%) (Fig. 2).

Fig. 2
figure 2

Changes in the survival of infective juveniles (IJs) of the heterogeneous foundation population (HFP) of Steinernema feltiae, which was subjected to selection for slow-desiccation tolerance. In each selection cycle, the IJs were exposed to 97% r.h. for 72 h, and then exposed to 85% r.h. for an additional 72 h. Y error bars represent the standard errors of the means and different letters above the different columns indicate significantly different means (P < 0.05)

Full size image

Downward dispersal

There was a gradual increase in the proportion of IJs found in the lower level of the column, in proximity to the insects (Fig. 3). After nine selection cycles, the proportion of IJs found in the lower layer increased by 10%. Following further selection downward dispersal increased and reached 75% (±3.7) of the population after 25 selection cycles. The proportion of nematodes found in the middle layer was reduced 2.3-fold following 25 selection cycles. The greatest reduction in proportional population, 3.3-fold, was observed in the upper layer.

Fig. 3
figure 3

Distribution of infective juveniles of the heterogeneous foundation population (HFP) of Steinernema feltiae, which was subjected to selection for downward dispersal in sand columns. Four last-instar Galleria mellonella larvae were placed at the bottom of each column. Y error bars represent the standard errors of each mean (P < 0.05). Arrow indicates when selection regime was intensified

Full size image

Fitness determination

Evaluations of the performance of the selected lines in the different assays indicated that none of these lines lost any of their ability to infect or survive stressful conditions (Table 3). Nevertheless, the nematode population selected for enhanced host-seeking behavior displayed significantly higher infectivity (P = 0.05) in the invasion-rate assay and in terms of LD50 values for T. molitor, as compared to the HFP. The population selected for its ability to withstand slow desiccation demonstrated enhanced tolerance of heat stress (Table 3).

Table 3 The results (± SEM) from bioassays conducted to evaluate the fitness of the selected populations following 15–20 selection cycles under the different regimes. The underlined data are significantly different (P < 0.05) from heterogeneous foundation population (HEP) performance in the various assays
Full size table

Discussion

Several studies have demonstrated the use of repeated selection genetic improvement regimes to enhance the host-finding ability of EPNs (Gaugler et al. 1989a), as well as their ability to withstand exposure to nematicides (Glazer et al. 1997) and extreme temperatures (Ehlers et al. 2005; Grewal et al. 1996). In the present study, we successfully enhanced desiccation tolerance and host-seeking ability. In a previous study, Salame et al. (2010) characterized the desiccation tolerance and downward dispersal of the populations that were used to establish the HFP. For all of the selected traits, the selected populations performed better than the best-performing steinernematid populations in the previous study (Salame et al. 2010). This demonstrates the value of genetic selection for improving the beneficial traits of EPN.

The fastest change in the selected population was recorded in the rapid desiccation regime (Fig. 1). The differential ability of EPN strains and species to withstand rapid desiccation has been attributed to variation in the thickness and water permeability of their cuticles (Patel et al. 1997; Patel and Wright 1998; Wright 1987). Further research is required to identify the differences between the HFP and the selected population that may be responsible for the observed differences in performance.

On exposed surfaces (soil or foliage), steinernematids can survive no longer than several hours, depending on species, temperature and relative humidity (Glazer 1992, 2002). The rapid removal of water from their bodies drastically reduces their viability, substantially decreasing their efficacy as biocontrol agents (Georgis et al. 2006; Shapiro-Ilan et al. 2006). Therefore, enhanced tolerance of conditions that would normally lead to rapid desiccation may increase nematode survival on exposed surfaces and, consequently, their efficacy as biocontrol agents.

Nematode efficacy is also dependent on their persistence in the soil for extended periods (Shapiro-Ilan et al. 2006; Susurluk and Ehlers 2008). In dry soil, EPNs can persist for 2–3 weeks (Kaya 1990; Kung and Gaugler 1990). Like several other animal and plant-parasitic nematodes, EPNs can survive exposure to desiccating conditions for long periods (Cooper and Van Gundy 1971; Glazer 2002). These nematodes are capable of reaching a dormant state of anhydrobiosis, which is usually attained following a period of gradual water loss (Crowe et al. 1992). Unlike rapid desiccation, survival under a slow-desiccation regime requires the physiological adaptation of the nematode. The biochemical mechanisms involved in the induction of anhydrobiosis are not fully understood (Glazer 2002; Zitman-Gal et al. 2005). One biochemical change that has been reported in anhydrobiotic nematodes is the accumulation of polyols and sugars, which are believed to act as protectants of biological membranes and intracellular proteins during dehydration. Further gene expression and molecular changes due to exposure to slow-desiccation conditions were described recently (see review, Zitman-Gal et al. 2005). Most of the studies that have investigated the desiccation survival ability of steinernematid nematodes have concentrated on the steinernematids S. carpocapsae and S. feltiae (Glazer 2002; Zitman-Gal et al. 2005). The general finding has been that, under slow drying conditions, various strains of S. carpocapsae can survive for several weeks. The improvement of the nematode’s ability to survive in dry soil will also increase its persistence and efficacy as a biocontrol agent.

In the present study, a gradual increase in slow-desiccation tolerance was observed. Glazer et al. (1991) studied the genotypic variation among inbred lines of Heterorhabditis bacteriophora (strain HP88). In that study, the heritability value (h2) for desiccation tolerance was low (0.11), suggesting that selective breeding for higher desiccation tolerance would be inefficient. Although, in the present study, the h2 values were not determined for the HFP, the gradual increase in tolerance of the gradual removal of water can be attributed to the low h2 value of this trait in the population.

A complementary approach to ensure the effectiveness of EPNs is the improvement of their ability to reach and invade the target host. We demonstrated here that under the selection regime, which involved a sand column, the proportion of nematodes that were found at the bottom layer increased 2.3-fold. Gaugler et al. (1989a) demonstrated a 20- to 27-fold increase in the host-finding ability of HFP of S. carpocapsae following a non-rigorous selection regime on agar plates. In the present study, we employed a harsher selection regime, forcing the IJs to pass through a 21-cm sand column before they could reach the target insects.

The host-seeking strategies of EPNs have been defined according to three categories: cruisers, ambushers and intermediates (see review by Lewis 2002). The cruisers allocate more of their foraging time to movement and scanning for resource-associated cues as they move through the environment, or during short pauses. Ambushers scan during long pauses that are interrupted by repositioning bouts of relatively short duration (Lewis et al. 1992). S. feltiae is known to be an intermediate host-seeker (Lewis 2002). By enhancing the cruising component of host-seeking behavior, as was done here, we can help the nematodes to reach their target host, thereby reducing the amount of time that they are exposed to unfavorable environmental conditions.

No reductions in overall fitness were recorded for the three selected populations. This is probably due to the high heterogeneity of the initial population. The initial hybridization of the different isolates, which were obtained from diverse environmental regions (Salame et al. 2010), ensures the maintenance of fitness despite the selections. The second factor is the mild selection pressures used in the desiccation selection regimes (3–7% survival in the stress assays, Fig. 1). This allowed a large number of individuals (thousands) to be selected for the next selection cycles, maintaining the overall fitness of the selected populations.

The fitness assays indicated that the selection for particular traits indirectly enhanced the ability of the selected populations in traits for which they were not selected. The increased infectivity of the host-seeking selected population, as measured by LD50 values, can be explained by the increasing ability of the nematode to reach and infect a target insect in the sand column. The improved heat tolerance of the slow-desiccation-tolerant population may be attributed to the effects of some fundamental tolerance mechanisms. Zitman-Gal et al. (2004) demonstrated that silencing of the gene from the LEA3 group reduced the desiccation as well as the heat tolerance of the nematode Cenorhabditis elegans. Glazer and Salame (2000) demonstrated the intertwined relations of different tolerance mechanisms. In that study, S. carpocapsae IJs which were desiccated osmotically or by evaporation, were highly resistant to heat stress, as compared with fresh nematodes in an aqueous suspension.

In conclusion, our successful enhancement of both S. feltiae HFP’s ability to survive desiccation and its host-seeking ability establish the foundation for improvement of this nematode for use as an effective biological control agent under less favorable conditions. This process is still in its infancy. Further improvement and crosses between selected lines are required, as well as validation under field conditions, before these nematodes can be used on a commercial basis. However, this classical genetic improvement approach may be more useful than other methods of improvement, such as gene transfer.