Introduction

Soil-dwelling nematodes of Steinernema species, the obligate parasites of insects, are symbiotically associated with the Gram-negative bacteria Xenorhabdus species residing in their digestive tract. These nematodes are very successful insect larvae killers (Arthurs et al. 2004; Grewal et al. 2005) and therefore used in agriculture as biological control agents against insect pests (Ehlers 2001; Ciche et al. 2006; Lewis et al. 2006). During their life cycle, they either live free in the soil where at some points their infective juveniles search for insect prey and infect it, or live within their host prey and complete their life cycle (Boemare 2002). The juvenile nematodes leaving the host’s cadaver carry bacteria inside of an intestine vesicle (Boemare 2002), protecting them from the soil environment and eventually transferring them to the host prey. Nematodes symbiotic bacteria cannot spread on their own and need vectors (Ehlers 2001). The bacterium’s main service for their symbiotic partner is killing the prey by inducing septicemia, providing nutrients for the nematodes by digesting prey’s tissues, and protecting prey’s cadaver from saprophytes (Boermare 2002; Dillman et al. 2012).

Steinernema feltiae Filipjev, 1934 carrying the symbiotic bacteria Xenorhabdus bovienii Akhurst, 1983 is widely spread in temperate regions (Tailliez et al. 2006). It is suspected to be highly thermally plastic as it is found on all continents except for Africa and Greenland (Campos-Herrera et al. 2012). It is known that S. feltiae infects its prey hosts at temperatures between 8 and 30 °C and reproduces between 10 and 25 °C (Richardson and Grewal 1993). Hazir et al. (2001), in their work on S. feltiae isolated from different climate zones, showed that the temperature directly affects the time of death, penetration rate, emergence time, and the number of emerging infective juveniles. They suggested that the geographic isolation of S. feltiae resulted in its adaptation to the given region (Hazir et al. 2001).

A relatively great amount of data are known about the thermal response of entomopathogenic nematodes (Kaya 1977; Grewal et al. 1994; Sulsuruk 2008; Ulu and Sulsuruk 2014; Evans et al. 2015). However, fewer data have been provided for associated bacteria, although bacterial enzymatic activity and growth rate depend strongly on temperature (Tailliez et al. 2006; Hapeshi et al. 2020). According to Tailliez et al. (2006), Xenorhabdus species differ in the ability to grow at high temperatures. They found that some Xenorhabdus species can grow at 35–42 °C, while the others grow only below 35 °C. They suggested that some Xenorhabdus species may be adapted to tropical or temperate regions (Talliez et al. 2006). This fact is not surprising as X. bovienii is associated with at least nine Steinernema spp. (Talliez et al. 2006; Lee and Stock 2010; Campos-Herrera et al. 2012; Murfin et al. 2015; Bisch et al. 2016).

However, S. feltiae is associated only with X. bovienii (Stock and Blair 2008; Murfin et al. 2015), making this symbiotic relationship a good research model in studies on intraspecific variability of symbiotic partners. We know that symbiotic bacteria are key elements for the nematode’s reproductive success. Some studies showed that temperature directly affects the reproductive success of S. feltiae (Hazir et al. 2001) and that the infectivity rate and soil survival time depend on the nematode’s origin: with wild strains showing higher infectivity rate but shorter survival time, and the laboratory-bred strains showing longer survival time but lower infectivity rate (Chapuis et al. 2011; Grewal et al. 1999; Wang and Grewal 2002) . Thus, we asked if X. bovienii bacteria isolated from the same nematode species living in different thermal conditions and originating from different environments, natural vs. laboratory, show differences in thermal sensitivity. The analysis of the literature showed a lack of such information. The answer to this question could help to estimate the partake of bacteria in the nematodes thermal sensitivity. It would also allow us to experiment with the nematodes’ reproductive success by creating new symbiotic connections. Using X. bovienii isolates with different thermal ranges, we could check S. feltiae reproductive success and infectivity rate, which could translate into its possible future applications in agriculture. However, since the question about X. bovienii thermal range remained unanswered, we focused on this basic science issue first.

The presented study aimed to evaluate if intraspecific variation in temperature sensitivity occurs in X. bovienii isolates and depends on local climatic conditions, measured as the mean annual temperature, of the place of S. feltiae origin. To do this, we examined the bacterial growth rate in a temperature gradient for eight X. bovienii isolates from S. feltiae: six collected at different geographical latitudes with different mean annual temperatures and two isolates cultured in constant laboratory conditions. We choose the mean annual temperature as a factor showing climatic differences between locations. Since air temperature strongly correlates with soil temperature on ground level (Dwyer et al. 1990; Xu et al. 2011; Wojkowski and Skowera 2017), it can be used as an indicator of habitat conditions of nematodes living in the soil. We suspected that X. bovienii isolated from S. feltiae originating from different habitats might differ in thermal range (minimum, optimal, and maximal growth temperatures), which would explain the nematodes’ ability to spread over a wide range of temperatures. We also suspected that the thermal range of isolates bred in laboratory conditions over a longer time is different from the thermal range of isolates from S. feltiae collected from natural habitats.

Materials and methods

Study material and its origin

The experiment was conducted on X. bovienii, entomopathogenic bacteria isolated from the nematode S. feltiae. The nematodes came from the collection of the Institute of Environmental Sciences of the Jagiellonian University (Kraków, Poland). They were identified as S. feltiae using standard 16S RNA sequencing and the BLAST method (Taillez et al. 2006).

Originally, the nematodes were collected from natural habitats by scientists across Europe and Asian part of Turkey (six samples) or were bred in a controlled laboratory environment (two samples). Table 1 shows the exact places of origin of the nematodes used for the experiment. Samples FRA45 and FRA44, as well as WG01 and WG02, were collected from the proximal locations in the same country of origin, France and Poland, respectively. All nematodes originating from the natural habitats were collected shortly before the experiment and stored in a storage solution (NaCl, CaCl2 × 2H2O, MgSO4, ascorbic acid, 0.25% formaldehyde) by two to three generations in 4 °C. Sample Commercial, purchased from Koppert Biological Systems (the Netherlands) and OBS III, collected from the natural habitat in 1980 (Noordoostpolder, the Netherlands) and stored under laboratory conditions since then (Scheepmaker et al. 1998), represent nematodes bred in a controlled environment for a long time.

Table 1 Location and local climatic conditions data (Copernicus Data Store 2021) of the place of origin of Steinernema feltiae used for isolation of Xenorhabdus bovienii used in the experiment
Full size table

Isolation of Xenorhabdus bovienii

Bacteria have been isolated from the nematodes using the Galleria mellonella (Lepidoptera: Pyralidae) method (Akhurst 1980). Final-stage larvae of the greater wax moth were infected with S. feltiae larvae on moist sand. After 24 h of incubation at 25 °C in the dark, the larvae of G. mellonella were dissected in sterile conditions, and samples of infected hemolymph were collected with an inoculation loop, streaked on a culture plate and incubated for 3–5 days at 25 °C in the dark. The culture plates with NTBA agar (Akhurst 1980) containing nutrient agar (Merck, Germany) and two dyes: bromothymol blue (Avantor Performance Materials Poland S.A., Poland) and triphenylotetrazolium chloride (Merck, Germany) that were used to identify the symbiotic bacteria. Bacteria of Xenorhabdus spp. have two forms: I and II. Form I enters the symbiotic relationship with nematodes, produces antibiotics and other substances that allow nematodes to propagate, and forms green–blue colonies on the NTBA medium. Form II, mostly found only in laboratory conditions (Adams et al. 2006), presents low metabolic activity and produces red colonies on the NTBA medium (Akhurst 1980; Thanwisai et al. 2012). Only bacteria forming green–blue colonies were used for further experiments. Additionally, the bacteria were identified as X. bovienii using 16SP1 and 16SP2 primers as described by Taillez et al. (2006).

After the incubation on the NTBA medium, the single green–blue colonies were propagated in LB Agar (Difco™, USA) for 48 h at 25 °C. Then, bacteria were frozen with 15% glycerol (1:1) and stored at − 80 °C until needed for the experiment. After defrosting, bacteria were first streaked on the NTBA medium, incubated for 3–5 days h at 25 °C in the dark, and then used in further experimental steps.

Experimental design

The thermal sensitivity of X. bovienii isolated from S. feltiae, originating from natural habitats located at different geographical latitudes, differing in annual temperatures and bred in artificial conditions for a long time, was estimated using the three-step method, designed for this study. The first step aimed to determine the minimal and maximal temperature at which the isolated X. bovienii grows. The second step aimed to measure the bacterial growth rate at several temperatures between the minimal and maximal temperatures determined in the first step. The third step aimed to estimate minimal, maximal, and optimal growth temperatures for X. bovienii isolated with high precision. The modeling of the bacterial growth curves performed at this step used data obtained at the first two steps of the experiment.

Minimal and maximal temperature of Xenorhabdus bovienii growth

Bacteria were propagated in the dark, at 25 °C for 48 h in the LB medium on an orbital shaker (200 rpm) (ELMI, DOS-10 M, Estonia). Samples of liquid culture (10 µl) were diluted tenfold and streaked on the LB agar medium and then incubated in the dark at temperatures ranging from 2 to 36 °C. Bacterial growth was observed daily for two weeks. Each day, growth was classified as “–” no growth, “(–)” very weak growth, “(+)“moderate growth, or “ + ” strong growth – colony highly visible. The experiment was repeated for 6 individual colonies obtained from one X. bovienii isolate. Bacteria originating from one colony were tested in triplicates in the whole temperature range.

Xenorhabdus bovienii specific growth rate

The optical density (OD) in the exponential phase of growth for each X. bovienii isolate at several temperatures was measured to estimate the bacterial specific growth rate (Sugar et al. 2012). Temperatures ranging between the minimal and maximal temperature of growth of each isolate were chosen based on the results of the previous step. If the bacterial growth at a given temperature was too slow and it was difficult to measure it in a short time, the bacteria were propagated in the LB medium at 25 °C for 48 h. Then they were serially diluted (1:10) and incubated at the chosen temperature in a 96-well plate on an orbital shaker (200 rpm). The OD measurements were taken at 600 nm every hour until the stationary phase was established (microplate reader Infinite® 200, TECAN, Austria). The OD measurements were taken for six individual colonies obtained from one X. bovienii isolate. Bacteria originating from one colony were tested in triplicates in the whole temperature range at least twice. The mean value of OD was calculated and then used for bacterial growth rate calculations and modeling.

Xenorhabdus bovienii growth modeling

Basic mathematical models for bacterial growth across temperatures were described by Zwietring et al. (1991). The growth curve is defined as the logarithm of the relative population size [y = ln (N/NO)] as a function of time (t). For bacteria, the growth rate shows a lag phase that is followed by an exponential phase and, finally, a stationary phase when the growth rate decreases to zero and the number of bacterial cells reaches a maximum. A growth model with three parameters can describe this growth curve: the maximum specific growth rate µm, which is defined as the tangent of the inflection point; the lag time X, which is defined as the t-axis intercept of this tangent; and the asymptote A, which is the maximal value reached (Zwietering et al. 1991). In our research, we used the extended Ratkowsky model (Ratkowsky et al. 1983), which allowed us to model thermal reaction norms across a wide range of temperatures. The bacterial growth rate across the thermal range was modeled using the square root model (Ratkowsky et al. 1983):

$$ \mu = \, \left[ {b\left( {T{-}T_{{{\text{min}}}} } \right)} \right]^{{2}} \left\{ {{1}{-}{\text{exp}}\left[ {c\left( {T{-}T_{{{\text{max}}}} } \right)} \right]} \right\}, $$

where µ is the bacterial specific growth rate, Tmin and Tmax is the theoretical maximum and minimum temperatures at which the growth rate is zero, b and c are the coefficients calculated during the modeling process.

The specific growth rate (µ) was calculated based on the OD measurements. To estimate Topt (the temperature of the highest µ), we calculated the maximum of the described function for each bacterial isolate. Growth curves were calculated for every single repetition of the experiment for all isolates of X. bovienii, then the mean value for each estimated point—Tmin, Tmax, Topt was calculated.

Data analysis

All statistical analyses were performed using the R environment (R Core Team 2017).

Bacterial growth rate at selected temperatures was estimated using a linear regression model (lm). Regression coefficients were tested using a t test with the confidence interval was set at 95%. The calculated growth rate values were then used for data modeling. Data were fit to model using a non-linear regression model (nls). All calculated parameters: Tmin, Tmax, b, and c coefficients were also tested using Student’s t test with the confidence interval set at 95%. Topt was calculated based on the calculated bacterial growth curves as the function’s maximum. The errors were estimated using the bootstrap function. The one-way ANOVA was calculated using aov function and used to test variance between X. bovienii isolates of S. feltiae. Pearson’s correlation coefficients were calculated using the cor function and tested using a t test with the confidence interval set at 95%.

Results

Xenorhabdus bovienii thermal range

The first step of determining the thermal sensitivity of X. bovienii isolated from S. feltiae aimed to find the minimal and maximal temperature at which the isolated bacteria grow. We observed no growth for any isolated X. bovienii at 2 °C (Table 2). At 8 and 10 °C, we observed no growth for the FRA44 sample, poor growth for X. bovienii isolated from laboratory-bred S. feltiae and those originating from Poland and the Czech Republic, and good growth for bacteria isolated from FRA45 and 09–38 samples. Between 16 and 30 °C, the bacterial growth was strong, and the colonies were large and visible for most of the samples. At temperatures above 30 °C, the bacterial growth was hardly observed, with only X. bovienii isolated from the commercially available S. feltiae forming visible colonies at 34 °C (Table 2).

Table 2 The results of the thermal range test of Xenorhabdus bovienii isolated from Steinernema feltiae of different origin
Full size table

Xenorhabdus bovienii specific growth rate and growth curves

Based on the results obtained in the first step of determining the thermal sensitivity of X. bovienii isolated from S. feltiae, we choose temperatures between 12 and 35 °C to determine bacterial specific growth rate based on the OD measurements and modeling.

The Ratkowsky model showed that the minimal temperature of X. bovienii growth varies between 0.9 ± 2.2 °C for X. bovienii isolated from S. feltiae WG02 sample and 7.1 ± 1.4 °C for X. bovienii isolated from S. feltiae FRA44 sample (Table 3). It is worth noting that 95% coefficient intervals for minimal growth temperature are much wider than optimal and maximal temperatures of growth. The minimal temperatures of growth estimated for X. bovienii isolated from S. feltiae FRA45, FRA44, OBSIII, and Prosenice samples were highly precise (p < 0.001), for WG01 and Commercial samples were moderately precise (p 0.001–0.1), and for 09–38 and WG02, samples were not strongly recognized (p > 0.1). The model also showed that X. bovienii stops to multiply at around 36 °C, and the maximal temperature of X. bovienii growth was estimated for all samples with a very narrow 95% confidence interval (p < 0.001).

Table 3 Minimal (Tmin), optimal (Topt), and maximal (Tmax) temperature of Xenorhabdus bovienii estimated based on the growth curves obtained using the Ratkowsky model for specific bacterial growth
Full size table

The optimal temperature for X. bovienii growth calculated with the Ratkowsky model varied between 25.1 ± 0.2 °C for the Prosenice sample and 30.5 ± 0.2 °C for the WG-02 sample (Table 3).

The calculated optimal temperature of X. bovienii growth did not correlate with the mean annual temperature (Tann) of S. feltiae place of origin (R = − 0.009, p = 0.986, n = 8). For this correlation, we assumed that the Tann for X. bovienii originating from S. feltiae bred in laboratory conditions (OBSIII) or for commercial purposes (Commercial) was 25 °C. ANOVA analysis showed that X. bovienii isolates show no difference between the isolates for the calculated Tmin (F = 2.371, p = 0.138), Topt (F = 0.057, p = 0.813) and Tmax (F = 0.002, p = 966). Also, the analysis showed no differences in Tmin (F = 0.165, p = 0.688), Topt (F = 2.412, p = 0.135), and Tmax (F = 0.108, p = 0.746) between X. bovienii isolated from S. feltiae originating from a natural environment and isolates from S. feltiae kept in laboratory conditions.

The calculated specific X. bovienii growth rate (Table 4) ranged between 2.0 ± 0.3 [h−1] for X. bovienii isolated from S. feltiae from Prosenice sample and 3.6 ± 0.5 [h−1] for X. bovienii isolated from S. feltiae from Commercial sample (Table 4). The modeled growth curves, based on the OD measurements, for X. bovienii isolated from S. feltiae of different origin, are shown in Fig. 1.

Table 4 Xenorhabdus bovienii specific growth rate for optimal temperature (Topt) calculated based on OD measurements (n = 241) using the Ratkowsky model, and X. bovienii specific growth rate for mean annual temperature (Tann) extrapolated from the growth curves
Full size table
Fig. 1
figure 1

Growth curves of Xenorhabdus bovienii isolated from Steinernema feltiae originating from different environments. The growth curves were obtained using the Ratkowsky model. The place of S. feltiae origin was as follows: 09–38 – Turkey, FRA45, FRA44 – France, Prosenice – Czech Republic, WG-01, WG-02 – Poland, OBSIII – The Netherlands, Commercial – The Netherlands. OBSIII and Commercial were bred in laboratory conditions for a longer time

Full size image

ANOVA analysis showed no differences in the specific growth rate at Topt between the examined X. bovienii isolates (F = 0.001, p = 0.977) and no differences between the isolates cultured from S. feltiae originating from a natural habitat and S. feltiae kept in laboratory conditions (F = 0.844, p = 0.394).

The calculated specific X. bovienii growth rate at Topt did not correlate with the Topt calculated for the X. bovienii isolates (R = 0.046, p = 0.914).

When looking at the growth curves, we noticed that the curves form two groups (Fig. 1). The first one is characterized by a lower maximal growth rate (below 3.0 h−1), lower Topt (< 27 °C), and also a lower difference between Tmin and Topt (between 18.2 and 20.9 °C) and higher difference between Topt and Tmax (between 9.3 and 10.9 °C). The second one is characterized by a higher maximal growth rate (above 3.0 h−1), higher Topt (> 29 °C), and also a higher difference between Tmin and Topt (between 22.6 and 29.6 °C) and a smaller difference between Topt and Tmax (between 5.5 and 6.6 °C) (Tables 3 and 4). The first subgroup is represented by X. bovienii isolated from S. feltiae of FRA45, Prosenice, WG-01, and OBSIII samples. The second subgroup is represented by X. bovienii isolated from S. feltiae of 09–38, FRA44, WG-02, and Commercial samples. The division into the two identified subgroups has been confirmed by pairwise comparison of Topt of X. bovienii growth (Table 5). However, the pairwise comparison of the X. bovienii specific growth rate showed that the difference exists only between X. bovienii isolated from S. feltiae from Prosenice sample (with the lowest specific growth rate) and X. bovienii isolated from S. feltiae of WG-02 and Commercial samples (with the highest specific growth rates) (p < 0.05 for both comparisons).

Table 5 Results of pairwise comparison of Topt and the specific bacterial growth rate of Xenorhabdus bovienii isolated from Steinernema feltiae originating from different environments using Student’s t test
Full size table

Discussion

The presented study aimed to check if X. bovienii isolated from S. feltiae originating from different environmental conditions, have different thermal sensitivity. We also aimed to check if the thermal sensitivity of X. bovienii correlates with the mean annual temperature of the place of S. feltiae origin. The minimal and maximal temperatures of X. bovienii growth are, respectively, much above and below the thermal range of S. feltiae reproduction (Richardson and Grewal 1993). We decided to discuss the optimal temperature of bacterial growth because it can impact the host’s infection and reproduction rate. We also decided to choose the mean annual temperature as the only factor showing differences between locations.

The lack of significant differences in the optimal, minimal, and maximal temperature of X. bovienii isolates growth, and the lack of correlation between the optimal temperature of X. bovienii isolates growth and the mean annual temperature of S. feltiae place of origin may have various reasons. Calculated optimal temperature for X. bovienii growth did not correlate with the mean annual temperature of S. feltiae place of origin. When looking at the reproductive cycle of the nematode and its symbiotic bacteria, it seems that X. bovienii we isolated are not necessarily exposed to extreme temperatures. In our experiment, we isolated X. bovienii from the nematode’s infective juvenile form. During this stage, the bacteria are carried in a specialized intestine vesicle, mainly in the quiescent state, when no significant bacterial propagation occurs (Adams et al. 2006).

It might be possible that the bacteria also are less prone to selection by other external factors, like temperature to which the nematodes are exposed during the infective juvenile stage. Therefore, we conclude that since X. bovienii are protected by the vesicle during the infective juvenile stage, the lack of correlation between X. bovienii thermal range and the mean annual temperatures of S. feltiae place of origin proves that they were not subjected to selection that would force them to adapt to growth at different temperature conditions.

The presence of S. feltiae with symbiotic X. bovienii in sites characterized by different climatic conditions may result from two mechanisms: both nematodes and bacteria are thermally plastic, or local adaptations occur. Our results show that X. bovienii are more thermally plastic than their host. The studies of others showed that S. feltiae can adapt to local conditions (Hazir et al. 2001). However, our results also indicate that some degree of adaptation may emerge between X. bovienii isolated from S. feltiae. We found that the examined X. bovienii isolates form two subgroups with different Topt. Each subgroup is represented by one of the isolates originating from proximal locations (FRA45 and FRA44, WG-01 and WG-02) or kept in laboratory conditions (OBSIII and Commercial), indicating that X. bovienii can show some variance. The observed segregation of X. bovienii isolated from proximally sourced S. feltiae into physiologically distinct subgroups might indicate that the seasonal and local variations in the temperature have had forced the changes between isolates. In the case of X. bovienii isolated from laboratory kept nematodes, the difference between the isolates is more pronounced. X. bovienii from the OBSIII sample was isolated from S. feltiae collected from a natural habitat in 1981 in the Netherlands (Ehlers et al. 1997) and kept in the laboratory conditions since then (Scheepmaker et al. 1998). It shows a lower specific growth rate and lower Topt than X. bovienii isolated from commercially available S. feltiae. In the case of commercially bred S. feltiae, the nematode is mass-produced in bioreactors, optimized for maximal efficiency to be sold worldwide as a biocontrol agent that can be applied under different climatic conditions in fields to protect the plants from insect pests (Gaugler and Kaya 2018). This isolate showed the highest growth rate and the widest thermal range (31.8 °C), when omitting 09–48 and WG-02 samples with not strongly recognized Tmin. It seems that the two laboratory-derived isolates, OBSIII and Commercial, are distinct from each other. The noted differences between X. bovienii isolated from S. feltiae originating from the laboratory conditions suggest the influence of different laboratory storage and culture conditions, focused on achieving different goals.

Conclusion

Our study showed that X. bovienii have a wider thermal range than their nematode host, S. feltiae and X. bovienii carried by S. feltiae live below their thermal optimum. In a broader perspective, this bacteria’s trait can impact S. feltiae survival. Hypothetically, the bacteria would be able to kill the prey faster in higher temperatures, but the nematodes would fail to reproduce in such conditions. The lack of differences in the thermal range of X. bovienii isolates, observed in our study, reflects no phenotypical differences in this fundamental and highly conserved bacterial trait.