Movement in different transmission stage(s) is a key aspect in parasite life cycles. Entomopathogenic nematodes have a specialized infective juvenile (IJ) stage, which is mobile, capable of withstanding environmental extremes and equipped to penetrate into the host insect hemocoel. Regardless of whether nematodes are applied to soil as inundative biological control agents, or are fulfilling their role as obligate insect parasites in more natural conditions, the IJ must move through the environment to find and then infect a host. Behaviors related to movement are an important aspect of EPN biology, and understanding how IJs move through space and interact in the soil will provide insights into their ecology and hopefully improve their utility as biological control agents.
In this series of experiments, we measured movement parameters of three species of EPN with different foraging strategies, applied alone or in combination.
Research into EPN movement has begun to explain many of the aspects of their group dynamics. It is well known that EPN populations aggregate due to both abiotic and biotic drivers. They cluster in response to environmental conditions, and their clumped distributions are reinforced by both mass emergence from host cadavers and the need for mass attack to effectively overcome insect defenses. Additionally, EPNs follow the trails of both conspecific and heterospecific IJs, especially in situations where the trial-making nematodes have had direct contact with insect hosts [2]; thus, clusters are also maintained via communication.
While the extent of heterospecific interaction (competition or cooperation) is not entirely understood, it is reasonable to assume that EPNs interact in field conditions. Some researchers have documented cases of niche separation by foraging habit [3], whereas others have found that similarity in host ranges and geography suggest that EPNs may interact in field conditions due to niche overlap [4] [5]. Mixed-species formulations are available commercially as a biological control product, and researchers have applied combinations containing multiple species with varying degrees of success [6] [7] [3] [8].
Previous work has indicated that aggregative behavior is a component of EPN dispersal [9] and that this aggregative pattern is continuous – i.e., that nematodes disperse together after emerging from a host, or are released into the soil from an aqueous solution, rather than dispersing separately and re-aggregating later at some point in space [10]. We investigated how these dynamics are affected when multi-species groups of EPN interact, either when co-applied at a single point or when two populations meet each other. We conducted these analyses in a series of experiments in sand-filled arenas that were maintained in uniform conditions. Because we were particularly interested in whether interspecific avoidance was occurring, we sampled the arena on a more extensive grid in later experiments, modifying protocols from earlier papers such as [9]. This would make it more likely for us to find instances where the species co-occurred in a single patch. Later experiments also focused on how IJs behaved when they were added at opposite corners of the arena, rather than after co-application in the arena center, as an opposite-corner condition would be a more precise way to observe avoidance, as opposed to interspecific differences in dispersal from a common introduction point.
We hypothesized that given the costs associated with mixed-species infections, where even the “winners” of such competitions show reduced fitness in terms of IJ production [2], aggregation should be more intense under mixed-species conditions than in single-species settings. Where two species were co-applied at a single point of introduction, we predicted that if clear species-specific signals enabled IJs to discriminate between species, the aggregations that would be established would be species-specific rather than mixed.
Entomopathogenic nematode species included a cruiser-type forager –
These experiments were conducted in arenas consisting of boxes fitted with a grid measuring 190mm × 190mm. The boxes were either made of polypropylene for the center application, or Pyrex (Corning Incorporated, Corning, NY) for the opposite-corner application. Boxes were filled with silica sand to a depth of 2.5 cm following previously described methods [9] [10]. The moisture level in the sand was brought to 10% to approximate field capacity with deionized water. Arenas were stored inside plastic bags at 25°C until nematode movement was assessed.
In single-species trials, nematodes were taken from culture flasks and concentrated onto filter paper using vacuum filtration. A single 60mm diameter filter paper (Whatman No. 1) was used to vacuum the IJs; the filter paper was then placed upside-down in the center of the sand arena (Fig. 1). Approximately 1,200 IJs were applied. Arenas were stored inside plastic bags at 25°C for three days prior to assessment.
After the three-day incubation, filter papers were removed, and sand cores measuring 20 mm in interior diameter and 25 mm in depth were taken at four points located at 45 mm from the center of the arena (diagonally and equidistant from each other, Fig. 1). Nematodes were extracted from the sand in each core, and from the filter paper, by triple-rinsing in deionized water; the number of IJs in each sample was then counted using a stereomicroscope.
In co-application trials, nematodes were taken from culture flasks and concentrated onto filter paper using vacuum filtration. A single 60-mm-diameter filter paper (Whatman No. 1) was used to vacuum the IJs of both species; the filter paper was then placed upside-down in the center of the sand arena (Fig. 1). Arenas were stored inside plastic bags at 25°C for three days prior to assessment.
After the 3-day incubation, filter papers were removed, and 25-mm deep sand cores with an interior diameter of 20 mm were taken at four points 45 mm from the center of the arena (diagonally and equidistant from each other, Fig. 1). Nematodes were extracted from the sand in each core and from the filter paper by triple-rinsing in deionized water; the precise number of IJs of each species in each sample was then counted using a stereomicroscope.
Nematodes were applied to separate-corner experimental arenas using two approaches: two different species placed at diagonal corners (Fig. 2A) and a single species at one corner (Fig. 2B). For the application of all treatments, 10,000 IJs from culture flasks were vacuum filtered onto 15-mm-diam. P8 filter paper (Fisher Scientific, Pittsburgh, PA). The filter papers were then placed upside-down in the corners of a sand arena, with nematodes in contact with the sand. Arenas were stored inside plastic bags at 25°C for five days. This extended period was chosen after experimentation because it allowed for IJs under one-corner treatments sufficient time to spread out across more than half of the arena, and time for possible interactions during heterospecific treatments.
After the 5-day incubation, filter papers were removed, and 11-mm-diam. interior sand cores with 25 mm depth were taken on a 25-point grid at points 35 mm from the center (Fig. 2C). Nematodes were extracted from the sand in each core and from the filter paper by triple-rinsing in deionized water; the precise number of IJs in each sample was then counted using a stereomicroscope.
For experiments conducted with nematodes that were applied to the center of the arena, four replicated dishes were tested and the entire experiment was conducted twice. In assays where applications were in one or more corners, each assay was conducted in time either three times (heterospecific two-corner assay) or four times (single-species single-corner assay).
The Index of Dispersion (D) was calculated in each of the center-placed nematode experiments. In these cases, as in previous experiments [9], [10], the null hypothesis we utilized predicted that nematodes would move independently, with an alternate hypothesis that they would tend to aggregate. We used a single approach for analyzing aggregation in single-species arenas, and two different approaches for assessing aggregation when species were added to the center together. In all cases, the Index of Dispersion [11] [9] was calculated as the ratio of the variance to the mean (D = σ2/µ) using the four equal-sized cores that were distributed around the introduction point (Fig. 1). In these cases, an index greater than 1 indicates aggregation, while an index lower than 1 indicates a more uniform distribution than expected given the null hypothesis. We tested whether the index of dispersion was significantly greater than 1 for a given species using a randomization test based on the sum of the indexes of dispersion across replicates [9].
In the corner-placed experiments, a null hypothesis of no aggregation predicted that expected counts would be the same for points equidistant from the origin. In the square grid layout used in the corner-placed nematode trials, there are two points at each distance from the origin. Since the grid is 5 × 5, this means that there are ten distances and ten corresponding pairs, excluding the diagonal.
According to the hypothesis, if the IJs did not aggregate, then the counts would tend to be similar between points in these pairs. If the IJs did aggregate, then there would tend to be larger differences in counts between points in the pairs. We used a test statistic that quantifies the degree of difference in count between paired points:
Where
Calculating
The expected value of T in this context is thus equal to 1. The following procedure was used to calculate [1] We calculated the observed test statistic [2] We estimated λ(d), which is the expected count in a core at distance [3] For each pair of off-diagonal points, we generated counts from Poisson λ(d). From these pairs of counts, we then calculated the test statistic [4] This third step was repeated 1000 times, and the test statistic values were saved, giving us a null hypothesis distribution for the test statistic. [5] Finally, we calculated the
We tested for differences in aggregation behavior between single-species and mixed-species assays in both center-placed and corner-placed experiments. If IJs moved via independent diffusive movement with no aggregation, then the counts in the cores would follow a Poisson distribution. One property of the Poisson distribution is that its mean is equal to its variance. If IJs aggregate, then we would expect to see that the variance in counts between cores is higher than the mean.
The negative binomial distribution is commonly used to model aggregated counts. This distribution has two parameters: the mean
The parameters of the negative binomial were estimated using the maximum likelihood for each assay. The mean parameters were estimated separately for each run, while the aggregation parameter was estimated jointly across all runs in an assay. We assumed that the mean count was determined by distance from the release point. For the center-placed nematodes, all cores were at the same distance. Thus, we estimated a single mean parameter for each run by maximum likelihood. For the corner-placed nematodes, the counts were on a grid and thus were at varying distances from the release point.
There was not enough data at each distance to reliably estimate mean parameters. Instead, LOESS regression was used to combine data across distances and estimate an overall-mean-count-versus-distance function, as described in the previous section. The mean value from this estimate was used in the likelihood function.
For each species, a likelihood ratio test was conducted to determine whether the aggregation parameter was higher when each species was paired with each other species than when a species was isolated. Under the null hypothesis, the aggregation parameter was the same between the two conditions; under the alternate hypothesis, the aggregation parameter was fit separately for each condition. Thus, there was a difference of one free parameter, and the likelihood ratio follows a chi-squared distribution with one degree of freedom.
Finally, we used Fisher’s method to combine
In the individual-species experiments, each of the three species showed strong evidence for aggregation (all
When species were paired and co-applied to the center of arenas, only
Using Fisher’s method to conduct an overall test of aggregation in single species vs. mixed species assays in corner-placed experiments showed that each species of nematode aggregated more in the presence of another species than when alone (
In general,
While prior research has shown that EPN display aggregative movement [9] [10], here we show that movement patterns are altered by the presence of other EPN species and the manner in which the EPN species are exposed to each other. Support for our initial hypothesis that aggregation should be more intense under mixed-species conditions than single-species settings depended on context. Where nematodes were co-applied in the center of arenas, we saw that EPN tended to aggregate less than they would when applied alone. In corner applications, where the nematodes started off at opposite sides of the arena, we saw strong evidence for greater levels of conspecific aggregation in mixed-species vs. single-species arenas, despite similar levels of EPN movement. In these corner-placed trials, then, nematodes moved just as much in mixed-species arenas as they did when alone, but tended to be more aggregated in the presence of another species.
Why did we see such differences between co-applied and corner-applied mixed-species aggregations? We believe this likely has to do with species-level differences in the ascarosides that are produced by these species. Ascarosides are biochemically active glycosides of the dideoxysugar ascarylose; ascaroside molecules play many roles in EPN and other nematodes, including regulating movement behaviors such as dispersal, trail following, etc. Each EPN species appears to produce species-specific blends of common ascarosides, with many differences between species being driven by the ratios among this common suite of chemicals [12]. These would be effective at maintaining species-specific groupings, but when two or more species are artificially mixed together (as we did here for center co-application, or as is done in some commercial formulations containing multiple EPN species), the ascaroside signals that would otherwise facilitate avoidance of heterospecifics and aggregation with conspecifics become muddled.
Wu and Duncan [14] investigated how mixtures of multiple species of EPN affected dispersal, using IJs of
While it is possible that compounds other than ascarosides are regulating these behaviors, given the previously-identified role of ascarosides in promoting dispersal [12], aggregation [16] and other social behaviors ([17], [18]), ascarosides seem the most likely suspect. We are presently investigating the chemistry that regulates species-specific aggregation of several species of entomopathogenic nematodes, with the objective of establishing the role of ascarosides in maintaining species-specific aggregative behaviors.
The results from the corner-placed portion of the experiments complement recent work on group behaviors in EPN. In tests of group joining behavior using six-arm olfactometers,
Extending those results to what we observed in the corner-placed trials suggests that without specific cues related to host presence, EPN IJs tend to avoid each other. This is supported by recent work on following behavior among these same three species, where there was no tendency for individual IJs to follow trails made by heterospecific IJs unless the “leader” on the trail had had prior contact with the host [2].
Scaling the avoidance behavior we observed in the corner-placed trials up to field conditions, it may hold the potential to mitigate interspecific competition and thereby promote diversity within the entomopathogenic nematode community. Many field surveys have found that multiple EPN species co-occur within fairly small areas [20], [21], [22]. While multiple EPN species can co-infect a host, there are costs to these co-infections, as early invaders may be attacked and killed by heterospecific or conspecific males [23], and even if one species succeeds within the host, there is generally a cost associated with this competition ([6] [24] [22] [2]). While niche separation based on factors such as foraging strategies, host specialization and depth preferences [25] [6] [4] certainly play a role in maintaining diversity, we show here that EPN appear to increase the level of conspecific aggregation in mixed-species environments.
It is important to consider that the current experiment used IJs stored in water prior to application, rather using IJs that emerged directly from cadavers into the arenas. Shapiro-Ilan et al. [9] compared aggregation behavior among six species in single-species arenas and found only one significant difference in aggregation behavior between aqueous-stored and cadaver-emerged IJs. In those trials,