Species-specific recognition of the carrier insect by dauer larvae of the nematode Caenorhabditis japonica

Phoresy, in which nematodes use larger invertebrates such as insects as vectors to transfer themselves to different sites for reproduction, is a common phenomenon among many nematodes. It is a form of commensalism in which the growth, fecundity or survival of the nematode is enhanced while the host remains unaffected. Although some host preferences exist, the same nematodes have often been isolated from different types of invertebrates (Barrière and Félix, 2005; Kiontke and Sudhaus, 2006; Herrmann et al., 2006), indicating low specificity. In addition, information on species-specific phoretic associations is limited (Kiontke, 1997; Herrmann et al., 2006), and there are very few reports in which the specificity of association has been demonstrated experimentally (Baird, 1999).

A key adaptation in Rhabditida that allows the development of such a phoretic relationship is the formation of dauer larvae (DL) (Sudhaus, 2008). DL are non-feeding larvae that have an arrested developmental stage specialized for dispersal and survival (Riddle, 1988). In phoretic nematodes, DL exhibit a special type of behavior called waving, i.e. lifting the anterior part, or more, of their body off a moist substrate and waving it in the air (Croll and Matthews, 1977). This behavior seems to improve the likelihood of phoresy or a host encounter (reviewed in Burr and Robinson, 2004). Although these typical phoretic behaviors are well known, the mechanisms used by DL to recognize their hosts have not been elucidated.

It is likely that DL use cues from their host insects, e.g. chemicals (volatile and water soluble) and physical contact (Grewal et al., 1993); however, there is little information on whether DL are able to recognize their host insects, and if so, the mechanism involved in this recognition. Chemoattraction has been studied in many nematodes, and the chemotaxis of Caenorhabditis elegans has been clarified in depth at both the molecular and cellular levels (Bargmann, 2006). In addition, host-specific chemoattraction has been reported in Pristionchus nematodes (Hong and Sommer, 2006). However, these studies in C. elegans and Pristionchus spp. used the adult stage, and thus no information is available on chemoattraction of DL to host insects, which is the actual phoretic stage, probably because DL are unresponsive to attractants in chemotaxis assays (Riddle, 1988). Recently, chemotaxis has been studied in DL of entomopathogenic nematodes (Rasmann et al., 2005; Hiltpold et al., 2010; Ali et al., 2010; Ali et al., 2011). However, these studies focused on herbivore-induced volatiles and used indirect host cues. CO₂ acts as a direct host cue for nematodes, but very little information is available on the chemicals involved in direct host recognition (Hallem et al., 2011). Moreover, no information is available on species-specific kairomones that are released from host insects and directly attract DL.

Caenorhabditis japonica is a bacterial-feeding nematode found on the burrower bug Parastrachia japonensis (Kiontke et al., 2002). Caenorhabditis japonica DL are always detectable throughout the year as aggregates on the body surface, particularly under the scutellum, of almost all female adults of P. japonensis but rarely on male P. japonensis. The biology and life history of C. japonica is currently under investigation, but according to our observations, this nematode appears to have a species-specific phoretic association because C. japonica has thus far never been detected on other invertebrates. To establish such an intimate association, C. japonica may have developed mechanisms to recognize and associate with its carrier bug P. japonensis. Our aim was to investigate whether C. japonica DL are able to recognize and associate with their carrier bugs in a species-specific manner. We used loading experiments in which C. japonica DL and a bug were incubated together in a small Petri-dish to demonstrate that C. japonica DL specifically associate with their carrier bugs. We also demonstrated the presence of kairomones, which are host-specific attractants, on the host bug for C. japonica DL.

Caenorhabditis japonica Kiontke, Hironaka and Sudhaus 2002 strain H1 was isolated from an adult female P. japonensis (Scott 1880) collected from Hinokuma Mountain Prefectural Park, Kanzaki City, Saga Prefecture, Japan. The nematode was maintained on dog food medium (Hara et al., 1981) seeded with Escherichia coli strain OP50. Phoretically active DL were collected as described by Tanaka et al. (Tanaka et al., 2010). Briefly, a sterile yellow 200 μl pipette tip (Watson Fukaekasei, Tokyo, Japan) was placed in the dog food agar in a 100 ml culture bottle, and DL waving on the top of the tip were collected using a worm picker (a 1 cm piece of platinum wire mounted on the tip of a Pasteur pipette), washed with distilled water three times and used for the experiments.

Adult P. japonensis were collected from Hinokuma Mountain Prefectural Park during 2007–2009. Females and males in reproductive diapause were collected between autumn and spring from aggregations of bugs. Two other species of shield bugs, Macroscytus japonensis Scott 1874 and Erthesina fullo (Thunberg 1783), were used for comparison. Macroscytus japonensis was chosen because it is a burrower bug that is present in the same habitat as P. japonensis in Hinokuma Mountain Park and it was relatively easier to collect. A phoretic nematode species (not C. japonica) was occasionally present on adult M. japonensis. Caenorhabditis japonica has never been found on M. japonensis (Yoshiga et al., 2013). Erthesina fullo was selected because it was a similar size to that of P. japonensis and it was readily available. No nematodes have been detected in this bug species (data not shown). A species of terrestrial isopod, Armadillidium vulgare Latreille 1804, was also used in the loading experiment because isopods were often present in the area where P. japonensis was found, and phoretic nematodes (not C. japonica) were often found in terrestrial isopods. Adult M. japonensis were collected in the autumn from below the litter near an aggregation of P. japonensis, whereas E. fullo were collected on the campus of Saga University, Saga City, Saga Prefecture, Japan, in autumn 2007. Armadillidium vulgare was collected on the campus of Saga University in July 2012.

Approximately 1000 DL were inoculated on a filter paper in a 3 cm plastic Petri dish, and then a C. japonica-free insect (male or female P. japonensis, M. japonensis, E. fullo or A. vulgare) or a pair of P. japonensis (male and female) was released in the dish. Twenty-four hours after inoculation at 25°C, the insects were dissected, their body parts were placed in water for 24 h to release DL, and the nematodes were counted under the stereomicroscope. Nematode numbers on an insect were compared among insect species and between male and female P. japonensis.

Because C. japonica DL are usually associated with P. japonensis, preparation of C. japonica-free P. japonensis was necessary before carrying out the loading experiments. To prepare nematode-free P. japonensis samples, adult P. japonensis were partly soaked in tap water for 3 days and rehydrated nematodes on the bugs were washed off (Tanaka et al., 2010). All DL were removed using this method (data not shown). The removal of nematodes was also confirmed by observations made using a stereoscopic microscope. Other shield bugs and the isopod were used after rinsing with distilled water.

We modified the chemotaxis assay method developed for Caenorhabditis elegans to investigate the nematodes’ response to the hexane extracts containing body surface components of shield bugs (Matsuura, et al., 2005). Test and control spots (1 cm circle) were set in a 6 cm nematode growth medium (NGM) plate, and 3 μl of hexane extracts containing body surface components of bugs and only hexane were spotted at the center of test and control spots, respectively. Next, ∼20–30 DL in 2–3 μl of water were placed at the center of the plate. Nematodes in the test and control spots were counted at intervals of 10 min for 60 min. The chemoattraction index (CI) value [(number of nematodes in the test spot–number of nematodes in the control spot)/total number of nematodes] was calculated as described in Bargmann et al. (Bargmann et al., 1993). Sodium azide, used in the original method to keep nematodes on the spots, was omitted in these experiments because C. japonica DL actively moved on the plate and were sometimes trapped by sodium azide during random movement. Hexane extracts containing body surface components of bugs were prepared by soaking a bug in hexane in a glass tube with a lid for 10 min. Bugs were weighed in advance, and only those with an average mass of 0.05±0.02 g for M. japonensis, 0.13±0.03 g for male P. japonensis, 0.17±0.03 g for female P. japonensis and 0.3±0.05 g for E. fullo were used in the assays. Based on the average insect sizes, 90, 300, 350 or 530 μl of hexane was used for M. japonensis, male P. japonensis, female P. japonensis or E. fullo, respectively. For fifth instar nymphs of P. japonensis, 350 μl of hexane was used for extraction.

We investigated the effects of hexane extracts containing body surface components of P. japonensis on the arrest of DL dispersal. One circle (1 cm diameter) was made at the center of an NGM plate, and 3 μl of hexane or hexane extracts was spotted at the center of the circle. As soon as the hexane evaporated and/or was absorbed on the plate, DL were inoculated directly onto the center of the circle. The number of nematodes left in the circle was counted every 10 min for 60 min.

An ANOVA with Bonferroni/Dunn tests was used for statistical analysis of chemoattraction (StatView Ver. 4.54; Abacus Concepts, Piscataway, NJ, USA).

During the loading experiments with P. japonensis, we observed DL crawling up the legs and abdominal parts of both male and female bugs. After dissection, aggregates of DL were found between the body segments and under the wings, where naturally associating DL are usually found. Up to 333 DL were found on P. japonensis. In contrast, nematodes were scarcely found on the body surfaces of M. japonensis, E. fullo and A. vulgare (max.=3, min.=0, max.=21, min.=0, and max.=16, min.=0, respectively), and there were significant differences between the nematode numbers on the insects (Fig. 1A).

The number of DL associating with male and female P. japonensis was not significantly different when a single male or female bug was placed in a 3 cm dish (Fig. 1B) and when a pair of male and female bugs was placed in a 3 cm dish (Fig. 1C).

The P. japonensis-specific embarkment in the loading experiments may imply the presence of some specific attractant cues. To test whether there were any differences in attraction toward the chemicals from the different bug species, chemoattraction of C. japonica DL to the hexane extracts containing body surface components of bugs was compared. Hexane extracts containing body surface components of P. japonensis moderately attracted C. japonica DL, and the CI values reached a plateau (∼0.3) within 60 min of the start of experiments (Fig. 2). In contrast, hexane extracts containing body surface components of M. japonensis and E. fullo did not affect nematode behavior and CI values were less than 0.03 (Fig. 2). The mean response to body surface components of P. japonensis was significantly higher than that to the other species at every time point (ANOVA, F=33.297, P<0.0001). When attraction was compared among the different stages, sexes and physiological conditions of P. japonensis, the CI values for nymphs were relatively high and were statistically higher than those for males (ANOVA, F=3.024, P=0.0429). However, no significant differences were observed in CIs among males, females in reproductive diapause (where C. japonica DL are usually associated) or provisioning females (where C. japonica DL are not found) (Fig. 3).

We released DL directly on the site where the hexane extracts were spotted to evaluate the arresting effects of the hexane extracts. When only hexane or hexane extracts containing body surface components of M. japonensis were used, DL rapidly dispersed on the plate (Fig. 4). In contrast, when hexane extracts containing the body surface components of P. japonensis were used, >50% of DL remained in the inoculation area even 60 min after nematode inoculation, and the mean response to P. japonensis was significantly higher than that to any of the other species at any time point (ANOVA, F=78.083, P<0.0001). Hexane extracts containing the body surface components of E. fullo paralyzed DL on the spot after nematode inoculation, possibly because of a toxin on its body surface, which resulted in high percentages of arrest of dispersal (82 and 62% at 10 and 20 min after nematode inoculation, respectively). Thus, these results were omitted from the analyses. When the effect of the hexane extracts on arrest of dispersal was compared among different P. japonensis stages, sexes and physiological conditions, no differences among the P. japonensis developmental stages or sexes, except between males and nymphs, were found (ANOVA, F=5.858, P=0.0024; Fig. 5).

In the present study, we demonstrated that C. japonica DL specifically associate with their carrier bug P. japonensis, but not with M. japonensis, E. fullo or A. vulgare. Hexane extracts containing body surface components of P. japonensis significantly attracted C. japonica DL and arrested their dispersal after contact, whereas DL were not attracted to hexane extracts containing body surface components of other insects. These results indicate that C. japonica DL recognize their carrier through some specific chemicals from the carrier, thereby enabling them to associate with the carrier. To the best of our knowledge, this is the first report demonstrating the species-specific embarkment of DL onto their host and the direct chemical recognition of a specific host by DL, thereby suggesting the presence of species-specific kairomones in nematodes.

Allelochemicals elicit a physiological or behavioral response between members of different species (Huettel, 1986; Riga, 2004). A kairomone is an allelochemical that elicits a positive response from the receiving organism. Although many nematodes form phoretic or parasitic associations with insects, very little information is available on kairomones other than CO₂ with regard to direct host recognition cues (Hallem et al., 2011). In the present study, we found that C. japonica DL were attracted to hexane extracts containing only body surface components of their host P. japonensis, indicating that the chemical components from their host appear to contain specific kairomones for the association of C. japonica DL with their carrier bug. The ecology of C. japonica is currently under investigation, but, to date, the nematode has been isolated only from P. japonensis. Parastrachia japonensis feeds only on the drupes of Schoepfia jasminodora and is one of only two species comprising the genus Parastrachia (Schaefer et al., 1991; Sweet and Schaefer, 2002). The specialized trophic ecology and evolutionary independence of P. japonensis may have helped to develop the specific kairomone for C. japonica. However, further studies on the characterization and identification of the kairomones are needed to understand the evolution of this species-specific phoresy and the mechanisms of host recognition.

Although phoresy is commonly found in nematodes, there are very few reports on sex-specific or sex-biased associations; female-biased association has been reported in the Fergusobia nematode/Fergusonia fly mutualism (Currie, 1937; Davies et al., 2001) and Sphaerularia nematode/Bombus bee or Vespa hornet parasitism (for reviews, see Bedding, 1984; Sayama et al., 2007). However, no information is available on the mechanisms of female-biased association. In the field, C. japonica DL are exclusively found on female P. japonensis. We expected that there would be some difference in attraction to male and female P. japonensis, but no significant difference was observed in either the loading or chemoattraction experiments. These results suggest that attraction of nematodes to both male and female bugs does not differ. The differences in nematode association in the field could be due to factors other than chemoattraction. One possibility is the difference in DL survival on male and female bugs. Male bugs are smaller than female bugs and DL on male bugs may face more severe desiccation than those on female bugs, resulting in the death of DL; these dead DL would then fall off the bug. Another possible reason is behavioral differences between the sexes. We observed grooming behavior in P. japonensis individuals in which they would open their wings and rub their body, antennae and legs with their hind legs. Thus, DL on male bugs could be removed during grooming. The frequency of the grooming behavior and/or the pattern of grooming may differ between sexes. Additional C. japonica and P. japonensis studies in the field are necessary to understand the female-specific association of DL, including studies on ecological and behavioral differences as well as on survivorship.

 
• CI

  chemoattraction index

 
• DL

  dauer larvae

 
• NGM

  nematode growth medium

C. elegans and E. coli OP50 strains were supplied by the Caenorhabditis Genetics Center.