Introduction
Different species of parasites utilizing the same host species may find themselves also infecting the same host individual and, hence, compete directly over host resources. For parasitic flatworms such as trematodes infecting a molluscan first intermediate host, the parasite–host biomass ratio is quite large, typically around 0.20 (Hechinger et al., Reference Hechinger, Lafferty, Manchini, Warner and Kuris2009). This predicts that interspecific trematode competition over host resources may be intensive, potentially affecting parasite fitness negatively. Accordingly, co-infecting colonies of larval trematodes (parthenitae) are frequently engaged in antagonistic interactions within the molluscan host, in which one species attempts to eliminate the other either directly by consumption or indirectly by producing substances that interfere with the growth and development of the competitor (Lie et al., Reference Lie, Basch and Umathevy1965; Basch et al., Reference Basch, Joe and Heyneman1970; Lim & Heyneman, Reference Lim and Heyneman1972; Walker, Reference Walker1979; Hechinger et al., Reference Hechinger, Wood and Kuris2011; Leung & Poulin, Reference Leung and Poulin2011; Miura, Reference Miura2012; Nielsen et al., Reference Nielsen, Johansen and Mouritsen2014; Garcia-Vedrenne et al., Reference Garcia-Vedrenne, Quintana, DeRogatis, Dover, Lopez, Kuris and Hechinger2017). In the former case, there exists a dominance hierarchy where trematodes that develop through a redial larva stage, equipped not only with a complete feeding apparatus (mouth, pharynx and stomach) but also appendages for locomotion, are competitively superior to species that develop through the largely immobile sporocyst stages lacking such structures (Sousa, Reference Sousa1993; Kuris & Lafferty, Reference Kuris and Lafferty1994; Galaktionov & Dobrovolskij, Reference Galaktionov and Dobrovolskij2003).
During the last decade, a range of trematode species originating from three different families have been shown to have their intramolluscan redial colonies divided into two morphologically and functionally distinct groups or castes (Poulin et al., Reference Poulin, Kamiya and Lagrue2019). The minor caste (soldiers) is primarily engaged in defending the colony against other invading parasites, whereas the considerably larger colony members are mostly engaged in producing the colony's cercarial dispersal stages (the reproductive caste). Although it is not generally accepted that these parthenitae colonies exhibit division of labour and, hence, emerge as socially organized (see Galaktionov et al., Reference Galaktionov, Podvyaznaya, Nikolaev and Levakin2015), available evidence suggests that the smaller colony members are those individuals mainly involved in antagonistic or predatory interactions with competitors (Hechinger et al., Reference Hechinger, Wood and Kuris2011; Leung & Poulin, Reference Leung and Poulin2011; Miura, Reference Miura2012; Nielsen et al., Reference Nielsen, Johansen and Mouritsen2014; Mouritsen & Halvorsen, Reference Mouritsen and Halvorsen2015).
If such a predatory role by the smaller parthenitae is, indeed, an adaptation resulting from natural selection under competitive pressures from co-infecting parasite species (see Poulin et al., Reference Poulin, Kamiya and Lagrue2019), it requires that the fitness of trematode colonies is negatively affected by co-infections. Surprisingly, this important evolutionary premise has been scrutinized in only two of the 15 socially organized systems so far identified, and demonstrated in just one of them. By using cercarial emission rate as a fitness parameter, Lloyd & Poulin (Reference Lloyd and Poulin2012, Reference Lloyd and Poulin2013) found markedly reduced fitness in redial colonies of Philophthalmus sp. when the host snail (Zeacumantus subcarinatus) was shared with the competitor Maritrema novaezealandensis. In contrast, the cercarial production by redial colonies of Himasthla elongata infecting common periwinkles (Littorina littorea) appeared unaffected by competition from co-occurring Renicola roscovita colonies (Mouritsen & Andersen, Reference Mouritsen and Andersen2017).
Albeit limited, these contradictive results raise concern about either the evolutionary premise of fitness costs of competition, or the application of cercarial emission rate as the only relevant fitness parameter. As stressed by Poulin et al. (Reference Poulin, Kamiya and Lagrue2019), other fitness components can be evolutionary targeted – for instance, the duration of infection rather than the parasite's immediate reproductive rate during competition events (i.e. lifetime rather than short-term reproductive success). Additional and equally relevant fitness components can be envisaged, such as the quality rather than the number of emerged cercariae. The amount of glycogen allocated to non-feeding cercariae during development will determine their functional longevity following emergence from the molluscan host and, hence, their likelihood of reaching the next host in the life cycle (Ginetsinskaya, Reference Ginetsinskaya1988; Galaktionov & Dobrovolskij, Reference Galaktionov and Dobrovolskij2003). Furthermore, the sheer size of the cercarial dispersal stage is known to correlate positively with adult body size and, in turn, the fecundity of adult worms in the definitive host (Poulin, Reference Poulin1997; Trouvé et al., Reference Trouvé, Sasal, Jourdane, Renaud and Morand1998; Poulin & Latham, Reference Poulin and Latham2003). Hence, reduced cercarial quality may have serious repercussions on the parasite's integrated lifetime reproductive success, and, insofar as cercarial quality is negatively affected by competition, natural selection may drive the evolution of a distinct redial caste dedicated to the elimination of competitors.
Here, we address this issue by revisiting the Himasthla–Renicola competition within periwinkle host snails (Mouritsen & Andersen, Reference Mouritsen and Andersen2017) in order to clarify whether or not cercarial quality (size and longevity) is influenced by co-infection. Because the energy available for the parasites may be influenced by host nutritional state (e.g. Seppälä et al., Reference Seppälä, Liljeroos, Karvonen and Jokela2008; Mouritsen & Andersen, Reference Mouritsen and Andersen2017), host size and season were also included in the analysis as potentially confounding variables.
Material and methods
Collection and acclimation of animals
Between September 2017 and January 2018, common periwinkles (L. littorea) were collected in the littoral zone of Knebel Vig, Denmark (56°13′41.2″N, 10°27′47.4″E). Here, the periwinkles are known to host infrapopulations of the trematodes R. roscovita and H. elongata, of which the former develops sporocysts (no division of labour) and the latter develops rediae (division of labour possible). Snails were distributed in storage containers (40 × 30 × 20 cm, approximately 25 individuals per container) kept in an illuminated temperature-controlled room. Each container was filled with 10 l of seawater that was replaced weekly (21–26 psu), an oxygen supply and a number of small rocks with microalgal films serving as food supply.
The snails were gradually acclimated after collection – that is, the water temperature in the storage containers was changed from approximately in situ water temperature at collection to 16°C. The final storage temperature of 16°C was chosen to minimize (1) temperature stress during the cercarial emission procedure (see below); and (2) unintentional emission of cercariae in storage containers (especially pronounced around 20°C and above; Galaktionov & Dobrovolskij, Reference Galaktionov and Dobrovolskij2003). The acclimation procedure also ensured a similar starting point in terms of temperature for all snails prior to experimentation regardless of date of collection. The length of the acclimation period as well as the temperatures involved depended on the in situ water temperature at the point of collection. Snails collected during autumn (late September to late October) were stored for five days at c. 12°C, followed by five days at 16°C. Snails collected during winter (mid-December to mid-January) were stored for five days at 5°C, followed by five days at 10°C and, finally, five days at 16°C.
Following acclimation, snails single-infected by R. roscovita or H. elongata as well as snails infected by both these trematode species (double infections) were identified by shedding of cercariae. Snails were incubated individually in seawater-filled glass jars (70 ml) at c. 20°C under light for a few hours and emerged cercariae were species determined according to Werding (Reference Werding1969) under a stereomicroscope. Infected snails were then uniquely marked (numbered) and returned to infection-specific storage containers at 16°C until further treatment. When planned measurements and experiments had been performed (see below), the shell height of each snail was determined to the nearest 0.1 mm using an electronic calliper (mean/range: 23.2/17.7–29.3 mm) followed by dissection to verify infection status.
Cercarial dimensions
Himasthla elongata and R. roscovita cercariae for size measurements were obtained by incubating infected snails individually in seawater-filled glass jars at 22°C for 3 h. Water and shed cercariae were then poured into Petri dishes and placed in an oven at 60°C for 1 h in order to heat-kill the trematode larvae. Corresponding body length (tail excluded) and width of 20 haphazardly chosen cercariae from each individual trematode colony (i.e. 20 Himasthla and 20 Renicola cercariae from double-infected hosts) were then measured under a stereomicroscope (×80) using an ocular micrometre. Also, the diameter of the relatively large ventral sucker of all H. elongata cercariae was measured. This procedure was carried out for infected snails collected both during autumn and winter, and, in total, 1620 cercariae from 20 single-infecting Himasthla colonies (14 in autumn, six in winter), 25 single-infecting Renicola colonies (ten in autumn, 15 in winter) and ten double infections (six in autumn, four in winter) were measured.
The exclusion of the tail in body measurements is justified because it is lost prior to metacercarial formation, and, hence, it is the tail-less cercarial body that transforms into the adult worm and determines its fecundity (our focus, see Introduction). Moreover, due to the high energy content of the tail that determines cercarial swimming stamina (Lawson & Wilson, Reference Lawson and Wilson1980; Ginetsinskaya, Reference Ginetsinskaya1988), properties of the tail will be an integral part of cercarial longevity (see below).
Cercarial longevity
Cercariae for survival experiments were obtained as described above. Petri dishes containing emerged cercariae (average age: 1.5 h) were subsequently placed in a temperature-controlled room at 24°C and inspected under a stereomicroscope for presence of irreversibly grounded cercariae every second hour. Irreversibly grounded and, hence, functionally dead cercariae were defined as individuals found on the bottom of the dish that either (1) presented markedly reduced tail movements; (2) had lost their tail; or (3) laid motionless. After 12 h, the total number of cercariae present in each Petri dish was recorded. Prior to the cercarial counts, host snails were measured (as above) and returned to their storage containers.
The above protocol was conducted for infected snails collected both during autumn and winter in order to account for different nutritional states of host snails. In total, 16 single-infecting Himasthla colonies (11 in autumn, five in winter), 17 single-infecting Renicola colonies (five in autumn, 12 in winter) and eight double infections (five in autumn, three in winter) were processed. On average, the longevity of 75.4 ± 14.4 (standard error) cercariae from each individual trematode colony was monitored.
Data analyses
All statistical analyses were carried out using Statistical Package for the Social Sciences (IBM SPSS 24.0; IBM Corporation, New York). The influence of competition (single-/double-infected), season (autumn/winter) and host size (shell height) on cercarial dimensions were analysed using ANCOVA (analysis of covariance), entering host size as covariate. The mean projected body area (length × width) and mean ventral sucker area across cercariae within individual colonies were used as the unit of analysis. Hence, mean values depicted in figures represent mean of mean. A similar ANCOVA model was applied for cercarial longevity by interring LT50 values (time passed at the point of 50% mortality) as the unit of analysis. Two separate ANCOVAs were conducted, one for each of the two trematode species involved (H. elongata and R. roscovita). LT50 values were obtained from cercarial survival curves elaborated for each individual trematode colony processed.
Preliminary full-model ANCOVAs were executed to evaluate the significance of two- and three-way interactions. As none of these interaction terms were statistically significant, neither regarding cercarial dimensions (P ≥ 0.188) nor cercarial longevity (P ≥ 0.576), we only present summary statistics from reduced-model ANCOVAs, including interaction mean squares and associated degrees of freedom in the error variance. All analyses were preceded by tests of homogeneity of error variance (Levene's test) and evaluation of normality. Because equality of error variance was violated in the case of Renicola body size, these data were rank-transformed prior to analysis in order to meet requirements. In any case, all mean values reported on in the following are based on untransformed data.
Results
Cercarial dimensions
The projected body area of H. elongata cercariae was significantly affected by both competition and season, whereas host size was uninfluential (fig. 1a, table 1). When the host snail was co-infected by R. roscovita, emerging H. elongata cercariae were roughly 20% smaller than cercariae emerging from a host infected solely by Himasthla. Similarly, cercariae emerging during winter were about 24% smaller than those emerging during autumn. On the other hand, the size of the ventral sucker was affected statistically only by competition; season as well as host size appeared uninfluential (fig. 1b, table 1). Interestingly, competition explained considerably more variation in cercarial body size (43%) than in sucker size (16%) (table 1).
Body dimensions (mean of mean ± standard error) of cercariae emerging from single- and double-infected periwinkles Littorina littorea during autumn (September–October) and winter (December–January). (A) Projected body area (mm2 × 10−2) of Himasthla elongata cercariae; (B) ventral sucker area (mm2 × 10−3) of H. elongata cercariae; (C) projected body area (mm2 × 10−2) of Renicola roscovita cercariae. NHimasthla, autumn: single/double = 14/6; NHimasthla, winter: single/double = 6/4; NRenicola, autumn: single/double = 10/6; NRenicola, winter: single/double = 15/4.
Summary statistics of reduced-model ANCOVAs, including body size (projected area) or ventral sucker area of Himasthla elongata or Renicola roscovita cercariae as dependent variable, competition (single/double infection) and season (autumn/winter) as independent fixed factors, and host size (shell height of Littorina littorea) as a covariate.
Statistically significant effects are indicated by P-values in bold. Partial η 2 denotes effect size, i.e. the proportion of variance explained.
Similar to H. elongata, the body size of R. roscovita was strongly negatively affected by competition (28% reduction), whereas host size had no bearing (fig. 1c, table 1). Opposite to H. elongata, however, no significant effect of season could be demonstrated on the cercarial size of Renicola.
Cercarial longevity
The functional longevity (LT50) of H. elongata cercariae was affected neither by competition, season or host size (fig. 2, table 2). In contrast, the longevity of R. roscovita cercariae was significantly negatively affected by competition with H. elongata (table 2). Competition reduced renicolid longevity by 37%, explaining about 25% of the variation in the species’ functional lifespan (table 2).
Survival rates (LT50) of Himasthla elongata and Renicola roscovita cercariae from single- and double-infected periwinkles Littorina littorea. NHimasthla: single/double = 16/8; NRenicola: single/double = 17/8.
Summary statistics of reduced-model ANCOVAs, including survival (LT50) of Himasthla elongata or Renicola roscovita cercariae as dependent variable, competition (single/double infection) and season (autumn/winter) as independent fixed factors, and host size (shell height of Littorina littorea) as a covariate.
Statistically significant effects are indicated by P-values in bold. Partial η 2 denotes effect size, i.e. the proportion of variance explained.
Overall, the functional longevity of H. elongata cercariae was 2–3-fold greater than that of R. roscovita cercariae (fig. 2). This substantially longer lifespan of Himasthla was highly significant for single infections (Student's t-test, t 20.7 = 3.901, P = 0.001) as well as double infections (t 14 = 5.513, P < 0.0005).
Discussion
The present results demonstrate that intramolluscan competition between trematode colonies may negatively affect size and longevity, and, thus, the quality of the emerged cercarial dispersal stages. However, these effects appear species-specific.
Cercariae emerging from H. elongata colonies co-occurring with R. roscovita were significantly smaller than those emerging from single infections. Similarly, the size of emerged Himasthla cercariae was smaller during winter when host food (mainly green algae) is scarce than during autumn when food is more abundant. Together, this suggests that colonies of H. elongata respond to nutrient limitation – either as a consequence of interspecific exploitative competition or reduced host food availability – by reducing the size of the already large dispersal stages (fig. 1). In contrast, cercarial longevity was unaffected by both competition and seasonally induced energy shortage, maintaining the already extraordinary long lifespan (fig. 2). Hence, under nutritional stress H. elongata appears to launch a transmission strategy that prioritizes the probability of reaching the next host in the life cycle by producing the same amount of dispersal stages (see Mouritsen & Andersen, Reference Mouritsen and Andersen2017) with unchanged energy density (i.e. unchanged longevity, this study) at the expense of cercarial size. Although larger cercariae may release greater fecundity among the resulting adult worms (Poulin, Reference Poulin1997; Trouvé et al., Reference Trouvé, Sasal, Jourdane, Renaud and Morand1998; Poulin & Latham, Reference Poulin and Latham2003), this promise of future egg reproduction is worthless if the transmission to second intermediate hosts fails.
The impact of competition and season on overall body size of emerged Himasthla cercariae was mirrored also in the size of the ventral sucker as a relatively rigid and perhaps more true-to-scale structure than the remaining body. Statistically, however, this was only significant for competition (table 1). Considering the limited sample size, this may, in the case of season, be due to a type II error. On the other hand, suckers are formed early during cercarial ontogenesis (Galaktionov & Dobrovolskij, Reference Galaktionov and Dobrovolskij2003) and, therefore, less likely to correlate tightly with the availability of host food at the point of emergence: cercariae emitted by the snails collected in winter may have been formed from redial germ balls already during late autumn.
Like H. elongata, interspecific competition reduced the body size of cercariae that emerged from R. roscovita colonies. As opposed to H. elongata, also the longevity of the Renicola cercariae was reduced under competition. Because competition also has a marked negative effect on cercarial production in Renicola (Mouritsen & Andersen, Reference Mouritsen and Andersen2017), R. roscovita colonies appear particularly sensitive to competitive interactions; the dispersal stages are affected negatively both quantitatively and qualitatively, which is bound to cause a substantial reduction in transmission success. That competition also negatively impacts renicolid cercarial quality, underlining that H. elongata does not diminish R. roscovita fitness solely through direct predation (see present Introduction and discussion in Mouritsen & Andersen, Reference Mouritsen and Andersen2017).
Opposite to H. elongata, there was no seasonal effect on the size of R. roscovita cercariae emerging from double-infected snails. Nor was cercarial longevity affected by season. Together, this suggests that the condition of the host (or the food available to the host) at the point of shedding has little bearing on renicolid cercarial quality, in agreement with similar findings on also the number of emerged cercariae (see Mouritsen & Andersen, Reference Mouritsen and Andersen2017). The pattern may follow from the fact that the entire renicolid sporocyst colony, as well as the contained cercariae, are formed more or less simultaneously early during infection (Wright, Reference Wright1956; Galaktionov & Dobrovolskij, Reference Galaktionov and Dobrovolskij2003; Galaktionov et al., Reference Galaktionov, Podvyaznaya, Nikolaev and Levakin2015), and, hence, cercariae shed by winter snails are likely to have developed to maturity earlier during the season.
Clearly, intramolluscan competition between co-infecting H. elongata and R. roscovita involves significant fitness losses to both parties. Renicola roscovita is especially afflicted, as the quantity as well as the quality of dispersal stages is reduced. However, also the fitness of the generally competitively superior H. elongata colonies is negatively influenced: cercarial quality in terms of body size is reduced, which is likely to affect this parasite's lifetime reproductive success (see Introduction). Aside from emphasizing that the number of dispersal stages is not the only relevant fitness parameter in colonies of larval trematodes, our study demonstrates that the evolutionary premise for natural selection to maintain and further optimize a distinct group of defenders in redial colonies of H. elongata is, indeed, in place.
Acknowledgements
We wish to thank two anonymous reviewers for improving the paper.
Financial support
The research received no specific grant from any funding agency or sector.
Conflicts of interest
None.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional guides on the care and use of laboratory animals.
