In single infections we found that P. ramosa isolates killed their hosts faster and produced fewer spores than P. ramosa clones. We also found that two similarly virulent isolates of P. ramosa differ considerably in their competitiveness when faced with coinfecting P. ramosa isolates and clones. While isolate P1 almost completely prevented the less virulent isolate P4 and the less virulent clones C1 and C14 from producing spores regardless of their relative dose (Figure 4A-C), in the case of isolate P2 the relative dose affected the competitive outcome (Figure 4D-F). Despite P1 being a better competitor, single and mixed infections with P1 resulted in the production of more host offspring than with P2 (Figure 2). Mixed infections were as virulent but not more virulent than single clone infections, and thus neither resulted in overexploitation of the host by the parasites (i.e., time-to-host-death and total spore production were not higher in mixed infections; Figures 1 and 4), nor entailed additional costs upon the host because there was no further reduction in host fecundity (Figure 2). The competitive ability in treatments with equal concentrations (50:50) appears to be transitive, i.e., against both reference isolates P1 and P2, isolate P4 competes better than clone C1 which competes better than clone C14. Although the P. ramosa isolate/clone with the higher starting dose has a higher likelihood to succeed, its success ultimately depends on its competitiveness (Figure 4). Based on spore counts 20 days post-infection, it appears that the competitive outcome is largely decided during the first half of the parasite’s growth phase (Figure 3).
Our results extend previous studies of multiple infections in the D. magna-P. ramosa host-parasite system, which were conducted using just parasite isolates [19, 27]. First, we show for the first time that the effects of multiple infections by parasite clones could be different than those previously reported for isolates, because P. ramosa clones were less virulent yet produced more transmission stages than P. ramosa isolates. Second, we show that P. ramosa isolates/clones vary in their within-host competitiveness and ability to induce host castration. Third, we show that epidemiology (i.e., relative dose) affects the outcome of within-host competition (previous studies in this system used the same dose but in equal concentrations). Taken together, these results highlight the need to investigate multiple infections using a wider range of host and parasite genotypes and under diverse epidemiological scenarios.
Parasites that castrate their hosts are expected to inhibit host reproduction early in the infection process, in order to divert host resources for parasitic use [44–46]. The higher fecundity of D. magna singly infected with P1 in comparison with other isolates/clones suggests that some P. ramosa clones are more successful at inducing castration. If inducing castration bears a cost to the parasite, in the form of slowing down spore development and growth, then it may affect its competitive ability with other clones. In mixed infections, the inability of P1 to castrate its host as quickly as P2 may be compensated by the greater competitiveness of P1. Put differently, castrating the host after it has reproduced once may be less costly to the parasite than doing so immediately after penetration, and may allow the parasite to focus on replicating itself to achieve a competitive edge . P1 might also be benefitting if coinfecting isolates/clones in mixed infection induce castration . It could be argued that P. ramosa sterilizes D. magna mechanistically, e.g., by growing around its ovaries. This is likely to bear no costs to the parasite, and may be supported by the fact that antibiotic treatment is sufficient to regain host reproduction . However, our day 20 post-infection data suggest that P1 grows faster than its competitors, despite delaying castration. Furthermore, it is not unusual for infected D. magna to release a clutch after a long period of castration.
The transitive relationship in competitiveness in mixed infections with equal concentrations (i.e., spore production of P4 > C1 > C14) is in line with their relative virulence in single infections (P4 was more virulent than C1 and C14). This suggests that when both parasite strains have equal chances to infect the host (50:50 concentration), their relative virulence in single infections may point to their competitive success in mixed infections. Similar results have been reported in a rodent malaria host-parasite system . Our study extends these results by showing that even in unequal concentrations P4 produced more or at least as many spores as C1 and C14 during mixed infections with P1/P2, despite its significantly lower spore throughput in single infections. Therefore, the ability of a parasite to transmit under conditions of frequent multiple infections ultimately depends on its competitiveness, and that a parasite’s relative virulence (but not its replication rate) in single infections serves as a good indicator of its competitive ability. Moreover, if more virulent parasite strains are more often better competitors, frequent multiple infections will lead to higher levels of virulence .
Pasteuria ramosa clones have been found to exhibit strong GxG interactions for infectivity . Some D. magna clones exhibit either complete resistance or complete susceptibility to infection that is governed by a simple genetic basis (i.e., one or few loci with dominance; ). Although the specificity of attachment to the host esophagus depends on both host and parasite genotypes , the specificity of P. ramosa proliferation within D. manga is poorly understood. It is also unknown whether the number of successful infections (i.e., number of spores attaching to the host esophagus) affects parasite replication rates within the host and the resulting spore load. Single-spore infection trials in the laboratory suggest that even though a single P. ramosa spore can cause disease, the likelihood of such an event is extremely low (circa 1 in 700; ). Spores that do not penetrate do not seem to be targeted by any innate immunity . It might very well be that if P. ramosa spores penetrate the host in small numbers, they are cleared by the host’s innate immune system before they are able to proliferate . Direct interference or apparent competition among different P. ramosa clones may also reduce proliferation . Since it is likely that P. ramosa isolates consist of more than one clone, some of which may be incompatible with the D. magna clone used in this experiment, we conjecture that a combination of proliferation specificity and inter-clone competition may explain why P. ramosa clones produced more spores than isolates. In other words, infection by a P. ramosa clone would maximize parasite fitness better than infection by a P. ramosa isolate. It remains to be determined whether the observed GxG interactions for infectivity also apply to within-host competitiveness and virulence (by examining the expression and evolution of virulence using additional D. magna clones).
Our finding that the competitive outcome is largely determined during the first half of the parasite's growth phase may be explained in several ways. First, the replication rates of successful competitors may be considerably higher than those of their counterparts, as evident from spore counts on day 20 post-infection. Second, direct interference or apparent competition might take place very early in the infection process e.g., , and clear out or considerably harm less competitive P. ramosa clones. Lastly, successful competitors might be able to facultatively upregulate their replication rates upon detection of another genotype within the same host, and thus express higher virulence [55–58]. To provide support to one or more of these conjectures would necessitate monitoring the competitive outcome during the initial growth phase while controlling for the number of successful infections.
Interestingly, the virulence of P. ramosa clones was lower than that of isolates. In theory under a scenario of resource competition, kin selection should reduce the increase in virulence per genotype in multiple infections by closely-related competing genotypes [59, 60]. However, the relationship between virulence and relatedness depends on the social behavior displayed by the parasites, i.e., prudent exploitation, public goods cooperation or spite . Evidence for reduced overall virulence in coinfections by closely-related parasite strains compared to unrelated strains is scarce [15, 16, 61]. Under the assumption that more than one P. ramosa spore penetrates the host during seven days of exposure, the present study provides additional support for the prediction that high relatedness selects for prudent exploitation and thus low virulence. This is because the difference in virulence between P. ramosa clones and isolates could be explained by <100% relatedness of genotypes in isolate infections. This latter statement assumes that a P. ramosa isolate consists of more than one P. ramosa clone. It remains to be seen whether the increase in overall virulence under multiple infections with potentially unrelated genotypes resulted from increased host exploitation or the inability of the D. magna immune system to cope with antigenic diversity .
The dose levels used in the present study were chosen to achieve high infection rates (>90%). Infection prevalence in natural populations of D. magna varies widely and may reach in certain ponds or years 100% [62–65]. However, it is unknown whether naturally occurring D. magna populations are exposed to concentrations of P. ramosa spores similar to those administered in our experiment. Lower spore concentrations may decrease the likelihood of multiple infections, and thus alter both within- and between-host dynamics. For example, if multiple infections are rare, less virulent P. ramosa clones that produce more transmission stages may be selected over more virulent clones that are less infective and/or produce fewer transmission stages . We do not expect different dose–response relationships for lower levels of infection, in terms of within-host competitiveness, overall virulence and parasite transmission. However, changes in the likelihood of multiple infections will affect the evolution of virulence.