The ecological success of a social parasite increases with manipulation of collective host behaviour

Many parasites alter the behaviour of their host to their own advantage, yet hosts often vary in their susceptibility to manipulation. The ecological and evolutionary implications of such variation can be profound, as resistant host populations may suffer lower parasite pressures than those susceptible to manipulation. To test this prediction, we assessed parasite‐induced aggressive behaviours across 16 populations of two Temnothorax ant species, many of which harbour the slavemaker ant Protomognathus americanus. This social parasite uses its Dufour's gland secretions to manipulate its hosts into attacking nestmates, which may deter defenders away from itself during invasion. We indeed find that colonies that were manipulated into attacking their Dufour‐treated nestmates were less aggressive towards the slavemaker than those that did not show slavemaker‐induced nestmate attack. Slavemakers benefited from altering their hosts’ aggression, as both the likelihood that slavemakers survived host encounters and slavemaker prevalence in ant communities increased with slavemaker‐induced nestmate attack. Finally, we show that Temnothorax longispinosus colonies were more susceptible to manipulation than Temnothorax curvispinosus colonies. This explains why T. curvispinosus colonies responded with more aggression towards invading slavemakers, why they were less likely to let slavemakers escape and why they were less frequently parasitized by the slavemaker than T. longispinosus. Our findings highlight that large‐scale geographic variation in resistance to manipulation can have important implications for the prevalence and host preference of parasites.     

The role of an attractive brood pheromone in the obligatory,slavemaking ant,Polyergus breviceps (hymenoptera: Formicidae)

Freeliving workers of Formica occulta, an ant species enslaved by the obligatory slavemaking ant Polyergus breviceps, retrieve and nurse Polyergus pupaejust as well as conspecific pupae in a choice test. No such attraction was found toward pupae of the facultative slavemaker; Formica wheeleri,which also enslaves F. occulta. Formica neogagates,a sympatric species which is not parasitized by either slavemaker, preferentially retrieves and tends conspecific brood over that of Polyergusand F. wheeleri.It is proposed that brood of obligatory slavemaking species must possess an attractive pheromone for slavemaker colony foundation to be successful, since slavemaker brood must be nursed by adult slave workers with no prior exposure to slavemaker brood. An attractive pheromone is not necessary in the brood of facultative slavemakers, since this brood is cared for by newly eclosed slave workers who imprint on the slavemaker brood.     

The influence of space and time on the evolution of altruistic defence: the case of ant slave rebellion

How can antiparasite defence traits evolve even if they do not directly benefit their carriers? An example of such an indirect defence is rebellion of enslaved Temnothorax longispinosus ant workers against their social parasite Temnothorax americanus, a slavemaking ant. Ant slaves have been observed to kill their oppressors' offspring, a behaviour from which the sterile slaves cannot profit directly. Parasite brood killing could, however, reduce raiding pressure on related host colonies nearby. We analyse with extensive computer simulations for the Temnothorax slavemaker system under what conditions a hypothetical rebel allele could invade a host population, and in particular, how host–parasite dynamics and population structure influence the rebel allele's success. Exploring a wide range of model parameters, we only found a small number of parameter combinations for which kin selection or multilevel selection could allow a slave rebellion allele to spread in the host population. Furthermore, we did not detect any cases in which the reduction of raiding pressure in the close vicinity of the slavemaker nest would substantially contribute to the inclusive fitness of rebels. This suggests that slave rebellion is not costly and perhaps a side‐effect of some other beneficial trait. In some of our simulations, however, even a costly rebellion allele could spread in the population. This was possible when host–parasite interactions led to a metapopulation dynamic with frequent local extinctions and recolonizations of demes by the offspring of few immigrants.     

Caste‐biases in gene expression are specific to developmental stage in the ant Formica exsecta

Understanding how a single genome creates and maintains distinct phenotypes is a central goal in evolutionary biology. Social insects are a striking example of co‐opted genetic backgrounds giving rise to dramatically different phenotypes, such as queen and worker castes. A conserved set of molecular pathways, previously envisioned as a set of ‘toolkit’ genes, has been hypothesized to underlie queen and worker phenotypes in independently evolved social insect lineages. Here, we investigated the toolkit from a developmental point of view, using RNA‐Seq to compare caste‐biased gene expression patterns across three life stages (pupae, emerging adult and old adult) and two female castes (queens and workers) in the ant Formica exsecta. We found that the number of genes with caste‐biased expression increases dramatically from pupal to old adult stages. This result suggests that phenotypic differences between queens and workers at the pupal stage may derive from a relatively low number of caste‐biased genes, compared to higher number of genes required to maintain caste differences at the adult stage. Gene expression patterns were more similar among castes within developmental stages than within castes despite the extensive phenotypic differences between queens and workers. Caste‐biased expression was highly variable among life stages at the level of single genes, but more consistent when gene functions (gene ontology terms) were investigated. Finally, we found that a large part of putative toolkit genes were caste‐biased at least in some life stages in F. exsecta, and the caste‐biases, but not their direction, were more often shared between F. exsecta and other ant species than between F. exsecta and bees. Our results indicate that gene expression should be examined across several developmental stages to fully reveal the genetic basis of polyphenisms.     

Gene expression patterns associated with caste and reproductive status in ants: worker‐specific genes are more derived than queen‐specific ones

Variation in gene expression leads to phenotypic diversity and plays a central role in caste differentiation of eusocial insect species. In social Hymenoptera, females with the same genetic background can develop into queens or workers, which are characterized by divergent morphologies, behaviours and lifespan. Moreover, many social insects exhibit behaviourally distinct worker castes, such as brood‐tenders and foragers. Researchers have just started to explore which genes are differentially expressed to achieve this remarkable phenotypic plasticity. Although the queen is normally the only reproductive individual in the nest, following her removal, young brood‐tending workers often develop ovaries and start to reproduce. Here, we make use of this ability in the ant Temnothorax longispinosus and compare gene expression patterns in the queens and three worker castes along a reproductive gradient. We found the largest expression differences between the queen and the worker castes (~2500 genes) and the smallest differences between infertile brood‐tenders and foragers (~300 genes). The expression profile of fertile workers is more worker‐like, but to a certain extent intermediate between the queen and the infertile worker castes. In contrast to the queen, a high number of differentially expressed genes in the worker castes are of unknown function, pointing to the derived status of hymenopteran workers within insects.     

Raiders from the sky: slavemaker founding queens select for aggressive host colonies

Reciprocal selection pressures in host-parasite systems drive coevolutionary arms races that lead to advanced adaptations in both opponents. In the interactions between social parasites and their hosts, aggression is one of the major behavioural traits under selection. In a field manipulation, we aimed to disentangle the impact of slavemaking ants and nest density on aggression of Temnothorax longispinosus ants. An early slavemaker mating flight provided us with the unique opportunity to study the influence of host aggression and demography on founding decisions and success. We discovered that parasite queens avoided colony foundation in parasitized areas and were able to capture more brood from less aggressive host colonies. Host colony aggression remained consistent over the two-month experiment, but did not respond to our manipulation. However, as one-fifth of all host colonies were successfully invaded by parasite queens, slavemaker nest foundation acts as a strong selection event selecting for high aggression in host colonies.     

Selective Acceptance of Alien Host Species Pupae by Slaves of the Dulotic Ant,Harpagoxenus sublaevis (Hymenoptera,Formicidae, Myrmicinae)

The European slave-making ant, Harpagoxenus sublaevis, has up to three host species which occasionally are found all together in one colony. Colonies with only Leptothorax acervorum slaves, or with both L. acervorum and L. muscorum, are frequent, whereas colonies with only L. muscorum slaves are rare, although young H. sublaevis queens found their colonies about equally frequently with either of the two main host species. We demonstrate here, that the two host species exhibit an asymmetric acceptance behavior towards each other's pupae: L. muscorum workers accept L. acervorum pupae, both in non-enslaved colonies and as slaves in H. sublaevis nests, whereas L. acervorum never accepted L. muscorum pupae. Origin and meaning of this behavioral difference are unknown. The findings, however, might be very important with respect to host specificity and host-parasite interactions also in other slave-making ants.     

FIRST EVIDENCE FOR SLAVE REBELLION: ENSLAVED ANT WORKERS SYSTEMATICALLY KILL THE BROOD OF THEIR SOCIAL PARASITE PROTOMOGNATHUS AMERICANUS

During the process of coevolution, social parasites have evolved sophisticated strategies to exploit the brood care behavior of their social hosts. Slave-making ant queens invade host colonies and kill or eject all adult host ants. Host workers, which eclose from the remaining brood, are tricked into caring for the parasite brood. Due to their high prevalence and frequent raids, following which stolen host broods are similarly enslaved, slave-making ants exert substantial selection upon their hosts, leading to the evolution of antiparasite adaptations. However, all host defenses shown to date are active before host workers are parasitized, whereas selection was thought to be unable to act on traits of already enslaved hosts. Yet, here we demonstrate the rebellion of enslaved Temnothorax workers, which kill two-thirds of the female pupae of the slave-making ant Protomognathus americanus . Thereby, slaves decrease the long-term parasite impact on surrounding related host colonies. This novel antiparasite strategy of enslaved workers constitutes a new level in the coevolutionary battle after host colony defense has failed. Our discovery is analogous to recent findings in hosts of avian brood parasites where perfect mimicry of parasite eggs leads to the evolution of chick recognition as a second line of defense.     

Oh sister,where art thou? Spatial population structure and the evolution of an altruistic defence trait

The evolution of parasite virulence and host defences is affected by population structure. This effect has been confirmed in studies focusing on large spatial scales, whereas the importance of local structure is not well understood. Slavemaking ants are social parasites that exploit workers of another species to rear their offspring. Enslaved workers of the host species Temnothorax longispinosus have been found to exhibit an effective post‐enslavement defence behaviour: enslaved workers were observed killing a large proportion of the parasites’ offspring. As enslaved workers do not reproduce, they gain no direct fitness benefit from this ‘rebellion’ behaviour. However, there may be an indirect benefit: neighbouring host nests that are related to ‘rebel’ nests can benefit from a reduced raiding pressure, as a result of the reduction in parasite nest size due to the enslaved workers’ killing behaviour. We use a simple mathematical model to examine whether the small‐scale population structure of the host species could explain the evolution of this potentially altruistic defence trait against slavemaking ants. We find that this is the case if enslaved host workers are related to nearby host nests. In a population genetic study, we confirm that enslaved workers are, indeed, more closely related to host nests within the raiding range of their resident slavemaker nest, than to host nests outside the raiding range. This small‐scale population structure seems to be a result of polydomy (e.g. the occupation of several nests in close proximity by a single colony) and could have enabled the evolution of ‘rebellion’ by kin selection.     

Parasite scouting and host defence behaviours are influenced by colony size in the slave-making ant Protomognathus americanus

In the slave-making ant Protomognathus americanus, scout workers leave their colony, discover host colonies, and initiate slave raids. Captured host pupae subsequently emerge in the slavemaker colony and replenish the slave workforce. The course of these antagonistic encounters can be influenced by the species, aggressivity, or size of the host colony. We asked how the demography of parasite and host colonies influences the initial raiding phase by observing the scouting behaviour of P. americanus slavemakers during 48 raiding attempts. Experiments were performed under controlled laboratory conditions in a Y-shaped experimental arena. The number of active scouts increased with increasing slavemaker worker numbers, but was unaffected by the slave to slavemaker ratio, showing that slavemaker worker numbers are a good indicator for the scouting workforce. Colonies with fewer slaves discovered host colonies faster (colonies with 15 or less slaves: median 9:53 min, colonies with 42 or more slaves: median 18:55 min), suggesting that small slave workforces lead to intensified scouting behaviour. The more scouts were active, the faster a host colony was discovered, but the time between discovery and trial completion was unaffected by slavemaker colony demography. Host colonies were successfully attacked in 79.2 % of the trials, and they fought off an intruding scout only once. Yet host aggression towards slavemaker scouts increased with host colony size, and higher aggression rates delayed a subsequent attack. Our study demonstrates that colony size influences the behaviour and the course of crucial interspecific interactions of a social parasite and its host.     

A simple genetic basis for complex social behaviour mediates widespread gene expression differences

A remarkable social polymorphism is controlled by a single Mendelian factor in the fire ant Solenopsis invicta. A genomic element marked by the gene Gp‐9 determines whether workers tolerate one or many fertile queens in their colony. Gp‐9 was recently shown to be part of a supergene with two nonrecombining variants, SB and Sb. SB/SB and SB/Sb queens differ in how they initiate new colonies, and in many physiological traits, for example odour and maturation rate. To understand how a single genetic element can affect all these traits, we used a microarray to compare gene expression patterns between SB/SB and SB/Sb queens of three different age classes: 1‐day‐old unmated queens, 11‐day‐old unmated queens and mated, fully reproductive queens collected from mature field colonies. The number of genes that were differentially expressed between SB/SB and SB/Sb queens of the same age class was smallest in 1‐day‐old queens, maximal in 11‐day‐old queens and intermediate in reproductive queens. Gene ontology analysis showed that SB/SB queens upregulate reproductive genes faster than SB/Sb queens. For all age classes, genes inside the supergene were overrepresented among the differentially expressed genes. Consistent with the hypothesized greater number of transposons in the Sb supergene, 13 transposon genes were upregulated in SB/Sb queens. Viral genes were also upregulated in SB/Sb mature queens, consistent with the known greater parasite load in colonies headed by SB/Sb queens compared with colonies headed by SB/SB queens. Eighteen differentially expressed genes between reproductive queens were involved in chemical signalling. Our results suggest that many genes in the supergene are involved in regulating social organization and queen phenotypes in fire ants.     

The influence of hybridization on colony structure in the ant species Temnothorax nylanderi and T. crassispinus

The parapatric sibling ant species Temnothorax nylanderi and T. crassispinus hybridize in the contact zone in the Franconian Jura, Southern Germany. Aim of our study was to investigate the impact of hybridization on colony composition and fitness. We classified colonies as either ‘pure’ or containing hybrids by determining their allozyme pattern at GPI, an enzyme that is fixed for different alleles in the two parental species, and quantified their reproductive output. Most colonies with hybrid workers had a T. crassispinus queen. Colonies with heterozygous, hybrid workers produced more young workers than colonies of the parental species but similar numbers of male and female sexuals. Female sexuals from colonies with heterozygous workers had a significantly lower weight than female sexuals from pure colonies. Only a single reproductive queen was found to be heterozygous, suggesting reduced fitness of hybrid queens. As in the parental species, hybrid colonies appear to be frequently taken over by alien queens, which obscures the genetic colony structure. Received 6 April 2006; revised 10 June 2006; accepted 15 June 2006.     

Egg Recognition and Social Parasitism in Formica Ants

Nests of social insects are an attractive resource in terms of nutrition and shelter and therefore targeted by a variety of pathogens and parasites that harness the resources of a host colony in their own reproductive interests. Colonies of the ants Formica fusca and F. lemani serve as hosts for mound‐building Formica species, the queens of which use host colonies during colony founding. Here, we investigate whether workers of the host species can mitigate the costs imposed on them by invading parasite queens by recognizing and selectively removing eggs laid by these queens. We used behavioural assays, allowing host workers to choose between con‐colonial eggs and eggs laid by the parasite species F. truncorum. We show that workers of both host species discriminate between the two types of eggs in favour of con‐colonial eggs. Moreover, workers of F. fusca rejected more con‐colonial eggs than F. lemani. This higher rate of error in F. fusca may reflect a greater selectivity or a greater difficulty in discriminating between the two egg types. Nevertheless, both host species removed parasite eggs at a similar rate, when these were artificially introduced into the colonies, although some eggs remained after 10 d. In addition, upon receiving parasite eggs, host workers started to lay unfertilized male‐destined eggs within 6 d, thus employing an alternative pathway to gain direct fitness when the resident queen is no longer present and the colony is parasitized.     

Social Control over the Survival and Selection of Winged Virgin Queens in an Ant without Nuptial Flight: Iridomyrmex humilis

Laboratory and field experiments show that the number of winged virgin queens of the Argentine ant Iridomyrmex humilis allowed to remain alive in the nest is under social control. When reared in their natal nest, survival of virgin queens depends on the presence of males. In the presence of mature alate males or even male pupae only 20% of the gynes are executed before they are 4 days old. By contrast, if pupal and adult males are absent, about 55% of the gynes are executed. The workers tend to kill the lightest ones. Although the Argentine ant is a unicolonial species exhibiting complete acceptance of foreign conspecific workers and mated queens, regardless of whether the recipient nest is queenright or queenless, the acceptance of gynes is variable: they are executed by foreign queenless workers but widely accepted in queenright nests. These results are discussed with regard to possible existence of queen and gyne pheromones allowing recognition and to the lack of nuptial flight in this species.     

Imprinting Effects on Host-selection Behavior of Slave-raiding Chalepoxenus muellerianus (Finzi) Workers (Hymenoptera: Formicidae)

Early experience is known to influence brood care behavior and to induce nesting preferences of several ant species. Scanty information is available with respect to imprinting effects on host selection of dulotic ants. In this paper we investigate host choice behavior of workers of the slave-making ant, Chalepoxenus muellerianus. As in a corresponding paper (Schumann & Buschinger 1994) on host specificity of C. muellerianus young queens during colony foundation, we demonstrate that a blend of innate factors and effects of early experience also influences host species selection of slave-raiding workers.     

Colony structure of a slavemaking ant. II. Frequency of slave raids and impact on the host population

The parasite pressure exerted by the slavemaker ant Protomognathus americanus on its host species Leptothorax longispinosus was analyzed demographically and genetically. The origin of slaves found in colonies of the obligate slavemaker was examined with nuclear and mitochondrial DNA markers to make inferences about the frequency and severity of slave raids. Relatedness of enslaved L. longispinosus workers in the same nest was very low, and our data suggest that, on average, each slavemaker nest raids six host colonies per season. Therefore, the influence of slavemaker species on their hosts is much stronger than simple numerical ratios suggest. We also found that slave relatedness was higher in small than in large slavemaker nests; thus, larger nests wield a much stronger influence on the host. We estimated that in the study population, on average, a host nest has a 50% chance of being attacked by a slavemaker colony per year. Free-living Leptothorax colonies in the vicinity of slavemaker nests did not represent the source of slaves working in P. americanus colonies, which suggests that raided nests either do not survive or migrate after being raided. Colony composition and intranest relatedness of free-living L. longispinosus colonies differed markedly between areas with slavemakers and those that are parasite-free. In the presence of slavemakers, host colonies were less likely to be polygynous and had fewer workers and a higher relatedness among worker brood. Host nests with slavemaker neighbors allocated more resources into sexuals, possibly caused by these shifts in nest demography. Finally, enslaved Leptothorax workers in P. americanus nests appeared to be less efficient than their counterparts in free-living colonies. Thus, slavemakers exert a much stronger impact on their hosts than had previously been suspected and represent an unique system to study parasite-host coevolution.     

Polygynous supercolonies of the acacia‐ant Pseudomyrmex peperi,an inferior colony founder

In ant–plant protection mutualisms, plants provide nesting space and nutrition to defending ants. Several plant–ants are polygynous. Possessing more than one queen per colony can reduce nestmate relatedness and consequently the inclusive fitness of workers. Here, we investigated the colony structure of the obligate acacia‐ant Pseudomyrmex peperi, which competes for nesting space with several congeneric and sympatric species. Pseudomyrmex peperi had a lower colony founding success than its congeners and thus, appears to be competitively inferior during the early stages of colony development. Aggression assays showed that P. peperi establishes distinct, but highly polygynous supercolonies, which can inhabit large clusters of host trees. Analysing queens, workers, males and virgin queens from two supercolonies with eight polymorphic microsatellite markers revealed a maximum of three alleles per locus within a colony and, thus, high relatedness among nestmates. Colonies had probably been founded by one singly mated queen and supercolonies resulted from intranidal mating among colony‐derived males and daughter queens. This strategy allows colonies to grow by budding and to occupy individual plant clusters for time spans that are longer than an individual queen’s life. Ancestral states reconstruction indicated that polygyny represents the derived state within obligate acacia‐ants. We suggest that the extreme polygyny of Pseudomyrmex peperi, which is achieved by intranidal mating and thereby maintains high nestmate relatedness, might play an important role for species coexistence in a dynamic and competitive habitat.