Ancestral Developmental Potential Facilitates Parallel Evolution in Ants

See allHide authors and affiliations

Science  06 Jan 2012:
Vol. 335, Issue 6064, pp. 79-82
DOI: 10.1126/science.1211451


Complex worker caste systems have contributed to the evolutionary success of advanced ant societies; however, little is known about the developmental processes underlying their origin and evolution. We combined hormonal manipulation, gene expression, and phylogenetic analyses with field observations to understand how novel worker subcastes evolve. We uncovered an ancestral developmental potential to produce a “supersoldier” subcaste that has been actualized at least two times independently in the hyperdiverse ant genus Pheidole. This potential has been retained and can be environmentally induced throughout the genus. Therefore, the retention and induction of this potential have facilitated the parallel evolution of supersoldiers through a process known as genetic accommodation. The recurrent induction of ancestral developmental potential may facilitate the adaptive and parallel evolution of phenotypes.

The wingless worker caste, a universal feature of ants (1, 2), has repeatedly expanded into a complex system of morphological and behavioral subcastes. The existence of these subcastes has long fascinated biologists (39), yet little is known about their developmental and evolutionary origin (7, 8). The ant genus Pheidole is one of the most species-rich genera, with 1100 species worldwide (10, 11). All Pheidole species have two worker subcastes: minor workers (Fig. 1C) that perform most tasks in the nest and forage and soldiers (Fig. 1B) that defend the nest and process food (11). This complex worker caste system is thought to have promoted the remarkable diversification of Pheidole by enhancing the division of labor (11).

Fig. 1

Wing polyphenism in P. morrisi: the ability of a single genome to produce (A) winged queens and wingless (B) soldiers and (C) minor workers (2). Caste determination occurs at two JH-mediated switch points in response to environmental cues (1, 15, 30). (D) Wing discs in queen larvae showing conserved hinge and pouch expression of sal. (E) Vestigial wing discs in soldier larvae showing a soldier-specific pattern of sal expression, where it is conserved in the hinge but down-regulated in the pouch. Asterisks represent the absence of visible wing discs and sal expression in (E) soldier and (F) minor worker larvae. Scale bars indicate the relative sizes of queen, soldier, and minor worker larvae and adults.

In a wild P. morrisi colony, we discovered several anomalous soldier like individuals. These individuals are anomolous because they are significantly larger than normal soldiers (Fig. 2, A and B, and fig. S1), and unlike normal soldiers, they have mesothoracic wing vestiges (Fig. 2, C and D) and rarely occur in nature. These anomalous soldiers are similar to a supersoldier subcaste, which is known to be continually produced in eight Pheidole species (fig. S2) (1012). These species co-occur with army ants and live exclusively in the deserts of the American southwest and northern Mexico (11). In one of these species, P. obtusospinosa (Fig. 2, E and I), a major function of the supersoldier subcaste is to block the nest entrance with their extra-large heads and engage in combat to defend against army ant raids (13).

Fig. 2

Comparison of P. morrisi (Pm) ants: (A, left) Normal adult soldier (SD) and (A, right) anomalous supersoldier (aXSD). (B) Mean and standard deviation of head width of normal SD (gray) and aXSD (black) (fig. S1), and thorax of (C) a normal SD and (D) an aXSD. Comparison of P. obtusospinosa (Po) ants: adults of (E) SD and (I) supersoldier (XSD); larvae of (F) SD and (J) XSD; and vestigial wing discs [stained with 4′,6′-diamidino-2-phenylindole (DAPI)] and sal expression in SD (G and H and fig. S4A) and XSD (K and L and fig. S4B). White arrowheads indicate the presence of mesothoracic vestigial wing buds or discs; asterisks denote their absence. Black arrowheads indicate sal expression in the wing pouch. Adult, larval, and vestigial wing disc images are all to scale. See fig. S3 for a comparison of P. rhea SD and XSD.

The similarity between the supersoldier-like anomalies in P. morrisi and the supersoldier subcaste suggests that they share a developmental origin. Normal soldier development in Pheidole may provide insight into how supersoldier-like anomalies may have originated. The soldier subcaste is determined late in larval development at a soldier–minor worker switch point (Fig. 1), which is largely controlled by nutrition (5) and mediated by juvenile hormone (JH) (14, 15). Soldier development is defined by two features: (i) Soldier-determined larvae grow larger than minor worker larvae; and (ii) they develop a pair of vestigial forewing discs in their mesothoracic segment (Fig. 1, E and F) (1416). These discs show a soldier-specific expression of spalt (sal) (Fig. 1E) (1), a key gene in the network underlying wing polyphenism in Pheidole. Sal is a key gene because its expression is spatiotemporally associated with the induction of apoptosis in these vestigial forewing discs (17, 18). Therefore, the supersoldier-like anomalies we found in P. morrisi were likely to have originated from the abnormal growth of soldier larvae and their vestigial wing discs. Based on this insight, we predicted that the evolution of the supersoldier subcaste in Pheidole occurred through developmental changes that elaborated these two features.

We tested this prediction in P. obtusospinosa and P. rhea, two species that have a supersoldier subcaste (11). As predicted, their supersoldier larvae grow larger (Fig. 2, F and J, and fig. S3, B and F) and develop two pairs of large vestigial wing discs relative to their soldier larvae (Fig. 2, G and K, and fig. S3, C and G). Furthermore, vestigial wing discs in supersoldier larvae show an elaborated pattern of sal expression in the wing pouch relative to those of soldier larvae (Fig. 2, H and L, and fig S3, D and H, black arrows). We then resolved the evolutionary history of their supersoldier subcaste by reconstructing a phylogeny of 11 Pheidole species (fig. S5). Of these, only P. obtusospinosa and P. rhea have a supersoldier subcaste. Our phylogenetic analysis suggests that the supersoldier subcaste has evolved independently, because P. rhea is one of the most basal species of this genus, whereas P. obtusospinosa is derived (fig. S5) (10). Therefore, the supersoldier subcaste has evolved in parallel, because similar developmental changes underlie its independent evolution in P. obtusospinosa and P. rhea. Furthermore, our phylogenetic analysis suggests that, relative to P. obtusospinosa, there are six basal and four derived species (fig. S5). We found that soldier larvae of these basal and derived species differ in their vestigial wing disc number and wing pouch expression of sal (fig. S6). This indicates that the supersoldier subcaste has evolved in parallel despite the evolutionary divergence of soldier development.

Application of methoprene (a JH analog) to Pheidole larvae has been shown to induce the development of unusually large soldier pupae (15). In P. morrisi, we found that methoprene can induce the development of larvae and adults that mimic the anomalous supersoldier-like individuals of P. morrisi and the supersoldiers of P. obtusospinosa and P. rhea. First, induced supersoldier larvae (Fig. 3G) and adults (Fig. 3B) are significantly larger than untreated controls (Fig. 3, D and A, and fig. S7), and several of the induced adult supersoldiers have mesothoracic wing vestiges (Fig. 3C). Second, the relative size ranges of induced supersoldiers overlap with those of anomalous and naturally produced supersoldiers (fig. S8). Finally, we found vestigial wing discs of induced supersoldier larvae (Fig. 3, H and I, and fig. S9, B to D and F to H) that mimic those of supersoldier larvae in P. obtusospinosa (Fig. 2, K and L) and P. rhea (fig. S3, G and H). Therefore, although P. morrisi lacks a supersoldier subcaste, there is a developmental potential to produce supersoldiers that can be induced through JH. Furthermore, the occurrence of supersoldier-like anomalies in P. morrisi (Fig. 2A) and other Pheidole species (16) suggests that this potential is recurrently induced in nature. This recurrent induction, which is probably mediated by JH, may be caused by nutrition, because it has been shown that environmental variation in nutrition (3) and experimentally increasing nutrition (5) produces supersoldier-like anomalies in Pheidole colonies.

Fig. 3

Methoprene-induced supersoldiers (iXSD) in P. morrisi (Pm): comparison of (A) normal SD with (B) iXSD. (C) Thorax of an iXSD. Comparison of (D) SD and (G) iXSD larvae, and of vestigial wing discs (stained with DAPI) and sal expression in (E and F) normal SD and (H and I) iXSD. White arrowheads indicate the presence of mesothoracic vestigial wing buds or discs; asterisks denote their absence. Black arrowheads indicate sal expression in the wing pouch. Adult, larval, and vestigial wing disc images are all to scale.

We discovered that this developmental potential to produce supersoldiers can be induced by methoprene in other derived (P. hyatti) and basal (P. spadonia) Pheidole species that lack a supersoldier subcaste. As in P. morrisi, we found vestigial wing discs of induced supersoldier larvae in P. hyatti (fig. S9, J and N) and P. spadonia (fig. S9, R and V) that mimic those of supersoldier larvae in P. obtusospinosa (Fig. 2, K and L) and P. rhea (fig. S3, G and H). Therefore, the developmental potential to produce supersoldiers has been retained and was probably present in the common ancestor of all Pheidole (Fig. 4). Without a priori knowledge of this ancestral developmental potential, we would have inferred that the supersoldier subcaste has evolved de novo: once in P. rhea and once in P. obtusospinosa (fig. S5) (10). However, our results support an alternative explanation for the parallel evolution of supersoldiers: The developmental potential and phenotypic expression of a novel supersoldier subcaste originated in the common ancestor of all Pheidole (Fig. 4, section i); the phenotypic expression of supersoldiers was subsequently lost, but the ancestral potential to produce them was retained (Fig. 4, section ii); and this potential was then actualized in P. obtusospinosa, leading to the re-evolution of a supersoldier subcaste (Fig. 4, section iii).

Fig. 4

Evolutionary history of ancestral developmental potential and phenotypic expression of supersoldiers (XSDs). MYA, million years ago. Purple represents the pattern of sal expression; asterisks indicate the absence of vestigial wing discs and sal expression. Green arrows and boxes represent the induction of XSD potential.

Finally, we showed that this ancestral potential was actualized in P. obtusospinosa through the re-evolution of a second JH-sensitive period mediating a soldier-supersoldier switch point (fig. S10). We applied methoprene to larvae that had passed the soldier–minor worker switch point but whose caste fate as either soldiers or supersoldiers was still undetermined. We found that applying methoprene to these larvae induced the development of a significantly greater proportion of supersoldiers (fig. S11). Collectively, our results indicate that the supersoldier subcaste in P. obtusospinosa re-evolved through genetic accommodation. This process occurs when: (i) a novel phenotype is induced and (ii) this phenotype is incorporated into the population through selection on genes that control its frequency and form of expression (19, 20). Environmental induction of the ancestral potential may have recurrently produced supersoldier-like anomalies in P. obtusospinosa. These anomalies would persist, because colonies of Pheidole generally care for and buffer anomalies against purifying selection (3). Selection on P. obtusospinosa colonies may have incorporated induced supersoldier-like anomalies by increasing their frequency through modification of the JH system (fig. S12) and by inhibiting the formation of any wing vestiges (fig. S13 and S14). Army ant raids may have been a selective pressure that incorporated these anomalies, because P. obtusospinosa supersoldiers currently use their extra-large heads to defend against these raids (13).

Selection for re-evolving supersoldiers may generally be reduced, because almost all Pheidole species lack a supersoldier subcaste (10, 11). P. hyatti provides insight into how this selective pressure can be reduced: Although P. hyatti retains the developmental potential (Fig. 4) and lives in an ecological environment similar to that of P. obtusospinosa, it has not re-evolved a supersoldier subcaste (11). Instead, P. hyatti uses nest evacuation behavior when attacked by army ants (21). The retention of this potential in P. hyatti and other Pheidole species that lack a supersoldier subcaste may therefore be due to a clade-specific constraint (22). This constraint may have arisen from having the same hormone (JH) mediate the determination of both soldiers and supersoldiers in the common ancestor of all Pheidole. Soldiers and supersoldiers are both defined by their larval size and the development of their vestigial wing discs, which indicates that their developmental programs share many modules. Therefore, the ancestral potential to produce supersoldiers cannot be lost without compromising the developmental program of soldiers.

Recurrent phenotypes reflecting ancestral potentials have long been recognized as widespread in plants and animals (6, 19, 2328). Because of the lack of empirical evidence, however, the evolutionary significance of these recurrent phenotypes has been underappreciated (19, 29). We uncovered an ancestral developmental potential to produce a novel supersoldier subcaste that has been retained throughout a hyperdiverse ant genus that evolved ~35 to 60 million years ago (10) (Fig. 4). Our results suggest that the recurrent induction of ancestral developmental potential is an important source of adaptive variation for selection that facilitates the adaptive and parallel evolution of novel phenotypes.

Supporting Online Material

Materials and Methods

Figs. S1 to S16

Tables S1 to S2

References (3146)

References and Notes

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 32.
  32. 33.
  33. 34.
  34. 35.
  35. 36.
  36. 37.
  37. 38.
  38. 39.
  39. 40.
  40. 41.
  41. 42.
  42. 43.
  43. 44.
  44. 45.
  45. 46.
  46. 47.
  47. Acknowledgments: We thank R. Sanwald, R. Johnson, S. Cover, M. Seid, W. Tschinkel, and J. King for help with ant collection or identification; Wellmark for methoprene; and T. Flatt, G. Wray, C. Metzl, S. Jolie, Abouheif lab members, and Konrad Lorenz Institute (KLI) fellows for comments. This work was supported by a Natural Sciences and Engineering Research Council grant and a KLI fellowship to E.A., as well as grants from NSF (0344946) and the Arizona Agricultural Experiment Station (ARZT 136321-H3106) to D.E.W. DNA sequences were deposited in GenBank (accession nos. JN205075 to JN205109).
View Abstract

Navigate This Article