J Comp Physiol A (2011) 197:1063–1071 DOI 10.1007/s00359-011-0667-0

ORIGINAL PAPER

Reproduction, dominance, and caste: endocrine proWles of queens and workers of the saltator

Clint A. Penick · Jürgen Liebig · Colin S. Brent

Received: 11 May 2011 / Revised: 24 June 2011 / Accepted: 30 June 2011 / Published online: 20 July 2011 © Springer-Verlag 2011

Abstract The regulation of reproduction within Introduction societies is a key component of the evolution of . DiVerential patterns of levels often underlie the At the heart of all eusocial societies is a reproductive divi- reproductive division of labor observed among colony sion of labor, where only one or a few individuals repro- members, and further task partitioning among workers is duce while the rest serve as a functionally sterile workforce also often correlated with diVerences in juvenile hormone (Wilson 1971). In simple societies, this division of labor (JH) and ecdysteroid content. We measured JH and ecdy- emerges as a result of dominance interactions among adults steroid content of workers and queens of the ant Harpegna- (West-Eberhard 1978; Fletcher and Ross 1985; Röseler thos saltator. In this species, new colonies are founded 1991; Heinze et al. 1994), while in highly eusocial species by a single queen, but after she dies workers compete in an this occurs as a consequence of the reproductive specializa- elaborate dominance tournament to decide a new group of tion of morphologically distinct queen and worker castes reproductives termed “.” Our comparisons (Oster and Wilson 1978; Hölldobler and Wilson 2009). revealed that queens, gamergates, and inside workers (non- Workers have reduced fecundity, although they often retain reproductive) did not diVer in levels of JH or ecdysteroids. functional ovaries and will begin to lay eggs in the absence However, increased JH and decreased ecdysteroid content of the queen. In order for a reproductive division of labor to was observed in outside workers exhibiting foraging behav- be maintained, activation of worker ovaries must be inhib- ior. Application of a JH analog to virgin queens of H. salta- ited so that they do not begin to lay eggs in the presence of tor, although eVective at inducing dealation, failed to an active reproductive. DiVerences in reproductive activity promote egg production. Together, these results support the observed among castes are thought to be under endocrine hypothesis that JH has lost its reproductive function in control. SpeciWcally, juvenile hormone (JH) and ecdyster- H. saltator to regulate foraging among the worker caste. oids have been implicated as regulators of reproductive physiology and behavior in social (Robinson 1987; Keywords Juvenile hormone · Ecdysone · Dominance · Robinson and Vargo 1997; Hartfelder 2000) although other Division of labor · Colony growth-rate factors, such as biogenic amines (Sasaki and Nagao 2001; Sasaki et al. 2007, 2009), have been increasingly shown to be involved in these processes. In solitary insects, JH and ecdysteroids generally function as gonadotropins and have been shown to aVect the synthe- C. A. Penick (&) · J. Liebig sis and uptake of vitellogenin, a precursor of yolk protein School of Life Sciences, Arizona State University, (Nijhout 1994; Klowden 1997; Hartfelder 2000; Raikhel Tempe, AZ 85287-4501, USA et al. 2005 Dong et al. 2008). This gonadotropic function e-mail: [email protected] has been conserved in the social bumble , Bombus ter- C. S. Brent restris, where increased JH and ecdysteroid levels are corre- US Department of Agriculture, Arid-Land Agricultural Research lated with reproduction by dominant workers and queens Center, Maricopa, AZ 85138, USA (Bloch et al. 2000; Geva et al. 2005). In contrast, JH appears 123 1064 J Comp Physiol A (2011) 197:1063–1071 to be dissociated from reproductive activity in adult queens For this study mature queens, gamergates, and workers of the highly eusocial , Apis mellifera (Robinson performing in-nest tasks or foraging were sampled from et al. 1991; Robinson et al. 1992). Instead, JH serves as a mature colonies of H. saltator to determine whole-body JH behavioral pacemaker within the worker caste, promoting and ecdysteroid content. In order to procure queen samples, the onset of foraging behavior (Jaycox et al. 1974; Robinson we established new, queenright colonies in the laboratory. 1987; Sullivan et al. 2000). This alternative role of JH as a Finally, we investigated the role of JH during the period regulator of foraging is thought to have evolved through following colony foundation by treating virgin queens with co-option of the endocrine signaling system associated with a juvenile hormone analog (JHA). We monitored the eVect reproduction in a solitary ancestor (West-Eberhard 1996; of JHA treatment on the timing of dealation and the initia- Amdam et al. 2004). Some evidence suggests a similar tion of egg laying by virgin queens. Comparisons among behavioral role for JH in other social insects (Robinson and groups allowed us to make inferences about the function of Vargo 1997; Giray et al. 2005), including , where JH and ecdysteroids with respect to caste, dominance, and increased JH levels correlate with foraging by workers of reproduction in H. saltator as well as to make comparisons the ant Myrmicaria eumenoides (Lengyel et al. 2007) as of hormonal patterns across species. well as co-founding queens of the harvester ant Pogonomyr- mex californicus (Dolezal et al. 2009). In species of social insects where colony members com- Methods pete over reproduction, diVerences in JH and ecdysteroids have been correlated with dominance. Examples in both Rearing conditions bumble bees and indicate that JH retains its gonado- tropic function, and dominant individuals are characterized Laboratory colonies of H. saltator were maintained at a by having higher levels of circulating JH (Röseler et al. constant temperature of 25°C and 12:12 light/dark cycle, 1984; Bloch et al. 2000; Tibbetts and Izzo 2009; Tibbetts and they were fed live crickets (Acheta domesticus) twice and Huang 2010). This pattern is reversed in two species of per week. The nests consisted of plastic boxes queenless ants— sp. and peet- (19 £ 27 cm) with a dental plaster Xoor, and each nest con- ersi—where JH levels are lowest in reproductive workers tained a preformed chamber covered with a glass plate and elevated in subordinates (Sommer et al. 1993; Brent (12 £ 15 cm). Over 75 colonies of H. saltator were origi- et al. 2006). These two Ponerine ant species are atypical in nally collected from Karnataka State in southern India that they lack true queens, and a single dominant worker is between 1994 and 1999, and colonies have been crossbred responsible for reproduction. The role of JH in these in the laboratory to maintain genetic diversity. worker reproductives contrasts with the role of JH found in true queens of the Wre ant, Solenopsis invicta, where JH still Determination of worker castes has a gonadotropic function (Barker 1978; Brent and Vargo 2003; Burns et al. 2007). Gamergates were identiWed in large colonies (200–300 Brent et al. (2006) proposed that queenless species workers) based on behavioral characteristics and direct evolved a separate gonadotropic regulatory mechanism observation of egg laying. Gamergates display a dominant when faced with the loss of the queen phenotype and the stance characterized by an elevated posture where the gas- physiological constraints of the worker phenotype. To test ter is raised above the level of the thorax, and subordinate this hypothesis we measured hormone content in the ant workers approach dominants with their head lowered , another species of Ponerine ant in (Liebig et al. 1998). Workers identiWed as gamergates were which both queens and fully reproductive workers occur. paint-marked (Testors Pactra® enamel, Rockford, IL, USA) Colonies are founded by a single queen, but after the queen with an individual code and observed over several days to dies, workers compete in a dominance tournament to estab- conWrm reproductive status. In all cases, colonies contained lish a new group of reproductives (Heinze et al. 1994; multiple gamergates. Gamergates used in this study had Peeters et al. 2000), and these are called “gamergates” been established at least 6 months prior to sampling, and (Peeters and Crewe 1984). Gamergates produce eggs at there was little or no observed among workers roughly half the rate of true queens (Peeters et al. 2000) and in these colonies. maintain their status by expressing a chemical fertility sig- Workers were selected from among 20 separate colo- nal similar to that of queens (Liebig et al. 2000). Because nies. Individuals classiWed as foragers were taken directly there is a low queen–worker dimorphism in this species and from the foraging arena outside the glass-covered nest they have similar behavioral repertoires (Peeters et al. chamber. These workers had a darker cuticle than in-nest 2000), a rare direct comparison of hormone levels and func- workers and responded defensively when provoked with tion can be made between castes. forceps (defensive or nest-guarding behavior), which are 123 J Comp Physiol A (2011) 197:1063–1071 1065 typical characteristics of foragers in other species (Hölldobler and Wilson 1990). Before selecting inside workers, all workers were removed from the foraging arena and held separately. One hour after their removal, any additional workers that entered the foraging arena were removed; inside workers were then chosen from those that remained in the nest chamber. Fully callow workers (newly emerged) were excluded from this study.

Queen rearing and colony foundation

Queen samples were derived from new colonies founded 5mm in the laboratory, which is relatively easy in this species (Liebig et al. 1998). Young, unmated females were taken Fig. 1 Illustration of queen (top) and worker castes of H. saltator. from their parent colonies and isolated with male pupae from There is limited size dimorphism between both castes, with queens a diVerent colony. A single queen was used to found each displaying increased thoracic development associated with wing mus- X colony except for one instance where a colony was founded culature necessary for a mating ight with eight queens grouped together. Observation of egg laying was used to conWrm that each queen was reproduc- tive. Over 2 years, 100 queen-founded colonies were estab- and the supernatant was removed. This extract was dried lished, though not all produced workers. In order to conWrm down, resuspended in 500 l of 90% methanol, and stored that queen colonies had reached a mature size at the time at ¡80°C until analysis for ecdysteroid content. queens were sampled for JH and ecdysteroid content, we The hexane portion was used to quantify JH using the censused successful foundations at several points over gas chromatography/mass spectrometry (GC–MS) method 3 years to approximate colony growth rate. Based on the of Bergot et al. (1981) as modiWed by Shu et al. (1997) and colony growth rate, it was determined that queens would be Brent and Vargo (2003). In order to Wlter-out contaminants, sampled for hormone analysis 8–15 months after colonies the hexane phase of the partitioned samples was eluted were established and contained at least 100 workers. through aluminum oxide columns successively with hex- There is a limited queen-worker dimorphism in H. salta- ane, 10% ethyl ether–hexane, and 30% ethyl ether–hexane. tor relative to other species (Fig. 1). Peeters et al. (2000) After drying, samples were derivatized by heating at 60°C found that both queens and gamergates have ovaries with for 20 min in a solution of methyl-D alcohol (Sigma- eight ovarioles, although these are more developed and Aldrich, St Louis, MO, USA) and triXuoroacetic acid active in queens than workers. While queens are capable of (Sigma-Aldrich, St Louis, MO, USA). Samples were dried ovipositing approximately two eggs per day, gamergates down then resuspended in hexane and again eluted through usually produce half that or less. These are minor diVer- aluminum oxide columns; 30% ethyl ether was used to ences relative to what is observed in most ant species remove non-derivatized components and 50% ethyl- (Hölldobler and Wilson 1990). Because these diVerences acetate–hexane was used to collect the JH derivative. After are minor, we report values that represent the total hormone drying, the sample was resuspended in hexane. The puriWed content per ant. In order to account for any diVerences and derivatized JH was then analyzed using an HP 7890A based on size with respect to caste (queen vs. worker), we Series GC (Agilent Technologies, Santa Clara, CA, USA) analyzed all hormone data using both non-adjusted values equipped with a 30 m £ 0.25 mm Zebron ZB-WAX col- as well as values standardized by dry-weight. umn (Phenomenex, Torrence, CA, USA) coupled to an HP 5975C inert mass selective detector. Helium was used as a Hormone quantiWcation carrier gas. JH form was conWrmed by Wrst running test samples in SCAN mode for known signatures of JH 0, JH I, For each sample, two individual ants of the same caste were JH II, JH III, and JH III ethyl; JH III was conWrmed as the placed in a vial containing 1 ml of methanol and crushed primary endogenous form in this species. Subsequent sam- with a glass tissue grinder. Hexane was mixed into the vial ples were analyzed using the MS SIM mode, monitoring at W with methanol (90% HPLC grade), centrifuged, and the m/z 76 and 225 to ensure speci city for the d3-methoxyhy- hexane layer was pulled oV to extract JH dissolved in this drin derivative of JH III. Total abundance was quantiWed layer while ecdysteroid remained in the methanol portion. against a standard curve of derivatized JH III. The detection The methanol layer was centrifuged at 3,500 rpm for 5 min, limit of the assay is approximately 1 pg.

123 1066 J Comp Physiol A (2011) 197:1063–1071

Ecdysteroid content was determined by a competition size diVerences (dry-weight) between queens and workers radioimmunoassay originally developed by Warren et al. as well as non-adjusted values. Both analyses converged on (1984) as modiWed by Brent et al. (2006). Duplicate 10 l the same pattern and did not aVect the outcome of our sta- aliquots of the methanol phase of each partitioned sample tistical comparisons, so we report p-values based on non- were incubated overnight with 100 l of (3H)-20-hydroxy- adjusted measurements and report these in all Wgures. Addi- ecdysone stock (85.2 Ci/mmol, Perkin-Elmer, Waltham, tionally, the non-parametric Mann–Whitney U test was MA, USA) in borate buVer, and 100 l of a polyclonal used to compare the timing of dealation and the initiation of ecdysteroid antiserum (H-22 antibody, L. Gilbert, UNC- egg laying among females treated with JHA or the control. CH) at 4°C on an orbital shaker. The speciWc ecdysteroid is Non-parametric tests were used in all cases due to hetero- unknown for H. saltator, but the H-22 antibody cross-reacts scedasticity (Brown-Forsythe test, p < 0.05) and a general with ecdysone, ecdysterone, 20-hydroxyecdysone and lack of normality. makisterone A (Warren and Gilbert 1986). Intra- and inter- assay variability was minimized by generating standard competition curves for each set of samples, using 20- Results hydroxyecdysone stock (Sigma-Aldrich, St Louis, MO, USA) over a range of 15.6–2,000 pg. After incubating Queen foundations and colony growth »18 h, 20 l of cleaned protein A solution (Pansorbin; CalBiochem, San Diego, CA, USA) was added to each To conWrm that the queens used for the hormone bioassays tube to precipitate the antibody–antigen complex. Tubes were mature and reproductively active, we tracked colony were incubated for 1 h at 27°C and then centrifuged at growth of our queen-founded laboratory colonies. Colony 5,000g. Supernatant was removed and the remaining pellet growth during the Wrst year after foundation was exponen-  V 0.0113t was washed twice with 100 l of borate bu er. The incor- tial (Fig. 2a), represented by the function Nt =1.94e . poration of microlabel was determined by a 2450 MicroBe- Over this Wrst year, colony age explained 75% of the varia- ta2 scintillation counter (Perkin-Elmer, Waltham, MA, tion in colony size. Over 3 years, growth-rate was best

USA) and ecdysteroid concentrations were estimated by approximated by the logarithmic function Nt = 117 £ nonlinear regression (Brent et al. 2006). ln(t) ¡ 595, where Nt represents the colony size at time t (Fig. 2b). During this period, colony age explained 69% JHA treatment of virgin alates of the variation in colony size. The Wrst workers did not eclose until 100 days after queens were isolated with male Virgin alate females were separated from their parent col- pupae in new nests. This was due in part to the time ony 3–6 weeks after eclosing. Individuals were treated top- required for females to mate, drop their wings, and begin ically (dorsal surface of the thorax) with 10g of JHA egg laying. In addition to workers and males, 5% of suc- (pyriproxyfen, technical grade, 98.9%) dissolved in 1 l of cessful colonies were observed to produce female sexuals acetone. Controls were treated with 1 l of the acetone sol- during their Wrst year, 10% of colonies by the second year, vent alone. Female alates were kept in an individual box and a total of 29% of colonies by the third year. These data with a dental plaster Xoor and were fed small crickets twice conWrmed that most colonies had reached a mature size at per week. Observations were made daily to determine when the end of the Wrst year, which corresponded to the time individuals dropped their wings and when they began egg when queens were sampled for hormone analysis. laying. An individual was considered to have begun egg laying when an egg was observed in their nest on each of JH and ecdysteroid hormone content two consecutive days to help exclude the possibility of counting non-viable, trophic eggs. JH content was elevated in foragers of H. saltator, but did not diVer among inside workers, gamergates, or queens Statistical analysis (Kruskal–Wallis, df =3, N =57, H = 26.24, p < 0.0001; multiple comparisons test: gamergate vs. forager, DiVerences in hormone content among groups were com- p = 0.0002; queen vs. forager, p = 0.0004; inside worker vs. pared using a Kruskal–Wallis ANOVA and post-hoc analy- forager, p = 0.009; vs. inside worker, p =0.94; ses in Statistica version 9 (StatSoft, Tulsa, OK, USA). gamergate vs. queen, p = 1.0; inside worker vs. queen, When outliers were excluded (points that fell more than 1.5 p = 0.79) (Fig. 3a). The opposite pattern was observed with times outside of the interquartile range), this did not change respect to ecdysteroids, where whole-body ecdysteroid con- signiWcant diVerences, and outliers are included in Wgures tent was depressed in foragers but did not diVer among where they occurred. For queen samples, statistical com- inside workers, gamergates, or queens (Kruskal–Wallis, parisons were made using concentrations standardized for df =3, N =60, H = 27.5, p < 0.0001; multiple comparisons 123 J Comp Physiol A (2011) 197:1063–1071 1067

120 (a) (a) 50 100 y = 1.94e0.0113x B, 21 R² = 0.75 80 20

15 60

10 40 JH content (pg/ant) A, 18 A, 8 Number of workers 5 A, 10 20 0 0 Queen Gamergate Inside worker Forager 120 160 200 240 280 320 360 Colony age (days) (b) 8000 400 (b) 350 6000 A, 10 A, 18 y = 117ln(x) - 595 300 R² = 0.69 A, 15 250 4000

200 2000 B, 17 150 Ecdysteroid content (pg/ant)

Number of workers 100 0 Queen Gamergate Inside worker Forager 50 Fig. 3 Median, 25–75% conWdence interval, and non-outlier range of 0 a JH and b ecdysteroid content of individual queens, gamergates, in- 100 200 300 400 500 600 700 800 900 1000 side workers, and foragers. Sample size is noted for each group, and Colony age (days) letters indicate signiWcant diVerences (p <0.05)

Fig. 2 Total worker number as a function of colony age (days since colonies were founded in the laboratory). a Represents growth during control group began egg laying up to 13 days before they the Wrst year, which is best approximated by an exponential function, and b represents growth over 3 years, which is best approximated by a dropped their wings. Of the 12 individuals treated with logarithmic function JHA, 3 died prior to egg laying, and these were excluded from our analysis. test: gamergate vs. forager, p = 0.0032; queen vs. forager, p < 0.00001; inside worker vs. forager, p = 0.0036; gamer- gate vs. inside worker, p = 1.00; gamergate vs. queen, Discussion p = 0.436; inside worker vs. queen, p =0.253) (Fig.3b). This study represents the Wrst time that JH and ecdysteroid JHA treatment of virgin alates content of queens and reproductive workers have been compared within the same ant species. Although JH is nor- To better understand the function of JH in adult queens, mally associated with egg production in solitary insects, as we treated virgin queens with JHA and measured the well as the highly eusocial ant Solenopsis invicta (Brent eVects on the timing of both dealation and oviposition. and Vargo 2003), it appears to have lost this function in Virgin alates treated with JHA dropped their wings earlier H. saltator. Queens and gamergates had low JH content, (median 1 day) compared with females that received the similar to that observed with inside workers. Instead, control treatment (median 5 days) (Mann–Whitney U test; increased JH content was associated with foraging behavior N =12, 11; Z = 2.89; p = 0.0034) (Fig. 4a); however, by non-reproductive workers. This pattern resembles what treatment did not aVect the time required for the initiation has been observed in the honey bee, Apis mellifera, where of egg laying (Mann–Whitney U test; N =9, 11; JH no longer drives oogenesis in adult queens, but has been Z = ¡0.84; p = 0.401) (Fig. 4b). In order to initiate egg implicated in modulating foraging behavior in workers laying, dealation was not required; 3 of 11 females in the (Robinson and Vargo 1997). 123 1068 J Comp Physiol A (2011) 197:1063–1071

40 (a) 40 (b) Therefore, hormone samples were collected roughly 1 year N.S. after colonies were established, when queens would be fully mature and regularly producing eggs. B, 11 11 30 30 9 Endocrine diVerences among reproductive and behavioral castes 20 20 The subsequent hormonal analyses compared JH and ecdy- steroid content between diVerent castes to determine if days until de-alation days 10 until egg-laying days 10 there were endocrine associations with reproductive and A, 12 dominance status. JH synthesis is inversely correlated with ovarian activity in the queenless ants Diacamma (Sommer 0 0 et al. 1993) and Streblognathus peetersi (Brent et al. 2006). JHA control JHA control This is the reverse pattern found in Solenopsis invicta, a Fig. 4 Median, 25–75% conWdence interval, and range of time (in species that has a true queen caste (Brent and Vargo 2003). days) until a dealation and b Wrst oviposition after virgin queens were A proposed explanation for this diVerence was that queen- separated from their natal nest and either treated with JHA or an ace- less species had to develop an alternative means for con- tone control. Sample size is noted for each group, and letters indicate trolling reproductive activity under constraints created by signiWcant diVerences (p <0.05) their worker phenotype (Brent et al. 2006). Because H. sal- tator has both queens and reproductive workers that regu- Colony foundation and growth rate larly replace the queen upon her death, we were able to test this hypothesis, but we found no support for reproductive Prior to examining the hormonal dynamics within H. salta- workers relying on an alternative regulatory mechanism. In tor colonies, it was Wrst necessary to determine an optimal this species, both queens and reproductive workers exhib- time to sample queens and to establish that dynamics within ited the same hormonal pattern—low JH and high ecdyster- laboratory colonies closely mimicked that of Weld colonies. oid content (Fig. 3)—as previously observed in After following the development of colonies founded by reproductive workers of queenless ants (Sommer et al. single queens for 3 years, it was determined that mature 1993; Brent et al. 2006). H. saltator belongs to the rela- colony size was reached after 1 year. At this time the num- tively basal subfamily , which also includes the ber of workers present (Fig. 2) fell within the range of 58– queenless species Diacamma and S. peetersi. This suggests 124 workers observed in mature Weld colonies (Peeters that the loss of JH’s typical gonadotropic function (Raikhel et al. 2000). et al. 2005) may be a common feature of Ponerine ants, not Similar to colony growth patterns observed in Solenopsis just the queenless species. invicta (Tschinkel 1988), early colony growth in H. saltator H. saltator colony dynamics are structured upon a stable was exponential during the Wrst year but was best described with clear rank diVerences in behav- by a logarithmic function when recorded over multiple ior, ovarian activity, and the expression of a chemical fertil- years. Several colonies in our study produced female sexu- ity signal (Liebig et al. 2000; Peeters et al. 2000). In S. als during the Wrst year they were established. This corre- peetersi, elevated JH is associated with lower rank (Brent sponds to data from the Weld where relatively small et al. 2006) and decreased expression of a cuticular fertility colonies were found to produce female sexuals (Peeters signal (Cuvillier-Hot et al. 2002). In H. saltator, there was et al. 2000). Many social insect species pass through an no endocrine diVerence between high-ranking gamergates ergonomic phase of colony growth where sexual produc- and the subordinate inside workers. Although we cannot tion is deferred several years until colonies reach a thresh- exclude the possibility of small hormonal diVerences that old size (Oster and Wilson 1978). From our data on colony were undetectable with our assay, it is unlikely that a subtle growth and queen production it appears that colonies of H. disparity could produce such divergent reproductive activ- saltator are capable of producing sexuals during their Wrst ity or fertility signaling. year and may not pass through an extended ergonomic The only signiWcant diVerence among castes was for phase. This rapid development may have evolved to workers venturing outside the nest to forage, which, rela- counter random factors that aVect colony survival (Liebig tive to the other groups, had elevated JH and depressed and Poethke 2004). The similar development patterns of ecdysteroid content. This pattern resembles what has been laboratory and Weld colonies also indicate that the hormone observed in the highly eusocial honey bee, Apis mellifera, proWles determined in this experiment are likely to be com- where JH has lost its reproductive function in adult queens parable to what would occur under natural conditions. and has a role in modulating foraging in the worker caste 123 J Comp Physiol A (2011) 197:1063–1071 1069

(Robinson and Vargo 1997). JH has also been implicated as that increased JH is more strongly correlated with forag- a releaser of foraging behavior in the young founding ing behavior among subordinate workers. This contrasts queens (Dolezal et al. 2009) and workers (Dolezal et al., with other societies that maintain dominance hierarchies, submitted) of the ant Pogonomyrmex californicus, and in such as wasps and bumble bees, where dominance and workers of Myrmicaria eumenoides (Lengyel et al. 2007). reproduction are associated with increased levels of JH Our results Wt this model, so it appears that JH may regulate and ecdysteroids (Bloch et al. 2000; Geva et al. 2005; foraging in H. saltator as well. With respect to ecdyster- Giray et al. 2005; Tibbetts and Huang 2010). Instead, oids, the fall in ecdysteroid content in foragers may be H. saltator resembles the pattern observed in other highly related to the degradation of the ovaries, which occurs as eusocial species, where JH has lost its reproductive func- workers age and begin foraging (Hölldobler and Wilson tion and regulates a strong temporal polyethism among 1990). Ecdysteroids are primarily produced by the ovaries the worker caste (Hartfelder 2000; Lengyel et al. 2007). in adult female insects (Nijhout 1994), including honey bee Our results show that JH may be involved in worker divi- workers (Amdam et al. 2010). In bumble bees, hemolymph sion-of-labor even in species where age-based polyethism ecdysteroid titers are low while ovarian ecdysteroid content is weak and where individuals still compete directly over is high (Geva et al. 2005). Similar to bumble bees, the pat- reproduction. Within the eusocial , regulation tern of ecdysteroid content we observed in H. saltator may of foraging in wasps, bees, and ants shows convergence in indicate that ecdysteroids serve an intra-ovarian function the pattern of JH associated with division-of-labor among (paracrine) rather than having an endocrine role. workers despite diVerent levels of caste specialization (Hartfelder 2000). JH during the founding stage Finally, we are able to show that the loss of a reproduc- tive function for JH in Ponerine ants is not a consequence While we did not Wnd diVerences in JH or ecdysteroid con- of specialization in worker reproduction, since queens tent between mature reproductives and young workers, we show the same relationship between JH and reproduction did Wnd evidence that JH is involved in changes that occur in H. saltator. While increased JH production is associ- during the founding stage, when queens Wrst activate their ated with reproduction by queens of the Wre ant Solenop- ovaries and begin egg laying (Fig. 4). Application of JHA sis invicta (Brent and Vargo 2003), our results clearly caused alate queens to drop their wings earlier than indicate that this pattern is not a general feature of all ant untreated females. A similar eVect was found in the Wre ant, species. S. invicta displays advanced characteristics Solenopsis invicta, where JHA induced early dealation and among ants, including large colony size and high queen- oviposition in virgin queens (Vargo and Laurel 1994). worker dimorphism, in contrast to Ponerine species which Likewise, Brent and Vargo (2003) found that dealation and more closely resemble the ancestral form. Therefore, it is the initiation of oviposition in queens of this species corre- possible that the association of increased JH with repro- sponded with peak rates of JH biosynthesis. Dealation is duction in S. invicta is a derived characteristic. The repro- known to trigger Xight muscle histolysis and oocyte activa- ductive function of JH in S. invicta may be related to the tion in other insect species (Tanaka 1986). Treatment with extreme fecundity of S. invicta queens, which have a JHA did not, however, aVect the timing of oviposition in H. dramatically increased egg-laying rate compared with saltator. This result provides further evidence that JH has queens or gamergates of Ponerine species. Also, repro- lost its reproductive function in H. saltator. The role of JH ductive conXict between adult workers and queens in during the founding stage may not correlate with reproduc- S. invicta is almost completely absent because workers tive activation directly, but because it aVects the timing of are sterile, and this may allow JH to regain a reproductive dealation it may still be an important regulator of behavior function in queens without aVecting the role of JH in and other physiological processes associated with colony workers. Future investigations across other subfamilies founding. While JHA treatment of reproductive workers of may help deWne a general pattern of caste-associated S. peetersi inhibited ovarian activity, we found no such diVerences in JH and ecdysteroid levels within the euso- eVect in virgin queens of H. saltator. This may reXect a cial Hymenoptera and help to elucidate the physiological diVerence in the function of JH in founding queens com- changes that occurred during the transition from solitary pared with mature reproductives. to social life.

Acknowledgments We thank Adam Dolezal, Charity Bashore, and Conclusion Dan Langhorst for help with sample processing, and Kevin Haight, Adrian Smith, and Dani Moore for comments and advice. Mention of trade names or commercial products in this article is solely for the pur- Overall, our results support the hypothesis that JH has lost pose of providing speciWc information and does not imply recommen- a gonadotropic function in the basal Ponerine ants and dation or endorsement by the US Department of Agriculture. 123 1070 J Comp Physiol A (2011) 197:1063–1071

References Liebig J, Poethke HJ (2004) Queen lifespan and colony longevity in the ant Harpegnathos saltator. Ecol Entomol 29:203–207 Adkins-Regan E (2005) and social behavior. Prince- Liebig J, Hölldobler B, Peeters C (1998) Are ant workers capable of ton University Press, Princeton colony foundation? Naturwissenschaften 85:133–135 Amdam GV, Norberg K, Fondrk MK, Page RE (2004) Reproductive Liebig J, Peeters C, Oldham NJ, Markstadter C, Hölldobler B (2000) ground plan may mediate colony-level selection eVects on indi- Are variations in cuticular hydrocarbons of queens and workers a vidual foraging behavior in honey bees. Proc Nat Acad Sci USA reliable signal of fertility in the ant Harpegnathos saltator? Proc 101:11350–11355 Natl Acad Sci USA 97:4124–4131 Amdam GV, Page RE, Fondrk MK, Brent CS (2010) Hormone response Nijhout HF (1994) Insect hormones. Princeton University Press, to bidirectional selection on social behavior. Evol Dev 12:428–436 Princeton Barker JF (1978) Neuroendocrine regulation of oocyte maturation in Oster GF, Wilson EO (1978) Caste and ecology in the social insects. imported Wre ant Solenopsis invicta. Gen Comp Endocrinol Princeton University Press, Princeton 35:234–237 Peeters C, Crewe R (1984) Insemination controls the reproductive divi- Bergot B, RatcliV M, Schooley D (1981) Method for quantitative sion of labour in a ponerine ant. Naturwissenschaften 71:50–51 determination of the four known juvenile hormones in insect Peeters C, Liebig J, Hölldobler B (2000) Sexual reproduction by both tissue using gas chromatography—mass spectroscopy. queens and workers in the ponerine ant Harpegnathos saltator. J Chromatogr A 204:231–244 Insect Soc 47:325–332 Bloch G, Borst D, Huang Z, Robinson G, Cnaani J, Hefetz A (2000) Raikhel AS, Brown M, Belles X (2005) Hormonal control of reproduc- Juvenile hormone titers, juvenile hormone biosynthesis, ovarian tive processes. In: Gilbert L, Gill S, Iatrou K (eds) Comprehensive development and social environment in Bombus terrestris. molecular insect science, vol 3, Elsevier Press, Amsterdam J Insect Physiol 46:47–57 pp 433–491 Brent CS, Vargo EL (2003) Changes in juvenile hormone biosynthetic Robinson GE (1987) Regulation of honey bee age polyethism by juve- rate and whole body content in maturing virgin queens of Solen- nile hormone. Behav Ecol Sociobiol 20:329–338 opsis invicta. J Insect Physiol 49:967–974 Robinson GE, Vargo EL (1997) Juvenile hormone in adult eusocial Brent C, Peeters C, Dietmann V, Crewe R, Vargo E (2006) Hormonal Hymenoptera: gonadotropin and behavioral pacemaker. Arch correlates of reproductive status in the queenless ponerine ant, Insect Biochem Physiol 35:559–583 Streblognathus peetersi. J Comp Physiol A 192:315–320 Robinson GE, Strambi C, Strambi A, Feldlaufer MF (1991) Compari- Burns SN, Vander Meer R, Teal PEA (2007) Juvenile hormone III con- son of juvenile hormone and ecdysteroid haemolymph titres in centrations in female reproductives of Solenopsis invicta Buren. adult worker and queen honey bees (Apis mellifera). J Insect J Entomol Sci 42:306–309 Physiol 37:929–935 Cuvillier-Hot V, Gadagkar R, Peeters C, Cobb M (2002) Regulation of Robinson GE, Strambi C, Strambi A, Huang ZY (1992) Reproduction in reproduction in a queenless ant: aggression, and worker honey bees is associated with low juvenile hormone titers reduction in conXict. Proc Roy Soc Lond B 269:1295–1300 and rates of biosynthesis. Gen Comp Endocrinol 87:471–480 Dolezal AG, Brent CS, Gadau J, Hölldobler B, Amdam GV (2009) Röseler P (1991) Reproductive competition during colony establish- Endocrine physiology of the division of labour in Pogonomyrmex ment. In: Ross KG, Matthews RW (eds) The social biology of californicus founding queens. Anim Behav 77:1005–1010 wasps. Cornell University Press, Ithaca, pp 309–335 X Dong S-Z, Ye G-Y, Guo J-Y, Hu C (2008) Role of ecdysteroid and Röseler P, Röseler I, Strambi A, Augier R (1984) In uence of insect juvenile hormone in vitellogenesis in an endoparasitic , hormones on the establishment of dominance hierarchies among Pteromalus puparum (Hymenoptera: Pteromalidae). Gen Comp foundresses of the paper wasp, Polistes gallicus. Behav Ecol Endocr 160:102–108 Sociobiol 15:133–142 Fletcher D, Ross K (1985) Regulation of reproduction in eusocial Sasaki K, Nagao T (2001) Distribution and levels of dopamine and its Hymenoptera. Ann Rev Entomol 30:319–343 metabolites in brains of reproductive workers in honeybees. Geva S, Hartfelder K, Bloch G (2005) Reproductive division of labor, J Insect Physiol 47:1205–1216 dominance, and ecdysteroid levels in hemolymph and ovary of Sasaki K, Yamasaki K, Nagao T (2007) Neuro-endocrine correlates of the bumble bee Bombus terrestris. J Insect Physiol 51:811–823 ovarian development and egg-laying behaviors in the primitively Giray T, Giovanetti M, West-Eberhard M (2005) Juvenile hormone, eusocial wasp (Polistes chinensis). J Insect Physiol 53:940–949 reproduction, and worker behavior in the neotropical social wasp Sasaki K, Yamasaki K, Tsuchida K, Nagao T (2009) Gonadotropic V Polistes canadensis. Proc Natl Acad Sci USA 102:3330 e ects of dopamine in isolated workers of the primitively eusocial Hartfelder K (2000) Insect juvenile hormone: from “status quo” to high wasp, Polistes chinensis. Naturwissenschaften 96:625–629 society. Braz J Med Biol Res 33:157–177 Shu S, Ihl Park Y, Ramaswamy S, Srinivasan A (1997) Hemolymph Heinze J, Hölldobler B, Peeters C (1994) ConXict and cooperation in juvenile hormone titers in pupal and adult stages of southwestern ant societies. Naturwissenschaften 81:489–497 corn borer [Diatraea grandiosella (Pyralidae)] and relationship Hölldobler B, Wilson EO (1990) . Harvard University Press, with egg development. J Insect Physiol 43:719–726 Cambridge Sommer K, Hölldobler B, Rembold H (1993) Behavioral and physio- Hölldobler B, Wilson EO (2009) The : the beauty logical aspects of reproductive control in a Diacamma species elegance and strangeness of insect societies. WW Norton & from Malaysia (Formicidae, Ponerinae). 94:162–170 Company, New York Sullivan JP, Jassim O, Fahrbach SE, Robinson GE (2000) Juvenile hor- Jaycox ER, Skowrone W, Guynn G (1974) Behavioral changes in mone paces behavioral development in the adult worker honey worker honey bees (Apis mellifera) induced by injections of a bee. Hormon Behav 37:1–14 X juvenile hormone mimic. Ann Entomol Soc Am 67:529–535 Tanaka S (1986) De-alation, ight muscle histolysis, and oocyte devel- Klowden MJ (1997) Endocrine aspects of mosquito reproduction. opment in the striped ground cricket, Allonemobius fasciatus. Arch Insect Biochem Physiol 35:491–512 Physiol Entomol 11:453–458 Lengyel F, Westerlund SA, Kaib M (2007) Juvenile hormone III inXu- Tibbetts EA, Huang ZY (2010) The challenge hypothesis in an insect: X ences task-speciWc cuticular hydrocarbon proWle changes in the juvenile hormone increases during reproductive con ict follow- ant Myrmicaria eumenoides. J Chem Ecol 33:167–181 ing queen loss in Polistes wasps. Am Nat 176:123–130

123 J Comp Physiol A (2011) 197:1063–1071 1071

Tibbetts EA, Izzo AS (2009) Endocrine mediated phenotypic plastic- Warren J, Smith W, Gilbert L (1984) SimpliWcation of the ecdysteroid ity: condition-dependent eVects of juvenile hormone on domi- radioimmunoassay by the use of protein A from Staphylococcus nance and fertility of wasp queens. Hormon Behav 56:527–531 aureus. Experientia 40:393–394 Tschinkel WR (1988) Colony growth and the ontogeny of worker West-Eberhard MJ (1978) Polygyny and the evolution of social behav- polymorphism in the Wre ant, Solenopsis invicta. Behav Ecol ior in wasps. J Kansas Entomol Soc 51:832–856 Sociobiol 22:103–115 West-Eberhard MJ (1996) Wasp societies as microcosms for the study Vargo E, Laurel M (1994) Studies on the mode of action of a queen of development and evolution. In: Turillazzi S, West-Eberhard primer of the Wre ant Solenopsis invicta. J Insect MJ (eds) Natural history and evolution of paper wasps. Oxford Physiol 40:601–610 University Press, New York, pp 290–317 Warren J, Gilbert L (1986) Ecdysone metabolism and distribution Wilson EO (1971) The insect societies. Belknap Press of Harvard during the pupal-adult development of Manduca sexta. Insect University Press, Cambridge Biochem 16:65–82

123