The role of juvenile hormone in dominance behavior, reproduction and cuticular pheromone signaling in the caste-flexible epiponine wasp, Synoeca surinama

Abstract

Background: The popular view on insect sociality is that of a harmonious division of labor among two morphologically distinct and functionally non-overlapping castes. But this is a highly derived state and not a prerequisite for a functional society. Rather, caste-flexibility is a central feature in many eusocial wasps, where adult females have the potential to become queens or workers, depending on the social environment. In non-swarming paper wasps (e.g., Polistes), prospective queens fight one another to assert their dominance, with losers becoming workers if they remain on the nest. This aggression is fueled by juvenile hormone (JH) and ecdysteroids, major factors involved in caste differentiation in most eusocial insects. We tested whether these hormones have conserved aggression-promoting functions in Synoeca surinama, a caste-flexible swarm-founding wasp (Epiponini) where reproductive competition is high and aggressive displays are common.

Results: We observed the behavioral interactions of S. surinama females in field nests before and after we had removed the egg-laying queen(s). We measured the ovarian reproductive status, hemolymph JH and ecdysteroid titers, ovarian ecdysteroid content, and analyzed the cuticular hydrocarbon (CHC) composition of females engaged in competitive interactions in both queenright and queenless contexts. These data, in combination with hormone manipulation experiments, revealed that neither JH nor ecdysteroids are necessary for the expression of dominance behaviors in S. surinama. Instead, we show that JH likely functions as a gonadotropin and directly modifies the cuticular hydrocarbon blend of young workers to match that of a reproductive. Hemolymph ecdysteroids, in contrast, are not different between queens and workers despite great differences in ovarian ecdysteroid content.

Conclusions: The endocrine profile of S. surinama shows surprising differences from those of other caste-flexible wasps, although a rise in JH titers in replacement queens is a common theme. Extensive remodeling of hormone functions is also evident in the highly eusocial bees, which has been attributed to the evolution of morphologically defined castes. Our results show that hormones which regulate caste-plasticity can lose these roles even while caste-plasticity is preserved.

Keywords: Cuticular hydrocarbons; Ecdysteroids; Endocrine; Epiponini; Juvenile hormone; Swarm founding; Wasps.

Figure 1

Nests of Synoeca surinama. (A) Worker performing the “queen dance” and the queen responding with a pronounced bending display. (B) Aggressive defense response in a three compartment nest. The hole in the side was an observation window, and the opening on the bottom was incidental, leading to the emergence of colony defenders. (C) A single compartment nest with the envelope removed. Most nests were studied this way. (D) A massive 6 or 7 compartment nest discovered in Chapada Diamantina, Bahia, Brazil. Nest expansion proceeds upwards (yellow arrow). Red arrows indicate nest entrance/exit in B-D.

Figure 2

Ovarian development in young females. (A) Time course of ovarian maturation in young queenright (QR) and queenless (QL) females from two nests. The average of the longest two primary oocytes is indicated, so that each point represents an individual ovary. Dotted line indicates filamentous ovaries lacking measurable oocytes (<8% LME). (B) Typical ovaries of a newly emerged wasp (day 0) and those of older queenright and queenless females. Oocytes and trophocytes (“trph”) are indicated.

Figure 3

JH titers according to female status. A Mixed Model analysis of samples taken from colonies 4–10, using up to two fixed factors (Female Status and Colony Condition) and a random one (Colony). Black, white and gray bars indicate samples from queenright (QR) conditions, queenless (QL) conditions and a combination of samples from QR and QL conditions, respectively. Least Squared Means (±SE) are shown along with sample size. (A) In QR conditions, JH differed between the female groups (F4,186=63.29, P<0.0001). Queens (Q) had higher titers than newly emerged (NE) females (t=−8.34), new benders without history of working (B) (t=−6.52), 1–3 day old pre-working and working females (pW) (t=−9.85), and workers >3 days old (W) (t=−15.35)). Also, pW females had higher JH titers than workers (t=−2.14). (B) In QL conditions, 1–3 days after queen removal, JH levels differed between the female groups (F3,109=20.32, P<0.0001). Workers that transitioned into benders (W➔B) had higher JH titers than NE females (t=−4.09), bending females which were never observed to work (t=−2.61), and workers (t=−7.53). The latter bending females also had higher JH titer than NE females (t=−2.13) and workers (t=−2.75). (C) In QL conditions, 7–8 days after queen removal, JH differed between the female groups (F3,36=5.73, P=0.003). Established benders (Est B) had higher JH titers than all female groups (vs. NE females (t=−2.42); vs. suppressed (S) females (i.e., attacked benders and huddlers) (t=−3.23); vs. workers (t=−3.85). (D) When QR and QL groups were pooled, with females classified as either queens, non-queen benders of types (ΣB), 0–3 day old non-benders (nB) and workers, JH differed according to Female Status (F3,341=65.83, P<0.0001), Colony Condition (F1,341=16.18, P<0.0001) and their interaction (F2,341=11.33, P<0.0001). Queens had higher JH titers than all other groups, and Benders had higher JH titers than nB females and workers. For all models, Colony had no effect.

Figure 4

Ecdysteroids levels in ovaries and hemolymph. Ovarian ecdysteroid contents correlate positively with: (A) ovary size (number of oocytes >70% length of mature egg) in queens from colonies 6–10 (N = 36, Spearman’s ρ = 0.51, P = 0.0014), and (B) oocyte length in workers (+) and benders (o) from colony 4 (N = 15; Spearman’s ρ = 0.87; P <0.0001). (C) Hemolymph ecdysteroid titers did not statistically differ between queens and workers in colony 8 (Kruskal-Wallis test: N = 21, H₁ = 0.18, P = 0.86, Z_r = 0.075). Box plots show the median and the middle two quartiles; the whiskers indicate the 1.5 interquartile range.

Figure 5

Hormone manipulation experiments. (A) JH III treatment of young queenless workers. JH III treated females showed no difference in bending propensity compared to cyclohexane-treated controls. (B) Effect of methoprene treatment on oocyte development in young queenright females from colonies 11 and 12. Each data point represents 1 of a possible 6 measurable primary oocytes from treated and non-treated females (the horizontal line represents the grand mean shown for each nest). In colony 11, methoprene-treated females had longer oocytes than both control types (Steel-Dwass all pairs: vs. non-treated: N = 111, Z = 3.25, P = 0.003 = **; vs. cyclohexane: N = 166, Z = 4, P = 0.0002 = ***). In colony 12, where a comparison was made with only a solvent control, the difference was less obvious but still significant (Mann–Whitney U test: N = 165, Z = 2.12, P = 0.03 = *).

Figure 6

Cuticular hydrocarbon (CHC) profiles of Synoeca surinama . (A) Representative chromatograms of whole-body extracts for main female types. For compound numbering and identification see Additional file 6: Table S1. (B) CHC profiles of newly emerged females (NE), queens (Q) and workers (>4 days since eclosion) (W) from colonies 4–8 (mean % ± SD). Discriminant analyses were performed for queens vs. workers (w/out PCA: Global Wilks’ λ =0.01489; F_16,145 = 599.68; p < 0.001; w/ PCA: Global Wilks’ λ = 0.03855; F_11,150 = 340.06; p < 0.001), and newly emerged females vs. pooled queens + workers (w/out PCA: Global Wilks’ λ = 0.00148; F_34,334 = 245.96; P < 0.001; w/ PCA: Global Wilks’ λ = 0.15546; F_15,170 = 61.570; p < 0.001). White and black stars indicate the main contributors of separation prior to and following PCA, respectively. “X” denotes a compound removed based on the PCA analysis. (C) Canonical scatterplot based on a discriminatory analysis for all three groups.

Figure 7

Age and status related changes in cuticular hydrocarbon (CHC) profiles of Synoeca surinama . (A) Relative percentages of CHC compounds in females in queenless (top) and queenright (bottom) conditions from colonies 5 and 7 (mean % ± SD). Linear alkanes are shown in white circles; alkenes are shown in black circles (for identities, see Additional file 6: Table S1). The surrounding arrows indicate evident changes in a compound’s proportional representation over the first 8 or 9 days of adult life and also show a difference between queen and workers (see Figure 6). Small arrows specify an age-specific shift, thick black arrows indicate a queen-like shift and thick gray arrows signify a worker-like shift. (B) Canonical scatterplot based on a discriminant analysis of CHC profiles for females belonging to colonies 6–8. daqr = days after queen removal.

Figure 8

Effect of the methoprene on the cuticular hydrocarbon (CHC) profile of Synoeca surinama. (A) Methoprene application induced a CHC profile similar to that of incipient reproductives. The CHC profile of methoprene-treated females was distinct from cyclohexane- and non-treated females (pooled from colonies 11 and 12). For compound identification, see Additional file 6: Table S1. White and black stars indicate the main contributors of separation prior to and following PCA, respectively. “X” denotes a compound removed based on the PCA analysis. (B) Canonical scatterplot based on a discriminatory analysis for all treatment groups from colonies 11 and 12, including established benders and queenright workers from colonies 6 and 7 showing a clearly significant difference, done with and without PCA (for both: Global Wilks’ λ = 0.185; F_13,51 = 17.3; p <0.001; 100% predicted classification). The non-treated controls group within the cyclohexane-treated wasps whereas methoprene-treated females overlap with the established benders, but not the queenright workers.