- Email: [email protected]
Cuticular lipid profiles of fertile and non-fertile Cardiocondyla ant queens
Journal of Insect Physiology 58 (2012) 1245–1249
Contents lists available at SciVerse ScienceDirect
Journal of Insect Physiology journal homepage: www.elsevier.com/locate/jinsphys
Cuticular lipid profiles of fertile and non-fertile Cardiocondyla ant queens Stefanie Will a, Jacques H.C. Delabie b, Jürgen Heinze a, Joachim Ruther c, Jan Oettler a,⇑ a
Universität Regensburg, Biologie I, D-93040 Regensburg, Germany Laboratório de Mirmecologia, CEPEC/CEPLAC, Cx.P. 7, Itabuna, BA, Brazil c Universität Regensburg, Chemical Ecology Group, D-93040 Regensburg, Germany b
a r t i c l e
i n f o
Article history: Received 22 March 2012 Received in revised form 18 June 2012 Accepted 19 June 2012 Available online 28 June 2012 Keywords: Cuticular hydrocarbons Queen pheromone Fertility signal Mating Formicidae
a b s t r a c t Both mating and reproduction strongly affect the physiology of insect females. In the ant Cardiocondyla obscurior, a comparison among virgin queens, mated queens, and queens mated with sterilized males (‘‘sham-mated’’) allows to separate the different effects of mating and egg laying. Here, we investigate whether and how different mating status is reflected in the cuticular lipid profiles of queens, i.e., the blend of chemicals that is thought to signal a queen’s fertility. Surprisingly, discriminant analyses failed to reliably distinguish among virgin, mated, and sham-mated queens. A generalized linear model on individual substances showed only very subtle differences. While mating appeared to be positively associated with the proportions of 3-MeC25, 11-/13-MeC27, 5-MeC27, 3-MeC27, and 12-/14-MeC28 and negatively with C27:1, fecundity was negatively associated with C29:1, C31:1, and a sterol derivative. We discuss these results in the light of the special life history of C. obscurior, with completely sterile workers and low egg laying rates in queens. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The ecological success of social insects (ants, bees, wasps, termites) is founded on their well-organized division of labor, with queens (and, in termites, kings) specializing in reproduction and workers in brood care, foraging, and nest defense. To function smoothly, division of labor requires that individuals can identify each others’ task in the society. Cuticular hydrocarbons are thought to convey information on an individual’s species, nest of origin, sex, caste, task, and age (Singer, 1998; Howard and Blomquist, 2005). The exact recognition of those individuals that reproduce is particularly important to maintain the reproductive division of labor. Numerous studies have shown that the reproductive status of queens and workers correlates with the qualitative and quantitative composition of the cuticular hydrocarbon profile (e.g., Monnin, 2006; Hefetz, 2007; Heinze and d’Ettorre, 2009; Liebig, 2010; Kocher and Grozinger, 2011). It appears that workers react to this fecundity signal by refraining from reproduction. For example, 3-MeC31 is a dominant compound of the cuticular profile of queens of the ant Lasius niger, and exposing queenless worker groups to synthetic 3-MeC31 significantly decreased worker ovarian activity (Holman et al., 2010). Not only egg laying, but also mating is associated with changes in the cuticular profile of social insect females. The sexual life of ant queens is limited to a brief episode soon after adult emergence, ⇑ Corresponding author. E-mail address: [email protected] (J. Oettler). 0022-1910/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jinsphys.2012.06.009
during which they mate with one or several males. Thereafter, they store all received sperm to later fertilize eggs they lay throughout their lives without ever re-mating again (Boomsma et al., 2005). Mating is associated with drastic changes in physiology and behavior: virgin queens of many species engage in nuptial flights or place themselves in exposed locations and actively attract males with sexual pheromones. Shortly after mating, however, they become unattractive to males, shed their wings, and try to hide. Their activity cycles and phototactic behavior change substantially (Sharma et al., 2004; Lone et al., 2012). This is accompanied by immediate alterations of their cuticular profiles, which might make them unattractive to males (Oppelt and Heinze, 2009) or modify the physico-chemical properties of cuticular waxes to adjust to a life in a more humid place in the ground (Johnson and Gibbs, 2004; Hora et al., 2008). Furthermore, these might reflect accelerated development of the ovaries and the maturation of eggs. Hence, both mating and reproduction strongly affect the composition of cuticular hydrocarbons. The effects, mating and egg laying have on the physiology of queens have previously been disentangled in the ant Cardiocondyla obscurior by comparing fecundity and longevity among virgin queens, mated queens, and queens mated to sterilized males (Schrempf et al., 2005). Virgin queens are capable of producing male offspring from unfertilized eggs but are significantly less fecund and also have a much shorter life span than mated queens when prevented from mating. Queens mated with sterilized males have a similarly low fecundity as virgin queens but live as long as mated queens, suggesting that mating itself rather than egg laying
1246
S. Will et al. / Journal of Insect Physiology 58 (2012) 1245–1249
affects queen life span. This might be a direct or indirect effect of the copulation or of beneficial substances transferred by the male in its seminal fluid (Schrempf et al., 2005). Here, we compare the cuticular lipid profiles among mated queens, virgin queens, and queens that mated with sterilized males. As honest queen signals should reliably indicate both the queen’s mating status and its fecundity (Keller and Nonacs, 1993; Heinze and d’Ettorre, 2009), we expected the cuticular profile to differ among the three different classes of queens. Furthermore, we examined whether specific cuticular compounds are consistently associated with mating status or egg laying rate of queens. We expected to find stronger correlates of cuticular compounds with mating than with fecundity because a low fecundity of C. obscurior queens is without consequences on survival (Schrempf et al., 2005). 2. Materials and methods 2.1. Animal collection and rearing Colonies for this study were collected from their nests between the bracts of aborted coconuts in an experimental coconut plantation in Una, Bahia, Brazil in July 2009. After transfer into the laboratory, individual colonies were housed in petridishes with a plaster floor and a glass-covered cavity serving as nest site in incubators at near-natural conditions (12 h 28 °C/12 h 24 °C). Ants were fed three times per week with a mixture of raw eggs, chopped cockroaches, honey, and agarose, and the plaster floor was regularly watered to keep humidity. For the experiments, we set up three series of 40 experimental colonies each, consisting each of 30 workers, some brood and a single virgin queen. To the colonies of one series (‘‘mated queens’’) we added a single wingless male, to the second series a single wingless male that had been sterilized by exposure to X-rays (120 Gy; Schrempf et al., 2005; ‘‘sham-mated queens’’). The third series did not receive any male (‘‘virgin queens’’). Queens of C. obscurior readily mate in the lab and quickly begin to lay eggs, whereas workers are completely sterile and do not lay eggs. We counted the eggs per colony once per week over 8 weeks. As eggs hatch within 7 days (Schrempf and Heinze, 2006), weekly egg counts reflect the fecundity of queens. Weekly egg counts per queen were summed and compared among mated queens, sham-mated queens and virgin queens by Kruskal–Wallis H-test and subsequent Mann– Whitney U-tests with Bonferroni correction. After 8 weeks, all queens were killed by freezing and stored at 20 °C for further analysis. Because some queens died during the experiment or did not lay eggs, egg counts were available only for 24 mated queens, 27 sham-mated queens, and 16 virgin queens. 2.2. Chemical analyses Cuticular hydrocarbons and some other lipids were analyzed by thermal desorption gas chromatography–mass spectrometry (TD– GC–MS) of whole ants. Eleven or 12 individual queens per series were introduced into conditioned thermal desorption tubes (89 mm length 5 mm inner diameter, Supelco, Bellefonte PA) and thermally desorbed for 8 min at 250 °C with a helium flow of 60 ml/min using an automated Shimadzu TD 20 thermal desorber (Shimadzu, Kyoto, Japan). Desorbed volatiles were cryofocused at 20 °C on an internal Tenax trap. Subsequently, volatiles were injected into the GC–MS by heating the internal trap to 280 °C for 5 min at a split ratio of 1:5. In addition, hydrocarbons were extracted from pooled 30–50 eggs by soaking in n-hexane for 15 min and analyzed by liquid injection. GC–MS analyses were conducted
using a Shimadzu QP 2010 GC–MS equipped with a non-polar 30 m 0.32 mm 0.25 lm BPX-5 capillary column (SGE, Melbourne). Helium was used as a carrier gas at a flow rate of 50 cm/s. The GC was programmed from 150 °C to a final temperature of 300 °C at 3 °C/min. Compounds were identified by comparing their linear retention indices with literature data (Carlson et al., 1998) and by interpretation of diagnostic ions obtained by mass spectrometry (Nelson, 1993). Peak areas were calculated using the GC–MS solution version 2.53 scientific software of the mass spectrometer. 2.3. Statistical analysis For statistical analysis we chose peaks with a relative area of at least 1% that were present in at least 80% of all samples. After filtering, peak areas of 0 were replaced by the minute amount 1 for subsequent log transformation. We reduced the number of variables for discriminant analysis (DA) by principle component analysis (PCA). For the PCA, the constant-sum constraint of the data was removed by a centered log ratio transformation following Reyment (1989): Zij = log [Xi, j/g(Xj)], where Xi, j is standardized peak area, i for the sample, j and g(Xj) is the geometric mean of all standardized peak areas of sample j. Williams and Titus (1988) suggested that the minimal number of samples in each group should be three times the number of discriminating variables in a DA. We consequently used only the first three principal components in DA to determine whether the different types of queens could be distinguished through their cuticular profiles. The similarity among groups was compared by squared Mahalanobis distances. In addition, we performed a DA with a full leaveone-out cross-validation method on all PCA factors with eigenvalues larger than 1. To determine whether particular substances are associated with a certain type of queens we compared the percentages by Kruskal– Wallis H-tests. However, this did not yield any significant differences after sequential Bonferroni correction following Holm (1979). We therefore analyzed the percentages of individual compounds in a generalized linear model with fecundity (egg count) and mating status (two levels, mated and unmated) as independent variables. In addition, we correlated the total number of eggs produced during the 8 weeks and the number of eggs present in the 8th week of the experiment with the percentages of individual compounds by a series of Spearman’s rank tests with sequential Bonferroni correction (Holm, 1979). Analyses were done with the statistical programs SPSS and R. 3. Results During the 8 weeks of our study, mated queens laid up to five times more eggs than sham-mated and virgin queens (summed weekly egg counts: mated queens, n = 24: median 11, quartiles 4, 26; sham-mated, n = 27, median 4, quartiles 2, 7; virgin queens, n = 16, median 2, quartiles 0, 9; Kruskal–Wallis H-test, H = 11.11, p = 0.004; Mann–Whitney U-test with Bonferroni correction, mated vs. sham-mated: U = 177.5, pcorr = 0.018; mated vs. virgin: U = 88.5, pcorr = 0.016; sham-mated vs. virgin: U = 183,0, pcorr = 1). Of the maximally 50 peaks present in the profile of individual queens, 23 peaks occurred consistently in more than 80% of the specimens and made up more than 1% of the total peak area of an individual (Table 1). These peaks represented 31 compounds, since some compounds overlapped in the chromatograms. Among these compounds were four n-alkanes, 19 monomethylalkanes, one dimethylalkane, five alkamonoenes, two alkadienes, and two unidentified sterol derivatives. Principal component analysis resulted in six eigenvalues larger than 1, which together accounted
1247
S. Will et al. / Journal of Insect Physiology 58 (2012) 1245–1249
Table 1 Median proportions (%) with upper and lower quartiles of substances from the cuticula of sham-mated queens, mated queens, and virgin queens of the ant C. obscurior. Substances in bold were shown to be positively associated with mating in a generalized linear model. Peak
LRIa
Compound
Diagnostic ions (m/z)
Sham-mated queens Median
a
1 2
2500 2531
3 4 5 6 7 8
2571 2662 2675 2683 2700 2730
9 10 11 12
2748 2773 2800 2830
13
2863
14 15
2876 2885
16 17
2900 2929
18 19 20 21
2938 3078 3117 3128
22
3326
23
3547
C25 13-MeC25 +11-MeC25 3-MeC25 C27:2; + 2-MeC26 C27:1 C27:1 C27 13-MeC27 +11-MeC27 5-MeC27 3-MeC27 C28 14-MeC28 +12-MeC28 C29:2 +2-MeC28 C29:1 C29:1 +unknown sterol derivative C29 15-MeC29 +13-MeC29 +11-MeC29 7-MeC29 C31:1 Unknown sterol derivative 13-MeC31 +11-MeC31 13-MeC33 +11-MeC33 13,23-DiMeC35
352 196/197 (sym) 168/169, 224/225 337 376; 337 378 378 380 196/197, 224/225 168/169, 252/253 84/85, 336/337 365 394 210/211, 224/225 182/183, 252/253 404 365 406 406 386, 368 408 224/225 (sym) 196/197, 252/253 168/169, 280/281 112/113, 336/337 434 386, 368 196/197, 280/281 168/169, 308/309 196/197, 308/309 168/169, 336/337 196/197, 350/351 (sym)
25%
Mated queens 75%
median
25%
Virgin queens 75%
Median
25%
75%
7.02 2.80
5.78 2.33
7.71 3.03
7.23 2.42
4.82 2.05
9.64 3.01
7.41 2.30
5.54 2.08
9.09 2.70
2.24 4.07 3.50 1.35 10.89 9.22
2.09 3.34 2.64 1.15 10.61 7.46
2.40 5.22 3.96 1.75 14.05 10.86
2.45 2.74 2.60 0.88 11.80 8.88
2.10 1.28 2.00 0.78 9.67 6.55
2.94 4.33 3.37 1.16 12.49 11.76
1.73 6.04 3.28 1.41 12.25 6.37
1.52 5.28 2.66 1.32 8.69 5.15
2.20 7.46 3.64 1.94 13.64 8.92
1.41 7.28 1.56 1.42
1.35 5.79 1.23 1.30
1.51 7.92 1.99 1.57
1.37 5.55 1.12 1.19
1.27 4.53 1.04 0.90
1.42 6.78 1.30 1.80
1.19 4.27 0.87 1.03
1.03 3.81 0.14 0.76
1.46 6.26 1.47 1.33
5.57
4.12
6.28
3.59
2.66
4.44
5.91
5.47
8.59
5.67 2.78
5.19 2.55
7.40 3.30
4.42 2.02
3.83 1.73
6.68 3.17
4.98 2.68
3.89 2.12
7.01 3.04
3.83 5.83
3.03 4.39
4.07 6.71
3.56 6.74
3.25 4.78
4.76 9.01
3.54 5.85
3.06 4.56
5.00 6.96
1.58 2.45 6.02 4.38
1.39 2.23 4.46 2.71
1.73 3.47 6.70 6.96
1.58 2.05 9.00 6.50
1.16 1.71 6.20 4.56
2.04 3.61 10.05 8.00
2.04 2.62 6.51 4.84
1.41 2.12 4.54 4.43
2.24 3.23 7.79 6.00
3.39
2.26
4.10
3.91
2.80
5.01
2.29
2.00
3.30
1.84
0.69
2.48
2.61
1.70
3.21
1.78
1.39
3.13
Linear retention index.
for 81.1% of the total variance. Discriminant analysis with the largest three principal components showed an overall significant differentiation of the three types of queens (Wilks’ k = 0.599, F6,58 = 2.816, p < 0.018). However, only 61.8% of the samples were correctly classified. Mismatches occurred particularly between the groups of shammated and mated queens and between sham-mated and virgin queens, suggesting that the chemical profiles of sham-mated queens are intermediate between those of virgin queens and mated queens. A comparison of squared Mahalanobis distances D2 did not indicate a clear separation of mated and sham-mated queens (D2 = 1.645, F3,29 = 2.687, p = 0.065) and sham-mated and virgin queens (D2 = 1.577, F3,29 = 2.458, p = 0.083). Mated and virgin queens were most clearly distinct (D2 = 2.031, F3,29 = 3.318, p = 0.034), but the difference becomes insignificant after Bonferroni correction for multiple testing. The large overlap among the three groups of queens is also indicated by a scatter plot with the two canonical functions (Fig. 1). Discriminant analysis on all six principal components with leave-one-out cross-validation qualitatively corroborates these results, with only 50.0% of all individuals correctly classified. Given that fecundity varied considerably within groups we tried to determine whether particular compounds were associated with mating and/or fecundity in a generalized linear model. Mating (mated and sham-mated queens) appeared positively correlated with percentages of five substances (3-MeC25, F = 6.177, p < 0.019; 11-/13-MeC27, F = 4.564, p < 0.041; 5-MeC27, F = 6.187, p < 0.019); 3-MeC27, F = 7.371, p < 0.011; 12-/14-MeC28, F = 7.109, p < 0.012) and negatively with peak 6, an isomer of C27:1 (F = 4.607, p < 0.040). Fecundity was negatively correlated with the proportions of peak 15, C29:1 and an unidentified sterol deriva-
Fig. 1. Plot of the two functions of the discriminant analysis of cuticular hydrocarbon profiles of sham-mated queens, mated queens, and virgin queens of the ant C. obscurior. Discriminant analysis does not reliably distinguish among the three different types of queens.
tive (F = 4.901, p < 0.035) and peak 19, C31:1 (F = 9.272, p < 0.005), but there was not any significant positive association between fecundity and the proportion of particular compounds. The cuticular profiles of eggs and queens overlapped only to some extent. We found ten substances on the eggs that were not present on the surface of queens. Eleven substances from the cuticular profile of queens could not be detected on the eggs. The major hydrocarbons present on the surface of eggs were the odd-numbered alkanes n-C25 (appr. 15%), n-C27 (appr. 32%), and n-C29 (appr.
1248
S. Will et al. / Journal of Insect Physiology 58 (2012) 1245–1249
13%). In addition, minor quantities of the respective methylbranched compounds were present, including some of the substances positively associated with mating status (3-MeC25, 3-MeC27, 11-/13-MeC27, Table 2). 4. Discussion Virgin queens of the ant C. obscurior differ from mated queens in fecundity and longevity and from queens mated to sterilized males in longevity (Schrempf et al., 2005). Despite of the presumably large physiological differences among the three types of queens, their cuticular hydrocarbon profiles are remarkably similar and can hardly be separated by discriminant analysis. The profiles of queens are dominated by n-alkanes, branched alkanes and alkenes with chain lengths between C25 and C33. Several of the most abundant compounds in our eight week old queens, i.e., C25, C27, 3-MeC27, and 11-/13-MeC27, have previously been found to dominate also the cuticular waxes of freshly eclosed virgin queens (Cremer et al., 2008). One of the nonacosenes (C29:1), which in our study appeared to be negatively correlated with fecundity, was the most abundant compound obtained from 1 day old virgin queens by solid sampling and also made up a considerable percentage of solvent-extracted hydrocarbons (Turillazzi et al., 2002; Cremer et al., 2008). Egg laying rate was also negatively associated with the proportion of C31:1, but our analysis did not give conclusive evidence for any positive association of specific substances with fecundity, i.e., the total number of eggs laid by the queens during the 8 weeks of the experiment. Instead, our study revealed five substances that were present in larger relative quantities on the cuticula of mated queens, regardless of whether these had mated with fertile or sterilized males, than on the cuticula of virgin queens. Among these were a few branched alkanes, which were also present on the surface of eggs.
Branched alkanes have previously been found to be associated with fecundity in other ants, e.g., Diacamma ceylonense (Cuvillier-Hot et al., 2001) and Camponotus floridanus (Endler et al., 2004), while particularly the proportion of 3-MeC27 on the cuticula of Leptothorax gredleri queens decreased rapidly after mating (Oppelt and Heinze, 2009). The absence of a positive correlation between the quantities of individual cuticular compounds and the total number of eggs laid by queens in our experiment is surprising. It contrasts with the growing number of studies in social insects according to which fecundity is associated with qualitative and quantitative changes in the cuticular profile (e.g., Monnin, 2006; Hefetz, 2007; Heinze and d’Ettorre, 2009; Liebig, 2010; Kocher and Grozinger, 2011). In a few cases, specific substances have been shown directly to elicit a response in the antennae of workers (d’Ettorre et al., 2004) or to serve as fecundity signals and regulate worker sterility (e.g., Holman et al., 2010). Such honest queen pheromones are important in maintaining the division or reproductive labor between queens and workers (e.g., Keller and Nonacs, 1993). What explains the apparent absence of fecundity-specific substances in C. obscurior and the resemblance of the profiles of egg laying virgin queens, sham-mated queens, and mated queens? On the one hand, the similarity of the profiles might reflect the fact that Cardiocondyla workers completely lack ovaries. Fertility signals therefore are not required for the regulation of reproduction between queens and workers. Worker sterility might also explain the simple hydrocarbon profile of C. obscurior eggs, which consists mostly of the linear alkanes n-C25, n-C27, and n-C29. These substances are among the most widespread hydrocarbons in insects (e.g., Lockey, 1980) and they are present in considerable quantities also on the surface of other social insect eggs (e.g., Endler et al., 2004; Bonckaert et al., 2012). Nevertheless, the egg profile of C. obscurior appears to be much simpler than those of other
Table 2 Median proportions (%) with upper and lower quartiles of egg surface hydrocarbons of the ant C. obscurior. ‘‘Queen peak ID’’ denotes substances identical to those on the cuticle of queens (see Table 1). Queen peak ID
1 3 7 8 9 10 11 14 15 16
LRIa
Compound
Diagnostic ions (m/z)
2400 2500 2571 2600 2700 2730
C24 C25 3-MeC25 C26 C27 13-MeC27 +11-MeC27 5-MeC27 3-MeC27 C28 C29:1 C29:1 C29 3-MeC29 C31 13-MeC31 +11-MeC31 3-MeC31 3,7-DiMeC31 C33 13-MeC33 +11-MeC33 15-MeC35 +13-MeC35 +11MeC35 13-MeC37 15-MeC39 +13-MeC39
338 352 337 366 380 196/197, 224/225 168/169, 252/253 84/85, 336/337 365 394 406 406 408 393 436 196/197, 280/281 168/169, 308/309 421 435, 126/127, 364/365 464 196/197, 308/309 168/169, 336/337 224/225, 308/309 196/197, 336/337 168/169, 364/365 196/197, 364/365 224/225, 364/365 196/197, 392/393
21
2748 2773 2800 2876 2885 2900 2974 3100 3128
22
3174 3206 3300 3326 3522
n.d.b n.d. a b
Linear retention index. Not determined (no reference hydrocarbons).
Eggs Median
25%
75%
0.54 14.83 0.55 1.60 32.40 0.72
0.52 14.48 0.45 1.58 29.97 0.59
0.55 15.42 0.63 1.65 34.51 0.85
0.43 3.47 1.61 0 0 13.01 1.39 4.58 1.37
0 3.24 1.61 0 0 12.94 1.18 4.26 1.21
0.43 3.52 1.81 0.98 0.47 15.18 1.55 5.75 1.74
0.76 0.64 1.22 1.91
0.52 0 1.13 1.88
0.78 0.92 1.59 1.98
1.39
1.36
1.59
3.71 0
3.53 0
3.82 1.01
S. Will et al. / Journal of Insect Physiology 58 (2012) 1245–1249
social insects, in that the three alkanes make up more than 60% of the egg surface hydrocarbons and branched alkanes are present only in minor quantities. On the other hand, observations in other species indicate that worker sterility alone is not sufficient to explain the similarities among queen profiles and the simple composition of egg hydrocarbons. Queen pheromones may have numerous additional functions other than keeping workers from reproducing (e.g., Vargo and Hulsey, 2000). In the Argentine ant, Linepithema humile, cuticular profiles of fertile and infertile queens differ considerably despite of complete worker sterility (de Biseau et al., 2004). Such variation in cuticular profiles might allow workers to preferentially take care of the most fecund queen. In C. obscurior, workers appear to treat the differently fecund types of queens equally (Schrempf et al., 2005), which matches the similarity of their cuticular profiles. In any case, our present data suggest that that fecundity signals play a less decisive role in this species than in other social insects. Acknowledgments Collecting and export of colonies was permitted by Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis (IBAMA, permit 20324-1) to J.H.C.D. Research was financed by DFG (He1623/23). Leonardo Dapporto and an anonymous referee made helpful comments on the manuscript. References Bonckaert, W., Drijfhout, F.P., d’Ettorre, P., Billen, J., Wenseleers, T., 2012. Hydrocarbon signatures of egg maternity, caste membership and reproductive status in the Common Wasp. Journal of Chemical Ecology 38, 42–51. Boomsma, J.J., Baer, B.C., Heinze, J., 2005. The evolution of male traits in social insects. Annual Review of Entomology 50, 395–420. Carlson, D.A., Bernier, U.R., Sutton, B.D., 1998. Elution parameters from capillary GC for methyl-branched alkanes. Journal of Chemical Ecology 24, 1845–1865. Cremer, S., d’Ettorre, P., Drijfhout, F.P., Sledge, M.S., Turillazzi, S., Heinze, J., 2008. Imperfect chemical female mimicry in males of the ant Cardiocondyla obscurior. Naturwissenschaften 95, 1101–1105. Cuvillier-Hot, V., Cobb, M., Malosse, C., Peeters, C., 2001. Sex, age and ovarian activity affect cuticular hydrocarbons in Diacamma ceylonense, a queenless ant. Journal of Insect Physiology 47, 485–493. De Biseau, J.-C., Passera, L., Daloze, D., Aron, S., 2004. Ovarian activity correlates with extreme changes in cuticular hydrocarbon profile in the highly polygynous ant, Linepithema humile. Journal of Insect Physiology 50, 585–593. D’Ettorre, P., Heinze, J., Schulz, C., Francke, W., Ayasse, M., 2004. Does she smell like a queen? Chemoreception of a cuticular hydrocarbon signal in the ant Pachycondyla inversa. Journal of Experimental Biology 207, 1085–1091. Endler, A., Liebig, J., Schmitt, T., Parker, J.E., Jones, G.R., Schreier, P., Hölldobler, B., 2004. Surface hydrocarbons of queen eggs regulate worker reproduction in a
1249
social insect. Proceedings of the National Academy of Sciences of the United States of America 101, 2945–2950. Hefetz, A., 2007. The evolution of hydrocarbon pheromone parsimony in ants (Hymenoptera: Formicidae) – interplay of colony odor uniformity and odor idiosyncrasy. A review. Myrmecological News 10, 59–68. Heinze, J., d’Ettorre, P., 2009. Honest and dishonest communication in social Hymenoptera. Journal of Experimental Biology 212, 1775–1779. Holman, L., Jørgensen, C.G., Nielsen, J., d’Ettorre, P., 2010. Identification of an ant queen pheromone regulating worker sterility. Proceedings of the Royal Society of London Series B: Biology 277, 3793–3800. Hora, R.R., Ionescu-Hirsh, A., Simon, T., Delabie, J.H.C., Robert, J., Fresneau, D., Hefetz, A., 2008. Postmating changes in cuticular chemistry and visual appearance in Ectatomma tuberculatum queens (Formicidae: Ectatomminae). Naturwissenschaften 95, 55–60. Holm, S., 1979. A simple sequential rejective multiple test procedure. Scandinavian Journal of Statistics 6, 65–70. Howard, R.W., Blomquist, G.J., 2005. Ecological, behavioral, and biochemical aspects of insect hydrocarbons. Annual Review of Entomology 50, 371–392. Johnson, R.A., Gibbs, A.G., 2004. Effect of mating stage on water balance, cuticular hydrocarbons and metabolism in the desert harvester ant, Pogonomyrmex barbatus. Journal of Insect Physiology 50, 943–953. Keller, L., Nonacs, P., 1993. The role of queen pheromones in social insects: queen control or queen signal? Animal Behaviour 45, 787–794. Kocher, S.D., Grozinger, C.M., 2011. Cooperation, conflict, and the evolution of queen pheromones. Journal of Chemical Ecology 37, 1263–1275. Liebig, J., 2010. Hydrocarbon profiles indicate fertility and dominance status in ant, bee and wasp colonies. In: Blomquist, G.J., Bagnères, A.-G. (Eds.), Insect Hydrocarbons. Biology, Biochemistry and Chemical Ecology. Cambridge University Press, Cambridge, pp. 254–281. Lockey, K.H., 1980. Insect cuticular hydrocarbons. Comparative Biochemistry and Physiology 65B, 457–462. Lone, S.R., Chakravarthi, A., Sharma, V.K., 2012. Sex and age related changes in the locomotor activity and phototactic behaviors of two closely related species of Camponotus ants. Journal of Insect Physiology 78, 75–82. Monnin, A., 2006. Chemical recognition of reproductive status in social insects. Annales Zoologici Fennici 43, 515–530. Nelson, D.R., 1993. Methyl-branched lipids in insects. In: Stanley-Samuelson, D.W., Nelson, D.R. (Eds.), Insect Lipids: Chemistry, Biochemistry, and Biology. University of Nebraska Press, Lincoln, pp. 271–316. Oppelt, A., Heinze, J., 2009. Mating is associated with immediate changes of the hydrocarbon profile of Leptothorax gredleri ant queens. Journal of Insect Physiology 55, 624–628. Reyment, R.A., 1989. Compositional data analysis. Terra Review 1, 29–34. Schrempf, A., Heinze, J., 2006. Proximate mechanisms of male morph determination in the ant Cardiocondyla obscurior. Evolution and Development 8, 266–272. Schrempf, A., Heinze, J., Cremer, S., 2005. Sexual cooperation: mating increases longevity in ant queens. Current Biology 15, 267–270. Sharma, V.K., Lone, S.R., Goel, A., 2004. Clocks for sex: loss of circadian rhythms in ants after mating? Naturwissenschaften 91, 334–337. Singer, T.A., 1998. Roles of hydrocarbons in the recognition systems of insects. American Zoologist 38, 394–405. Turillazzi, S., Sledge, M.F., Cremer, S., Heinze, J., 2002. A method for analysing smallsize specimens in GC–MS. Insect Social Life 4, 169–175. Vargo, E.L., Hulsey, C.D., 2000. Multiple glandular origins of queen pheromones in the fire ant Solenopsis invicta. Journal of Insect Physiology 46, 1151–1159. Williams, B.K., Titus, K., 1988. Assessment of sampling stability in ecological applications of discriminant analysis. Ecology 69, 1275–1285.
Contents lists available at SciVerse ScienceDirect
Journal of Insect Physiology journal homepage: www.elsevier.com/locate/jinsphys
Cuticular lipid profiles of fertile and non-fertile Cardiocondyla ant queens Stefanie Will a, Jacques H.C. Delabie b, Jürgen Heinze a, Joachim Ruther c, Jan Oettler a,⇑ a
Universität Regensburg, Biologie I, D-93040 Regensburg, Germany Laboratório de Mirmecologia, CEPEC/CEPLAC, Cx.P. 7, Itabuna, BA, Brazil c Universität Regensburg, Chemical Ecology Group, D-93040 Regensburg, Germany b
a r t i c l e
i n f o
Article history: Received 22 March 2012 Received in revised form 18 June 2012 Accepted 19 June 2012 Available online 28 June 2012 Keywords: Cuticular hydrocarbons Queen pheromone Fertility signal Mating Formicidae
a b s t r a c t Both mating and reproduction strongly affect the physiology of insect females. In the ant Cardiocondyla obscurior, a comparison among virgin queens, mated queens, and queens mated with sterilized males (‘‘sham-mated’’) allows to separate the different effects of mating and egg laying. Here, we investigate whether and how different mating status is reflected in the cuticular lipid profiles of queens, i.e., the blend of chemicals that is thought to signal a queen’s fertility. Surprisingly, discriminant analyses failed to reliably distinguish among virgin, mated, and sham-mated queens. A generalized linear model on individual substances showed only very subtle differences. While mating appeared to be positively associated with the proportions of 3-MeC25, 11-/13-MeC27, 5-MeC27, 3-MeC27, and 12-/14-MeC28 and negatively with C27:1, fecundity was negatively associated with C29:1, C31:1, and a sterol derivative. We discuss these results in the light of the special life history of C. obscurior, with completely sterile workers and low egg laying rates in queens. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction The ecological success of social insects (ants, bees, wasps, termites) is founded on their well-organized division of labor, with queens (and, in termites, kings) specializing in reproduction and workers in brood care, foraging, and nest defense. To function smoothly, division of labor requires that individuals can identify each others’ task in the society. Cuticular hydrocarbons are thought to convey information on an individual’s species, nest of origin, sex, caste, task, and age (Singer, 1998; Howard and Blomquist, 2005). The exact recognition of those individuals that reproduce is particularly important to maintain the reproductive division of labor. Numerous studies have shown that the reproductive status of queens and workers correlates with the qualitative and quantitative composition of the cuticular hydrocarbon profile (e.g., Monnin, 2006; Hefetz, 2007; Heinze and d’Ettorre, 2009; Liebig, 2010; Kocher and Grozinger, 2011). It appears that workers react to this fecundity signal by refraining from reproduction. For example, 3-MeC31 is a dominant compound of the cuticular profile of queens of the ant Lasius niger, and exposing queenless worker groups to synthetic 3-MeC31 significantly decreased worker ovarian activity (Holman et al., 2010). Not only egg laying, but also mating is associated with changes in the cuticular profile of social insect females. The sexual life of ant queens is limited to a brief episode soon after adult emergence, ⇑ Corresponding author. E-mail address: [email protected] (J. Oettler). 0022-1910/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jinsphys.2012.06.009
during which they mate with one or several males. Thereafter, they store all received sperm to later fertilize eggs they lay throughout their lives without ever re-mating again (Boomsma et al., 2005). Mating is associated with drastic changes in physiology and behavior: virgin queens of many species engage in nuptial flights or place themselves in exposed locations and actively attract males with sexual pheromones. Shortly after mating, however, they become unattractive to males, shed their wings, and try to hide. Their activity cycles and phototactic behavior change substantially (Sharma et al., 2004; Lone et al., 2012). This is accompanied by immediate alterations of their cuticular profiles, which might make them unattractive to males (Oppelt and Heinze, 2009) or modify the physico-chemical properties of cuticular waxes to adjust to a life in a more humid place in the ground (Johnson and Gibbs, 2004; Hora et al., 2008). Furthermore, these might reflect accelerated development of the ovaries and the maturation of eggs. Hence, both mating and reproduction strongly affect the composition of cuticular hydrocarbons. The effects, mating and egg laying have on the physiology of queens have previously been disentangled in the ant Cardiocondyla obscurior by comparing fecundity and longevity among virgin queens, mated queens, and queens mated to sterilized males (Schrempf et al., 2005). Virgin queens are capable of producing male offspring from unfertilized eggs but are significantly less fecund and also have a much shorter life span than mated queens when prevented from mating. Queens mated with sterilized males have a similarly low fecundity as virgin queens but live as long as mated queens, suggesting that mating itself rather than egg laying
1246
S. Will et al. / Journal of Insect Physiology 58 (2012) 1245–1249
affects queen life span. This might be a direct or indirect effect of the copulation or of beneficial substances transferred by the male in its seminal fluid (Schrempf et al., 2005). Here, we compare the cuticular lipid profiles among mated queens, virgin queens, and queens that mated with sterilized males. As honest queen signals should reliably indicate both the queen’s mating status and its fecundity (Keller and Nonacs, 1993; Heinze and d’Ettorre, 2009), we expected the cuticular profile to differ among the three different classes of queens. Furthermore, we examined whether specific cuticular compounds are consistently associated with mating status or egg laying rate of queens. We expected to find stronger correlates of cuticular compounds with mating than with fecundity because a low fecundity of C. obscurior queens is without consequences on survival (Schrempf et al., 2005). 2. Materials and methods 2.1. Animal collection and rearing Colonies for this study were collected from their nests between the bracts of aborted coconuts in an experimental coconut plantation in Una, Bahia, Brazil in July 2009. After transfer into the laboratory, individual colonies were housed in petridishes with a plaster floor and a glass-covered cavity serving as nest site in incubators at near-natural conditions (12 h 28 °C/12 h 24 °C). Ants were fed three times per week with a mixture of raw eggs, chopped cockroaches, honey, and agarose, and the plaster floor was regularly watered to keep humidity. For the experiments, we set up three series of 40 experimental colonies each, consisting each of 30 workers, some brood and a single virgin queen. To the colonies of one series (‘‘mated queens’’) we added a single wingless male, to the second series a single wingless male that had been sterilized by exposure to X-rays (120 Gy; Schrempf et al., 2005; ‘‘sham-mated queens’’). The third series did not receive any male (‘‘virgin queens’’). Queens of C. obscurior readily mate in the lab and quickly begin to lay eggs, whereas workers are completely sterile and do not lay eggs. We counted the eggs per colony once per week over 8 weeks. As eggs hatch within 7 days (Schrempf and Heinze, 2006), weekly egg counts reflect the fecundity of queens. Weekly egg counts per queen were summed and compared among mated queens, sham-mated queens and virgin queens by Kruskal–Wallis H-test and subsequent Mann– Whitney U-tests with Bonferroni correction. After 8 weeks, all queens were killed by freezing and stored at 20 °C for further analysis. Because some queens died during the experiment or did not lay eggs, egg counts were available only for 24 mated queens, 27 sham-mated queens, and 16 virgin queens. 2.2. Chemical analyses Cuticular hydrocarbons and some other lipids were analyzed by thermal desorption gas chromatography–mass spectrometry (TD– GC–MS) of whole ants. Eleven or 12 individual queens per series were introduced into conditioned thermal desorption tubes (89 mm length 5 mm inner diameter, Supelco, Bellefonte PA) and thermally desorbed for 8 min at 250 °C with a helium flow of 60 ml/min using an automated Shimadzu TD 20 thermal desorber (Shimadzu, Kyoto, Japan). Desorbed volatiles were cryofocused at 20 °C on an internal Tenax trap. Subsequently, volatiles were injected into the GC–MS by heating the internal trap to 280 °C for 5 min at a split ratio of 1:5. In addition, hydrocarbons were extracted from pooled 30–50 eggs by soaking in n-hexane for 15 min and analyzed by liquid injection. GC–MS analyses were conducted
using a Shimadzu QP 2010 GC–MS equipped with a non-polar 30 m 0.32 mm 0.25 lm BPX-5 capillary column (SGE, Melbourne). Helium was used as a carrier gas at a flow rate of 50 cm/s. The GC was programmed from 150 °C to a final temperature of 300 °C at 3 °C/min. Compounds were identified by comparing their linear retention indices with literature data (Carlson et al., 1998) and by interpretation of diagnostic ions obtained by mass spectrometry (Nelson, 1993). Peak areas were calculated using the GC–MS solution version 2.53 scientific software of the mass spectrometer. 2.3. Statistical analysis For statistical analysis we chose peaks with a relative area of at least 1% that were present in at least 80% of all samples. After filtering, peak areas of 0 were replaced by the minute amount 1 for subsequent log transformation. We reduced the number of variables for discriminant analysis (DA) by principle component analysis (PCA). For the PCA, the constant-sum constraint of the data was removed by a centered log ratio transformation following Reyment (1989): Zij = log [Xi, j/g(Xj)], where Xi, j is standardized peak area, i for the sample, j and g(Xj) is the geometric mean of all standardized peak areas of sample j. Williams and Titus (1988) suggested that the minimal number of samples in each group should be three times the number of discriminating variables in a DA. We consequently used only the first three principal components in DA to determine whether the different types of queens could be distinguished through their cuticular profiles. The similarity among groups was compared by squared Mahalanobis distances. In addition, we performed a DA with a full leaveone-out cross-validation method on all PCA factors with eigenvalues larger than 1. To determine whether particular substances are associated with a certain type of queens we compared the percentages by Kruskal– Wallis H-tests. However, this did not yield any significant differences after sequential Bonferroni correction following Holm (1979). We therefore analyzed the percentages of individual compounds in a generalized linear model with fecundity (egg count) and mating status (two levels, mated and unmated) as independent variables. In addition, we correlated the total number of eggs produced during the 8 weeks and the number of eggs present in the 8th week of the experiment with the percentages of individual compounds by a series of Spearman’s rank tests with sequential Bonferroni correction (Holm, 1979). Analyses were done with the statistical programs SPSS and R. 3. Results During the 8 weeks of our study, mated queens laid up to five times more eggs than sham-mated and virgin queens (summed weekly egg counts: mated queens, n = 24: median 11, quartiles 4, 26; sham-mated, n = 27, median 4, quartiles 2, 7; virgin queens, n = 16, median 2, quartiles 0, 9; Kruskal–Wallis H-test, H = 11.11, p = 0.004; Mann–Whitney U-test with Bonferroni correction, mated vs. sham-mated: U = 177.5, pcorr = 0.018; mated vs. virgin: U = 88.5, pcorr = 0.016; sham-mated vs. virgin: U = 183,0, pcorr = 1). Of the maximally 50 peaks present in the profile of individual queens, 23 peaks occurred consistently in more than 80% of the specimens and made up more than 1% of the total peak area of an individual (Table 1). These peaks represented 31 compounds, since some compounds overlapped in the chromatograms. Among these compounds were four n-alkanes, 19 monomethylalkanes, one dimethylalkane, five alkamonoenes, two alkadienes, and two unidentified sterol derivatives. Principal component analysis resulted in six eigenvalues larger than 1, which together accounted
1247
S. Will et al. / Journal of Insect Physiology 58 (2012) 1245–1249
Table 1 Median proportions (%) with upper and lower quartiles of substances from the cuticula of sham-mated queens, mated queens, and virgin queens of the ant C. obscurior. Substances in bold were shown to be positively associated with mating in a generalized linear model. Peak
LRIa
Compound
Diagnostic ions (m/z)
Sham-mated queens Median
a
1 2
2500 2531
3 4 5 6 7 8
2571 2662 2675 2683 2700 2730
9 10 11 12
2748 2773 2800 2830
13
2863
14 15
2876 2885
16 17
2900 2929
18 19 20 21
2938 3078 3117 3128
22
3326
23
3547
C25 13-MeC25 +11-MeC25 3-MeC25 C27:2; + 2-MeC26 C27:1 C27:1 C27 13-MeC27 +11-MeC27 5-MeC27 3-MeC27 C28 14-MeC28 +12-MeC28 C29:2 +2-MeC28 C29:1 C29:1 +unknown sterol derivative C29 15-MeC29 +13-MeC29 +11-MeC29 7-MeC29 C31:1 Unknown sterol derivative 13-MeC31 +11-MeC31 13-MeC33 +11-MeC33 13,23-DiMeC35
352 196/197 (sym) 168/169, 224/225 337 376; 337 378 378 380 196/197, 224/225 168/169, 252/253 84/85, 336/337 365 394 210/211, 224/225 182/183, 252/253 404 365 406 406 386, 368 408 224/225 (sym) 196/197, 252/253 168/169, 280/281 112/113, 336/337 434 386, 368 196/197, 280/281 168/169, 308/309 196/197, 308/309 168/169, 336/337 196/197, 350/351 (sym)
25%
Mated queens 75%
median
25%
Virgin queens 75%
Median
25%
75%
7.02 2.80
5.78 2.33
7.71 3.03
7.23 2.42
4.82 2.05
9.64 3.01
7.41 2.30
5.54 2.08
9.09 2.70
2.24 4.07 3.50 1.35 10.89 9.22
2.09 3.34 2.64 1.15 10.61 7.46
2.40 5.22 3.96 1.75 14.05 10.86
2.45 2.74 2.60 0.88 11.80 8.88
2.10 1.28 2.00 0.78 9.67 6.55
2.94 4.33 3.37 1.16 12.49 11.76
1.73 6.04 3.28 1.41 12.25 6.37
1.52 5.28 2.66 1.32 8.69 5.15
2.20 7.46 3.64 1.94 13.64 8.92
1.41 7.28 1.56 1.42
1.35 5.79 1.23 1.30
1.51 7.92 1.99 1.57
1.37 5.55 1.12 1.19
1.27 4.53 1.04 0.90
1.42 6.78 1.30 1.80
1.19 4.27 0.87 1.03
1.03 3.81 0.14 0.76
1.46 6.26 1.47 1.33
5.57
4.12
6.28
3.59
2.66
4.44
5.91
5.47
8.59
5.67 2.78
5.19 2.55
7.40 3.30
4.42 2.02
3.83 1.73
6.68 3.17
4.98 2.68
3.89 2.12
7.01 3.04
3.83 5.83
3.03 4.39
4.07 6.71
3.56 6.74
3.25 4.78
4.76 9.01
3.54 5.85
3.06 4.56
5.00 6.96
1.58 2.45 6.02 4.38
1.39 2.23 4.46 2.71
1.73 3.47 6.70 6.96
1.58 2.05 9.00 6.50
1.16 1.71 6.20 4.56
2.04 3.61 10.05 8.00
2.04 2.62 6.51 4.84
1.41 2.12 4.54 4.43
2.24 3.23 7.79 6.00
3.39
2.26
4.10
3.91
2.80
5.01
2.29
2.00
3.30
1.84
0.69
2.48
2.61
1.70
3.21
1.78
1.39
3.13
Linear retention index.
for 81.1% of the total variance. Discriminant analysis with the largest three principal components showed an overall significant differentiation of the three types of queens (Wilks’ k = 0.599, F6,58 = 2.816, p < 0.018). However, only 61.8% of the samples were correctly classified. Mismatches occurred particularly between the groups of shammated and mated queens and between sham-mated and virgin queens, suggesting that the chemical profiles of sham-mated queens are intermediate between those of virgin queens and mated queens. A comparison of squared Mahalanobis distances D2 did not indicate a clear separation of mated and sham-mated queens (D2 = 1.645, F3,29 = 2.687, p = 0.065) and sham-mated and virgin queens (D2 = 1.577, F3,29 = 2.458, p = 0.083). Mated and virgin queens were most clearly distinct (D2 = 2.031, F3,29 = 3.318, p = 0.034), but the difference becomes insignificant after Bonferroni correction for multiple testing. The large overlap among the three groups of queens is also indicated by a scatter plot with the two canonical functions (Fig. 1). Discriminant analysis on all six principal components with leave-one-out cross-validation qualitatively corroborates these results, with only 50.0% of all individuals correctly classified. Given that fecundity varied considerably within groups we tried to determine whether particular compounds were associated with mating and/or fecundity in a generalized linear model. Mating (mated and sham-mated queens) appeared positively correlated with percentages of five substances (3-MeC25, F = 6.177, p < 0.019; 11-/13-MeC27, F = 4.564, p < 0.041; 5-MeC27, F = 6.187, p < 0.019); 3-MeC27, F = 7.371, p < 0.011; 12-/14-MeC28, F = 7.109, p < 0.012) and negatively with peak 6, an isomer of C27:1 (F = 4.607, p < 0.040). Fecundity was negatively correlated with the proportions of peak 15, C29:1 and an unidentified sterol deriva-
Fig. 1. Plot of the two functions of the discriminant analysis of cuticular hydrocarbon profiles of sham-mated queens, mated queens, and virgin queens of the ant C. obscurior. Discriminant analysis does not reliably distinguish among the three different types of queens.
tive (F = 4.901, p < 0.035) and peak 19, C31:1 (F = 9.272, p < 0.005), but there was not any significant positive association between fecundity and the proportion of particular compounds. The cuticular profiles of eggs and queens overlapped only to some extent. We found ten substances on the eggs that were not present on the surface of queens. Eleven substances from the cuticular profile of queens could not be detected on the eggs. The major hydrocarbons present on the surface of eggs were the odd-numbered alkanes n-C25 (appr. 15%), n-C27 (appr. 32%), and n-C29 (appr.
1248
S. Will et al. / Journal of Insect Physiology 58 (2012) 1245–1249
13%). In addition, minor quantities of the respective methylbranched compounds were present, including some of the substances positively associated with mating status (3-MeC25, 3-MeC27, 11-/13-MeC27, Table 2). 4. Discussion Virgin queens of the ant C. obscurior differ from mated queens in fecundity and longevity and from queens mated to sterilized males in longevity (Schrempf et al., 2005). Despite of the presumably large physiological differences among the three types of queens, their cuticular hydrocarbon profiles are remarkably similar and can hardly be separated by discriminant analysis. The profiles of queens are dominated by n-alkanes, branched alkanes and alkenes with chain lengths between C25 and C33. Several of the most abundant compounds in our eight week old queens, i.e., C25, C27, 3-MeC27, and 11-/13-MeC27, have previously been found to dominate also the cuticular waxes of freshly eclosed virgin queens (Cremer et al., 2008). One of the nonacosenes (C29:1), which in our study appeared to be negatively correlated with fecundity, was the most abundant compound obtained from 1 day old virgin queens by solid sampling and also made up a considerable percentage of solvent-extracted hydrocarbons (Turillazzi et al., 2002; Cremer et al., 2008). Egg laying rate was also negatively associated with the proportion of C31:1, but our analysis did not give conclusive evidence for any positive association of specific substances with fecundity, i.e., the total number of eggs laid by the queens during the 8 weeks of the experiment. Instead, our study revealed five substances that were present in larger relative quantities on the cuticula of mated queens, regardless of whether these had mated with fertile or sterilized males, than on the cuticula of virgin queens. Among these were a few branched alkanes, which were also present on the surface of eggs.
Branched alkanes have previously been found to be associated with fecundity in other ants, e.g., Diacamma ceylonense (Cuvillier-Hot et al., 2001) and Camponotus floridanus (Endler et al., 2004), while particularly the proportion of 3-MeC27 on the cuticula of Leptothorax gredleri queens decreased rapidly after mating (Oppelt and Heinze, 2009). The absence of a positive correlation between the quantities of individual cuticular compounds and the total number of eggs laid by queens in our experiment is surprising. It contrasts with the growing number of studies in social insects according to which fecundity is associated with qualitative and quantitative changes in the cuticular profile (e.g., Monnin, 2006; Hefetz, 2007; Heinze and d’Ettorre, 2009; Liebig, 2010; Kocher and Grozinger, 2011). In a few cases, specific substances have been shown directly to elicit a response in the antennae of workers (d’Ettorre et al., 2004) or to serve as fecundity signals and regulate worker sterility (e.g., Holman et al., 2010). Such honest queen pheromones are important in maintaining the division or reproductive labor between queens and workers (e.g., Keller and Nonacs, 1993). What explains the apparent absence of fecundity-specific substances in C. obscurior and the resemblance of the profiles of egg laying virgin queens, sham-mated queens, and mated queens? On the one hand, the similarity of the profiles might reflect the fact that Cardiocondyla workers completely lack ovaries. Fertility signals therefore are not required for the regulation of reproduction between queens and workers. Worker sterility might also explain the simple hydrocarbon profile of C. obscurior eggs, which consists mostly of the linear alkanes n-C25, n-C27, and n-C29. These substances are among the most widespread hydrocarbons in insects (e.g., Lockey, 1980) and they are present in considerable quantities also on the surface of other social insect eggs (e.g., Endler et al., 2004; Bonckaert et al., 2012). Nevertheless, the egg profile of C. obscurior appears to be much simpler than those of other
Table 2 Median proportions (%) with upper and lower quartiles of egg surface hydrocarbons of the ant C. obscurior. ‘‘Queen peak ID’’ denotes substances identical to those on the cuticle of queens (see Table 1). Queen peak ID
1 3 7 8 9 10 11 14 15 16
LRIa
Compound
Diagnostic ions (m/z)
2400 2500 2571 2600 2700 2730
C24 C25 3-MeC25 C26 C27 13-MeC27 +11-MeC27 5-MeC27 3-MeC27 C28 C29:1 C29:1 C29 3-MeC29 C31 13-MeC31 +11-MeC31 3-MeC31 3,7-DiMeC31 C33 13-MeC33 +11-MeC33 15-MeC35 +13-MeC35 +11MeC35 13-MeC37 15-MeC39 +13-MeC39
338 352 337 366 380 196/197, 224/225 168/169, 252/253 84/85, 336/337 365 394 406 406 408 393 436 196/197, 280/281 168/169, 308/309 421 435, 126/127, 364/365 464 196/197, 308/309 168/169, 336/337 224/225, 308/309 196/197, 336/337 168/169, 364/365 196/197, 364/365 224/225, 364/365 196/197, 392/393
21
2748 2773 2800 2876 2885 2900 2974 3100 3128
22
3174 3206 3300 3326 3522
n.d.b n.d. a b
Linear retention index. Not determined (no reference hydrocarbons).
Eggs Median
25%
75%
0.54 14.83 0.55 1.60 32.40 0.72
0.52 14.48 0.45 1.58 29.97 0.59
0.55 15.42 0.63 1.65 34.51 0.85
0.43 3.47 1.61 0 0 13.01 1.39 4.58 1.37
0 3.24 1.61 0 0 12.94 1.18 4.26 1.21
0.43 3.52 1.81 0.98 0.47 15.18 1.55 5.75 1.74
0.76 0.64 1.22 1.91
0.52 0 1.13 1.88
0.78 0.92 1.59 1.98
1.39
1.36
1.59
3.71 0
3.53 0
3.82 1.01
S. Will et al. / Journal of Insect Physiology 58 (2012) 1245–1249
social insects, in that the three alkanes make up more than 60% of the egg surface hydrocarbons and branched alkanes are present only in minor quantities. On the other hand, observations in other species indicate that worker sterility alone is not sufficient to explain the similarities among queen profiles and the simple composition of egg hydrocarbons. Queen pheromones may have numerous additional functions other than keeping workers from reproducing (e.g., Vargo and Hulsey, 2000). In the Argentine ant, Linepithema humile, cuticular profiles of fertile and infertile queens differ considerably despite of complete worker sterility (de Biseau et al., 2004). Such variation in cuticular profiles might allow workers to preferentially take care of the most fecund queen. In C. obscurior, workers appear to treat the differently fecund types of queens equally (Schrempf et al., 2005), which matches the similarity of their cuticular profiles. In any case, our present data suggest that that fecundity signals play a less decisive role in this species than in other social insects. Acknowledgments Collecting and export of colonies was permitted by Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis (IBAMA, permit 20324-1) to J.H.C.D. Research was financed by DFG (He1623/23). Leonardo Dapporto and an anonymous referee made helpful comments on the manuscript. References Bonckaert, W., Drijfhout, F.P., d’Ettorre, P., Billen, J., Wenseleers, T., 2012. Hydrocarbon signatures of egg maternity, caste membership and reproductive status in the Common Wasp. Journal of Chemical Ecology 38, 42–51. Boomsma, J.J., Baer, B.C., Heinze, J., 2005. The evolution of male traits in social insects. Annual Review of Entomology 50, 395–420. Carlson, D.A., Bernier, U.R., Sutton, B.D., 1998. Elution parameters from capillary GC for methyl-branched alkanes. Journal of Chemical Ecology 24, 1845–1865. Cremer, S., d’Ettorre, P., Drijfhout, F.P., Sledge, M.S., Turillazzi, S., Heinze, J., 2008. Imperfect chemical female mimicry in males of the ant Cardiocondyla obscurior. Naturwissenschaften 95, 1101–1105. Cuvillier-Hot, V., Cobb, M., Malosse, C., Peeters, C., 2001. Sex, age and ovarian activity affect cuticular hydrocarbons in Diacamma ceylonense, a queenless ant. Journal of Insect Physiology 47, 485–493. De Biseau, J.-C., Passera, L., Daloze, D., Aron, S., 2004. Ovarian activity correlates with extreme changes in cuticular hydrocarbon profile in the highly polygynous ant, Linepithema humile. Journal of Insect Physiology 50, 585–593. D’Ettorre, P., Heinze, J., Schulz, C., Francke, W., Ayasse, M., 2004. Does she smell like a queen? Chemoreception of a cuticular hydrocarbon signal in the ant Pachycondyla inversa. Journal of Experimental Biology 207, 1085–1091. Endler, A., Liebig, J., Schmitt, T., Parker, J.E., Jones, G.R., Schreier, P., Hölldobler, B., 2004. Surface hydrocarbons of queen eggs regulate worker reproduction in a
1249
social insect. Proceedings of the National Academy of Sciences of the United States of America 101, 2945–2950. Hefetz, A., 2007. The evolution of hydrocarbon pheromone parsimony in ants (Hymenoptera: Formicidae) – interplay of colony odor uniformity and odor idiosyncrasy. A review. Myrmecological News 10, 59–68. Heinze, J., d’Ettorre, P., 2009. Honest and dishonest communication in social Hymenoptera. Journal of Experimental Biology 212, 1775–1779. Holman, L., Jørgensen, C.G., Nielsen, J., d’Ettorre, P., 2010. Identification of an ant queen pheromone regulating worker sterility. Proceedings of the Royal Society of London Series B: Biology 277, 3793–3800. Hora, R.R., Ionescu-Hirsh, A., Simon, T., Delabie, J.H.C., Robert, J., Fresneau, D., Hefetz, A., 2008. Postmating changes in cuticular chemistry and visual appearance in Ectatomma tuberculatum queens (Formicidae: Ectatomminae). Naturwissenschaften 95, 55–60. Holm, S., 1979. A simple sequential rejective multiple test procedure. Scandinavian Journal of Statistics 6, 65–70. Howard, R.W., Blomquist, G.J., 2005. Ecological, behavioral, and biochemical aspects of insect hydrocarbons. Annual Review of Entomology 50, 371–392. Johnson, R.A., Gibbs, A.G., 2004. Effect of mating stage on water balance, cuticular hydrocarbons and metabolism in the desert harvester ant, Pogonomyrmex barbatus. Journal of Insect Physiology 50, 943–953. Keller, L., Nonacs, P., 1993. The role of queen pheromones in social insects: queen control or queen signal? Animal Behaviour 45, 787–794. Kocher, S.D., Grozinger, C.M., 2011. Cooperation, conflict, and the evolution of queen pheromones. Journal of Chemical Ecology 37, 1263–1275. Liebig, J., 2010. Hydrocarbon profiles indicate fertility and dominance status in ant, bee and wasp colonies. In: Blomquist, G.J., Bagnères, A.-G. (Eds.), Insect Hydrocarbons. Biology, Biochemistry and Chemical Ecology. Cambridge University Press, Cambridge, pp. 254–281. Lockey, K.H., 1980. Insect cuticular hydrocarbons. Comparative Biochemistry and Physiology 65B, 457–462. Lone, S.R., Chakravarthi, A., Sharma, V.K., 2012. Sex and age related changes in the locomotor activity and phototactic behaviors of two closely related species of Camponotus ants. Journal of Insect Physiology 78, 75–82. Monnin, A., 2006. Chemical recognition of reproductive status in social insects. Annales Zoologici Fennici 43, 515–530. Nelson, D.R., 1993. Methyl-branched lipids in insects. In: Stanley-Samuelson, D.W., Nelson, D.R. (Eds.), Insect Lipids: Chemistry, Biochemistry, and Biology. University of Nebraska Press, Lincoln, pp. 271–316. Oppelt, A., Heinze, J., 2009. Mating is associated with immediate changes of the hydrocarbon profile of Leptothorax gredleri ant queens. Journal of Insect Physiology 55, 624–628. Reyment, R.A., 1989. Compositional data analysis. Terra Review 1, 29–34. Schrempf, A., Heinze, J., 2006. Proximate mechanisms of male morph determination in the ant Cardiocondyla obscurior. Evolution and Development 8, 266–272. Schrempf, A., Heinze, J., Cremer, S., 2005. Sexual cooperation: mating increases longevity in ant queens. Current Biology 15, 267–270. Sharma, V.K., Lone, S.R., Goel, A., 2004. Clocks for sex: loss of circadian rhythms in ants after mating? Naturwissenschaften 91, 334–337. Singer, T.A., 1998. Roles of hydrocarbons in the recognition systems of insects. American Zoologist 38, 394–405. Turillazzi, S., Sledge, M.F., Cremer, S., Heinze, J., 2002. A method for analysing smallsize specimens in GC–MS. Insect Social Life 4, 169–175. Vargo, E.L., Hulsey, C.D., 2000. Multiple glandular origins of queen pheromones in the fire ant Solenopsis invicta. Journal of Insect Physiology 46, 1151–1159. Williams, B.K., Titus, K., 1988. Assessment of sampling stability in ecological applications of discriminant analysis. Ecology 69, 1275–1285.