Counter-perfume: using pheromones to prevent female remating

(1) Coleoptera

The first record of an AA pheromone was in the mealworm beetle Tenebrio molitor (Happ, 1969). This is a highly polygynandrous species in which the AA pheromone acts shortly after mating by reducing long-distance attractiveness of females (Carazo, Fernández-Perea & Font, 2012). The only other coleopteran known to express an AA is the rove beetle Aleochara curtula, which transfers the pheromone through a spermatophore (capsule encompassing the spermatozoa and containing nutrients transferred from males to females during copulation; Schlechter-Helas, Schmitt & Peschke, 2011). Both species have very high last-male sperm precedence (Simmons, 2001; Drnevich, 2003), i.e. most of the eggs are fertilized by the sperm of the last male to mate with a female before oviposition. Some other species of Coleoptera have been argued to use AAs, but they do not act directly as a repelling signal; instead, they mask female pheromones (Trypodendron lineatum; Nijholt, 1973) or induce a refractory period in females that makes them unreceptive (Leptocarabus procerulusi; Takami et al., 2008).

(2) Lepidoptera

AAs have been best documented in lepidopterans. Gilbert (1976) first recorded them in Heliconius erato, with recent studies investigating further Heliconius species (Schulz et al., 2008; Estrada et al., 2011). The chemical compound was identified as beta-ocimene in H. melpomene (Schulz et al., 2008). In the most comprehensive study of AA evolution to date (Estrada et al., 2011) AAs compounds were identified and compared across 11 Heliconius species. This study showed an absence of phylogenetic signal, suggesting a rapid diversification among species. Heliconius butterflies encompass two clades: one is monoandrous with pupal mating and the other polyandrous. Estrada et al. (2011) found a faster rate of divergence in AA chemicals in the polyandrous group with more sperm competition as opposed to the pupal-mating group. Another comparison of AAs among species of a single genus was performed by Andersson et al. (2000); Andersson, Borg-Karlson & Wiklund (2003) in pierid butterflies. Pieris napi uses methyl salicylate as an AA (Andersson et al., 2000), which is also used by P. rapae in association with indole, whereas P. brassicae uses benzyl cyanide (Andersson et al., 2003). In the latter species, the fragrance transferred is a male-specific pheromone, suggesting that mated females become unattractive by expressing a male signal. The three pierid butterflies seem to transfer their AA through the spermatophore (Andersson et al., 2000, 2003). Both genera of butterflies described here exhibit last-male sperm precedence (Gilbert, 1976; Simmons, 2001; Andersson et al., 2003).

The most recent lepidopteran recorded to express AA pheromones is the armyworm moth, Pseudaletia unipuncta (Noctuidae; Hirai, Shorey & Gaston, 2013). Unlike other lepidopteran species using AAs, this moth has relatively low last-male sperm precedence (the proportion of offspring sired by the last male in a two-males experiment p2 = 0.47; Simmons, 2001). The specific AA compound has not been identified, but mated females have a scent which is associated with the expression of benzaldehyde, making this molecule the suspected AA. In other moth species, females become unattractive or unreceptive after mating because male-transferred molecules switch off the production of sex pheromones in the female, therefore acting only indirectly as an AA (Raina, 1989; Giebultowicz et al., 1991; Gillott, 2003).

(3) Diptera

AAs have also been studied in species of Drosophila and in the tsetse fly Glossina morsitans (Table 1; Thomas, 2011). Both D. melanogaster and G. morsitans have cuticular pheromones transferred by contact during mating that function as AAs (Scott, 1986; Carlson & Schlein, 1991). In addition to these, D. melanogaster transfers two other AAs in the seminal fluid: the compound cis-vaccenyl acetate and the long-chained acetate CH₅O₃ (Ejima et al., 2007; Yew et al., 2009). Cis-vaccenyl acetate is typically found in the ejaculatory bulb and cuticle of males only, making it a male-specific signal. Thus, recently mated females become unattractive by expressing a male fragrance, similar to the situation in the lepidopteran P. brassicae. However, the cis-vaccenyl acetate is only expressed for up to two days after copulation so the only AA that female fruit flies express for a longer time is the CH₅O₃, which is present for at least 10 days post-copulation on the surface the anogenital region (Yew et al., 2009). D. melanogaster is the only species recorded to use a combination of different AAs that are chemically distinct and are transmitted through different modes. This may be due to the unusually high last-male sperm precedence in this species (p2 = 0.93 compared to p2 = 0.11–0.45 in G. morsitans; Simmons, 2001) imposing a strong selection pressure on males to be the last to mate with a female before fertilization.

The three other Drosophila species found to use AAs are part of the D. repleta group – drosophilids adapted to desert environments that feed on fermented cacti. Their AA compounds are a mixture of triacylglycerides (lipids typically found in fat bodies that function in energy storage) acting synergistically with long-chain acetyldienyl acetates (Chin et al., 2014). These are unusual chemicals to be used as pheromones, particularly when compared to the volatile compounds used as AAs in most other species, but this may be related to the unique diet of these desert-living flies. Alternatively, such chemical compounds may be present as AAs in other species too, but have not been identified in pheromone studies because they are not detectable with the typical methods used to identify volatile compounds (gas chromatography and mass spectrometry; Yew et al., 2009). Chin et al. (2014) found that the triacylglycerides used as pheromones are conserved across the D. repleta group as well as the related D. virilis and D. robusta taxa. It is thus likely that AA pheromones are used in all 11 species making up these three taxa.

(4) Hymenoptera

AAs have been recorded in several hymenopteran families. The male bee Osmia rufa (Megachilidae) uses cuticular transfer to transmit 7-ethylhexadecenoate to the female by rubbing the female's wings with its stermites after copulation (Ayasse, Paxton & Teng, 2001). Other bees in which AA pheromones are known include Centris adani (Ayasse et al., 2001), the sweat bee Lasioglossum malachurum, and possibly Lasioglossum zephyrum, which was initially reported to use an AA (Kukuk, 1985) although behavioural experiments failed to confirm this (Wcislo, 1987). Finally, the bumble bee Bombus terrestris uses linoleic acid as an AA in a mating plug (Baer & Boomsama, 2004). The mating plug itself is not enough to preclude subsequent mating, as other males can displace it, whereas the presence of linoleic acid is effective. Linoleic acid has also been suspected to act as an AA in the fire ant Solenopsis invicta. The same fatty acids are produced in this ant as in the bumble bee B. terrestris, including linoleic acid which is transferred to the female during mating (Mikheyev, 2003). Its use as an AA remains to be proven but seems highly plausible.

(5) Hemiptera

Only one hemipteran species, Lygus hesperus, has been shown to use an AA (Brent & Byers, 2011). As in pierid butterflies, this compound – myristyl acetate – is transmitted through the spermatophore (Brent & Byers, 2011). Two species of the same genus, L. lineolaris and L. elisus, are also suspected to use myristyl acetate as an AA but this remains to be confirmed. Myristyl acetate is present in males and mated females but not in virgin females, which is similar to the pattern observed in L. hesperus.

(6) Orthoptera

A pheromone similar to an AA has been identified in the desert locust Schistocerca gregaria and the term AA has been used to describe it (Seidelmann & Ferenz, 2002). Because there is total sperm precedence in this species (p2 = 1; Simmons, 2001), gregarious males remain in close proximity to the female to prevent remating until egg laying. But in crowded situations other males are able to separate the pair and do so if the repelling pheromone phenylacetonitrile is not emitted. However, this phenylacetonitrile is secreted directly by the male, therefore not falling strictly into the definition of an AA pheromone which must be transferred to the female. A similar finding has been reported in the sweat bee Lasioglossum malachurum (Ayasse, 1994) where an ester emitted from the cuticle of males during copulation repels other males. The inhibitory effect on other males is specifically efficient with individuals from the same population.

(7) Squamata

The only records of AA pheromones outside of the Insecta are in the garter snakes Thamnophis radix (Ross & Crews, 1978) and T. sirtalis parietalis (Friesen et al., 2013). These pheromones are transferred through a mating plug in T. radix, whereas the transfer is more likely in the seminal fluid and independent of the mating plug in T. sirtalis parietalis (Shine, Olsson & Mason, 2000). The AA compound has yet to be identified, although Mason & Parker (2010) suggested that it might be squalene. Squalene is a potential sex identifier since it is found in the skin of male garter snakes but not females, and Shine et al. (2005) showed that experimentally applying squalene on females repels males. However, artificially adding a male-specific molecule to females would make them unattractive even if this compound is not used as an AA in nature, and thus the role of squalene as an AA awaits further investigation.

(8) Convergent evolution?

The above review shows that AAs are present in many different taxa within the insects, with garter snakes the sole example outside this group. Although AAs are found in multiple insect orders, only a few species within each order use these pheromones, suggesting multiple cases of independent evolution of AAs. This is also supported by the diversity of transmission modes: seminal fluid, cuticular contact, mating plug, and spermatophore. Indeed, transmission mode does not correlate strictly with taxon, although closely related species often use the same transmission mode.

In some instances, the AA compound is the same as a male-specific pheromone, thus making mated females signal themselves with a male identifier to repel potential mates. This strategy seems less costly and easier to evolve than the use of a new chemical, which may be why it arose in three different taxa. In particular, using a male-specific odour as an AA can be highly beneficial in species where the AA is not an honest signal of female receptivity. In that case, selection pressures would favour other males to evolve resistance to AA signal, which could be avoided if the signal is also used as honest information in male–male communication.

The convergent evolution of AA pheromones among different taxa suggests substantial benefits to their acquisition. If such benefits are present, why are AAs restricted to these few taxa, and are the benefits the same for males and females?

(1) Males

In promiscuous species, as well as in socially monogamous species where females exhibit extra-pair matings, males are subject to strong intra-sexual competition in order to access females and secure paternity. This leads to numerous morphological, physiological, chemical and behavioural adaptations. Examples of such adaptations include the evolution of weapons to fight other males for access to females (Emlen, 2008), increased testosterone levels (Martínez-Padilla et al., 2014), sperm competition between the seminal fluid of different males (Edward, Stockley & Hosken, 2015) and physical mate guarding (Edward et al., 2015). Sperm competition can either affect the evolution of the sperm itself, for instance by increasing its size or longevity (Parker, 1998), or it can affect the other molecules present in the ejaculate (Cameron, Day & Rowe, 2007). These seminal fluid contents transmitted along with the sperm increase reproductive success in three different ways: they have an ‘offence’ effect when they replace sperm already present in the female; a ‘defence’ one when they protect their sperm from being displaced by future males; or an ‘avoidance’ effect when they prevent sperm competition from happening altogether (Dhole & Servedio, 2014). Antiaphrodisiacs fall into the avoidance category: they make the female appear unattractive and thus prevent future sperm competition from happening. This is a strong benefit for males, particularly in species with a high degree of last sperm precedence.

If the selection pressure to secure paternity through access to females is strong enough to lead to extreme morphological adaptations like the eyestalks of the stalk-eyed fly, or the horns of Hercules beetles (Emlen, 2008), we can expect a similar pressure on the evolution of the ejaculate compounds transmitted to females during mating. Any such compound that provides reduced competition from the sperm of other males would thus be highly beneficial to the individual producing it.

Additionally, AAs can benefit males in two other ways. First, when females have a refractory period after mating AAs can function as honest signals of females' lack of receptivity and prevent subsequent males from wasting time and energy courting unreceptive females (Daly, 1978; Ross & Crews, 1978; Andersson et al., 2003; Estrada et al., 2011). This is especially important in species where copulation cannot be forced, such as some species of Lepidoptera (e.g. Pieridae and Heliconius species; Gilbert, 1976; Andersson et al., 2003). Second, if males court females despite the expression of AAs and copulation ensues, they can use these pheromones as information that there is likely to be sperm competition due to the female's previous mating. Having that information can influence their mating behaviour to maximize fitness, as has been observed in P. napi. The males of this butterfly species use the expression of the AA methyl salicylate to assess the female's mating status, and transmit an ejaculate on average 26% larger to mated females than to virgin ones to improve their chances of fertilizing the eggs in the presence of competition with the sperm of a previous male (Larsdotter-Mellström et al., 2015).

The use of AA pheromones thus has multiple advantages for males, the major one being an increased chance of fertilizing eggs by reducing sperm competition. The effects on females, however, are more complex.

(2) Females

Even though the optimal number of partners is lower for females than for males (Bateman's principle), they can still benefit from multiple matings (Eberhard, 1998). These benefits might be direct by providing fertility insurance and nutrients to the female (Hosken & Blanckenhorn, 1983), as is the case in some lepidopteran species. For example, female pierid butterflies gain nuptial gifts during copulation events (Andersson et al., 2003), and in Heliconius erato additional spermatophores increase female fecundity (Gilbert, 1976). Benefits of multiple matings might also be genetic by compensating for low-quality partners (Hosken & Blanckenhorn, 1983), avoiding genetic incompatibilities (Hosken & Blanckenhorn, 1983), acquiring ‘good genes’ (Yasui, 1998), and enhancing the genetic diversity of offspring – which favours different gene by environment interactions (Yasui, 1998). AAs incur a cost to females by reducing their chances of remating and thus of accessing these various benefits of multiple mating.

Nevertheless, most females have a refractory period after mating. In this case, it is advantageous for them to have an honest signal of their mating status that will reduce harassment from males, therefore saving females time and energy (Gilbert, 1976; Andersson et al., 2003; Estrada et al., 2011). In garter snakes, for example, removing the mating plug containing AAs makes the female attractive again, but she is still unreceptive. In this species, as in many others, avoiding male harassment is especially important because mating increases exposure to predators. Other organisms might experience similar trade-offs, such as between mating and laying eggs (e.g. pierid butterflies; Andersson et al., 2003). In addition to reducing male harassment, AAs can benefit females in species where mating damages females' organs, as is for example the case in D. melanogaster where the male wounds the female's vaginal wall during copulation (Mattei et al., 2015). The expression of AA pheromones in this species might allow females to avoid unnecessary matings and the ensuing internal damage. This could explain why D. melanogaster has several different molecules used as AAs; a sex pheromone is more likely to evolve if it has significant benefits for both sexes. Alternatively, this apparent higher diversity of AA pheromones in the fruit fly may simply result from the higher number of studies on this model organism.

In Heliconius erato, females seem to have evolved specific structures called ‘stink-clubs’ to disseminate AAs (Eltringham, 1925; Gilbert, 1976), suggesting selection pressures favouring AAs in this sex. The impact of AAs on female fitness then depends on the control females have on the release of the pheromones. An honest signal is beneficial for females, but if AAs are expressed for longer than the female is unreceptive this becomes a cost (Andersson et al., 2003; Estrada et al., 2011). In most species, it seems that females do not have such control on the release of AAs, but further research is needed to explore this. There is no evidence of evolution of resistance to AAs by females, which would be expected if these pheromones were highly detrimental to their fitness, but this has not been investigated systematically.

In several species, there is evidence for both evolutionary advantages and disadvantages to AAs in females. In H. erato for instance, females suffer the cost of a reduced number of spermatophores, and the resources they contain, if they mate only once, but have nonetheless evolved structures to disseminate male-transferred AA. Similarly, it is unclear whether females of the rove beetle Aleochara curtula benefit or suffer a cost from AAs: mated females do not alter their production of aphrodisiac pheromones, which suggests that they are still willing to mate after copulation, but females without AA refuse to do so and spend 60% of their time repulsing males (Peschke, 1987; Schlechter-Helas et al., 2011).Whether AAs are favourable or costly to females would influence how these pheromones interact with other components of the ejaculate transferred during mating, notably in situations where AAs are transferred through a mating plug or spermatophore. Mating plugs are detrimental to females' fitness since they prevent the females from remating without any apparent female control on the presence of the plug. Thus, the expression of AA pheromones transmitted through this route is probably not controlled by females and is presumably deleterious to them. Alternatively, AAs may act as a signal when transferred with a plug, preventing subsequent males from wasting time courting a female that cannot mate. Even if the female would ultimately benefit from a second mating, the plug does not allow this and her best strategy would be to not waste time courting or rejecting other males. Spermatophores, on the other hand, often benefit females by providing them with resources (Boggs, 1995). We could therefore expect AAs to be beneficial for females in species where they are transmitted through this route, as females may have more control on their expression. But it could also be a successful evolutionary strategy for males to transmit costly pheromones with beneficial resources since that would make the evolution of resistance in females less likely.