Pollination by sexual deception via pro-pheromone mimicry?

Most plant species worldwide depend on insects for pollination (Ollerton et al., 2011), with volatile organic compounds being pivotal for mediating pollinator attraction in many of these plants (Raguso, 2008; Dötterl & Gershenzon, 2023). Among plants, orchids are exceptional in their extraordinary range of pollinators, pollination strategies, and floral volatiles (Ackerman et al., 2023; Perkins et al., 2023). One of the most remarkable pollination strategies is that of sexual deception, where the flower imitates female insects to attract male pollinators, with sex pheromone mimicry typically being key to pollinator attraction (Schiestl 2005; Ayasse et al., 2011). While the chemical basis of the sexual mimicry and the extreme pollinator specificity has been confirmed by field bioassays with synthetic compounds for a growing number of sexually deceptive orchids (see Bohman et al., 2016a; Bohman et al., 2020a; Peakall et al., 2020), these examples represent just a tiny fraction of the hundreds of known cases of orchids employing this pollination strategy (Johnson and Schiestl 2016; Peakall, 2023).

Australia is home to a high proportion of sexually deceptive orchids, where several hundred species spanning 11 genera are now known to use this strategy (Gaskett, 2011; Peakall, 2023). Cryptostylis was the first Australian orchid genus discovered to be sexually deceptive (Coleman, 1927), with all five Australian species dependent on the same pollinator, the orchid dupe wasp, Lissopimpla excelsa Costa (Ichneumonidae) (Coleman, 1927, 1929, 1930a, 1930b; Nicholls, 1938). While attempted copulation (pseudocopulation) is not always necessary for pollination (Peakall, 2023), Cryptostylis represents an extreme amongst sexually deceptive plants as one of only two confirmed cases (the other being the beetle-pollinated Disa forficaria (Cohen et al., 2021)) where flowers induce ejaculation by some male pollinators (Coleman, 1930b; Gaskett et al., 2008). While it is almost 100 yr since Coleman conducted simple experiments with Cryptostylis revealing that wasps could locate hidden flowers, leading to her astute conclusion that scent and mimicry were involved in this case of pollination by sexual deception (Coleman, 1930a), the compounds responsible for pollinator attraction have only just started to be elucidated. In previous experiments with (S)-2-(tetrahydrofuran-2-yl)acetic acid from Cryptostylis ovata R.Br, only close approaches by L. excelsa have been observed (Bohman et al., 2019). As such, it is still unknown what induces attempted copulation in male L. excelsa, suggesting that additional chemical cues remain to be discovered.

While the Ichneumonidae is one of the most diverse families of Hymenoptera, with over 25 000 species known (Yu et al., 2016), sex pheromones released by females have been structurally elucidated for just three ichneumonid species (Bohman et al., 2019). In Itoplectis conquisitor, a blend of neral and geranial elicits male sexual activity (Robacker & Hendry, 1977). Eller et al. (1984) found that Syndipnus rubiginosus uses ethyl (Z)-9-hexadecenoate as its sex pheromone, while in Campoletis chlorideae, tetradecanal and 2-heptadecanone have been identified as the sex pheromones (Guo et al., 2022). Limited data for ichneumonids make it difficult to predict the likely compounds involved in inducing attempted copulation in L. excelsa with Cryptostylis.

Sex pheromones are typically female-produced volatile compounds that underpin the chemical sexual communication among conspecifics (Witzgall et al., 2010). Some insects, however, produce precursors (pro-pheromones) that are subsequently modified by external processes to become bioactive compounds. For example, relatively nonvolatile unsaturated long-chain hydrocarbon pro-pheromones have been shown to be oxidatively cleaved in air into attractive volatile aldehydes, as demonstrated in sawflies (Bartelt et al., 1982, 2002; Bartelt & Jones, 1983; Cossé et al., 2002; Staples et al., 2009), flies (Collignon, 2011; Lebreton et al., 2017), cockroaches (Hatano et al., 2020), beetles (Wickham et al., 2012) and wasps (Swedenborg & Jones, 1992; Xu et al., 2020; Faal et al., 2022). Recently, it has also been shown that in poplar and corn, instead of short-chain aldehydes being directly biosynthesised by the plant, unsaturated waxes are produced that are oxidatively cleaved to yield the bioactive aldehydes, such as nonanal (Chen et al., 2023). So far, no evidence of pro-pheromone mimicry by plants has been presented.

The objective of this study was to investigate whether pro-pheromone mimicry may be involved in the sexual attraction of the ichneumonid wasp pollinator L. excelsa to the orchid Cryptostylis ovata, thereby providing the first evidence of the involvement of pro-pheromone mimicry in pollination.

After demonstrating that (S)-2-(tetrahydrofuran-2-yl)acetic acid derivatives constitute attractants that C. ovata uses to lure male L. excelsa as pollinators (Bohman et al., 2019), we have been working to discover the missing compound(s) required to elicit the pseudocopulatory behaviour that the pollinators frequently exhibit at the flowers. To shorten the list of candidate floral volatiles from whole flowers, we dissected C. ovata flowers into four sections (Fig. 1c), which were tested individually in field bioassays to see which parts were most attractive to pollinators. In total, 382 wasp responses were recorded, with average responses per trial to the centre (3.53 ± 0.43 SE) and tip (4.18 ± 0.53 SE) of the labellum being significantly higher (ANOVA: P < 0.01, pairwise t-tests: P < 0.05) than to the labellum base (1.82 ± 0.36 SE) and sepals/petals (1.71 ± 0.30 SE). Furthermore, attempted copulation was observed at both the large centre and tip sections of the labellum, which contrasts with most other sexually deceptive orchids (Perkins et al., 2023), where the attraction is pinpointed to small sections of the flower. For example, among the Australian species, the small callus structure of the labellum is the attractive tissue in Chiloglottis (de Jager & Peakall, 2016) and Drakaea (Phillips et al., 2013), while the glandular sepal tips are the source of attraction in many Caladenia (Phillips et al., 2024).

The chemical compositions of the various floral parts were analysed with GC-MS, and the components of the more attractive floral tissue were compared with those of less attractive floral tissue. We detected C₂₃- and C₂₅-unsaturated hydrocarbons in the attractive floral tissue (Fig. 1e). Meanwhile, in females of L. excelsa, C₂₃-unsaturated hydrocarbons represented the largest peak in the solvent extracts of all dissections, whereas only traces of C₂₅-alkenes were detected. After DMDS derivatisation to determine alkene positional isomers (Bohman et al., 2020c), GC-MS analysis revealed that two main compounds, 8-tricosene (characteristic fragment ions of m/z = 159/257) and 8-pentacosene (characteristic fragment ions of m/z = 159/285), were present in the flowers, with traces of 10-pentacosene (characteristic fragment ions of m/z = 187/257). In female L. excelsa, 8-tricosene was the prominent alkene, with a minor second alkene identified as 6-tricosene (characteristic fragment ions of m/z = 131/285). Therefore, (Z)-8-Tricosene (2) and (Z)-8-pentacosene (3) were synthesised and, together with the previously identified (S)-2-(tetrahydrofuran-2-yl)acetic acid (1) (Fig. 1b), formulated into blends for field bioassays.

Consistent with our previous findings (Bohman et al., 2019), tests with the tetrahydrofuran acid 1 only, on 3D plastic models of C. ovata flowers, elicited rapid approaches (A) from male wasps, but only one landing (L) and no attempted copulation (C) at the spiked dummy (1 only: A = 21, LC = 1, %C = 0, 12 trials over 3 d). By contrast, in preliminary experiments involving 1 and the alkenes 2 and 3, in ratios of 1 : 2 : 3 varying from 1 : 10 : 5 (in total c. 0.2 mg), 1 : 100 : 50 (in total c. 2 mg) to 10 : 100 : 50 (in total c. 2 mg), landings (L) were induced and one attempted copulation (C) was observed (combined results: A = 124, LC = 18, %C = 6, 36 trials over 4 d). Further preliminary tests over 2 d, with threefold more concentrated alkenes, 2 and 3, indicated increased landings and attempted copulation (1 : 2 : 3 at 3 : 300 : 150: A = 45, LC = 33, %C = 27, 18 trials over 2 d). Over this same period, a negative control, consisting of dummies with only solvent, did not induce any approaches, landings or attempted copulation (12 trials over 2 d). In the light of these preliminary results, a 3 : 300 : 150 blend of tetrahydrofuran acid and alkene compounds 1 : 2 : 3 was established as the baseline control blend in subsequent experiments.

Although similar mono-unsaturated long-chain hydrocarbons are pollinator attractants in some sexually deceptive Ophrys orchids (Schiestl et al., 1999, 2000, 2003; Stökl et al., 2005), the very large amounts required to induce frequent sexual behaviour of L. excelsa led us to conclude that these rather nonvolatile compounds were unlikely to be the active compounds. Therefore, we hypothesised that these alkenes, present primarily in the active floral tissue, might be precursors to other, more volatile sexual attractants. Apart from the requirement of unusually large amounts of alkenes for activity, this hypothesis was also based on: (1) previous discoveries of aldehydes as sex pheromones from other parasitic wasps, which are proposed to be products of oxidative cleavage of unsaturated cuticular hydrocarbons (Swedenborg & Jones, 1992; Xu et al., 2020); (2) our observation that the pollinator-attractive parts of C. ovata make up an unusually large area of the flower for a sexually deceptive orchid (Fig. 1a,c), potentially providing a relatively large surface area where oxidative cleavage can take place; and (3) that the alkenes are uncommon in nature, with double bonds located on even-numbered carbons, potentially giving rise to unusual aldehydes.

In a targeted analysis for the putative aldehyde products from our identified alkenes, proposed to be pentadecanal (4), heptadecanal (5) and octanal (6), we found small amounts of compound 4 and traces of 5 in the attractive floral tissues, although we were unable to detect any octanal (6). In the solvent extracts of females of L. excelsa (n = 3), traces of all aldehydes 4–6 were found, thus warranting the testing of these aldehydes in combination with 1 for pollinator attraction.

The combination of 1 and the aldehydes 4–6 (c. 0.003 mg), despite 1000-fold lower levels of the aldehydes relative to the alkenes, elicited both significantly higher total counts per trial and significantly more LC counts than the baseline control blend (Fig. 2c; Supporting Information Video S1: video of copulatory behaviour on the dummies; 1, 4–6: A = 19, LC = 27, %C = 29.6, 12 trials over 2 d, 1 : 2 : 3 at 3 : 300 : 150, A = 23, LC = 14, %C = 14.3, 19 trials over 2 d, Fig. 2c). Furthermore, the 1, 4–6 blend yielded the highest percentage of attempted copulation observed across the study (29%).

When 1 was removed from the blend with alkenes 2 and 3, there were significantly fewer total and LC counts per trial, and no attempted copulation (2 : 3 at 300 : 150, A = 12, LC = 2, %C = 0, 12 trials over 2 d, 1 : 2 : 3 at 3 : 300 : 150, A = 27, LC = 13, %C = 23, 12 trials over 2 d, Fig. 2a). Similarly, when compound 1 was removed from a blend of the aldehydes 4–6, there were significantly fewer total counts per trial than the baseline control (4–6: A = 5, LC = 1, %C = 0, 12 trials over 2 d, 1 : 2 : 3 at 3 : 300 : 150, A = 28, LC = 15, %C = 13.3, 24 trials over 2 d, Fig. 2b).

To summarise our results, the tetrahydrofuran acid 1 on its own rapidly elicits approaches but rarely lands, and no attempted copulation. In the absence of 1, total wasp responses were significantly reduced in experiments with alkenes 2 and 3 only (Fig. 2a) and the aldehydes 4–6 only (Fig. 2b). These findings indicate a key role of the acid 1 in longer-range pollinator attraction, as previously predicted (Bohman et al., 2019). The blend eliciting the strongest attraction (total count), strongest sexual response (LC counts) and the highest observed attempted copulation of 29.6% was the combination of the acid 1 with small amounts of the aldehydes 4–6 (Fig. 2c). While the percentage of wasps engaging in attempted copulation with our synthetic blend of (1 + 4–6) was close to the maximum of 33% reported at patches of multiple inflorescences (Weinstein et al., 2016), attempted copulation rates might be enhanced at the flower by additional visual and physical cues.

Overall, despite the alkenes 2 and 3 being dominant peaks in the floral and wasp extracts (Fig. 1e–g), the required amounts to achieve pollinator attraction was unnaturally high, even when allowing for their low volatility. However, the aldehydes 4–6, derived from 2 and 3, were present in trace amounts in the female wasps, but were far more attractive in bioassays and induced attempted copulation at 1000-fold lower quantities than the alkenes in the baseline blend. A plausible explanation is that 2 and 3 are pro-pheromones for L. excelsa, which are oxidatively cleaved to produce the aldehydes 4–6 as the active attractants, alongside 1. Given the abundance of 2 and 3 in C. ovata flowers, we suggest this may be the first known example of a pro-pheromone mimicry pollination system, in this case involving the ‘passive’ oxidation of alkenes to aldehydes. Given that all five Australian Cryptostylis species use the same male wasp pollinator, L. excelsa, we expect this strategy will apply to all of them.

To test the pro-pheromone mimicry hypothesis in Cryptostylis and in sexually deceptive orchids more broadly, it will be imperative to analyse flowers from different developmental stages and ages, as exposure to heat, UV radiation and other environmental factors would affect the ratio of alkenes and aldehydes. Application of antioxidants on flowers, similar to what has previously been experimented with for insect pro-pheromones (Bartelt & Jones, 1983), and transcriptomic studies aiming to elucidate the biosynthetic pathways in C. ovata (e.g. Xu et al., 2017), are needed to further test this hypothesis. In particular, transcriptomic studies could provide evidence of whether or not any or all aldehydes are biosynthesised by the flower. To detect other cases of possible pro-pheromone mimicry in sexually deceptive orchids, it may be worth re-opening investigations into orchids that use alkenes as pollinator attractants and/or have low rates of attempted copulation in bioassays.

To test whether orchid pollination systems suggested to utilize pro-pheromone mimicry operate in a similar fashion to other sexually deceptive orchids, requirements for structural specificity of the alkenes and aldehydes should be investigated. It is noteworthy that the alkenes in C. ovata and L. excelsa have the unusual position of the double bonds on even-numbered carbons, giving rise to the uncommon aldehydes pentadecanal and heptadecanal. Structurally specific and/or unusual aldehydes may be required to achieve the specific communication with the targeted pollinator. Alternatively, specificity could also rely on the uniqueness of other attractants, such as 1, acting in concert with aldehydes more broadly.

Oxidative cleavage of cuticular wax alkenes to aldehydes in plant leaves was only discovered very recently (Chen et al., 2023). Here, we show that a sexually deceptive plant may also rely on this mechanism for pollinator attraction. Pro-pheromones and pro-pheromone mimicry may actually be major, but often overlooked, strategies of mate attraction in insects (Bartelt et al., 2002; Hatano et al., 2020) and the plants that mimic them, particularly since many aldehydes have extremely low detection thresholds by their receivers (Lebreton et al., 2017; Becher et al., 2018), allowing these compounds to operate at very low concentrations, thereby going undetectable in routine analysis.

None declared.

RP, RDP and BB conceptualised the study; BB, SvK, AMW and BO collected the data; RP, BB, AMW and GRF analysed the data. BB and GRF synthesised and confirmed chemical structures. BB and RDP wrote the first version of the manuscript. All authors edited the manuscript.

The New Phytologist Foundation remains neutral with regard to jurisdictional claims in maps and in any institutional affiliations.