Emission of the aphid alarm pheromone could serve as a defense against aphids, but its value to plants depends on both the costs and benefits of emission. In previous work with the transgenic A. thaliana lines studied here, Beale et al.  could detect no metabolic costs based on the similar size, growth rate and flowering time of EBF-emitting lines and wild-type controls. Here it was shown that the growth (Figure 4) of EBF-emitting and wild-type plants were also not different from each other under two fertilizer regimes where the low fertilizer regime caused a sharp reduction in plant growth. In addition, we demonstrated that seed production also did not differ among lines emitting EBF and wild-type controls (Figure 5). Thus, there is no evidence for any metabolic costs associated with EBF production in these lines. This conclusion may apply to other plant species as well since the rate of EBF emission from the A. thaliana transgenic lines is in the same range as that of EBF emission reported for other plant species [19–21]. The lack of metabolic costs may simply be due to the low rate of production. The amount emitted in 24 hours corresponds to less than 0.1 ‰ of the fresh weight of the above ground biomass. However, metabolic costs of EBF emission might conceivably be observed under other stress conditions, such as greater nutrient limitation, low light, drought or various biotic stresses [33–35]. Volatile emission can also attract additional herbivores and can therefore generate ecological costs . The biosynthesis of EBF and other sesquiterpenes requires substantial amounts of fixed carbon and energy for substrates and cofactors . Additionally, the diversion of farnesyl diphosphate (FPP) to EBF synthesis may directly reduce the supply of sterols (a major group of membrane components) and other isoprenoid metabolites produced from FPP. Beale et al.  noted that, when flowering, EBF-producing A. thaliana plants emit lower amounts of other sesquiterpenes than wild-type plants.
The transgenic, EBF-producing A. thaliana lines expressed the EBF synthase gene under the control of a constitutive promoter. Hence the enzyme should be present in almost every cell and produce EBF whenever and wherever FPP is available. From this perspective, the level of EBF produced is an indicator of the size of the FPP pools. Given the increase in emission during the day as compared to the night period (additional file 3), these pools appear to be larger during the light phase, possibly due to the action of photosynthesis. Of the two basic isoprenoid pathways operating in plant cells, the MEP pathway is strongly stimulated by light . A direct relationship between the rate of photosynthesis and the rate of EBF formation is consistent with the trend observed for greater emission from larger plants (Figure 1), which presumably have more active photosynthetic leaf area. Fertilization also promoted EBF formation (Figure 3), probably by increasing plant size (Figure 4).
Since no metabolic costs of EBF emission were detectable, any effect of EBF against aphids should be advantageous for the plant fitness as long as there are no high ecological costs involved. It was tested whether EBF emission could protect plants from aphids in several ways, by (1) repelling insects, (2) reducing their reproduction and (3) inducing wing formation. In tests for repulsion, both winged and wingless aphids were used since it is known, that winged morphs (migrants) can respond differently from wingless morphs to visual and olfactory stimuli  due to differences in their sensory systems . However, neither morph was repelled by constant EBF emission from the transgenic lines, and neither morph distinguished between emitting and control plant lines (Figure 7) under the conditions employed. One reason for this behavior could be that aphids chose the plant at night when EBF emission is strongly reduced. However, aphids are known to be active during the day time , and were observed to start searching for a host plant shortly after the beginning of the experiment in the light period.
In addition to olfaction, visual cues are important in aphid host choice, especially for winged aphids  and landing involves a phototactic response to the wavelengths reflected by the plant . For M. persicae, three spectral types of photoreceptors are known (near UV: 330 - 340 nm; blue - green: 490 nm; and green: 530 nm)  and yellow colors act as very strong stimuli [42, 44]. If the plants tested had differed in their leaf reflectance, this cue could have influenced the aphid's host plant choice. However, leaf reflectance of EBF emitting plants did not differ from that of wild type plants (Figure 6). The size of the color stimulus is also important in aphid attraction , but again EBF emitting plants did not differ in size from wild type plants. Thus, visual cues are not likely to have influenced the outcome of our experiments.
That Myzus persicae was not repelled by EBF emission is contrary to previous findings where EBF released from wild potato (Solanum berthaultii) was shown to deter aphids [22, 23]. This discrepancy might be ascribed to differences between the species in their mode of EBF release. S. berthaultii contains two types of glandular trichomes (A and B), with type A trichomes having a four-lobed head which contains EBF and other volatile compounds, while type B trichomes bear a sticky droplet on their tops [23, 45, 46]. When aphids contact the leaf surface, their tarsi get coated by the sticky exudate of the type B trichomes. While extricating themselves, they destroy the head of type A trichomes  which then release the EBF. In this system, EBF is released in pulses whenever a trichome is ruptured, mimicking the EBF emission by aphids when they get attacked by natural enemies. In contrast, the transgenic A. thaliana lines tested here appear to release EBF continuously over the diurnal period based on our volatile collection data. The difference between continuous vs. pulsed EBF release might be crucial for aphid deterrence.
Once on a plant, aphids may be disturbed by the release of alarm pheromone and interrupt feeding more often than aphids on wild type plants would do [47, 48]. As a consequence aphids on EBF-emitting plants would likely grow and reproduce less. However, in the present study aphid performance on EBF-emitting plants was the same as on wild type plants (Figure 8). This is not due to aphid feeding causing a reduction in EBF release. The emission of EBF was not significantly affected by aphid infestation (Figure 2). The undiminished performance of aphids on EBF-releasing plants is contrary to the findings of Gut et al.  who noticed a reduced aphid growth when Myzus persicae feeding on cabbage plants covered with a plastic bag were exposed to 10 mg of EBF applied on a filter paper placed next to the plant. But, 10 mg EBF is a very high dose, much more than our transgenic A. thaliana plants emitted during their whole lifetime (Figure 3), and 200,000 fold more than an aphid would emit when attacked by an enemy [50, 51]. Such a high dose of EBF may have been toxic to the aphids, and therefore reduced their growth.
EBF emission might also benefit plants by leading to the production of more winged offspring which would leave the plant. EBF-caused wing induction has been demonstrated previously for the pea aphid (Acyrthosiphon pisum) and arises naturally as a consequence of EBF release by aphids during an enemy attack. The ability of enemy attacks to trigger wing induction has been observed frequently for the pea aphid [52–55] as well as the cotton aphid (Aphis gossypii) . The phenomenon is thought to arise as a result of the higher aphid activity on the plant caused by EBF which leads to more frequent encounter rates, similar to what happens when aphid density is high . However, wing formation in this study was not significantly induced in the green peach aphid by EBF emitted from the transgenic A. thaliana plants, even though this aphid species is in fact capable of producing more winged offspring when it perceives EBF (Figure 9), which is not universally true for all aphid species [52, 57] or clones . The fact that the green peach aphid did produce more winged offspring when treated with two EBF pulses a day for three days compared to aphids which were treated just with the control solvent (Figure 9) was not due to different aphid densities, which have been previously shown to influence wing induction [59, 60]. This and the observed predator avoidance behavior of the aphids when EBF was applied demonstrate that the aphids in the experimental colony were still sensitive to EBF. Thus the green peach aphid was able to produce more winged offspring when it perceived pulses of synthetic EBF, but it did not produce more winged offspring if the EBF was being emitted from plants at a constant rate. Taking all the results together, it seems that EBF emitted from transgenic A. thaliana was not effective in causing changes in the physiology and behavior of the green peach aphid that have been shown to be triggered by EBF in other studies. This discrepancy may be due to the fact these transgenic plants release EBF constantly as opposed to the pulsed release caused by natural enemy attacks on individual aphids. There are two non-exclusive possibilities for why aphids do not react to continuous emission of EBF. First they might get habituated to the compound after extended exposure which is also known for other insects responding to pheromones [61, 62]. In fact, Wohlers  determined that pea aphids could become habituated to their alarm pheromone since they did not show typical escape behavior after some time of EBF exposure. Petrescu et al.  reached the same conclusion but only applied EBF once in 24 h, and it is unclear whether EBF remained in the vicinity of the aphids long enough so that they could become habituated. The other possibility is that aphids react to EBF only if it is emitted in pulses, which would mimic the release caused by attack on individual members of an aphid colony. This could explain why the green peach aphid reacted to S. berthaultii where the EBF was only released as individual EBF-containing trichomes were destroyed . The mode of EBF release, whether pulsed or continuous, might therefore be an important cue in informing aphids whether the EBF is coming from attacked conspecifics (so it is necessary to take evasive action) or from a plant (so there is no immediate predation risk).