Title: Author response: Feeding-induced rearrangement of green leaf volatiles reduces moth oviposition
Abstract: Full text Figures and data Side by side Abstract eLife digest Introduction Results Discussion Material and methods Data availability References Decision letter Author response Article and author information Metrics Abstract The ability to decrypt volatile plant signals is essential if herbivorous insects are to optimize their choice of host plants for their offspring. Green leaf volatiles (GLVs) constitute a widespread group of defensive plant volatiles that convey a herbivory-specific message via their isomeric composition: feeding of the tobacco hornworm Manduca sexta converts (Z)-3- to (E)-2-GLVs thereby attracting predatory insects. Here we show that this isomer-coded message is monitored by ovipositing M. sexta females. We detected the isomeric shift in the host plant Datura wrightii and performed functional imaging in the primary olfactory center of M. sexta females with GLV structural isomers. We identified two isomer-specific regions responding to either (Z)-3- or (E)-2-hexenyl acetate. Field experiments demonstrated that ovipositing Manduca moths preferred (Z)-3-perfumed D. wrightii over (E)-2-perfumed plants. These results show that (E)-2-GLVs and/or specific (Z)-3/(E)-2-ratios provide information regarding host plant attack by conspecifics that ovipositing hawkmoths use for host plant selection. https://doi.org/10.7554/eLife.00421.001 eLife digest Plants have developed a variety of strategies to defend themselves against herbivorous animals, particularly insects. In addition to mechanical defences such as thorns and spines, plants also produce compounds known as secondary metabolites that keep insects and other herbivores at bay by acting as repellents or toxins. Some of these metabolites are produced on a continuous basis by plants, whereas others—notably compounds called green-leaf volatiles—are only produced once the plant has been attacked. Green-leaf volatiles—which are also responsible for the smell of freshly cut grass—have been observed to provide plants with both direct protection, by inhibiting or repelling herbivores, and indirect protection, by attracting predators of the herbivores themselves. The hawkmoth Manduca sexta lays its eggs on various plants, including tobacco plants and sacred Datura plants. Once the eggs have hatched into caterpillars, they start eating the leaves of their host plant, and if present in large numbers, these caterpillars can quickly defoliate and destroy it. In an effort to defend itself, the host plant releases green-leaf volatiles to attract various species of Geocoris, and these bugs eat the eggs. One of the green-leaf volatiles released by tobacco plants is known as (Z)-3-hexenal, but enzymes released by M. sexta caterpillars change some of these molecules into (E)-2-hexenal, which has the same chemical formula but a different structure. The resulting changes in the ‘volatile profile’ alerts Geocoris bugs to the presence of M. sexta eggs and caterpillars on the plant. Now Allmann et al. show that adult female M. sexta moths can also detect similar changes in the volatile profile emitted by sacred Datura plants that have been damaged by M. sexta caterpillars. This alerts the moths to the fact that Geocoris bugs are likely to be attacking eggs and caterpillars on the plant, or on their way to the plant, so they lay their eggs on other plants. This reduces competition for resources and also reduces the risk of newly laid eggs being eaten by predators. Allmann et al. also identified the neural mechanism that allows moths to detect changes in the volatile profile of plants—the E- and Z- odours lead to different activation patterns in the moth brain. https://doi.org/10.7554/eLife.00421.002 Introduction Insects rely on olfaction in most aspects of life: volatile signals guide them to food sources, mating partners and oviposition hosts. Especially for herbivorous insects, plant volatiles provide important cues to locate and identify appropriate host plants for their offspring. Upon herbivory, plants respond with an increased release and de novo synthesis of several volatile compounds from their vegetative tissues (Mumm and Dicke, 2010). These so-called herbivore induced plant volatiles can provide significant information to the surrounding environment as composition and abundance reflect several biotic and abiotic factors (Takabayashi et al., 1995; De Moraes et al., 1998; Gouinguené et al., 2001; Schuman et al., 2009; Hare, 2010). Due to the context dependent composition of plant volatile signals, the ability to detect and discriminate volatile compounds is crucial for insects to generate appropriate behavioral responses. In insects and more specifically in the hawkmoth Manduca sexta (Lepidoptera/Sphingidae), olfactory sensory neurons (OSNs) located on the antennae detect odorant molecules (Kalinová et al., 2001; Shields and Hildebrand, 2001; Fraser et al., 2003; Spaethe et al., 2013) and convey this information to the antennal lobe (AL), the first olfactory processing center. The AL of M. sexta females consists of about 70 structural and functional subunits called olfactory glomeruli (Grosse-Wilde et al., 2011). OSNs expressing the same receptor, and thus responding to the same set of odorants, converge onto the same glomerulus in the AL (Gao et al., 2000; Vosshall, 2000) as has been demonstrated for Drosophila melanogaster and indirectly also in several moth species (Hansson, 1997). Spatio-temporal patterns of neuronal activity representing sensory input to the AL can be visualized by optical imaging methods (Hansson et al., 2003; Skiri et al., 2004; Carlsson et al., 2005; Silbering and Galizia, 2007) enabling identification of compound- and blend-specific responses in the AL of M. sexta (Hansson et al., 2003; Bisch-Knaden et al., 2012; Kuebler et al., 2012). Green leaf volatiles (GLVs) constitute a large group of herbivore-induced plant volatiles characterized by a C6-backbone. While emitted only in trace amounts from healthy, undamaged plant tissue, they are emitted instantly after cell disruption (Turlings et al., 1995; D’Auria et al., 2007). GLVs are generated from C18-fatty acids via the enzymes lipoxygenase (LOX) and hydroperoxide lyase (HPL; Allmann et al., 2010). One of the most abundant GLVs, (Z)-3-hexenal, originates from the cleavage of α-linolenic acid through the activity of HPL and it partly rearranges to (E)-2-hexenal. Both alkenals can be further metabolized by an alcohol dehydrogenase (ADH) and alcohol acyltransferase (AAT; D’Auria et al., 2007) to the corresponding alcohols and their esters (Matsui, 2006). GLVs have been assigned various plant defense-associated functions by directly inhibiting phytopathogens (Hamilton-Kemp et al., 1992; Nakamura and Hatanaka, 2002; Prost et al., 2005) and repelling several herbivore species (De Moraes et al., 2001; Kessler and Baldwin, 2001; Vancanneyt et al., 2001; Zhang and Schlyter, 2004). Remarkably, GLVs also function as indirect plant defenses by attracting foraging predators and host-seeking parasitoids to the plant and its attacker (Kessler and Baldwin, 2001; Shiojiri et al., 2006; Halitschke et al., 2008; Schuman et al., 2012) reminiscent of the role of other herbivore induced plant volatiles. Due to their ubiquity and instant release, GLVs are thought to act as nonspecific signals of plant damage (Hatanaka et al., 1987; Hoballah et al., 2002). We recently showed that an enzymatic component of the oral secretions (OS) of M. sexta larvae adds an herbivory-specific feature to the GLV signal. Mechanically damaged leaves of Nicotiana attenuata released large amounts of (Z)-3-GLVs and low amounts of (E)-2-GLVs. However, when the plant was attacked by M. sexta caterpillars or when puncture wounds of plant leaves were treated with M. sexta’s OS, the amount of (E)-2-GLVs released increased, while the amount of (Z)-3-GLVs decreased, resulting in a distinct change in the (Z)-3/(E)-2-ratio of GLV emissions. This herbivore-induced change in the (Z)-3/(E)-2-ratio attracted the generalist hemipteran predator Geocoris spp., which decreased the herbivore load on the plant by feeding on herbivore eggs (Allmann and Baldwin, 2010). Our discovery of a (3Z):(2E)-enal isomerase in the OS of M. sexta larvae raises many questions. Why does Manduca produce an enzyme that generates volatiles which betray the insect to its enemies, and why did evolution not select against this isomerase? The enzyme might be maladaptive and therefore is, or will be, under negative selection. The occurrence of this specific isomerase activity in at least two other lepidopteran species (Allmann and Baldwin, 2010) however, suggests that it may have a beneficial function that outweighs the larva’s net costs of maintaining such an enzyme. It is well known that plants exchange information above ground by releasing volatiles into the air (Baldwin, 2010), which can be perceived by insects as well. Insects can use plant derived volatiles for communication by giving the herbivore induced volatile blend a ‘personal’ note—in our case, by converting (Z)-3-GLVs to their structural isomers and by changing the (Z)-3/(E)-2 ratio. Which message could M. sexta larvae thereby communicate? In this study we hypothesized that the altered GLV emission might serve to reduce the number of competitors on their host plant by informing conspecific ovipositing moths that this plant is already occupied and, possibly, receiving increased predation. Reduced oviposition of Manduca moths in response to feeding damage has been shown in field experiments with M. quinquemaculata (Kessler and Baldwin, 2001) as well as under laboratory conditions with M. sexta (Spaethe et al., 2013). Deviating from the previous study, we chose Datura wrightii (Solanaceae) for our experiments. Datura is a highly preferred host plant of both M. sexta and the congeneric M. quinquemaculata for both nectar feeding (Alarcón et al., 2008; Kessler, 2012) and oviposition (Spaethe et al., 2013). Its distribution covers southwestern USA (Avery, 1959; Munz, 1973) overlapping with the occurrence of both Manduca species. The perennial shrub is repeatedly described to quickly regrow leaves after herbivore damage (Bronstein et al., 2009; Reisenman et al., 2010, 2013). Laboratory experiments failed to find reduced oviposition on damaged D. wrightii (Reisenman et al., 2013; Spaethe et al., 2013) suggesting flexibility in oviposition choice of Manduca females. As the previously examined N. attenuata (Gaquerel et al., 2009), D. wrightii, respond to Manduca herbivore attack by emitting GLVs (Hare and Sun, 2011). While we investigated GLV emission during the day when focusing on the diurnal egg predator Geocoris ssp., Manduca moths oviposit at twilight and night (Madden and Chamberlin, 1945; Lingren et al., 1977). Therefore, we decided to collect volatiles during these times instead. We expected the shift to occur also during the night, as in several plant species GLV emission has been shown to occur also in the dark period (Loughrin et al., 1994; Arimura et al., 2008), and the respective shift in the (Z)-3/(E)-2-ratio is caused by M. sexta oral secretions and not by the plant itself (Allmann and Baldwin, 2010). However, volatile emissions vary with light regime (Halitschke et al., 2000; De Moraes et al., 2001; Gouinguene and Turlings, 2002; Morker and Roberts, 2011), and we therefore chose two nocturnal light conditions differing by moonlight intensity to examine whether light intensity affects GLV emission in D. wrightii. We performed functional imaging in the antennal lobe of female M. sexta moths asking whether (Z)-3- and (E)-2-structural isomers of any of the tested GLVs can be discriminated by the olfactory system. In classical host recognition experiments the Colorado potato beetle Leptinotarsa decemlineata has been shown to recognize and avoid altered ratios of (Z)-3- and (E)-2-GLVs emitted by its host plant Solanum tuberosum (Visser and Avé, 1978). Furthermore, enantioselectivity has been reported for projection neurons in the female AL of M. sexta in response to (+)- and (−)-linalool (Reisenman et al., 2004). Thus, we hypothesized that M. sexta females would be able to differentiate between (Z)-3 and (E)-2-isomers of at least one GLV. If so, ovipositing M. sexta should avoid plants with increased levels of (E)-2-GLVs as they indicate host plants with increased larval feeding competition and predation risk (Allmann and Baldwin, 2010). Here we show by combining field studies with neurophysiological imaging techniques that (i) OS-induced D. wrightii plants have altered (Z)-3/(E)-2-ratios also during the night under both laboratory and field conditions, (ii) Manduca females detect and discriminate the (Z)-3- and (E)-2-isomers and (iii) show ovipositional preference for high (Z)-3/(E)-2-GLV ratios. Results Application of M. sexta OS to leaf wounds triggers pronounced changes in the GLV profile of Datura wrightii To investigate whether application of M. sexta’s OS onto wounded leaves of Datura wrightii plants causes a similar shift in the (Z)-3/(E)-2-ratio as observed in Nicotiana attenuata, we compared the emissions of mechanically wounded D. wrightii plants that were treated with either water as a control (w + w) or with M. sexta’s OS (w + OS) in growth chamber experiments. During the day, application of OS onto wounds caused a significant decrease in the (Z)-3/(E)-2-ratio of the GLVs released from Datura plants compared with control plants (Figure 1A, day). Figure 1 Download asset Open asset Diurnal changes in the emitted (Z)-3/(E)-2-ratios of GLVs in Datura wrightii plants. (Z)-3/(E)-2-ratios of GLVs in Datura wrightii plants represented as box plots. (A) Growth chamber experiment: a single not yet fully developed leaf of each D. wrightii plant was mechanically wounded and treated with water (w + w) or M. sexta OS (w + OS) during three different light conditions to mimic day, sunset, and night. (B) Field experiment: Three single previously undamaged leaves per plant were chosen and randomly assigned to a treatment (control, w + w or w + OS). Values of the control leaf were subtracted from the values of treated leaves. As (Z)-3-hexenal was not detectable in any of the field samples (E)-2-hexenal values are displayed in ng*cm−2*2h−1 (adsorbents used in field collection are not accountable for the absence of (Z)-3-hexenal; Table 6). For visual simplifications (Z)-3/(E)-2-ratios <1 are represented as their negative reciprocal. Values above ‘1’ (red dotted line) thus represent treatment-groups that produced more of the (Z)-3-isomer and values below ‘1’ represent treatment-groups that produced more of the (E)-2-isomer. Asterisks indicate significant differences between treatments (A: Mann–Whitney U test, **p≤0.01, *p≤0.05; n = 5), (B: Wilcoxon signed-rank test, *p<0.05; n = 8). ADH: alcohol dehydrogenase; AAT: alcohol acyl-transferase. The median is represented as a line in each box, box outlines mark the 25% and 75% percentiles; outliers are depicted as circles (if value > 1.5× the interquartile range). For raw data, see F1AB_AllmannSpaethe2012_volatiles.xlsx (Dryad: Allmann et al., 2012). https://doi.org/10.7554/eLife.00421.003 Since Manduca moths are crepuscular and nocturnal insects (Theobald et al., 2010), we repeated the experiment under low light and no-light conditions to mimic sunset and night (Figure 2). The (Z)-3/(E)-2-ratio of the aldehydes differed significantly between treatments also at sunset and night light intensities (Figure 1A, sunset). However, the (Z)-3/(E)-2-ratio of w + w treated plants also decreased with decreasing light intensities, which was mainly caused by increased (E)-2-hexenal emissions (Figure 3 and Table 1). Correspondingly, treatment-dependent differences in (Z)-3/(E)-2-ratios for the alcohol and the hexenyl acetate decreased under lower light conditions and were not found during the night (Figure 1A, sunset, night). Figure 2 Download asset Open asset Light conditions during laboratory volatile collection. Light composition and intensity changed within 24 hr to simulate day, sunset and night condition. Photosynthetically active radiation (PAR, μmol photons*m−2*s−1, orange line) was measured for every light composition and ranged from 0.39 ± 0.01 SE at night to 138.37 ± 0.09 SE at full day conditions. Blue lines denote PAR values measured in the field during the respective volatile collection event (during the night samplings, PAR was below detection limit). For the graph values were logarithmized. Grey areas denote volatile collection events; respective light spectra are shown on the right. For representational reasons time scale starts at 2 am. Flight activity, related to nectar feeding and oviposition (Madden and Chamberlin, 1945; Lingren et al., 1977), is indicated on top of the graph. For raw data, see F2_AllmannSpaethe2012_light.xlsx (Dryad: Allmann et al., 2012). https://doi.org/10.7554/eLife.00421.004 Figure 3 Download asset Open asset Total amounts of GLVs released from Datura wrightii plants at different times of the day in laboratory and field experiments. Mean release of major GLVs from Datura wrightii plants at different times of the day and at different light intensities. Grey and white bars represent (Z)-3- and (E)-2-GLVs, respectively. Single leaves were mechanically damaged and volatiles were trapped for 2 hr immediately after wounds had been treated with either water (w + w) or with M. sexta’s OS (w + OS). (A) GLV emissions of D. wrightii plants under controlled light conditions in a growth chamber. Light conditions are explained in this figure. Quantities are given in nmol/g fresh mass (FM)/2 hr; n = 5. (B) GLV emissions of D. wrightii plants naturally grown in the field. Quantities are given in pmol/cm2/2 hr; n = 8. For an approximate comparison between (A) and (B): 50 cm2 leaf area ≈ 1 g FM. Colored bars mark the emission of aldehydes (light green), alcohols (green) and acetates (dark green). For raw data, see F1AB_AllmannSpaethe2012_volatiles.xlsx (Dryad: Allmann et al., 2012). https://doi.org/10.7554/eLife.00421.005 Table 1 GLV emission of Datura wrightii plants in the growth chamber during the first 2 hr after w + w or w + OS treatment with 100% light (day), 20–10% light (sunset) or 0% light (night) https://doi.org/10.7554/eLife.00421.006 ClassCommon nameRTvolatile release in µg / g leaf fresh massw + ww + OSDayAldehyde(Z)-3-hexenal8.540.64 ± 0.2930.097 ± 0.027(E)-2-hexenal10.490.22 ± 0.1090.7 ± 0.17Alcohol(Z)-3-hexenol14.981.30 ± 0.5111.06 ± 0.275(E)-2-hexenol15.570.058 ± 0.0340.195 ± 0.034Hexenylester(Z)-3-hexenyl acetate13.281.59 ± 0.4421.92 ± 0.244(E)-2-hexenyl acetate13.750.017 ± 0.0040.105 ± 0.018(Z)-3-hexenyl butyrate17.070.028 ± 0.0090.051 ± 0.016(E)-2-hexenyl butyrate17.440.01 ± 0.0020.017 ± 0.004SunsetAldehyde(Z)-3-hexenal8.541.62 ± 0.50.26 ± 0.118(E)-2-hexenal10.491.28 ± 0.7751.69 ± 0.697Alcohol(Z)-3-hexenol14.981.62 ± 0.4330.93 ± 0.308(E)-2-hexenol15.570.45 ± 0.3150.44 ± 0.183Hexenylester(Z)-3-hexenyl acetate13.281.62 ± 0.4311.28 ± 0.511(E)-2-hexenyl acetate13.750.18 ± 0.120.158 ± 0.067(Z)-3-hexenyl butyrate17.070.039 ± 0.0110.031 ± 0.003(E)-2-hexenyl butyrate17.440.013 ± 0.0040.01 ± 0.001NightAldehyde(Z)-3-hexenal8.541.71 ± 0.7320.28 ± 0.118(E)-2-hexenal10.492.43 ± 0.5211.22 ± 0.697Alcohol(Z)-3-hexenol14.981.18 ± 0.350.81 ± 0.308(E)-2-hexenol15.570.79 ± 0.140.37 ± 0.183Hexenylester(Z)-3-hexenyl acetate13.280.63 ± 0.2680.71 ± 0.511(E)-2-hexenyl acetate13.750.093 ± 0.040.083 ± 0.067(Z)-3-hexenyl butyrate17.070.036 ± 0.0020.033 ± 0.003(E)-2-hexenyl butyrate17.440.01 ± 0.0020.014 ± 0.001 Mean (±SEM; n = 5) release of GLVs in D. wrightii plants. A single not yet fully developed leaf of each plant was mechanically wounded and treated with water (w + w) or M. sexta OS (w + OS) during the day (A, 100% light), sunset (B, 20–10% light) and night (C, 0% light). Volatiles are listed by chemical classes and in order of their retention time. To evaluate whether w + w and w + OS treated plants release GLVs in distinguishable ratios under normally variable conditions found in nature, we trapped volatiles during daylight and repeatedly at night from a native D. wrightii population in the Utah desert during the 2011 field season. We performed the experiments on two different days using eight plants for each sampling. Three equally sized leaves of each plant were selected and randomly assigned to one of the treatments (control, w + w or w + OS). Similar to previous experiments with N. attenuata (Allmann and Baldwin, 2010) we were unable to detect (Z)-3-hexenal in any of the samples. During the day the application of OS to the wounds caused a significant increase in (E)-2-hexenal emissions compared with w+w treated leaves (Figure 1B, day, and Table 2, day). As seen from the climate chamber experiment, average (Z)-3/(E)-2-ratio of the hexenyl acetates decreased (Figure 1B, day), but this change was not significant. Table 2 GLV emission of native Datura wrightii plants in the field (2011) during the first 2 hr after w + w or w + OS treatment; during day (1:30–3:30 pm), first or second night (0–2 am) https://doi.org/10.7554/eLife.00421.007 ClassCommon nameRTVolatile release in ng/cm2 leafControlw + ww + OSDayAldehyde(E)-2-hexenal10.870.062 ± 0.0061.02 ± 0.2332.43 ± 0.597Alcohol(Z)-3-hexenol15.380.137 ± 0.0671.21 ± 0.282.07 ± 0.465(E)-2-hexenol15.970.248 ± 0.0350.368 ± 0.0880.57 ± 0.148Hexenylester(Z)-3-hexenyl acetate13.660.26 ± 0.08311.1 ± 1.88112.3 ± 2.067(E)-2-hexenyl acetate14.130.01 ± 0.0020.87 ± 0.3961.34 ± 0.564(Z)-3-hexenyl butyrate17.440.011 ± 0.0020.22 ± 0.1810.19 ± 0.142(E)-2-hexenyl butyrate17.80.004 ± 0.0010.007 ± 0.0020.006 ± 0.001First nightAldehyde(E)-2-hexenal10.870.103 ± 0.01324.6 ± 7.84422.5 ± 5.312Alcohol(Z)-3-hexenol15.380.032 ± 0.0069.6 ± 2.0288.2 ± 3.734(E)-2-hexenol15.970.296 ± 0.0234.12 ± 0.9552.89 ± 0.855Hexenylester(Z)-3-hexenyl acetate13.660.165 ± 0.02810.7 ± 3.62111.53 ± 4.291(E)-2-hexenyl acetate14.130.009 ± 0.0011.14 ± 0.3711.15 ± 0.306(Z)-3-hexenyl butyrate17.440.007 ± 0.0010.022 ± 0.0080.04 ± 0.022(E)-2-hexenyl butyrate17.80.002 ± 00.006 ± 0.0020.007 ± 0.003Second nightAldehyde(E)-2-hexenal10.870.055 ± 0.0094.7 ± 1.8779.5 ± 4.009Alcohol(Z)-3-hexenol15.380.034 ± 0.0184.0 ± 1.2252.94 ± 0.522(E)-2-hexenol15.970.177 ± 0.0210.99 ± 0.4271.47 ± 0.554Hexenylester(Z)-3-hexenyl acetate13.660.089 ± 0.0244.8 ± 2.1149.4 ± 4.708(E)-2-hexenyl acetate14.130.01 ± 0.0020.74 ± 0.5051.77 ± 0.972(Z)-3-hexenyl butyrate17.44bld.0.032 ± 0.0190.039 ± 0.013(E)-2-hexenyl butyrate17.8bld.0.005 ± 0.0030.007 ± 0.002 Mean (±SEM; n = 5) release of GLVs in D. wrightii plants in nature. A single not yet fully developed leaf of each plant was mechanically wounded and treated with water (w + w) or M. sexta OS (w + OS) during the day (A, 1:30–3:30 pm) and during night (B, first night, C, second night, 0–2 am). Volatiles are listed by chemical classes and in order of their retention time; bld.: below the limit of detection. During the first night-experiment (first night, average temperature 17.6 ± 0.7°C, wind speed 1.1 ± 0.8 m/s, waxing crescent lunar illumination with 9% of the moon illuminated), plants of both treatments released very high but similar amounts of (E)-2-hexenal, and the (Z)-3/(E)-2-ratios of the alcohols and hexenyl acetates were low, but did not differ between treatments, resembling the results of the night trapping in the growth chamber (Figure 1B, first night, and Table 2, first night). During the second experiment (second night; average temperature 24.6 ± 0.8°C, wind speed 0.7 ± 0.8 m/s, full moon), approximately 2 weeks later, w + OS-treated plants released significantly higher amounts of (E)-2-hexenal (twofold increase compared with w + w treated plants) and the (Z)-3/(E)-2-ratios of the hexenols and hexenyl acetates were significantly lower compared with mechanically wounded plants that were treated with water only (Figure 1B, second night). (Z)-3- and (E)-2-GLVs evoke different activation patterns in the antennal lobes of Manduca sexta To evaluate if female M. sexta moths are physiologically able to discriminate between (Z)-3- and (E)-2-GLVs and between different (Z)-3/(E)-2-ratios we performed functional calcium imaging in the antennal lobes (AL) of females. Odor-evoked calcium changes in response to exposure to the pure (E)-2- and (Z)-3-isomers of hexenal, hexenol and hexenyl acetate led to activity in discrete regions corresponding to specific glomeruli in the AL of M. sexta females (Figure 4A,B). Aldehyde and alcohol structural isomers activated one single specific region (region of interest 2 [ROI 2], green), with significantly stronger responses to the (E)-2- compared with (Z)-3-isomers (Figure 4B). (Z)-3-hexenyl acetate and its structural isomer activated three different regions in the female AL: a significantly (Z)-3-specific (ROI 3, blue), a significantly (E)-2-specific (ROI 4, pink) and an isomer-unspecific region (ROI 1, grey, Figure 4B). The differences in activation patterns caused by stimulations with (Z)-3- or (E)-2-hexenyl acetate (Figure 4C) strongly suggest that the two odors activated OSNs expressing different sets of odorant receptor types on the female antennae. Of all tested GLVs, hexenyl acetate was the only compound eliciting isomer-specific responses in the AL, therefore we focused on (Z)-3- and (E)-2-hexenyl acetate for all further experiments. Figure 4 Download asset Open asset Calcium activity patterns of the (Z)-3- and (E)-2-isomers in the M. sexta antennal lobe (AL). (A) View onto the AL (marked by outline) of a Manduca sexta female after bath application with the calcium-sensitive dye calcium-green-AM. Stimulations with the six tested GLVs resulted in the activation of four regions in the AL most probably corresponding to single glomeruli (four ROIs, regions of interest). (B) Representative false color-coded images show calcium responses in the AL after odor stimulation. Images are individually scaled to the strongest activation (given by the max value in each image). Time traces show activity of ROI 1, 2, 3 and 4 (n = 10) in response to odor stimulation (2 s; grey bar). Error bars represent standard errors of means. For hexenal and hexenol, stimulations with the (E)-2-isomers activated ROI 2 significantly stronger than did stimulations with the (Z)-3-isomers (Wilcoxon signed-rank test: hexenal: p<0.01, hexenol: p<0.05). ΔF: change in fluorescence; F: background fluorescence. For raw data, see F4B_AllmannSpaethe2012_timetracesGlvs.xlsx (Dryad: Allmann et al., 2012). (C) Comparison of response pattern similarity for repeated stimulations of one structural isomer ((Z)-3 vs (Z)-3 Or (E)-2 vs (E)-2, white boxes) and for both structural isomers ((E)-2 vs (Z)-3, grey boxes); sample size is given above the boxes (Mann–Whitney U test: hexenal: p>0.05; hexenol: p>0.05, hexenyl acetate: p<0.001). For raw data, see F4C_AllmannSpaethe2012_correlationcoefficientsGlvs.xlsx (Dryad: Allmann et al., 2012). https://doi.org/10.7554/eLife.00421.008 As plants do not emit isomerically pure odors but rather mixtures, we studied AL representation of the acetate structural isomers in more detail by stimulating the antenna with blends of (Z)-3- and (E)-2-hexenyl acetate in different ratios (given as Z/E: 100/0, 80/20, 50/50, 20/80, 0/100). In ROI 3 (blue) calcium signals evoked by (Z)-3-hexenyl acetate-containing mixtures were significantly higher compared with stimulations with pure (E)-2-hexenyl acetate, which in turn did not differ from the mineral oil control (Figure 5A,B). For the (E)-2-specific ROI 4 (pink) stimulation with pure (Z)-3-hexenyl acetate led to significantly lower calcium responses when compared with pure (E)-2-hexenyl acetate and the 20/80 ratio, but was not different from stimulation with mineral oil (Figure 5B). Calcium responses of the unspecific ROI 1 (in grey) did not differ between the structural isomers and their mixtures (Figure 5B). Figure 5 Download asset Open asset Female antennal lobe (AL) shows isomer-specific calcium responses to (Z)-3- and (E)-2-hexenyl acetate. (A) Representative false color-coded images show calcium responses in the AL after odor stimulation with isomeric mixtures of a total dose of 250 ng. Images are individually scaled to the strongest activation (given by the max value in each image). Time traces show activity of ROI 1, 3 and 4 (n = 10) in response to odor stimulation (2 s; grey bar). Error bars represent standard error of mean. For raw data, see F5A_AllmannSpaethe2012_timetraceshexenylacetate.xlsx (Dryad: Allmann et al., 2012). (B) Change in fluorescence in ROI 1, 3 and 4 to the pure structural isomers and their mixtures, normalized to the highest activation in every animal. Filled boxes represent responses significantly different from the mineral oil (MO) control; different letters denote significantly different calcium responses (Kruskal–Wallis and Dunn’s multiple comparison test). For raw data, see F5BCE_AllmannSpaethe2012_imaginghexenylacetate.xlsx (Dryad: Allmann et al., 2012). https://doi.org/10.7554/eLife.00421.009 When comparing odor-evoked activation by different (Z)-3/(E)-2-ratios in ROI 3 and 4, stimulations with pure structural isomers as well as the 20% (Z)-3/80% (E)-2 mixture led to significantly different levels of neural activity in these (E)-2/(Z)-3-specific regions (Figure 6A). Activation patterns differed signifi