Defense of pyrethrum flowers: repelling herbivores and recruiting carnivores by producing aphid alarm pheromone

Summary (E)‐β‐Farnesene (EβF) is the predominant constituent of the alarm pheromone of most aphid pest species. Moreover, natural enemies of aphids use EβF to locate their aphid prey. Some plant species emit EβF, potentially as a defense against aphids, but field demonstrations are lacking. Here, we present field and laboratory studies of flower defense showing that ladybird beetles are predominantly attracted to young stage‐2 pyrethrum flowers that emitted the highest and purest levels of EβF. By contrast, aphids were repelled by EβF emitted by S2 pyrethrum flowers. Although peach aphids can adapt to pyrethrum plants in the laboratory, aphids were not recorded in the field. Pyrethrum's (E)‐β‐farnesene synthase (EbFS) gene is strongly expressed in inner cortex tissue surrounding the vascular system of the aphid‐preferred flower receptacle and peduncle, leading to elongated cells filled with EβF. Aphids that probe these tissues during settlement encounter and ingest plant EβF, as evidenced by the release in honeydew. These EβF concentrations in honeydew induce aphid alarm responses, suggesting an extra layer of this defense. Collectively, our data elucidate a defensive mimicry in pyrethrum flowers: the developmentally regulated and tissue‐specific EβF accumulation and emission both prevents attack by aphids and recruits aphid predators as bodyguards.

Comparison of representative total ion chromatograms of the headspace of flowers with and without beetles using detached flowers from the fields. Indicated compounds are those statistically having higher emission rates in flowers with or without beetles respectively (Table   1). (Z)-3-hexen-1-ol (HO), (Z)-3-hexenyl acetate (HA) and (E)-β-farnesene (EβF) were the dominant volatiles in the headspace of the detached flowers. A '+' indicates that this compound has a higher abundance in flowers with/without beetles and the percentage relative to the ion abundance value of flowers without beetles (See Table 1).

Fig. S8 Alignment of the promoter sequences of (E)-β-farnesene synthase genes of T. cinerariifolium and A. annua
Alignments were performed with ClustalW (http://www.genome.jp/tools/clustalw/) and were shaded with BoxShade version 3.21 software. The 1.5kb upstream sequence of (E)-β-farnesene synthase gene (Aa-PEbFS, gi|523967140|) was from A. annua. The sequences with bright yellow background are part of the coding sequence.

Fig. S9 Aphid movement on early stage pyrethrum flowers in the field
Aphids reared on cabbage and pyrethrum plants were individually used to observe their first response behavior by recording their position on a pyrethrum flower every five seconds. Early stage flowers (S1, S1/2, S2) from the same genotype were used to do the assay. In total sixty aphids (twenty aphids for each stage flower) were tested. Y axis represents the movements of aphids on the flower at different recording times (X axis). 0 represents the joint part of the flower head and flower peduncle; 1 represents the upper receptacle part; 2 represents the upper ray floret part; -1 represents the lower flower peduncle.  plants were used in this report. The other field had a 20-fold lower density of beetles which was too low to include for statistical analysis. Monitoring was done each week for 3 consecutive weeks from a stage when flowers were predominantly in stage 2 to finally predominantly stages 4-5.

Methods S2 Headspace collection and volatile analysis of field samples by thermodesorption GC-MS
A metal tube (140 × 4 mm) filled with Tenax adsorbent (200 mg Tenax; 20/35 mesh; Markes Ltd) was connected to the inlet to purify incoming air. A second Tenax cartridge was connected to the outlet to trap volatiles. Headspace volatiles from different samples were collected at an airflow of approximately 100 ml/min using an aquarium pump to push the air through two serially connected cuvettes with two Tenax cartridges in between. Serial connection was done to ensure identical collected volumes between sample pairs because the flow from the pump was known to be not completely regular under those field conditions. The first cuvette contained 10 freshly harvested S2 flowers with 10 cm peduncles derived from 10 independent plants on which a beetle had been observed, and the second cuvette contained 10 S2 flowers from 10 independent plants from the same row without beetles visiting.
For analysis of volatiles released by the peduncle, whole flower, with and without thrips or aphid herbivory, on day 1 (15:00 hrs) 60 S2 flowers of mixed genotypes were harvested from the field. Flower peduncles were all cut to the same length (20 cm) and randomly placed in 12 sets (4 treatments) of five in 100 ml cuvettes containing 80 ml water. From three sets the flower buds were removed leaving the peduncle only. At 10:40 on day 2, from all 12 sets a pretreatment headspace was collected for two hours after first overlaying the water with 20 ml paraffin oil to prevent any volatiles from the cut peduncle to be released via the water. Each cuvette was placed inside a 2.5 l glass container with tubes filled with Tenax connected at the inlets and outlets as described above for thermodesorption GC-MS except that in this laboratory set-up all collections were in parallel at 100 ml per minute. Next, 20 onion thrips (Thrips tabaci) and 10 green peach aphids (M. persicae) per flower were inoculated on two times three sets of flowers and immediately at 13:40 a second 2-hour headspace collection was started for the four treatments

Methods S3 Headspace collection and analysis of intact pyrethrum plants and on-plant analysis of a single leaf upon mechanical damage by GC-MS
The collection vessels were sealed 10-liter glass containers with two connector plugs on the top. Headspace was collected for 24 h (12 h light and 12 h dark). Volatiles were eluted from the liner in 1 ml hexane (HPLC grade) and then methyl laurate (8.7 ng/μl, Sigma-Aldrich, Co., LLC.,

USA) was added as an internal standard and analyzed by GC-MS on a Thermo Trace GC
(Thermo Fisher Scientific, USA) coupled to a DSQ mass-selective detector (Thermo Fisher Scientific, USA). Separation was performed on a HP-5MS column (30 m × 0.25 mm i.d. × 0.25 μm film thickness, Agilent Technologies, USA) with helium as the carrier gas. The detector voltage was 1700 V, and ion spectra were obtained with 70 eV with a scan range from 40 to 700 atomic mass units with a data acquisition rate of 20 Hz. One μl of eluate was injected in splitless mode into the GC at an injection port temperature of 280 °C. The GC oven ramp was 45 °C for 1 min, 10 °C min -1 to 280 °C, and held there for 3.5 min. The transfer line temperature was 220 °C.
A single leaf was enclosed in a 50 ml glass extraction bottle with its petiole sticking through the hole of a soft cap (HM-0020A, http://www.nbhmyq.com/cn/index.php) preventing tissue damage and gas exchange. Bottles were replaced with clean ones at the end of each interval.
Volatiles were collected by inserting a purged clean SPME fiber into the bottle through the lids for 30 min at room temperature. For the mechanical damage treatment, each time half of a fully expanded leaf of a 3-month-old genotype '39' pyrethrum was cut off to create a wounded surface.
Secondary metabolites stored in the different organs at various plant developmental stages were extracted. Sampling was done as follows. For extraction of flower parts from one genotype, stage 2 flowers and 3 × 3 leaves (young, medium, old) were harvested. Pools of 5 flowers derived from 5 different plants were dissected into 4 cm subsequent peduncle sections (upper, middle, lower, starting right under the receptacle), receptacles, ray florets and disc florets. For flower development, 9 genotypes were used to harvest 9 flowers of all flowering stages (S1-5).
Three flowers were used per replication. Young S1 flowers were subdivided into two developmental stages of a small bud (1A) and a bud (1B) immediately prior to opening. Each detached flower was dissected into a peduncle part (12 cm below bud) and a flower part (the bud). All materials were folded into aluminum sheeting and immediately frozen in liquid nitrogen and weighed. A Retsch bullet grinder was used to grind the samples. 100 mg of material was extracted in 2 ml of dichloromethane (DCM) and vortexed for 15 sec, subsequently centrifuged at 3000 rpm for 15 sec and the supernatant was passed through a column filled with sodium sulphate to remove water. Flow through was collected in glass 2 ml vials and used for GC-MS analysis as described before (Ramirez et al., 2012). The identification and relative quantitation of pyrethrins and lactone were performed by comparing with commercial pyrethrin oil (2 mg/ml, Honghe Senju Biological, China) and using a selective mass ion (specific mass 123 is used for pyrethrins and specific mass 43 for lactones). EβF was identified and quantified by comparing with a dilution series of an EβF standard.

Methods S7 Aphid behavior assay in response to early stage pyrethrum flowers
A pyrethrum field with nearly ten hundreds flowering pyrethrum plants (mostly S0-S3 stages flowers) planted in a company yard (Yuxi, Yunnan province, China) was used to do this experiment. Three early-stage (S1, S1/2, S2) pyrethrum flowers were selected for observation of aphid response behavior. Aphids at fourth instar or young adult reared on cabbage plants or pyrethrum flowers were individually used to observe their first-response behavior by continuously recording their position on a pyrethrum flower every five seconds. The total recording lasted for five minutes. For each experiment, aphids reared on cabbage plants or pyrethrum flowers were individually tested. Each aphid was used one time on one flower. In total, sixty cabbage-reared and sixty pyrethrum-reared aphids (twenty aphids for each flower stage) were tested by releasing them individually on the top of a flower head (S1) or the ray florets (S1/2, S2). We defined the joint part Pyrethrum flower volatiles were collected by inserting five S2 flowers from one genotype into the neck of a 500 ml headspace jars without cutting the flower peduncles and sealing it by fitting moistened cotton wool around the peduncle. The jars were sealed 500 ml glass containers with two connector plugs on the top and bottom. EβF or hexane solvent control was introduced by adding 1 μl on a filter paper disc and sealing it into the jars for 2 h. A 50 ml plastic syringe was connected to the jar via a metal tube containing Tenax TA at the inlet to clean the entering air. By pressing the syringe, an amount of 50 ml clean air was blown into the jar and the headspace of the test substances was blown out through the upper Teflon outlet tubing. To check for unforeseen asymmetries, a standard solution of hexane alone and clean air were included as two control treatments. For each test, a fresh aphid settled on a cabbage leaf in a Petri dish with moistened absorbent cotton was used. 10 aphids were repeated for each 500 ml jar. In total 60 aphids reared on pyrethrum and 60 aphids reared on cabbage plants were tested for each odor.

Methods S8 Aphid honeydew collection and volatile analysis
Fresh honeydew droplets were collected by placing the plants horizontally so that the flower peduncles with aphids were hovering just above a large glass Petri dish (Ø = 15 cm). Landing droplets were picked up immediately using a micro-capillary with rubber balloon for suction and pressure and deposited into 50 μl hexane (containing 1.67 ng/μl carvone). A minimum of 40 droplets were collected. For the control, aphids were inoculated on fresh N. benthamiana plants and honeydew was collected in the same way. The volume of the droplets was calculated by assuming a half sphere shape and measuring the average droplet diameter.
One microliter of pooled aphid honeydew extract was injected in splitless mode on an Agilent 7890A GC coupled to an Agilent 5973 mass selective detector. The GC was equipped with an HP-5MS (Agilent Technologies, USA) capillary column (30 m × 0.25 mm i.d. × 0.25 μm). The GC oven temperature program was initiated at 45 °C, held for 2.25 min then raised with 40 °C min -1 to 300 °C with 5 min hold. Other operating conditions were as follows: He carrier gas with a constant flow rate of 1 ml/min; injector temperature, 260 °C. Mass spectra were obtained by electron impact at 70 eV.