Identification of a male-produced sex-aggregation pheromone for a highly invasive cerambycid beetle, Aromia bungii

The longhorned beetle Aromia bungii (Coleoptera: Cerambycidae) is a major pest of stone fruit trees in the genus Prunus, including cherries, apricots, and peaches. Its native range includes China, Korea, Mongolia, and eastern Russia, but it has recently invaded and become established in several countries in Europe, and Japan, and it has been intercepted in shipments coming into North America and Australia. Here, we report the identification of its male-produced aggregation pheromone as the novel compound (E)-2-cis-6,7-epoxynonenal. In field trials in its native range in China, and in recently invaded areas of Japan, the pheromone attracted both sexes of the beetle. Thus, the pheromone should find immediate use in worldwide quarantine surveillance efforts to detect the beetle in incoming shipments. The pheromone will also be a crucial tool in ongoing efforts to eradicate the beetle from regions of the world that it has already invaded.

Current quarantine surveillance methods for the beetle are limited to visual inspections of host plant materials and wooden products, which are of limited efficacy in detecting larvae deep within wood 1 . In the field, infestations can be detected in their later stages by the accumulation of larval frass and exit holes in the tree trunks, and for heavy infestations, visual observation of the large, colourful, and diurnally active adults (Fig. 1). Use of fermenting liquids as trap baits has been reported 9 , but the efficacy of these baits is unclear. Because of its recent invasion of Japan and Europe, its likely invasion of additional countries, and its economic importance, including in its native range, sensitive methods of sampling this species are urgently needed, both for detecting new infestations, and for use in ongoing control and eradication efforts in areas in which it has established.
As part of an ongoing program to elucidate the chemical ecology of cerambycid beetles, we report here the identification of the male-produced sex-aggregation pheromone (sensu Cardé) 10 of A. bungii. The identification was carried out simultaneously and independently by a Japanese team, and a joint Chinese and American team. The synthesized pheromone was field-tested in China in the beetles' native range, and in Japan in a newly-infested region where the beetle is already causing substantial damage. The pheromone was attractive to beetles of both sexes, and so should provide a valuable tool for use in ongoing worldwide efforts to detect, contain, and eradicate this economically important invasive species.

Results
Identification of the male-produced pheromone. Extracts of headspace volatiles collected from male and female A. bungii in Japan were analyzed by coupled gas chromatography-electroantennogram detection (GC-EAD, Fig. 3), which showed that some extracts from males contained a sex-specific compound which elicited strong responses from beetle antennae. The compound had a Kovat's index of 2016 on the polar INNOWax column used in these analyses. Further analyses by coupled gas chromatography-mass spectrometry (GC-MS) determined that the compound had a Kovat's index of 1332 on a lower polarity DB-5 GC column. In addition, extracts of males contained small amounts of (2E,6Z)-nona-2,6-dienal (9.8 ± 7.3% of major compound, n = 3), and extracts from both sexes contained small amounts of benzyl alcohol and benzaldehyde. The major male-produced compound was detected in six of the eight extracts prepared in Japan by rinsing jars in which beetles had been held. The compound also was detected in all 12 of the extracts of male beetle headspace odours prepared in Japan, in amounts estimated at ~250 ng to 10 µg per extract, and in eight of the 23 aeration extracts prepared in China. The compound was not detected in any of the extracts prepared from female beetles in either country.
The electron impact ionization mass spectrum of the major compound detected in extracts from both Japan and China was minimally informative (Fig. 4), but suggested a molecular weight of 154 daltons based on a trace ion visible in some analyses, and a slightly larger ion at m/z 136 corresponding to loss of water from the molecular ion. Hydrogenation over a palladium catalyst resulted in the disappearance of the parent compound, suggesting that there was at least one C = C double bond. However, none of the products from hydrogenation could be identified. In addition, during removal of the catalyst by filtration of the hydrogenation products through a pad of celite filtering aid, eluting with pentane, the reduction products remained adsorbed on the celite, indicating that they were of at least moderate polarity, i.e., that they contained one or more oxygens or other heteroatoms.
The proton NMR spectrum obtained on ~10 µg of material purified by preparative GC was much more informative. Thus, a one-proton doublet at 9.51 ppm with a large coupling constant (J = 7.8 Hz) was assigned to an aldehyde conjugated to a trans-1,2-disubstituted alkene, with the alkene protons appearing at 6.13 (ddt, 1 H, J = 15.6, 7.8, 1.5 ppm) and 6.90 ppm (dt, 1 H, J = 15.6, 6.8 Hz). The chemical shifts and coupling patterns of two single-proton multiplets at 2.92 (apparent td, 1 H, J = 7.3, 4.7 Hz) and 2.88 ppm (apparent td, 1 H, J = 6.5, 4.7 Hz) suggested an epoxide flanked by methylene groups with diastereotopic protons at 1.76 and 1.67 ppm (both m, 1 H) and 1.52 ppm (m, 2 H). The protons at 1.76 and 1.67 ppm were coupled to an allylic methylene at 2.51 ppm (m, 2 H), whereas the methylene at 1.52 ppm was coupled to a methyl group at 1.03 ppm (t, 3 H, J = 7.5 Hz). 1 H-1 H COSY confirmed the connectivity, and suggested that the structure was the monoepoxide  of (2E,6Z)-nona-2,6-dienal. The magnitude of the coupling constant between the two epoxide protons (4.7 Hz) supported the assignment of the epoxide as cis rather than trans. The identification was confirmed by synthesis of the monoepoxides of both (2E,6Z)-nona-2,6-dienal and (2E,6E)-nona-2,6-dienal; the proton NMR spectrum of the former was a close match with that of the insect-produced compound, whereas the epoxide and methyl protons were shifted significantly upfield to 2.67 and 0.96 ppm respectively in the spectrum of (E)-2-trans-6,7-epoxynonenal. Furthermore, the mass spectra of the synthetic compound and the insect-produced compound, and their retention times/retention indices on both the polar INNOWax and nonpolar DB-5 columns, matched exactly.

Field bioassays in China.
A total of 59 male and 69 female A. bungii were caught in the first field trial in China (Fig. 5A). Significantly more A. bungii were caught in traps baited with (E)-2-cis-6,7-epoxynonenal lures than those with isopropanol control lures (P < 0.0001). When males and females were analyzed separately, both sexes were significantly attracted to the pheromone lures ( Fig. 5A; males, P < 0.0001; females, P < 0.0001). There was no significant difference between the sex ratios of beetles caught in traps baited with pheromone (X 2 = 0.36, P = 0.55) or the control (X 2 = 0.043, P = 0.83).
In the second bioassay (Fig. 5B), a total of 47 A.bungii were captured (18 males and 29 females). Pheromone-baited traps again caught significantly more beetles of both sexes combined than either the isopropanol solvent (P < 0.0001) or completely blank control (P = 0.0003) treatments. Examining the sexes separately, traps with pheromone and blank traps captured significantly more males than those with isopropanol (P = 0.0034 and 0.012, respectively), suggesting that isopropanol might be inhibitory to males. Significantly more females were captured by traps with pheromone lures compared to either traps with isopropanol (P < 0.0001) or blank traps (P < 0.0001). Traps with pheromone caught significantly more females than males (X 2 = 9.13, P = 0.0025), while there was no significant sex bias among the beetles caught in either isopropanol or blank traps (X 2 = 2.78, P = 0.096; X 2 = 0, P = 1.000, respectively).
Field bioassays in Japan. A total of 9 males and 27 females were trapped (Fig. 5C). Pheromone baited traps captured significantly more A. bungii (sexes combined) than either the isopropanol solvent (P = 0.0005) or the completely blank control (P = 0.0003). When the sexes were analyzed separately, marginally more male beetles were caught in traps with pheromone than traps with either isopropanol or blank traps ( Fig. 5C; P = 0.056 for both). Significantly more females were caught in traps with pheromone than traps with either isopropanol or blank traps (P = 0.0032 and 0.0012, respectively). Significantly more females than males were caught by pheromone-baited traps (X 2 = 9.32, P = 0.0023). Overall, these results indicated attraction of both sexes of A. bungii to (E)-2-cis-6,7-epoxynonenal, but with a female bias.

Discussion
The analytical and bioassay data presented above provide strong evidence that (E)-2-cis-6,7-epoxynonenal is the major component of the male-produced sex-aggregation pheromone of A. bungii. In GC-EAD analyses, antennae of female beetles responded strongly and specifically only to this compound, even though some extracts contained several other compounds in roughly equal proportions. Similarly, in field bioassays, beetles of both sexes were significantly more attracted to pheromone-baited traps than to control traps in field trials conducted in the beetle's native region, and a country which it has recently invaded. It also must be emphasized that in the field trial in Japan, to minimize further infestation, all visible beetles were being collected by hand daily throughout the trial. Thus, trap catches would likely have been much higher in the absence of these control efforts.
The structure of the pheromone component, (E)-2-cis-6,7-epoxynonenal, is unlike that of any previously reported cerambycid pheromones 11 , and is the first pheromone reported from the tribe Callichromatini. Other Aromia species, such as A. moschata (L.) are commonly known as musk beetles because of the scents which they produce when threatened. The main compounds in the secretions of A. moschata were identified as isomers of rose oxide and iridodial 12 , but it is not known whether these compounds are strictly defensive, or whether they may also have roles as pheromones. Furthermore, these oxygenated monoterpenoid structures are clearly products of a different biosynthetic pathway than the unbranched (E)-2-cis-6,7-epoxynonenal, which could conceivably be produced by oxidative cleavage of an omega-3 polyunsaturated fatty acid such as linolenic acid to give (3Z,6Z)-nona-3,6-dienal, followed by isomerization to the thermodynamically more stable (2E,6Z)-nona-2,6-dienal and epoxidation of the unconjugated double bond.
The effectiveness of the lures and the trapping system can almost certainly be improved by optimizing operational factors such as the dose/release rate of the lures and their effective field lifetimes, and possible additive or synergistic effects of (2E,6Z)-nona-2,6-dienal, the minor component released only by males. It also remains to be determined which enantiomer of (E)-2-cis-6,7-epoxynonenal is produced, and whether the "unnatural" enantiomer causes any inhibition of attraction, although this seems unlikely given the significant attraction to the racemic compound. Nevertheless, because the beetle is currently invading Japan and several countries in Europe, we deemed it important to reveal the identity of the major component of the pheromone immediately, so that it can be incorporated into ongoing efforts to detect, manage, and hopefully eradicate this pest in the countries which it has recently invaded. Various aspects of trap and lure optimization are the subject of ongoing work. The Porapak Q was initially cleaned by Soxhlet extraction with dichloromethane, and traps were rinsed with dichloromethane before each use. Trapped volatiles were eluted from traps with dichloromethane (0.5 ml). In total, 23 aeration extracts were prepared from male beetles, and 7 extracts from females.

Collection of insect-produced volatiles in Japan.
No official permits were required for collection of insects in Japan. Adult A. bungii were hand collected at Soka City (35°49′30.3″N 139°48′27.9″E), Saitama Prefecture, Japan during late June to July, 2015, soon after their emergence from cherry blossom trees, "Somei-yoshino" [Cerasus × yedoensis (Matsum.) A. V. Vassil], which had been netted at the base for collection. Each adult was kept individually in a plastic cup and provided a cotton ball soaked with diluted sucrose water under conditions of 15 °C and 15 h/9 h light/dark (lights on 05:00, off 20:00) lighting conditions. Single beetles were held in 100 ml glass beakers covered with aluminum foil and kept 2 h at 25 °C. After removal of the beetles, the beakers were rinsed with 1 ml of n-hexane. The hexane extracts were kept at −30 °C until analyzed. In total, eight extracts were prepared from males, and two from females. Twelve aeration extracts also were prepared from single males or pairs of males held in 1 L glass flasks, collecting volatiles on Porapak Q adsorbent similar to methods described above.

Analysis of extracts by coupled gas chromatography-electroantennogram detection (GC-EAD).
The extracts of beetle volatiles were analyzed by GC-EAD using a Hewlett-Packard 5890 series II GC fitted with a HP-INNOWax column (30 m × 0.32 mm ID × 0.25 μm film thickness, Agilent Technologies, Santa Clara, CA, USA). The injector and detector temperatures were 250 °C, and injections were made in splitless mode. The oven was programmed from 50 °C for 1 min, then 10 °C per min to 250 °C, using helium carrier gas at a linear velocity of 50 cm/sec. The FID detector temperature was 250 °C. Helium makeup gas (10 ml per min) was added to the column effluent via a stainless steel T-union, after which the flow was split equally between the GC's flame ionization detector (FID) and the EAD with a press-fit Y splitter (Agilent). The GC effluent for EAD was directed to a glass transfer tube (15 mm ID) mounted on the GC and was mixed with humidified air (300 ml per min, 20 °C), and then passed over the antennal preparation. An antenna including the basal segment was gently removed from a live beetle with scissors and forceps, and was mounted on a Syntech EAG probe (Syntech, SciEntific RepoRts | 7: 7330 | DOI:10.1038/s41598-017-07520-1 Kirchzarten, Germany), making connections with electrode gel (Spectra ® 360, Parker Lab. Inc., NJ, USA). EAG and FID signals were fed into a computer through an analogue-to-digital conversion board (IDAC-232, Syntech), and the signals were displayed and analyzed with Syntech GC-EAD software. In total, analyses were replicated with antennae from four females and two males, with each antennal preparation being reused for 2-3 analyses.

Identification of insect-produced compounds. Extracts of insects from China were analyzed on an
Agilent 78020 A GC interfaced to an Agilent 5977E mass selective detector (MSD, Agilent). The GC was fitted with a low polarity DB-5MS column (30 m × 0.25 mm ID × 0.25 μm film; J&W Scientific, Folsom CA, USA), and the oven temperature was programmed from 40 °C for 1 min, then 10 °C per min to 280 °C, hold for 10 min, with helium carrier gas. The injector temperature was 250 °C, and the transfer line 280 °C. Injections were made in splitless mode. Mass spectra were taken in EI mode (70 eV). Extracts of insects from Japan also were analyzed by GC-MS on a polar INNOWax column, using oven temperature conditions as described above. Some compounds were tentatively identified by mass spectral matches with database spectra, with identifications then confirmed by matching retention times and mass spectra of the insect-produced compounds with those of authentic standards, obtained as follows: benzyl alcohol, benzaldehyde, (2E,6E)-nona-2,6-dienal all from Aldrich Chemical (Milwaukee WI, USA), and (2E,6Z)-nona-2,6-dienal (Bedoukian Research, Danbury CT, USA). To calculate Kovat's indices, a set of linear alkanes from C 8 -C 28 was analyzed under the same conditions.
An aliquot of an extract (~50 μl) was hydrogenated to remove all nonaromatic C = C bonds. Thus, the aliquot was diluted with ~100 μl pentane, ~1-2 mg of 5% Pd on carbon was added, and the mixture was stirred under hydrogen atmosphere for 1 h. The mixture was then filtered through a plug of celite, rinsing successively with pentane and dichloromethane, and the filtrates were concentrated and analyzed by GC-MS.
The major male-specific compound was isolated by preparative GC. Thus, the three extracts containing the largest amounts of the main male-specific compound were combined and concentrated to ~10 µl under a gentle stream of nitrogen. The entire sample was then injected onto a Megabore column (DB-5, 25 m × 0.53 mm ID, 5 µm film thickness; J&W Scientific) over ~30 sec, with an injector temperature of 250 °C, and an oven program of 40 °C for 1 min, then 10 °C per min to 250 °C for 10 min. The column effluent was split ~30:1, with the majority going to a heated outlet port (200 °C). A fraction enriched in the major compound was collected in a dry-ice cooled glass capillary, which, after warming to room temp, was rinsed into a conical vial with ~20 µl of deuterated methylene chloride. The resulting solution was transferred to a 1 mm diameter NMR tube for microbore NMR analyses, which were carried out on a Bruker Avance spectrometer at 600 MHz. Spectra were referenced to the signal from residual CDHCl 2 in the deuterated solvent.

Synthesis of monoepoxides of (2E,6Z)-and (2E,6E)-nona-2,6-dienals.
A solution of (2E,6Z)nona-2,6-dienal (27.6 g, 200 mmol; Bedoukian Research) in 1 liter of methylene chloride was cooled to 0 °C, and meta-chloroperbenzoic acid (MCPBA, ~70% purity, 54.4 g, ~220 mmol; Aldrich Chemical) was added in 5 portions over 30 min. The resulting mixture was stirred overnight at 0°, by which time all the starting material had been consumed. The resulting white slurry was filtered cold with a Buchner funnel to remove the bulk of the precipitated meta-chlorobenzoic acid, and the filtrate was concentrated by rotary evaporation. The residue was taken up in 500 ml hexane and filtered again to remove more precipitated excess MCPBA and meta-chlorobenzoic acid, and the filtrate was extracted twice with 2 M aqueous NaOH. The resulting colourless hexane layer was washed with water and brine, dried over anhydrous Na 2 SO 4 , and concentrated under reduced pressure. The residue was purified by Kugelrohr distillation (bp~80 °C at 0.3 mm Hg), producing the monoepoxide as a colorless oil (14.7 g, 48%). 1  (diluted 10-fold before application; DuPont Chemical Co., Wilmington, DE, USA) were hung in a grid orientation with alternating treatments and controls, 1-2 m above the ground, and with traps spaced > 10 m apart. The trap collection cups contained a 1:1 solution of automobile antifreeze and water to kill and preserve trapped insects.
The second bioassay used pheromone and control lures prepared as in the first experiment, but added a treatment consisting of traps with no lures at all (N = 14). Traps were deployed on 29 June 2016, and A. bungii were collected from traps on July 3, 6, 8, 13, 15, 16, and 24 for a total of 7 temporal replicates. Lures were replaced once, on July 15. SciEntific RepoRts | 7: 7330 | DOI:10.1038/s41598-017-07520-1 Field bioassays in Japan. Field bioassays were carried out in a peach orchard (ca. 50 m × 23 m) (34°08′55.5″N 134°27′27.5″E) in Itano county, Tokushima prefecture, Japan. Teflon ® coated black flight-intercept panel traps (Alpha Scents Inc., Portland OR, USA) were hung in a grid orientation with alternating treatments and controls, 1-2 m above the ground, and with traps spaced > 5 m apart. The trap collection cups contained a 1:10 solution of unscented detergent and water to kill trapped insects.
Statistical analyses. The numbers of male, female, and combined male plus female A. bungii collected were modeled using Poisson Regression followed by pairwise contrast tests for assessing differences between treatment pairs. Multiple comparisons were adjusted using Tukey's range test to control the family-wise error rate. Residual analysis and influence diagnostics were applied to detect potential outliers among the data. Sex ratios were assessed using the Chi-square test. All statistical analyses were carried out using SAS 9.4 software.