Unusual macrocyclic lactone sex pheromone of Parcoblatta lata, a primary food source of the endangered red-cockaded woodpecker

Abstract

Wood cockroaches in the genus Parcoblatta, comprising 12 species endemic to North America, are highly abundant in southeastern pine forests and represent an important prey of the endangered red-cockaded woodpecker, Picoides borealis. The broad wood cockroach, Parcoblatta lata, is among the largest and most abundant of the wood cockroaches, constituting >50% of the biomass of the woodpecker's diet. Because reproduction in red-cockaded woodpeckers is affected dramatically by seasonal and spatial changes in arthropod availability, monitoring P. lata populations could serve as a useful index of habitat suitability for woodpecker conservation and forest management efforts. Female P. lata emit a volatile, long-distance sex pheromone, which, once identified and synthesized, could be deployed for monitoring cockroach populations. We describe here the identification, synthesis, and confirmation of the chemical structure of this pheromone as (4Z,11Z)-oxacyclotrideca-4,11-dien-2-one [= (3Z,10Z)-dodecadienolide; herein referred to as “parcoblattalactone”]. This macrocyclic lactone is a previously unidentified natural product and a previously unknown pheromonal structure for cockroaches, highlighting the great chemical diversity that characterizes olfactory communication in cockroaches: Each long-range sex pheromone identified to date from different genera belongs to a different chemical class. Parcoblattalactone was biologically active in electrophysiological assays and attracted not only P. lata but also several other Parcoblatta species in pine forests, underscoring its utility in monitoring several endemic wood cockroach species in red-cockaded woodpecker habitats.

Many animals—especially nocturnal insects—have evolved sexually dimorphic sex pheromones as an efficient and relatively private communication channel for mate attraction and mate choice (1). Sex pheromones also function in species discrimination, and therefore pheromone blends also play prominent roles in premating reproductive isolation of closely related species and in speciation (1, 2).

Most cockroach species are nocturnal, and they profoundly rely on pheromones for intraspecific communication. For example, females of the broad wood cockroach, Parcoblatta lata, a species endemic to pine forests of the southeastern United States, have short wings, are incapable of flight, and therefore have limited dispersal capability. However, by emitting a volatile sex pheromone sexually receptive females are able to recruit males, which are excellent flyers, efficiently (3). Females and nymphs thus are able to remain in the relatively protected habitat beneath the sloughing bark of decaying logs, whereas males incur the greater energetic cost of mate-finding and predation that is associated with greater movement. Our earlier behavioral observations suggested the presence of a female sex pheromone, emitted during a “calling” display, and anatomical and electrophysiological studies confirmed that female-specific tergal pheromone glands produced the pheromone (3).

Identification of the sex pheromone of P. lata has important implications in biological conservation and forest management practices. This species and seven related species in the genus Parcoblatta inhabit standing pines, woody debris, logs, and snags in pine forests of the southeastern United States, and they represent the most abundant arthropod biomass in this habitat (4). Most importantly, P. lata constitutes a significant portion (>50%) of the diet of the endangered red-cockaded woodpecker, Picoides borealis (5–7). The red-cockaded woodpecker is especially sensitive to habitat disturbance, and suitable habitats include old-growth pine with nest cavities, dying and dead pines, and no hardwood understory and overgrowth (8). Moreover, reproduction in the red-cockaded woodpecker is highly dependent on seasonal and spatial changes in the arthropod community on bark, with successful reproduction occurring when arthropod production is high and failure to produce viable clutches occurring in poorer foraging habitats (9). Because P. lata nymphs develop beneath the bark of dead standing or recently fallen trees, this species also has very specific habitat requirements. These requirements may result in spikes in the abundance of wood cockroaches several years after environmental disturbances such as hurricanes or ice storms, which can knock down large numbers of pines. Availability of a synthetic sex pheromone for P. lata and other Parcoblatta species therefore could serve as a useful, low-cost, and ecologically sound tool (i) to monitor cockroach populations, (ii) to assess the suitability of foraging habitats, and (iii) to guide management decisions in Southern forests to maximize habitat for arthropod communities. The sex pheromone could be used to monitor spatial and especially temporal changes in the major food source of the red-cockaded woodpecker, ensuring a robust and healthy overall food web.

Results

Our previous analysis showed that the sex pheromone is produced in the anterior seven tergites of sexually mature virgin females, and biological activity is contained in hexane extracts or collections of headspace volatiles (3). Gas chromatography (GC)-electroantennographic detection (GC-EAD) analyses of hexane extracts of tergites from virgin female P. lata consistently revealed four EAD-active peaks, denoted 1, 2, 3, and 4 in Fig. 1. The electrophysiologically active compounds eluted in a nonpolar fraction (second hexane fraction) when this extract was fractionated on a silica-gel column. This fraction then was chromatographed on a silver nitrate-impregnated silica-gel column, and the resulting fractions were monitored by electroantennography (EAG). Compound 1 was found in the 3% and 4% diethyl ether in hexane-fractions, which were combined and then separated on normal-phase HPLC, yielding compound 1 in a semipure fraction that eluted at 22–23 min. This combined fraction (∼100 μL hexane representing the extract of 1,400 females) was separated in several sequential preparative GC procedures (10), accumulating all collections of compound 1 in the same megabore capillary trap. Finally, the ∼2 μg of compound 1 collected in the capillary trap was transferred directly into an NMR microcapillary tube using benzene D6, as described by Nojima et al. (11). The NMR sample was prepared at North Carolina State University, Raleigh, NC, and then was hand-delivered immediately to the College of Environmental Science and Forestry, State University of New York, Syracuse, NY for NMR analyses. In this study, we did not pursue the purification and chemical identification of GC peaks 2, 3, and 4, which gave relatively weak EAD responses and would require many more females because they were present in minute amounts in the extracts.

Fig. 1.

Characterization of an extract of the sex pheromone glands of Parcoblatta lata virgin females by coupled GC-EAD. FID is flame ionization detector response; EAD electroantennographic response; numbers 1–4 denote electrophysiologically active peaks consistently found in P. lata extracts.

The GC-electron impact (EI)-mass spectrum of compound 1 showed a base peak at m/z 67 and characteristic fragments at m/z (% intensity relative to the base peak) 54 (87), 134 (27), 152 (2.4), and 194 (7.3) (Fig. 2). The chemical ionization spectra of compound 1 with either ammonia or isobutane as the reagent gas showed m/z 195 [M+H]⁺ as the base peak (Fig. S1), indicating a molecular weight of 194, in good agreement with the EI spectrum. The fragmentation pattern did not offer many clues to the structure but retrospectively provided support for the final structure. The vapor-phase IR spectrum revealed a strong carbonyl peak at 1,752 cm⁻¹, consistent with a nonconjugated ester function. The IR spectrum also showed good evidence of both aliphatic and olefinic C—H stretching with multiple peaks around 3,000 cm⁻¹ (Fig. S2). The final structure of compound 1 was elucidated largely from NMR experiments.

Fig. 2.

EI mass spectrum of compound 1, also showing its molecular weight and deduced molecular formula.

The proton spectrum was remarkably clean (Fig. 3), highlighting the importance of minimizing solvent and the advantage of using benzene-D6 to elute compound 1 in preparative GC. If we assume that the olefinic protons are in a 2:1:1 ratio, then there are a total of 18 hydrogen atoms. The 18 hydrogen atoms taken with the probable ester function (i.e., two oxygen atoms) gave a molecular formula of C₁₂H₁₈O₂, with four sites or degrees of unsaturation. The carbonyl accounted for one unsaturated site and two carbon–carbon double bonds (obtained from the correlation spectroscopy (COSY) NMR spectrum; Fig. S3) accounted for two other unsaturated functions. Because neither the IR nor the NMR spectrum indicated other multiple bonds, we assumed a ring, and, because it is an ester, a lactone structure seemed likely.

Fig. 3.

NMR of compound 1 purified from an extract of the sex pheromone glands of P. lata virgin females. (Top) Olefinic region (Left) and homo decoupling experiment with both sets of protons 6 and 10 decoupled simultaneously (Right). (Middle) Natural compound, (Bottom) Synthetic parcoblattalactone.

The ¹H-NMR spectrum conspicuously lacked a methyl group and nonolefinic methine groups (Fig. 3). Also, the COSY spectrum indicated that each carbon–carbon double bond was disubstituted and that each olefinic carbon was bonded to one hydrogen atom (Fig. S3). Taken together, this information led to the conclusion that the compound is a macrocyclic lactone with two isolated double bonds and no branching; in other words, the compound is a 13-member ring ester. The positions and configurations of the double bonds were the only remaining questions.

The position of one double bond (carbons numbered 11 and 12 in Fig. 3) could be determined directly from the COSY spectrum (Fig. S3) and the chemical shift of the methylene doublet at 4.41 ppm. The COSY spectrum showed correlation between the proton on carbon 12 at 5.64 ppm and the methylene group at 4.41 ppm. The chemical shift of this methylene group was consistent with its being “sandwiched” between the ester oxygen and a carbon–carbon double bond. Olefinic proton 11 showed correlation to methylene group 10.

The position of the other double bond was determined analogously. The COSY spectrum showed a correlation between olefinic proton 4 at 5.62 ppm and methylene group 3 at 2.84 ppm (Fig. S3). The chemical shift of methylene group 3 must be between the ester carbonyl and the other double bond. The COSY and total correlation spectroscopy (TOCSY) spectra (Figs. S3–S6) allowed assignment of all of the protons listed in Table S1.

The configuration of double bonds at positions 4 and 11 was determined by simultaneous decoupling of methylene group 6 and methylene group 10. When methylene group 6 was decoupled from proton 5, the only remaining coupling was to proton 4. The coupling constant was 10.5 Hz, which allowed only the Z-configuration. Similar reasoning was applied to the decoupling of methylene group 10 from proton 11, which gave a coupling constant of 10.9 Hz and therefore also a Z-configuration.

The structure of compound 1 was proposed as (4Z,11Z)-oxacyclotrideca-4,11-dien-2-one (= (3Z,10Z)-dodecadienolide), which we named “parcoblattalactone” to account for its origin from Parcoblatta and lactone structure. This compound was synthesized, as detailed in Fig. S7. The synthesis took advantage of the symmetric placement of the double bonds and the oxygen atoms. The key step was the deconjugation of the triple bond from the carboxylic acid. Thus, the diyne (compound 7) was cyclized to the corresponding diyne lactone and semihydrogenated to the desired compound 1 (Fig. S7). The identity of the sex pheromone of P. lata was proved unambiguously: The NMR spectra of compound 1 and the natural compound were identical (Fig. 3).

Synthetic parcoblattalactone was tested in EAG dose–response experiments with antennae of adult male P. lata, Parcoblatta virginica, and Parcoblatta pennsylvanica that were freshly collected in a pine forest in Raleigh, NC. Both P. lata and P. virginica responded to the pheromone at 0.1 ng to ∼1 μg on filter paper in a log dose–response manner, but P. pennsylvanica antennae responded only at very high loadings of parcoblattalactone (Fig. 4). At high doses the EAG amplitude declined, but EAG recovery to baseline was considerably slower (i.e., the EAG peak was broader at high doses).

Fig. 4.

EAG responses of isolated antennae of three Parcoblatta species. The photograph shows P. lata (Upper) and P. pennsylvanica (Lower) nymphs (Left), adult females (Center), and adult males (Right). The small squares on the grid are 1/8 inch (3.17 mm) on a side, and the large squares are 1 inch (25.4 mm) on a side.

In a mixed pine-hardwood forest (Fig. S8), parcoblattalactone was significantly more attractive in overnight trapping using rubber septa than filter-paper dispensers (t test, t = 2.398, n = 7, P = 0.035) (Fig. S9). Adhesive-covered (sticky) traps also were significantly more effective than pitfall traps (t test, t = 3.438, n = 8, P = 0.011) (Fig. S9). In June 2011 we conducted dose–response studies at Lake Johnson Park, Raleigh, NC. Sticky traps baited with 10, 100, and 1,000 ng parcoblattalactone in rubber lures attracted significantly more P. lata males than did control traps that were loaded with hexane solvent only (Fig. 5). Parcoblattalactone also attracted three other Parcoblatta species, most commonly P. virginica and Parcoblatta caudelli. Although P. pennsylvanica nymphs and adults were found in this forest, none were trapped by parcoblattalactone-baited traps. Interestingly, rubber septa baited with headspace volatiles from aerations of virgin P. lata females attracted P. lata males and the other Parcoblatta species in approximately the same ratio as synthetic parcoblattalactone. No immature Parcoblatta were found on any of the traps; one or two adult females were trapped occasionally, but equally in baited and unbaited traps.

Fig. 5.

Field trapping of Parcoblatta species using synthetic parcoblattalactone loaded in rubber septa dispensers.

Discussion

The recovery, conservation, and management of habitats of an endangered species necessitate detailed knowledge of its ecological requirements. The red-cockaded woodpecker, for example, requires open pine woodlands and savannahs with large old pines as cavity trees for nesting and roosting (8) as well as abundant, nonfragmented, contiguous foraging habitat, including mature pines with an open canopy, few or no midstory trees, abundant native bunchgrass and forb groundcovers, and coarse woody debris that supports arthropod availability to bark-foraging birds (4). The assessment of habitat quality and prey availability entails exceptional investment in broad-scale monitoring with nonselective traps, such as burlap bands on trees or cardboard, pitfall and light traps. Because the availability of large prey to feed nestlings significantly increases the reproductive success of red-cockaded woodpecker breeding groups, and P. lata and related wood cockroaches comprise the most common large item fed to nestlings—26–62% of the diet (5)—the newly identified P. lata sex pheromone holds promise as an important, economically and ecologically sound tool to monitor the quality and suitability of foraging habitats for red-cockaded woodpeckers.

Sex pheromones may be separated functionally into two broad classes: long-distance volatile pheromones that attract mates from a distance, and short-range or contact sex pheromones used in courtship, mate assessment, and mate choice (1). Despite their great biological and ecological diversity, and the recognition that several species are major pests with public health and veterinary impacts, the sex pheromones of only a small number of cockroach species have been identified chemically so far (12). In most cockroach species, sexual attractants are emitted by the female—the heavier and shorter-winged sex—attracting the slimmer, long-winged male. The volatile sex pheromones identified to date represent three divergent classes of chemicals. Several highly attractive epoxygermacranoid compounds have been identified from various Periplaneta and Blatta species (family Blattidae), and these compounds, in various blends, also appear to function in reproductive isolation (13–17). An exceptionally attractive α-pyrone (supellapyrone) was identified from the brown-banded cockroach Supella longipalpa (Blattellidae, Pseudophyllodromiinae), and this pheromone appears to function as a single compound and not as part of a pheromone blend (18–20). A substituted quinone (blattellaquinone), from the German cockroach Blattella germanica (Blattellidae, Blattellinae) (21), represents another chemical class. The only volatile sex attractant identified in the family Blaberidae is a mixture of rather simple hydroxyketones, phenols, and thiazolidines (collectively named “seducins”), produced by male lobster cockroaches, Nauphoeta cinerea (22). Thus, cockroaches have undergone adaptive radiations in their sexual communication signals, producing highly diverse and often novel sexual attractants and consequently fashioning relatively “private” channels of olfactory communication. Our understanding of their contact-based chemosensory sexual communication is limited to only a few species, but in this context, too, cockroaches evolved a diverse chemical arsenal of “aphrodisiacs” to elicit courtship and to bias mate choice and acceptance (23, 24).

The P. lata pheromone (4Z,11Z)-oxacyclotrideca-4,11-dien-2-one, a 13-member macrocyclic lactone, represents a newly identified natural product and a new class of cockroach pheromones that is radically different from blattellaquinone, the only other sex pheromone identified in the subfamily Blattellinae. Although the P. lata pheromone shares a lactone structure with supellapyrone, also isolated from the family Blattellidae, the latter includes only a six-member ring structure. Lactones are ubiquitous in microbes, plants, and animals, where they serve diverse biological functions including communication and quorum-sensing (25), as defensive compounds, toxins, and insecticides (26), and even as structural elements of insect nests (27).

However, macrocyclic lactones similar to parcoblattalactone have been found in only a handful of other animals, where they have multiple functions. A closely related macrolide, but with a double bond between carbons 7 and 8 rather than 11 and 12 [(4Z,7Z)-oxacyclotrideca-4,7-dien-2-one = (3Z,6Z)-3,6-dodecadienolide] was isolated as an aggregation pheromone of the sawtoothed grain beetle, Oryzaephilus surinamensis, an important stored-product pest (28). Aerations of the flat grain beetle, Cryptolestes pusillus, also contained this 13-member macrolide, but apparently it was not part of the aggregation pheromone blend; instead, a similar lactone with only one double bond, (4Z)-oxacyclotrideca-4-en-2-one [=(3Z)-3-dodecenolide)], is the major aggregation pheromone component in this species (29). Related 12- and 14-member unsaturated macrolides have been isolated as aggregation pheromones of stored-products beetle pests (28, 29). Interestingly, (3Z)-3-dodecenolide also is thought to serve as a sex pheromone in the emerald ash borer, Agrilus planipennis (30), a devastating invasive beetle that has killed tens of millions of ash trees in the United States. During copulation, male Heliconius butterflies transfer several 12- and 14-member macrolides to the female, including both the 13R- and 13S-configured (9Z,11E)-octadeca-9,11-dien-13-olide, and these lactones apparently serve as antiaphrodisiacs that allow the male to monopolize the female (31, 32). Stritzke et al. (33) found an unusual 13-member sesquiterpene macrolide in the abdominal androconial organs of a danaid butterfly. Saturated 14-member lactones also have been characterized as sex pheromones of the stink bug Piezodorus hybneri (34) and aggregation pheromones in ants (35). Much larger macrocyclic lactones, up to 24-member rings, have been isolated as defensive secretions of Amitermes termites (36) and from the Dufour's glands of halictine (sweat) bees, where they appear to function in recognition of females by males (37). Very large macrolides produced by many bacteria and fungi also have antibiotic and broad antiparasitic spectra (e.g., avermectins) (38). The possibility that parcoblattalactone might have biological characteristics other than its role as a sex pheromone deserves further investigation.

Although we isolated parcoblattalactone from P. lata, it also elicited antennal EAG responses from P. virginica and attracted males of P. virginica, P. caudelli, and a third unidentified Parcoblatta species. Interestingly, the antennae of P. pennsylvanica did not respond to parcoblattalactone, and this species was not attracted to pheromone-baited traps, although it co-occurs in pine-hardwood forest with the other species. Parcoblattalactone is a major component of the P. lata pheromone, but male antennae also responded to three additional components in extracts of P. lata female pheromone glands. These observations suggest that the 12 species in the genus Parcoblatta likely use species-specific multicomponent pheromone blends, as is common for many other insects. Moreover, P. pennsylvanica, and possibly other Parcoblatta species that were not attracted to parcoblattalactone, either have greater fidelity to their multicomponent blend that might include parcoblattalactone or may use a different major component in their pheromone blend. The latter possibility is suggested by the observation that P. pennsylvanica antennae were relatively unresponsive to parcoblattalactone in EAG assays. Unfortunately, there are no phylogenetic studies of this genus, so we cannot speculate on whether the cross-attraction of P. lata, P. virginica, and P. caudelli has a basis in a common evolutionary history. Interestingly, however, P. virginica and P. caudelli also were attracted to volatile emissions of P. lata females that contained the full blend of pheromone components. This unusual observation suggests that other species-isolating mechanisms may operate in this genus, such as temporal and spatial partitioning of sexual activity of different species within the forest. It also is possible that, although we loaded the volatile emissions P. lata females into the rubber septa dispensers, they might have released an “off-blend” of the pheromone blend, thus compromising species specificity.

Parcoblattalactone thus contributes to our fundamental understanding of the evolution of chemical communication in cockroaches. It also provides a practical tool for monitoring populations of insects that comprise an important food source for an endangered bird species, whose protection hinges on habitat conservation through appropriate management. Future research should elucidate the pheromone blends of all Parcoblatta species on which the red-cockaded woodpecker preys. These blends would serve as valuable tools in determining the richness and abundance of prey and, together with other habitat features (e.g., dead trees, boles, snags, downed trees, understory complexity), could be essential in preserving appropriate habitat for this endangered species. By attracting three of the most common Parcoblatta species within the red-cockaded woodpecker habitat, parcoblattalactone represents a significant advance toward this goal.

Materials and Methods

Insects and Pheromone Extraction.

Parcoblatta lata nymphs were collected during the spring in the years 2003–2007 from felled pines and snags in forests in and around Raleigh, Wake County, NC. The nymphs were reared to adulthood and separated by sex. We were able to raise 1,400 female nymphs to the adult stage, and abdominal tergites 1–7 of 6- to 7-d-old virgin females were dissected and immediately extracted in n-hexane for 30 min, followed by a second hexane extraction. The two extracts were combined for further pheromone purification.

Flash Chromatography.

Hexane extracts (∼200 female equivalents) were reduced under a gentle stream of N₂ to ∼200 μL and were loaded onto Pasteur pipette minicolumns filled with 200 mg of chromatographic silica gel (100–200 mesh; Fisher Scientific). The column was eluted successively with two rinses of hexane (2 mL each) and 2%, 5%, and 10% (vol/vol) diethyl-ether in hexane. Each fraction was tested for biological activity with EAG, and the active fraction from different batches (second hexane fraction) was reduced again under N₂ and refractionated on a silica-gel column impregnated with 20% (wt/wt) silver nitrate. This column was eluted with three rinses each of 2 mL hexane, 1%, 3%, 4%, 5%, 10%, and 100% diethyl ether. Each fraction was bioassayed with EAG, and the active fractions (3% and 4% ether) were combined for further purification by HPLC and preparative GC.

HPLC.

Active fractions from silver nitrate silica-gel columns were purified further on a HP1050 HPLC (Hewlett-Packard) equipped with a 1-mL sample loop and two silica-gel columns connected in series (Econosphere Silica, 5 μm, 4.6 mm i.d. × 250 mm; Alltech Associates, Inc.). The solvent system was hexane and diethyl ether programmed as follows: 100% hexane for 3 min, then a linear gradient of 0.5% diethyl ether/min to 10% diethyl ether, then a linear gradient of 20% diethyl ether/min to 50% diethyl ether, and a 5-min hold at 50% diethyl ether. The flow rate was set at 1 mL/min; the effluent was monitored at 210 and 254 nm, and 1-min fractions were collected in glass vials.

GC-EAD.

An HP 5890 Series II gas chromatograph, equipped with either an EC-5 or EC-WAX capillary column (30 m × 0.25 mm, 0.25-μm film thickness; Alltech Associates, Inc.), was modified and used for the GC-EAD analyses (39). Helium was used as the carrier gas at a head pressure of 115 kPa and a flow rate of 1.5 mL/min. Samples were injected in splitless mode, and the purge valve was opened after 1 min. Oven temperature was set at 50 °C for 2 min, then increased at 15 °C/min to 250 °C and held for 10 min. The temperature of the injector, FID, and EAD transfer line was set at 270 °C. The column effluent was combined with N₂ makeup gas (20 mL/min) using a thermal conductivity detector capillary column adapter and then was split 1:1 to the FID and EAD. Charcoal-filtered and humidified air at 250 mL/min flushed the EAD outlet over the antennal preparation within a cold-water–jacketed condenser (∼10 °C).

A custom acrylic antennae holder modified from Nojima et al. (21, 39) was used for EAD recordings. A male cockroach head was mounted in a plastic pipette tip, and the distal ends of both antennae were inserted into a glass capillary. The pipette tip and capillary tube were filled with Eagle's cell culture medium that made contact with gold electrodes. The output signal from the antennae was amplified 10× by a custom high-input impedance DC amplifier and filtered by a high-pass filter with a cutoff frequency of 0.5 Hz (21, 39, 40). The amplifier output was routed through a signal acquisition board within the GC and displayed along with the FID signal in Agilent ChemStation. Adult males, ∼30 d post eclosion, were used for GC-EAD.

Preparative GC.

An HP5890 gas chromatograph equipped with a cool on-column injector was modified as a preparative GC according to Nojima et al. (10, 11). Briefly, a split-splitless injection port assembly was installed adjacent to the FID port and modified as a sample collection port. A 30-m deactivated column (0.53 mm i.d., no stationary phase) (Alltech Associates, Inc.), followed by a separation column, was installed to permit large-volume cool on-column injections (41). An EC-5 megabore capillary column (5.0-μm film thickness, 0.53 mm i.d. × 30 m) (Alltech Associates, Inc.) was used as the separation column and programmed at 30 °C for 2 min, at 10 °C/min to 260 °C, then at 20 °C/min to 280 °C, and held for 30 min. The cool on-column injector temperature was set to track 3 °C higher than the oven temperature. Sections of megabore capillary columns (DB-1, 20 cm long, 0.53 mm i.d., 5.0-μm film thickness) (Agilent Technologies) were used as fraction collection traps (10). Before use, these collection traps were rinsed twice with 100 μL of methylene chloride (HPLC and pesticide residue analysis grades; Fisher Scientific) and dried overnight at room temperature. Traps were not cooled during collections, and the same trap was used repeatedly to collect the same GC fraction.

The 20-cm capillary traps were eluted directly into microcapillary NMR tubes with ∼7 μL of benzene-D6 (99.96% deuteration; Cambridge Isotope Laboratories), as described by Nojima et al. (11).

GC-MS Analyses.

An Agilent 5975 mass selective detector coupled to an Agilent 6890 GC was used for GC-MS analyses of EAD-active compounds. The GC was operated in splitless injection mode and fitted with a DB-5MS column (30 m × 0.25 mm × 0.25 μm; Agilent). The oven was programmed from 50–250 °C at 15 °C/min after an initial delay of 2 min and held at 250 °C for 10 min. Injector temperature was 270 °C; MS quadrupole temperature was 150 °C; MS source temperature was 230 °C; and transfer line temperature was 250 °C.

GC-IR Spectroscopy.

Vapor-phase IR spectroscopy was acquired with a Win GC/IR Pro (Varian) with a GC/IR interface and a Scimitar FTS 2000 linked to a 6890 Network GC system (Agilent). The transfer line and light pipe were operated at 250 °C. Pheromone extract was separated on a HP-5 column (30 m × 0.32 mm × 0.25 μm; Agilent) operated at 100 °C for 1 min, increased to 250 °C at a rate of 20 °C/min, and held at this temperature for 5 min.

NMR Spectroscopy.

A Bruker AVANCE 600-MHz spectrometer equipped with a 1-mm Bruker triple resonance inverse (TXI) microprobe was used for NMR analysis. The signal of a trace amount of undeuterated benzene (7.16 ppm), which is an unavoidable contaminant in NMR samples, was used as a reference. Capillary NMR tubes were obtained from Bruker BioSpin.

EAG.

Adult males were freshly collected in the field, and four antennae of each species were exposed sequentially to the full range of doses. An IDAC-4 amplifier (Syntech) was used to amplify the responses of an isolated antenna to odorants. Parcoblattalactone was loaded in 20 μL hexane onto a strip of filter paper. Then hexane was allowed to evaporate for 2 min, and the filter-paper dispenser was placed into a glass Pasteur pipette. Purified air was delivered over the antenna at 250 mL/min and was diverted through the stimulus cartridge for 0.5 s.

Field Trapping.

Parcoblattalactone was loaded in 100 μL hexane into red rubber septa (Wheaton). The hexane was allowed to evaporate; then 100 μL of clean hexane was loaded into the same septum to carry the pheromone into the septum matrix and was allowed to evaporate. The septa were positioned in the centers of adhesive-coated traps vertically attached by thumb tacks to pine trees at a height of 1.5–2 m (Fig. 5 and Fig. S8). Trapping was conducted overnight in a mixed pine-hardwood forest at Lake Johnson Park, Wake County, NC. Six transects of the five treatments were set with traps 12–15 m apart.

To evaluate the species specificity of their sex-pheromonal signal, lures baited with volatile collections of virgin P. lata females were included also. Virgin females were placed in a large glass jar and aerated with purified air. Volatiles were trapped on SuperQ, which was eluted with hexane, concentrated, and loaded on rubber septa at an approximate dose representing 10 female-day equivalents.