Spray Mechanism of the Most Primitive Bombardier Beetle (Metrius Contractus)

Bombardier beetles comprise two evolutionary branches within the beetle family Carabidae (Fig. 1). One branch, the brachinoid branch, includes the familiar bombardiers of the genus Brachinus, in which the workings of the bombarding mechanism were first elucidated. The second branch, the paussoid branch, includes beetles that also bombard, but that are of cryptic habit and more restricted distribution and, therefore, known well only to specialists. The relationship between the brachinoid and paussoid branches is subject to some dispute. Whereas some investigators, ourselves included, consider bombardiers to represent a monophyletic lineage (Aneshansley et al., 1983; Bousquet, 1986; Crowson, 1981; Deuve, 1988; Eisner et al., 1977; Erwin and Sims, 1984), in other words, that the paussoids and brachinoids are sister groups, and that the ability to bombard evolved only once in these beetles, others take the view that the bombarding mechanism evolved independently in paussoids and brachinoids (Ball and McCleve, 1990; Forsyth, 1972; Moore and Wallbank, 1968; Moore et al., 1987). Either way, there appears to be no argument that the beetles of one genus, Metrius (tribe Metriini), are primitive members of the paussoid branch. Metrius, therefore, could provide some indication of what the bombarding mechanism might have been like in ancestral paussoids. But if bombardiers are monophyletic and brachinoids are derived from paussoids, Metrius could also shed light on bombardier ancestry generally, including that of brachinoids. Our purpose here is to provide some details of the defensive discharge mechanism of Metrius contractus, one of only three species in the single genus of the Metriini.

The most thoroughly investigated bombarding mechanism is that of the brachinines (e.g. Brachinus, Stenaptinus). As has long been known, brachinines respond promptly to assault by ejecting jets of defensive secretion (Westwood, 1839). The fluid is produced by a pair of capacious glands that lie side by side in the abdomen and open at the abdominal tip (Dierckx, 1899). The discharges are accurately aimed (Eisner, 1958; Eisner and Aneshansley, 1999) and they occur with audible detonations, hence the name ‘bombardiers’. The active components of the secretion are 1,4-benzoquinones (Schildknecht, 1957), compounds known to be potently irritating (Thomson, 1971). Both arthropod and vertebrate predators are deterred by the discharges (Eisner, 1958, 1970; Aneshansley et al., 1969; Dean 1980a,b).

The most remarkable property of the brachinine spray is that it is hot (100 °C) when discharged (Aneshansley et al., 1969). This is because the quinones are generated explosively at the moment of ejection by the mixing of two sets of chemicals ordinarily stored separately in the glands. Each gland consists of two confluent compartments. One compartment (the reservoir or storage chamber) contains hydroquinones and hydrogen peroxide. The other compartment (the reaction chamber) contains a mixture of enzymes (catalases, had reason to believe M. contractus to be anomalous because those that we observed discharging in our hands seemed to emit bursts of mist rather than discrete jets. Here, we provide detailed information on the chemistry of the secretion of M. contractus and show that the discharge mechanism of this beetle is indeed unlike that of any other bombardier that has been studied. peroxidases). To activate the spray, the beetle squeezes fluid from the reservoir into the reaction chamber. This results in the instantaneous liberation of oxygen from hydrogen peroxide, and the oxidation of the hydroquinones to quinones by the freed oxygen. The liberated oxygen also provides the propellant, causing the mixture to pop out (Dean et al., 1990; Schildknecht, 1957; Schildknecht and Holubek, 1961; Schildknecht et al., 1968, 1970). The heat that accompanies the formation of the spray is perceptible: brachinines feel hot when they discharge into one’s hand (Aneshansley et al., 1969; Westwood, 1839).

We had previously come into possession of some live M. contractus and had reported on certain characteristics of their discharge mechanism. Thus, we had earlier shown their glands to be two-chambered and their discharged secretion, in common with that of other paussoids, to be quinonoid and hot. We had also shown the reservoirs of the glands to contain the typical precursors of the quinones, hydrogen peroxide and hydroquinones, and, as part of a separate lipoidal phase, a mixture of hydrocarbons (principally pentadecane) (Aneshansley et al., 1983; Eisner et al., 1977, 1989, 1992; Eisner and Aneshansley, 1982; Roach et al., 1979). This indicated that the discharged secretion is a two-phase system, in which the quinones are being formed in the aqueous phase, while at the same time being partitioned (in part at least) into the hydrocarbon phase. It is not known whether the beetles aimed their discharges and ejected them in the form of jets. We

Metrius contractus Escholtz were collected in California and maintained in our Cornell laboratories in small plastic cages on sand or peat moss. They were given chopped mealworms (larvae of Tenebrio molitor) and water (soaked cotton wads). They tended to be gregarious (Fig. 2A) and slow-moving, and were active primarily at night. Some survived for nearly 2 years.

M. contractus lack hindwings and, consequently, are flightless. Their elytra are fused mid-dorsally and form a broadly convex shield covering the back of the abdomen.

For scanning electron microscopy, specimens were sonically cleaned, first in aqueous detergent solution, then in a 1:1 methanol/chloroform mixture. For study of the cuticular components of the glands, beetles were dissected following overnight immersion in dilute aqueous KOH solution.

For experimental purposes, beetles were fitted with a small aluminum hook affixed to their elytra with a droplet of wax. The hook provided a handle by which the beetles could be picked up without being induced to discharge. The hook also provided the means by which a beetle could be attached to the end of an adjustable rod and tethered in any desired orientation.

Tethered beetles were induced to discharge by pinching their individual legs with fine-tipped forceps. The beetles survived the procedures uninjured. Samples of secretion for chemical analysis were obtained by holding pieces of filter paper beside the gland openings when the beetles were stimulated. The papers were promptly transferred to solvent.

We had demonstrated previously that filter paper impregnated with an acidified potassium iodide/starch solution discolors instantly in the presence of 1,4-benzoquinones and can serve as an indicator paper for registering discharges of bombardier beetles (Eisner, 1958). We used this technique to depict the discharges of M. contractus.

Since the secretion ejected by M. contractus is hot, it is possible to photograph the discharges simply by causing the beetles to eject onto a heat-sensing device (thermocouple) that acts as a trigger to an electronic flash unit. We have provided details of this technique elsewhere (Aneshansley et al., 1969) and used it here to photograph the discharges of M. contractus.

A number of M. contractus discharges were filmed using a high-speed motion picture camera (400 frames s⁻¹; Photo-Sonics 16 mm camera, model 1PD).

Five samples of secretion from two beetles were collected on pieces of filter paper. Each sample was extracted with 50 μl of dichloromethane or hexane, and 0.5 μl of each sample was analyzed by gas chromatography–mass spectrometry (GC-MS) using an HP MSD linked to a gas chromatograph equipped with a 25 m×0.22 mm fused-silica column coated with DB-5 (J. & W. Scientific). Gas-phase infrared (GC-IR) spectra were obtained with an HP IRD linked to a gas chromatograph (using a 25 m×0.32 mm fused-silica column coated with DB-5). Dimethyl disulfide (DMDS) and 4-methyl-1,2,4-triazoline-3,5-dione (MTAD) derivatives were prepared as described previously (Attygalle, 1998).

The location of the glands in M. contractus is shown in Fig. 3A. As is typical for bombardiers and evident from Figs 2B, 3B, the reservoir of the glands (R) is substantially larger than the reaction chamber (r) and, unlike the latter, is enveloped by muscles and presumably compressible. The glandular tissue (g) that secretes the reservoir contents forms a series of connected lobes lying apressed against the reservoir wall. This tissue is drained by a long, highly coiled efferent duct (d) that feeds directly into the reservoir. The glands of M. contractus, like the defensive glands of arthropods generally, are integumental in origin and therefore lined throughout with cuticle. Even the efferent ducts and the microscopic tubules draining the secretory cells of the glandular tissue possess such a lining. This is readily demonstrated by treating the glands with dilute aqueous KOH, which dissolves away all tissue components and preserves cuticle only (Fig. 2B and right-hand gland in Fig. 3B). As expected, the lining of the compressible reservoir is thin and flexible, while that of the non-deformable reaction chamber is thick and rigid.

Each gland opens to the outside by way of the slit-like orifice of the reaction chamber.

Pinching the legs of individual M. contractus, tethered and positioned in normal stance on indicator paper, revealed that the beetles can exercise some control over the directionality of their emissions. Thus, stimulation of the hindleg of one side invariably resulted in a discharge of the gland of that side (Fig. 4A,B). The responses were similarly ipsilateral when individual forelegs were pinched. Moreover, stimulation of a foreleg resulted in discoloration of the indicator paper around that leg rather than the hindleg, indicating that the beetles are able somehow to direct their discharges anteriorly. As a rule, we could elicit no more than five consecutive discharges per side from any one M. contractus.

The evenness of the discoloration induced by the beetle’s discharges on the indicator paper was notable. In other bombardiers, where the discharges occur as jets, the impact patterns on paper are characteristically spotty (Eisner, 1958; Eisner and Aneshansley, 1982; Eisner et al., 1989, 1992). M. contractus, it seemed, emitted its secretion diffusely, rather than as a coarse spray.

A number of photographs were obtained depicting the moment of secretory emission in M. contractus. Such pictures (Fig. 4C,D) confirmed that the beetle’s discharges were unilateral when the stimulus was confined to one side, and showed further that the emissions did not occur as jets. Instead, the secretion seemed to ooze from the glands as a bubbling, mist-engendering froth that clung to the beetle’s body. One can easily envisage the atomized fall-out from such froth inducing the evenly discolored markings that M. contractus had made on indicator paper.

The film clarified how secretion is conveyed forwards in anteriorly directed ejections and provided data on the time course of discharges. Two film sequences from the same M. contractus are discussed here (Fig. 5).

In the forwardly directed discharge (upper sequence in Fig. 5), secretion made its appearance externally (4 ms) as a faint yellow streak on the margin of the elytron. As further secretion poured out, it was relayed along the elytral margin to the very base of the elytron (14 ms), where it built up as a bubbling mass (14–34 ms). The bubbling eventually ceased, but not before it had engendered considerable mist (61–184 ms).

In the posteriorly directed discharge (lower sequence in Fig. 5; note that the elytral margin bore a yellow secretory residue from an earlier anteriorly directed discharge), the events proceeded similarly, except that the bubbling mass built up directly over the gland opening. There was again emanation of mist (34–184 ms).

The time courses of the two discharges were similar. Secretory outflow proceeded over the first 34 ms, by which time the bubbling itself appeared to have reached its peak.

It was clear that forwardly emitted discharges were conveyed along the elytral margin. Scanning electron microscopy showed such a margin (Fig. 6A) to take the form of a smooth continuous track (tr) extending the full length of the elytron (henceforth, we refer to the elytral margin as the elytral track). Anteriorly, where the secretion builds up when it is directed forwards, the elytral track is appropriately widened.

Scanning electron microscopy also revealed the structural relationship between the elytral track and the gland opening. This opening is ordinarily hidden from view. What is visible is the outermost portion of the channel (white arrow, Fig. 6B) by which secretion is emitted from the gland (the channel is shown magnified, in stereo view, in Fig. 7A,B).

Whether secretion is conveyed forwards after emission appears to depend on whether the abdominal tip is deflected downwards or pressed upwards against the elytra. When deflected downwards (Fig. 6B), the emission channel is disengaged from the elytral track, thereby ensuring that, if a discharge occurs, the fluid builds up at the site of emergence. When the tip is pressed upwards (Fig. 6C), the channel is lined up with the elytral track, thereby forming a continuous pathway by which secretion can be routed onto the track. We suggest that the former condition prevails when the beetle is responding to attacks from the rear and the latter when it is repelling assaults from the front. Indeed, abdominal deflective motions could be readily induced (and observed with a microscope) in M. contractus held by their tethering hooks and stimulated with forceps. In response to brief pinching of a hindleg, the beetles deflected the abdominal tip downwards, whereas they pressed it upwards in response to stimulation of a foreleg.

We did not succeed in preparing a M. contractus for scanning electron microscopy with its abdominal tip pressed upwards. The photograph in Fig. 6C was generated by taking a print of Fig. 6B, cutting it in half along the pleural elytral margin, and then rejoining the two halves after the abdominal tip had been rotated upwards. Evidence for the actual occurrence of the deflections was also provided by the motion picture film. Close examination of the first frames (0 ms) of the two sequences in Fig. 5 shows that, for the forward discharge (upper sequence), the abdominal tip was pressed upwards, while for the posterior discharge (lower sequence) it was deflected downwards.

Two tongue-in-groove devices ensure that the beetle’s abdominal deflections can proceed with reliable precision. The tongues of these devices are two blade-like structures projecting upwards from the abdominal margins, just in front of the emission channels (stars, Figs 6B, 7C). The corresponding grooves are two slit-like inflections on the inside of the elytra (Fig. 7D; broken line in Fig. 6B).

Twenty-seven components (quinones and hydrocarbons) of the M. contractus secretion were characterized by GC-MS and/or GC-IR techniques, in some instances in combination with preliminary chemical transformations. These results are summarized in Fig. 8 and Table 1. There were no qualitative differences in the composition of the secretions of the two beetles examined. Quantitative differences pertained to minor components only. 1,4-Benzoquinone, pentadecane and (7Z,9Z)-7,9-heptadecadiene were the major components in all five secretion samples examined.

The principal quinone in the secretion, as noted previously (Eisner et al., 1977), was 1,4-benzoquinone, although very small amounts of four other benzoquinones could also be detected on the basis of their mass spectra (Attygalle et al., 1993). The finding that 2-chlorobenzoquinone is the second most important quinone in the secretion (although present only at 1 %) comes as a surprise, since chlorinated secondary metabolites are rare in nature (Gribble, 1998). Although we cannot rule out that chloroquinone was a contaminant, we cannot envisage how contamination at the 1 % level could have occurred.

Mass spectral data served to identify n-pentadecane as the predominant hydrocarbon, along with eight C₁₃–C₁₉ straight-chain and branched saturated hydrocarbons.

The remaining 13 characterized constituents were three unbranched C₁₅–C₁₉ alkenes and 10 alkadienes. Dimethyl disulfide adducts (Francis and Veland, 1981) were helpful in identifying three of these unsaturated hydrocarbons. Eight of the dienes proved to be conjugated. It was therefore possible to determine their structure from the highly characteristic mass spectral fragmentation patterns of their Diels–Alder adducts with 4-methyl-1,2,4-triazoline-3,5-dione (MTAD) (Young et al., 1990). Finally, double-bond configurations could be assigned to those components present in sufficient quantity on the basis of characteristic infrared absorption bands observed by GC-IR analysis (Attygalle et al., 1994, 1995) (see Table 1).

In several fundamental respects M. contractus is typically bombardier-like. Its glands conform to the norm, being constructed of the two basic compartments, the reservoir and the reaction chamber, and its secretion is quinonoid. It ejects its secretion hot (Aneshansley et al., 1983), indicating that it is synthesizing the quinones by chemical explosion, and it discharges its secretion strictly in response to disturbance.

What is unique about M. contractus is its discharge mechanism (Fig. 9). No bombardiers other than M. contractus eject their secretion as an ooze that builds up on their bodies. Nor do any other bombardiers route emitted fluid forwards along tracks extending the length of their elytra. The bubbling undergone by the M. contractus secretion upon discharge may indicate that hydrogen peroxide breakdown and quinone formation, as well as the accompanying heat liberation, are still ongoing after emission of the fluid, raising the question of how M. contractus withstands direct exposure to such material. Notable in this respect is that M. contractus ejects its secretion at the relatively moderate temperature of 55 °C (recorded maximum 77 °C) (Aneshansley et al., 1983), rather than at 100 °C, as do some of the jet-ejecting species (Aneshansley et al., 1969).

Several features of the glandular apparatus bind M. contractus unmistakably to the paussoid lineage. The glands of M. contractus, in common with those of all paussoids, open laterally on the body at some distance from the abdominal tip, rather than on the tip itself as they do in brachinoids. The glands of M. contractus also are typically paussoid in appearance. They are, for instance, virtually indistinguishable from the glands of Mystropomus regularis (Eisner et al., 1992) and closely similar to those of ozaenines (Eisner et al., 1989). They also share some of the features that set paussoids apart from brachinoids, such as, peculiarities of the structure of the reaction chamber, of the morphology of the wall of the efferent ducts and of the microstructure of the intracellular tubules that drain the secretory cells in the glandular tissue associated with the reservoir (T. Eisner, unpublished data).

Despite its anomaly, the discharge mechanism of M. contractus is basically paussoid. Both M. contractus and other paussoids (Eisner and Aneshansley, 1982; Eisner et al., 1992) direct their discharges by deflection of the abdominal tip. They deflect the tip downwards for posterior discharges and press it upwards for forwardly directed emissions. Both types of bombardier engage the elytra for forwardly directed ejections, but their elytra differ structurally, and they put them to use in different ways. In M. contractus, the elytra bear tracks along which the secretion glides forward without physically detaching from the beetle. In other paussoids, the tracks are modified in that they possess a pair of grooved protruding flanges (Fig. 10) that serve to deflect the secretory stream away from the elytral surface and forward as an aerial jet (Eisner and Aneshansley, 1982; Eisner et al., 1992). All paussoids other than M. contractus possess these flanges, including mystropomines, ozaenines and paussines (Eisner and Aneshansley, 1982; Eisner et al., 1992), accounting for why all are able to eject jets of fluid. We consider the flangeless condition of M. contractus to be ancestral in paussoids vis à vis the flanged condition and illustrative, perhaps, of the earliest stage in bombardier beetle evolution. Froth emission, as exemplified by M. contractus, might therefore be indicative of how the bombarding mechanism originated among carabids. The evolution of flanges represented a major advance for paussoids. It enabled the beetles to aim forwards in parallel to their bodies and past their heads, and to deliver the secretory heat more fully upon the target. For M. contractus, retention of the ancestral discharge mechanism evidently was no handicap. For an exceptionally sluggish beetle such as M. contractus, unable to outrun its enemies, being left enshrouded by secretion and its vapors following discharges could clearly be helpful in keeping assailants from closing in again after an attack.

Nothing that we learned about M. contractus suggests that this beetle is also ancestral to brachinoids. Whether brachinoids and paussoids are sister groups or independent lineages will therefore need to be resolved on the basis of other evidence.

Some comments are in order regarding the chemistry of the secretion of M. contractus. First, it should be noted that both the quinones and hydrocarbons in the fluid, rather than only the quinones, can be expected to fulfill a defensive role. The quinones, as widespread components of arthropod defensive secretions, have long been known to be potently repellent to insects and vertebrates alike (Blum, 1981), but the fact that hydrocarbons, which often accompany the quinones in these secretions, might play a similar role, at least vis à vis insects, had mostly escaped notice. Both alkanes and alkenes have been shown to be repellent and irritant in tests with ants and cockroaches (Peschke and Eisner, 1987).

The secretion of M. contractus appears to be specially adapted in other ways as well. M. contractus is the only bombardier that emits secretion as a topical ooze, such as can be expected to confer protracted deterrency. It is also the only bombardier that produces 1,4-benzoquinone as its primary defensive quinone. Interestingly, 1,4-benzoquinone is less volatile, and therefore more persistent after ejection, than 2-methyl-1,4-benzoquinone, the primary quinone produced by other bombardiers (Eisner et al., 1977; Roach et al., 1979).

There appear to be adaptive reasons also for the particular blend in which the hydrocarbons occur in the secretion of M. contractus. Aside from their defensive function, one can assume the hydrocarbons to play a double physical role. First, one would expect them to act as solvents for uptake of the quinones, as these are formed by oxidation of hydroquinones in the aqueous phase of the expelled secretory mixture. For such partitioning purposes, alkenes have been shown to be more effective than alkanes (Peschke and Eisner, 1987). Second, one would expect the hydrocarbons to act as surfactants that facilitate the spread of secretion over the body of the beetle (as in forwardly directed ejections). For such wetting purposes, a combination of alkanes and alkenes has been shown to be potentially more effective than either type of hydrocarbon alone (Peschke and Eisner, 1987). It makes sense, therefore, that the secretion of M. contractus should contain both alkanes and alkenes. It is, of course, conceivable that the hydrocarbons function also as frothing agents.

The secretion of M. contractus is unusual in that it contains conjugated dienes, of which one, (7Z,9Z)-7,9-heptadecadiene, figures prominently. This is striking not only because conjugated dienes are uncommon in nature, but because one would not expect to find dienes in coexistence with 1,4-benzoquinone, a compound usually regarded as a reactive dienophile. Why does 7,9-heptadecadiene not react with the 1,4-benzoquinone to form a Diels–Alder adduct? The answer may have a stereochemical basis. The 7,9-heptadecadiene we characterized has the Z,Z configuration, and this configuration renders it a particularly poor partner in Diels–Alder reactions, for well-understood steric reasons (Attygalle, 1998). (We suggest, therefore, that the other conjugated dienes in the secretion are likely to have Z,Z configurations as well). M. contractus thus appears to make use of the kind of dienes in its secretion that are least likely to have a diminishing effect on the 1,4-benzoquinone upon which it depends for protection.

This study was supported in part by grants AI02908 and GM53830 from the National Institutes of Health. John Doyen and Kipling Will kindly provided the beetles. Frances Fawcett made the drawings in Figs 1 and 9. Janis Strope was most helpful in the preparation of the manuscript. This is paper number 165 in the series ‘Defense Mechanisms of Arthropods’; paper number 164 is C. Rossini, A. González, J. Farmer, J. Meinwald and T. Eisner (2000). Anti-insectan activity of epilachnene, a defensive alkaloid from the pupa of the Mexican bean beetle (Epilachna varivestis). J. Chem. Ecol. 26, 391–397.