Introduction

Approximately 300 million years ago, insects with complete metamorphosis, such as beetles, flies and moths, diverged from direct-developing hemimetabolous ancestors to form the monophyletic Holometabola (Kukalova-Peck 1991). Following this event, the Holometabola radiated so extensively that today there are more insect species than all other animal, plant and fungal species combined (Wheeler 1990). Many hypotheses have arisen to account for the emergence of the Holometabola. In the earliest of these, Berlese proposed that the holometabolous larval stage arose by de-embryonization, whereby the embryo emerges from the egg at an earlier stage of ontogeny with respect to its direct-developing ancestor (Imms 1931).

Molting and metamorphosis are directed by the sesquiterpenoid hormone, juvenile hormone (JH), and the steroid hormone 20-hydroxyecdysone (20E). Peaks of 20E trigger molting, while the presence or absence of JH determines the type of molt. High JH through larval life maintains the larval state and its disappearance at the last larval instar allows the onset of metamorphosis. In the moth, Manduca sexta, where these processes have been studied extensively, JH then reappears at the prepupal peak of 20E, and its presence is required to prevent precocious adult development of some imaginal structures (Kiguchi and Riddiford 1978). JH is also required to maintain the nymphal state in hemimetabolous insects, and its disappearance at the final nymphal instar allows development of wings and genitalia (Wigglesworth 1934, 1936). Because it is required to prevent precocious maturation during postembryonic development, JH has been called the “status quo” hormone.

Although much is known of the role of the metamorphic hormones in postembryonic development, their embryonic function is less clear. In holometabolous insects, two cuticles are typically produced during embryogenesis. A peak of JH appears at about 50% of embryonic development (Bergot et al. 1981), coinciding with formation of the second cuticle, which is that of the first larval stage. JH then persists through the larval stages until it drops at metamorphosis. In hemimetabolous insects, by contrast, three cuticles are produced during embryogenesis (Sbrenna 1990). In Orthoptera, these three cuticles coincide with three peaks of ecdysteroids (Lagueux et al. 1979). JH appears at about 70% of embryonic development and is present during the formation of the last cuticle, the first instar nymphal cuticle (Temin et al. 1986). Treatment of Schistocerca gregaria embryos with JH mimics prior to this endogenous peak causes both morphological and growth defects (Novak 1969), and nymphal features on the second embryonic (pronymphal) cuticle (Sbrenna-Micciarelli 1977).

Recently, Truman and Riddiford (1999, 2002) have proposed an endocrine-based explanation for Berlese’s hypothesis in which the progress of embryonic development is altered by an advancement of JH production into earlier stages of embryonic development. In this scenario, such an advance would induce precocious tissue differentiation but inhibit nymphal patterning. To better understand the developmental consequences of an advance in JH production, and extend the results of previous studies, we have simulated the holometabolous endocrine profile in embryos of the direct-developing cricket, Acheta domesticus, by treating with JH III and a JH mimic at early or mid-embryogenesis. We provide evidence that JH treatment redirects pronymphal cuticle formation to a nymphal fate and affects cuticle composition in a stage-dependent way. Interestingly, the loss of the pronymphal cuticle has little effect upon the viability of the first instar nymph. In addition, we provide evidence that suggests that the morphogenetic effects of JH may be confined to embryonic molting periods when ecdysteroids are present.

Materials and methods

Cricket husbandry and egg collections

Adult crickets were obtained from Fluker’s Cricket Farm (Port Allen, La.). Eggs were collected from dishes of damp sand that had been left with adult crickets for 24 h and incubated on wet paper towels in 100×15 mm petri dishes at 25°C. Under our conditions, the embryos hatch 18 days later. Therefore, to match the staging system described by Lauga (1969) where embryos develop in 16 days, we multiplied the day of development in our experiments by 16/18.

JH treatment

A stock solution of pyriproxifen (a JH mimic, JHM; Sumitomo; a gift of Dr. H.R. Oouchi of Sumitomo Chemical Company, Osaka, Japan) was kept in cyclohexane (Aldrich, Milwaukee, Wis.). Stock solutions of JH III [natural all-trans enantiomer of JH III prepared from the Malaysian sedge Cyperus iria (Toong et al. 1988); a gift of Dr. Toong] were prepared in cyclohexane and the concentration was checked spectrophotometrically at 217 nm. The stock solution was diluted in acetone (Aldrich, Milwaukee, Wis.) and applied in 0.2-μl doses with a 10-μl Hamilton syringe to individual embryos immobilized on double-stick tape. We found that embryos under 4.5 days old had to be cultured in wet conditions, and in these cases the embryos were treated on glass slides and then moved to culture directly on damp paper towels in petri dishes.

Morphometrics

Images of cricket limbs were taken at ×20 with a Nikon microscope and Apple Video software. Segment lengths were measured manually on printed images using multiple straight lines parallel to the axis of the segment.

Cuticle analysis

Ecdysed cuticles were mounted in Fluoro-mount (Southern Biotechnology Associates, Birmingham, Ala.). First instar nymphal cuticles were first treated overnight in a concentrated NaOH solution to remove tissues.

Immunocytochemistry

The rat anti-BAB II antiserum was a generous gift from Frank Laski (University of California, Los Angeles). We used it at 1:1,000 in 1- to 2-day incubations with embryos that had been fixed for 15–30 min in 4% formaldehyde (Fisher, Pittsburgh, Pa.) in phosphate-buffered saline, washed several times in PBS with 1.0% Triton X-100 (Sigma, St. Louis, Mo.), and blocked for 30 min in 5% normal donkey serum. For secondary labeling, a donkey anti-rat FITC conjugate was used at 1:500 in overnight incubations (Jackson Immuno Research Laboratories, West Grove, Pa.). The embryos were dehydrated in a series of increasing concentrations of ethanol, cleared through xylene and mounted in DPX. The immunostaining was examined using a Biorad Radiance 2000 MP confocal microscope.

The rabbit anti-Distal-less antibody was a generous gift of Grace Panganiban (University of Wisconsin). We used it at 1:1,000 with tissue that had been fixed for 1 h at room temperature or overnight at 4°C. To localize the Distal-less antibody we used a donkey anti-rabbit secondary conjugated with horseradish peroxidase, also at 1:1,000 (Jackson Immuno Research, West Grove, Pa.).

Results

Embryonic development and cuticle formation in Orthoptera

After formation of the germ band, embryonic development in Orthoptera is divided into three stages separated by molts that occur within the egg: the E1, pronymphal and nymphal stages (Fig. 1). The E1 stage begins at stage 15, about 4 days after fertilization, with the appearance of the first embryonic cuticle (E1). A few days after E1 cuticle formation the embryo undergoes the movements of katatrepsis, a process by which the embryo, which is initially facing the posterior pole of the egg, inverts around the posterior pole to face the anterior end (Fig. 1). In both crickets and grasshoppers, the E1 cuticle detaches from the epidermis around the time of katatrepsis (Edwards and Chen 1979; Lagueux et al. 1979). However, in crickets we last see an obvious E1 cuticle as late as stage 27. Katatrepsis is followed by dorsal closure, during which the embryo first grows around the yolk dorsally and then extends anteriorly to finally envelop the yolk (Fig. 1). Once dorsal closure is completed, the embryonic epidermis secretes the pronymphal cuticle at stage 24–25. This is followed several days later by secretion of the first nymphal cuticle at stage 28–29 (Edwards and Chen 1979). The nymph hatches while still enclosed by the pronymphal cuticle, which facilitates the hatchling’s movement through soil as it digs to the surface. Upon reaching the surface, the pronymphal cuticle is quickly shed (Bernays 1971).

Fig. 1
figure 1

Major developmental events during embryonic development of the cricket. Arrows indicate egg deposition and hatching/nymphal ecdysis, and the appearance of the embryonic cuticles. The persistence of the cuticles is indicated by gray lines. Days and stages are taken from Lauga (1969). The times of deposition and duration of the embryonic cuticles are adapted from Edwards and Chen (1979). The inset shows a developmental series from katatrepsis to dorsal closure, with anterior to the top. The red bar represents the expected time of juvenile hormone appearance in cricket embryos based on data from grasshoppers (Temin et al. 1986)

Full size image

Since detailed titers of the metamorphic hormones during embryonic development of crickets are not available, we have inferred these from detailed work on embryos of the desert locust, Locusta migratoria. During embryogenesis there are four peaks of ecdysone and its metabolites that correlate with deposition of the three embryonic cuticles as well as the serosal “cuticle”, a thin layer secreted by the serosal cells that underlie the chorion (Lagueux et al. 1979). A peak of JH III appears at about 70% of embryogenesis, coinciding with terminal differentiation and persisting through the production of the first nymphal cuticle (Temin 1986).

The effect of JHM on embryonic cuticle progression

The shell of cricket eggs is transparent, and distinct features of the pronymphal and nymphal cuticles can be seen as they appear in development. The pronymphal cuticle is distinguished by pigmented labral teeth. The nymphal cuticle, by contrast, has sensory bristles, sclerotized mandibles (Fig. 2), and becomes pigmented just prior to hatching. We compared the appearance of these characters after treatment of embryos at stage 17 (about day 6) with 200 ng JH mimic (JHM) pyriproxifen in acetone to that seen in controls treated with acetone only (Fig. 3). Pyriproxifen is not readily metabolized by insects and is therefore a persistent source of JH. In acetone-treated control embryos, labral teeth appeared at 10.5 days, with the suite of nymphal characters appearing 4–5 days later. Labral teeth were never observed in embryos treated with JHM. Instead, they formed sclerotized mandibles by day 10.5, followed by cercal sensillae on day 12, and pigmented cuticle and body wall bristles on day 13. The accumulation of red pigment in the eye, a character that is not associated with cuticle formation, was not affected by JHM treatment.

Fig. 2
figure 2

Ventro-lateral views showing a pronymph (left) at stage 27, and nymph (right) at stage 30. The pronymphal cuticle has “teeth” on the labrum (arrow). An arrowhead indicates the sclerotized mandibles of the nymph (right), and open arrows point to sensory bristles

Full size image
Fig. 3
figure 3

A comparison of the effects of juvenile hormone mimic (JHM) on the appearance of cuticular features during development. Embryos were treated with 200 ng JHM or 0.2 μl acetone on day 6. The symbols indicate the point in development where 50% of the group possess the particular character

Full size image

In addition to changing the progression of cuticle features, we found that JHM treatment affected growth of the embryo. Figure 4a shows a JHM-induced precocious nymph that was treated at katatrepsis alongside a normal nymph at full egg-length. The precocious nymph was compressed along the anterior-posterior axis of the body and proximal-distal axis of the limbs. In addition, the embryo was open anteriorly, as it failed to complete dorsal closure.

Fig. 4
figure 4

a JHM affects growth while it induces precocious nymphal features. A control nymph and a JHM-induced precocious nymph; anterior is to the left. Arrows indicate the induced mandibles and cercal bristles. The induced nymph is approximately two-thirds the length of the untreated nymph, and is open anteriorly, since it failed to complete dorsal closure (white arrowhead). b An embryo treated prior to E1 cuticle formation and dissected from the egg at the end of embryonic development. The axis of the embryo is U-shaped, with the abdomen directly behind the head, and the thoracic segments at the right of the image. An arrow indicates the sclerotized mandibles (Ab abdomen, T1 first thoracic segment, T3 third thoracic segment)

Full size image

Effects of JHM treatment throughout development

When we treated embryos at katatrepsis, which just precedes the molt to the pronymphal stage, we observed the appearance of precocious nymphal characters. To determine what effect JHM treatment would have at the E1 molt, we treated embryos on consecutive days, framing this event. Embryos treated at the earliest times failed to complete katatrepsis (Fig. 5), and remained with the head, head and abdomen, or abdomen only at the posterior of the egg, so that the axis of the embryo was U-shaped (Fig. 4b). These embryos continued to develop beyond this stage and precociously expressed nymphal characters such as pigmented mandibles and cercal bristles (Fig. 4b). The blockage at katatrepsis was seen in embryos that were treated up to day 4.8 (Fig. 5). Embryos treated after day 4.8 completed katatrepsis but did not complete dorsal closure (Fig. 5), leaving one-third to one-half of the yolk unengulfed (Fig. 4a). These embryos also showed precocious nymphal characters.

Fig. 5
figure 5

Embryos treated with JHM on consecutive days of development show two alternative final phenotypes that are dependent on the day of treatment. Filled ellipses indicate the percentage of embryos that had completed katatrepsis after treatment with 50 ng JHM. Open ellipses indicate the percentage of embryos that had completed dorsal closure after 10 ng JHM. We found that larger doses were necessary when treating early embryos since these had to be cultured wet. Between 5 and 25 embryos were scored for each point. Arrows show the normal timing of katatrepsis and dorsal closure (KAT katatrepsis, DC dorsal closure)

Full size image

To identify the window of JH action in inducing the latter phenotype, we separated embryos according to developmental stage, treated each egg with 10 ng JHM and scored embryos for percent egg length and precocious nymphal development (Fig. 6). Embryos treated at stages 17–22 (approximately days 6–8.5) showed similar responses to JHM. All showed the precocious expression of the suite of nymphal characters (i.e. pigmented mandibles, cercal hairs and body bristles). In addition, they failed to complete dorsal closure, occupying only about two-thirds of the length of the egg. In contrast, all embryos treated at stage 23 (about day 10) grew to occupy the full length of the egg and made a normal pronymphal cuticle.

Fig. 6
figure 6

The effect of JHM applied during the E1 stage on growth and cuticle progression. Black circles indicate the mean percent egg length, indicated by the percentage of the egg occupied by the embryo at the end of development (100% is normal). Open circles indicate the percentage of embryos with the normal progression of cuticle formation (first pronymphal, then nymphal). The number of embryos used for each point is indicated just above the x-axis

Full size image

Effect of JH III on cuticle progression

Because the initial dose of JHM we used was very high, we wanted to know whether similar effects could be obtained using physiological doses of the naturally occurring JH, JH III. To determine this, we treated embryos with different doses of JH III at katatrepsis, and monitored the appearance of the cuticle markers.

At higher doses, 20–500 ng, the effect produced by JH III resembled those seen with JHM and the embryos produced precocious nymphal cuticle characters without producing labral teeth. Growth was also affected at the highest doses; the mean percent egg length occupied by embryos was 74±4% after 200 ng JH III, and 66±1% after 500 ng JH III.

At lower doses of JH III, however, a variable response was seen. Figure 7a shows the dose-response relationship for labral teeth, sclerotized mandibles, and cercal sensory bristles. A comparison of ED50s reveals that suppression of labral teeth and induction of cercal bristles occurred at about 4 and 4.5 ng JH III, respectively, while sclerotized mandibles were induced at slightly lower doses of about 2.5 ng.

Fig. 7a, b
figure 7

Dose-response and stage-response curves for the effect of JH III on selected cuticle characters. a N =24, 49, 41, 31, 49, 37 and 16 were treated with 0.1, 0.5, 1.0, 2.0, 6.0, 10.0, 20.0 and 60.0 ng JH III, respectively, and the characters of the next cuticle produced were recorded. The ED50s of each character overlap: labral teeth, which are characteristic of the pronymphal cuticle, are inhibited at about 4 ng/embryo, cercal bristles are induced at 4.5 ng/embryo, and mandibles are induced at about 2.5 ng/embryo. b The response of different cuticle characters to 2 ng JH III at different points in development. Embryos of the same egg batch were separated according to stage and treated with 2 ng JH III, and the ensuing cuticle was scored for the presence of labral teeth, sclerotized mandibles and cercal sensillae. N =6, 18, 14, 16, and 18 for stages 18, 19, 20, 21 and 22, respectively. The data points for stage 23 (open squares) are taken from data where 10 ng JHM was used, N =13

Full size image

To determine if the variation in cuticle transformation we observed after treatment at katatrepsis was due to differences in JH III concentration, or differential degradation prior to some JH-sensitive critical period, we treated staged embryos with a low dose of JH III (2 ng), and recorded the character of the cuticle produced at the pronymphal molt. Unlike pyriproxifen, JH III is actively degraded in cricket embryos by endogenous esterases, and this activity lasts through katatrepsis (Roe et al. 1987). Therefore, treatment with JH III at different times prior to the pronymphal molt should reveal the temporal windows of JH action. We found that the effect of 2 ng JH III increased when the time of treatment approached the onset of the pronymphal molt (Fig. 7b). When given at stage 21, 2 ng JH III suppressed the formation of labral teeth and induced precocious differentiation of cercal bristles. When given only a few hours earlier (stage 20), this dosage had no effect upon these characters. Yet at this time, 2 ng JH III induced the precocious sclerotization of the mandibles, indicating an earlier sensitivity of this tissue. Moreover, as seen with the JHM treatment, the embryos lost sensitivity to treatment with JH by stage 23, midway through dorsal closure.

Interestingly, we found that with doses between 0.6 and 20 ng JH III applied at katatrepsis (stage 18–19), the embryo produced second embryonic cuticles that had both pronymphal and nymphal characteristics. Figure 8 shows the relationship between the concentration of JH III and the composition of the second embryonic cuticle in the head. Labral teeth alone predominated after treatment with between 0.1 and 1 ng JH III (Fig. 8a). At 2 and 6 ng JH III, the embryo produced a mosaic cuticle bearing both labral teeth and sclerotized mandibles, and above 20 ng, only mandibles were found (Fig. 8c).

Fig. 8a–d
figure 8

Composite cuticles appear after treatment with intermediate doses of JH III. a A summary of the cuticle composition in the head after different doses of JH III. The width of the bars represents the percentage of embryos in each class. The same group is scored as in Fig. 7 (LT labral teeth, M mandibles). b–d Frontal view of the labrum/mandible area of second embryonic cuticles. b An ecdysed pronymphal cuticle from an untreated cricket. Arrows indicate the labral teeth. c A pronymphal/nymphal composite showing both labral teeth and mandibles together. An arrow indicates the labral teeth; an asterisk marks the sclerotized mandibles. d Ecdysed cuticle from an embryo treated with 10 ng JH III that has sclerotized mandibles (asterisk), and bristles on the labrum but no labral teeth (la, labrum)

Full size image

Juvenile hormone mimic-treated precocious nymphs die around the time of hatching. Consequently, we were surprised when many of the JH III-treated embryos that produced mosaic or precocious nymphal cuticles subsequently underwent their third embryonic molt, hatched and ecdysed from the affected second embryonic cuticle. The ecdysis that followed hatching was frequently complete, and we have raised the subsequent nymphs to adulthood. Some of the ecdysing nymphs were stuck with their appendages still within the transformed second embryonic cuticle. Embryos with mosaic second cuticles (with labral teeth and mandibles) were slightly more successful at hatching and ecdysis than those with a more complete nymphal transformation. Of 25 pronymphal-nymphal mosaics induced after treatment with between 0.6–10 ng JH III, over one-half hatched and at least partially ecdysed, while about one-third hatched without ecdysing. In contrast, of the 34 with precocious nymphal cuticles, over one-third hatched without ecdysing while only one-quarter were able to ecdyse. Even at doses of JH III as high as 100 ng where JH III inhibits growth, embryos that completed dorsal closure were still able to hatch and ecdyse (data not shown).

Analysis of the ecdysed second cuticles shows an array of pronymphal to nymphal characteristics. At lower doses of JH III, such as 0.6 and 1 ng, cuticles with both labral teeth and sclerotized mandibles had scale-like denticles typical of the pronymphal cuticle (Fig. 9a). However, at 6 ng and higher, the nymphal transformation was more complete; these cuticles bore bristles on the cercus, legs and tergites, and the cuticle was covered with non-sensory microtricheae, which are characteristic of the first nymphal cuticle (Edwards and Chen 1979). In addition, intermediate bristles were frequently observed. Figure 9 compares the cercal cuticle from embryos treated with various dosages of JH III. The pronymphal cuticle is covered with scale-like denticle rows (Fig. 9a), whereas the nymphal cercus has filiform hairs in sockets and microtricheae (Fig. 9b). After treatment with 2 ng JH III, the denticles typical of the pronymphal cuticle persisted, but tanned stubs resembling partial bristles appeared without a corresponding socket (Fig. 9c). After treatment with 10 ng JH III, the cercal bristle size and morphology were almost indistinguishable (Fig. 9d) from that seen on the first nymphal cuticle (Fig. 9b).

Fig. 9a–d
figure 9

The degree of pronymphal-nymphal transformation of the cercal cuticle is dependent on the dose of JH III. a An untreated ecdysed pronymphal cuticle bearing the posterior-pointing denticles typical of this cuticle. b An untreated first nymphal cuticle bearing complete sensory bristles and socket. The non-sensory microtricheae between the bristles are characteristic of the first nymphal cuticle. c An ecdysed pronymphal cercal cuticle partially transformed towards a nymphal fate after treatment with 2 ng JH III at katatrepsis. Arrows indicate the tanned stubs that suggest partially differentiated bristles. d After 10 ng JH III at katatrepsis, the pronymphal cuticle is completely transformed to a nymphal fate, with near-perfect sensory bristles and sockets. In addition, the pronymphal scales are replaced by non-sensory microtricheae

Full size image

Effect of JHM on growth along the leg proximo-distal axis

To better understand how JH affects growth, we focused on the leg proximal-distal axis. Embryos were treated with 200 ng JHM at katatrepsis (stage 18–19), then fixed after 12, 24, 36 or 48 h, and the length of the third thoracic leg was measured from joint to joint. Figure 10a summarizes the mean segment lengths over time. Because the joint that separates the coxa and trochanter was difficult to identify, we measured the coxa and trochanter together.

Fig. 10
figure 10

a Effects of 200 ng juvenile hormone mimic (JHM) on growth of the T3 limb. JH affects distal segments more than proximal ones. Between 5 and 18 segments were measured in limbs fixed after treatment with 200 ng JHM or acetone after 12, 24, 36 and 48 h. The mean segment lengths of acetone-treated and JHM-treated groups are summarized here. b Growth of each segment is determined by the difference between the 12- and 48-h segments divided by their dimensions at 12 h, so that 100% is a doubling of the 12-h length. c Differential inhibition by JHM calculated by dividing the difference between the change in the control segment and the change in the JHM-treated segment by the change in the control segment

Full size image

Acetone-treated control limbs began to grow rapidly between 24 and 36 h after treatment (Fig. 10a). By 48 h, the acetone-treated limb had increased in length by 47%. In contrast, the JHM-treated limb grew significantly less in the proximal-distal direction, although the difference between the two treatment groups did not appear until 36 h after treatment. At 48 h, the JHM-treated limb had only increased in length by about 16%, longer than its 12-h dimensions, but about 200 μm shorter than control limbs.

The growth in length of each of the segments was affected differently by JHM treatment (Fig. 10b). For instance, the tarsus, which increased in length by 116% in control limbs, grew only 7% in JHM-treated embryos. By contrast, the increase in length of the coxa/trochanter was the same in both treatment groups (Fig. 10b). To express these numbers as the percent inhibition by JHM, we took the difference between the change in segment lengths of control and JH-treated embryos, and divided it by the change in the length of the control segment. Calculated this way, we find a proximal-distal gradient of growth inhibition by JH: the coxa and trochanter were least affected (26% inhibition), while the femur, tibia, and tarsus were inhibited by 75%, 76% and 94%, respectively (Fig. 10c). Note that this gradient is not simply an outcome of even growth inhibition acting upon segments growing at variable rates because the tibia, which experienced the greatest P/D growth, was less inhibited than the tarsus. We observed the same trend in the T2 limb, although the differences were less pronounced (data not shown).

Because JHM treatment affected the length of distal segments more than proximal ones, we asked whether JHM inhibited the expression of genes that pattern the P/D axis of the limbs. Distal-less is required for proximal-distal growth in Drosophila, and its expression in crickets resembles that seen in the fly imaginal disc. Briefly, Distal-less is first detected in the developing limb bud, at stage 13/14 (Inoue et al. 2002). Thereafter, it is expressed in the distal three-quarters of the growing limb until it divides into a proximal “band” and a distal “sock” at stage 17 (Fig. 11a). We treated embryos with as much as 200 ng JHM at stage 13/14 and examined the effect upon Distal-less expression later in development. Treatment with 200 ng JHM at stage 13 had no effect on the appearance of this “sock and band” pattern. Although the limbs of these embryos were severely misshapen, they still possessed a clear distal sock and proximal band (Fig. 11b, c). We were also unable to detect changes in the pattern of Bric-a-brac, which acts downstream of Distal-less to specify tarsal segments 2–5 in Drosophila (Godt et al. 1993). We found that Bric-a-brac was first expressed in the distal tip of the limb bud; then as the segments formed, Bric-a-brac expression was lost in the distal tip, so that a thick band was expressed in the medial tarsus (data not shown). By stage 22, two distinct rings of enhanced expression were seen within the thick band, with a third emerging distally (Fig. 11d). We could not prevent the formation of the Bric-a-brac bands by treatment with 200 ng JHM on the second day of development, although the limb was severely stunted and compressed (Fig. 11e).

Fig. 11a–e
figure 11

The progression of Distal-less and Bric-a-brac expression is not inhibited by treatment with large doses of juvenile hormone mimic (JHM) during early development. a Distal-less expression on day 11 (about stage 25) after treatment with acetone on day 3.5 (stage 14) showing the “sock and band” pattern. Treatment with 200 ng JHM at stage 14 does not inhibit sock and band formation by day 10.5 (b) although by 13.5 days of development (c) the segments are stunted and compressed. d After acetone treatment, two distinct bands of Bric-a-brac are seen in the tarsus (arrows) by stage 22 (about 8.5 days), with a third forming distally (arrowhead). e The same pattern of Bric-a-brac expression is seen in embryos treated with 200 ng JHM on day 2–3.5 (stage 6–14), and fixed on day 9 (Fe femur, Ti tibia, Ta tarsus). The scale bar is 50 μm

Full size image

Discussion

JH during embryonic development of insects

Juvenile hormone titers have been reported in a range of hemi- and holometabolous insect embryos (Sbrenna 1990), but they differ in the onset of JH production. This difference falls along the lines of developmental mode with direct developers producing a peak of JH at about 70% embryogenesis, while metamorphosing insects produce JH at about 50% of embryogenesis (Truman and Riddiford 1999). The effects of exogenously applied JH also differ according to developmental mode. Embryos of metamorphosing insects are the least affected by JH treatment, while ametabolous insects are most dramatically affected. In this latter group, in addition to the compressed dimensions we have reported here for the cricket, JH treatment causes regression of the limb buds (James W. Truman and Lynn M. Riddiford, unpublished data). By contrast, JH treatment of holometabolous hawkmoth embryos blocks katatrepsis, but does not affect the embryo’s proportions or cuticle type (James W. Truman, unpublished data).

JH action at molts

During postembryonic development of holometabolous insects, JH acts as a “status quo” hormone, as it works in concert with ecdysteroids to prevent precocious development at molts (Riddiford 1996). For instance, in the hawkmoth, Manduca sexta, the presence of JH during the larval molts prevents initiation of metamorphosis, and a small peak of JH during the pupal molt prevents precocious adult differentiation of some imaginal tissues (Kiguchi and Riddiford 1978). Here we show that JH redirects the pronymphal molt to become nymphal, thereby acting in a seemingly opposite manner. Since these two modes of action involve molting, this would suggest that the salient feature of JH action throughout insects may be its modulation of ecdysteroid action, and not its “status quo” activity.

Because JH generally acts at molts with 20E, we wanted to know if, in addition to its effects at the pronymphal molt, JH was acting in conjunction with ecdysteroids at the E1 molt. We found that treatment of embryos with JHM prior to 4.6 days of development resulted in suppressed katatrepsis, which normally begins 3–4 days later. When treated after this time, the embryos successfully completed katatrepsis but then failed to complete dorsal closure (Fig. 5). Since the loss of the ability of JH to block katatrepsis occurs at about the time that the E1 cuticle starts to form, we suggest that the early presence of JH may shift the nature of this cuticle, perhaps making it less pliant than the thin E1 cuticle, and thereby restrict the movements of katatrepsis. Once the E1 molt was initiated, this cuticle could no longer be affected by JH but the hormone could then act later in development at the pronymphal molt.

JH and nymphal differentiation

Novak first showed that treating grasshopper embryos with JH evoked precocious development (Novak 1969). However, the ability of a juvenoid to redirect the pronymphal molt towards a nymphal fate was first demonstrated by Sbrenna-Micciarelli (1977). She found that treatment of Schistocerca gregaria embryos with the JH-mimic, farnesyl methyl ether, produced a pronymphal cuticle endowed with hairs and setae, characteristics of the nymphal cuticle (Sbrenna-Micciarelli 1977).

Juvenile hormone generally acts as a binary switch at molts. However, we found that at certain concentrations, JH III treatment could produce mosaic cuticles comprised of pronymphal and nymphal characters, suggesting that different structures had different sensitivities to JH. Previous studies of Manduca larval-pupal mosaics (Truman et al. 1974) indicated that regions of the abdominal epidermis differed in the time that 20E caused the cells to become committed to larval differentiation. This pattern also reflected their sensitivity to JH so that after a tissue passed its commitment point, JH could no longer influence its fate but the hormone could still influence the development of neighboring uncommitted tissues. A similar phenomenon may account for the mosaic response of cricket embryos to low dosages of JH III. JH III applied during katatrepsis may only persist to the onset of the pronymphal molt. Since the mandibles became responsive to JH III before the bristles and labral teeth (Fig. 7), it seems likely that sufficient JH III was present to commit the mandibles to a nymphal fate but that JH levels then declined below threshold before the labral teeth could be suppressed or the bristles induced.

Whether JH simply induces differentiation of existing structures or whether it induces precocious formation of those structures can be better understood by following the origin and fate of cercal sensory bristles. All sensory bristles in insects arise by a series of differential divisions. A sensory organ precursor (SOP) undergoes two differential divisions, where one pair of cells gives rise to the bristle and socket cells, and the other pair becomes the associated neuron and glial cell (Hartenstein and Posakony 1989). In the developing cercus of the cricket, a small bout of cell division, probably corresponding to SOP divisions, appears at stage 25–26, during the pronymphal stage, but enlarged epidermal cell groups do not appear until stage 28, after the apolysis of the pronymphal cuticle (Edwards and Chen 1979). Moreover, using an antibody directed against horseradish peroxidase to detect neuron birth and outgrowth, Shankland and Bentley (1983) found in S. americana embryos that neurons that innervate the bristles appear between 50% and 70% of development (stage 17–25), which is prior to and during pronymphal cuticle secretion. The neurons began morphogenesis immediately, while the other cells of the SO quartet, the bristle and socket cells, became quiescent and did not differentiate until 80% embryogenesis (stage 28), during the nymphal molt. These data suggest that at least some of the trichogen and tormogen cells are already present at the time of the pronymphal molt but do not make cuticular structures at that time. JH treatment apparently induces the precocious differentiation of these pre-existing cells. Interestingly, in ametabolous insects such as the firebrat, Thermobia, the cuticle of the free-living pronymph already bears bristles. Although embryonic JH titers are not available for these insects, we speculate, based on the latent developmental pathway that we uncovered by JH treatment of the pronymphal cricket, that JH causes sensory organ differentiation at the pronymphal stage of this ametabolous insect. As nymphal development became embryonized in the hemimetabolous insects, JH production would have moved to coincide with production of the final embryonic cuticle to induce the terminal features required for life outside the egg.

Studies using the allatocidal agent precocene, a compound that is converted to a cytotoxic agent in cells of the corpora allata, have generated contradictory results. Chemical allatectomy of Locusta migratoria embryos, which nearly abolished JH levels in late stage embryos, had no effect on development until the 2nd–3rd instar molt when the nymphs died as “precocious adultiforms” (Aboulafia-Baginsky et al. 1984). In contrast, treatment of the cockroach, Nauphoeta cinerea, embryo with precocene III delayed late development and reduced pigmentation of mandibles and leg bristles, suggesting a role for JH in terminal differentiation (Bruning et al. 1985). In the cricket, we found that treatment with 40 μg precocene I at dorsal closure could delay formation of pigmented mandibles by 3 days and formation of sensory bristles on the tergites by 3.5 days without changing the onset of labral teeth appearance (data not shown). We could not, however, rescue the delay with a JH mimic, and therefore cannot exclude the possibility that the delay was not due to the allatocidal action of the precocenes.

The effect of JH on growth

Novak (1969) first showed the morphogenetic effects of a juvenoid upon direct-developing embryos. By treating S. gregaria eggs with a range of juvenoids, he found a spectrum of morphogenetic defects. These included the dorsal closure and katatrepsis arrests that we have described. In addition, he reported more severe phenotypes that included loss of abdominal and thoracic segments. In these cases, he found that the timing of JH treatment influenced the segments affected: early treatment correlated with thoracic segment loss while slightly later treatment resulted in loss of abdominal segments (Novak 1969). We have observed, with very high doses of JH given on the first few days of development, embryos arrested during the germ band stage with highly reduced abdomens. However, despite treatment with JHM on the first day of development, we could not detect changes in the segmental patterning of either engrailed or an epitope shared by Ultrabithorax and Abdominal-A at the segmented germ band stage (data not shown). These data suggest that the apparent segmental deletions observed by Novak are due to a lack of segment growth after their formation.

The outgrowth of distal segments in the developing T3 leg was more affected by JH treatment than that of proximal ones. However, we did not detect a change in the pattern or progression of proximal-distal patterning genes, such as Distal-less or Bric-a-brac, even in grossly misshaped limbs. One explanation for this discrepancy might be that JH affects other proximal-distal patterning genes that we have not examined. Alternatively, JH may act at the interface between these patterning genes and effector genes that regulate growth and morphogenesis. In support of this, the limbs of embryos treated with JH are greatly reduced, suggesting a suppression of growth. A third alternative is, as discussed earlier, that JH could have altered the E1 or pronymphal cuticles so that they were less pliant, causing the limbs to compress. We have observed this in our nymphs which ecdyse from transformed pronymphal cuticles after JH III treatment. In these cases, the limbs of the precocious nymphs were also compressed in their distal segments (data not shown), but this compression was not overtly carried over to the true first nymphal cuticle.

Evolutionary implications and the pronymph hypothesis

Several years ago, Truman and Riddiford (1999) advanced a new idea about the origins of insect metamorphosis that was based on the fate of the pronymph, the first free-living instar in apterygote insects and some other arthropods. In hemimetabolous insects, this stage has become embryonized so that the first nymphal instar is the first free-living stage. With embryonization, much of the terminal differentiation previously associated with the pronymphal stage is deferred to the first nymphal stage. They suggest that de-embryonization of the pronymph prefigured the evolution of true metamorphosis. In their view, the evolution of metamorphosis followed the advancement of JH production so that it would coincide with (what was) pronymphal cuticle formation. In this paper, we have confirmed that such an advance in the presence of JH does indeed advance terminal differentiation of cuticle features, endowing the pronymph with characters associated with life outside the egg. In addition, the ability of the pronymphal cuticle to differentiate these terminal nymphal characters suggests that this pathway may be a vestige of an ancient developmental mode, and thus could have been reactivated in the line that led to the Holometabola.

It has also been suggested that the advancement of JH that caused terminal differentiation of the pronymphal cuticle in the line of insects that led to the Holometabola suppressed tissue patterning of the adult body plan (Truman and Riddiford 2002). Here we show that high levels of JHM do affect the dimensions of appendages, as discussed. However, we expected that JH would alter these proportions by altering the expression of patterning genes, and account for the differences in morphology that distinguish the nymph of hemimetabolous insects from the larva of Holometabola. However, we were unable to show a change in the expression of either Distal-less or Bric-a-brac in severely misshaped limbs, which suggests that the difference between the morphology of nymphs and the JH-induced precocious nymphs is not determined by changes in the most basic pattern elements that lie upstream of Distal-less and Bric-a-brac.

In both locusts (Sbrenna-Micciarelli 1977; Truman and Riddiford 1999) and crickets, the application of JHM prior to the pronymphal molt results in a precocious nymph that does not hatch or ecdyse. Therefore, we were surprised to see that the precocious nymphs caused by JH III treatment underwent another nymphal molt within the egg (to the first nymphal stage) and subsequently hatched, shed the induced mosaic nymphal cuticle, and grew to adulthood. Consequently, it appears that attaining a nymphal condition is not, by itself, sufficient to turn off the embryonic endocrine system.