A family of neuropeptides that inhibit juvenile hormone biosynthesis in the cricket, Gryllus bimaculatus.

Four nonapeptides that inhibit juvenile hormone synthesis have been isolated by four high performance liquid chromatographic steps from extracts of the brain of the field cricket, Gryllus bimaculatus. The primary structures of these peptides were assigned by Edman degradation and mass spectrometry as Gly-Trp-Gln-Asp-Leu-Asn-Gly-Gly-Trp-NH2 (Grb-AST B1), Gly-Trp-Arg-Asp-Leu-Asn-Gly-Gly-Trp-NH2 (Grb-AST B2), Ala-Trp-Arg-Asp-Leu-Ser-Gly-Gly-Trp-NH2 (Grb-AST B3), and Ala-Trp-Glu-Arg-Phe-His-Gly-Ser-Trp-NH2 (Grb-AST B4). Each of the peptides shows high sequence similarity to the locustamyoinhibiting peptide (Lom-MIP), but is structurally different from all the allatostatins so far identified. The synthetic allatostatins Grb-AST B1-4 are potent inhibitors (50% inhibition at 10 to 7 10M) of juvenile hormone III biosynthesis by corpora allata from 3-day-old virgin females of G. bimaculatus using an in vitro bioassay. At 10M, Grb-AST B1 also strongly inhibits juvenile hormone III biosynthesis by corpora allata from 2-day-old adult males and 1-day-old (males and females) and 4-day-old (females) last instar larvae of G. bimaculatus. The inhibitory effect of Grb-AST B1 was also evident on corpora allata from a related species, Acheta domesticus. Inhibition of juvenile hormone synthesis by Grb-AST B1-4 is reversible.

Inhibition of juvenile hormone (JH) 1 biosynthesis by corpora allata (CA) in vitro has provided the basis for the isolation of a family of allatostatins (Tyr-Xaa-Phe-Gly-Leu-NH 2 ) from brain extracts of various cockroach species (Diploptera punctata, Periplaneta americana, and Blattella germanica) (1)(2)(3)(4)(5)(6). In D. punctata and P. americana, molecular cloning has led to the isolation of cDNA that encodes a precursor polypeptide containing 13 and 14 potential allatostatic sequences, respectively, including those formerly identified through conventional purification techniques (7,8). In D. punctata, the expression of the allatostatin gene in midgut endocrine cells was also demonstrated (9). In the blowfly, Calliphora vomitoria, six peptides with sequence similarity to cockroach allatostatins have been identified; these inhibit JH III biosynthesis in cockroaches, but do not inhibit synthesis of JH III bisepoxide in the fly itself (10,11). An allatostatin that is structurally unrelated to the cockroach allatostatins has been purified and characterized from heads of pharate adults of the moth Manduca sexta (12).
Very recently, we identified two peptide inhibitors of JH III biosynthesis from extracts of the brain of the field cricket, Gryllus bimaculatus, which were designated Grb-AST A1 and Grb-AST A2, respectively (13). Each of the peptides shows C-terminal amino acid sequence similarity to cockroach allatostatins and blowfly callatostatins. In this report, we describe the isolation and primary structure of four nonapeptides from brains of G. bimaculatus that are potent inhibitors of JH III synthesis in vitro by CA from virgin adult females, males, and last instar larvae of the species. These peptides are structurally different from all allatostatins so far identified, but are homologous to locustamyoinhibiting peptide (Lom-MIP), a peptide that inhibits the spontaneous contractions of the hindgut and oviduct of Locusta migratoria (14). We propose that this family of peptides represents a novel type of neurohormones that inhibit JH biosynthesis in crickets.

Insects and Tissue Dissection
Virgin adult females and males of G. bimaculatus (de Geer) (Ensifera, Gryllidae) reared at 27°C and staged according to chronological age (15) were used in the experiments. Brains (2700) of 2-4-dayold females were dissected free of optic lobes, corpora cardiaca-CA complexes, and adhering fat body and stored in extraction medium (methanol/water/acetic acid (100:10:1, v/v/v)) at Ϫ25°C prior to purification. Single CA of 3-day-old females were used to test chromatographic fractions for allatostatic activity and to measure dose response and reversibility of inhibition of synthetic peptides. These CA exhibit the highest and most consistent rate of JH III biosynthesis from animal to animal (13). Inhibition of JH III synthesis was also tested on CA from 2-day-old adult males and 1-day-old (males and females) and 4-day-old (females) last instar larvae of G. bimaculatus and from 5-day-old adult females of the house cricket, Acheta domesticus. Adult A. domesticus females were removed from stock cultures on the day of emergence and maintained at 27°C until use.

Radiochemical Assay for Allatostatic Activity
Juvenile hormone III biosynthesis in vitro was determined by the radiochemical assay described previously (16 -18), with some modifications. The radiolabeled precursor for JH III was L-[methyl-14 C]methionine (Amersham Buchler, Braunschweig, Germany; final specific activity of 1.11-1.30 ϫ 10 9 Bq⅐mmol Ϫ1 ; final concentration of L-methionine of 0.24 -0.28 mM). After a 1.5-h preincubation in medium without [ 14 C]methionine, single CA were incubated for 2 h in 20 l of medium TC199 (Sigma M7653 (Deisenhofen, Germany); with Hanks' salts and sodium bicarbonate, without L-glutamine; buffered with 25 mM HEPES; supplemented with CaCl 2 to a final concentration of 3 mM; fortified with 1% Ficoll 400 (Pharmacia, Uppsala, Sweden); sterilized by suction through a 0.2-m filter; adjusted to pH 7.0; containing 12,500 -18,000 Bq of L-[methyl-14 C]methionine) to establish rates of JH synthesis by untreated glands; subsequently, CA were transferred to medium with the extract/synthetic peptide to be tested for a second 2-h period. The percentage change in rates of synthesis was calculated. Following incubation, the medium was processed for extraction and JH III quantification as described (18). * This work was supported by Landesforschungsschwerpunkt Baden-Wü rttemberg "Molekulare Grundlagen der zellulä ren Differenzierung, Musterbildung und Morphogenese" Project C2 and by Deutsche Forschungsgemeinschaft Grant Ho 631/15-1. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Brain Extracts and Prepurification by Solid-phase Extraction
A total of 54 batches, each batch representing 50 brains, was extracted as described previously (13). The extracts were subsequently prepurified on Sep-Pak cartridges (13). Allatostatic material eluted in the 40% CH 3 CN fraction. A total of 13 of these Sep-Pak separations were required to process the entire material.

Chromatography
HPLC separations were carried out on a Shimadzu HPLC system (13). The eluted material was detected at 215 nm.
Second HPLC Run-The conditions were as follows: column: 250 ϫ 4.6-mm Shiseido CAPCELL PAK C 18 SG 300 with 10 ϫ 4.6-mm guard column (same material; Grom, Herrenberg-Kayh, Germany); solvent A: 0.13% heptafluorobutyric acid in water; solvent B: 0.13% heptafluorobutyric acid in CH 3 CN; gradient: 0 -2 min of 5% solvent B and 2-52 min of 5-60% solvent B (linear gradient, 1.1%/min) followed by a 7-min wash at 100% solvent B; and flow rate: 1 ml/min. The pooled fractions 23 and 24 from the first HPLC runs (2400 brain equivalents) were dried down to 500 l, diluted with 500 l of 0.13% heptafluorobutyric acid in 10% CH 3 CN, and loaded onto the column. Peaks were collected manually into reaction tubes and tested for allatostatic activity (33.3 brain equivalents/assay). Each of the tubes was preloaded with 5 l of a 0.1% bovine serum albumin solution (Sigma A3294; purified by HPLC) as a carrier peptide to reduce loss of material. Four fractions showed allatostatic activity and were subjected to further purification.

Sequence Analysis
The active material from the final HPLC separations was loaded onto a Polybrene-coated glass-fiber filter and sequenced by automated Edman degradation using a Model 477A sequenator connected to a Model 120A on-line phenylthiohydantoin analyzer (Applied Biosystems, Weiterstadt, Germany).

Mass Analysis
Mass spectra were acquired using a matrix-assisted laser desorption ionization spectrometer (Bruker Reflex, Bruker Franzen, Bremen, Germany). The acceleration voltage was set to 30 kV for the linear mode. The matrix was a saturated solution of ␣-cyano-4-hydroxycinnamic acid dissolved in water/acetonitrile (7:3, v/v) (19). Peptide solutions (0.5 l, ϳ1 pmol) were mixed on target with matrix solution (1:1, v/v) and left to dry. Each spectrum is the average of ϳ50 -200 single shot spectra acquired in sets of five shots. Mass analysis was kindly performed by G. Talbo (EMBL, Heidelberg, Germany).

Peptide Synthesis
Peptide synthesis was performed on a Model 9050 peptide synthesizer (Milligen, Eschborn, Germany) using Fmoc/HOBt chemistry. Peptides were synthesized in the amide form using an Fmoc-peptide amide linker polyethylene glycol-polystyrene resin (Milligen). Synthetic peptides were purified after cleavage from the resin by reversed-phase HPLC. Confirmation of the correct sequence of synthetic peptides was attained by mass analysis and coelution with the native peptides.

Computer Sequence Analysis
Computer sequence analysis was carried out performing a blitz E-mail search at EMBL and a tblastn search at the NCBI Blast network server. SwissProt release 30 (December 1994), EMBL release 41 (December 1994), and GenBank release 88 (April 1995) were searched for similarity to peptides Grb-AST B1-4 using program default parameters.

HPLC Separation of Allatostatic Neuropeptides-Bioactive
material from the first HPLC runs (pooled fractions 23 and 24) eluted from the second HPLC system between 30 and 40 min (Fig. 1 Amino Acid Sequence Analysis-Aliquots from HPLC fractions B1-4 were subjected to automated Edman degradation (fraction B1: 210 pmol/286 brain equivalents; fraction B2: 103 pmol/551 brain equivalents; fraction B3: 125 pmol/740 brain equivalents; and fraction B4: 86 pmol/604 brain equivalents). Sequence analyses indicated four nonapeptides. The primary structures were elucidated for fraction B1 as Gly-Trp-Gln-Asp-Leu-Asn-Gly-Gly-Trp, for fraction B2 as Gly-Trp-Arg-Asp-Leu-Asn-Gly-Gly-Trp, for fraction B3 as Ala-Trp-Arg-Asp-Leu-Ser-Gly-Gly-Trp, and for fraction B4 as Ala-Trp-Glu-Arg-Phe-His-Gly-Ser-Trp (Table I). Mass spectral analysis indicated that the C terminus was amidated in all the peptides. Values for mass ([M ϩ H] ϩ ) determined by spectrometry match precisely the calculated molecular sizes (Table I). Peptide contents per brain were calculated from the data of the sequence analysis (loss of material during chromatographic steps not taken into account; Table I).
Synthetic Peptides-All four peptides were synthesized with amidated C termini. Synthetic peptides showed the same retention times as their native counterparts in all four HPLC systems, and coelution gave a single sharp peak, indicating the amidated peptides to be the naturally occurring form (data not shown). The dose responses for inhibition of JH III synthesis by Grb-AST B1-4 are shown in Fig. 2. Fifty percent inhibition of JH III synthesis was achieved with 7 ϫ 10 Ϫ8 M Grb-AST B1, 2 ϫ 10 Ϫ8 M Grb-AST B2, 7 ϫ 10 Ϫ8 M Grb-AST B3, and 1 ϫ 10 Ϫ8 M Grb-AST B4 (Table I).
A computer search for similar peptides revealed sequence similarities with locustamyoinhibiting peptide (Lom-MIP) ( Table I) (14). Sequence similarity between Grb-AST B1 and Lom-MIP is ϳ78%. Synthetic Lom-MIP, however, showed a maximal inhibitory activity on JH III biosynthesis of crickets that was significantly lower than that of Grb-AST B1-4 (compare values for 50% inhibition; Fig. 2 and Table I). On the other hand, half-maximal inhibition (which is different from 50% inhibition since the glands are not inhibited up to 100%) is reached at ϳ5 ϫ 10 Ϫ8 M. This is only slightly higher than the respective doses of Grb-AST B1-4 required for half-maximal inhibition. This means that Lom-MIP and Grb-AST B1-4 may bind equally effectively to the cricket allatostatin receptor, but that Lom-MIP causes a weaker response. The adipokinetic peptide of G. bimaculatus (Grb-AKH) (20), which shares the two C-terminal amino acids with Grb-AST B1-3 (Table I), showed ϳ30% inhibition at 10 Ϫ5 M, the highest concentration soluble in the incubation medium (data not shown).
Reversibility of Inhibition-After 1 h of incubation in medium without [ 14 C]methionine, single CA were transferred to radioactive medium and incubated for 2 h to establish rates of synthesis by untreated glands; then CA were transferred to medium containing 10 Ϫ7 M allatostatic peptides B1-4 and incubated for another 2 h. Finally, CA were transferred to medium without addition of peptides and incubated again for 2 h. Glands recovered completely from inhibited rates (Fig. 3).
Sex and Species Specificity-Grb-AST B1 was tested on single CA from 2-day-old adult males of G. bimaculatus at concentrations ranging from 10 Ϫ9 to 10 Ϫ6 M (Fig. 4). Fifty percent inhibition was achieved with 4 ϫ 10 Ϫ8 M Grb-AST B1. Grb-AST B1 was also tested on 1-day-old male and female last instar larvae. At a concentration of 10 Ϫ7 M, Grb-AST B1 inhibited JH III biosynthesis by CA from these larvae by 55.1 Ϯ 7.7% (mean Ϯ S.E.; Table II). No sex differences could be found. When tested on single CA from 4-day-old female last instar larvae, 10 Ϫ7 M Grb-AST B1 reduced JH III biosynthesis by 43.3 Ϯ 12.8% (Table II). Grb-AST B1 (10 Ϫ7 M) also inhibited JH III synthesis by CA from 5-day-old adult female house crickets, A. domesticus, by 70.3 Ϯ 2.6% (Table II). DISCUSSION The search for neuropeptides with allatostatic activity in G. bimaculatus has led to the isolation of four nonapeptides with sequence similarity at the C terminus (Gly-Xaa-Trp-NH 2 ; Xaa ϭ Gly in B1-3 and Ser in B4) and a common amino acid at position 2 (Trp). In addition, three of the four peptides (B1-3) have in common amino acids at positions 4 and 5 (Asp-Leu). These peptides have been designated G. bimaculatus allatostatic neuropeptides B1-4 (Grb-AST B1-4) in accordance with their biological activity and with the widely accepted nomenclature for invertebrate neuropeptides (21). The B designation has been used because another class of allatostatic neuropeptides has been recently presented in G. bimaculatus, Grb-AST A1 and A2 (13), which are members of the Tyr-Xaa-Phe-Gly-Leu-NH 2 allatostatin family first found in D. punctata (1).
The primary structures of the four allatostatic neuropeptides show homology to the locustamyoinhibiting peptide (Lom-MIP), which was isolated from brain-corpora cardiaca-CAsubesophageal ganglion extracts of the locust L. migratoria (14). Nonapeptides Grb-AST B1-3 resemble Lom-MIP in having Trp at position 2 (also in Grb-AST B4), Asp-Leu at positions 4 and 5, and Gly-Trp-NH 2 at the C terminus. Lom-MIP suppresses the spontaneous contractions of the hindgut and oviduct of L. migratoria and of the hindgut of a cockroach, Leucophaea maderae (14). The C terminus Gly-Trp-NH 2 in Lom-MIP and in Grb-AST B1-3 is in common with the myoinhibiting peptide Ala-Pro-Gly-Trp-NH 2 isolated from the snail Lymnea stagnalis (22). In addition, Grb-AST B1-3 show some carboxylterminal sequence similarity to Grb-AKH, which ends in Thr-Gly-Trp-NH 2 (20). None of Grb-AST B1-4 shows sequence similarity to either of the known allatoactive neuropeptides.
The ability of Grb-AST B1-4 to inhibit JH III biosynthesis in crickets is almost an order of magnitude lower than that of Grb-AST A1 and A2 (Table I). Maximal inhibition (reached at an allatostatin concentration of ϳ10 Ϫ6 M), however, ranges from 60 to 80% for both classes. In the assay on CA from 3-day-old virgin females, the peptides showed 50% inhibition of JH III synthesis at 1 ϫ 10 Ϫ8 to 7 ϫ 10 Ϫ8 M. This suggests that cricket CA are highly sensitive to both types of cricket allatostatic neuropeptides when they are at the height of activity during middle-late vitellogenesis (day 3 after imaginal moult). In D. punctata, an inverse relationship between the activity of the CA and their sensitivity to allatostatins was observed (3,23). Also in crickets, CA with lower basal rates of JH III biosynthesis (e.g. CA from 5-day-old females) might be more sensitive to allatostatins, but the rate of JH synthesis by these CA is often very inconsistent, which makes these CA unsuitable for testing allatoinhibiting substances.
Testing synthetic peptides or HPLC fractions for only 2 h was necessary because otherwise total incubation time would have been too long and often resulted in reduced rates of JH III biosynthesis during the last incubation period. 2 Compared with the 3-4-h incubation used by many other authors (1,5,17), the 2-h incubation may slightly reduce sensitivity of the assay, but was sufficient to show fast response to and reversibility of allatostatic material.
From a structural point of view, it is interesting to note that a change at position 7 (Ala instead of Gly as found in Lom-MIP) reduces the inhibitory activity. On the other hand, a change at position 8 (Ser instead of Gly as in Grb-AST B4) has no negative effect on the allatostatic activity: Grb-AST B4 turned out to be the most active of the four B allatostatins. Grb-AKH, which shares only the two C-terminal amino acids with Grb-AST B1-3, had only weak allatostatic activity. The occurrence of Trp at position 2, which is a common feature of all B allatostatins and of Lom-MIP, might indicate the importance of this residue with regard to the biological activity of these peptides. The information about the bioactivity therefore might be encoded not only in the C-terminal region, as shown for the Tyr-Xaa-Phe-Gly-Leu-NH 2 allatostatin family (3,24).
One may ask whether Grb-AST B1-4 can be termed "allatostatins" since the first isolation of a member of this peptide family (Lom-MIP) used a bioassay based on myoinhibiting activity of the HPLC fractions (14). In crickets, the allatostatic activity is the only effect demonstrated so far; therefore, the term allatostatin seems to be justified.
Grb-AST B1-4 proved to be neither sex/stage-nor speciesspecific. Although extracted only from brains of adult females, they inhibit JH III biosynthesis by CA from adult males and last instar larval males and females as well. The allatostatic activity of the B allatostatins was also shown in a closely related cricket species, the house cricket, A. domesticus. Interspecific allatostatic activity was also shown for the Tyr-Xaa-Phe-Gly-Leu-NH 2 allatostatin family (25,26).
The existence of two types of allatostatic neuropeptides in G. bimaculatus makes it more difficult to understand their modes of function. As yet, nothing is known about changes in brain and hemolymph titer of allatostatic neuropeptides, e.g. during the female reproductive period of G. bimaculatus. Isolation and quantification of allatostatic receptors will be needed, the levels of which may be largely attributed to changes in CA sensitivity for allatoregulating peptides (8). The occurrence of multiple allatostatin species might suggest the existence of individual receptors for each species of molecule. It remains to be determined, however, whether each allatostatin species is associated with a different receptor. It even remains to be determined whether allatostatins are the principal regulators of JH biosynthesis in vivo. The immunocytochemical localization of Tyr-Xaa-Phe-Gly-Leu-NH 2 allatostatin-like peptides in the central nervous system as well as in midgut cells of many invertebrates (8,27) suggests that allatostatins are multifunctional neuropeptides, even though inhibition of JH biosynthesis by CA in vitro was the only bioassay utilized in their isolation. Sequence homology, as found between Lom-MIP and the family of allatostatic neuropeptides presented in this paper, may indicate a myomodulatory role of these peptides, as was also discussed for the Tyr-Xaa-Phe-Gly-Leu-NH 2 allatostatin family (28,29).