J Mol Neurosci (1992) 3:213-218
Joumalof Molecular Neuroscience ~
Birkhauser Boston 1992
Isolation and Structural Characterization of Drosophila TDVDHVFLRFamide and FMRFamide-Containing Neural Peptides R. Nichols Departments of Biological Chemistry and Biology. University of Michigan. 830 N, University Street. II:!I Natural Science Building. Ann Arbor. MI 48109-1048, USA
Abstract. An extract of adult Drosophila melanogaster was separated by gel exclusion, ion exchange. and reversed-phase chromatography. Four peptides. each with an -ArgPheNH 2 C-terminal sequence. were identified by radioimmunoassay. The primary sequences were determined by Edman degradation and confirmed by mass spectrometry and sequence-specific radioimmunoassay. Three of the peptides are encoded by Drosophila proFMRFamide: AspProLysGlnAspPheMetArgPheN H z (DPKQDFMRFamide), ThrProAlaGluAspPheMetArgPheNH 2 (TPAEDFMRFamide), and SerAspAsnPheMetArgPheNH 2 (SDNFMRFamide). A novel Drosophila peptide ThrAspValAspHisValPheLeuArgPheNH 2 (TDVDHVFLRFamide) was also isolated. TDVDHVFLRFamide is structurally related to peptides isolated from chicken, cockroach, locust, and snail; the cockroach, fruitfly, and locust peptides differ only by the N-terminal amino acid residue. Two Drosophila neural genes, dsk and FMRFamide. are known to encode -ArgPheNH 2 -containing peptides; however, neither encodes TDVDHVFLRFamide'; indicating that Drosophila contains another precursor encoding -ArgPheNH z peptides.
The tetrapeptide PheMetArgPheNH 2 (FMRFamide) was first identified and purified as a cardioexcitatory peptide from mollusk (Price and Greenberg, 1977). Since then, FMRFamide-like immunoreactive materials have been isolated from several invertebrates and are N-terminally extended peptides with common C-termini -XArgPheNH 2 (where X = Leu or Met). These structurally related peptides belong to a family of peptides thought to act as transmitters, regulators, and modulators in the central nervous system and function in a broad range of important physiological processes. The Drosophila drosulfakinin (dsk) and FMRFamide genes encode multiple -MetArgPheNH 2 containing peptides. The dsk precursor encodes two
-MetArgPheNH 2 peptides (Nichols, 1987; Nichols et aI., 1988), and the FMRFamide gene encodes a protein that may be processed to five different FMRFamide-containing peptides (N ambu et aI., 1988; Schneider and Taghert, 1988). Although peptides with the C-terminal sequence -LeuArgPheNH 2 have been isolated from cockroach (Holman et aI., 1986), snail (Ebberink et aI., 1987), crab (Krajniak et al., 1990), locust (Robb et aI., 1989), and chicken (Dockray et aI., 1983), no DrosophilaLeuArgPhe-NH 2 peptide has been isolated, and none are encoded in either the dsk or FMRFamide precursor. The precursors of biologically active peptides often encode several peptides and, in general, potential peptide sequences can be predicted by examining the precursor sequence since conversion of a precursor to final products most commonly occurs at adjacent basic amino acid residues or at single arginyl residues (Wold, 1981; Loh et aI., 1984). However, both primary and secondary structural features appear to be important for determining processing sites since not all of these possible cleavage sites are utilized. As a result, it is not yet possible to predict with certainty what the final products of processing will be solely on the basis of the deduced sequence of the precursor. Although deduction of amino acid sequence from DNA has become a common first step and an important method for predicting putative bioactive peptides, it is necessary to experimentally isolate and structurally characterize the naturally occurring peptides. The isolation of the naturally occurring peptides will verify that the peptides are indeed processed, confirm the predicted primary sequence, and identify posttranslational modifications. Lack of structural data for the naturally occurring peptides may result in errone-
Drosophila Neural Peptides
ous assignment of function due to similarities in sequence and/or invalid localization data due to crossreactivities. Although Drosophila melanogaster has proven to be a powerful system for molecular and genetic studies of nervous system development, the study of Drosophila neural peptides has been limited by the lack of structural characterization of the naturally occurring peptides. As a step to determine the function and expression of Drosophila FMRFamide-related neural peptides, this article describes the identification and structural characterization of four Drosophila -ArgPheNH 2 peptides.
plied to a PLRPS reverse-phase column (I00A.) 4.6 mm x 25 cm (Polymer Labs), and peptides were eluted with a linear gradient of 0-40% acetonitrile plus 0.1 % TF A in 60 minutes. Flow rate was 0.7 mIlmin, and fractions were collected based on absorbance measured at A255 . Five-microliter aliquots of each fraction collected were analyzed by radioimmunoassay. The final stage in purification was frequently a reverse-phase microbore HPLC using a I-mm x 22.5-cm, 300-A. column. The flow rate was 50 J.Ll/min. Typical gradients were from 0 to 40% acetonitrile with 0.1 % TF A in 60 minutes. Peptides were detected by monitoring absorbance at 215 nm.
Materials and Methods
Radioimmunoassay Antibody was raised in two New Zealand white rabbits to FMRFamide conjugated to succinylated thyroglobulin via carbodiimide coupling (Bauminger and Wilchek, 1980). The rabbits were immunized by intradermal injections of I mg antigen emulsified in Freund's complete adjuvant and boosted every two weeks by subcutaneous injections of 0.5 mg antigen in Ribi Adjuvant System (Ribi Immunochem Research). The titer of the antisera was monitored using a solid-phase dot-blot assay for peptides (Andrews, 1987). Antisera from both rabbits were found suitable for a general radioimmunoassay to detect peptides containing a C-terminal -ArgPheNH 2 • Synthetic drosulfakinin-I (DSK-I; PheAspAspTyrGlyHisMetArgPheNH 2 ) was iodinated using 10dogen Pierce) (Markwell, 1982). DSK-I is a Drosophila neural peptide (Nichols, 1987; Nichols et aI., 1988) and was chosen as the competitive antigen because it contains a C-terminal -ArgPheNH 2 and residues (His and Tyr) suitable for iodination; authentic FMRFamide does not contain a suitable site for iodination. Antigens were detected using a competition radioimmunoassay (Harlow and Lane, 1988) with the solid-phase matrix provided by adding Staphylococcus aureus membranes (Calbiochem). To quantitative the level of antigen in the column fractions, serial dilutions of the antigen test solution were made and a standard curve generated for each radioimmunoassay. Each experimental and standard curve sample was done in duplicate. A typical assay was: 1120 of the column fraction dried to remove the column buffer, 20 J.Ll of washed S. aureus membranes, 10 J.Ll of antiserum diluted I: 1000, 2 J.Ll of 125I-labeled DSK-I (-40,000 cpm/J.L1), and 175 J.Ll buffer (3% BSA/PBS with 0.02% sodium azide) mixed thoroughly and incubated at 4°C overnight. The samples were centrifuged, the supernatant re-
Extraction and purification of peptides Six hundred grams of adult Oregon-R Drosophila melanogaster were homogenized in methanolwater-acetic acid, 900:90: 10, containing 25 J.Lg/ml of pepstatin A. After the extract was freeze-dried and washed with diethyl ether, it was resuspended in I N formic acid and applied to a Sephadex G-25 column (4.5 x 100 cm), which had been equilibratea with I N formic acid. Peptides were eluted from the column at room temperature using I N formic acid. Two-milliliter fractions were collected at a flow rate of 35 ml/hr and assayed for immunoreactivity as described below. Fractions from each of the two major immunoreactive peaks were combined separately, diluted, and subjected to a de salting, concentrating step using SepPak cartridges (Waters). Peptides were eluted from the SepPak using 50% acetonitrile, 0.1 % TF A with no loss of immunoreactivity. The de salted fractions. from the Sephadex G-25 column were diluted 3: I with 5 mM sodium phosphate, pH 3.0, 25% acetonitrile, and further purified using two HPLC columns having different separation characteristics. Immunoreactive materials were applied to a polysulfoethyl aspartamide ion exchange column, 9.3 mm x 20 cm (PolyLC) (Alpert and Andrews, 1988) equilibrated with buffer A (5 mM sodium phosphate, pH 3.0, 25% acetonitrile). The operating conditions were: buffer A for 10 minutes after injection, to 20% buffer B (200 mM sodium phosphate, pH 3.0, 25% acetonitrile) at 80 minutes, to 40% buffer B at 90 minutes, and held at 40% butTer B for 5 minutes with a flow rate of 1.2 mIlmin. Fractions were collected on the basis of absorbance measured at A 225 . Five microliter aliquots of each fraction collected were analyzed by radioimmunoassay. Immunoreactive materials were individually ap-
moved, and the membranes counted using a Searle Analytic Inc. model 1190 gamma counter. Background was determined by including a sample without labeled antigen and one without antibody. Sequence and mass analysis The purified peptides were sequenced on an Applied Biosystems model 470 automated protein sequencer with online detection of PTH amino acids using standard operating conditions. The masses were determined by peak matching to an appropriate CsI cluster ion by fast atom bombardment mass spectrometry using a Kratos MS-50 mass spectrometer as described previously (Andrews et aI., 1987). The molecular ion calculations were based on composition using PROCOMP version 1.2, a computer program developed by P.C. Andrews, PhD (University of Michigan) for peptide data manipulations on an IBM Pc.
Results The first step in the purification of peptides from the Drosophila extract was separation based on size using a Sephadex G-25 column. Two major peaks of -ArgPheNH 2-like immunoreactivity were identified by radioimmunoassay after gel exclusion fractionation of the Drosophila extract (Fig. I). Another peak of immunoreactive materials containing larger species, which may represent incompletely processed products or as yet unidentified -ArgPheNH 2 peptides, was not further studied. Since the FMRFamide-related peptides isolated to date and those that can be predicted from the FMRFamide-like precursors are fairly similar in size, it was necessary to take advantage of other j
Table 1. Automated Edman Sequencing: Amino Acid Yields of Drosophila -ArgPheNH 2 Peptides at Each Step Cycle Number
Amino Acid Residue Assignment
Thr Asp Val Asp His Val Phe Leu Arg Phe
36.5 140 58 119 51 31 35 32 13 4.2
Asp Pro Lys Gin Asp Phe Met Arg Phe
164 65 81 57 49 41 39 22
Thr Pro Ala Glu Asp Phe Met Arg Phe
47 100 67 38 19 16 25 18 10
TDVDHVFLRFamide I 2 3 4 5 6 7 8 9 10 11
2 3 4 5 6 7 8 9
independent physical parameters (charge and hydrophobicity) to ensure separation of these structurally related peptides. The fractions comprising each of the two major peaks containing immunoreactive material were combined separately, freezedried to reduce volume, and the -ArgPheNH 2 peptides were purified by ion exchange and reversedphase chromatography and then subjected to automated Edman sequencing (Table I). Frequently, the final material was further purified by
Drosophila Neural Peptides
"oc: 2 t ", .§E j
Fig. 1. Sephadex G-25 profile of Drosophila homogenate. A 280 elution profile (closed circles) and immunoreactive material (bar graph) from size exclusion chromatography of an acidic methanol extract of 600 g of adult D. melanogaster on Sephadex G-25. Fractions 161 (void volume) to 531 (salt volume) are illustrated.
2 3 4 5 6 7 8 9 10 SDNFMRFamide I 2 3 4 5 6 7 8
Ser Asp Asn Phe Met Arg Phe I N.Q. Observed but not quantitated.
2.0 6.8 6.1 3.6 0.3
Nichols: Drosophila Neural Peptides
microbore reverse-phase HPLC before sequencing. Representative chromatograms are illustrated for the purification of TDVDHVFLRFamide: ion exchange (Fig. 2) and reverse-phase HPLC (Fig. 3). Although Edman sequencing does not provide any information regarding the presence of a C-terminal amide, the radioimmunoassay for FMRFamide requires a C-terminal amide as part of the recognition site and the presence of a C-terminal amide results in a decrease in mass by one mass unit relative to a free carboxyl terminus. Mass spectral analyses of the naturally occurring peptides confirmed the presence of a C-terminal amide and were consistent with the sequences observed (Table 2). Multiple peaks containing TDVDHVFLRFamide or DPKQDFMRFamide were observed as a result of gel exclusion, ion exchange, and reverse-phase chromatography. Some of the duplication is due to overlap in the initial, low-resolution purification procedure, while some of the mUltiple peaks were clearly the result of oxidized Met residues (identified by an increase in mass by 16 mass units over the nonoxidized peptides). The data presented identify three FMRFamidecontaining peptides: DPKQDFMRFamide, TPAEDFMRFamide, and SDNFMRFamide, all contained within Drosophila proFMRFamide and flanked by typical processing sites. The structural characterization of the naturally occurring Droso20
.~ i ijl
~ 1.0_ c:: co .c
~ij . '11'
o~--~UL~----~--~--,_---r--_r--~ o 10 20 30 40 50 60 ~o
Fig. 2. Polysulfoethyl aspartamide column profile: purification ofTDVDHVFLRFamide. Separation of Sephadex G-25 fractions 359-379 on polysulfoethyl aspartamide column. Column fractions collected at 52 minutes (peak indicated by arrow) contained immunoreactive material.
_ _ _ _ _ _. -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
05 C CD