Identification of RFamide neuropeptides in the medicinal leech

Peptides, Vol. 12. pp. 897-908. ©PergamonPress plc, 1991. Printedin the U.S.A.

0196-9781/91 $3.00 + .00

Identification of RFamide Neuropeptides in the Medicinal Leech B R U C E D. E V A N S , J A N P O H L , * N I C H O L A S A. K A R T S O N I S 1 A N D R O N A L D L. C A L A B R E S E

Department of Biology, Emory University, 1510 Clifton Road, Atlanta, GA 30322 *Microchemical Facility, Emory University, Atlanta, GA 30322 Received 13 March 1991 EVANS, B. D., J. POHL, N. A. KARTSONIS AND R. L. CALABRESE. Identification ofRFamide neuropeptides in the medicinal leech. PEPTIDES 12(5) 897-908, 1991.--Using a four-step reverse phase HPLC separation and RIA, five RFamide peptides were purified from CNS extracts of the leech Hirudo medicinalis. YMRFamide, FMRFamide, YLRFamide, FLRFamide, and GGKYMRFamide were identified by a combination of antiserum specificity in RIA, Edman degradation, and mass spectrometry. At least three of these five endogenous peptides can modulate neuromuscular interactions in the leech (38). FMRFamide-like immunoreactivity was selectively released from neural processes on isolated heart tubes in the presence of calcium and depolarizing levels of potassium. Annelida FMRFamide Radioimmunoassay

Hirudomedicinalis

HPLC

Microsequence analysis

Neuropeptide

quenced from several species (27, 29, 30, 49-53). These genes encode some peptides, such as FLRFamide and FMRFamide (53), which are known to be expressed. Gene sequence analyses also suggest the presence of RFamide peptides not yet biochemically isolated. For example, six peptide sequences terminating with the sequence PNFLRFamide are encoded in a gene isolated from the nematode C. elegans (27), but these peptides have not yet been identified in tissue extracts. The leech nervous system is particularly suitable for studying neuropeptides. The CNS comprises anterior and posterior brains, and a ventral chain of 21 segmental ganglia linked by axon bundles called connectives. Each segmental ganglion contains approximately 400 neurons [700 in ganglia 5 and 6 (31)], many of which are identified and have known physiological roles (35). These neurons are accessible for immunocytochemical labeling, intracellular recording and staining, microdissection for radioimmunoassay (RIA), and reverse phase high pressure liquid chromatography (rpHPLC) analysis (24, 25, 28). FMRFamide-like immunoreactivity (FLI) is present in cell bodies and neuronal processes throughout the leech CNS, including both head and tail brains, and segmental ganglia 1-21 (11,24). FLI is localized to heart motor (HE) neurons, heart modulatory (HA) neurons (24), and several motor neurons to longitudinal and medial dorsoventral muscle (37). HA neurons and neuronal processes on the heart contain a peptide indistinguishable from FMRFamide on two different rpHPLC systems (28). Both HA and HE neurons form synapses on heart muscle, and their axon terminals contain both small clear vesicles and large dense core vesicles (32). Activity in HA neurons increases the amplitude and duration of heartbeat tension in a heart entrained to the rhythm of the HE

RFamide neuropeptides (peptides with the carboxyl terminal sequence Arg-Phe-amide) have been identified in several species from throughout the animal phyla (5, 22, 46). These peptides exert diverse physiological effects on neurons and muscles in the animals where they are found. While this family of peptides shows considerable diversity in amino acid sequence, some similarities have emerged. Several identified peptides from different phyla, including the molluscs (5, 10, 44, 45) and crustaceans (54), share the carboxyl terminal sequence Phe-Leu-Arg-Phe-amide. Three decapeptides isolated from different insect species are identical at 7 of the 10 amino acid residues (X-D-V-X-H-X-FLRFamide) (21, 22, 47). Other peptides of the form -Leu-Arg-Phe-amide include KHEYLRFamide from the nematode Ascaris suum (8) and LPLRFamide from the chicken (9). The tetrapeptide FMRFamide is apparently found in all molluscs (42, 43, 45), and also in the annelid Nereis (23). An N-terminally extended form of FMRFamide, DPKQDFMRFamide, is present in the nervous system of Drosophila melanogaster (36). Several other RFamide peptides have been isolated and identified from various vertebrate and invertebrate species, including three RFamide peptides in the mosquito with the carboxyl terminal sequence Thr-Arg-Phe-amide [(34), M. R. Brown, personal communication]. Two peptides from coelenterates (18,19) and one from a mollusc (13) terminate with the sequence GlyArg-Phe-amide. Two related peptides isolated from bovine brain share the carhoxyl terminal sequence Pro-Gln-Arg-Pheamide (57), while Ascaris contains the heptapeptide KNEFIRFamide (7). Genes encoding RFamide peptides have been isolated and se-

~Present address: School of Medicine, Emory University, Atlanta, GA 30322.

897

898

EVANS, POHL, KARTSONIS AND CALABRESE

neurons (4); bath application of FMRFamide mimics this modulatory effect of the HA neurons (25). The HE neurons are cholinergic (24,55), and elicit myogenic activity in quiescent hearts in the presence of curare, an effect which is also mimicked by bath-applied FMRFamide (28). Bath-applied FMRFamide and several related peptides, YMRFamide, FLRFamide, pQDPFLRFamide, YGGFMRFamide, and LPLRFamide, elicit contractions in leech longitudinal muscle (38). Since different RFamide peptides may have different physiological actions (5, 6, 26) or potencies (12, 14, 38, 48), it is important to know the identities of the endogenous peptides in each system where they occur. RFamide peptides in Hirudo have been chromatographically characterized (28), using an antiserum [231 (39)] which, in RIA, is highly selective for peptides containing the carboxyl terminal sequence -Met-Arg-Phe-amide. We have characterized RFamide peptides in the leech by using rpHPLC and RIA with a less specific antiserum [671C (33)], allowing us to detect peptides, such as FLRFamide, that do not cross-react with antiserum 231 to any great extent (39). A comparison to identified RFamide peptides in other animal phyla may now be made, and the physiological roles for each of the endogenous RFamide peptides in Hirudo may now be thoroughly investigated. METHOD

Animals Adult specimens of the leech Hirudo medicinalis were purchased from Blutegelimport and Versand, Leeches USA, or Eduardo Macagno at Columbia University. Animals were kept in the dark at 15°C in artificial pond water. They were occasionally fed defibrinated bovine blood (Carolina Biological).

Dissections Leeches were pinned out, dorsal side up, in ice-cold leech saline (35), and a dorsal midline cut was made to expose the nerve cord or hearts. For biochemical experiments, lateral nerve roots were cut close to each segmental ganglion and the anterior and posterior brains were carefully dissected from the ventral blood sinus. Nerve cords, including anterior and posterior brains and 21 segmental ganglia, were removed by bisecting and then pulling them out of the blood sinus at either end. For peptide release experiments, heart tubes were removed with fine scissors and incubated in normal leech saline for I0 rain before use.

Tissue Extraction Immediately after removal, nerve cords were immersed in acidified methanol solution [90% methanol, 9% glacial acetic acid, 1% water, 1 mM dithiothreitol, - 1 0 ° C , modified from (28)] and stored at - 1 0 ° C . Nerve cords were homogenized using a straight wall glass grinder, and centrifuged at 15,000 x g for 10 min at room temperature. The supernatants were removed, and the pellets resuspended in 100% methanol. The resuspended pellets were recentrifuged, and the supernatants from both centrifugations were pooled. The pellets were discarded. Samples were next concentrated in a Speed Vac concentrator (Savant) and water added to give a volume of 1 ml. This volume was then loaded onto a C18 Sep-Pak cartridge (Waters) which had been equilibrated with 5 ml each of methanol and water. For large samples (100 nerve cords or greater), two SepPak cartridges were used. Samples were desalted with 15 ml water and eluted with 10 ml methanol. The pooled methanol eluant was collected, concentrated in a Speed Vac concentrator,

and water (Fisher) was added to give the appropriate volume (100-500 ixl).

Tracer lodination YMRFamide was used as the tracer (~25I-YMRFamide) in RIA (28,33). A modified chloramine T method (17, 28, 41) was used to iodinate the peptide. The reaction product was initially desalted with 30 ml water on a preequilibrated C18 Sep-Pak cartridge, and eluted with 5 ml methanol. The methanol fractions were partially dried in a Speed Vac concentrator, combined, and water added to give a volume of 400-500 Ixl. This solution was then loaded onto a 3.9 mm x 30 cm C18 rpHPLC column, eluted at 1 ml/min using a gradient of water/acetonitrile (see High Pressure Liquid Chromatography under the Method section), and 1 ml fractions collected. Radioactivity [counts/rain (cpm)] in each fraction was detected on a Packard RIAStar 5410 gamma counter. Fractions with significant radioactivity above background were tested for maximum binding (Bo) and precipitation control in RIA. The fraction used for tracer was precipitated well by dextran-coated charcoal, and showed a large difference between this precipitate control and B o. Tracer was concentrated in a Speed Vac concentrator, diluted with RIA buffer to - 1 x 10 6 cpm/ml, and stored at - 10°C until use.

Radioimmunoassay The RIA for FMRFamide using antiserum 671C has been previously described and characterized (33). Each assay tube contained 5000-9000 cpm of tracer diluted in RIA buffer (0.1% bovine serum albumin, 0.2% sodium azide in 0.1 M phosphate buffer, pH 7.5) and 100 ILl of a dilution of sample or an amount (2.5 fmol to 5.0 pmol) of one of the following synthetic peptides. Peptides used included FMRFamide (Bachem or donated by D. A. Price); YMRFamide, pQDPFLRFamide (Bachem); F L R F a m i d e (Sigma); F M R F ( P e n i n s u l a ) ; Y L R F a m i d e , GGKYMRFamide, GGKYMRF (Microchemical Facility, Emory University). Antiserum 671C (50 ill diluted in RIA buffer) was added to each tube to give a final antiserum concentration of 1:10,000 after all reagents had been added. Total volume in each tube was 200 I~1. Tubes were then vortexed briefly and stored at 4°C for 18-24 hours. Unbound tracer was precipitated by the addition of 100 ixl dextran-coated charcoal suspension [20 mg activated charcoal (Sigma)/5 mg dextran (Kodak)/ml RIA buffer]. Tubes were vortexed, centrifuged (15,000xg) for 10 min at 4°C, and 200 I~1 of supernatant removed. Both precipitate and supernatant were counted for one rain on a gamma counter, and a weighted ratio {[1.5A/(A + B)] x 100, where A = c p m in 200 ~1 supernatant, B = cpm remaining in assay tube} used to determine the percentage of tracer bound by each sample. Duplicate tubes were used for unknowns and standards. Samples were diluted to adjust the amount of immunoreactivity in a sample to an amount which would lie along the linear portion of the standard curve of FMRFamide dilutions. The minimum detectable amount of FMRFamide was 20-50 fmol.

High Pressure Liquid Chromatography Reverse phase high pressure liquid chromatography (rpHPLC) was performed on a Rainin system controlled by an Apple Macintosh Plus computer using a Dynamax HPLC Method Manager program. HPLC grade solvents and HPLC or sequenal grade counterions were used throughout the purification process. Solvents were delivered at 1 ml/min by two Rainin HP pumps equipped with 5 ml pump heads, and mixed with a Dynamax high pressure mixer. Samples (100-500 ILl) were injected via a

RFamide PEPTIDES IN LEECH

Rheodyne 7125 sample injection valve (1 ml sample loop). One-ml fractions were collected with a Pharmacia FRAC-100 fraction collector. Two columns were used with this system. The first was a 3.9 mm x 30 cm C18 i.tBondapak (10 lJ.m particle size, Waters). The second column was a 4.6 mm × 10 cm C18 Microsorb (3 tam particle size, Rainin). Two chromatographic separation protocols were used on each of the two columns. The v.Bondapak column was used for initial separation of nerve cord extracts, and also for all separations involving radioactive tracer (see Radioimmunoassay under the Method section). Desalted nerve cord extracts or peptide standards were injected in 100-500 V.1 water and eluted with the following solvent gradient. Solvent A was 0. I% trifluoroacetic acid (TFA), v/v, in water; solvent B was 0.1% TFA, v/v, in acetonitrile. Initial conditions were 5% B; a linear gradient to 10% B over 5 min was followed by a linear gradient to 45% B at 45 min, then to 50% at 45.5 min and isocratic elution (50% B) for 4.5 min. Radioactive tracer separations were performed under identical conditions except 0.1% heptafluorobutyric acid (HFBA) replaced TFA as the counterion in both solvents. Aliquots (10-20 ~1) of each HPLC fraction were removed for RIA and diluted with RIA buffer. Immunoreactive fractions, as indicated by RIA, were pooled and concentrated in a Speed Vac concentrator. Water was added to achieve the appropriate injection volume. The Microsorb column was used for two rpHPLC separations of nerve cord extract (under both reduced and oxidized conditions, see next paragraph). In each case, the elution gradient used the same solvents. Solvent A was 0.1% HFBA, v/v, while solvent B was 0.1% HFBA, v/v, in acetonitrile. Peptides were eluted with a linear gradient from 5% B to 60% B over 70 min, followed by an increase to 80% B over 16 min. Immunoreactive fractions from rpHPLC separations of nerve cord extracts were separated twice on the Microsorb column according to this protocol. Immunoreactive fractions from the first of these two separations were evaporated almost to dryness, diluted with water, and subjected to mild oxidation prior to being rechromatographed. In order to convert methionine in the methionine-containing peptides to methionine sulfoxide [Met(O)], immunoreactive fractions were treated with hydrogen peroxide (1.5%, 1.5-2 hours) and then rechromatographed separately under the same conditions. In comparison to the unoxidized peptides, the Met(O)containing peptides, due to their increased polar character, eluted earlier in the gradient (Figs. 3B and 4). To demonstrate that immunoreactivity from nerve cord extracts had the same retention time as a synthetic peptide on one of the two columns described above, extracts and synthetic peptides were separately chromatographed within 24 h of one another. Between each chromatographic separation, the column used was washed by pumping solvents in a recurring cycle from 100% solvent B to 0% solvent B and back to 100% solvent B (4 cycles over 45 min at 1.5 ml/min), and then equilibrated for 15 min at 5% solvent B. Final purification of peptides was done by microbore rpHPLC on an Applied Biosystems Micro Separation System equipped with a microbore LC flow cell and 0.005" i.d. tubing. An Aquapore OD-300 C18 silica column (1 mm x 25 cm, Applied Biosystems) was equilibrated at 30°C in solvent A (0.1%, v/v, TFA) and developed isocratically at 90 p,/min for 10 min, followed by a linear gradient (0.5%/rain) of solvent B (acetonitrile/ water/TFA, 4:1:0.004, v/v/v). To assure ion-pairing with TFA, samples were injected in 1% (v/v) TFA. Material absorbing at 214 nm was manually collected, assayed for immunoreactivity, and stored at - 2 0 ° C . Immunoreactive fractions were sequenced.

899

TABLE 1 AMOUNTSOF IDENTIFIEDRFamidePEPTIDESAND TOTAL FMRFamide-LIKEIMMUNOREACTIVITYPRESENTIN LEECH NERVECORD EXTRACTS Peptide

N

#

YMRFamide FMRFamide YLRFamide FLRFamide GGKYMRFamide Total FLI

4 4 4 4 4 5

600 600 600 600 600 640

pmol/n.c. (mean --- SE) 3.15 5.76 4.45 4.51 6.72 22.72

--- 0.97 _+ 1.30 +_ 1.38 __- 1.43 __- 2.26 __- 3.54

Recovery 38.8% 58.0% 41.6% 41.0% 10.4%

The amounts of each identified RFamide peptide shown were determined by RIA after the third rpHPLC purification step and corrected for the cross-reactivity of each peptide with antiserum 671C in RIA (Fig. 1) and recovery from rpHPLC (listed in column on far right; see the Method section for computation). The value for total FLI was determined by RIA immediately after extraction and desalting, and recovery was assumed to be 100%. N denotes the number of times each determination was made; # denotes the total number of nerve cords used in all determinations. Abbreviations: FLI=FMRFamide-like immunoreactivity; n.c.=nerve cord.

Peptide Identification by Automated Edman Degradation and Mass Spectrometry Each rpHPLC-purified peptide was subjected to twelve cycles of Edman degradation on an Applied Biosystems Model 477A Protein Sequencer with on-line Model 120A PTH (phenylthiohydantoin) Analyzer using a described protocol (40). Plasma desorption mass spectrometry (PDMS) of rpHPLC-purified peptides was performed by Dr. Ling Chen at Applied Biosystems (Foster City, CA).

Estimation of Peptide Quantities in Nerve Cord Extracts The quantity of FLI per nerve cord was measured after the desalting step. RIA was performed on aliquots of nerve cord extracts and multiplied by the dilution factor to determine the amount of immunoreactivity present in each extract. This value was then divided by the number of nerve cords used to prepare that sample to determine the immunoreactivity per nerve cord. The quantity of each immunoreactive peptide found in nerve cord extracts was determined after the third (after oxidation) rpHPLC separation. Immunoreactivity from nerve cord extracts with the same retention time as each of the five synthetic RFamide peptides was measured in RIA, and these values were corrected for the cross-reactivity of each putative peptide in RIA (Fig. 1) to determine the quantity of each immunoreactive peptide present in each peak. These values were further corrected for percent recovery during the purification process and normalized per nerve cord (Table 1). Percent recovery for each peptide was determined after the first three rpHPLC purification steps. Known amounts of synthetic peptide were subjected to these purification steps and the amount recovered from the process as determined by RIA was divided by the original quantity and multiplied by 100 to determine percent recovery for each peptide. The average percent recovery of three replicates was used to estimate the recovery of each endogenous leech RFamide peptide.

Release of FMRFamide-Like Immunoreactivity From Neural Processes on Heart Tissue Dissected hearts with their neural processes were constantly superfused with various solutions designed to elicit or inhibit the

900

EVANS, POHL, KARTSONIS AND CALABRESE

release of neurotransmitters from nerve terminals. The apparatus and procedure were similar to those described previously (1). Twenty ml of each of three solutions were peristaltically pumped through a small chamber containing leech heart tubes and then through a C18 Sep-Pak cartridge. A new Sep-Pak was used for each of the superfused solutions to collect the immunoreactive material released by each step of the process. Peptides were desalted with 3 ml water and eluted from the Sep-Pak with 1 ml methanol. The methanol eluants were partially dried, reconstituted in RIA buffer, and the amounts of FLI determined by RIA. The three solutions in the order used in these experiments were: A) normal saline (1.8 mM CaC12, 4.0 mM KC1, 115 mM NaC1); B) high potassium saline with Co + +/EGTA [ethylene glycol-bis-(B-aminoethyl ether) N,N'-tetra-acetic acid] (0 CaCI 2, 100 mM KC1, 14.8 mM NaC1, 5.0 mM CoC12, 2.0 mM EGTA); C) high potassium saline (1.8 mM CaCI2, 100 mM KCI, 19.0 mM NaC1). All three solutions also contained 10 mM glucose and 10 mM Tris-HC1, and were adjusted to pH 7.4.

RESULTS Characterization of the Radioimmunoassay Several synthetic peptides had previously been used to characterize the specificity of antiserum 671C in RIA (33). Antiserum 671C reacts with peptides of the form X-Met-Arg-Pheamide and X-Leu-Arg-Phe-amide. Since peptides of these two forms have been described in CNS extracts of other invertebrates [see (5,16) for reviews], we decided to use this antiserum for all RIAs in this study. After five RFamide peptides from nerve cord extracts of Hirudo were identified, the specificity of antiserum 671C in RIA was further characterized by testing six synthetic peptides for inhibition of tracer binding (Fig. 1). YMRFamide (135% cross-reactivity), FMRFamide (100% cross-reactivity), and GGKYMRFamide (75% cross-reactivity) inhibited tracer binding more efficiently than either FLRFamide (45% cross-reactivity) or YLRFamide (40% cross-reactivity). As previously shown (33), FMRF displayed no significant cross-reactivity with 671C (<0.0009%). Nerve cord extracts showed similar displacement behavior to synthetic FMRFamide in RIA (Fig. 2), as the slopes of the two inhibition of binding curves were similar (-0.577+--0.041 for nerve cord extracts, and - 0 . 5 3 4 ± 0 . 0 2 3 for FMRFamide). This observation indicates that this RIA can be used effectively for quantification of FLI in nerve cord extracts.

Characterization of FMRFamide-Like Immunoreactivity by HPLC The first rpHPLC separation of nerve cord extracts showed that the majority of the immunoreactivity present eluted in a broad band from fractions 19-31, and that synthetic FMRFamide eluted within this range (Fig. 3A). Three other immunoreactive peaks were detected. Two of these peaks, fraction 16 and fractions 38-39, were not always present in similar chromatograms and were not pursued further. A third peak (fractions 44 and 45) was reproducibly present and was rechromatographed, eluting in fractions 68-70 on the second system (Fig. 3C). No further characterization of this material was possible due to substantial losses (>95%) during purification. The major immunoreactive band from Fig. 3A (fractions 1931) was further purified on the second system as shown in Fig. 3B. Four major and two minor peaks of immunoreactivity were detected in this chromatogram. The first minor peak (fraction 34) had the same retention time as the sulfoxide form of YMRFamide, and the second (fractions 37-38) had the same retention

100 -

i ~ i ~ - ~ - - - - _ a .----.---"II ~ m

~

ai

0 0

%

50- 0 FMRFarnide

1:13 t213

X~O\\\

~'X~.

g GGKYMRFomide

• YLRFamide

x ~ X

• FLRFamide • FMRF -12

~~lb~...li,~ -11

-10

-g -8 -7 LOG CONC (M)

~ -6

-5

-4

FIG. 1. Specificity of anti-FMRFamide antiserum 671C. Cross-reactivity of five RFamide peptides found in Hirudo and the nonamidated peptide FMRF. Tracer was 12SI-YMRFamide. Cross-reactivity for each peptide (except FMRF) was measured at the concentration at which B/Bo × 100 = 50 (75 for FMRF). Each data point represents the average of two replicates, except FLRFamide, which is the average of six. Only YMRFamide showed greater cross-reactivity (135%) than FMRFamide (100%), while GGKYMRFamide (75%), FLRFamide (45%), YLRFamide (40%), and FMRF (<0.0009%) all showed less. The abscissa shows the log of the molar concentration for each peptide. Abbreviations: B = specific tracer binding; Bo = maximum specific tracer binding.

time as the sulfoxide form of FMRFamide. To further purify the immunoreactive peptides, we oxidized each of the four major immunoreactive peaks (Fig. 3B) with hydrogen peroxide, and rechromatographed them under the same conditions. Since the retention time of methionine-containing

Dilution 1/4oo 1/2oo 1/loo '~/5o 1/25 lO0

o x mo

1:/3

i

I

i

t

h

NERVE CORD ~

~

EXTRACTS

FMRFamide " ~

~ 0

10 . . . . . , 40 IO0

.

.

.

.

.

.

.

.

• ,

1000 frnol

FIG. 2. Displacement of tracer binding by nerve cord extracts and synthetic FMRFamide in radioimmunoassay. The dilution of nerve cord extract is indicated along the upper X axis as a fraction of one nerve cord. The quantity of FMRFamide is indicated along the lower X axis. Serial dilutions of both samples show parallel inhibition of binding curves, thus indicating that antiserum 671C reacts with antigens from the leech nerve cord and synthetic FMRFamide in similar fashion. Curves were fitted using least squares regression analysis. The slope of the nerve cord extract line is -0.577---0.041 (S.E.), while the line for FMRFamide is -0.534___0.023 (S.E.). Maximum specific binding was 49.0%. Abbreviations: B-specific tracer binding; Bo=maximum specific tracer binding.

RFamide PEPTIDES IN LEECH

9ol

A. FblRFomidl

(5v

_v

'~

400-

.80

300-

60

200 -

40

100

::il,l;ili

20

o ................. 10

2O

~

40 "

30

"-r

' 0

FRACTION NUMBER

B.

C.

_? 0

500-

"5

A

25.

D

1

300-

g

............................

60

.

' 60

,.......................•....

40

15.

E

o

-20

I00

0

II

.... •

200m~

0

80

400-

10

20

30

40

50 FRACTION NUMBER

60

-0 70

5lO

20

30

40

50

60

70

FRACTIONNUMBER

FIG. 3. Chromatography of nerve cord extracts by rpHPLC. Immunoreactivity in each fraction is indicated by filled bars, and the acetonitrile/water gradient is shown by dashed lines (see the Method section for conditions)• (A) The desalted extract from 100 nerve cords was chromatographed on a txBondapak C18 column with 0.1% TFA as counterion. The majority of immunoreactivity eluted in fractions 19-31. Synthetic FMRFamide eluted in fraction 25 in similar chromatograms, as indicated by the arrow. A minor immunoreactive peak eluted in fractions 44 45 (*). Fractions 38-39 also contained a small immunoreactive peak which was not further characterized (see the Results section). (B) Fractions 19-31 from (A) were combined and rechromatographed on a Microsorb C18 column with 0.1% HFBA as counterion. Four major immunoreactive peaks, A, B, C, and D, were recovered. Each of the four peaks had the same retention time as one of the following synthetic peptides: A, YMRFamide; B, FMRFamide; C, YLRFamide; D (fraction 50), FLRFamide. (C) Fractions 43-45 from (A) were combined with similar fractions from other nerve cord extract separations and rechromatographed on a Microsorb C18 column with 0.1% HFBA as counterion. Note the difference in retention time between synthetic pQDPFLRFamide (45 min, arrow) and the immunoreactivity (68-70 min). peptides decreases upon oxidation, further purification is achieved. A similar approach has been used to purify FMRFamide from Nereis (23) and to facilitate chromatographic separation of the closely migrating rat peptides insulin I and II (20). Fraction 41 from Fig. 3B, which had the same retention time as unoxidized synthetic YMRFamide, also had the same retention time after oxidation (Fig. 4A, fraction 34, peak A) as oxidized synthetic YMRFamide. Similarly, fraction 46 from Fig. 3B now eluted in fraction 38 (Fig. 4B, peak B), as did oxidized synthetic FMRFamide. The retention times of both fraction 44 from Fig. 3B and synthetic YLRFamide were unaffected by oxidation (Fig. 4C, peak C). Fraction 49-50 from Fig. 3B was resolved into two major immunoreactive peaks (Fig. 4D, peaks D1 and D2) after oxidation. Peak D1 and synthetic FLRFamide were unaffected by the oxidation step (Fig. 4D). Peak D2, however, eluted earlier (fraction 43) after oxidation and had the same retention

time (43 min) as the oxidized form of synthetic GGKYMRFamide (Fig. 4D). Final purification of the immunoreactive fractions from Fig. 4 was achieved by microbore rpHPLC. Each of the five immunoreactive peaks obtained from the third HPLC separation (Fig. 4) was purified individually prior to peptide sequencing. Figure 5A shows the final purification of peak A from Fig. 4A. The peaks of absorbance (214 nm) were each tested for immunoreactivity; only the absorbance peak indicated by the arrow was immunoreactive. Peaks B, C, D1 and D2 from Fig. 4 B - D were purified in a similar fashion, and again only one immunoreactive absorbance peak was found in each chromatogram, as indicated by the arrows (Fig. 5B-D2).

Identification of HPLC-Purified RFamide Peptides The purified immunoreactive peaks, as indicated by arrows

902

EVANS, POHL, KARTSONIS AND CALABRESE

B.

A. "5

FMRFomidl

2ooi

yMRForniOo

f 80

80

4-00 -

v _.e "B

b..

1501

A~

lOO I

soo-

.............................. 60 ............................................

4.o

sol ..............'""""'""•"-"'"i

200 I00-

20

0

10

20

30

40

50

60

70

C.

10

.................................

2'0

~

30

JL

-.

_ ,

40

.

,

50

60

.

,

70

D. YLRFamlda

FLRFomide

1 O0 -

-80

60

.80 BO

el

C

o> '5

ioo -,~o 12o

B

.. .............-."" ........

GGK~MRF~mido

-60

91

6o

40

2 ....

- 40 2O

o 10

2~

30

4o

so

o~

-20

2O

o

0

70

FRACTION NUMBER

..... 6o

D2

2o ,

1o

20

, 30

.

t

. 40

50

I 60

70

FRACTION NUMBER

FIG. 4. Chromatography of oxidized immunoreactive fractions from Fig. 3B by rpHPLC. Immunoreactivity in each fraction is indicated by filled bars, and the acetonitrile/water gradient is shown by dashed lines (see the Method section for conditions). Arrows indicate the elution position of synthetic standards. (A) Fractions 41 and 42 from Fig. 3B eluted as a single peak (A), which had the same retention time (34 rain) as synthetic YMRFamide. (B) Fractions 45--47 from Fig. 3B eluted as a single peak (B), which had the same retention time (38 min) as synthetic FMRFamide. (C) Fraction 44 from Fig. 3B eluted as a single peak (C), which had the same retention time (44 min) as synthetic YLRFamide. (D) Fractions 48-51 from Fig. 3B were separated into two major immunoreactive peaks, DI and D2, which had the same retention time as FLRFamide (50 min) and GGKYMRFamide (44 rain), respectively.

in Fig. 5, were subjected to automated Edman degradation. Five successive cycles of degradation are shown for Fig. 6A-D1, and eight cycles for Fig. 6D2. In Fig. 6A-D1 the most prevalent PTH amino acid for each of the first four cycles is clearly distinguishable from background amino acid levels. The fifth and succeeding cycles showed no PTH amino acids present above background. In some of the cycles, minor amounts of contamination were present. Figure 6D2 shows the sequence data for peak D2 from Fig. 5D2. In each of the first seven cycles of degradation, one PTH amino acid per cycle can easily be distinguished from background levels. Each of the RFamide peptides (peaks A-D2) was sequenced at least twice and similar results were obtained each time. In one of the two sequence runs of purified peak D2, the sixth amino acid residue was not detected, but was clearly shown to be arginine in the other sequence run of this peak. Low recoveries of PTH-arginine may be due to variable efficiencies of extraction from the sequencer (2). The specificity in RIA of antiserum 671C for the amidated forms of FMRFamide-like peptides (Fig. 1) indicates that the peptides we have detected by RIA for sequencing are amidated. The sequence data thus indicate that peaks A, B, C, D1, and D2, respectively, can be identified as the peptides YMRFamide,

FMRFamide, YLRFamide, FLRFamide, and GGKYMRFamide. Each of these five peptides had the same retention time on rpHPLC as its synthetic counterpart under reduced and oxidized conditions (Fig. 3B, 4; data for reduced GGKYMRFamide not shown). In addition, the nonamidated synthetic peptide FMRF was chromatographed together with synthetic FMRFamide (Fig. 7A); synthetic FMRF was also chromatographed together with purified FMRFamide from leech nerve cord extracts (Fig. 7B). In each chromatogram, two absorbance peaks were detected, only one of which was immunoreactive as determined with RIA. These data indicate that the amidated form of this peptide is clearly distinguishable from the nonamidated form on our rpHPLC system, but the amidated form is the only one which can be detected in our RIA. Similar results were obtained using synthetic GGKYMRF and GGKYMRFamide (data not shown). Plasma desorption mass spectrometry (PDMS) analysis of two of the purified RFamide peptides was performed to determine their molecular masses, rpHPLC-purified FMRFamide (sulfoxide form, Fig. 8) had a mass of 615.7, virtually identical to the calculated mass of 615.78. Similarly, rpHPLC-purified FLRFamide had a mass of 581.6 (Fig. 9), with a calculated mass of 581.74. The nonamidated forms of these peptides (FMRF and

R F a m i d e P E P T I D E S IN L E E C H

903

BII

AI

TI:

E C 0.03-

C

/

C 0.02"

I//

0

£

0°00o 40

0.070.08 0.05

0.04 •"

1.20

2O

0.03

o 0.02

o

Lfloo

0,01 0.00,

0 tO

20

0.00

,o

0

30

Time (minutes)

0.07

0.04- -

r 0,06 -

40

A 0.05 •

,2o

0.03 30

.~. 004

i-

0.02 i

40

D1.

CI

q¢

2'o ~'o Time(minutes)

20

0.03 -

//

0.02 lO

O.Ol

o

o.oo

lO

o.ol o.oo

,o

2'o

~'0

40

,

o

Time (minutes)

10

20

Time

•

,

30

•

,

J

0

40

(minutes)

D2. ,40

o

• =o . ~ ¢-

o

I0

i = xi

Time(minutes) FIG. 5. Chromatography of the immunoreactive fractions from Fig. 4 by rnicrobore rpHPLC. Each absorbance peak (214 nm) was collected manually and assayed for immunoreactivity. In each case, only one immunoreactive peak was detected (arrows). The acetonitrile/water gradient is indicated by dashed lines (see the Method section for conditions). (A) The immunoreactivity in fraction 34 from Fig. 4A had a retention time of 28.2 minutes (arrow). (B) The immunoreactivity in fraction 38 from Fig. 4B had a retention time of 31,0 minutes (arrow). (C) The immunoreactivity in fraction 44 from Fig. 4C had a retention time of 34.9 minutes (arrow). (D1) The immunoreactivity in fraction 50 from Fig, 4D had a retention time of 38.5 minutes (arrow). (D2) The immunoreactivity in fraction 43 from Fig. 4D had a retention time of 29.1 minutes (arrow).

904

EVANS, POHL, KARTSONIS AND CALABRESE

A.

B. O Gin

Met

80

O Leu

Phe •

g .o 6o

vk~= 0 0

:et

100-

Asn Gkl

oE ~0 50.

o 20

•

0

1

2

•

3

4

5

C. ~o • 15o

1

2

Phe

Leu

I

2

3

4

5

D1 Tyr

Leu

Z~ Ala V lie

•

E] Gly

B0.

o

60

E 100 40 ~ 50 "6

20

I

2

3

,$

0

5

Cycle number

3 Cycle number

h 4

5

D2. 4O

Gly •

.o 3o

O Gin V Set

Met Cly

Lys

2

3

Tyr

2o

4

.5

6

7

8

Cycle number

FIG. 6. Amino acid sequences of the five immunoreactive peaks obtained in Fig. 5. Data points indicate the amount of each PTH amino acid recovered from each successive cycle of Edman degradation. The major amino acid at each cycle is shown by its three-letter designation. Contaminating amino acids are as indicated in the upper right comer of each graph. (A) Sequence of peptide A from Fig. 5A. (B) Sequence of peptide B from Fig. 5B. (C) Sequence of peptide C from Fig. 5C. (D1) Sequence of peptide D1 from Fig. 5DI. (D2) Sequence of peptide D2 from Fig. 5D2.

FLRF) would have calculated masses of 616.78 and 582.74, respectively. No ions with these masses were detected above background levels in either analysis. The mass spectra provide physical evidence that these two peptides are amidated. Release of FLI From Neural Processes on Heart Tubes Heart tubes were superfused with a series of three solutions (A, B, C) which varied in K +, Ca + ÷ , Co + +, and EGTA concentrations. The experiments were designed to test whether neural processes on the hearts would preferentially release neu÷+ rotransmitters upon depolarization in a Ca -dependent manner when supeffused with solution C. The amount of FLI released in the presence of each solution was quite variable (Fig. 10), but paired t-tests show that a significant difference exists be-

tween the amount of FLI released in the presence of solution C and either solution A or B. When the concentration of K ÷ was raised to 100 mM in the presence of Co + ÷ and EGTA (solution B), the amount of FLI released was not significantly different from the amount released in the presence of 4 mM K + , 0 Co ÷ + , 0 Ca + +, 0 EGTA (solution A) (0.15
Using RIA and a four-step rpHPLC separation followed by peptide sequencing, we have isolated and identified five RF-

RFamide PEPTIDES IN LEECH

905

A.

12000-

•~=~ ~, 0.4.

0 t"q ~E tJ"3 t.~

FMRFamide

40 ~.

0.3

R S z

z

r-: tn

~ 0.2 20

Co

800O Z 0

OJ I--

2

0.0

4O00

B. 0.4

40

0.3

o

z m~ 0 . 2

800

0

,

10

20

30

11 O0

M/Z

0.1

0.0

800

J

40

TIME (mlnules)

FIG. 7. Coelution of FMRFamide from Hirudo medicinalis with synthetic FMRFamide and not with synthetic FMRF. An Aquapore C18 column was used and each absorbance peak was collected and assayed for immunoreactivity. In each chromatogram, only one immunoreactive peak was eluted, as indicated by arrows. The dashed lines indicate the acetonitrile/water gradient. All peptides were oxidized prior to injection. (A) Coinjection of synthetic FMRFamide and synthetic FMRF. FMRFamide eluted at 30.8 minutes; FMRF eluted at 32.4 minutes. (B) Coinjection of leech FMRFamide and synthetic FMRF. Leech FMRFamide eluted at 30.8 minutes, just as synthetic FMRFamide did in (A). Synthetic FMRF did not comigrate with either synthetic or leech FMRFamide.

amide peptides in extracts of leech nerve cord: YMRFamide, FMRFamide, YLRFamide, FLRFamide, and GGKYMRFamide. Carboxyl terminal amidation was verified by antibody specificity and, in the cases of FLRFamide and FMRFamide, structural proof of amidation was obtained by mass spectrometry. A very hydrophobic FMRFamide-like immunoreactive substance is also present in nerve cord extracts (fractions 44 45 in Fig. 3A). It is possible that this molecule is an RFamide peptide precursor which has not been fully processed to release the peptide(s). The specificity of antiserum 671C in RIA suggests that at least one amidation site has been enzymatically created from an Arg-Phe-Gly sequence in the precursor. Other RFamide peptides not identified in this study may also be present in leech nerve cord. Figure 3 shows a peak of FLI (fractions 38-39) not detected in other similar chromatograms. This immunoreactivity may be due to a peptide which is easily lost during extraction or chromatography, or it may represent a peptide that is preferentially degraded in most animals and therefore is not usually present in nerve cord extracts. Isolation of larger quantities of the peptide will be necessary to pursue its identity. Analysis of RFamide peptide genes in Hirudo medicinalis may suggest the identity of one or more as yet unidentified RFamide peptides in leech.

FIG. 8. Mass spectrogram for rpHPLC purified Pbe-Met(O)-Arg-Pheamide. The abscissa shows the mass/charge ratio of the ionized molecules in the sample. The ordinate shows the number of positive ions which were detected• The mass of the major molecular species is 615.7. This peak corresponds to the calculated mass of FMRFamide (sulfoxide form). A minor peak was also detected at 599.6 (M/Z), which is the calculated mass of reduced FMRFamide. Abbreviations: M=mass; Z = charge.

A Comparison of RFamide Peptides in Annelids and Molluscs FMRFamide has now been identified in two annelids and numerous molluscs. There are striking similarities between the

3000

FLRFamide

To

2000

to

1000

•

500

I

•

i

,

800

i

-

!

•

i

•

i

•

!

)

11 O0

M/Z FIG. 9. Mass spectrogram for rpHPLC-purified Phe-Leu-Arg-Phe-amide. The abscissa shows the mass/charge ratio of the ionized molecules in the sample. The ordinate shows the number of positive ions which were detected. The mass of the major molecular species is 581.6. This peak corresponds to the calculated mass of FLRFamide. Abbreviations: M = mass; Z = charge.

906

EVANS, POHL, KARTSONIS AND CALABRESE

Variations Among Leech RFamide Tetrapeptides

1200 o

E 1000-

p ( 0.05

800-

g 600E 400. ,= 200"

Solution A 1,8 C a + + 4.0 K + 115 N a +

Solution B 0 Ca ++ 100 K + 14.8 N~ + 5.0 Co ++ 2,0 EGTA

Solution C 100 K + 19.0 Na +

FIG. 10. FMRFamide-like immunoreactivity released from neural processes on heart muscle during consecutive exposure to three different bathing media. Solution A was normal leech saline; solution B contained 0 Ca + +, a high concentration (100 mM) of K + , and also contained 5.0 mM Co ÷+ and 2.0 mM EGTA; solution C contained a normal (1.8 raM) level of Ca + +, high K + , and no Co-* ÷ or EGTA. The data represent the results of 6 different experiments. The bars indicate standard arrows. Paired t-tests indicate that A and B are not significantly different (0.15
The four RFamide tetrapeptides isolated from Hirudo medicinalis contain either phenylalanine or tyrosine at the amino terminal, and leucine or methionine as the next amino acid. Only one point mutation is necessary to change a methionine to a leucine residue, or a phenylalanine to a tyrosine. The combinations of two point mutations in a primitive RFamide peptide gene could account for the variations seen in leech RFamide tetrapeptides. Thus the four RFamide tetrapeptides identified in Hirudo possibly derive from a common genetic sequence which encoded one of the four peptide sequences. The repetition of tetrapeptide sequences in the RFamide peptide genes in Aplysia (53) and Lymnaea (29) suggests the possibility that multiple RFamide peptide sequences may be encoded by an RFamide peptide gene in Hirudo.

Calcium-Dependent Release of RFamide Peptides Proctolin is released from neural processes on insect muscle in a Ca + +-dependent manner upon high K ÷ depolarization (1). In similar experiments, we have shown that FMRFamide-like immunoreactive substances are released from neural processes on heart muscle. When these preparations were depolarized with high potassium solutions without Ca +÷ , but with the calcium chelator EGTA and the Ca +÷ channel blocker Co +÷ , no increase in detectable FLI release was seen between this solution and normal saline. A significant increase in release of FLI was detected, however, when depolarizing saline with normal levels of Ca + ÷ and no chelators or channel blockers were used. These data indicate that Ca ÷ ÷ is necessary to cause an increase in FLI release upon depolarization, which is consistent with the hypothesis that FMRFamide and related peptides act as neurotransmitters in Hirudo.

RFamide Peptides as Putative Neurotransmitters in the Leech Five RFamide peptides in Hirudo medicinalis can now be tested as possible neurotransmitters of identified motor neurons, with a knowledge that the peptides being tested are endogenous. The bioactivity of the untested endogenous leech RFamide peptides may be different from that exhibited by synthetic FMRFamide. For example, YMRFamide is more potent on longitudinal muscle than FMRFamide or FLRFamide (38). N-terminal extensions of FLRFamide or FMRFamide elicit longitudinal muscle contractions in the leech equal to (YGGFMRFamide) or greater than (pQDPFLRFamide) those elicited by FMRFamide (38). Thus it is possible that GGKYMRFamide may show a significant difference from FMRFamide in potency on leech muscle. GGKYMRFamide may even act in an opposite manner to FMRFamide, as is the case for YGGFMRFamide in some molluscan muscle preparations, including rectum of the clam Mercenaria (15). The identification of five RFamide peptides in nerve cord extracts of Hirudo medicinalis provides a firm foundation for further investigations concerning the neuronal distribution of each RFamide peptide, and what role each one plays at specific synapses, such as cardiac and longitudinal neuromuscular junctions. Activities requiring fast or strong movements, such as swimming, may depend on peptide modulators to increase the frequency or strength of muscular contractions. Electrical stimulation of cell 204, a swim-initiating interneuron which contains FLI, leads to acceleration of the heartbeat central pattern generator, as does bath-applied FMRFamide. Cell 204's action on heartbeat thus may be due to the central release of RFamide peptides

RFamide PEPTIDES IN LEECH

907

(3). RFamide peptides may act in conjunction with serotonin, which is released centrally from Retzius neurons during swimming (56) and also accelerates the heartbeat central pattern generator (3). RFamide peptides released by motor neurons containing FLI may also play a role in modulating longitudinal muscle contractile strength during swimming or other behaviors due to a decrease in the relaxation rate of neurally evoked contractions in response to RFamide peptides (38). If RFamide peptides are, in addition, blood-borne factors, a coincident increase in heart rate

would facilitate the distribution of these peptides or their sites of action. ACKNOWLEDGEMENTS We wish to thank Dr. David Price for donation of FMRFamide and helpful discussions, Drs. Mark Brown and John McDonald for helpful comments on HPLC and ILIA techniques, and Dr. Patricia Whitten for use of her gamma counter. We are especially grateful to Dr. Ling Chen for mass spectrometric analysis of peptides.

REFERENCES 1. Adams, M. E.; O'Shea, M. Peptide cotransminer at a neuromuscular junction. Science 221:286--289; 1983. 2. Applied Biosystems Protein Sequencer User Bulletin 21:3; 1986. 3. Arbas, E. A.; Calabrese, R. L. Rate modification in heartbeat central pattern generator of the medicinal leech. J. Comp. Physiol. 155:783-794; 1984. 4. Calabrese, R. L.; Maranto, A. R. Neural control of the hearts in the leech, Hirudo medicinalis. III. Regulation of myogenicity and muscle tension by heart accessory neurons. J. Comp. Physiol. [A] 154:393--406; 1984. 5. Cottrell, G. A. The biology of the FMRFamide-series of peptides in molluscs with special reference to Helix. Comp. Biochem. Physiol. [A] 93:41-45; 1989. 6. Cottrell, G. A.; Greenberg, M. J.; Price, D. A. Differential effects of the molluscan neuropeptide FMRFamide and the related Met-enkephalin derivative YGGFMRFamide on the Helix tentacle retractor muscle. Comp. Biochem. Physiol. [C] 75:373-375; 1983. 7. Cowden, C.; Stretton, A. O.; Davis, R. E. AF1, a sequenced bioactive neuropeptide isolated from the nematode Ascaris suum. Neuron 2:1465-1473; 1989. 8. Cowden, C.; Stretton, A. O. W. AF2, a nematode neuropeptide. Soc. Neurosci. Abstr. 16:305; 1990. 9. Dockray, G. J.; Reeve, J. R., Jr.; Shively, J.; Gayton, R. J.; Barnard, C. S. A novel active pentapeptide from chicken brain identified by antibodies to FMRFamide. Nature 305:328-330; 1983. 10. Ebberink, R. H.; Price, D. A.; vanLoenhout, H.; Doble, K. E.; Riehm, J. P.; Geraerts, W. P.; Greenberg, M. J. The brain of Lymnaea contains a family of FMRFamide-like peptides. Peptides 8:515522; 1987. 11. Evans, B. D.; Calabrese, R. L. Small cardioactive peptide-like immunoreactivity and its colocalization with FMRFamide-like immunoreactivity in the central nervous system of the leech Hirudo medicinalis. Cell Tissue Res. 257:187-199; 1989. 12. Evans, P. D.; Myers, C. M. The modulatory actions of FMRFamide and related peptides on locust skeletal muscle. J. Exp. Biol. 126:403-422; 1986. 13. Fujimoto, K.; Ohta, N.; Yoshida, M.; Kubota, I.; Muneoka, Y.; Kobayashi, M. A novel cardio-excitatory peptide isolated from the atria of the African giant snail, Achatina fulica. Biochem. Biophys. Res. Commun. 167:777-783; 1990. 14. Greenberg, M. J.; Price, D. A. Cardioregulatory peptides in molluscs. In: Bloom, F. E., ed. Peptides: Integrators of cell and tissue function. New York: Raven Press; 1980:107-126. 15. Greenberg, M. J.; Painter, S. D.; Dobie, K. E.; Nagle, G. T.; Price, D. A.; Lehman, H. E. The molluscan neurosecretory peptide FMRFamide: Comparative pharmacology and relationship to the enkephalins. Fed. Proc. 42:82-86; 1983. 16. Greenberg, M. J.; Price, D. A.; Lehman, H. K. FMRFamide-like peptides of molluscs and vertebrates: Distribution and evidence of function. In: Kobayashi, H.; Bern, H. A.; Urano, A., eds. Nearosecretion and the biology of neuropeptides. Berlin: Japan Sci. Soc. Press/Springer Verlag; 1985:370-376. 17. Greenwood, F. C.; Hunter, W. M.; Glover, L. S. The preparation of 13tI-labeled human growth hormone of high specific radioactivity. J. Biochem. 89:114--123; 1963. 18. Grimmelikhuijzen, C. J. P.; Graff, D. Isolation of
20.

21.

22.

23. 24.

25.

26. 27.

28.

29.

30.

31. 32. 33.

34.

35. 36.

N. Isolation of
908

neuropeptide gene. Neuron 1:55-61; 1988. 37. Norris, B. J.; Calabrese, R. L. Identification of motor neurons that contain a FMRFamidelike peptide and the effects of FMRFamide on longitudinal muscle in the medicinal leech Hirudo medicinalis. J. Comp. Neurol. 266:95-111; 1987. 38. Norris, B. J.; Calabrese, F. L. Action of FMRFamide on longitudinal muscle of the leech, Hirudo medicinalis. J. Comp. Physiol. [A] 167:211-224; 1990, 39. O'Donohue, T. L.; Bishop, J. F.; Chronwall, B. M.; Groome, J.; Watson, W. H. Characterization and distribution of FMRFamide immunoreactivity in the rat central nervous system. Peptides 5:563568; 1984. 40. Pohl, J.; Pereira, H. A.; Martin, N. M.; Spitznagel, J. K. Amino acid sequence of CAP37, a human neutrophil granule-derived antibacterial and monocyte-specific chemotactic glycoprotein structurally similar to neutrophil elastase. FEBS Lett. 272:200-204; 1990. 41. Price, D. A. The FMRFamide-like peptide of Helix aspersa. Comp. Biochem. Physiol. [C] 72:325-328; 1982. 42. Price, D. A. Evolution of a molluscan cardioexcitatory neuropeptide. Am. Zool. 26:1007-1015; 1986. 43. Price, D. A.; Greenberg, M. J. Structure of a molluscan cardioexcitatory peptide. Science 197:670-671; 1977. 44. Price, D. A.; Cottrell, G. A.; Doble, K. E.; Greenberg, M. J.; Jorenby, W.; Lehman, H. K.; Riehm, J. P. A novel FMRFamiderelated peptide in Helix: pQDPFLRFamide. Biol. Bull. 169:256266; 1985. 45. Price, D. A.; Davies, N. W.; Doble, K. E.; Greenberg, M. J. The variety and distribution of the FMRFamide-related peptides in molluscs. Zool. Sci. 4:395-410; 1987. 46. Raffa, R. B. The action of FMRFamide (Phe-Met-Arg-Phe-NH2) and related peptides on mammals. Peptides 9:915-922; 1988. 47. Robb, S.; Packman, L. C.; Evans, P. D. Isolation, primary structure and bioactivity of schistoFLRF-amide, a FMRF-amide-like neuropeptide from the locust, Schistocerca gregaria. Biochem. Biophys.

EVANS, POHL, KARTSONIS AND CALABRESE

Res. Commun. 160:850--856; 1989. 48. Roth, B. L.; Disimone, J.; Majane, E. A.; Yang, H. Y. Elevation of arterial pressure in rats by two new vertebrate peptides FLFQPQRFNH2 and AGEGLSSPFWSLAAPQRF-NH2 which are immunoreactive to FMRF-NH2 antiserum. Neuropeptides 10:37-42; 1987. 49. Schaefer, M.; Piciotto, M. R.; Kreiner, T.; Kaldany, R. R.; Taussig, R.; Scheller, R. H. Aplysia neurons express a gene encoding multiple FMRFamide neuropeptides. Cell 41:457-467; 1985. 50. Schneider, L. E.; Taghert, P. H. Isolation and characterization of a Drosophila gene that encodes multiple neuropeptides related to PheMet-Arg-Phe-NH2 (FMRFamide). Proc. Natl. Acad. Sci. USA 85: 1993-1997; 1988. 51. Shyamala, M.; Fisher, J. M.; Scheller, R. H. A neuropeptide precursor expressed in Aplysia neuron L5. DNA 5:203-208; 1986. 52. Taghert, P. H.; Schneider, L. E. Interspecific comparison of a Drosophila gene encoding FMRFamide-related neuropeptides. J. Neurosci. 10:1929-1942; 1990. 53. Taussig, R.; Scheller, R. H. The Aplysia FMRFamide gene encodes sequences related to mammalian brain peptides. DNA 5:453-461; 1986. 54. Trimmer, B. A.; Kobierski, L. A.; Kravitz, E. A. Purification and characterization of FMRFamidelike immunoreactive substances from the lobster nervous system: Isolation and sequence analysis of two closely related peptides. J. Comp. Neurol. 266:16-26; 1987. 55. Wallace, B. G.; Gilion, J. W. Characterization of acetylcholinesterase in individual neurons in the leech central nervous system, J. Neurosci. 2:1108-1118; 1982. 56. Willard, A. L. Effects of serotonin on the generation of the motor program for swimming by the medicinal leech. J. Neurosci. 1:936-944; 1981. 57. Yang, H. Y. T.; Fratta, W.; Majane, E. A.; Costa, E. Isolation, sequencing, synthesis, and pharmacological characterization of two brain neuropeptides that modulate the action of morphine. Proc. Natl. Acad. Sci. USA 82:7757-7761; 1985.