The peptide pQFYRFamide is encoded on the FMRFamide precursor of the snail Helix aspersa but does not activate the FMRFamide-gated sodium current.
The tetrapeptide FMRFamide (Price and Greenberg, 1977) is undoubtedly present in all molluscs (Greenberg and Price, 1992), but the mRNA encoding the FMRFamide precursor has only been completely described from an opisthobranch, Aplysia californica (Schaefer et al., 1985; Taussig and Scheller, 1986), and a pulmonate, Lymnaea stagnalis (Linacre et al., 1990). Although only a few stretches of amino acid sequence are common to the precursors of these two species [i.e., a single copy of FLRFamide, a sequence SEEPLYRKRRS which includes a tetrabasic proteolytic processing site, and the multiple copies of FMRFamide], the general organization of these two precursors is similar (see Greenberg and Price, 1992).
Pulmonate snails, but not other molluscs, also contain heptapeptides of the form XDPFLRFamide (X = N, S, G, pQ) (Price et al., 1990). These heptapeptide analogs of FLRFamide are encoded on a different precursor from FMRFamide (Saunders et al., 1991; Lutz et al., 1992) and are expressed in different neuronal populations (Bright et al., 1993; Cottrell et al., 1992) in both Lymnaea stagnalis and Helix aspersa. In L. stagnalis, the two precursors are alternatively spliced variants of a single gene which share the same first exon (Saunders et al., 1992). Whether a similar alternative splicing mechanism occurs in H. aspersa is unclear because only a partial cDNA encoding the tetrapeptides had been previously isolated (Lutz et al., 1992), notwithstanding that the cDNA encoding the heptapeptide precursor in H. aspersa has been fully sequenced (Cottrell et al., 1994).
Because conservation of structure is often associated with functionally important regions of a protein, and because the FMRFamide precursor from only one other pulmonate is fully known, we have determined the complete sequence of the FMRFamide precursor in H. aspersa, showing in the process that alternative splicing occurs in this species as well as in L. stagnalis. We have also broadened the comparison by examining two additional pulmonate snail species and have established, thereby, landmark sequences that seem to be common features of the FMRFamide precursors of pulmonate and opisthobranch molluscs.
The resulting sequences have revealed a putative novel peptide (pQFYRFamide) that has now been purified from extracts. The pharmacology of FMRFamide-related peptides has already been studied extensively in H. aspersa, and we know that the heptapeptide analogs of FLRFamide have some biological actions on peripheral organs and neurons that are distinct from those of FMRFamide and FLRFamide (Lehman and Greenberg, 1987; Cottrell and Davies, 1987). We have therefore surveyed the pharmacology of synthetic pQFYRFamide and report here that, although this pentapeptide has some of the same actions as FMRFamide on hearts and neurons of H. aspersa, it has no effect on the neuronal sodium channel that is directly gated by FMRFamide (Cottrell et al., 1990; Green et al., 1994).
Materials and Methods
Three species of snails from the pulmonate order Stylommatophora, superfamily Helicoidae, were used in this study. Two species, Helix aspersa and Cepaea nemoralis, are members of the family Helicidae and the third, Polydontes acutangula, is in the family Camaenidae. Note that Lymnaea stagnalis, another frequently studied pulmonate, is in the order Basommatophora.
Those specimens of H. aspersa used in the purification of pQFYRFamide and in the pharmacological testing of that peptide on snail hearts were collected in Fullerton, California, and shipped to St. Augustine, Florida, by Robert A. Koch. The cDNA library was made from H. aspersa collected in St. Andrews, Scotland, and these animals were also used in electrophysiological experiments. Specimens of Cepaea were collected in Bellmore, New York, whereas Polydontes were obtained in El Yunque rain forest in Puerto Rico.
RNA was isolated from clusters of neurons located in the parietal ganglia of H. aspersa. These clusters consist almost entirely of neurons immunoreactive to FMRFamide peptides (Cottrell et al., 1992). A directional cDNA library was constructed using the Lambda-Zap kit (Stratagene).
Preparation of template DNA. Aliquots (typically 5 [[micro]liter], but always less than 10 [[micro]liter]) of the amplified cDNA library were used directly in the PCR. For amplification of genomic DNA from Cepaea or Polydontes, a small piece of tissue (typically 1 ganglionic ring, 10 mg) was heated for 10 min at 95 [degrees] C in about a 5-fold excess of 100 mM NaOH (50 [[micro]liter] for a ganglion) and was then centrifuged for 5 min at maximum speed in an Eppendorf centrifuge. The supernatant was transferred to a clean tube and diluted 10-fold with water. Aliquots (0.3 to 5 [[micro]liter]) were used for PCR.
Standard amplification. All of the PCR components, except the DNA polymerase, were assembled in a total volume of 80 [[micro]liter] in a 0.5-ml tube and overlaid with mineral oil. The tube was heated to 95 [degrees] C for 10 min, cooled to 80 [degrees] C, the DNA polymerase (Taq, Promega, 2.5 U in 20 [[micro]liter]) added through the oil, and the PCR cycles (Perkin Elmer Cetus thermal cycler) started (30 to 40 cycles of 1 min at 94 [degrees] C, 1 min at 50 [degrees] C, 2 min at 72 [degrees] C). The PCR reaction mixture contained buffer (Promega: 50 mM KCl, 10 mM Tris-HCl, 0.1% Triton X-100, pH 9.0 at 25 [degrees] C), 1.5 mM Mg, 200 [[micro]molar] of each dNTP (Pharmacia), template and primers (20 to 200 [[micro]molar]). An aliquot (10 [[micro]liter]) was run on an agarose gel, and reactions that showed products of the expected size were purified for cloning as follows. First, the aqueous phase containing the PCR product was removed from under the oil layer and rolled around on a sheet of Parafilm to remove residual oil droplets (Whitehouse and Spears, 1991). The DNA was then precipitated by the addition of an equal volume of 5 M ammonium acetate followed by two volumes of isopropanol (storage overnight at -20 [degrees] C). The precipitate was then washed with 70% ethanol, dried in a Speed-Vac, and resuspended in 10 [[micro]liter] water. Aliquots (1-7 [[micro]liter]) were ligated into the pGEM-T vector (Promega), which was then transformed into E. coli (JM 109). Colonies were analyzed by PCR with primers taken from either side of the cloning site. Clones containing an insert of the appropriate size were sequenced with Sequenase (US Biochemicals, Cleveland, Ohio) kits.
Nested PCR. For the initial amplification, 5 [[micro]liter] from the library was used as the template for a pair of primers. [For an example, consider M13R and PD 12; see [ILLUSTRATION FOR FIGURE 1 OMITTED] for location.] From the initial reaction product, 0.1% to 1% was used as template for the secondary amplifications with two additional primers chosen from the targeted PCR product, and only 25 cycles of amplification were done. Continuing the example, see T3 and PD34 in [ILLUSTRATION FOR FIGURE 1 OMITTED].
Primers. The following oligonucleotide primers were synthesized by the DNA synthesis facility of the University of Florida Interdisciplinary Center for Biotechnology Research. Seven of these primer sequences were taken directly from known DNA sequences: PD7, PD9, and PD12 are from the FMRFamide-encoding clone HF1 (Lutz et al., 1992); PD64 is from the heptapeptide-encoding clone HF4 (also Lutz et al., 1992); M13 rev and T3 promoter are from near the insertion site of the plasmid cloning vector Bluescript (Stratagene); and PD92 is from the first FMRFamide-encoding PCR product obtained from Polydontes. The remaining eight primers were designed from amino acid sequences rather than DNA sequences and are degenerate; the mixed bases are parenthetical in the following list. Inosine (I) was used in positions where all four bases could encode the desired amino acid. Some primers to repeated sequences were designed to be only partially degenerate so that they would not match every repeat exactly. For example, PD11, an antisense primer to some DNA sequences encoding RFMRFGK (R or S) exactly matches only one repeat in the Cepaea FMRFamide gene. The primer binding sites for all of these oligonucleotides are shown in Figure 1.
PD7 (sense): CGTGGGGATTCAGAAACATCA-TGA
PD9 (anti-sense): TAGCGCTAGAACCCTACTACC-GAT
PD11 (anti-sense): CTTTTCCC(AG)AA(TC)CTCAT(GA)-AA(TC)CG
PD12 (anti-sense): AGATTGTTGGTTAGCGTCAGC-TGA
PD18 (sense): GAIAAG(AC)GITTC(TC)TG(AC)-GGTTCGG
PD34 (anti-sense): A(AG)(AG)AAIC(GT)(CT)TT(AG)-TC(TC)TG(AG)TAIGG
PD57 (sense): TCAGA(GC)(GC)AG(GC)CI(CT)-TITA(TC)(AC)GIAA(AG)(AC)G
PD64 (sense): ACTAGTCTGTGCCTCACCATC
PD75 (anti-sense): TTCCC(AG)AA(CT)CTCAT(AG)-AAIC(GT)(CT)TT
PD90 (sense): (CA)G(GA)CAGTT(TC)TA(TC)CGA-(TA)T(TC)GG(CT)(CA)G
PD92 (anti-sense): CTCGGCTGTTTGGTTGGCACT
PD122 (sense): TT(TC)ATG(CA)GITT(TC)GGIAG-(AG)GGCGA
PD123 (anti-sense): CTT(CG)CCGAACCTCATGAATCT-TTT
M13 rev (NEB # 1233): AGCGGATAACAATTTCACACA-GGA
T3 promoter (NEB # 1228): ATTAACCCTCACTAAAGGGA
Ganglia were added to acetone to give a final concentration of 2 g/10 ml, and left at -20 [degrees] C until processed further, but at least overnight. The acetone was decanted from the tissue, centrifuged, and the supernatant filtered through plain nylon (0.45-[[micro]meter] pore size; MSI, Westboro, Massachusetts). The clarified extract was reduced in volume to about 10% on a rotary evaporator with heating to 50 [degrees] C, leaving a primarily aqueous phase that was again clarified by filtration.
In all cases, the samples were pumped directly onto the reverse phase column (2.1 x 200 mm, Brownlee Aquapore Octyl) through the aqueous solvent pump. The absorbance of the effluent was monitored at two wavelengths (214 and 280 nm), and fractions (about 0.25 ml) were collected every 30 s. The flow rate was 0.5 ml/min throughout. Fractions to be rerun for further purification were pooled and diluted several-fold with the new aqueous solvent before being loaded onto the column. Three solvent systems were used: 0.1% trifluoroacetic acid (TFA) in water/0.1% TFA in 80% acetonitrile; the same with heptafluorobutyric acid (HFBA) substituted for TFA; and 5 mM sodium phosphate (pH 7.0) in water/5 mM sodium phosphate in 60% acetonitrile.
Radioimmunoassay, with either of two antisera (S253 or Q2), was used to analyze the fractions as previously described (Lesser and Greenberg, 1993).
The peptide pQFYRFamide was custom synthesized by Research Genetics (Huntsville, Alabama). Other peptides were synthesized by the peptide synthesis laboratory of the University of Florida Interdisciplinary Center for Biotechnology Research or purchased from Sigma Chemical Co.
Liquid chromatography-mass spectrometry (LC-MS)
Mass spectra were recorded in the positive ion mode on a TSQ-700 triple quadrapole instrument (Finnigan-MAT, San Jose, California) equipped with an electro-spray ion source operating at atmospheric pressure. The electrospray needle was operated at a voltage differential of 3-4 kV, and a sheath flow of 2 [[micro]liter]/min of methoxymethanol was used. Scans were acquired every 3 s (Swiderek et al., 1996).
The chromatography was performed on a microcapillary high-performance liquid chromatography (HPLC) system built at the Beckman Research Institute, City of Hope (Davis et al., 1994, 1995). Fused silica columns with an inner diameter of 250 [[micro]meter] were packed with Vydac 3-[[micro]meter] C18 RP support. The sample was eluted with a gradient from 98% solvent A (0.11% TFA) to 62% solvent B (90% acetonitrile, 0.07% TFA) over 30 min with a flow of 2 [[micro]liter]/min. The UV absorbance (200 nm) was monitored before the sample was introduced into the mass spectrometer. Spectra were generated by averaging scans containing the peak, and the masses were calculated with data reduction software (Finnigan MAT BIOMASS). The determined mass values of the molecular ions are accurate to within a few tenths of a dalton, reported values are rounded down to the nearest integral value.
The isolated heart. The perfused ventricle of H. aspersa was used as described previously (Lesser and Greenberg, 1993) to assay the cardioactivity of synthetic pQFYRFamide.
Recording from identified neurons. Electrical recordings were made from two identified neurons (C1 and C2) in the cerebral ganglia of H. aspersa; neither neuron is a known follower of a FMRFamide-containing cell (see Cottrell and Davies, 1987). The physiological solution had the following composition (mM): NaCl (90), KCl (5), Mg[Cl.sub.2] (5), Ca[Cl.sub.2] (7), HEPES (20), and the pH was adjusted to 7.4 with NaOH. Neurons were voltage clamped with a Dagan 8100 single electrode clamp system with low-resistance micropipettes containing 200 mM KCl. Peptide solutions (100 [[micro]molar]) were prepared in an appropriate physiological medium and were usually pressure applied (Picospritzer II, General Valve Corporation). However, a barium-containing physiological solution (10 mM KCl, 25 mM Ba[Cl.sub.2], 5 mM Mg[Cl.sub.2], 75 mM tetraethylammonium chloride, 3 mM 4-amino pyridine, 5 mM Tris HCl, pH 7.5) was used during the measurement of currents passing through calcium channels, and the peptide was added to the bathing solution prior to the generation of a current-voltage curve.
The complete sequence of the mRNA encoding the FMRFamide precursor
A partial sequence had previously been isolated from a Helix aspersa cDNA library; it extended from an EcoRI site 21 nucleotide residues upstream of the tetra-basic sequence (RKRR; indicated on [ILLUSTRATION FOR FIGURE 1 OMITTED]), through the 3[prime] noncoding region (Lutz et al, 1992). We have used the PCR-based strategy illustrated in Figure 1 to determine the sequence of the missing 5[prime] end of the cDNA. This additional sequence (shown in [ILLUSTRATION FOR FIGURE 2 OMITTED]), taken together with that of Lutz et al. (1992), completes the sequence of the FMRFamide precursor of H. aspersa. The precursor is diagrammed in Figure 1, and the complete sequence has been deposited in GenBank, Accession L20768.
The 5[prime] end of the cDNA and the peptide precursor that it encodes [derived from these data and those of Lutz et al. (1992)] contain four noteworthy features [ILLUSTRATION FOR FIGURE 1 OMITTED]:
First, the 243 bases at the 5[prime] terminal (including the noncoding region and the signal sequence) are identical to those of the cDNA encoding the precursor of the heptapeptide analogs of FMRFamide (Lutz et al., 1992). Therefore, the two transcripts (i.e., those encoding the FMRFamide and heptapeptide precursors, respectively) must be derived by alternative RNA splicing, a finding also reported for Lymnaea stagnalis (reviewed by Benjamin and Burke, 1994).
Second, downstream from the signal sequence, the 5[prime] cDNA segment also encodes two copies of a previously unknown pentapeptide - pGlu-Phe-Tyr-Arg-Phe-N[H.sub.2] (pQFYRFamide) - as well as a copy of FLRFamide.
Third, 60 base pairs downstream from this FLRFamide, the cDNA encodes the tetrabasic cleavage site (Arg-Lys-Arg-Arg; RKRR) reported previously (Lutz et al., 1992). The RKRR site also occurs in Aplysia californica (Taussig and Scheller, 1986) and L. stagnails (Benjamin and Burke, 1994.)
Finally, 111 base pairs downstream, about 3/5 of the distance between the tetrabasic (RKRR) site and the first copy of FMRFamide, the cDNA encodes a dibasic cleavage site (KR). This site is also found in Aplysia californica and L. stagnalis (Taussig and Scheller, 1986; Linacre et al. 1990), but its relative position between the tetrabasic sequence and the first FMRFamide repeat varies.
In summary, the complete, derived precursor includes 10 copies of FMRFamide, 2 copies of FLRFamide (one in the midst of the FMRFamide repeats), 2 copies of the novel pentapeptide pQFYRFamide, and 2 conserved cleavage sites in addition to those flanking the FMRFamide-related peptides.
PCR amplification of other FMRFamide gene fragments
DNA from ganglia of the helicid snail Cepaea nemoralis and the camaenid snail Polydontes acutangula was prepared by crude alkaline lysis and amplified by PCR. In addition to the primers that had been employed for amplifications in Helix, we used another (PD92) that was synthesized to match exactly the P. acutangula sequence [ILLUSTRATION FOR FIGURE 1 OMITTED].
From C. nemoralis we obtained overlapping fragments covering much of the precursor, as shown in Figure 1 (for sequence see GenBank Accession U02488). The sequence of the Cepaea precursor is very similar to that of the two species of Helix ([ILLUSTRATION FOR FIGURES 1 and 3 OMITTED]; Lutz et al., 1992). Within the portion of the precursor that has now been sequenced from all three species, the Cepaea precursor has 88% nucleotide (nt) and 84% amino acid (aa) identity to H. aspersa and 90% nt and 81% aa identity to H. pomatia, and the two Helix species have 89% nt and 86% aa identity to each other.
The linear arrangement - in the precursors of Cepaea and H. aspersa - of the landmarks described above is virtually identical. In particular, the Cepaea precursor also encodes two copies of pQFYRFamide, a copy of FLRFamide, a tetrabasic site, and a dibasic cleavage site. Moreover, all of these features are in the same order and in virtually the same positions as in H. aspersa; the exception is that the distance between the dibasic cleavage site and the first copy of FMRFamide is one amino acid longer in Cepaea [ILLUSTRATION FOR FIGURES 1 and 3 OMITTED]. As expected, the segments between these landmarks are also very similar at both the amino acid and nucleotide levels [ILLUSTRATION FOR FIGURE 3 OMITTED].
Although we have amplified only a small portion of the FMRFamide gene from P. acutangula [ILLUSTRATION FOR FIGURE 1 OMITTED], most of the landmarks found in the cDNA from H. aspersa and Cepaea are present and recognizable. The 5[prime] end of the sequence encodes a copy of FLRFamide, the tetrabasic site, and the dibasic cleavage site. Upstream from the FLRFamide is the site where PCR primer PD90 annealed; this primer was designed to hybridize to DNA sequences encoding either pQFYRFamide or pQFYRIamide. Although the PCR product that was cloned and sequenced contains the code for Ile, we cannot be sure which amino acid is actually coded for in the native DNA.
In contrast to the landmarks themselves, the segments between them are poorly conserved in P. acutangula [ILLUSTRATION FOR FIGURE 3 OMITTED]. The precursor in this species is only 41% identical to H. aspersa at the nucleotide level, and 34% at the amino acid level. Moreover, although the overall length of the region between the two outside landmarks - pQFYR (I/F) amide and FMRFamide [ILLUSTRATION FOR FIGURE 1 OMITTED] - is exactly the same as for H. aspersa, three off-setting gaps were required to align the tetrabasic and dibasic cleavage sites in the two species. For another example, the peptide just 5[prime] of the tetrabasic site is especially dissimilar to that of H. aspersa: i.e., it is longer (23 amino acids instead of 20) and, even with the gaps, only 17% conserved.
Isolation of pQFYRFamide
The predicted peptide pQFYRFamide was synthesized; it was as reactive as FMRFamide in an RIA with antiserum S253 which facilitated the purification. The synthetic peptide eluted at 12-13 min in the TFA/ACN solvent system, i.e., coincident with a large peak of immunoreactivity observed with H. aspersa ganglion extracts (see Price et al., 1990, for a typical pattern). Since this large peak also contains SDPFLRFamide and NDPFLRFamide, we sought a second solvent system that could separate pQFYRFamide from these two heptapeptides; the phosphate/ACN system proved to be satisfactory.
This fractionation procedure, with one modification, was applied separately to two H. aspersa extracts, one made from 50 pairs of cerebral ganglia and another from 30 subesophageal ganglia. The cerebral ganglion extract was given a preliminary fractionation with the HFBA/ACN system. Thereafter, each extract was fractionated on a TFA/ACN gradient, and the immunoreactive peaks from this run were further resolved with the phosphate/ACN solvent system. On this system, the cerebral ganglion extract showed a large peak at 19 min corresponding to synthetic pQFYRFamide, and a small peak at 16 min corresponding to the two heptapeptides (which are not separated in this system). The subesophageal ganglion extract contained peaks of heptapeptide (16 min) and pQFYRFamide (19 min) of roughly equal size (not shown). The pQFYRFamide peaks from both purifications were combined and run through a final TFA/ACN system. The major peak (12 to 12.5 min; [ILLUSTRATION FOR FIGURE 4A OMITTED]) was subjected to combined HPLC/mass spectrometry, revealing a molecular ion at 742 [ILLUSTRATION FOR FIGURE 4B OMITTED] in agreement with the calculated value of 742. Since this peptide eluted from the HPLC column at the same time as synthetic pQFYRFamide and had the same mass as pQFYRFamide, we conclude that it is pQFYRFamide.
Cardioactivity of pQFYRFamide
The effects of synthetic pQFYRFamide on the isolated heart of H. aspersa were compared with those of FMRFamide [ILLUSTRATION FOR FIGURE 5 OMITTED]. FYRFamide was about half as potent as FMRFamide, whereas pQFYRFamide was about 5 times more potent. Thus, blocking of the N-terminal by addition of pyroglutamic acid increases the potency of FYRFamide by 10-fold. Payza (1987) also noted that FMRFamide analogs with blocked N-terminals were more potent on the heart than those without.
Responses of identified neurons of Helix aspersa
Potassium current evoked in the C1 neuron. FMRFamide induces a slowly developing hyperpolarization of the C1 neuron that results from an increase in potassium conductance (Cottrell and Davies, 1987). In normal physiological solution (see Methods) pQFYRFamide evoked a response similar to that of FMRFamide (n = 5); e.g., under voltage clamp at -45 mV, an outward current accompanied by an increased membrane conductance was recorded [ILLUSTRATION FOR FIGURE 6A OMITTED]. At a holding potential of -100 mV the direction of the current response was inward (bottom record, [ILLUSTRATION FOR FIGURE 6A OMITTED]). The relationships between evoked current and holding potential were compared in normal physiological solution (containing 5 mM K) and in a saline with reduced potassium (1.5 mM), as shown in Figure 6B. The shift in reversal potential produced by reducing the potassium concentration was about 30 mV, close to that predicted for a response mediated entirely by a change in potassium conductance. FMRFamide and pQFYRFamide both had E[C.sub.50] values of about 2 [[micro]molar], but pQFYRFamide consistently produced a larger maximum effect (about 50% greater than that of FMRFamide).
Suppression of a voltage-dependent calcium current in the C1 neuron. FMRFamide reduces the voltage-dependent calcium current by up to 30% (Colombaioni et al., 1985; Cottrell and Lesser, 1987). pQFYRFamide also reversibly reduced the amplitude of the Ca current and by a similar amount (n = 4; data not shown). The E[C.sub.50] for pQFYRFamide was similar to that for FMRFamide, i.e., about 5 [[micro]meter].
Fast sodium current evoked in the C2 neuron. FMRFamide was shown to produce a rapidly developing inward current in the C2 neuron (Cottrell and Davies, 1987), an effect mediated by the direct activation of sodium channels (Cottrell et al., 1990; Green et al., 1994). A comparison of the effects of FMRFamide and pQFYRFamide, recorded under voltage clamp from the whole cell [ILLUSTRATION FOR FIGURE 7 OMITTED], clearly shows that pQFYRFamide fails to evoke this fast inward current (n = 7). Apparently, pQFYRFamide cannot activate the FMRFamide-gated sodium channel.
Potassium current evoked in the C2 neuron. Some C2 neurons, but not the example shown in Figure 7, responded to FMRFamide with a weak outward current in addition to the fast response. This outward current, like the slow hyperpolarization observed in many H. aspersa neurons, is due to an increased potassium conductance (see C1 neuron above). pQFYRFamide also, and consistently (n = 7), evoked a slow outward current [ILLUSTRATION FOR FIGURES 7 and 8A OMITTED] that was accompanied by an increase in membrane conductance [ILLUSTRATION FOR FIGURE 8A OMITTED]. This response inverted close to the potassium equilibrium potential [ILLUSTRATION FOR FIGURE 8B OMITTED] in both normal and reduced potassium solutions (based on [[K.sub.i]] = 98 mM Alvarez-Leefmans and Gamino, 1982). It is, therefore, very similar to the slow increase in potassium conductance induced by FMRFamide and described above for the C1 neuron ([ILLUSTRATION FOR FIGURE 6 OMITTED]; and Cottrell et al., 1984); the E[C.sub.50] values for pQFYRFamide and FMRFamide were similar - about 1 [[micro]molar].
We have sequenced the 5[prime] end of the cDNA encoding the FMRFamide precursor of Helix aspersa, and so the entire sequence is now complete. The general organization of this precursor is similar to that of the two other completely sequenced FMRFamide precursors: Lymnaea stagnalis (reviewed by Benjamin and Burke, 1994) and Aplysia californica (see Taussig and Scheller, 1986). The similarity is manifest in both the splicing pattern and in the linear arrangement of landmark sequences in the precursor.
The mRNA encoding the FMRFamide precursor in all three species is composed of at least two exons, and in all three the splice junction is in a roughly similar position. In both Lymnaea and Helix the first exon of the FMRFamide precursor is alternatively spliced to give another neuropeptide precursor (that of the FLRFamide-related heptapeptides), but no alternatively spliced product has ever been found in Aplysia.
The amino acid segments that are processed to become the individual mature peptides are not distributed at random in the precursors, rather, they are clustered according to their sequences in two domains. First, the C-terminal end of the precursor (always downstream of the tetrabasic sequence, RKRR) is the site of repeated segments, each comprising a FMRF sequence with its processing signals and acidic spacers. Second, the N-terminal domain (upstream of the tetrabasic sequence) is devoid of FMRFamide but always contains one or two copies of FLRFamide, as well as of related peptides. The related peptide in H. aspersa and C. nemoralis is the newly discovered pentapeptide pQFYRFamide [ILLUSTRATION FOR FIGURE 1 OMITTED]. In Lymnaea, one copy of pQFYRFamide is encoded at the level of the more 3[prime] of the two copies of pQFYRF-amide in Helix. And even in Aplysia, the 5[prime] end of the FMRFamide mRNA encodes a copy of FLRFamide and a long, related peptide (Greenberg and Price, 1992).
The functional significance of the tetrabasic site and the sequences surrounding it has been considered from time to time (reviewed by Linacre et al., 1990). In Lymnaea, a physiological role has been advanced for the acidic dodecapeptide segment SEQPD . . . SEEPLY that occurs just upstream from the tetrabasic site (Linacre et al., 1990; Burke et al., 1993). Our comparison of the Polydontes and H. aspersa precursors bears on this notion of function. Of the 23 amino acids preceding the tetrabasic site in these two land snails (and corresponding to the SEQPD . . . SEEPLY peptide of Lymnaea), only 5 are identical. Moreover, of the 23 residues following the tetrabasic site, only 3 are identical. These non-conserved sections of the precursor are, therefore, not likely to have a specific function common to all pulmonates, let alone all molluscs.
In contrast, the sequence of amino acids in the tetra-basic site seems to be conserved among pulmonates and opisthobranchs; moreover, it fits the consensus motif (Arg-Xaa-Lys/Arg-Arg) of a site of action for a furin-like endopeptidase (Hosaka et al., 1991). The involvement of furin-like cleavage in neuropeptide processing was first demonstrated for the egg-laying hormone of A. californica (Newcomb et al., 1988; Fisher et al., 1988). Putative furin sites have since been observed in the myomodulin precursors of Aplysia (Miller et al., 1993) and L. stagnalis (Kellett et al., 1996); and the first cleavage seems to occur at such a site during the processing of the hepta-peptide precursor in H. aspersa (Cottrell et al., 1994). As for the tetrapeptide precursor, Linacre et al. (1990) had already suggested that the FLRFamide and FMRFamide domains could be separated by cleavage at the RKRR sequence. This conjecture was later elaborated by Greenberg and Price (1992), and it appears to be a reasonable working hypothesis.
If, however, this hypothesis is correct, and molluscan FMRFamide precursors are first processed into N- and C-terminal domains, then the products of the two domains might be functionally, and not merely structurally, distinct. In search of such a distinction, we compared the actions of pQFYRFamide (restricted to the N-terminal domain) with those of FMRFamide (restricted to the C-terminal domain) on the heart and selected neurons of H. aspersa.
The activity of pQFYRFamide on the isolated Helix ventricle was qualitatively similar to that of FMRFamide and FLRFamide, the other two peptides derived from the tetrapeptide precursor. But pQFYRFamide, like FLRFamide, is about 5-fold more potent than FMRF-amide (Price et al., 1990).
On neurons, pQFYRFamide has some, but not all of the actions of FMRFamide. It mimics FMRFamide in evoking a potassium-dependent outward current and also in suppressing a voltage-activated calcium current. But pQFYRFamide - unlike FMRFamide - does not activate the fast inward sodium current in neuron C2, and thus it must not activate the FMRFamide-gated sodium channel. In fact, neither the heptapeptides (Cottrell and Davies, 1987), nor acetyl-Phe-Nle-Arg-Phe-N[H.sub.2], FYRFamide, or FLRF evoke this fast sodium current in the C2 neuron; and even FLRFamide is less potent than FMRFamide (K.A. Green, pers. comm.). Recently, the cDNA encoding this FMRFamide-gated sodium channel was sequenced and expressed in frog oocytes (Lingueglia et al., 1995); these expressed channels had properties similar to those observed in the C2 neuron (Cottrell et al., 1990; Green et al., 1994).
The effectiveness of FMRFamide in evoking the inward sodium current - in contrast to the ineffectiveness of pQFYRFamide and FLRFamide - indicates that the recognition site for the fast ligand-gated response has a high level of specificity. More important, it shows that the pharmacological activity of the major peptide at the N-terminal portion of the precursor protein is different from the activity of the major peptides at the C-terminal portion, at least at one site.
In summary, we have sequenced the 5[prime] end of the cDNA for the FMRFamide precursor of H. aspersa. When the FMRFamide precursors from different species were compared, the overall level of similarity varied with the phyletic distance between the species. But in every case, the sequences incorporating the FMRFamide-related peptides and the tetrabasic recognition site, as well as the organization of the precursor, were strongly conserved. Noting the putative peptide pQFYRFamide encoded on this cDNA segment, we searched for and subsequently identified this peptide in extracts of ganglia. Mass spectroscopy confirmed the predicted sequence and post-translational modifications (i.e., amidation and cyclization of glutamine to pyroglutamic acid). pQFYRFamide had effects similar to those of FMRF-amide on the heart and to an extent on neurons. But this novel pentapeptide did not activate the neuronal FMRFamide-gated sodium channel.
This study was supported by grants from NIH (HL28440 to MJG) and the SERC (to GAC). We thank Lynn Milstead for her help in preparing the figures.
Alvarez-Leefmans, F. J., and S. M. Gamino. 1982. Intracellular free potassium concentration in snail neurones measured with ion-sensitive micro-electrodes. Soc. Neurosci. Abst. 8: 684.
Benjamin, P. R., and J. F. Burke. 1994. Alternative mRNA splicing of the FMRFamide gene and its role in neuropeptidergic signalling in a defined neural network. Bioessays 16: 335-342.
Bright, K., E. Kellett, S. E. Saunders, M. Brierley, J. F. Burke, and P. R. Benjamin. 1993. Mutually exclusive expression of alternatively spliced FMRFamide transcripts in identified neuronal systems of the snail Lymnaea. J. Neurosci. 13:2719-2729.
Burke, J. F., N. Santama, E. Kellett, S. E. Saunders, M. Yoeman, and P. R. Benjamin. 1993. Expression of the neuropeptide gene FMRFamide and myomodulin in Lymnaea. Soc. Neurosci. Abst. 19:1274.
Colombaioni, L., D. Paupardin-Tritsch, P.P. Vidal, and H.M. Gerschenfeld. 1985. The neuropeptide FMRF-amide decreases both the Ca2+ conductance and a cyclic 3[prime], 5[prime]-adenosine mono-phosphate-dependent K+ conductance in identified molluscan neurons. J. Neurosci. 5: 2533-2538.
Cottrell, G. A., and N. W. Davies. 1987. Multiple receptor sites for a molluscan peptide (FMRFamide) and related peptides of Helix. J. Physiol. 382:51-68.
Cottrell, G. A., and W. Lesser. 1987. A comparison of the effects of some molluscan (FMRFamide-like) peptides on the calcium current of a snail neurone. J. Physiol. 392: 33P.
Cottrell, G. A., N. W. Davies, and K. A. Green. 1984. Multiple actions of a molluscan cardioexcitatory neuropeptide and related peptides on identified Helix neurones. J. Physiol. 356:315-333.
Cottrell, G. A., K. Green, and N. W. Davies. 1990. The neuropeptide Phe-Met-Arg-Phe-N[H.sub.2] (FMRFamide) can activate a ligand-gated ion channel in Helix neurones. Pflugers Arch. 416: 612-614.
Cottrell, G. A., D. A. Price, K. E. Doble, S. Hettle, J. Sommerville, and M. Macdonald. 1992. Identified Helix neurons: mutually exclusive expression of the tetrapeptide and heptapeptide members of the FMRFamide family. Biol. Bull. 183:113-122.
Cottrell, G. A., E. M. Lutz, D. A. Price, and J. Sommerville. 1994. N-terminally extended FMRFamide-related peptides of Helix aspersa: processing of the precursor protein and distribution of the released peptides. Mol. Cell. Neurosci. 5:632-641.
Davis, M. T., D. C. Stahl, and T. D. Lee. 1995. Low flow HPLC solvent delivery system designed for tandem capillary liquid chromatography mass spectrometry. J. Amer. Soc. Mass Spectrom. 6: 571-577.
Davis, M. T., D. C. Stahl, K. Swiderek, and T. D. Lee. 1994. Capillary liquid chromatography mass spectrometry for peptide and protein characterization. Pp. 304-314 in Methods, a Companion to Methods in Enzymology 6, J. E. Shively, ed. Academic Press, San Diego.
Fisher, J. M., W. Sossin, R. Newcomb, and R. H. Scheller. 1988. Multiple neuropeptides derived from a common precursor are differentially packaged and transported. Cell 54:813-822.
Green, K. A., S. W. P. Falconer, and G. A. Cottrell. 1994. The neuropeptide Phe-Met-Arg-Phe-N[H.sub.2] (FMRFamide) directly gates two ion channels in an identified Helix neurone. Pflugers Arch. 428: 232-240.
Greenberg, M. J., and D. A. Price. 1992. Relationships among the FMRFamide-like peptides. Prog. Brain Res. 92: 25-37.
Hosaka, M., M. Ngahama, W.-S. Kim, T. Watanabe, K. Hatsuzawa, J. Ikemizu, K. Murakami, and K. Nakayama. 1991. Arg-X-Lys/Arg-Arg motif as a signal for precursor cleavage catalyzed by furin within the constitutive secretory pathway. J. Biol. Chem. 266: 12127-12130.
Kellett, E., S. J. Perry, N. Santama, B. M. Worster, P. R. Benjamin, and J. F. Burke. 1996. Myomodulin gene of Lymnaea: structure, expression and analysis. J. Neurosci. 16: 4949-4957.
Lehman, H. K., and M. J. Greenberg. 1987. The actions of FMRFamide-like peptides on visceral and somatic muscles of the snail Helix aspersa. J. Exp. Biol. 131: 55-68.
Lesser, W., and M. J. Greenberg. 1993. Cardiac regulation by endogenous small cardioactive peptides (SCPs) and FMRFamide-related peptides in the snail Helix aspersa. J. Exp. Biol. 178: 205-230.
Linacre, A., E. Kellett, S. Saunders, K. Bright, P. R. Benjamin, and J. F. Burke. 1990. Cardioactive neuropeptide Phe-Met-Arg-Phe-N[H.sub.2] (FMRFamide) and novel related peptides are encoded in multiple copies by a single gene in the snail Lymnaea stagnalis. J. Neurosci. 10:412-419.
Lingueglia, E., G. Champigny, M. Lazdunski, and P. Barbry. 1995. Cloning of the amiloride-sensitive FMRFamide peptide-gated sodium channel. Nature 378: 730-733.
Lutz, E. M., M. Macdonald, S. Hettle, D. A. Price, G. A. Cottrell, and J. Sommerville. 1992. Structure of cDNA clones and genomic DNA encoding FMRFamide-related peptides (FaRPs) in Helix. Mol. Cell. Neurosci. 3: 373-382.
Miller, M. W., S. Beushausen, A. Vitek, S. Stamm, I. Kupfermann, J. Brosius, and K. R. Weiss. 1993. The myomodulin-related neuropeptides: characterization of a gene encoding a family of peptide cotransmitters in Aplysia. J. Neurosci. 13: 3358-3367.
Newcomb, R., J. M. Fisher, and R. H. Scheller. 1988. Processing of the egg-laying hormone (ELH) precursor in the bag cell neurons of Aplysia. J. Biol. Chem. 263: 12514-12521.
Payza, K. 1987. FMRFamide receptors in Helix aspersa. Peptides 8: 1065-1074.
Price, D. A., and M. J. Greenberg. 1977. Structure of a molluscan cardioexcitatory neuropeptide. Science 197:670-671.
Price, D. A., W. Lesser, T. D. Lee, K. E. Doble, and M. J. Greenberg. 1990. Seven FMRFamide-related and two SCP-related cardioactive peptides from Helix. J. Exp. BioL 154:421-437.
Saunders, S. E., K. Bright, E. Kellett, P. R. Benjamin, and J. F. Burke. 1991. Neuropeptides Gly-Asp-Pro-Phe-Leu-Arg-Phe-amide (GDPFLRFamide) and Ser-Asp-Pro-Phe-Leu-Arg-Phe-amide (SDPFLRFamide) are encoded by an exon 3[prime] to Phe-Met-Arg-Phe-N[H.sub.2] (FMRFamide) in the snail Lymnaea stagnalis. J. Neurosci. 11: 740-745.
Saunders, S. E., E. Kellett, K. Bright, P. R. Benjamin, and J. F. Burke. 1992. Cell-specific alternative RNA splicing of an FMRFamide gene transcript in the brain. J. Neurosci. 12: 1033-1039.
Schaefer, M., M. R. Picciotto, T. Kreiner, R.-R. Kaldany, R. Taussig, and R. H. Scheller. 1985. Aplysia neurons express a gene encoding multiple FMRFamide neuropeptides. Cell 41: 457-467.
Swiderek, K. M., T. D. Lee, and J. E. Shively. 1996. Trace structural analysis of proteins. Pp. 68-86 in Methods in Enzymology, vol. 271, B. L. Karger and W. S. Hancock, eds. Spectrum, Publisher Services, York, PA.
Taussig, R., and R. H. Scheller. 1986. The Aplysia FMRFamide gene encodes sequences related to mammalian brain peptides. DNA 5: 453-462.
Whitehouse, E. H., and T. Spears. 1991. A simple method for removing oil from cycle sequencing reactions. Biotechniques 11:617-618.