Purification, sequence analysis, and cellular localization of a prodynorphin-derived peptide related to the alpha-neo-endorphin in the rhynchobdellid leech Theromyzon tessulatum.

Cells immunoreactive to an antiserum specifically directed against vertebrate alpha-Neo-endorphin (alpha-NE) were detected in the internal wall of anterior and posterior suckers of the rhynchobdellid leech Theromyzon tessulatum. These cells have morphological and ultrastructural characteristics close to the "releasing gland cells" of adhesive organs. The epitope recognized by anti-alpha-NE was contained in granules having a diameter of 0.2-0.3 microm. Previous works involving the brain of this leech demonstrate the existence of approximately 14 neurons immunoreactive to the anti-alpha-NE. Following an extensive purification including high pressure gel permeation and reversed-phase high performance liquid chromatography, epitopes contained in both suckers and central nervous system were isolated. Purity of the isolated peptides was controlled by capillary electrophoresis. Their sequences were determined by a combination of automated Edman degradation, electrospray mass spectrometry measurement, and coelution experiments in reversed-phase high performance liquid chromatography with synthetic alpha-NE. The results demonstrate that epitopes recognized by the anti-alpha-NE in the suckers and the central nervous system are identical to vertebrate alpha-NE (YGGFLRKYPK). This finding constitutes the first biochemical characterization of a prodynorphin-derived peptide in invertebrates. Moreover the isolation of this peptide in the annelida establishes the very ancient phylogenetic origin of alpha-NE as well as its conservation in evolution.

In vertebrates, all known opioids peptides are cleavage products of three different precursors, i.e. proopiomelanocortin (POMC), 1 proenkephalin, and prodynorphin (1). Processing of the prodynorphin yields a number of bioactive peptides including leucine-enkephalin, Neo-endorphins (␣ and ␤), and dynorphins (A and B) (Patey and Rossier, 1986). Among these peptides, ␣-Neo-endorphin (␣-NE) has been isolated from all vertebrates phyla (Sei et al., 1989;Dores et al., 1993a;Gold-smith, 1992). The ␣-NE isolated from different tetrapods revealed by reversed-phase HPLC similar biochemical properties, reflecting the great conservation of this peptide in tetrapods (Sei et al., 1989, Goldsmith, 1992. NE immunocytochemical probes in cyclostomes, holostean and teleostean fish, proved negative (Dores and Gorbman, 1989;Dores et al., 1993b;Dores and McDonald, 1992). This result may be due to the following: 1) ␣-NE sequence changes that render this opioid undetectable to heterologous mammalian antisera, 2) a unique set of posttranslational processing reactions in which ␣-NE is not liberated from fragments of the prodynorphin precursor, or 3) the possibility that during evolution the prodynorphin precursor may be absent in fishes (see Dores et al. (1993b)).
If these three genes encoding opioid peptides are related (Gubler, 1987), one of the central questions is in what order did these genes evolve from a hypothetical ancestral gene? In this context research on different opioid peptides in invertebrates is essential (Stefano, 1991).
Given this amount of information in invertebrates concerning opioid peptides, little is known about ␣-NE in these animals. This peptide was not detected by immunocytochemistry in the mollusk Lymnaea stagnalis (Boer and van Minnen, 1985). By contrast, immunoreactive material was present in annelids i.e. polychaeta (Dhainaut-Courtois et al., 1986) and achaeta (Verger-Bocquet et al., 1987a, 1987b. In the leech T. tessulatum, a great number of cells immunoreactive to the antiserum raised against ␣-NE were found in the proboscis (Verger-Bocquet et al., 1987a) and in the brain (6 -14 neurons) (Verger-Bocquet et al., 1987a, 1987b. The aim of the present study is to fully characterized the ␣-NE-like peptide(s) in the leech T. tessulatum. Here we report for the first time that such material is present in the central nervous system and in a novel type of cell in the suckers.

Animals
Mature specimens of the rhynchobdellid leech T. tessulatum were reared under controlled laboratory conditions as described in detail by Malecha et al. (1989).

Dissections and Surgical Procedures
After anesthesia with chloretone, animals were pinned flat, ventral side up in leech saline solution (Muller et al., 1981). Central nervous * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Light Microscopy
Anterior and posterior parts of T. tessulatum (including suckers) were fixed overnight at 4°C in Bouin-Hollande fixative (ϩ10% HgCl 2 saturated solution). They were then embedded in paraffin and serially sectioned at 7 m. After removal of paraffin with toluene, the sections were successively treated with the anti-␣-NE diluted 1:200 and with goat anti-rabbit IgG conjugated to horseradish peroxidase as described elsewhere (Verger-Bocquet et al., 1992). The specificity of the antiserum was tested on consecutive sections mounted on different slides by preadsorbing the antiserum overnight at 4°C with the homologous antigen (synthetic ␣-NE, Sigma) at a concentration of 500 g/ml pure antiserum.

Purification
A four-step procedure was employed (Table I).

Central Nervous Systems
Step I: Sep-Pak Prepurification-Batches of 200 CNS were homogenized and extracted with 200 l of 1 M acetic acid at 4°C. After centrifugation at 12,000 rpm for 30 min at 4°C, the pellet was reextracted once. The two supernatants were combined and loaded onto Sep-Pak C 18 cartridges (500 l of extract/cartridge; Waters). After washing the cartridges with 5 ml of 1 M acetic acid, elution was performed with 5 ml of 50% acetonitrile in water acidified with 0.1% trifluoroacetic acid (Pierce). The eluted fractions were reduced 20-fold in a vacuum centrifuge (Savant) to remove acid and organic solvent. The total amount of anti-␣-NE-like material was quantified using ELISA.
Step III: Reversed-phase HPLC-Fractions immunoreactive in DIA to anti-␣-NE were concentrated before separated on a C 18 -peptide protein column (250 mm ϫ 4.6 mm; Vydac) equilibrated with acidified water (0.1% trifluoroacetic acid). Elution was performed with a discontinuous linear gradient of acetonitrile in acidified water over 0 -15% for 10 min and over 15-45% for 30 min at a flow rate of 1 ml/min. The column effluent was monitored by absorbance at 226 nm and the presence of ␣-NE-like material detected in aliquots of each fraction by DIA. Fractions that contained the immunoreactive material were applied on the same column with a shallower gradient. Elution was performed with a discontinuous linear gradient of acetonitrile in acidified water over 0 -15% for 10 min and over 15-45% for 40 min at a flow rate of 1 ml/min. After a 20-fold concentration by freeze-drying, fraction aliquots of 0.5 l were tested by DIA.
Step IV: Final Purification-The ␣-NE-like material was finally purified on an ODS C 18 reversed-phase column (Ultrasphere, 250 mm ϫ 2 mm; Beckman) developed with a linear gradient of 0 -60% acetonitrile in acidified water for 60 min at a flow rate of 50 l/min. The column effluent was monitored by absorbance at 226 nm and the immunoreactive material detected as above.

Suckers
Suckers underwent same purification procedure, except an additional prepurification step was added. The supernatant obtained after acidic extraction contained green and yellow pigments as well as mucus. They were removed early on the procedure by precipitation with water/acetone (20/80, v/v). The acetonic fraction was reduced 20-fold before subjected to the Sep-Pack prepurification step procedure.
All HPLC purifications were performed with a Beckman Gold HPLC system equipped with a photodiode array detector (Beckman 168).

Spectral Scanning
A comparison between first derivatives of absorbance of scan profiles obtained from synthetic peptides and from the endogenous immunoreactive peptides, at identical concentration (100 pmol) was performed as noted in detail elsewhere (Salzet et al., 1993b).

Capillary Electrophoresis
Prior to microsequencing, the purity of the peptide was controlled by capillary electrophoresis. Separation was performed on an Applied Biosystems model 270A-HT capillary electrophoresis system. Silica capillary (72 cm length) was used. Under these conditions, separation was achieved from anode to cathode in a citrate buffer (20 mM) at pH 2.5. Detection was at 200 nm, temperature was 30°C, and the volume injected was 2 nl.

Microsequencing
Automated Edman degradation of the purified peptide and detection of phenylthiohydantoin-derivatives were performed on a pulse liquid automatic sequenator (Applied Biosystems, model 473A).

Electrospray Mass Spectrometry
The purified peptide was dissolved in water/methanol (50/50, v/v) containing 1% acetic acid and analyzed on a VG Biotech BioQ mass spectrometer (Manchester, UK). Details of the method have been described elsewhere (Salzet et al., 1993a).

Immunocytochemical Investigations
In addition to the cells previously described in the proboscis (Verger-Bocquet et al., 1987a) and in the brain (Verger-Bocquet et al., 1987b of T. tessulatum, numerous ␣-NE immunoreactive cells are observed in the epithelium of anterior and posterior suckers (Figs. 1, panels 1 and 2). In T. tessulatum, the ␣-NE-like material exhibited subepidermal bodies located among muscular and mucous gland cells. These small cells, 6 -10 m in diameter, possess a large nucleus (3-5 m) for their size (Fig. 1, panels 4 and 5). They bear elongated processes that terminate immediately beneath the cuticle (Fig. 1,  panel 3) and are abundant (Ͼ12,000/mm 2 ). None of the processes had pores through the cuticle (Fig. 1, panel 7). At the ultrastructural level, numerous granules can be noted in the cell body and in the processes. These ovoid or somewhat crescent-shaped granules (0.2-0.3 m) are characterized by a homogenous electron-dense material (Fig. 1, panel 7). After immunogold labeling, numerous colloidal gold particles are observed in these granules (Fig. 1, panels 6 and 8).

Biochemical Investigations
Isolation of the ␣-NE Substance-Central nervous systems (1000) or suckers (400) were subjected to a peptide extraction in 1 M acetic acid, pH 2. ELISA revealed in crude extract 6.32 Ϯ 1.8 pmol of ␣-NE-like material/CNS and 15.45 Ϯ 4.6 pmol of ␣-NE-like material/sucker (after acetonic precipitation). Crude extracts were prepurified using Sep-Pak C 18 cartridges. The fraction eluted by 50% of acetonitrile was reduced 20-fold by freeze-drying and applied to a HPGPC column. Eluted fractions tested in DIA revealed a single immunoreactive zone in the two cases (CNS or suckers) corresponding to peptides with molecular mass of ϳ1-4.5 kDa (data not shown). An amount of 5.45 Ϯ 0.75 pmol of ␣-NE-like material/CNS (recovery of ϳ87%) and of 13.44 Ϯ 2.25 pmol of ␣-NE-like material/sucker (recovery of ϳ82%) was obtained. Results obtained after preadsorption of the antiserum by synthetic ␣-NE established the specificity of the immunodetection. Each immunoreactive fraction was then concentrated 20-fold and applied to a reversed-phase HPLC.
In a first step of reversed-phase HPLC on a Vydac C 18 column, ␣-NE-like substances (suckers or CNS) eluted at a same retention time (RT) comprised between 21-22 (corresponding to 26 -27% of acetonitrile) (Fig. 2). In these conditions the vertebrate ␣-NE eluted at a RT of 21.85 min. Total amount of ␣-NE-like material determined by ELISA at this step of purification was 4.85 Ϯ 0.92 pmol of ␣-NE-like material/CNS (recovery of ϳ80%) and of 12.63 Ϯ 4.25 pmol of ␣-NE-like material/sucker (recovery of ϳ81%). The immunoreactive zone containing this material (suckers or CNS) was analyzed on the  (6) and its process (8). Scale bar, 0.5 m. Panel 7, ending of the process with numerous electron dense granules. Scale bar, 0.5 m; a.j., adherens junction; s.j., septate junctions

FIG. 2. Reversed-phase HPLC separation of the immunoreactive zone to the anti-␣-NE antibody detected after HPGPC.
The immunoreactive material to the anti-␣-NE antibody eluted from the HPGPC column was loaded onto a C 18 -peptide protein column (250 mm ϫ 4.6 mm; Vydac). Elution was performed with a discontinuous linear gradient of 0 -15% acetonitrile in acidified water (0.1% trifluoroacetic acid) for 10 min, followed by a gradient of 15-45% acetonitrile in acidified water (0.1% trifluoroacetic acid) for 30 min at a flow rate of 1 ml/min. ␣-NE-like material was detected on aliquots of each fraction by the ␣-NE-DIA. The bar indicates the immunoreactive material. same column with a shallower gradient. A peak immunoreactive to anti-␣-NE, at a RT comprised between 25 and 25.3 min (corresponding to 26.25-26.48% of acetonitrile), was resolved in both cases. At this step of purification, quantification by ELISA indicated an amount of 4.15 Ϯ 0.75 pmol of ␣-NE-like material/ CNS (recovery of ϳ66%) and of 10.25 Ϯ 3.75 pmol of ␣-NE-like material/sucker (recovery of ϳ66%). In both cases (suckers or CNS), a peak was then purified to homogeneity on an ODS C 18 reversed-phase column and gave in each case a single peak at a RT of 28.3 min. Purity of the immunoreactive material (suckers, CNS) was established by capillary electrophoresis (Fig. 3). Quantification by ELISA at this step of purification indicated 3.75 Ϯ 0.86 pmol of ␣-NE-like material/CNS (recovery of ϳ60%) and of 9.86 Ϯ 2.62 pmol of ␣-NE-like material/sucker (recovery of ϳ64%).
In order to ensure that the ␣-NE-like peptide purified either from CNS or from suckers were the same compound, a coinjection on ODS C 18 reversed-phase HPLC column of the two purified peptides was performed. A single peak was eluted at a RT of 28.3 min. Moreover, a comparison of the first derivatives of absorbance of the purified ␣-NE-like peptide from CNS and from suckers at a same concentration revealed a total spectral overlapping between 190 and 300 nm and a ratio equal to 1, reflecting a very similar homology.
Characterization of the ␣-NE-like Peptide-After the final purification step, a fraction aliquot of the immunoreactive material was evaluated by Edman degradation. The sequence, established on 139 pmol of purified ␣-NE-like peptide with a sequencing yield of 95%, was Tyr-Gly-Gly-Phe-Leu-Arg-Lys-Tyr-Pro-Lys (Table II). The molecular mass of the CNS ␣-NElike peptide measured by electrospray mass spectrometry (m/z ϭ 1228.04 Ϯ 0.06 Da) is in agreement with the calculated isotopic mass (1228.68 Da) of the ␣-NE (Fig. 4). The same result was obtained with the ␣-NE-like peptide isolated from suckers (m/z ϭ 1228.3 Ϯ 0.35 versus 1228.68). Moreover, coinjection on ODS C 18 reversed-phase HPLC column of each purified peptides (CNS, suckers) and synthetic ␣-NE revealed a single peak at a RT of 28.5 min. This procedure established that the primary structure of T. tessulatum ␣-NE-like peptide is fully superposable on that of vertebrate ␣-NE peptide. Its amount was estimated to 2.45 Ϯ 1.15 pmol/CNS and 8.95 Ϯ 2.15 pmol/sucker.

DISCUSSION
The present study demonstrates that to the peptide ␣-NE isolated from the CNS and suckers of the leech T. tessulatum is structurally identical with the one identified in vertebrates (YGGFLRKYPK). The characterization of ␣-NE in the leech constitutes the first report of the presence of a prodynorphinderived peptide in an invertebrate. The same true ␣-NE was isolated and identified both from CNS and suckers of mature T. tessulatum.
␣-Neoendorphin immunoreactivity is expressed by specific cells in CNS (Verger-Bocquet et al., 1987b, neuroendocrine cells of the foregut (Verger-Bocquet et al., 1987a), and by epithelial cells at the level of the suckers in this animal. Interestingly, this is not a novel observation in that in leeches certain monoclonal antibodies reacted with cells in the CNS and with epithelial cells (Hogg et al., 1983). Such an observation of a common localization among neurons and epithelial cells is very frequent in vertebrates.
The present study also shows that ␣-NE immunoreactive cells in T. tessulatum suckers have morphological and ultrastructural characteristics close to the "releasing gland cells" of adhesive organs of the Branchiobdellids (Farnesi et al., 1981;Hogg et al., 1983;Gelder and Rowe, 1988;Weigl, 1994). These cells might constitute with the viscid gland cells a duo-gland adhesive organ (Tyler, 1976). In this regard, the sticky secretion of the viscid cells is used for substrate attachment and the secretion of the releasing cells may assist in detachment. However, we never found pores for these cells or releasing granules on the surface of the leech suckers. The same observation was made by Weigl (1994) in Branchiobdellids. The function of these releasing cells is currently not understood. Interestingly, they may be derived from nerve cells (Tyler, 1976;Gelder and Rowe, 1988), an observation that is based on similarities in structure and staining properties between them. In Branchiobdellids, the interpretation of Farnesi et al. (1981) is that these cells are neurons and the granules might be considered neurosecretory granules. Farnesi et al. (1981) suggest that the releasing cells are actually neurons with junctions on the viscid adhesive cell ducts to control the release of the viscid secretion granules. In T. tessulatum, the hypothesis of the nerve cell nature of these releasing cells is sustained by our observations showing that their intracytoplasmic granules (0.2-0.3 m) contain ␣-NE, identical to the one present in specific neurons. It is known that in leeches, neurons and epithelial cells are derived from the same blastomeres (Weisblat et al., 1980). In Hirudinea, Stewart et al. (1985) suggest that the epidermal cells recognized by monoclonal antibodies specific to the CNS are indeed peripheral neurons. Their study with the monoclonal antibody Lan 3-6 shows that during the embryonic development the same labeled epithelial cells possess an apical dendrite and a basal process; therefore, they are neurons. Stewart et al. (1985) suggest that the expression of the antigens in the axon, is either absent or below the level detectable with these technique in this cell. To answer this question anatomical experiences, e.g. anterograde transport, could be conducted. In T. tessulatum, as in Branchiobdellids (Weigl, 1994), continuity between these cells and necks of the viscid cells are scarce, which reinforces the improbable nervous control as proposed by  Leech ␣-Neoendorphin Farnesi et al. (1981). If the nerve nature of these ␣-NE positive sucker cells is confirmed and considering their abundance, we could propose that they exert an important role either in the control of the adhesivity or in one of the multiple sensorial functions of the suckers. The existence of secretory granules suggests that ␣-NE can be released in the intracellular spaces and acts via a paracrine mechanism. Although that relationship between Branchiobdellids and the other clitellates remains unclear (Holt, 1989;Brinkhurst and Gelder, 1989), the discovery in T. tessulatum of cells resembling the releasing cells of Branchiobdellids argues in favor of a link between adhesive organs of Branchiobdellids and suckers of Hirudinea. Furthermore, in the CNS of the leech, the ␣-NE secretion of neurons into the circulatory system at the neurohemal site suggests a hormonal role. Alternatively, an action as neurotransmitter or neuromodulator at the level of nerve endings located in the dorsal commissure or the neuropile is also possible, as in the presence of both mechanisms in these animals.
The fact that ␣-NE is highly conserved in course of evolution from annelids to vertebrates suggests an essential function for this peptide. Actually we know that peptides conserved during evolution appear to keep an action on a same physiological function. This argument is sustained by two neuropeptides, acting on the osmoregulation, isolated in both T. tessulatum and vertebrates, i.e. the lysine-conopressin (Salzet et al., 1993a) and the angiotensin II (Salzet et al., 1995b). In vertebrates, it is known that opioids are involved in variety of physiological functions and interact with endocrine system. The endogenous opioids peptides appear to have a role in the interaction between the CNS and the immune system (Scharrer, 1990;Scharrer et al., 1994). In invertebrates, until now, by contrast, nothing is known about the role of ␣-NE. Consequently, the leech provides for a good model system in which to study this conserved peptide as well as the presence of a possibly conserved precursor, i.e. prodynorphin.
The existence of an ancestral proenkephalin gene is supported by the isolation of enkephalin peptides in invertebrate taxa, i.e. the crustacea Carcinus maenas (Luschen et al., 1991), in the mollusk Mytilus edulis (Leung and Stefano, 1984), and in the leech T. tessulatum (Salzet et al., 1994). However, the ratio of Leu-enkephalin and Met-enkephalin in C. maenas and T. tessulatum is 3:1 and 2:1, respectively, whereas in vertebrates and Mytilus edulis Met-enkephalin is the major opioid peptide. Moreover, immunocytochemical studies performed at the level of T. tessulatum brains revealed that Leu-enkephalin and Metenkephalin are not detected in the same cells (Verger-Bocquet et al., 1987b). From these observations several hypothesis can be given. First, unlike in vertebrates, there are two separated genes, one coding for Met-enkephalin and the other for Leuenkephalin. Second, the two pentapeptides come from a unique ancestral proenkephalin precursor and the expression of Metor Leu-enkephalin is due to different posttranslation processing mechanisms; Leu-enkephalin is not expressed in cells expressing Met-enkephalin and vice versa. If we consider the first hypothesis, the question that is raised is: what can be the second opioid precursor, i.e. a prodynorphin or a POMC-like precursor?
Peptides yielded from POMC processing have been identified by immunocytochemistry and radioimmunoassay in invertebrates, i.e. ACTH and ␤ endorphin in the mollusk Planorbarius corneus (Franchesi and Ottaviani, 1992) and in insects (Duve and Thorpe, 1988), ␤ endorphin in the flatworms (Reuter and FIG. 4. Electrospray mass spectrum of the purified ␣-NE-like peptide from the central nervous system of T. tessulatum. Peaks at m/z ϭ 308, m/z ϭ 410.4, and m/z ϭ 615 are multiply charged ions with four, three, or two charges corresponding to a mass of 1228.04 Ϯ 0.06 Da. The peak at m/z ϭ 571.3 corresponds to an internal mass standard (gramicidin). Gustafsson, 1989) and MSH, and endorphins in annelids (Dhainaut-Courtois et al., 1986;Verger-Bocquet et al., 1987a, 1987b. These results envisage the existence of a POMClike precursor in invertebrates. Complete molecular studies have been performed in the trematod Schistosoma mansoni, where a gene related to the vertebrate POMC has been cloned . However, according to Duvaux-Miret and Capron (1991), the extremely high homology between the genes of S. mansoni and vertebrates may be the result of a transfer of genetic material from the host toward the parasite, probably by a viral mechanism. The other possibility is a high evolutionary conservation through selective pressure, which has allowed S. mansoni to synthesize molecules highly similar to the host endogenous signals; this could be used to avoid the host defense (Duvaux-Miret and Capron, 1991). Recently, a peptide belonging to the POMC vertebrate family has been isolated in T. tessulatum brains. This peptide is related to the Vertebrate ␥-MSH and presented ϳ80% sequence homology with vertebrates ␥ 1 -MSH (Salzet et al., 1994). However, in the leech T. tessulatum, the ␥-MSH-like peptide seems be localized in a multipeptidic precursor different from POMC (Salzet et al., 1993b(Salzet et al., , 1995b. Other peptides present in this multiple hormones precursor have recently been isolated in leeches, i.e. AI and AII (Laurent et al., 1995;Salzet et al., 1993bSalzet et al., , 1995b. Therefore, these results tend to favor the absence in leeches of a true POMC precursor.
If we consider that POMC is absent in annelids, the hypothesis of a prodynorphin-like precursor can be suggested. Furthermore as suggested by Stefano et al. (1989), detection by immunocytochemistry of the prodynorphin end-products makes it difficult to imagine that proenkephalin is the only source of Leu-enkephalin in neural tissue. It could also derive from a prodynorphin. Although epitopes of the different prodynorphin end-products are immunologically recognized in nervous system of different invertebrates, their chemical nature is unknown. However, in the leech T. tessulatum, ␣-NE and the dynorphin-like peptides are expressed in different neurons (Verger-Bocquet et al., 1987b. Only the characterization of the ␣-NE precursor would demonstrate precisely if a prodynorphin-like gene existed in annelids. Our discovery of the ␣-NE confirms the very ancient stature of opioids in metazoans and the hypothesis emitted by Stefano et al. (1989) that the highly regulated vertebrate immune system probably had its origin in the invertebrates. These molecules would be used since the beginning of the evolution to start a type of integrated reply in order to maintained the body homeostasis (Franchesi and Ottaviani, 1992) and notably in neuroimmunity reactions (Stefano, 1992), as well as documented neuroregulatory actions (Kream et al., 1980;Stefano, 1982).