Báo cáo khoa học: Identification of proNeuropeptide FFA peptides processed in neuronal and non-neuronal cells and in nervous tissue potx

Identification of proNeuropeptide FF A peptides processed in neuronal and non-neuronal cells and in nervous tissue Elisabeth Bonnard, Odile Burlet-Schiltz, Bernard Monsarrat, Jean-Philippe Girard and Jean-Marie Zajac Institut de Pharmacologie et de Biologie Structurale, Toulouse, France Peptides which should be generated from the neuropeptide FF (NPFF) precursor were identified in a neuronal (human neuroblastoma SH-SY5Y) cell line and in COS-7 cells after transient transfection of the human proNPFF A cDNA and were compared with those detected in the mouse spinal cord. After reverse-phase high performance liquid chromatogra- phy of soluble material, NPFF-related peptides were im- munodetected with antisera raised against NPFF and identified by using on-line capillary liquid chromatography/ nanospray ion trap tandem mass spectrometry. Neuronal and non-neuronal cells generated different peptides from the same precursor. In addition to NPFF, SQA-NPFF (Ser-Gln-Ala-Phe-Leu-Phe-Gln-Pro-Gln-Arg-Phe-amide) and NPAF were identified in the human neuroblastoma while only NPFF was clearly identified in COS-7 cells. In mouse, in addition to previously detected NPFF and NPSF, SPA-NPFF (Ser-Pro-Ala-Phe-Leu-Phe-Gln-Pro-Gln-Arg-Phe- amide), the homologous peptide of SQA-NPFF, were characterized. These data on intracellular processing of proNeuropeptide FFA are discussed in regard to the known enzymatic processing mechanisms. Keywords: neuropeptide FF; electrospray tandem mass spectrometry; precursor processing; neuroblastoma. Neuropeptide FF (NPFF, FLFQPQRFamide) is a mam- malian amidated neuropeptide, originally isolated from bovine brain and characterized as a modulator of endo- genous opioid functions [1,2]. Two precursors, proNPFF A and proNPFF B encoding peptides possessing the PQRF- amide sequence, have been cloned in mammals [3,4]. The proNPFF A precursor at basic proteolytic sites should generate two PQRFamide containing peptides [3] and the proNPFF B , also called RFamide-related peptides precursor [5], contains a PQRFa sequence and an LPLRFa-contain- ing peptide. There is a large body of evidence that NPFF exhibits antiopioid properties; in rodents, morphine-induced anal- gesia decreased following administration of NPFF or NPFF analogues and increased, as stress-induced analgesia, in response to anti-NPFF antibody administration [6–8]. In contrast, intrathecal injections of NPFF analogues induced a long-lasting analgesia [9,10] by increasing opioid peptide release in the spinal cord through the functional blockade of presynaptic delta-opioid autoreceptors [11,12]. Recent data provided evidence that opioid and NPFF endogenous systems exert a tonic activity, NPFF counteracting tonic opioid analgesia under resting conditions [13]. NPFF is also implicated in morphine tolerance, morphine abstinence and also in several physiological processes, such as body thermoregulation, food intake and blood pressure regula- tion [7,14–21]. These pharmacological effects are mediated by two G-protein-coupled receptors, NPFF 1 and NPFF 2 , cloned in human and rat [22–25]. Pharmacological characteriza- tion of these receptors in recombinant cell lines showed a better selectivity of peptides deduced from proNPFF A sequence for NPFF 2 receptors binding, whereas proN- PFF B -derived peptides displayed a greater affinity for NPFF 1 receptors [26]. Autoradiographic studies per- formed on rat CNS with highly selective radioligands revealed the localization of both receptors in central nervous areas implicated in pain transmission [27]. The existence of two peptidergic systems of neurotransmission, mediated through NPFF 1 and NPFF 2 receptor stimula- tion by peptides generated by proNPFF B and proNPFF A processing, respectively, could explain the complex pharmacological effects of NPFF. The characterization of NPFF-related peptides generated by NPFF precursors processing is essential to identify peptides candidate to the role of neurotransmitter. This study focused on proNPFF A processing. In humans, bovines and rodents, proNPFF A contains consensus sequences for a processing by protein convertases [28,29] (Fig. 1). A sequen- tial recruitment of carboxypeptidases and peptidylglycine-a amidating monooxygenase could be implicated in the production of amidated active NPFF-related peptides. According to these processing rules, two families of NPFF- related peptides should be generated by proNPFF A process- ing: (a) N-terminal extended NPFF undecapeptides and (b) N-terminal extended NPSF (SLAAPQRFamide)-derived peptides, 11 or 18 amino acids long. In previous studies, NPFF and NPAF (AGEGLSSPFWSLAAPQRFamide) were isolated from bovine brain [6]. More recently, NPFF and NPSF were identified in rodents [30] and a longer Correspondence to J M. Zajac, Institut de Pharmacologie et de Bio- logie Structurale, 205 route de Narbonne, 31077 Toulouse, France. Fax: + 33 5 61175994, Tel.: + 33 5 61175911, E-mail: jean-marie.zajac@ipbs.fr Abbreviations: MS/MS, tandem mass spectrometry; NPFF, neuro- peptide FF (FLFQPQRFamide); CNS, central nervous system. (Received 23 June 2003, revised 22 August 2003, accepted 3 September 2003) Eur. J. Biochem. 270, 4187–4199 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03816.x peptide, NPA-NPFF (NPAFLFQPQRFamide), was quan- tified as the most abundant in rat spinal cord [31]. NPFF and NPSF were also identified in the mouse spinal cord [31], NPAF and NPSF in human cerebrospinal fluid [32]. These reports suggested a processing of proNPFF A to both octapeptides (NPFF and NPSF) despite the absence of consensus processing sites at their N-terminal end. The identification of NPFF-related peptides actually synthesized in neurones should help to understand their role in neurotransmision. Thus, we investigated the processing of proNPFF A in SH-SY5Y human neuroblastoma cells and COS-7 (nonneuronal) cells transiently transfected with the human proNPFF A cDNA. The pattern of intracellular NPFF-related peptides isolated from cell extracts was compared with that observed in mouse spinal cord. NPFF-related peptides from cells and tissues were extracted and purified by RP-HPLC, assessed by radioimmunoassay and identified by using on-line capillary HPLC/nanospray ion trap tandem mass spectrometry (nanospray MS/MS). Materials and methods Chemicals NPFF-related peptides (Table 1) were synthesized by the solid-phase method using Fmoc chemistry with an automatic synthesizer (430 A Applied Biosystem) and purified by reverse-phase HPLC as described previously [33]. Fmoc amino acid derivatives were purchased from Bachem, France. Iodination of 1DMe ([D.Tyr1(NMe)Phe3]NPFF) was performed according to Dupouy et al. [34]. hproNPFF A cDNA cloning and vector construction The SMART PCR cDNA synthesis kit (Clontech, Palo Alto, CA, USA) was used to generate high yields of full- length cDNA from 1 lg human lymph node total RNA (Clontech). Amplification of the hproNPFF A cDNA from human LymphNode cDNA was performed by PCR with an Advantage cDNA PCR kit (Clontech), using 100 ngÆmL )1 of cDNA and 400 nm of each primer: 5¢Bgl II-hproNPFF A :5¢-CGCAGATCTAGCATGGATT CTAGGCAGGCTGCTGC-3¢ and 3¢Apa-hproNPFF A : 5¢-GCGGGGCCCTTCTTCCCAAAGCGTTGAGGGG CAG-3¢, targeted to the 5¢-and3¢-end of the hproNPFF A coding sequence, respectively, in a PTC-150 MiniCycler (MJ Research Inc.), with 25 cycles consisting of 30 s at 94 °Cand30sat68°C. PCR products were cloned in pEGFPn3 (Clontech) and sequenced on both strands. Tissue extraction Mouse spinal cord tissue. Animals were handled in accordance with standard ethical guidelines (NIH Guide for Care and Use of Laboratory Animals,1985).Threemice were killed by decapitation and the cervical segment of spinal cord was dissected in ice-cold 0.9% NaCl (106 mg of tissue), and frozen. All tissues were stored at )80 °C until used. Cervical segments were chosen for their high NPFF- like immunoreactivity content [31]. The extraction proce- dure was performed on frozen tissue: sonication in 0.1 M HCl, followed by boiling for 10 min. Tissue homogenates were buffered to pH 7.4 with 2 M Tris pH 7.4, at a final concentration of 25 mg tissueÆmL )1 and centrifuged at 10 000 g for 10 min at 4 °C. The supernatant was stored at )80 °C until radioimmunoassay. SH-SY5Y cells. SH-SY5Y cells were grown in Dulbecco’s modified Eagle’s medium supplemented with Glutamax-1, glucose (4.5 gÆL )1 ), 10% fetal bovine serum, penicillin (100 UÆmL )1 ) and streptomycin (100 lgÆmL )1 ). Cells were seeded into 35 mm Petri dishes, at a density of 7.5 10 5 cells per dish. Twenty-four hours later, the medium was removed; cells were washed with cold NaCl/P i and scraped in 0.1 M HCl. After sonication, lysates were centrifuged at Fig. 1. Partial amino acid sequence of proNPFF A in human and mouse. NPFF-related peptides predicted by consensus dibasic processing sites are shown in bold. NPFF and NPSF are eight amino acid peptides common in mammals (boxed). In mice, an eleven amino acid long NPSF-derived peptide could be processed: QFW-NPSF. Table 1. Analytical parameters of synthetic NPFF-related peptides. The IC 50 values were obtained from independent experiments (n ‡ 3). The RIA detection limit was 10 fmol for NPFF-derived peptides, 55 fmol for hNPAF and 520 fmol for QFW-NPSF. HPLC fractions were collected every 60 s for gradient 1(HPLC-1) and 2 (HPLC-2), every 30 s for gradient 3 (HPLC-3). Theoretical monoisotopic mass (Da) RIA IC 50 (fmol) HPLC retention time (min) Observed [M + 2H] 2+ (m/z)HPLC-1 HPLC-2 HPLC-3 SQA-NPFF (SQAFLFQPQRFamide) 1366.7 32.8 ± 1.3 – 41.85 54.91 684.5 NPFF (FLFQPQRFamide) 1080.6 28.5 ± 1.3 33.7 42.79 56.42 541.3 hNPAF (AGEGLNSQFWSLAAPQRFamide) 1977.0 264.4 ± 29.5 – 44.42 66.33 990.0 SPA-NPFF (SPAFLFQPQRFamide) 1335.7 29.9 ± 3.2 36.78 – – 668.9 QFW-NPSF (QFWSLAAPQRFamide) 1348.7 1500.2 ± 149.3 36.12 – – 675.4 4188 E. Bonnard et al. (Eur. J. Biochem. 270) Ó FEBS 2003 10000 g for 20 min at 4 °C. Surpernatant was stored at )20 °C until radioimmunoassay. COS-7 cells. COS-7 cells were grown in 35 mm Petri dishes in Dulbecco’s modified Eagle’s medium supplemented with Glutamax-1, glucose (1 gÆL )1 ), 10% fetal calf serum, penicillin (100 UÆmL )1 ) and streptomycin (100 lgÆmL )1 ). When cells reached 50–80% confluence, they were transi- ently transfected with hpro-NPFF A cDNA using Lipofect- AMINE Reagent (Invitrogen). Cells were transfected with 4 lg per dish of pEGFP-hpro-NPFF A according to the manufacturer’s instruction. Forty hours later, the medium was removed; cells were washed with cold NaCl/P i , extractedin1 M acetic acid, sonicated and centrifuged at 10 000 g for 20 min at 4 °C. Supernatant was stored at )20 °C until radioimmunoassay was performed. Radioimmunoassay The procedure was carried out as described previously [13]. Dilutions of synthetic peptides, cell or tissue extracts were incubated overnight at 4 °C with NPFF antiserum (1 : 150 000 final dilution) and [ 125 I][D.Tyr1(NMe)- Phe3]NPFF (40 pmol). Non-specific binding was deter- mined with synthetic NPFF (100 pmol per assay). The limit of detection of NPFF-IR material was estimated to be between 8 and 12 fmol. Analysis of the binding characteristics of the NPFF antiserum indicated that, among all the possible derivatives of proNPFF A and proNPFF B , only NPAFLFQPQRF- amide, SQAFLFQPQRFamide and SPAFLFQPQRF- amide interfered in the assay (100, 100 and 86% cross-reactivity as compared with 100% with NPFF, respectively). SLAAPQRFamide-derived peptide cross- reacted at 3–10%. Other RFamide peptides, like Met- EnkRFamide, FMRFamide and the nonamidated NPFF (NPFF-OH) showed less than 0.1% cross-reactivity. Degradation of synthetic SQA-NPFF in COS-7 cells COS-7 cells that had reached 50–80% confluency were washed with NaCl/P i andincubatedwithNaCl/P i plus EDTA, 1 m M at 37 °C for 5 min. Cells were then centrifuged at 500 g, 5 min, resuspended in NaCl/P i and disrupted by nitrogen cavitation. Briefly, cells are intro- duced in a bomb and equilibrated with nitrogen gas at 30 atm pressure for 10 min. Sudden decompression resul- ted in a complete disruption of cells with minimum damage of intracellular organelles [35,36]. Cells lysates were incu- bated at 37 °C for 10, 30, 60 120, 180 min with or without protease inhibitors (phenylmethylsulfonyl fluoride, 2 m M , bestatin, 0.1 m M ) and 50 pmol of synthetic SQA-NPFF. At the end of the incubation period, cell lysates were extractedinto1 M acetic acid and centrifuged at 4 °Cand 8000 g for 20 min. Supernatant was stored at )20 °C until analytic procedure. Reverse-phase high pressure liquid chromatography (RP-HPLC) Mouse spinal cord. The procedure was carried out as described previously [13]. Tissue extracts were purified on C18 Sep-Pak cartridges (Waters). Samples were loaded on cartridge and washed with 0.088% trifluoroacetic acid in H 2 O/CH 3 CN (80 : 20 v/v) and eluted with 0.088% trifluoroacetic acid in H 2 O/CH 3 CN (25 : 75 v/v). The eluates were lyophilized, diluted in 500 lL of mobile phase and applied to a C8 Aquapore RP300 Brownlee (4.6 · 220 mm, Perkin Elmer) equilibrated previously with 70% A and 30% B at a flow rate of 400 lLÆmin )1 . Solvent A consisted of 0.088% trifluoroacetic acid in H 2 O and solvent B was 0.088% trifluoroacetic acid in H 2 O/ CH 3 CN (25 : 75 v/v). Separation was performed by using isocratic elution at 30% B for 6 min, followed by a linear gradient of 30–60% B for 50 min (gradient 1). HPLC fractions (HPLC-1) corresponding to the retention time of synthetic NPFF-related peptides were collected and concentrated for radioimmunoassay. NPFF-IR fractions were subjected to a second lHPLC separation on a C18 column, as previously described [31] before MS/MS analyses. Cell extracts Cell extracts were lyophilized, diluted in 500 lL of mobile phase and applied to a C8 Spheri-5 RP-8S 5 lmBrownlee (2.1 · 220 mm) previously equilibrated with 98% of mobile phase A and 2% of mobile phase B, at a flow rate of 400 lLÆmin )1 . Separation of SH-SY5Y cells extract was achieved using isocratic elution at 2% B for 6 min, followed by a linear gradient of 2–80% B for 50 min (gradient 2). Fractions (HPLC-2) co-eluted with SQA- NPFF, NPFF and hNPAF were subjected to a second HPLC separation using a linear gradient of 0–44% B for 45 min, followed by an isocratic elution for 15 min. B reached 54% by 1 min and an isocratic elution was achieved for 10 min (gradient 3). HPLC fractions (HPLC-3) corresponding to the retention time of synthetic NPFF-related peptides were collected and concentrated for radioimmunoassay. Separation of COS-7 cells extract was achieved by gradient 1 procedure, followed by gradient 3 procedure. To ensure that tissue and cell extracts were not contaminated by NPFF-IR material, a blank run on the RP-HPLC column was performed before each sample RP-HPLC run and assessed by RIA. On-line capillary HPLC/nanospray ionization MS/MS HPLC fractions were concentrated under vacuum and analyzed by on-line capillary HPLC/nanospray ionization MS/MS. The sample was injected onto a C18 PepMap TM (LC Packings) column (75 lm · 150 mm). The separation was performed using an isocratic elution at 0% B for 2 min, followed by a linear gradient of 0–40% B in 30 or 40 min, at a flow rate of 150 nLÆmin )1 . Two different gradient slopes were used in 40 min. Solvent A consisted of 0.1% formic acid in H 2 O/CH 3 CN (99 : 1 v/v) and B was 0.1% formic acid in H 2 O/CH 3 CN (10 : 90 v/v). The eluent was injected into an LCQ Deca ion trap mass spectrometer (ThermoFinnigan, San Jose, CA, USA) through a nano- flow needle (New Objective, Cambridge, MA, USA) at 2.0 kV. MS/MS data were acquired using a three m/z unit ion isolation window and a relative collision energy of 35%. Ó FEBS 2003 Processing of proNeuropeptide FF A (Eur. J. Biochem. 270) 4189 Results RP-HPLC and mass spectrometry analyses of NPFF-related synthetic peptides In an attempt to identify NPFF-related peptides in cell and tissue extracts, analytical characteristics of synthetic peptides were initially determined on RP-HPLC and mass spectrometry. Table 1 shows the retention times of human and mouse synthetic peptides in three different RP-HPLC procedures. Each peptide was further analyzed by on-line capillary HPLC/nanospray MS/MS (Fig. 2). The fragmen- tation of each synthetic NPFF-related peptide (double- charged precursor ion reported in the Table 1) gave rise to a Fig. 2. Mass spectrometry analyses of synthetic NPFF-related peptides. One hundred femtomoles of each synthetic NPFF-related peptide were analyzed by on-line capillary HPLC/nanospray ion trap MS/MS. The MS/MS spectra of NPFF (A), SPA-NPFF (B), SQA-NPFF (C), QFW- NPSF (D) and hNPAF (E) were acquired from the [M + 2H] 2+ precursor ion at m/z 541.3 (NPFF), m/z 668.9 (SPA-NPFF), m/z 684.5 (SQA- NPFF), m/z 675.4 (QFW-NPSF) and m/z 990.0 (hNPAF). Fragment ion peaks are labelled according to Biemann’s nomenclature [49]. *Loss of NH 3 from the y and b ions. The peptide sequence and fragmentation pattern for each peptide is indicated at the top. 4190 E. Bonnard et al. (Eur. J. Biochem. 270) Ó FEBS 2003 series of y and b fragment ions, from which the most abundant was in each case the singly charged y 4 fragment ion at m/z 546.3, corresponding to the C-terminal tetrapep- tide PQRFamide. These observations led us to consider the y 4 fragment ion in the MS/MS spectrum of each precursor ion, as a criterion for the identification of NPFF-related peptides in biological samples. Other criteria for the identification of NPFF-related peptides were the retention time in each HPLC system, the MS/MS fragmentation pattern of the double-charged precursor ion which is characteristic of each peptide and the RIA signal. RP-HPLC profiles of NPFF-IR in cell extracts SH-SY5Y cell extracts were applied on gradient 2 and fractions containing NPFF-immunoreactivity were separ- ated on gradient 3 (Fig. 3A). Three immunoreactive peaks corresponding to retention time of 51, 55 and 67 min were observed. The second peak coeluted with synthetic SQA- NPFFandthethirdwithhNPAF(Table1). COS-7 cell extracts were separated on gradient 1 and NPFF-IR fractions were subjected to a second HPLC procedure on gradient 3 (Fig. 3B). Three immunoreactive peaks were obtained with retention times of 55.5, 60 and 67 min. Two corresponded to the retention time of synthetic peptides: SQA-NPFF or NPFF for peak 1 and hNPAF for peak 3. Quantification of NPFF-IR in HPLC fractions was assessed by radioimmunoassay (Table 2). No NPFF-IR material was detected in COS-7 cells either non transfected or transiently transfected with pEGFPn3 vector (data not shown). The identification of NPFF-IR molecular forms in SH-SY5Y (peaks 2 and 3) and COS-7 transfected cells (peaks 1 and 3) was provided using MS/MS analyses. Identification of SQA-NPFF and hNPAF by capillary HPLC/nanospray MS/MS HPLC fractions corresponding to the NPFF-IR peak 2 from SH-SY5Y cells extract were pooled, concentrated under vacuum and analyzed by on-line capillary HPLC/ nanospray MS/MS. No peak at m/z 684.5 corresponding to the expected double-charged ion of SQA-NPFF could be detected in the MS spectrum. However, the search for the specific y 4 fragment ion at m/z 546.3 in the MS/MS spectrum allowed to extract a signal of low intensity from the background noise at the retention time of synthetic SQA-NPFF (Figs 4A,B). The corresponding MS/MS spec- trum (Fig. 4C) displayed, in addition to the y 4 fragment ion, b and y fragment ions compatible with the fragmentation Fig. 3. HPLC profile of NPFF-IR in SH-SY5Y and COS-7 cell extracts. Acid SH-SY5Y extracts were applied on a C8 column and separated first on gradient 2 at a flow rate of 400 lLÆmin )1 . Collected NPFF-immunoreactive fractions were pooled, concentrated and sep- arated on gradient 3 (A). Acid COS-7 extracts were first separated on a C8 column on gradient 1. Collected fractions corresponding to the retention time of synthetic NPFF-related peptides were pooled and subjected to a second HPLC separation on gradient 3 (B). Elution positions of synthetic NPFF-related peptides are indicated by arrows. The gradient is represented by a dotted line. Table 2. Separation by RP-HPLC of cell and tissue extracts and quantitative analyses of NPFF-IR peaks by RIA. SH-SY5Y cell extract was applied on gradient 2 and collected fractions corresponding to the retention time of synthetic NPFF-related peptides were pooled and separated on the gradient 3. COS-7 cells transiently transfected by hpro-NPFF A were extracted and separated on gradient 1 and gradient 3. Mouse cervical spinal cord extracts were purified on a Sep-Pack Cartridge before separated on gradient 1. NPFF-related peptides were estimated in HPLC-3 and HPLC-1 fractions by RIA. SH-SY5Y neuroblastoma hpro-NPFF A transfected COS-7 Mouse cervical spinal cord HPLC-3 retention time (min) Cell extract (fmolÆmL )1 ) HPLC-3 retention time (min) Cell extract (fmolÆmL )1 ) HPLC-1 retention time (min) Tissue (fmolÆmg )1 ) SQA-NPFF 54.5–55.5 54.6 54.5–55.5 195.7 – – NPFF 56–57 <10.4 56–57 189.6 34–35 2.36 hNPAF 66–67 34.0 66–67 31.9 – – SPA-NPFF/QFW-NPSF – – – – 38 1.43 Ó FEBS 2003 Processing of proNeuropeptide FF A (Eur. J. Biochem. 270) 4191 pattern of the synthetic SQA-NPFF (Fig. 2C). HPLC fractions corresponding to the peak 1 from COS-7 cell extract were subjected to the same procedure. In this case, neither MS nor MS/MS analyses showed SQA-NPFF. Further analyses were performed on the HPLC fractions flanking the peak 1, without identifying SQA-NPFF. HPLC fractions corresponding to the NPFF-IR peak 3 from SH-SY5Y cell extract were pooled, concentrated under vacuum and analyzed by on-line capillary HPLC/nanospray MS/MS. The reconstructed ion chroma- togram of the specific y 4 fragment ion generated by the fragmentation of the double-charged ion at m/z 990.0 of the expected hNPAF (Fig. 4E) shows a peak at the retention time of synthetic hNPAF (Fig. 4D). The corresponding MS/MS spectrum (Fig. 4F) shows a fragmentation pattern superimposable to that obtained with synthetic hNPAF (Fig. 2E) thus unambiguously identifying this peptide in SH-SY5Y cell extracts. HPLC fractions corresponding to Fig. 4. Identification of SQA-NPFF and hNPAF in SH-SY5Y cell extracts. Reconstructed ion chromatograms of the fragment ion at m/z 546.3 generated during the MS/MS analysis of double-charged precursor ions at m/z 684.5 from 100 fmol of synthetic SQA-NPFF (A), at m/z 684.5 from HPLC fractions of NPFF-IR peak 2 (B), at m/z 990.0 from 100 fmol of synthetic hNPAF (D) and at m/z 990.0 from HPLC fractions of NPFF-IR peak3(E)andMS/MSspectraofthe[M+2H] 2+ precursor ions at m/z 684.5 (C), and at m/z 990.0 (F) in HPLC fractions of SH-SY5Y extract. Fragment ion peaks are labelled according to Biemann’s nomenclature. *Loss of NH 3 from the y and b ions. The peptide sequence and fragmentation pattern for each peptide is indicated at the top. 4192 E. Bonnard et al. (Eur. J. Biochem. 270) Ó FEBS 2003 the peak 3 from COS-7 cell extract were subjected to the same procedure. Neither MS nor MS/MS analyses unam- biguously identified hNPAF. Identification of NPFF by capillary HPLC/nanospray MS/MS NPFF was searched in HPLC fractions corresponding to the NPFF-IR peak 1 from COS-7 cell extract. The capillary HPLC/nanospray MS/MS analysis of these fractions allowed to detect the y 4 fragment ion at m/z 546.3 in the MS/MS spectrum of the double-charged ion at m/z 541.3 corresponding to NPFF (Fig. 5B) at the retention time of synthetic NPFF (Fig. 5A). The MS/MS spectrum obtained from the cell extract (Fig. 5C) was identical to the fragmentation pattern of the synthetic NPFF (Fig. 2A) identifying NPFF in the samples. Fig. 5. Identification of NPFF in SH-SY5Y and COS-7 cell extracts. Reconstructed ion chromatograms of the fragment ion at m/z 546.3 generated during the MS/MS analysis of double-charged precursor ions at m/z 541.3 from 100 fmol of synthetic NPFF (A,D), HPLC fractions of NPFF-IR peak 1 of COS-7 cell extract (B), HPLC fractions of NPFF-IR peak 2 of SH-SH5Y extracts (E), and MS/MS spectra of the [M + 2H] 2+ precursor ions at m/z 541.3, in HPLC fractions of COS-7 (C) and SH-SY5Y (F) cell extracts. Ó FEBS 2003 Processing of proNeuropeptide FF A (Eur. J. Biochem. 270) 4193 The same procedure was applied to the HPLC fractions corresponding to the NPFF-IR peak 2 of SH-SY5Y cell extract. A weak signal of the m/z 546.3 specific y 4 fragment ion was detected at the retention time of synthetic NPFF (Fig. 5D,E). The corresponding MS/MS analysis of this peak (Fig. 5F) revealed, in addition to the y 4 fragment ion, a fragmentation pattern compatible with that of the synthetic NPFF (Fig. 2A). Taking into account the retention times and the fragmentation pattern observed, these results indicate that NPFF is present in the SH-SY5Y cell extracts. Degradation of SQA-NPFF Synthetic SQA-NPFF was incubated in COS-7 cell lysates in order to investigate the putative enzymatic degradation of SQA-NPFF into NPFF. Time-course experiments per- formed at 37 °C indicated that synthetic SQA-NPFF Fig. 6. Degradation of synthetic SQA-NPFF in COS-7 cells. COS-7 cell lysate was incubated with 50 pmol of synthetic SQA-NPFF, 2 m M phenylmethanesulfonyl fluoride and 0.1 m M bestatin for 10 min at 37 °C. After extraction and separation on gradient 1, 34–35 min HPLC fraction was analyzed by capillary HPLC/nanospray ion trap MS/MS. Reconstructed ion chromatograms of the fragment ion at m/z 546.3 generated during the MS/MS analysis of double-charged precursor ions at m/z 541.3 from 100 fmol of synthetic NPFF (A), at m/z 684.5 from 100 fmol of synthetic SQA-NPFF (D), at m/z 541.3 (B) and m/z 684.5 (E) in HPLC fraction of COS-7 cell extract. MS/MS spectra of the [M + 2H] 2+ precursor ions at m/z 541.3 (C) and m/z 684.5 (F) in HPLC fractions of COS-7 cell extract. 4194 E. Bonnard et al. (Eur. J. Biochem. 270) Ó FEBS 2003 (50 pmol) was quickly degraded in the absence of protease inhibitors, as no NPFF-IR was detected 10 min after incubation with cell lysates. At this time, the incubation in the presence of 2 m M phenylmethanesulfonyl fluoride and 0.1 m M bestatin prevented partially SQA-NPFF degrada- tion, as 208 fmol of NPFF-IR were detected in HPLC fraction coeluting with both synthetic peptides SQA-NPFF and NPFF. After 30-min incubation with SQA-NPFF and cell lysates in the presence of phenylmethanesulfonyl fluoride and bestatin, no NPFF-IR was detected. The capillary HPLC/nanospray MS/MS analysis of cell lysates incubated for 10 min with peptidase inhibitors was performed and allowed to detect the specific y 4 fragment ion at m/z 546.3 in the MS/MS spectrum of the double-charged ion at m/z 541.3 (Fig. 6B) at the retention time of synthetic NPFF (Fig. 6A). The corresponding MS/MS fragmenta- tion pattern (Fig. 6C) was identical to the fragmentation pattern of synthetic NPFF (Fig. 2A). Similarly, the MS/MS analysis of the double-charged ion at m/z 684.5, corres- ponding to the SQA-NPFF (Fig. 6D–F) indicated that SQA-NPFF was not totally degraded. Even though the signal intensity of SQA-NPFF was weak, signal intensities observed for the m/z 541.3 and 684.5 ions also indicated that NPFF was detected in an approximately 20-fold higher level than SQA-NPFF. Identification of NPFF-related peptides in mouse spinal cord tissue extracts The proNPFF A processing was investigated in mouse spinal cord extracts. We have reported previously the presence of NPFF-related octapeptides NPFF and NPSF in mouse spinal cord [31] but larger peptides should be present in this tissue as the mouse proNPFF A contains cleavage sites predicted to generate the undecapeptide SPA-NPFF and the N-terminal extended form of NPSF, QFW-NPSF. The RP-HPLC profile of NPFF-IR in mouse cervical spinal cord extract is reported on Fig. 7. Three immuno- reactive peaks with retention times of 34, 38 and 41 min were obtained (Table 1). Peaks 1 and 2 coeluted with synthetic NPFF and SPA-NPFF, respectively. Because the difference between the retention time of synthetic SPA-NPFF and QFW-NPSF was very tight, endogenous QFW-NPSF was also searched in the peak 2. On line capillary HPLC/nanospray MS/MS analyses of NPFF-immunoreactive peaks 1 and 2 are reported in Fig. 8. The data obtained from the NPFF-IR peak 1 HPLC fraction show that NPFF is identified in mouse spinal cord (Fig. 8A–C). Similarly, data obtained from the NPFF-IR peak 2 HPLC fraction unambiguously identified SPA-NPFF (Fig. 8D–F). The identification of QFW- NPSF in peak 2 was more difficult. However, the detection of the characteristic y 4 fragment ion at m/z 546.3 in the MS/MS spectrum of the precursor ion of QFW-NPSF at m/z 675.4 and the signal retention time corresponding to the synthetic peptide (Fig. 8G–I) were convincing data for the identification of QFW-NPSF in the sample. The coelution of SPA-NPFF and QFW-NPSF did not allow the quantification of each peptide in the HPLC fractions. Considering the poor affinity of QFW-NPSF for the antibody used (Table 1), it seems likely that the immuno- reactivity detected in the peak 2 corresponds to SPA- NPFF. Discussion It is well documented that pro-neuropeptides are synthe- sized as inactive precursors that are processed during intracellular transport [37–40]. At the present time, the enzymatic pathway responsible for the conversion of NPFF precursors NPFF A and NPFF B [41] to smaller biologically active peptides is completely unknown. The first step in this knowledge is the description of the peptides actually generated in neurones before extracellular degradation processing by a great variety of peptidases. The key finding of the present study is that the pattern of NPFF-related peptides processed from the proNPFF A is similar in neuronal cell line and nervous tissue. The SH- SY5Y human neuroblastoma cell line, used in this study as an in vitro model for human neurones, expressed and processed the hproNPFF A to generate SQA-NPFF, NPFF and NPAF. These results showed for the first time that three different NPFF-related active peptides, NPFF, SQA-NPFF and NPAF, could be generated by intracellular processing of hproNPFF A in the human neuroblastoma. The presence of some of these peptides has not been described previously. Only NPFF was clearly detected in COS-7 cells while in the mouse spinal cord, SPA-NPFF was detected in addition to NPFF. These data were obtained by the combination of sensitive complementary methods, in particular mass spectrometry, which has permitted the precise identification of the different NPFF-related peptides. Nanospray ionization and MS/MS analyses allowed the identification of femto- moles quantities of deduced NPFF-related peptides enco- ded by mouse proNPFF A , in particular SPA-NPFF, which was not previously detected in spinal cord with MS analyses [31]. This study exemplified that on-line capillary HPLC/ nanospray ion trap tandem mass spectrometry was a powerful analytical technique, giving rise to the character- ization of minute amounts of endogenous neuropeptides [32]. Fig. 7. HPLC profile of NPFF-immunoreactivity in mouse cervical spinal cord extract. Acid extract prepared from three cervical spinal cord segments (106 mg) was purified on a Sep-Pack Cartridge, applied on a C8 column and separated on gradient 1 at a flow rate of 400 lLÆmin )1 . NPFF-IR was assessed by radioimmunoassay. Elution positions of synthetic NPFF-related peptides are indicated by arrows. Ó FEBS 2003 Processing of proNeuropeptide FF A (Eur. J. Biochem. 270) 4195 The physiological relevance of these observations is that the three peptides, NPFF, SQA-NPFF and hNPAF, could act as neurotransmitters in human as they exhibit a high affinity and a high activity towards NPFF 2 receptors [26,42]. We have compared previously the affinities and antiopioid activities of the different peptides putatively produced by the rat NPFF precursor and reveal that the undecapeptides are likely to be the physiologically active Fig. 8. Identification of SPA-NPFF, NPFF and QFW-NPSF in mouse spinal cord extract. HPLC fractions corresponding to the NPFF-IR peaks 1 and 2 on gradient 1 were separated on a C18 column, as previously described [31]. Collected fractions corresponding to the retention time of NPFF, SPA-NPFF and QFW-NPQF were analyzed by capillary HPLC/nanospray ion trap MS/MS. Reconstructed ion chromatograms of the fragment ion at m/z 546.3 generated during the MS/MS analysis of double-charged precursor ions at m/z 541.3 from 100 fmol of synthetic NPFF (A), at m/z 668.9 from 100 fmol of synthetic SPA-NPFF (D), at m/z 675.4 from 100 fmol of synthetic QFW-NPSF (G), at m/z 541.3 from HPLC fractions of NPFF-IR peak 1 (B), at m/z 668.9 (E) and at m/z 675.4 (H) from HPLC fractions of NPFF-IR peak 2 of tissue extracts. MS/MS spectra of the [M + 2H] 2+ precursor ions at m/z 541.3 (C), at m/z 668.9 (F) and at m/z 675.4 (I) of tissue extract. 4196 E. Bonnard et al. (Eur. J. Biochem. 270) Ó FEBS 2003 [...]... key for the processing of proNPFFA to the amidated NPFF-related peptides In this respect, two enzymatic pathways for the production of NPFF should be considered Clearly COS-7 and SH-SY5Y cells generate different proNPFFA fragments indicating that a neuronal cell line possess a capacity of processing not found in non -neuronal cells The undecapeptide generated in a neuronal cell line such as SH-SY5Y... necessity of an arginine residue at the P4 position and the fact that an alanine at P1 or phenylalanine at P¢1 position did not affect the cleavage further support this possibility [44] NPFF was also isolated from extracts of non -neuronal COS-7 cells indicating that processing of hproNPFFA at a specific nonbasic residue should be considered to explain the presence of NPFF in cell extracts In contrast,... explain the generation of NPFF: NPFF could be generated by processing of the hproNPFFA by enzymes recognizing a nonbasic motif This hypothesis is supported by the presence of a conserved motif -RXXAFL- that extend the N-terminal sequence of NPFF in mammals and recent reports demonstrating the importance of a protein convertase, the subtilisin/kexin isozyme SKI-1, in the processing of prohormones at specific... Chan, S.M.T., Moore, A.N.J., Ganellin, C.R & schwartz, J.C (1996) Characterization and inhibition of a cholecystokinin-inactivating serine peptidase Nature 380, 403–409 Eipper, B.A., Milgram, S.L., Husten, E.J., Yun, H.Y & Mains, R.E (1993) Peptidylglycine alpha-amidating monooxygenase: a multifunctional protein with catalytic, processing, and routing domains Protein Sci 2, 489–497 Biemann, K (1990)... concentration of SQA-NPFF observed in SH-SY5Y cells, relative to NPFF, further supports the possibility that the long peptides are synthesized before degradation into NPFF In contrast in non -neuronal cells, NPFF is present at a concentration similar to that of SQANPFF, suggesting that another processing pathway is responsible for NPFF synthesis in non -neuronal cells In mice, the long form SPA-NPFF contributes... decrease morphine tolerance and dependence in mice Eur J Pharmacol 358, 203–206 8 Kavaliers, M & Yang, H.Y (1989) IgG from antiserum against endogenous mammalian FMRF-NH2-related peptides augments morphine- and stress-induced analgesia in mice Peptides 10, 741–745 9 Gouarderes, C., Sutak, M., Zajac, J.M & Jhamandas, K (1993) Antinociceptive effects of intrathecally administered F8Famide and FMRFamide in the... FEBS 2003 Processing of proNeuropeptide FFA (Eur J Biochem 270) 4197 neurotransmitters in brain as they exhibit high affinity for NPFF2 receptor, functional activity and they could be generated from proNPFFA precursor in neuronal cells NPFF and NPA-NPFF exhibited a very high affinity for NPFF2 receptor of the rat spinal cord in contrast to shorter peptides such as NPSF [42] Similarly NPFF and NPANPFF maximally... activity of known maturation enzymes As these undecapeptides are active peptides, this processing pathway is likely to be similar to the one present in neurones in vivo In contrast, NPFF, which is observed in mouse and rat tissues as well as in both cell lines tested, could correspond to a metabolite of the undecapeptide or to a non -neuronal production by tripeptidylpeptidase The high concentration of SQA-NPFF... (1999) Proprotein and prohormone convertases: a family of subtilases generating diverse bioactive polypeptides Brain Res 848, 45–62 Yang, H.-Y & Martin, B (1995) Isolation and characterization of a neuropeptide from brain and spinal cord of rat Soc Neurosci Abstr 21, 760 Bonnard, E., Burlet-Schiltz, O., Frances, B., Mazarguil, H., Monsarrat, B., Zajac, J.M & Roussin, A (2001) Identification of neuropeptide... of mature NPFF-related peptides [47] The processing of proneuropeptide and prohormones occurs in most cases in subcellular compartments within the intracellular secretory pathways [39,40] It seems likely that NPFF, isolated from SH-SY5Y and COS-7 cells, was produced in specific compartments, where are localized peptidylglycine-a amidating monooxygenase, carboxypeptidases and proteins convertases [29,48] . Identification of proNeuropeptide FF A peptides processed in neuronal and non -neuronal cells and in nervous tissue Elisabeth Bonnard, Odile Burlet-Schiltz,. from extracts of non -neuronal COS-7 cells indicating that processing of hproNPFF A at a specific nonbasic residue should be considered to explain the presence of NPFF in cell extracts. In contrast,. of NPFF should be consid- ered. Clearly COS-7 and SH-SY5Y cells generate different proNPFF A fragments indicating that a neuronal cell line possess a capacity of processing not found in non -neuronal cells.