Discovery of Neuropeptides in the Nematode Ascaris suum by Database Mining and Tandem Mass Spectrometry

Abstract

Liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) was used to discover peptides in extracts of the large parasitic nematode Ascaris suum. This required the assembly of a new database of known and predicted peptides. In addition to those already sequenced, peptides were either previously predicted to be processed from precursor proteins identified in an A. suum library of expressed sequence tags (ESTs), or newly predicted from a library of A. suum genome survey sequences (GSSs). The predicted MS/MS fragmentation patterns of this collection of real and putative peptides were compared with the actual fragmentation patterns found in the MS/MS spectra of peptides fractionated by MS; this enabled individual peptides to be sequenced. Many previously identified peptides were found, and 21 novel peptides were discovered. Thus, this approach is very useful, despite the fact that the available GSS database is still preliminary, having only 1X coverage.

Introduction

Across animal phyla, neuropeptides have been shown to have strong effects on a wide variety of behaviors ¹. In order to understand the cellular mechanisms of such behaviors, a detailed knowledge of the circuitry is required. This includes a description of the neuromodulators present in the system, and of their effects on the neurons and synapses in each circuit. The parasitic nematode Ascaris suum, with only 298 neurons, is a simple system that allows for such detailed analysis. Extensive physiological characterizations of motorneurons and muscles have been performed, as have descriptions of classical neurotransmitters such as acetylcholine, gamma-amino butyric acid (GABA), dopamine, and serotonin ^2-8. It is also clear that there is a large array of neuropeptides present. Already, forty members of the Ascaris FMRFamide-related family of peptides (AF peptides, typically containing an –RFamide C-terminus) have been isolated, and have been shown to have potent effects on body posture, muscle tone, neuronal physiology, or second messenger systems ^9-20. However, it is clear from mass spectrometry (MS) ^21-23, and from immunochemical studies with anti-peptide antibodies ²⁴, that there are still many peptides to be discovered – we estimate that there are at least 250 neuropeptides in A. suum ²³, and a comparable number has been predicted in the small free-living nematode Caenorhabditis elegans ²⁵.

There are several methods available for neuropeptide discovery. Some are direct, such as the Edman degradation of pure peptides from chemical fractionation of low molecular weight extracts of tissues, or the mass spectrometric analysis of nervous tissue or single neurons, both of which have yielded the sequences of peptides that are expressed in vivo ^{12, 13, 20-22, 26}. Indirect methods involve prediction of peptides that are cleaved from precursor protein sequences deduced from cDNA (including expressed sequence tags – ESTs ^{27, 28}) or from genomic sequences ^{29, 30}. The validity of these predictions depends on several factors. The first is the ability to recognize neuropeptide precursor genes. Typically these genes encode proteins with an N-terminal signal peptide sequence that delivers them to the secretory pathway, and with predictable cleavage sites for proteolytic processing into shorter peptides (usually basic amino acids, in pairs or singly, flanking each processed peptide) ²⁵. The second factor is the validity of the predictions about sites of cleavage, which depends on assumptions about the nature and cellular expression patterns of the cleavage enzymes. Unfortunately, these enzymes exist in multiple forms with differing specificity, and their cellular expression patterns are usually unknown. This problem is compounded by any covalent post-translational modifications of the original sequences, and by the existence of unusual cleavages that are exceptions to the basic amino acid rule ³¹. Consequently, the indirect prediction of peptide sequences must remain tentative until proven by one or more of the direct methods.

In principle, recent advances in liquid chromatography-tandem mass spectrometry (LC-MS/MS) should make the description of the peptidome of any organism relatively easy. This approach has been applied very successfully for protein identification in the field of proteomics by enzymatically digesting the complex mixtures of proteins into peptides prior to LC-MS/MS analysis ^32-34. The pattern of peptide fragment peaks in the MS/MS spectra are matched to theoretical peptide spectra, which have been predicted from a protein database derived from a sequenced genome, annotated so that the possible protein products have been identified. This process of database searching relies heavily upon knowledge of the cleavage characteristics of the enzyme used to digest the proteins into peptides - trypsin is the most common, but other enzymes are suitable as well. Assigning a peptide sequence to an experimental spectrum by using this database searching algorithm is much easier than interpreting the MS/MS spectra with no previous sequence information, a process termed de novo sequencing.

Applying this technique of database searching to the discovery of neuropeptides in A. suum is not straightforward, however. Genomic sequences have only recently become available for A. suum, and they consist of genome survey sequences (GSSs), with only 1 X coverage (M. Mitreva, personal communication). The sequences are not assembled, and sequencing errors have not been corrected. Nevertheless, GSSs have been extremely useful in aiding de novo sequencing, and we have used them in conjunction with MS/MS data to sequence many novel peptides from single neurons ^{21, 35}. In addition, neuropeptides are not products of tryptic cleavage, so the algorithms typically used to simplify large database searches cannot be used. Instead, we assembled a smaller database to improve peptide identification. This new database consisted of translated GSS or cDNA sequences that encoded predicted or known peptides. These sequences were either known transcripts from A. suum ^{20, 21, 23, 28, 35-37, 44}, were identified as sequelogs of known or predicted C. elegans peptide-encoding ESTs ^{27, 28}, or were discovered from BLAST searches of A. suum GSSs. The goal of the current study was to explore whether the predicted peptides are expressed at detectable levels in A. suum.

This general approach has already been used very successfully in exploring the neuropeptidomes of Drosophila melanogaster ³⁸ and C. elegans ³⁹, taking advantage of completely sequenced genomes, and extensive predictions of putative peptides ^{29, 30}; in C. elegans the number of sequenced peptides was increased from 10 to 64 ³⁹. Relatively, we are at a disadvantage in A. suum, because of the restricted nature of the available nucleic acid sequence data, and we wished to determine whether these data were sufficient to give useful information (see also ref. 40). The initial use of the new database with LC-MS/MS data, reported here, has discovered 21 novel peptides in A. suum, a substantial addition to the previous total of 50 peptides sequenced in the last two decades ^{12, 13, 20-23, 26, 35}. In addition to the new peptides, many previously identified peptides were also found. Because it is relatively rapid, this method is indeed a useful complement to other methods of peptide discovery. Among the novel sequenced peptides are 15 peptides that are not members of the AF peptide family, since they lack the common C-terminal sequence (-RFamide) that defines the AF peptides. This makes the further exploration of the nature of this class of non-AF peptides in A. suum, and an analysis of their functional roles, an important priority.

Materials and Methods

Animals

A. suum were harvested from the small intestines of pigs at a regional slaughterhouse. The worms were maintained for up to 3 days at 37 °C in phosphate-buffered saline (PBS; 8.5 mM sodium phosphate, 150 mM sodium chloride, pH 7.4). The head region (anterior 1-2 cm) was removed and immediately frozen on dry ice. The heads were stored at -80 °C until use.

Peptide Extraction and Sample Preparation

Heads (N = 2,500) were freeze-powdered, and then peptides extracted with 500 mL ice-cold methanol: water: acetic acid, 90:9:1. The extract was centrifuged to remove solids, and the supernatant was then concentrated by rotary evaporation. The concentrated extract (4 mL) was diluted with 500 mL 0.1% aqueous trifluoroacetic acid (TFA). Lipids were removed by extraction with 4 × 100 mL 15:85 ethyl acetate:hexanes. Aliquots (60 mL) of the remaining aqueous solution were applied to 8 activated C18 cartridges (360 mg, SepPak, Waters Associates, Inc., Milford, MA). After washing each cartridge with 0.1% aqueous TFA, peptides were eluted with 5 mL 1:1 acetonitrile (ACN): water containing 0.1% TFA. The fractions were concentrated to ca. 500 μL in a vacuum centrifuge,. and a Millipore YM3 Size Exclusion column was used to remove proteins. The resultant “low molecular weight fraction” was dried almost to completion and then diluted in 50 μL 0.1% formic acid in 1% aqueous ACN. However, proteins were still prevalent so additional size-exclusion chromatography was performed. A 25 μL aliquot was separated on a 200 × 4.6 mm PolyHYDROXYETHYL A column (PolyLC, Columbia, MD) using a high performance liquid chromatograph (HPLC; Beckman, Fullerton, CA) controlled by System Gold software. The mobile phase was 50 mM aqueous ammonium acetate, with a flow rate of 0.1 mL/min. All fractions eluting prior to 25 min. were collected by hand, combined, dried almost to completeness, and resuspended in 20 μL 0.1% formic acid in 1% aqueous ACN.

LC- MS/MS Analysis

The neuropeptide sample was diluted 3-fold with 0.1% formic acid:2% aqueous ACN, and 3 μL injected into the LC-MS/MS capillary system consisting of an HPLC (Waters nanoAcquity, Milford, MA) connected to an electrospray ionization FT/ion-trap mass spectrometer (LTQ orbitrap Velos, Thermo Fisher Scientific, San Jose, CA). The sample loading occurred for 20 min at a flow-rate of 1 μL/min onto a trapping column prepared by packing 5 cm of 5 μm-diameter 300 Å-pore C18 beads (Western Analytical Products, Inc. Murrieta, CA) into a 75 × 365 μm fused silica capillary having a frit at one end. The analytical column (50 × 365 μm) was packed with 15 cm of the same C18 beads, but instead of packing against a frit, the capillary tip was pulled to ~1 μm with a P-2000 pipette puller (Sutter Instruments Co.). The peptides were eluted at a flow rate of 0.2 μL/min in an aqueous mobile phase containing 0.1% formic acid with a gradient of increasing ACN: 2% to 15% in the first 15 min. and then up to 45% in the next 75 min (Run 1), 90 min (Run 2), or 105 min (Runs 3 and 4).

A survey mass scan (150–2000 m/z – mass to charge ratio) was performed in the FT orbitrap at a resolution of 60,000. Various types of MS/MS scans were performed: collision induced dissociation (CID) at 30% relative collision energy, high energy collision dissociation (HCD) at 35 or 45% relative collision energy, and electron transfer dissociation (ETD) for 70 ms. These dissociation methods are somewhat complementary in that different methods lead to optimal fragmentation for various types of peptides depending upon their mass, ion charge, and amino acid sequence. The isolation width for the parent ions was 3.0 m/z, and the product ions from CID, HCD, or ETD fragmentation were analyzed in the FT orbitrap detector at a resolution of 7500.

The MS/MS conditions for the four runs were as follows (percentages refer to relative collision energy):

• Run 1: CID at 30% and HCD at 45% on all +2 and higher charge ions. The 5 most intense peaks from the MS survey scan were chosen for fragmentation by both CID and HCD.

• Run 2: CID at 30%, HCD at 35%, HCD at 45%, and ETD 70 ms on all assigned charge state ions +1 and higher. The 4 most intense peaks were chosen for fragmentation by all 4 of these MS/MS types.

• Run3: CID at 30% and HCD at 35% on +1 and +2 ions. The 6 most intense peaks were chosen for both MS/MS types.

• Run 4: HCD at 45% and ETD 70 ms on +3 and higher charge ions. The 6 most intense peaks were chosen for both MS/MS types.

Database creation

Translated Basic Local Alignment Search Tool (tBLASTn) searches of A. suum GSSs were performed in September 2010 using the National Center for Biotechnology Information (NCBI) BLAST server (http://www.ncbi.nlm.nih.gov/BLAST) as described previously by McVeigh et al., (2008) ²⁷. Lists of C. elegans nlps ^{25, 27, 30} were used to develop search strings. Genes encoding multiple peptides were concatenated into a single string. The Expect value was set to 30,000, with a word size of 2, no compositional adjustments, and complexity filter off. Results were translated (http://www.expasy.ch/tools/dna.html) and examined by eye for neuropeptide motifs and cleavage sites. Previously cloned and/or predicted afps (Ascaris FMRFamide-like precursor proteins) and As-nlps (Ascaris neuropeptide-like precursor proteins) were also added to the database ^{20, 21, 23, 27, 28, 35-37, 44}.

Peptide identification

The MS/MS data of all 4 runs were searched against the A. suum peptide database using SEQUEST or Mascot on in-house servers. These bioinformatics programs match the fragmentation data against any FASTA format protein database ^{40, 41}. To use Mascot, data files were transformed to mzXML using ReAdW and then to mgf using TPP (http://tools.proteomecenter.org/software.php). Multiple MS/MS ion searches were performed using no C-terminal modifications or a fixed C-terminal amidation in combination with variable modifications of pyroglutamic acid (Glu), acetylation (Lys or N-terminus), and oxidation (Met), which are all common post-translational modifications of neuropeptides. Tentative peptide identifications, which had Expect values <0.5 and a significance threshold of p<0.05 (Mascot) or XCorr values >2 (SEQUEST) and a parent mass within 5 ppm of its expected value, were further explored manually. To confirm these identifications as naturally occurring peptides, stringent requirements had to be met: 1) the peptide is flanked in the precursor by basic cleavage sites or N-terminally flanked by the signal peptide, and 2) if the peptide sequence contains a C-terminal amidation, the precursor peptide sequence includes a C-terminal glycine.

RESULTS AND DISCUSSION

This study attempts to identify novel neuropeptides in A. suum through BLAST (Basic Local Alignment Search Tool) searches of genomic survey sequences (GSSs) coupled with MS/MS analysis of a peptide extract.

GSS Database search

Previously, mining of the nematode expressed sequence tag (EST) databases has allowed the cross-species identification of a large number of predicted -RFamide peptides, as well as members of the neuropeptide-like protein (NLP) family (peptides that do not contain a C-terminal –RFamide) including several in A. suum ²⁷. In this study, the A. suum GSS database was mined using C. elegans peptide-encoding transcripts to discover novel A. suum putative peptides. This was done both to identify possible novel peptides and to generate a reference database for MS analysis. Known translated A. suum peptide-encoding transcripts and ESTs were also added to this database. When analyzing GSSs as compared to ESTs, there is an added complication since ESTs are derived from the fully processed mRNA and GSSs are derived from genomic DNA. The GSSs contain introns and are not assembled, unlike completely sequenced genomes, making prediction of the final peptide far more difficult. With this in mind, more BLAST search results were included in the database used for MS analysis than those that included predicted peptides. Despite these complications, we predicted 12 novel putative - RFamide peptides from 9 GSSs (Table 1). Seven of these GSSs are sequelogs of known C. elegans flps (FMRFamide-like peptide transcripts). Two peptides, predicted from a search using Ce-FLP-10 as the query, have no obvious sequelog in C. elegans. We also found 44 nlps predicted from 21 GSSs (Tables 2 and 3). This number does not include sequelogs of the anti-microbial nlps, Ce-nlp-24 - Ce-nlp-33, because it is difficult to predict peptide cleavage in these sequences. However, the 32 related GSSs that encode these antimicrobial nlps were included in the database used for MS analysis. Several searches returned more than one GSS encoding a peptide related to the peptide query. For example, the search using Ce-NLP-15 as the query returned two GSSs, As-nlp-15a and As-nlp-15b, each encoding the peptide AFDSLTGAGFTGFD and two different additional sequence-related peptides. It is unclear if the transcripts from these two GSSs are spliced together to form one transcript or if this is a gene duplication. Similarly, searches with Ce-NLP-18 and Ce-NLP-20 each returned two distinct peptide-encoding GSSs, but the high degree of similarity between first the C. elegans peptides and then the newly predicted A. suum peptides can cause difficulty in naming these peptides and peptide-encoding sequences by similarity, as has become the convention. Once the transcripts are cloned, more permanent names for transcripts and peptides will be assigned.

Table 1.

Putative A. suum flp peptides predicted from tBLASTn searches of the Genbank GSS database. The numbers below each gene are accession numbers.

gene	predicted sequence		cleavage site
As–flp–3	KR	GGPLGTMRFG	KR
ED071134
As–flp–4	RK	ASFIRFG	KR
ED359465	KR	KASFIRLG	–
As–flp–10a	RR	LPHMEKISGFIRFG	KR
ED453735
As–flp–10b	R	SCYMIRFG	KR
ED395851
As–flp–10c	KR	TVYAVIRFG	K
ED291226
As–flp–17	KR	KSAFVRFG	R
ED065987	KR	KSSYIRFG	–
As–flp–19	KK	WASQLRLG	KR
ED386389	KR	ASWASKVRFG	–
As–flp–25	K	TSYQYFRFG	R
ED280015
As–flp–31	KK	GRPRGPLRFG	KR
ED243419

Table 2.

Putative A. suum nlp-1 -- nlp-17 peptides predicted from tBLASTn searches of the Genbank GSS database.

As–nlp–1	KR	MDASRFYLGFG	KR
ED075198
As–nlp–5	A	TSISRSKDAKKPSLDILEGAGFSPL	KK
ED209100	KK	PSLDILEGAGFSPL	KK
	KR	ALDTLEGSGFGFS	KR
	KR	ALDDLEGVGFGGML	RK
	KR	ALDSLEGTDFGL	KK
	KR	ALDYLEGSGFGLM	KR
	KR	AFDAIEGADIGFH	KR
	KR	ALDMLEGSGFGL	KK
	KR	ALDSLEGTGFGL	KK
As–nlp–6	KR	YASRGGFGQ	KR
ED364076	KR	ASMRFSDRVAAQ	K
As–nlp–9	KR	AGARSFGPRIYDVESYLYPTN	KR
ED378410	KR	FGSFSPYYFYQQPK	R
	KR	GGGRTFASYWIPPSGNG	R
As–nlp–10	KR	SSIPFHGGIYG	KR
ED196733	KR	TAALPFSGGLYG	KR
	KR	AGWIPFSGGLYG	KR
As–nlp–15a	KR	AFDSLTGAGFTGFD	KR
ED281083	KR	AFDSLTGSGFTGFD	KR
	KR	SLDSPANRGFTHFD	KR
As–nlp–15b	R	AFDSLTGAGFTGFD	KR
ED054924	KR	SFDSLNSRGFTGFD	KR
	KR	AFDSLVGHGFTGFD	–
As–nlp–16	R	NNDISSSGEHEHQ	KR
ED433984
As–nlp–17	RK	NILSNMMRIG	RR
ED472655

Table 3.

Putative A. suum nlp-18 - nlp-23 and nlp-34 - nlp-46 peptides predicted from tBLASTn searches of the Genbank GSS database.

As–nlp–18a	RR	FAYSFAFA	KR
ED276235
As–nlp–18b	KK	AGRFARHASGQQNELSFAFA	KR
ED229080	KR	APFA	KR
As–nlp–20a	KR	NKDANSVYAFSFA	KR
ED418549
As–nlp–20b	K	AFANFAFA	KR
ED054056
As–nlp–21	KR	KHYLQTDE	KR
ED417768	KR	GGGRSFRLSSLGE	KR
	KR	AGARSFPGAGGLT	R
As–nlp–22	KR	SIALGRFSLRPG	KR
ED456305
As–nlp–23a	KR	SMALGRLAFRPG	KR
ED282494	KR	SLALGRVDFRPG	KR
	KR	NVDIADKDIEYLMEDSMG	KR
	KR	SAAFGRFHFRPG	KR
	KR	SLALGRSGFRPG	KR
	KR	SLALGRVGFRPG	KR
As–nlp–23b	K	GIAPYAGMRPG	KR
ED431569
As–nlp–35	RR	DSKRAFLQQLMQRLKPRF	RR
ED247130
As–nlp–44	RR	NDRPQSSAAVLVPYPRVG	KR
ED420546
As–nlp–46	KR	NIAIGRGDGFRPG	K
ED469442

Mass spectrometric analysis of peptide extracts

Peptides extracted from freeze-powdered A. suum heads were analyzed in four runs by LC-MS/MS (liquid chromatography-tandem mass spectrometry). Each run produced a complex base-peak chromatogram containing numerous overlapping peaks. Survey mass spectra were automatically acquired to find suitable targets for MS/MS, and the appropriate m/z ions were selected for fragmentation (Figure 1). The fragmentation patterns in the MS/MS spectra were then compared, by Mascot or SEQUEST search engines, to predicted fragmentation patterns from the database described above. The selection and fragmentation conditions were varied in the four runs in an attempt to maximize the number of peptide identifications (See Methods for details).

Figure 1.

Schematic overview of MS data acquisition. (a) Base peak chromatogram of Run 1. The red line indicates the time of the MS scan in b. (b) Example of MS survey scan in which the doubly charged ion 594.84 (AF19) was selected for fragmentation. (c) MS/MS spectrum of AF19 with the observed b and y fragment ions labeled. These fragment ions were matched to those predicted from the AF19 sequence.

AF peptide identification

In this study, members of 6 AF peptide families were identified although we did not identify any of the newly predicted AF peptides (Table 4). In all but one of these families, more peptides are known to exist than were identified here; these peptides may have been lost during sample preparation or due to insufficient chromatographic separation. The sequences of several peptides were confirmed with our results, although they were previously predicted and analyzed by MS (not MS/MS) ²³. These include a peptide resulting from an incomplete cleavage of AF3, ENEKKAVPGVLRFa (AF3*), and several members of the afp-5 (As-nlp-13) family of peptides: SNAFDRNFmNFa, SDAFSRNFmNFa, and ESQFSRDFLNFa (The “m” refers to an oxidized methionine). Although these peptides are not actually RFamides, we have included them as AF peptides here because another peptide encoded by this transcript, AF17, which has an HFamide rather than an RFamide C-terminus, was originally identified by its RFamide immunoreactivity, and therefore named an AF peptide ⁴³. Peaks with m/z values corresponding to the predicted mass of all of these peptides in an unoxidized form have been seen in MALDI-MS of whole A. suum ganglia and single neurons ^{21, 23}. We believe that the oxidation is an artifact of the isolation procedure, since no effort was made to keep the peptides reduced during the preparative steps prior to LC-MS/MS.

Table 4.

A. suum peptides identified by either Mascot (M) or Sequest (S) searches in 4 different MS runs.

			Run1		Run2		Run3		Run4
Gene	Peptide	Sequence	M	S	M	S	M	S	M	S
afp–1	AF3*	ENEKKAVPGVLRFa	+	+	+		+	+
	AF4	GDVPGVLRFa	+	+			+	+
	AF13+O	SDmPGVLRFa	+		+	+	+	+
	AF10+O2	GFGDEmSmPGVLRFa			+		+	+
	AF10+O	GFGDEMSmPGVLRFa	+	+			+	+
	AF10–+O2	FGDEmSmPGVLRFa	+	+
	AF10–+O2	GDEmSmPGVLRFa	+			+	+	+
	AF10–+O2	DEmSmPGVLRFa					+	+
afp–5	As–NLP–13–2+O	SNAFDRNFmNFa	+	+	+	+		+
	As–NLP–13–6+O	SDAFSRNFmNFa	+	+	+	+
	As–NLP–13–3	ESQFSRDFLNFa	+	+	+	+
	AF17+O	DFDRDFmHFa	+		+
afp–6	AF22	NGAPQPFVRFa	+	+		+
	PepSE	SGRVDHIHDILSTLQRLQLANE				+				+
afp–11	AF29	NAEPNFLRFa	+	+	+	+	+	+
	AF11	SDIGISEPNFLRFa	+	+			+	+
	AF28	SAEPNFLRFa	+		+	+	+	+
	AF32	GSDPNFLRFa	+		+		+	+
	AF27	PADPNFLRFa					+	+
afp–12	AF36+O2	VPSAADmmIRFa	+	+	+	+	+	+
afp–13	AF19	AEGLSSPLIRFa	+	+	+	+	+	+
	AF34+O	DSKLmDPLIRFa	+	+	+	+
	AF34–+O	LmDPLIRFa					+	+
	PepTT	TPPEEDLLGRFT	+	+	+	+	+
	PeptTL+O	TNImGENRLNRNL							+
As–nlp–3	As–NLP–3–1	snfdgreaslprf	+	+		+				+
	As–NLP–3–2	AINPFTDSIa				+		+
	As–NLP–3–4*	NTGIAAERDLPKRYFDALAGQSLa		+		+
	As–NLP–3–4	YFDALAGQSLa					+
As–nlp–5	As–NLP–5–2*	TSISRSKDAKKPSLDILEGAGFSPL			+	+				+
	As–NLP–5–4+O	ALDDLEGVGFGGmL			+	+
As–nlp–7	As–NLP–7	NHLIGFDDPRLFSSSYa	+	+			+	+
As–nlp–8	As–NLP–8–2*	SFDRIDGSAFGPHRH	+	+					+	+
	As–NLP–8–5	AFDRIEGSGFGLD					+	+
	As–NLP–8–4	AFDRIEGAGFGLS					+	+
	As–NLP–8–4–	AFDRIEGAGFG					+	+
As–nlp–9	As–NLP–9–2	FGSFSPYYFYQQPK		+	+	+		+
As–nlp–15	As–NLP–15a–3	SLDSPANRGFTHFD	+	+	+	+		+
	As–NLP–15b–2–	SFDSLNSRGFTG					+	+
	As–NLP–15b–3–	AFDSLVGHGFTG					+
As–nlp–21	As–NLP–21–2	GGGRSFRLSSLGE				+		+	+	+
As–nlp–23	As–NLP–23–2	SLALGRVDFRPG	+	+

^*extended form of peptide; - truncated form of peptide; m oxidized methionine; a amidated terminus; novel peptides in bold.

A peptide with sequence LmDPLIRFa was identified, which is interesting because it is a fragment of peptide AF34 (DSKLMDPLIRFa) encoded by the afp-13 transcript. The AFP-13 precursor protein, which encodes several sequenced peptides, is known to have both single and double basic amino acid cleavage sites ²¹, so it is plausible that AF34 undergoes a second cleavage at the internal K. However, MS of the isolated ALA neuron, which contains AF34 as well as the other members of the afp-13 peptide family, did not reveal a peak corresponding to the shorter peptide ²¹. This could be a case of cell-specific cleavage of these peptides; alternatively, this peptide could arise from degradation during storage, extraction, or fractionation preceding MS analysis. We have had previous indications that proteolytic degradation may occur: AF10 (GFGDEMSMPGVLRFa) was originally discovered by Edman degradation of highly purified peptides ²⁶; in the same purification, a related peptide, named AF12 (FGDEMSMPGVLRFa), was also isolated. However, when the afp-1 transcript was subsequently cloned and sequenced, there was no separately encoded AF12 sequence - the AF12 sequence occurs only within AF10 ⁴⁴. Furthermore, AF12 was not detected in MS of whole ganglia ²³, so we believe that AF12 is a degradation product of AF10, probably due to the action of an aminopeptidase. Interestingly, in addition to AF12, we have now found two more truncated forms of AF10 (Table 4). In earlier fractionations of peptide extracts, truncated forms of other peptides derived from afp-1 were detected (with single amino N-terminal truncations of peptides AF3, AF13, AF14, and AF20 – Cowden and Stretton, unpublished). It is striking that, so far, N-terminal loss is particularly prevalent among the products of the afp-1 transcript; it is unclear whether these changes occur in the cells expressing these peptides, or whether they represent unusual susceptibility to the action of extracellular aminopeptidases, which have been shown to be present in A. suum ⁴⁵. One other case of N-terminal truncation has been observed, the loss of the N-terminal D from DFDRDFMHFa to generate AF17 (FDRDFMHFa). However, in this case there is strong evidence that this is due to an unusual susceptibility of the N-terminal D to loss under the acidic conditions of the extraction ³⁶.

As-NLP peptide identification

A total of five As-NLP peptides (Ascaris neuropeptide-like peptides), encoded by sequences found in the EST library, were identified (Table 4). One of these arises from As-nlp-7 (NHLIGFDDPRLFSSSYa, Figure 2a) and four are encoded by As-nlp-3, two of which were predicted by McVeigh et al., (2008) ²⁷. Of the two not predicted, As-NLP-3-4* is the result of an incomplete peptide cleavage. The other peptide appears to be the result of an atypical cleavage after an alanine, which we believe might be due to cleavage of the signal peptide. The flaw in this hypothesis is that the EST does not include an initiating methionine, but if one were added at the beginning of the translated gene and analyzed by SignalP ⁴⁶, the signal peptide is predicted to be cleaved to form the N-terminus of As-NLP-3-1. We are currently attempting to clone the full-length transcript to resolve this issue.

Figure 2.

Four novel peptide MS/MS spectra with b and y fragment ions denoted. (a) CID spectrum of NHLIGFDDPRLFSSSYa, m/z = 933.962, z = +2. (b) CID spectrum of SLDSPANRGFTHFD, m/z = 782.363, z=+2. (c) CID spectrum of FGSFSPYYFYQQPK, m/z = 879.912, z = +2. (d) HCD spectrum of SLALGRVDFRPG, m/z = 644.363, z = +2.

Peptides encoded by 6 distinct GSSs were also identified (Table 4). Nine peptides were previously described from the As-nlp-8/anp-1 GSS ³⁵, and three of those were found in this study along with an additional peptide fragment. This fragment is likely the result of a procedural artifact, as was seen with several AF peptides. Similar results were observed for the highly related peptides from As-nlp-15: one full length peptide (SLDSPANRGFTHFD, Figure 2b), and two truncated peptides missing two C-terminal amino acids in this case. From As-nlp-21, one of three predicted peptides was identified. From the GSS As-nlp-5, two peptides were detected (TSISRSKDAKKPSLDILEGAGFSPL and ALDDLEGVGFGGmL), although more were predicted. The former is possibly the result of a signal peptide cleavage and a missed internal cleavage, and the latter is a more conventional peptide. Two peptides were also identified from As-nlp-9 and As-nlp-23, which were not the products that would have originally been predicted, although they could be products of conventional cleavages. One of these (FGSFSPYYFYQQPK, Figure 2c) would more likely have been processed to remove the C-terminal lysine, according to conventional wisdom ²⁵. However, it is possible that the proline before the lysine interfered with this “conventional” processing by carboxypeptidase. For the other peptide, SLALGRVDFRPG (Figure 2d), the typical prediction of the final peptide product would be SLALGRVDFRPa. This may result from proline interference with the amidating enzymes; alternatively, it is possible that both forms exist but we did not detect the amidated form. In our recent single cell MALDI-MS experiments, the non-amidated C-terminal glycine form of several peptides have been detected, and they are presumed to be intermediates in the processing cascade ²¹. Interestingly, in experiments with C. elegans peptide extracts, no Pamide peptides have been detected, although many are predicted ²⁵.

Prediction versus chemical sequencing

As has been shown with several of the peptides identified in this study, peptide prediction can be helpful but also occasionally inaccurate. We found peptide prediction essential for the type of MS analysis performed here because the very large number of spectra generated made manual interpretation infeasible—Mascot and/or SEQUEST searches were required. However, the extensive size of a total GSS or EST database, when used in these searches, either caused the program to crash or returned very few identifications. Thus, having removed non-coding genomic sequences and sequences not identified as putative peptides, the smaller database containing only predicted peptides of interest allowed a more practical and confident route to the identification of peptides. An analogy may be to consider the greater ease of searching for peptide “needles” within a sequence “haystack” when much of the non-interesting “hay” has been removed.

Are there more peptides?

Although many novel peptides were discovered in this study, it is clear that there are many more peptides that remain to be sequenced. The mass spectra in these experiments show complex mixtures of ions entering the instrument at one time; thus, better chromatographic separation prior to MS analysis would help identify a larger number of peptides. It is striking that AF2 was not detected in this study, even though it is believed to be the most abundant peptide in A. suum and it has been detected repeatedly in whole ganglia and single cell MALDI-MS ^{21, 23, 35}. Interestingly, no peptides without basic amino acids have been detected by MALDI-MS, but three were detected in this study.

CONCLUSIONS

The use of GSS database mining in combination with MS yielded a dramatic increase in the number of predicted and identified peptides in A. suum. This was especially true of the As-nlps, of which only the As-nlp-8 (anp-1) family of peptides had been sequenced previously ³⁵. The prediction and identification of all of these peptides is an important first step in understanding the vital role neuropeptides play in the nervous system of A. suum.

ABBREVIATIONS

ACN

acetonitrile

AF

Ascaris FMRFamide-related

afp

Ascaris FMRFamide-like precursor protein

BLAST

Basic Local Alignment Search Tool

CID

collision induced dissociation

EST

expressed sequence tag

ETD

electron transfer dissociation

flp

FMRFamide-like peptide

FT

Fourier transform

GABA

gamma-amino butyric acid

GSS

genome survey sequence

HCD

high energy collision dissociation

HPLC

high performance liquid chromatography

LC-MS/MS

liquid chromatography coupled with tandem mass spectrometry

m/z

mass to charge ratio

NLP

neuropeptide-like protein

TFA

trifluoroacetic acid