Four day inhibition of prolyl oligopeptidase causes significant changes in the peptidome of rat brain, liver and kidney

Biochimie 94 (2012) 1849e1859

Contents lists available at SciVerse ScienceDirect

Biochimie journal homepage: www.elsevier.com/locate/biochi

Research paper

Four day inhibition of prolyl oligopeptidase causes significant changes in the peptidome of rat brain, liver and kidney Jofre Tenorio-Laranga a, Pekka T. Männistö a, Markus Storvik b, Pieter Van der Veken c, J. Arturo García-Horsman a, * a b c

Division of Pharmacology and Toxicology, University of Helsinki, P.O. Box 56 (Viikinkaari 5E), 00014 Helsinki, Finland Department of Pharmacology and Toxicology, University of Eastern Finland, Kuopio, Finland Laboratory of Medicinal Chemistry, University of Antwerp, Wilrijk (Antwerp), Belgium

a r t i c l e i n f o

a b s t r a c t

Article history: Received 5 January 2012 Accepted 3 April 2012 Available online 21 April 2012

Prolyl oligopeptidase (PREP) cleaves short peptides at the C-side of proline. Although several proline containing neuropeptides have been shown to be efficiently cleaved by PREP in vitro, the actual physiological substrates of this peptidase are still a matter of controversy. The aim of this study was to evaluate the changes in the peptidome of rat tissues caused by a repeated 4-day administration of the potent and specific PREP inhibitor KYP-2047, using our recently developed iTRAQ-based technique. We found tissuedependent changes in the levels of specific subsets of peptides mainly derived from cytosolic proteins. Particularly in the kidney, where the levels of cytochrome c oxidase were found decreased, many of the altered peptides originated from mitochondrial proteins being involved in energy metabolism. However, in the hypothalamus, we found significant changes in peptides derived from hormone precursors. We could not confirm a role of PREP as the metabolising enzyme for b-endorphin, galanin, octadecaneuropeptide, neuropeptideeglutamic acideisoleucine, substance P, somatostatin, enkephalin and neuropeptide Y. Furthermore, changes in the degradation patterns of some of these neuropeptides, and also most of those derived from other larger proteins, did not follow specificity to proline. After a 4-day treatment, we found a significant amount of peptides, all derived from secreted pro-proteins, being cleaved with pair of basic residue specificity. In vitro experiments indicated that PREP modifies the endogenous dibasic residue specific proteolysis, in a KYP-2047 sensitive way. These findings suggest that PREP may act indirectly within the routes leading to the specific peptide changes that we observed. The data reported here suggest a wider tissue specific physiological role of PREP rather than the mere metabolism of proline containing active peptides and hormones. Ó 2012 Elsevier Masson SAS. All rights reserved.

Keywords: Peptidomics Peptide metabolism Protein convertases Prolyl oligopeptidase

1. Introduction Prolyl oligopeptidase (PREP) is a serine protease which cleaves peptides shorter than 30 amino acids at the carboxyl side of an internal proline. Although PREP is found in all tissues, it is localised only in specific cell types particularly in the brain, kidney and liver [1,2].

Abbreviations: ACBP, Acyl-CoA-binding protein; NEI, Neuropeptideeglutamic acideisoleucine; PC, Pro-protein convertase; PREP, Prolyl oligopeptidase; VEGF, Vascular endothelial growth factor. * Corresponding author. Tel.: þ358 9 191 59459; fax: þ358 9 191 59471. E-mail address: Arturo.Garcia@helsinki.fi (J.A. García-Horsman). URL: http://www.helsinki.fi/farmasia/farmakologia/english/mainpage.htm 0300-9084/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.biochi.2012.04.005

PREP has become an interesting drug target since administration of specific inhibitors has shown improvements in cognitive functions in different animal models [3]. Thus, PREP is probably involved in the cleavage of peptides involved in memory and learning like melanocyte-stimulating hormone, vasopressin, oxytocin, angiotensin II, or substance P [4]. Indeed, these peptides, and many other proline containing neuropeptides [4], are digested by PREP in vitro. However, with the current information, it has been difficult to relate the effects of PREP inhibitors in vivo with the proposed in vitro functions of the peptidase [5]. From previous peptidomic analysis of PREP inhibition [6e8], and circumstantial evidence on the effect of PREP inhibitors on peptide function [9], it can be suggested that PREP may have a role in the degradation of intracellular active peptides and proteins [10].

1850

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859

New efforts to identify novel peptides have relied on the development and application of peptidomic approaches based on mass spectrometry (MS) [37,38]. These methods are able to identify changes in the peptidome followed by experimental manipulations, for example after modulation of the activity of a particular enzyme [11e13]. MS-based techniques have the capacity to sensitively and rapidly detect multiple peptides without a priori knowledge of their identities, thus unbiased. This makes it a more suitable tool for the identification of unknown enzyme regulated peptides [14,15]. Previous peptidomic studies have demonstrated the value of this approach on PREP research, and have suggested the involvement of PREP in the degradation of hormones and mitochondrial protein precursors [6e8]. Pioneering studies analysed peptides generated from the in vitro PREP digestion of fractionated peptides form porcine brain homogenates [6]. Although, the treatment did not reflect physiological conditions, interesting peptides from intracellular proteins were identified, like those derived from myelin basic protein, a- and b-synucleins, haemoglobin, and from proteins connected with the inositol phospholipid signalling pathway and the Ca2þ-calmodulin system. Nolte et al. [7] analysed the peptidome of 1e4 h PREP inhibited mice and found many additional peptides, highlighting proline rich peptides, confirming PREP preference for shorter peptides and suggesting a previously unknown cleavage specificity. We have validated a peptidomic approach which considerably minimises postmortem and tissue processing associated proteolysis, and increases peptide resolution by using a 2-dimensional RP-HPLC separation technique of the iTRAQ labelled peptides [8]. In that study, we reported changes due to an acute PREP inhibition, and found alterations in the levels of peptides from secreted proteins but also from other large proteins which included histones, proteasome components and respiratory complexes. All previous studies have been limited to the analysis of an acute PREP inhibition in the brain, and have drawn conclusions on few peptides found. These studies have not provided a comprehensive functional analysis of the data in search for clues on the biological or metabolic processes where PREP might be involved. In an attempt to make the differences between control and PREP inhibited animals more obvious, and to take advantage of the bioinformatics tools, the present study reports the analysis of the peptidome of several brain areas and peripheral tissues from 4-day PREP inhibited rats. Accordingly, we have applied our iTRAQ labelling peptidomic approach [8] to analyse changes on peptide levels, using the most potent PREP inhibitor available, KYP-2047 [18]. We also performed a functional analysis of the data, and additional biochemical experiments to challenge some of the suggestions derived from this analysis. Finally we present a discussion in order to conclude on the physiological role of PREP. 2. Materials and methods

a good brain penetration and activity [16,18]. The measured compound levels in several brain areas are enough to inhibit PREP nearly 100% for several hours [19e21]. Effects of the 4-day twicea-day treatment with KYP-2047 in general behaviour, food or water intake, relative to vehicle administrated animals, were not observed. Altogether 15 animals were used, 9 of them were treated with KYP-2047 and 6 with vehicle. Four hours after the last injection, the animals were sacrificed and quickly dissected by three experienced surgeons taking several samples of the liver, kidney, hypothalamus, striatum, frontal cortex, cerebellum and the rest of the brain. Tissues were immediately flash frozen (liquid nitrogen). Samples were kept at 80  C until processed for peptide extraction, or to assay peptidase activity or western immunoblotting. 2.3. Peptide extraction Peptide extraction was performed as described before [8]. Briefly, each frozen rat tissue sample was powdered in a mortar at 80  C and transferred to a pre-cooled tube. A volume of 300 ml/25 mg of tissue of boiling water was added, and boiled for 10 min. After centrifugation (16,000 g) the supernatant was separated and one volume of 0.25% acetic acid was added to the pellet, sonicated and spun at 20,000 g during 20 min at 4  C. The resulting supernatant was decanted, combined with the boiling water fraction from above and filtered through a 10 kDa cut-off filter (Millipore, Billerica, MA, USA). After the extraction, the relative peptide concentration for each sample was measured at 280 nm (NanoDrop, Thermo Fisher, Waltham, MA, USA). 2.4. Sample pooling and iTRAQ labelling After the peptide extraction, the different samples obtained were pooled in 4 groups as defined in Fig. 1. All the control samples (from 6 different animals) were pooled in one single group. Since, it is considered that pooling of sample is an appropriate biological average in proteomics studies, specially recommended when there is an interest of general characteristics of a population and when the sample obtained from one individual is insufficient [22]. The samples from treated animals were pooled in 3 groups. The 4 groups contained the same relative amount of peptides before the iTRAQ labelling. We performed the iTRAQ labelling according to the protocol recommended by the manufacturer (Applied Biosystems, Framingham, MA, USA) as reported before [8]. Basically, vacuum dried samples (from 260 ml of a peptide solution of 1 UA) were reconstituted in 25 ml iTRAQ dissolution buffer. An aliquot (70 ml) of iTRAQ-reporter 114 was added to the control sample, and same aliquots of each iTRAQ-reporter, 115, 116 and 117, were added to the three different samples from treated animals. Mixtures were incubated at room temperature for 1 h and vacuum dried.

2.1. Chemicals 2.5. 2-D RP-HPLCeMS/MS analysis Chemicals were purchased from SigmaeAldrich (St. Louis, MO, USA) unless otherwise specified in the text. The PREP inhibitor KYP2047 (4-phenylbutanoyl-L-prolyl-2(S)-cyanopyrrolidine) was synthesised at the Laboratory of Medicinal Chemistry (University of Antwerp) as previously described [17]. 2.2. Animals Male Wistar rats (Harlan, CPB, Zeist, The Netherlands) weighting 180e280 g were administrated intraperitoneally twice-a-day with 20 mg/kg of KYP-2047 in vehicle (0.5% DMSO in saline) or vehicle (control animals). The inhibitor is highly specific to PREP and has

The dual-pH reversed phase liquid chromatography coupled to tandem mass spectrometry (2-D RP-HPLCeMS/MS) was applied as described before [8]. Briefly, the peptides were separated by differential pH 2-dimensional-reverse phase-HPLC (2-D RP-HPLC), where the first dimension was at pH 11, and the second at pH 2, and the elution was directly applied to a nanospray source of a QSTAR XL instrument (Applied Biosystems, Framingham, MA, USA). Information-dependent acquisition analysis was carried out with acquisition cycles in mass spectrometry (MS) and tandem mass spectrometry (MS/MS) modes along all the chromatogram. The QSTAR XL was operated in information-dependent acquisition

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859

1851

Fig. 1. Scheme of the experimental set-up to study the differential peptidome of the PREP inhibited rats, showing the distribution and pooling of the samples before the labelling of the iTRAQ 4-plex. C1eC6, peptide samples from control extracts; T1eT9, peptide samples from treated extracts. *Performed by duplicate.

mode, in which a 1-s TOF MS scan from 350 to 2000 m/z was performed, followed by 3-s product ion scans from 100 to 2000 m/z on the three most intense doubly or triply charged ions. 2.6. Data analysis

bioinformatic resources David’s annotation (http://david.abcc. ncifcrf.gov/) and Babelomics (http://babelomics.bioinfo.cipf.es/). The list of the proteins containing altered peptide levels after KYP-2047 treatment, were inspected for the enriched canonical pathways of the gene ontology (GO) terms (cellular compartment and molecular function) among the genes coding the proteins, using David Bioinformatic Resources [23,24] and FatiGo application from the Babelomics web portal [25,26]. In order to reduce noise, we used lists of gene products expressed in each tissue as the background list, when analysing whether the proteins containing altered peptides levels after KYP-2047 treatment are statistically distinct from the background. Our analysis identified GO terms that were most significant to the data set.

All the MS/MS spectra were analysed through ProteinPilot v 3.0 package (Applied Biosystems, Framingham, MA, USA) considering all possible posttranslational modifications included in this version of the programme. The sequences were searched against SwissProt protein database. ProteinPilot software reported relative quantification of each peptide based on the areas under the peaks at 114, 115, 116, and 117 Da, which were the masses of the corresponding iTRAQ reporters (the reporter 114 corresponded to the control and those 115, 116 and 117 to the three different pools from treated animals). Only the sequences identified with a minimum confidence of 95%, or with 90% in case of an early 95% identification of such sequence, were considered. All the areas corresponding to the iTRAQ-reporter peaks were corrected by the Bias value given by the software. This was based on the assumption that the sum of all the areas quantified for P115 P116 P each reporter should be equal ð 114 1.X A ¼ 1.X A ¼ 1.X A P117 ¼ 1.X AÞ, since the relative amount of peptides for each group was the same. The peptide levels measured on the pool of samples are equal to the average of those levels from each individual sample in the pool. This assumption has been considered appropriated for proteomic studies [22]. Furthermore, variations on the levels of peptides in controls were found not significant, as we reported earlier [8]. Therefore, each group of treated samples was referred to the pool of controls as a relative treated/control ratio. After the calculation of all the individual ratios of each sequence and its associated measure error, the final ratio was the weighted average of all individual common sequences and the p-value, for ratios different from 1, was calculated using the same methodology applied by the ProteinPilot software. We did verify this manually for all the peptides reported in this paper.

For enzyme activity measurements, frozen tissues were homogenised in seven volumes of assay buffer (0.1 M NaeK-phosphate buffer, pH 7.0) with a mechanical homogeniser (Kontes Microtube Pellet Pestle Rod with Motor, Vineland, NJ, USA). Protease activity specific to pair of basic residues (pro-hormone convertase activity) was assayed by cleavage of the synthetic substrate pyroglutamyl-arginyl-threonyl-lysyl-arginyl-7-amido4-methyl-coumarin (Pyr-RTKR-AMC, Bachem AG, Bubendorf, Switzerland), measuring the liberation of the fluorescent product AMC. Around 5 mg of tissue homogenate was pre-incubated (20 min) in 500 ml of assay buffer (1 mM CaCl2, 50 mM acetate, pH 7) in the presence or absence of 10 nM purified recombinant PREP (from a 100 stock in assay buffer) and/or 100 nM KYP-2047 (from a 10 stock in assay buffer which contained <0.01% DMSO). The reaction was started with the addition of Pyr-RTKR-AMC (to 100 mM) and the fluorescence of the free AMC released was read every minute for 60 min at 37  C by a Victor-II fluorescence reader (PerkinElmer, Waltham, MA, USA). Activity was calculated from the slope of the linear part of the curve. Free AMC was used as standard, and the basal fluorescence of Pyr-RTKR-AMC (less than 1%) was subtracted.

2.7. Bioinformatic and enrichment analysis

2.9. Protein determination

We classified the proteins which contained the altered peptides identified, by the functionally annotation, using web-based

Protein content was determined by the Bradford method (Bio-Rad, Hercules, CA, USA) using albumin from bovine serum as standard.

2.8. Tissue homogenization and enzymatic assays

1852

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859

2.10. Purification of PREP Recombinant porcine prolyl oligopeptidase was expressed in Escherichia coli and purified as described previously [27]. 2.11. Cytochrome c oxidase subunit I SDS-PAGE and western blotting Tissue homogenate samples were diluted to 1 mg/ml with loading buffer (63 mM TriseHCl, 10% Glycerol, 2% SDS, 0.0025% Bromophenol Blue, 5 mM b-mercaptoethanol, pH 6.8) and 20 mg of protein were separated in a Triseglycine 4e10% discontinuous gel. Proteins in the gel were transferred to nitrocellulose membrane (Millipore Billerica, MA, USA), by electro-transfer using a trans-blot cell (Bio-Rad, Hercules, CA, USA) for 20 min at 100 V in transfer buffer (25 mM Tris, 192 mM Glycine, 20% methanol, pH 8.3). PBS rinsed blots were incubated 1 h in 5 ml of blocking buffer (5% skim milk in 0.1% Tween 20, 20 mM TriseHCl, 500 mM NaCl, pH 7.2) and incubated overnight with anti COXI primary antibody (Abnova, Walnut, CA, USA) diluted to 1:500. After three washes, blots were incubated with an anti-rabbit-horseradish-peroxidase complex (Pierce, Rockford, IL, USA) diluted 1:3000 in blocking buffer. Protein visualization was performed using ECL-kit (Amersham-bioscience, Little Chalfont, UK) following manufacturer’s instructions. 3. Results 3.1. Peptide extraction The amounts of peptides extracted from control tissues, measured by the absorption at 280 nm (expressed as absorption units per ml, AU/ml), were not significantly different from those measured in extracts from treated animals, except for hippocampus where a w30% decrease in peptide content was observed in the KYP-2047 treated animals. In general, the peptide levels varied from tissue to tissue but roughly within the same order of magnitude (Fig. 2). In control and treated animals, the level of peptides from the liver was the highest, with around 12 UA/ml. Those from the cerebellum and cortex followed with 6e8 UA/ml, and the levels from the hypothalamus were around 2e3 UA/ml. The extracted peptides levels from the striatum and cortex were lower, just below 2 UA/ml. We used the same specific amount of peptides for every sample for iTRAQ labelling and subsequent 2-D HPLC/MS/MS analysis. 3.2. The peptidome of the 4-day KYP-2047 treated rats Considering all the runs for all areas and tissues analysed (duplicates of a total of 7 different tissues/areas), we were able to

Fig. 2. Average peptide levels in the extracts from liver (Liv) kidney (Kid) and 5 brain areas (Str, striatum; Cx, frontal cortex; Hy, hypothalamus; Cb, cerebellum; R.b., rest of brain) from control and KYP-2047 treated rats. Peptide levels are reported as the absorbance at 280 nm per ml of extract (mean  SEM). *p-value < 0.05 of control vs. treated conditions in the hypothalamus.

identify around 1600 different peptides with at least 95% of confidence. From all those peptides, only 391 were altered due to treatment (p-values < 0.05). These peptides are listed in supplementary materials (Tables S1eS6). The p-value was the only parameter we used to select these peptides (set as <0.05), at the condition that the ratio in peptide abundance between treated and control conditions would be different from 1 (see Discussion). All modified peptides per tissue area found are listed according to the ratio treated/control (Tables S1aeS7b) and to the precursor protein (Tables S1beS7b) in the supplementary material. In general, most of the peptides, altered by the PREP inhibitor treatment, were fragments without possessing known biological activity by themselves, and coming from relatively large proteins. In addition, the sequence of most of those peptides did not follow a post-proline cleaving fragmentation, as it has been expected by the inhibition of the proline specific peptidase, PREP. 3.2.1. Kidney and liver It could be considered that the relatively high amounts of peptides extracted from the kidney and liver might have been due to their high endogenous proteolytic activity (Fig. 2). However, as our study took into account just those peptides that showed altered levels over PREP inhibition, the products of the background protease activity are subtracted out upon standardisation to control levels. A total of 213 peptides were found significantly changed in the kidney and 19 peptides in the liver after the 4-day treatment with KYP-2047. The peptides found altered in the kidney were part of a total of 107 different proteins, several of them containing more than one peptide, i.e. 19 peptides identified were for acyl-CoAbinding protein, and 10 peptides were part of the ATP-synthasecoupling factor 6. It is of interest that from all 213 modified peptides by PREP inhibition in the kidney, the majority were found to be decreased by the treatment, and only 9 were found increased. The highest decrease was one of the peptides part of the acyl-CoAbinding protein (4.04), and the highest increase was a peptide derived from para-thymosin (þ3.78) (Tables S1eS7). In the liver, the 19 peptides found modified by PREP inhibition were fragments of 14 different proteins. Compared with the kidney, in the liver all the peptides changed were increased by the treatment with the exception of one peptide from the elongation factor 1-a 1 (1.7 fold change). The biggest change (þ3.0) was in a peptide of a mitochondrial heat-shock protein (Table S2b). The altered peptide levels in the kidney and liver were mostly from mitochondrial proteins, but also an important number of proteins were related to protein translation or gene transcription. We performed a bioinformatic evaluation to know whether there was a concerted change of any specific subset of proteins according common cellular location or having similar function. We used the annotations on the respective protein data bases. The GO terms, cellular compartment and molecular function were inspected for enriched characteristics (see Methods) and the results were classified in terms of the tissue source (Fig. 3). The analysis shows that more than 40% of the source proteins in the kidney and over 20% of those proteins found modified in the liver were of mitochondrial origin (Fig. 3A). An important proportion of proteins from both the liver and the kidney (over 30% and 20% respectively) were soluble cytosolic proteins. The third most abundant protein location, especially for the kidney, was nuclear (Fig. 3A). Functional analysis indicated that 50% of the proteins, which contained altered peptides upon PREP inhibition, were annotated as mitochondrial oxidoreductases/transporters or nucleic acid binding proteins (Fig. 3C). None of these peptides followed a post-proline cleaving pattern. 3.2.2. Brain The 4-day treatment with KYP-2047 altered levels of 159 specific peptides across the brain; 48 in the hypothalamus, 77 in the

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859

1853

Fig. 3. Bioinformatic analysis of the proteins containing altered peptide levels upon PREP inhibition, based on different gene ontology (GO) terms for peripheral tissues (A, C) and brain areas (B, D). Cellular compartment (A, B), and molecular function clustering (C, D).

frontal cortex, 31 in the cerebellum and 3 in the striatum, in addition to the 24 peptides found altered in other structures of the brain (Tables S3eS7). In general, the peptide levels were decreased by PREP inhibition: in the hypothalamus two thirds, in the cortex more than 75% and in the cerebellum all the peptides were decreased. In the rest of the brain tissue showed equal changes in both directions. In general, the peptides altered in the brain were also found to be part of large mitochondrial proteins, but also a large proportion was from secreted protein precursors. Interestingly, several peptides from fibrinogen a-chain, a secreted protein involved in coagulation, were found mostly down-regulated by PREP inhibition across the brain. The striatum behaved peculiarly since only 3 altered peptides were identified after the KYP-2047 treatment, among all peptides identified, and therefore the data are not included in the bioinformatic analysis run with the other brain parts. In these areas, subunit 5A of the cytochrome oxidase complex was increased more than 5 times, and two peptides from calmodulin and protachykinin-1 were only moderately changed. 3.2.2.1. Hypothalamus. In the hypothalamus, 50% of the peptides altered by PREP inhibition were part of secreted proteins, but also an important proportion (20%) were from mitochondrial proteins (Fig. 3B). The rest of proteins that contained altered peptides were cytoplasmic, membranal, lyzosomal and nuclear. The functional analysis showed that an important proportion of the secreted protein precursors, containing peptides altered by PREP inhibition, were hormones. Almost 30% of the hypothalamic proteins were classified as oxidoreductases and ion-transport proteins (Fig. 3D). 3.2.2.2. Frontal cortex. In the frontal cortex, 35% of the proteins containing the peptides altered by PREP inhibition, were secreted proteins (Fig. 3B). Cytoplasmic and membrane-bound proteins accounted for about 40%, and mitochondrial proteins constituted slightly more than 10% (Fig. 3B) of the total. As to biological

function, the peptides found altered in frontal cortex by PREP inhibition, most often, were part of hormone precursors, followed by oxidoreductases/transporters and by nucleic acid binding proteins, a group that was not represented in the hypothalamus (Fig. 3D). 3.2.2.3. Cerebellum. A 4-day PREP inhibition altered cerebellar peptides derived mainly from cytoplasmic proteins, notably from mitochondria and nucleus while secreted proteins remained in a minority (Fig. 3B). Functionally, most of the proteins were classified as oxidoreductases/ion-transport proteins, without significant hormonal derived peptides, resembling the peripheral tissues (Fig. 3D). An important number of proteins were nucleic acid and calcium and calmodulin binding proteins (Fig. 3D). 3.2.3. Neuropeptides in the brain and peripheral tissues One of the main interests of our study was to specifically look for proline containing neuropeptides that had been found to be PREP substrates in vitro (for a comprehensive list of potential natural substrates of PREP, see [4]). With our technique, we were able to detect and identify 25 internal peptides contained in neuroactive peptides, derived from substance P, b-endorphin, galanin, somatostatin, enkephalin, dynorphin, neuropeptide Y, neuropeptidee glutamic acideisoleucine (NEI), secretoneurin, secretogranin-3 and somatostatin in the brain as shown in Table 1. The peptides from octadecaneuropeptide were detected only in the liver and kidney. In general, the levels of the parent peptides in the various brain areas were not significantly modified by PREP inhibition, with some exceptions. A fragment from somatostatin was found significantly decreased in the hypothalamus and cortex. One peptide from enkephalin, and another one from secretoneurin were found increased in the hypothalamus, where a peptide from neuropeptide Y was decreased. In the cortex, one peptide from NEI was found significantly decreased, but another one from secretogranin-3 showed almost three-fold increase (Table 1). Notably, no post-

1854

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859

Table 1 Internal peptides of several neuroactive peptides identified in rat tissue extracts, and the variation of their levels caused by the 4-day KYP-2047 treatment. Neuropeptide

Tissue

Sequence or fragment

Ratioa

Substance P b-endorphin

Hypothalamus Hypothalamus Hypothalamus Hypothalamus Liver

RPKPQQF YGGFMTSEKSQTPLVTLFKNAIIKNVHKKGQ YGGFMTSEKSQTPLVTL GWTLNSAGYLLGPHAIDNHRSFSDKHGLT QATVGDVNTDRPGLLDLK GDVNTDRPGLLD GDVNTDRPGLLDL TVGDVNTDRPGLLDLK NTDRPGLLDLK GDVNTDRPGLLD DVNTDRPGLLDLK DVNTDRPGLLD SANSPAMAPRE SANSPAMAPRE SANSPAMAPRE SLPSDEEGESYSKEVPEM SLPSDEEGESYSKEVPEM GGFLRKYPK PKLKWNQ SSPETLISDLLMRESTENAPRT EIGDEENSAKFPI TNEIVEEQYTPQSLATLESVFQELGKLTGPSNQ ELSAERPLNEQIAEAEADKI

1.11 1.11 1.00 1.00 1.11 1.25 1.11 1.67*** 1.67* 1.11 1.25 1.25 1.43*** 2.5* 1.25 1.4** 1.11 1.1 1.0 1.3* 1.9** 1.8** 2.800

Galanin Octadecaneuropeptide

Kidney

Somatostatin

Enkephalin Dynorphin A Neuropeptide Y NEI Secretoneurin Secretogranin-3 a

Hypothalamus Frontal cortex Rest of the brain Hypothalamus Frontal cortex Frontal cortex Rest of the brain Hypothalamus Frontal cortex Hypothalamus Frontal cortex

The ratio changes of statistical significance are indicated by asterisks: *p-value < 0.05, **p-value < 0.01. ***p-value < 0.005.

proline pattern was found in the peptides that were found decreased. On the other hand, all increased peptides contained an internal proline. 3.2.4. PREP and hormone processing In the hypothalamus and frontal cortex, a significant proportion of the affected peptides found altered by PREP inhibition originated from hormone precursors, but also from pro-proteins of secreted mature proteins. Accordingly, alignments of these particular peptide sequences were inspected in order to find common features within these peptides. All increased peptides upon PREP inhibition have an internal proline but none of the decreased peptides had this residue at the C-termini. It can be claimed that other proteases, perhaps carboxypeptidases, were degrading the peptides from the C-terminus after PREP action, erasing the terminal proline. We noticed consistently that all these peptides were immediately flanked by pairs of basic residues. The Table 2 lists all the peptides derived from hormonal precursors, and secreted pro-proteins, whose levels were found altered after PREP inhibition treatment. Arginine and lysine are found immediately before or after the first or the last amino acid of each identified peptide (underlined in Table 2). Most of these peptides were altered in the hypothalamus, but some also in the frontal cortex. Furthermore, it was remarkable that in the kidney we found as many as 19 peptides from Acyl-CoA-binding protein, which is the precursor of octadecaneuropeptide, flanked by basic residues (Table 3). The fact that peptides flanked by dibasic residues are found modified by KYP-2047 treatment could be explained by several ways: 1) PREP has different sequence specificity in vivo than in vitro; 2) KYP-2047 directly inhibits other proteases than PREP, or; 3) PREP activity is required for the regulation of secretory pathways. As all of these peptides were found derived from large proteins, and it is generally believed that PREP cleaves peptides of less than 30 amino acids, with well demonstrated preference to proline at least in vitro, the first possibility seems unlikely. Although there is a possibility of the existence of KYP-2047 off-targets, this compound has no significant inhibitory effect on large variety of different peptidases and isomerases tested [19,20]. It is known that convertases PC1/3, PC2, furin, PC4, PC5/6, PACE4, and PC7

(specifically cleaving proteins at pairs of basic residues) are activators of pro-hormones and pro-neuropeptides within the regulatory secretory pathway of neural and endocrine cells [28]. We therefore designed an in vitro experiment to test if the endogenous generic PC activity was sensitive to PREP, or PREP inhibition. We thus measured the cleavage of the synthetic peptide PyrRGKR-AMC, by tissue homogenates. Fig. 4 shows that a preincubation of brain homogenates with purified PREP reduced basal generic dibasic specific protease activity by 50% and that this inhibitory effect was prevented by the addition of KYP-2047, which itself did not affect the endogenous convertase activity (Fig. 4). 3.2.5. PREP and mitochondrial protein degradation The KYP-2047 treatment in rats produced changes in the levels of peptides that were part of mitochondrial proteins, especially in the kidney. In all tissues analysed, the mitochondrial proteins identified were quite often part of the oxidative phosphorylation route, especially subunits of the cytochrome oxidase, the bc1complex and the ATP-synthase (Table 4). We hypothesised that the 4-day KYP-2047 treatment affected the turnover of proteins involved in the oxidative phosphorylation. We assayed the levels of the subunit I of the cytochrome oxidase complex by western blot and observed a consistent decreasing trend in the liver of animals treated with KYP-2047 (Fig. 5). 4. Discussion The physiological role of PREP and the nature of its biological substrates have been critically discussed [4,6e8,29]. Due to the ability of this enzyme to cleave prolyl-containing biologically active short peptides in vitro, it has been considered that PREP directly metabolizes certain memory relevant neuropeptides in vivo and that PREP-specific inhibitors elevate their levels in the brain (for recent reviews see [3e5]). Although this thesis has received some support [30e32], it has been more challenged [20,33,34]. If the hypothesis is true, the levels of intact neuropeptides (substrates) would increase, and those of the digestion products (presumably with proline at the C-termini) would be decreased upon PREP inhibition.

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859

1855

Table 2 Peptides from secreted pro-proteins identified in the peptidome of rat hypothalamus (H) and frontal cortex (FC) and found significantly modified in their levels by the 4-day KYP-2047 treatment. Neuropeptide precursor

Ratioa H

Pro-thyroliberin 27e50 EAEQEEGAVTPDLPDLENVQVRPERRb 25e50 LPEAEQEEGAVTPDLPDLENVQVRPERR Pro-corticoliberin 127e140 RGAEDALGGHGALERERR 127e142 RGAEDALGGHGALERERR Pro-neuropeptide Y 68e90 KRSSPETLISDLLMRESTENAPRTR Cocaine- and amphetamine-regulated transcript protein 62e77 RAPGAVLQIEALQEVLKKLKSKR 62e79 RAPGAVLQIEALQEVLKKLKSKR Pro-MCH 131e143 KREIGDEENSAKFPIGRR Somatostatin 89e101 RSANSNPAMAPRERK Proenkephalin-A 242e259 KRFAESLPSDEEGESYSKEVPEMEKR 201e209 KRSPQLEDEAKELQKR 198e207 KRSPQLEDEAKELQKR 198e209 KRSPQLEDEAKELQKR Secretogranin-1 383e395 KRNHPDSELESTANR 434e451 KRLLDEGHDPVHESPVDTAKR 435e451 KRLLDEGHDPVHESPVDTAKR Secretogranin-2 168e181 RFPLMYEENSRENPFKR 184e216 KRTNEIVEEQYTPQSLATLESVFQELGKLTGPSNQKR 594e611 KVLEYLNQEQAEQGREHLAKR Secretogranin-3 38e57 RELSAERPLNEQIAEAEADKIKK Neuroendocrine protein 7B2 (Secretogranin-5) 198e210 KKSVPHFSEEEKEPE Neurosecretory protein VGF 220e235 RVPERAPLPPSVPSQFQAR 489e507 KRKRKKNAPPEPVPPPRAAPAPTHVR

1.6[ 2.1[

a b

Thyroliberin, TRH (5 identical tripeptides, QHP, starting at amino acid 77, 109, 154, 172, and 202)

1.4Y 1.2Y

Corticotropin-releasing factor (145e185)

1.3Y

Neuropeptide Y (30e65), C-flanking peptide of NPY (69e98)

1.4Y 1.2Y

CART(1e52), CART(55e102), CART(62e102)

1.4Y

1.9Y

Neuropeptide GeE (110e128), Neuropeptide EeI (131e143), MCH (147e165)

2.9Y

Antrin (25e34), Somatostatin-28 (89e116), Somatostatin-14 (103e116)

1.4[ 3.4[ 5.1[ 1.7Y 2.2Y 2.0Y 1.8Y 1.3Y 1.8[ 2.0[

1.3Y

ProSAAS 44e59 RSCSAASAPLAETSTPLRLRR Neurogranin 55e78 RKGPGPGGPGGAGGARGGAGGGPSGD Fibrinogen a-chain 19e36 TADTGTTSEFIEAGGDIR 20e36 ADTGTTSEFIEAGGDIR 27e36 EFIEAGGDIR

Active peptide FC

Secretogranin-1 (21e675), CCB peptide short form (615e674), CCB peptide long form (615e675)

Secretogranin-2 (31e619), Secretoneurin (184e216)

2.8Y

Secretogranin-3 (23e471)

1.8Y

Neuroendocrine protein 7B2 (25e210)

3.8Y 1.9Y

VGF(24e63), VGF(180e194), VGF(375e407), NERP-1 (285e309), NERP-2, (313e350), TLQP-62 (556e617), TLQP-30 (559e587), TLQP-21 (556e576), TLQP-11 (556e566), HFHH-10 (567e576), AQEE-30 (588e617), LQEQ-19 (599e617)

1.7Y

KEP (34e40), Big SAAS (34e59), Little SAAS (42e59), Big PEN-LEN (221e260), PEN (221e242), PEN-20 (221e240), Little LEN (245e254), Big LEN (245e260)

3.3[ 2.5Y 1.8Y

Synenkephalin (25e97), Met-enkephalin (100e104, 107e111, 136e140, 212e216), PENK(114e133), PENK(143e185), Met-enkephalin-Arg-Gly-Leu (188e195), Leu-enkephalin (232e236), PENK(239e260), Met-enkephalin-Arg-Phe (263e269)

NEUG (55e78)

4.6Y 2.2[

All changes showed p-values < 0.05. Ratios in [ and Y indicate increase or decrease on the levels in PREP inhibited animals relative to controls. Flanking basic amino acid residues are shown in underline, residues not part of the identified peptide sequence are shown in smaller bold italics.

In the present study, we identified several intact neuropeptides but none of those were increased. In fact b-endorphin, galanin, and octadecaneuropeptide were not significantly changed at all, while NEI and secretoneurin were even decreased. We found significant decrease of peptides that were part of octadecaneuropeptide, somatostatin, enkephalin, neuropeptide Y and secretogranin-3, as products of PREP digestion, but none of those followed a proline specific pattern. We were not able to find other neuropeptides, reported to be cleaved by PREP in vitro, we cannot reach a definitive conclusion on the importance of PREP in neuropeptide metabolism. Previously, we have validated a peptidomic approach to study the changes after an acute PREP inhibition, and found only indications that PREP might have effects on the levels of peptides derived from mitochondrial proteins and hormone precursors [8]. The aim of the present study was to evaluate whether a repeated administration of KYP-2047 could make differences in peptide levels even more evident, reveal more peptides, not detected in the previous study, and allow bioinformatic analysis. In comparison with our previous peptidomic study of acute PREP inhibitor (ZPP)

administration, we were able to find around 90% of the peptides identified also in the new study (see Supplementary material Table S8). In the present study, we were able to detect much more peptides; at least one order of magnitude more than our preliminary studies, except for the striatum. Under the sub-chromic treatment reported here, the differences were indeed bigger for most of these peptides. Besides we now find also few contiguous peptides, to those found before, which were found to vary in accordance. Indeed, the long treatment causes a new steady-state which is globally similar to the short inhibition, but we could now identify much more peptides than before. Interestingly, we observed an enrichment of a number of altered peptides originating from proteins of specific location and function. When classifying the precursor proteins according to their function, five significant groups could be found (Fig. 3B). A special subset of peptides, found to be increased by a 4-day administration of KYP-2047, was derived from mitochondrial proteins. This was particularly evident in the kidney. Our previous peptidomic study after an acute PREP inhibition already reported alteration of

1856

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859

Table 3 Internal peptides of acyl-CoA-binding protein, the 87 amino acid octadecaneuropeptide pro-protein (first row), having modified levels by the 4-day KYP-2047 treatment in rat kidney.

peptides derived from NADH dehydrogenase-1 alpha subunit 4 and cytochrome c oxidase subunit 5A [8]. Here, several peptides derived from the respiratory complexes II (succinate dehydrogenase), III (cytochrome c oxidase) and IV (ATP-synthase) were found to be affected by the 4-day PREP inhibitor treatment. We also detected an apparent decrease of cytochrome oxidase subunit I in the KYP-2047 treated rats. This may suggest that PREP inhibition could affect the turnover of mitochondrial proteins. However, further research is needed to elucidate the mechanisms or consequences of these findings. Owing to the substrate size limitation of PREP, the whole precursors of the peptides cannot be direct substrates of PREP, and thus a prior protease activity is necessary to generate fragments under <30 amino acids, a presumable substrate size limit. On the other hand, and due to the lack of consistent proline specific cleavage pattern on most of the identified peptides, it could be hypothesized that the changes observed are indirectly driven by modulating PREP activity. In some cases, peptides products of these putative peptidases were identified, as in the case of thymosin b-4 (Tb4). We found that several overlapping peptides from this

Fig. 4. Basal basic amino acid pair specific protease activity in brain homogenates (left white bar) or pre-incubated in presence of recombinant PREP (right white bar), and with the presence of PREP and KYP-2047 (black bar). The release of fluorescent AMC from the substrate Z-RGKR-AMC (see Methods) was followed during 60 min at 37  C. PREP activity added was w20 times higher than the endogenous tissue PREP activity. *p-value < 0.05, **p-value < 0.01. Mean  SEM, n ¼ 6 each group.

precursor were modified in their levels to opposite directions. For example, Tb4 fragments (2e20) and (21e43) were found decreased (1.6 and 2.0 respectively), but other overlapping fragments (2e18) and (19e43) of the same protein where increased (þ2.0 and þ3.0) (Fig. 6). It could be speculated that PREP may modulate the peptidases in the route which cleave the precursor Tb4. We have recently demonstrated that PREP is able to contribute to

Table 4 Mitochondrial proteins containing peptides with levels altered by the KYP-2047 treatment in the peptidome of the liver or/and kidney. Protein

Cytochrome c oxidase subunit 5A 95e107 NDFASAVRILEV 126/134e146 LRPTLNELGISTPEELGLDKV ATP-synthase-coupling factor 6 70e79 EVDRELFKLK 84e108 KGEMDKFPTFNFEDPKFEVLKDKPQS 83e92 GKGEMDKFPT 91/100e108 PTFNFEDPKFEVLDKPQS Cytochrome b-c1 complex subunit 11 32e48 SAVPATSEPPVLDVKRP 52/56e66/68 RESLSGQAATRPLVATV 68e79 VGLNVPASVRY ATP synthase subunit b 42e57 PLPPLPEYGGKVRLG Cytochrome c oxidase subunit 4 40/43e54 DRRDYPLPDVAHVKL 155/158e169 GFSAKWDYNKNEWKK Cytochrome c oxidase subunit 6A1 63e82 SRHEEHERPEFVAYPHLRIR NADH dehydrogenase 1 alpha 4 67e76/82 SVNVDYSKLKKEGPDF Cytochrome c oxidase subunit 6B1 2e12/13 AEDIKTKIKNYK 74e85/86 DDRIAEGTFPGKI Cytochrome c 2e14 GDVEKGKKIFVQK 84/85e100 AGIKKKGERADLIAYLK ATP synthase subunit e 58e71 ERELAEAEDVSIFK Cytochrome c oxidase subunit 7 55/56e63 IVRHQLLKK Cytochrome c oxidase subunit 6C2 56e76 NYDSMKDFEEMRQAG VFQSAK a

Ratioa Liver

Kidney

2.7[ 1.9[

2.5Y

2.1[

1.7Y 2.2Y 1.4Y 2.1Y 1.4Y 1.8Y 2.3Y

1.4[ 1.5Y 1.5Y 1.4Y 2.0Y 2.2Y 1.7Y 1.6Y 2.1Y 1.7Y 1.6Y 1.5Y

Ratios in [ and Y indicate increase or decrease on the levels in PREP inhibited animals relative to controls.

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859

Fig. 5. Cytochrome c oxidase subunit I levels, measured by western immunoblotting, in the liver of control and KYP-2047 treated rats. Optical density of cytochrome c oxidase subunit I bands were normalized to GADPH levels. Mean  SEM, n ¼ 6 for controls and n ¼ 9 for KYP-2047 treated animals. A representative western blot is also shown.

a production of an active peptide derived from this cytosolic protein [9]. Incubation of recombinant PREP with Tb4 revealed that PREP itself was indeed unable to cleave the full 43 mer-length Tb4. However, when a tissue homogenate was added to the mixture, the PREP product (Ac-SDKP) was produced. This indicated that PREP needed a preceding cleavage of the 43-amino acid peptide to produce fragments small enough to be hydrolysed by PREP. Surprisingly, when we followed Tb4 initial degradation by tissue homogenates, an addition of recombinant PREP significantly inhibited this initial step. We also found more examples of these antagonistic changes on levels of peptides of some other precursor proteins, as pro-enkephalin A, secretogranin-2, cytochrome c oxidase 5A, glucose-6-phosphatase and para-thymosin (see Supplementary material). Previous peptidomics studies using specific PREP inhibitors, have also reported alterations of inactive peptides derived from hormone precursors, such as pro-SAAS precursor, VGF nerve growth factor inducible and somatostatin [7,8], but no further explanations have been offered. We have here identified sequences of pro-hormone derived peptides, altered by PREP inhibition characterized by the presence of pairs of basic amino acids at the C-

1857

or N-termini or both, depending only if the peptide was internal or not. One of many explanations is that the proteases, responsible to process those pro-hormones, are sensitive to active PREP (Fig. 7). In fact, we found that pre-incubation of brain homogenates with recombinant PREP caused a 50% reduction of the specific dibasic amino acid pair activity. Furthermore, this inhibition was reverted by KYP-2047 (Fig. 4). It was remarkable to find that in the kidney, as a result of PREP inhibition treatment, there were found altered levels of many peptides derived from Acyl-CoA-binding protein (ACBP). This is a secreted protein, which binds Acyl-CoA esters, with GABA receptor modulator activity, and the precursor of octadecaneuropeptide and triakontatetraneuropeptide. All ACBP derived peptides found altered by PREP inhibition in the rat kidney are shown in Table 3. It can be observed that contiguous peptides, out of the sequence of the mature octadecaneuropeptide were found decreased. This indicates that the whole processing of ACBP to the mature neuropeptide was inhibited but the action is not proline specific but rather basic residue specific. Proteases known to have specificity towards pairs of basic residues are furin, PC1/3 and PC2, which are the principal activators of pro-hormones and pro-neuropeptides within the regulated secretory pathway of neural and endocrine cells: proopiomelanocortin (precursor of ACTH and b-endorphin), proinsulin, pro-melanin concentrating hormone, secretogranin II, pro-enkephalin, pro-dynorphin, pro-somatostatin, chromogranins A and B, pro-PACAP, pro-enkephalin, and the neurotrophic factor pro-BDNF, vascular endothelial growth factor (VEGF), among others. Interestingly, most of those were found in our analysis. There is evidence that convertases are regulated by proteins and peptides. Of interest is that PC2 is regulated by a w30 kDa neuroendocrine PC2-binding protein 7B2 that is co-regulated with it [35]. We have found that PREP inhibition caused changes in peptide levels from regulatory peptides like 7B2. On the other hand, we found peptides from a protease inhibitor; pro-SAAS, also called proprotein convertase subtilisin/kexin type 1 inhibitor. These findings support the speculation that PREP activity could be modulating PCs. However, more research should be done before any conclusion can be drawn. Our peptide extraction approach was designed to minimise unspecific post-mortem proteolysis. However, there are other technical limitations that have to be considered. Although the isotopic labelling improves the quantification sensitivity, it also limits the range of peptides that can be quantified since only the tagged peptides can be measured, especially short peptides are unlikely to be labelled with a technique that has a low-mass cut-off limitation [10,36]. Thus, peptides which were altered by PREP inhibition, but were not labelled, are invisible to our analysis. Furthermore, there is a detection limit, and changes in peptide

Fig. 6. Scheme of initial degradation steps of thymosin b-4 based on the altered internal peptides found in frontal cortex. Ratios of the changes of peptide levels due to PREP inhibition are shown. KYP-2047 decreases the levels of the products of peptidase X. As a result, the alternative route, via peptidase Y, is activated as reflected by the relative increase of the levels of the products of this second peptidase.

1858

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859

Fig. 7. Scheme showing a hypothetical explanation of PREP relation to pro-protein convertase activity. Some dibasic specific pro-protein convertase(s) (Convertase) interact directly, or indirectly (?), with PREP producing their activation or inactivation. Addition of KYP-2047 dampers the interaction resulting on the return of convertase(s) to basal activity. The symbols “- -” denote pairs of basic residues in the pro-protein sequence.

levels under this limit are obviously left unidentified, and thus unaccounted. Additionally, some peptides may be unstable to the acid-boiling extraction method used here, and this could also reduce their concentration probably to undetectable levels. There are also some other methodological aspects that need to be taken into account. We want to point out that the analysis conditions were non-stringent. We set a statistical significance for a particular peptide change, where the p-value was <0.05, and included all those cases in the analysis regardless of the absolute value of the change. This may raise a concern than minor changes may lack biological relevance. Since the tissue dissections were done macroscopically, differences in peptide levels represent only an average of the whole tissue sample. Although PREP is found in all tissues, it is concentrated in fairly specific cell types, groups or cellular layers, and the neighbouring groups of cells may contain low or undetectable amounts of PREP that dilute the changes [2]. Although conditions and methods were different among the reports on peptidomic changes upon PREP inhibition (recently reviewed by Lone et al., 2010 [36]), we could observe significant overlap of our results and those other MS-based analysis performed before using different, or no PREP inhibitors, in different animals [6e8]. Brandt et al. (2005) [6] analysed the effect of recombinant PREP on a fractionated pig brain peptide pool, Nolte et al. (2009) [7] used S17092-treated mice, and the samples were collected at 1 or 4 h, after the inhibitor, and Tenorio-Laranga et al. (2009) [8] used ZPro-Prolinal in rats, and the samples were collected at 4 h, using the similar snap-freezing procedure as in the present study. Several identical peptides were found by Nolte et al. (2009) [7] and us, like those derived from neurogranin, pro-thyroliberin, and Tb4. Moreover, Nolte et al. (2009) [7] found peptides derived from the same proteins, like pro-enkephalin, pro-SAAS, and protein VGF as we did. Like Brandt et al. (2005) [6], we also identified proteins derived from identical proteins, i.e. actin, haemoglobins, fructosebisphosphate aldolase c (very same peptide), PEP19 (here called Purkinje cell protein 4), neurogranin and b-synuclein. Compared to our previous acute peptidomic analysis, we found the same peptides derived from proteins like somatostatin, pro-SAAS, a-subunit 4 of the NADH dehydrogenase complex, breast carcinoma-amplified sequence homologue 1, calmodulin, fibrinogens, vitamin K epoxide reductase, Tb4, cerebelins, cytochrome c oxidase subunit 5, histone H2b and haemoglobin b chain. We found, in the present study, around 90% of the peptides modified by the acute treatment reported earlier (see Supplementary material

Table S8). A 4-day KYP-2047 treatment revealed much more peptides than the acute treatment, particularly those related to mitochondrial function. The advantage of our analysis is that we were able to obtain a significant number of peptides in order to apply statistically confident bioinformatics methods, to characterize the physiological relevance of PREP inhibition. In conclusion, our peptidomic study reveals that a 4-day PREP inhibition affects specific subset of peptides related to concrete biological processes. In the hypothalamus, repeated PREP inhibition alters levels of fragments derived from hormone precursors. Overall, our results indicate that PREP activity influence on the levels of peptides derived from cytosolic proteins, but this effect could not be direct. The finding that some of the peptides are increased in one organ and decreased in another is not coherent with the idea of these peptides being substrates and products of PREP. Thus the effects of PREP inhibition on peptide levels must be indirect. This is supported by the broad lack of proline specific pattern of cleavage of the identified peptides. It can be speculated that PREP activity may affect the activity of other proteases changing in turn the levels of substrates and products of these enzymes. However, our current data do not provide any evidence for this thesis. On the other hand, repeated PREP inhibition seems to have a special influence on the turnover of mitochondrial protein, particularly in the kidney. As to the classical PREP substrate peptides, our results are not conclusive, since most of them were not detected, and those that were identified, were not found modified by PREP inhibition under our assay conditions. Acknowledgements We thank Ms. Marjo Vaha and Ms. Anna Niemi for excellent technical support, and to the Proteomic facility of the Centro de Investigación Príncipe Felipe, member of the Spanish ProteoRed, for outstanding proteomic analysis and support. These studies were financed by the European Commission 7th Work Programme of Health to JAGH (FP7-Health-F2-2008-223077), Marie Curie Actions to JTL (FP7-PEOPLE-2009-IEF-254127) and the Academy of Finland (No. 210758 and 1131915) and Sigrid Juselius Foundation to PTM. Appendix A. Supplementary material Supplementary material related to this article can be found online at doi:10.1016/j.biochi.2012.04.005. References [1] T.T. Myöhänen, J.I. Venäläinen, J.A. García-Horsman, M. Piltonen, P.T. Männistö, Distribution of prolyl oligopeptidase in the mouse whole-body sections and peripheral tissues, Histochem. Cell. Biol. 130 (2008) 993e1003. [2] T.T. Myöhänen, J.I. Venäläinen, E. Tupala, J.A. García-Horsman, R. Miettinen, P.T. Männistö, Distribution of immunoreactive prolyl oligopeptidase in human and rat brain, Neurochem. Res. 32 (2007) 1365e1374. [3] P.T. Männistö, J. Venäläinen, A. Jalkanen, J.A. Garcia-Horsman, Prolyl oligopeptidase: a potential target for the treatment of cognitive disorders, Drug News Perspect. 20 (2007) 293e305. [4] J.A. García-Horsman, P.T. Männistö, J.I. Venäläinen, On the role of prolyl oligopeptidase in health and disease, Neuropeptides 41 (2007) 1e24. [5] A.M. Lambeir, Translational research on prolyl oligopeptidase inhibitors: the long road ahead, Expert Opin. Ther. Pat. 21 (2011) 977e981. [6] I. Brandt, K. De Vriendt, B. Devreese, J. Van Beeumen, W. Van Dongen, K. Augustyns, I. De Meester, S. Scharpé, A.M. Lambeir, Search for substrates for prolyl oligopeptidase in porcine brain, Peptides 26 (2005) 2536e2546. [7] W.M. Nolte, D.M. Tagore, W.S. Lane, A. Saghatelian, Peptidomics of prolyl endopeptidase in the central nervous system, Biochemistry 48 (2009) 11971e11981. [8] J. Tenorio-Laranga, M.L. Valero, P.T. Männistö, M. Sanchez del Pino, J.A. GarciaHorsman, Combination of snap freezing, differential pH two-dimensional reverse-phase high-performance liquid chromatography, and iTRAQ technology for the peptidomic analysis of the effect of prolyl oligopeptidase inhibition in the rat brain, Anal. Biochem. 393 (2009) 80e87.

J. Tenorio-Laranga et al. / Biochimie 94 (2012) 1849e1859 [9] T. Myöhänen, J. Tenorio-Laranga, B. Jokinen, R. Vazquez-Sanchez, M. MorenoBaylach, J. Garcia-Horsman, P. Männistö, Prolyl oligopeptidase induces angiogenesis both in vitro and in vivo in a novel regulatory manner, Br. J. Pharmacol. 163 (2011) 1666e1678. [10] J. Tenorio-Laranga, P.T. Männistö, J.A. Garcia-Horsman, Hunting for peptide substrates of prolyl oligopeptidase: classical versus non-classical bioactive peptides, CNS Neurol. Disord. Drug Targets 10 (2011) 319e326. [11] F.Y. Che, L.D. Fricker, Quantitative peptidomics of mouse pituitary: comparison of different stable isotopic tags, J. Mass Spectrom. 40 (2005) 238e249. [12] H. Pan, F.Y. Che, B. Peng, D.F. Steiner, J.E. Pintar, L.D. Fricker, The role of prohormone convertase-2 in hypothalamic neuropeptide processing: a quantitative neuropeptidomic study, J. Neurochem. 98 (2006) 1763e1777. [13] X. Zhang, F.Y. Che, I. Berezniuk, K. Sonmez, L. Toll, L.D. Fricker, Peptidomics of Cpe(fat/fat) mouse brain regions: implications for neuropeptide processing, J. Neurochem. 107 (2008) 1596e1613. [14] D.M. Desiderio, Quantitative analytical mass spectrometry of endogenous neuropeptides in human pituitaries, Life Sci. 51 (1992) 169e176. [15] K. Skold, M. Svensson, A. Kaplan, L. Bjorkesten, J. Astrom, P.E. Andren, A neuroproteomic approach to targeting neuropeptides in the brain, Proteomics 2 (2002) 447e454. [16] A.J. Jalkanen, T.P. Piepponen, J.J. Hakkarainen, I. De Meester, A.-M. Lambeir, M.M. Forsberg, The effect of prolyl oligopeptidase inhibition on extracellular acetylcholine and dopamine levels in the rat striatum, Neurochem. Int. 60 (2012) 301e309. [17] E.M. Jarho, J.I. Venäläinen, J. Huuskonen, J.A. Christiaans, J.A. Garcia-Horsman, M.M. Forsberg, T. Järvinen, J. Gynther, P.T. Männistö, E.A. Wallen, A cyclopent2-enecarbonyl group mimics proline at the P2 position of prolyl oligopeptidase inhibitors, J. Med. Chem. 47 (2004) 5605e5607. [18] A.J. Jalkanen, J.J. Hakkarainen, M. Lehtonen, T. Venäläinen, T.M. Kääriäinen, E. Jarho, M. Suhonen, M.M. Forsberg, Brain pharmacokinetics of two prolyl oligopeptidase inhibitors, JTP-4819 and KYP-2047, in the rat, Basic Clin. Pharmacol. Toxicol. 109 (2011) 443e451. [19] A.J. Jalkanen, K.A. Puttonen, J.I. Venäläinen, V. Sinerva, A. Mannila, S. Ruotsalainen, E.M. Jarho, E.A. Wallén, P.T. Männistö, Beneficial effect of prolyl oligopeptidase inhibition on spatial memory in young but not in old scopolamine-treated rats, Basic Clin. Pharmacol. Toxicol. 100 (2007) 132e138. [20] A.J. Jalkanen, K. Savolainen, M.M. Forsberg, Inhibition of prolyl oligopeptidase by KYP-2047 fails to increase the extracellular neurotensin and substance P levels in rat striatum, Neurosci. Lett. 502 (2011) 107e111. [21] J.I. Venäläinen, J.A. Garcia-Horsman, M.M. Forsberg, A. Jalkanen, E.A. Wallen, E.M. Jarho, J.A. Christiaans, J. Gynther, P.T. Männistö, Binding kinetics and duration of in vivo action of novel prolyl oligopeptidase inhibitors, Biochem. Pharmacol. 71 (2006) 683e692. [22] N.A. Karp, K.S. Lilley, Investigating sample pooling strategies for DIGE experiments to address biological variability, Proteomics 9 (2009) 388e397.

1859

[23] G. Dennis Jr., B.T. Sherman, D.A. Hosack, J. Yang, W. Gao, H.C. Lane, R.A. Lempicki, DAVID: Database for Annotation, Visualization, and Integrated Discovery, Genome Biol. 4 (2003) P3. [24] W. Huang da, B.T. Sherman, R.A. Lempicki, Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources, Nat. Protoc. 4 (2009) 44e57. [25] F. Al-Shahrour, J. Carbonell, P. Minguez, S. Goetz, A. Conesa, J. Tárraga, I. Medina, E. Alloza, D. Montaner, J. Dopazo, Babelomics: advanced functional profiling of transcriptomics, proteomics and genomics experiments, Nucleic Acids Res. 36 (2008) W341eW346. [26] F. Al-Shahrour, P. Minguez, J. Tárraga, I. Medina, E. Alloza, D. Montaner, J. Dopazo, FatiGO þ: a functional profiling tool for genomic data. Integration of functional annotation, regulatory motifs and interaction data with microarray experiments, Nucleic Acids Res. 35 (2007) W91eW96. [27] J.I. Venäläinen, R.O. Juvonen, M.M. Forsberg, A. Garcia-Horsman, A. Poso, E.A. Wallen, J. Gynther, P.T. Männistö, Substrate-dependent, non-hyperbolic kinetics of pig brain prolyl oligopeptidase and its tight binding inhibition by JTP-4819, Biochem. Pharmacol. 64 (2002) 463e471. [28] N.G. Seidah, What lies ahead for the proprotein convertases? Ann. NY Acad. Sci. 1220 (2011) 149e161. [29] I. Brandt, S. Scharpé, A.M. Lambeir, Suggested functions for prolyl oligopeptidase: a puzzling paradox, Clin. Chim. Acta. 377 (2007) 50e61. [30] K. Toide, T. Fujiwara, Y. Iwamoto, M. Shinoda, K. Okamiya, T. Kato, Effect of a novel prolyl endopeptidase inhibitor, JTP-4819, on neuropeptide metabolism in the rat brain, Naunyn Schmiedebergs Arch. Pharmacol. 353 (1996) 355e362. [31] K. Toide, K. Okamiya, Y. Iwamoto, T. Kato, Effect of a novel prolyl endopeptidase inhibitor, JTP-4819, on prolyl endopeptidase activity and substance Pand arginine-vasopressin-like immunoreactivity in the brains of aged rats, J. Neurochem. 65 (1995) 234e240. [32] K. Toide, M. Shinoda, A. Miyazaki, A novel prolyl endopeptidase inhibitor, JTP4819eits behavioral and neurochemical properties for the treatment of Alzheimer’s disease, Rev. Neurosci. 9 (1998) 17e29. [33] M.A. Cavasin, T.D. Liao, X.P. Yang, J.J. Yang, O.A. Carretero, Decreased endogenous levels of Ac-SDKP promote organ fibrosis, Hypertension 50 (2007) 130e136. [34] I. Peltonen, A.J. Jalkanen, V. Sinerva, K.A. Puttonen, P.T. Männistö, Different effects of scopolamine and inhibition of prolyl oligopeptidase on mnemonic and motility functions of young and 8- to 9-month-old rats in the radial-arm maze, Basic Clin. Pharmacol. Toxicol. 106 (2010) 280e287. [35] M. Mbikay, N.G. Seidah, M. Chretien, Neuroendocrine secretory protein 7B2: structure, expression and functions, Biochem. J. 357 (2001) 329e342. [36] A.M. Lone, W.M. Nolte, A.D. Tinoco, A. Saghatelian, Peptidomics of the prolyl peptidases, AAPS J. 12 (2010) 483e491. [37] A. Brockmann, S.P. Annangudi, T.A. Richmond, S.A. Ament, F. Xie, B.R. Southey, S.R. Rodriguez-Zas, G.E. Robinson, J.V. Sweedler, Quantitative peptidomics reveal brain peptide signatures of behavior, Proc. Natl. Acad. Sci. U S A 106 (2009) 2383e2388. [38] M. Svensson, K. Skold, P. Svenningsson, P.E. Andren, Peptidomics-based discovery of novel neuropeptides, J. Proteome Res. 2 (2003) 213e219.