Proteome Analysis of Erythrocytes Lacking AMP-Activated Protein

Jan 10, 2011 - The PAK2 kinase, previously implicated in apoptosis, was detected as downregulated in erythrocytes of ampk−/− mice, pointing to its...
0 downloads 0 Views 3MB Size
Download PDF
ARTICLE pubs.acs.org/jpr

Proteome Analysis of Erythrocytes Lacking AMP-Activated Protein Kinase Reveals a Role of PAK2 Kinase in Eryptosis Christine Zelenak,† Michael F€oller,† Ana Velic,† Karsten Krug,‡ Syed M. Qadri,† Benoit Viollet,§,||,^ Florian Lang,*,† and Boris Macek*,‡ †

Department of Physiology, University of T€ubingen, Germany INSERM, U1016, Institut Cochin, Paris, France CNRS, UMR8104, Paris France ^ Universite Paris Descartes, Paris, France ‡ Proteome Center Tuebingen, University of T€ubingen, Germany

)

§

bS Supporting Information ABSTRACT: Activation of AMP-activated protein kinase (AMPK) upon energy depletion stimulates energy production and limits energy utilization. Erythrocytes lacking AMPK are susceptible to suicidal cell death (eryptosis). A hallmark of eryptosis is cell membrane scrambling with phosphatidylserine exposure at the erythrocyte surface, which can be identified from annexin V-binding. AMPKR1-deficient mice (ampk-/-) suffer from anemia due to accelerated clearance of erythrocytes from circulating blood. To determine the link between AMPK and the eryptotic phenotype, we performed a global proteome analysis of erythrocytes from ampk-/- mice and wild-type mice using high-accuracy mass spectrometry and label-free quantitation and measured changes of expression levels of 812 proteins. Notably, the p21-activated kinase 2 (PAK2), previously implicated in apoptosis, was detected as downregulated in erythrocytes of ampk-/- mice, pointing to its potential role in eryptosis. To validate this, we showed that specific inactivation of PAK2 with the inhibitor IPA3 in human and murine ampkþ/þ erythrocytes increases the binding of annexin V and augments the stimulating effect of glucose deprivation on annexin V-binding. Inhibition of PAK2 failed to significantly modify annexin V-binding in ampk-/- erythrocytes, showing that AMPK and PAK2 exert similar phenotypes upon inactivation in erythrocytes. This study presents the first large-scale analysis of protein expression in erythrocytes from AMPKR1-deficient mice and reveals a role of PAK2 kinase in eryptosis. KEYWORDS: AMPK, PAK2, eryptosis, proteomics, orbitrap

’ INTRODUCTION The AMP-activated protein kinase (AMPK) is activated upon increase in the cytosolic AMP/ATP concentration ratio and thus senses the energy status of the cell.1,2 It modulates several cellular functions such as stimulation of cellular glucose uptake, glycolysis, fatty acid oxidation and expression of enzymes required for ATP production.2-5 AMPK further downregulates energy-consuming cellular mechanisms including protein synthesis, gluconeogenesis and lipogenesis.2,3,6 By the regulation of cellular metabolism AMPK supports cell survival during energy depletion,6,7 and it is also known to inhibit cell growth and to trigger suicidal cell death or apoptosis.7 Similar to apoptosis of nucleated cells, erythrocytes may undergo suicidal cell death or eryptosis,8 which is characterized by cell shrinkage, cell membrane blebbing and scrambling that leads to exposure of phosphatidylserine at the erythrocyte surface.9-11 The r 2011 American Chemical Society

latter can be triggered by cytosolic Ca2þ increase9,10 following activation of Ca2þ-permeable cation channels.12,13 Ca2þ further activates Ca2þ-sensitive Kþ channels,14,15 causing exit of KCl and water which leads to cell shrinkage.11 In addition to its antiapoptotic effect,7 AMPK also inhibits eryptosis.16 Unlike most other cells, erythrocytes exclusively express the AMPKR1 isoform, the absence of which cannot be compensated by AMPKR2.16 Accordingly, AMPKR1-deficient mice suffer from severe anemia due to excessive eryptosis.17 Although the eryptosis was extensively studied at the molecular and physiological level and its phenotype is well described, little is known about the exact molecular mechanisms underlying the action of AMPKR1 in this vital process. To gain insight into the global protein changes affected by this kinase, here Received: October 4, 2010 Published: January 10, 2011 1690

dx.doi.org/10.1021/pr101004j | J. Proteome Res. 2011, 10, 1690–1697

Journal of Proteome Research we perform the first proteome analysis of erythrocytes lacking AMPKR1 and compare them to wild-type erythrocytes. Mass spectrometry is the method of choice for global analysis of protein levels in erythrocytes, since they lack genetic material needed to assess gene expression at the transcription level. Proteomics has already been used successfully for the qualitative characterization of protein profiles of normal erythrocytes.18-20 Pasini et al. used high-accuracy mass spectrometry to identify 314 membrane and 252 soluble proteins21 in human erythrocytes and several other studies have followed suit. The number of proteomic studies on erythrocytes and blood components is rising, but so far none of the studies addressed quantitative changes of protein levels in eryptosis. Here we present the first large-scale analysis of protein expression in red blood cells undergoing eryptosis. We used high-accuracy mass spectrometry and label-free quantitation to measure the erythrocyte proteome of AMPKR1-deficient mice (ampk-/-) and compared it with that of wild-type mice. We quantified 812 erythrocyte proteins and performed the bioinformatic enrichment analysis of protein classes regulated in eryptosis. Among the regulated proteins, we detected the PAK2 kinase, previously reported to be involved in apoptosis. We confirmed the downregulation of PAK2 by Western blotting and showed that inactivation of PAK2 in human and murine ampkþ/þ erythrocytes leads to a typical eryptotic phenotype.

’ MATERIALS AND METHODS Erythrocytes

Isolation of Murine Erythrocytes from Full Blood. Murine erythrocytes from three 8-12 week-old ampk-/- mice22 and three wild-type mice were isolated and pooled separately. For the isolation of erythrocytes from murine full blood, density centrifugation with two polysucrose solutions of different concentrations (Histopaque 1083 and 1119, Sigma) was used according to the protocol provided by the company with slight modifications. Briefly, erythrocytes were pelleted at the bottom of the tube by centrifugation whereas other cells were captured in layers over the histopaque solutions. The erythrocyte pellet was isolated and washed three times with PBS. The purity of the erythrocytes was checked by FACS analysis using an antibody against CD45 to detect remaining lymphocytes. To confirm the FACS analysis of the antibody-stained erythrocytes, unstained samples were used and erythrocytes were excluded by gating according to the forward scatter dot plot analysis. The percentage of erythrocytes was ascertained to be 98.3 ( 1.0% in the samples. In addition, two markers were used in a double staining to estimate the number of premature erythroid cells in the erythrocyte preparations: CD71, a marker for reticulocytes and erythroid progenitors (1:50; PE Pharmingen), and Ter119, a marker for mature erythrocytes and erythroid precursor cells (1:20; Milteny Biotec). The percentage of CD71þ cells in the population of Ter119þ cells was determined in the population of histopaque-isolated red blood cells, which was gated in the forward scatter dot plot (see Results). All animal experiments were conducted according to the German law for the welfare of animals and were approved by local authorities. Human Erythrocytes. Leukocyte-depleted human erythrocytes from concentrates provided by the blood bank of the University of T€ubingen were used for annexin V studies. The volunteers providing erythrocytes gave informed consent. The study was approved by the Ethical commission of the University of T€ubingen.

ARTICLE

Activation of AMPK in Isolated Erythrocytes. For Western blots and mass spectrometric measurements, the isolated erythrocytes were incubated in glucose-free Ringer solution for 9 hours to activate AMPK. This treatment also resulted in the induction of phosphatidylserine exposure on erythrocytes of 38% in ampkþ/þ and of 69% in ampk-/- erythrocytes as determined in FACS analysis. Sample Preparation for LC-MS/MS Measurement

The isolated erythrocytes were dissolved in denaturation buffer consisting of 6 M urea, 2 M thiourea, 1% N-octylglucoside (w/v) in 10 mM HEPES buffer, pH 8.0. Equal amounts of proteins from wild-type (WT) and knock-out (KO) erythrocytes were separated in a NuPAGE Bis-Tris 4-12% gradient gel (Invitrogen) and stained with Coomassie Brilliant Blue (BioRad). Each lane was then cut into 10 equal pieces, always covering the same MW range for both groups. After Coomassie destaining by washing with 50% ammonium bicarbonate (ABC)/50% acetonitrile (ACN), the slices were dehydrated by incubation in 100% ACN. Disulfide bonds in proteins were reduced by incubation in 10 mM dithiotreitol (DTT) in 20 mM ABC for 45 min and cysteine residues were alkylated with 55 mM iodoacetamide (IAA) in 20 mM ABC for 30 min in the dark. Following washing in 50% ABC/50% ACN and dehydration in 100% ACN, proteins were in-gel digested using 30 μL of 12.5 ng/μL sequencing grade trypsin (Promega) in 20 mM ABC. Peptide extraction followed in three steps: with 30% ACN/3% trifluoric acid (TFA), with 80% ACN/0.5% acetic acid, and finally with 100% ACN. The ACN was evaporated from the samples using a vacuum centrifuge. Before measurement, 1 volume of 2% ACN/1% TFA was added to the peptide samples and the peptides were desalted on C18-phase stage-tips.23 For LC-MS/MS-measurement, the peptides were eluted with 80% ACN in 0.5% acetic acid, dried in a vacuum centrifuge to a volume of approximately 10 μL, and spiked with 1 μL of 2% ACN/1% TFA. LC-MS/MS Measurement

Samples were measured on an LTQ Orbitrap XL mass spectrometer (Thermo Fisher Scientific) coupled to an EasyLC nano-HPLC system (Proxeon Biosystems) through a nanoelectrospray LC-MS interface (Proxeon Biosystems). Peptides were separated on a house-made, 15 cm  75 μm nano-HPLC column emitter filled with 3 μm C18 beads (Dr. Maisch) using a 120 min segmented gradient of 80% ACN/0.5% acetic acid and ionized directly after elution. The mass spectrometer was operated in the positive ion mode; the peptide ion masses were measured in the Orbitrap mass analyzer, whereas the peptide fragmentation was performed by collision-induced dissociation (CID) in the linear ion trap analyzer using default settings. Five most intense ions were selected for fragmentation in each scan cycle; fragmented masses were excluded from further sequencing for 90 s. Mass Spectrometry Data Processing and Quantitation

MS/MS data were processed with MaxQuant version 1.1.1.624,25 (http://www.maxquant.org/) and searched using the integrated, probability-based Andromeda search engine against a decoy protein database (ipi.MOUSE.fasta v3.72) consisting of 56 957 mouse protein entries, 248 commonly observed protein contaminants and 57 205 reversed protein sequences. Reversal of all target sequences was done by MaxQuant. The initial database search mass tolerance was set to 6 ppm for the Orbitrap data (full scans), and 0.5 Da for the ion trap data (MS/ 1691

dx.doi.org/10.1021/pr101004j |J. Proteome Res. 2011, 10, 1690–1697

Journal of Proteome Research

ARTICLE

Figure 1. Proteomics workflow. Erythrocytes from three ampk-/- and three ampkþ/þ mice were purified, lysed and their proteins separated on a 1D gel. Gel lanes, containing equal amounts of ampkþ/þ or ampk-/- protein extracts, were cut into 10 equal slices and in-gel digested with trypsin. After extraction from the gel, the resulting peptide mixtures were analyzed by LC-MS/MS; precursor ion (peptide) masses and intensities were measured in the Orbitrap (MS), and fragmented in the linear ion trap analyzer (MS/MS). Peptides/proteins were identified and quantified in subsequent bioinformatic analysis.

MS scans). Full trypsin specificity was required and two missed cleavages were allowed; carbamidomethylation on cysteine was defined as fixed modification; methionine oxidation and N-terminal (protein) acetylation were defined as variable modifications in database search. The database search results were parsed by the MaxQuant software and the protein groups table with FDR 1% was assembled based on the number of reverse hits in database search results. The label-free quantitation module included in the MaxQuant software suite was used for quantitation. The reported ratios were obtained by division of protein intensities normalized by the label-free algorithm.24 Downstream Bioinformatics Analysis of Proteomics Data

To determine significantly regulated proteins, we assumed a Gaussian distribution of protein ratios. Under this assumption, the complementary error function of the normal distribution can be used to asses a measure of significance of outlier ratios. The MaxQuant implements this significance value and reports it as “Significance B”.25 A Significance B threshold of p < 0.05 was used to determine significantly regulated protein groups. Gene Ontology (GO)26 terms based on the Uniprot identifiers were retrieved using MaxQuant. Fisher’s exact test was used to test whether specific terms were significantly enriched among the set of down- and up-regulated proteins. The distribution of GO terms was visualized by pie charts using the “Plotrix” “R” package and the GO terms were mapped to a common hierarchy level in the GO graph using the “R” package GO.db. (http://www.R-project.org.). Comparison of the quantified proteins to the proteins reported by Pasini et al.27 was performed by mapping of protein identifiers from the two IPI database versions used in the studies using the IPI history file (ftp://ftp.ebi.ac.uk/pub/databases/IPI/ current/ipi.MOUSE.history.gz). In total, 584 out of 644 protein identifiers reported by Pasini et al. could be unambiguously mapped to the IPI database version used in this study.

FACS Analysis

For the in vitro experiments on eryptosis, incubations of erythrocytes were carried out at 37 C in Ringer or glucose-free Ringer solution at a hematocrit of 0.4% in a total volume of 500 μL with or without 5 μM of PAK2 inhibitor IPA3. FACS analysis was performed as described earlier.28 After incubation, erythrocytes were washed once in Ringer solution containing 5 mM CaCl2. The cells were then stained with Annexin V-Fluos Staining Kit (Roche, Mannheim, Germany) or Annexin V-FITC (Immunotools, Friesoythe, Germany) at a 1:500 dilution. After 20 min, samples were measured by flow cytometric analysis (FACS-Calibur, Becton Dickinson). The intensity of the annexin V fluorescence was measured in fluorescence channel FL-1 with an excitation wavelength of 488 nm and an emission wavelength of 530 nm. Data are expressed as arithmetic means ( SEM. For statistical analysis, ANOVA was made with p < 0.05 considered statistically significant. Western Blotting

Isolated human erythrocytes (200 μL) were lysed in 50 mL of 20 mM N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES, pH 7.4) and the ghost membranes were pelleted at 15 000 rpm for 20 min at 4 C. The supernatant was removed and the procedure repeated, until pellet and supernatant lost their red color. The pellet was finally resolved in 200 μL of lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Triton X-100, 0.5% SDS, 1 mM NaF, 1 mM Na3VO4, 0.4% β-mercaptoethanol, protease inhibitor cocktail (Sigma)) and stored at -80 C. The murine erythrocytes were washed after isolation from full blood by the histopaque procedure and then lysed in the same lysis buffer as human erythrocytes. Lysate of HEK293 cells was prepared as a positive control for PAK1 in the same lysis buffer. 1692

dx.doi.org/10.1021/pr101004j |J. Proteome Res. 2011, 10, 1690–1697

Journal of Proteome Research

ARTICLE

Figure 2. Histogram of ratios measured in the proteomics experiment. The distribution of 812 measured protein ratios showed a shift toward higher ratios (a general “upregulation” in the ampk-/- erythrocytes). Gene Ontology (GO) terms retrieved for the 22 significantly downregulated and 37 significantly upregulated proteins are shown in insets.

For each lane, 80 μg of the protein was mixed with loading buffer (Roti-Load 1, Roth), boiled at 95 C for 5 min, and resolved by 10% SDS-PAGE. For immunoblotting, proteins were electrotransfered onto a nylon or PVDF membrane and blocked with 5% nonfat milk in TBS-0.1% Tween 20 (TBST) at room temperature for 1 h. The membrane was incubated with affinitypurified rabbit anti-PAK2 antibody (1:500; 62 kDa, Cell Signaling) or rabbit anti-PAK1 antibody (1:500; 68 kDa, Cell Signaling) at 4 C overnight. After washing with TBST, the blots were incubated with secondary anti-rabbit antibody (1:2000; Cell Signaling) for 1 h at room temperature. After washing, antibody binding was detected with the ECL detection reagent (Amersham). In the same way, GAPDH (1:1000; 37 kDa, Cell Signaling) was detected to serve as loading control.

’ RESULTS In this study, we investigated the AMPKR1-sensitive proteins which are potentially relevant for erythrocyte survival. To perform a proteome-wide analysis of erythrocytes undergoing apoptosis in the absence of AMPKR1 and to uncover novel

proteins regulating cell survival during energy depletion, we applied mass spectrometry-based proteomics to measure the overall effect of AMPKR1 on protein levels in the murine red blood cells. Erythrocytes from three WT and three ampk-/animals were lysed and their proteins separated on a 1D SDSPAGE gel. The gel lanes were each cut into 10 equal slices and ingel digested with trypsin. The resulting peptide mixtures were extracted from the gel and analyzed in triplicates by nano-LCMS/MS on an LTQ Orbitrap mass spectrometer (Figure 1). Proteins were identified by searching the MS/MS spectra against a mouse protein database using Andromeda search engine. Relative changes in protein abundance between the WT and ampk-/- samples were determined using a label-free quantitation algorithm integrated in the MaxQuant processing software. To verify the purity of the studied cell population, the percentage of Ter119-positive (erythroid) cells was determined by FACS analysis. As a result, 97.9 ( 1.0% (n = 4) of the studied cells from ampk-/- mice and 94.8 ( 2.3% (n = 4) from ampkþ/þ mice were Ter119-positive and thus of erythroid origin. Moreover, 7.92 ( 0.82% (n = 4) of the erythroid cells from ampkþ/þ mice and 33.74 ( 6.16% (n = 4) from ampk-/- mice were 1693

dx.doi.org/10.1021/pr101004j |J. Proteome Res. 2011, 10, 1690–1697

Journal of Proteome Research

ARTICLE

Figure 4. Phosphatidylserine exposure of human erythrocytes incubated with or without glucose depletion in the presence and absence of IPA3. Arithmetic means ( SEM (n = 6) of the percentage of annexin V-binding erythrocytes exposed for 48 h to glucose-containing or glucose-depleted Ringer in the absence (white bars) or presence (black bars) of the PAK2 inhibitor IPA3 (5 μM). ***(p < 0.001), significant difference from presence of glucose; ###(p < 0.01), significant difference from absence of IPA3. Figure 3. Expression of p21-activated kinase in human and murine erythrocytes. (A) Western blot confirmation of casein kinase 2 (CK2) expression in erythrocytes from ampkþ/þ and ampk-/- mice. Equal volumes of erythrocytes from three ampkþ/þ and three ampk-/mice each were used. Erythrocytes were isolated and treated as described in 3B. (B) Western blot confirmation of p21-activated kinase 2 (PAK2) expression in membrane preparations of human erythrocytes (RBC) and in membrane preparation of whole blood (left panel), as well as in erythrocytes and whole blood from ampkþ/þ (WT) and ampk-/- (KO) mice (right panel). Equal volume of blood from six ampkþ/þ and six ampk-/- mice each was pooled. Erythrocytes were isolated by the histopaque procedure and incubated in glucose-depleted Ringer solution to activate AMPK and to induce eryptosis. Subsequently they were pelleted, washed and lysed. (C) Western blot demonstrates a lack of PAK1 expression in erythrocytes from ampkþ/þ (WT) and from ampk-/- mice (KO). A lysate of HEK293 cells served as a positive control (HEK).

CD71-positive, indicative of an increased number of erythroid progenitor cells in ampk-/- mice. Representative microphotographs of erythrocytes from ampkþ/þ and ampk-/- animals are shown in Supplementary Figure 1. Proteome Analysis of the ampk-/- Erythrocytes

Our approach led to the identification of 13 594 peptides that mapped to 1390 protein groups from both experiments at a false discovery rate of 1.8%; 1127 proteins were identified in the WT and 1344 in the ampk-/- sample. The average absolute mass measurement deviation at the precursor ion level was 0.3 ppm and 99% of the precursor ion masses were measured with a deviation of 1.48 ppm or better. Among the detected proteins, 28 were known contaminants and 812 were present in both data sets and fulfilled the requirements of having at least two quantitation events in each experiment. The list of identified and quantified proteins is presented in the Supplementary Table 1. Owing to the increased number of the erythroid progenitor cells in ampk-/- mice, we compared our data set with the data set of Pasini et al.,27 considered to be the most accurate murine

erythrocyte proteome to date. At least 352 proteins (60.3%) reported in Pasini et al. were quantified in our data set, representing the subset of proteins that is highly specific to mature erythrocytes (Supplementary Figure 2 and Supplementary Table 3). Additional 40 proteins reported by Pasini et al. could not be mapped due to ambiguities in the IPI identifiers between the versions of the IPI database used in the studies. The distribution of measured ratios in our data set showed a shift toward higher ratios (upregulation in ampk-/- samples) (Figure 2). Of all quantified proteins, 37 were significantly upand 22 were significantly downregulated (p < 0.05). To gain insights into the functions of regulated proteins, we performed a bioinformatic functional enrichment analysis of GO terms of all significantly regulated proteins (Supplementary Table 2). Interestingly, the upregulated set of proteins showed an enrichment in functions related to protein synthesis and its regulation; indeed, most of the detected upregulated proteins were ribosomal subunits and elongation factors, that are known to be present but not functional in erythrocytes.18 The reason for their apparent upregulation in erythrocytes of ampk-/- mice is at present not clear, but downregulation of the proteasomal subunits (see below) may reflect the slower clearance of ribosomal proteins from the cytosol and hence their higher abundance in the ampk-/- erythrocytes. One of the upregulated proteins of potential functional importance in ampk-/erythrocytes was Casein kinase 2 (CK2). The upregulation of CK2 in ampk-/- erythrocytes was confirmed by Western blotting (Figure 3A). More biological insight provided the enrichment analysis of downregulated proteins, whose membrane localization and functions related to protein kinase activity and integrin binding were significantly enriched. A closer look at the downregulated set of proteins revealed that proteasomal subunits, as well as the integrin and PI3 kinase were downregulated by >4-fold in ampk-/erythrocytes. Evidence for functional expression of integrin in erythrocytes has been provided earlier.29 In nucleated cells, integrins are multifuntional proteins fostering cell survival.30 The PI3 kinase 1694

dx.doi.org/10.1021/pr101004j |J. Proteome Res. 2011, 10, 1690–1697

Journal of Proteome Research

ARTICLE

Figure 5. Phosphatidylserine exposure of erythrocytes from AMPKR1-deficient and wild-type mice incubated with or without glucose depletion in the presence and absence of IPA3. Arithmetic means ( SEM (n = 9) of the percentage of annexin V-binding erythrocytes from ampk-/- mice (KO) and ampkþ/þ mice (WT) exposed for 9 h to either glucose containing or glucose-depleted Ringer in the absence (white bars) or presence (black bars) of the PAK2 inhibitor IPA3 (5 μM). Significant differences from the presence of glucose was always given independently of the presence or absence of Ipa3 and is not further shown in the plot. #(p < 0.05), significant difference from the absence of IPA3; ***(p < 0.001), significant difference between ampk-/- and ampkþ/þ erythrocytes.

has earlier been shown to confer survival of erythroid precursor cells.31 The downregulation of those proteins is thus in agreement with the suicidal phenotype of ampk-/- erythrocytes.16 The PAK2 kinase, previously not connected to eryptosis, was also measured as downregulated albeit not to the same extent. To validate the accuracy of quantitation and to investigate the potential role of PAK2 in eryptosis, we focused on this kinase to follow up our proteomics results. Expression levels of PAK2 were therefore further assessed by Western blotting. As shown in Figure 3B, PAK2 expression was confirmed in both, human and murine erythrocytes and, in agreement with MS results, ampkþ/þ mouse erythrocytes expressed higher levels of PAK2 than ampk-/- erythrocytes. Effect of PAK Inhibitor IPA3 on Cell Membrane Scrambling

Energy depletion provoked by glucose deprivation is known to trigger scrambling of the erythrocyte cell membrane resulting in exposure of phosphatidylserine that can be measured by annexin V-binding to the erythrocyte surface. To test whether inactivation of PAK2 has an effect on membrane scrambling in ampkþ/þ erythrocytes, we measured annexin V-binding in the presence of the PAK2 inhibitor IPA3 and/or glucose. Since IPA3 can also inhibit the PAK1 kinase, we tested the PAK1 expression in erythrocytes by Western blotting. PAK1 expression could not be detected in erythrocytes from ampk-/- or from ampkþ/þ mice (Figure 3C), neither could it be detected by MS. As expected, the annexin V-binding in human erythrocytes was significantly increased following glucose removal (Figure 4).

However, the binding did not change significantly after the addition of the PAK2 inhibitor IPA3 in the presence of glucose. Rather, IPA3 led to a significant increase in annexin V-binding only in glucose-depleted human erythrocytes. These results demonstrate that the inhibition of PAK2 augments the eryptotic effects of glucose deprivation in human erythrocytes. Similar results were obtained with murine erythrocytes: the PAK2 inhibitor IPA3 had no significant effect on annexin V-binding in normal, glucose-containing Ringer solution (Figure 5). Following glucose removal, annexin V-binding was significantly increased in both, ampk-/- and ampkþ/þ erythrocytes, and again to a higher extent in ampk-/- erythrocytes. Importantly, addition of IPA3 led to a significant increase in annexin V-binding only in ampkþ/þ erythrocytes under glucose deprivation and failed to significantly increase the annexin V-binding in ampk-/- erythrocytes under the same conditions. Taken together, these results demonstrate that PAK2 inhibition has a similar effect on annexin V-binding as AMPKR1 and is therefore involved in the regulation of eryptosis.

’ DISCUSSION Eryptosis, or erythrocyte programmed cell death, is the process in which the phosphatidylserine asymmetry of the erythrocyte plasma membrane breaks down, leading to the exposure of phosphatidylserine at the erythrocyte surface.11 This can be caused by osmotic shock, oxidative stress, or energy 1695

dx.doi.org/10.1021/pr101004j |J. Proteome Res. 2011, 10, 1690–1697

Journal of Proteome Research depletion.32 Conversely, eryptosis can be inhibited by erythropoietin, nitric oxide, catecholamines and high concentrations of urea.33 Eryptosis is increased in different pathological conditions, such as sepsis, malaria, sickle cell anemia, and iron deficiency,17 and excessive eryptosis is also observed in erythrocytes lacking the AMP-activated protein kinase AMPK.16 Unlike most other cells, erythrocytes express exclusively AMPKR1, whose absence cannot be compensated by AMPKR2.16 Accordingly, AMPKR1 deficiency is followed by severe anemia due to excessive eryptosis.16 To gain insight into the molecular processes taking place in erythrocytes undergoing eryptosis, we performed a global proteome analysis of erythrocytes from ampk-/- mice and from wild-type mice using high-accuracy mass spectrometry and label-free quantitation. Here we report, for the first time, the proteome of ampk-/- murine erythrocytes undergoing eryptosis and compare relative abundance of more than 800 proteins between erythrocytes of ampk-/- and WT mice. This reported erythrocyte proteome size is somewhat larger than in previous studies27 and may reflect the fact that the blood of ampk-/- mice contains a higher proportion (33%) of reticulocytes that cannot be completely removed. As documented by >95% of Ter119-positive cells in our samples, it is evident that a vast majority of cells analyzed in our study were of erythroid origin in both analyzed genotypes. Since the high percentage of reticulocytes among the peripheral blood erythrocytes is the typical phenotype of the AMPK KO mouse16 and since our study aimed at the analysis of the peripheral blood RBCs predominant in the respective genotype, we did not focus on a particular CD71-negative (mature erythrocyte) or CD71-positive (reticulocyte) population. The bioinformatic analysis of the subset of regulated proteins showed characteristic features of apoptosis—especially downregulation of integrin, PI3 kinase and the proteasome—confirming the molecular phenotype of apoptosis. In addition, serotransferrin was found to be highly upregulated in the erythrocytes from ampk-/mice. This carrier protein, also known as transferrin, is known to recirculate the iron which is released into the plasma following eryptosis.34 Another example of proteins downregulated in erythrocytes of ampk-/- mice is p21-activated kinase 2 (PAK2) that was chosen for further investigation. PAK2 is a member of the p21-activated serine-threonine kinase family. It can be activated by binding of small G proteins, which is then followed by autophosphorylation of its kinase domain. Activated full-length PAK2 has been reported to foster cell survival, enhance cell growth and exert antiapoptotic effects.35 Despite this, the possibility that PAK2 modulates the programmed death of erythrocytes has not been examined so far.36,37 The full-length PAK2 can alternatively be cleaved by caspase-3 in cells undergoing apoptosis and cells exposed to stress, such as hyperosmolarity or serum deprivation, which results in fragmentation into PAK-2p34 and PAK-2p27. Both fragments become highly autophosphorylated,35 of which the former is constitutively active and has proapoptotic effects.35 We showed that PAK2 is expressed in both, human and murine erythrocytes and whole blood. The mass spectrometric analysis measured PAK2 downregulation in the erythrocytes from ampk-/- mice, a finding which was additionally confirmed by Western blotting (Figure 3B). The measurement of annexin V-binding in human red blood cells showed a significant increase upon glucose deprivation, both in the presence and in the absence of PAK2 inhibitor, IPA3. However, the increase in binding was higher when IPA3 was present, emphasizing the effect of PAK2 in eryptosis.

ARTICLE

Additional measurements on erythrocytes from ampk-/- and ampkþ/þ mice showed the same trend in annexin V-binding caused by glucose deprivation, as in human erythrocytes. Addition of IPA3 abolished the difference between ampk-/- and ampkþ/þ erythrocytes by increasing the annexin V-binding in ampkþ/þ erythrocytes nearly as much as in ampk-/- erythrocytes by glucose withdrawal itself. Accordingly, addition of IPA3 had nearly no impact on annexin V-binding in ampk-/- erythrocytes. The inability of IPA3 to additionally increase the number of annexin V-binding cells under glucose deprivation in ampk-/erythrocytes confirms our findings and demonstrates a lower expression level of PAK2 in the erythrocytes from ampk-/mice. AMPK deficiency can lead to a decreased amount of PAK2 in erythrocytes only at the developmental stage of reticulocytes, since erythrocytes lack genetic material. Since a difference in annexin V-binding upon addition of IPA3 is measurable only when eryptosis is already induced and addition of IPA3 leads to a further increase in eryptosis, IPA3 must inhibit a protein that has an antieryptotic effect and is only activated upon some survival-threatening conditions, in this case energy depletion. As IPA3 in our study specifically inhibits the PAK2, these results demonstrate an antieryptotic effect of PAK2. Taken together, our findings lead to the conclusion that AMPKR1-deficiency leads to a decrease of PAK2 in erythrocytes, which makes these erythrocytes more vulnerable toward programmed cell death under energy depletion. The impact of the PAK2 inhibition on erythrocyte survival further implies that PAK2 could be an important factor through which AMPK deficiency leads to anemia onset already a couple of weeks after birth. It will be interesting to investigate the exact mode of action of AMPKR1 on PAK2.

’ ASSOCIATED CONTENT

bS

Supporting Information Supplementary Figure 1: Microphotographs of the erythrocytes from ampkþ/þ and ampk-/- mice. Supplementary Figure 2: Comparison of the data set reported in this study with Pasini et al.27 Supplementary Table 1: Lists of proteins and peptides detected and quantified in this study. Supplementary Table 2: Lists of up- and downregulated proteins and bioinformatic enrichment analysis of GO terms among significantly regulated (p < 0.05) proteins. Supplementary Table 3: Lists of proteins detected in this study and in Pasini et al.27 This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Prof. Dr. Boris Macek, Proteome Center Tuebingen, University of Tuebingen, Auf der Morgenstelle 15, D-72076 Tuebingen, Germany, Phone: þ49/(0)7071/29-70556, Fax: þ49/(0)7071/ 29-5779. E-mail: [email protected]. Prof. Dr. Florian Lang, Department of Physiology, University of T€ubingen, Gmelinstr. 5, D-72076 T€ubingen, Germany. Tel: þ49/(0)7071/29-72194. Fax: þ49/(0)7071/29-5618. E-mail: fl[email protected].

’ ACKNOWLEDGMENT This study was supported by the Carl-Zeiss-Stiftung, the Deutsche Forschungsgemeinschaft, GRK 1302, SFB 773, La 315/13-3 (to F.L.) and a Juniorprofessorenprogramm grant of the Landesstiftung BW (to B.M.). 1696

dx.doi.org/10.1021/pr101004j |J. Proteome Res. 2011, 10, 1690–1697

Journal of Proteome Research

’ ABBREVIATIONS: AMPK, 50 AMP-activated protein kinase; PAK2, p21-activated kinase 2; ppm, parts per million ’ REFERENCES (1) Towler, M. C.; Hardie, D. G. AMP-activated protein kinase in metabolic control and insulin signaling. Circ. Res. 2007, 100 (3), 328–41. (2) Winder, W. W.; Thomson, D. M. Cellular energy sensing and signaling by AMP-activated protein kinase. Cell Biochem. Biophys. 2007, 47 (3), 332–47. (3) Carling, D. The role of the AMP-activated protein kinase in the regulation of energy homeostasis. Novartis Found Symp 2007, 286, 72– 81; discussion 81-5, 162-3, 196-203. (4) Horie, T.; Ono, K.; Nagao, K.; Nishi, H.; Kinoshita, M.; Kawamura, T.; Wada, H.; Shimatsu, A.; Kita, T.; Hasegawa, K. Oxidative stress induces GLUT4 translocation by activation of PI3-K/Akt and dual AMPK kinase in cardiac myocytes. J. Cell. Physiol. 2008, 215 (3), 733–42. (5) Jensen, T. E.; Rose, A. J.; Hellsten, Y.; Wojtaszewski, J. F.; Richter, E. A. Caffeine-induced Ca(2þ) release increases AMPKdependent glucose uptake in rodent soleus muscle. Am. J. Physiol.: Endocrinol. Metab. 2007, 293 (1), E286–92. (6) McGee, S. L.; Hargreaves, M. AMPK and transcriptional regulation. Front. Biosci. 2008, 13, 3022–33. (7) Hardie, D. G. The AMP-activated protein kinase pathway--new players upstream and downstream. J. Cell Sci. 2004, 117 (Pt. 23), 5479–87. (8) Lang, F.; Gulbins, E.; Lerche, H.; Huber, S. M.; Kempe, D. S.; Foller, M. Eryptosis, a window to systemic disease. Cell. Physiol. Biochem. 2008, 22 (5-6), 373–80. (9) Berg, C. P.; Engels, I. H.; Rothbart, A.; Lauber, K.; Renz, A.; Schlosser, S. F.; Schulze-Osthoff, K.; Wesselborg, S. Human mature red blood cells express caspase-3 and caspase-8, but are devoid of mitochondrial regulators of apoptosis. Cell Death Differ. 2001, 8 (12), 1197–206. (10) Bratosin, D.; Estaquier, J.; Petit, F.; Arnoult, D.; Quatannens, B.; Tissier, J. P.; Slomianny, C.; Sartiaux, C.; Alonso, C.; Huart, J. J.; Montreuil, J.; Ameisen, J. C. Programmed cell death in mature erythrocytes: a model for investigating death effector pathways operating in the absence of mitochondria. Cell Death Differ. 2001, 8 (12), 1143–56. (11) Lang, P. A.; Kaiser, S.; Myssina, S.; Wieder, T.; Lang, F.; Huber, S. M. Role of Ca2þ-activated Kþ channels in human erythrocyte apoptosis. Am. J. Physiol.: Cell Physiol. 2003, 285 (6), C1553–60. (12) Duranton, C.; Huber, S. M.; Lang, F. Oxidation induces a Cl(-)dependent cation conductance in human red blood cells. J Physiol 2002, 539 (Pt 3), 847–55. (13) Foller, M.; Kasinathan, R. S.; Koka, S.; Lang, C.; Shumilina, E.; Birnbaumer, L.; Lang, F.; Huber, S. M. TRPC6 contributes to the Ca(2þ) leak of human erythrocytes. Cell. Physiol. Biochem. 2008, 21 (1-3), 183–92. (14) Bookchin, R. M.; Ortiz, O. E.; Lew, V. L. Activation of calciumdependent potassium channels in deoxygenated sickled red cells. Prog. Clin. Biol. Res. 1987, 240, 193–200. (15) Brugnara, C.; de Franceschi, L. Effect of cell age and phenylhydrazine on the cation transport properties of rabbit erythrocytes. J. Cell. Physiol. 1993, 154 (2), 271–80. (16) Foller, M.; Sopjani, M.; Koka, S.; Gu, S.; Mahmud, H.; Wang, K.; Floride, E.; Schleicher, E.; Schulz, E.; Munzel, T.; Lang, F. Regulation of erythrocyte survival by AMP-activated protein kinase. FASEB J. 2009, 23 (4), 1072–80. (17) Foller, M.; Bobbala, D.; Koka, S.; Huber, S. M.; Gulbins, E.; Lang, F. Suicide for survival--death of infected erythrocytes as a host mechanism to survive malaria. Cell. Physiol. Biochem. 2009, 24 (3-4), 133–40. (18) D’Alessandro, A.; Righetti, P. G.; Zolla, L. The red blood cell proteome and interactome: an update. J. Proteome Res. 2010, 9 (1), 144–63. (19) Pasini, E. M.; Lutz, H. U.; Mann, M.; Thomas, A. W. Red blood cell (RBC) membrane proteomics—Part II: Comparative proteomics and RBC patho-physiology. J. Proteomics 2010, 73 (3), 421–35.

ARTICLE

(20) Pasini, E. M.; Lutz, H. U.; Mann, M.; Thomas, A. W. Red blood cell (RBC) membrane proteomics—Part I: Proteomics and RBC physiology. J. Proteomics 2010, 73 (3), 403–20. (21) Pasini, E. M.; Kirkegaard, M.; Mortensen, P.; Lutz, H. U.; Thomas, A. W.; Mann, M. In-depth analysis of the membrane and cytosolic proteome of red blood cells. Blood 2006, 108 (3), 791–801. (22) Jørgensen, S. B.; Viollet, B.; Andreelli, F.; Frøsig, C.; Birk, J. B.; Schjerling, P.; Vaulont, S.; Richter, E. A.; Wojtaszewski, J. F. Knockout of the alpha2 but not alpha1 5'-AMP-activated protein kinase isoform abolishes 5-aminoimidazole-4-carboxamide-1-beta-4-ribofuranosidebut not contraction-induced glucose uptake in skeletal muscle. J Biol. Chem. 2004, 279 (2), 1070–9. (23) Rappsilber, J.; Mann, M.; Ishihama, Y. Protocol for micropurification, enrichment, pre-fractionation and storage of peptides for proteomics using StageTips. Nat. Protoc. 2007, 2 (8), 1896–906. (24) Luber, C. A.; Cox, J.; Lauterbach, H.; Fancke, B.; Selbach, M.; Tschopp, J.; Akira, S.; Wiegand, M.; Hochrein, H.; O’Keeffe, M.; Mann, M. Quantitative proteomics reveals subset-specific viral recognition in dendritic cells. Immunity 2010, 32 (2), 279–89. (25) Cox, J.; Mann, M. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteomewide protein quantification. Nat. Biotechnol. 2008, 26 (12), 1367–72. (26) Ashburner, M.; Ball, C. A.; Blake, J. A.; Botstein, D.; Butler, H.; Cherry, J. M.; Davis, A. P.; Dolinski, K.; Dwight, S. S.; Eppig, J. T.; Harris, M. A.; Hill, D. P.; Issel-Tarver, L.; Kasarskis, A.; Lewis, S.; Matese, J. C.; Richardson, J. E.; Ringwald, M.; Rubin, G. M.; Sherlock, G. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet. 2000, 25 (1), 25–9. (27) Pasini, E. M.; Kirkegaard, M.; Salerno, D.; Mortensen, P.; Mann, M.; Thomas, A. W. Deep coverage mouse red blood cell proteome: a first comparison with the human red blood cell. Mol. Cell. Proteomics 2008, 7 (7), 1317–30. (28) Lang, P. A.; Kasinathan, R. S.; Brand, V. B.; Duranton, C.; Lang, C.; Koka, S.; Shumilina, E.; Kempe, D. S.; Tanneur, V.; Akel, A.; Lang, K. S.; Foller, M.; Kun, J.F.; Kremsner, P. G.; Wesselborg, S.; Lauffer, S.; Clemen, C. S.; Herr, C.; Noegel, A. A.; Wieder, T.; Gubins, E.; Lang, F.; Huber, S. M. Accelerated clearance of Plasmodium-infected erythrocytes in sickle cell trait and annexin-A7 deficiency. Cell Physiol. Biochem. 2009, 24 (5-6), 415–28. (29) Carvalho, F. A.; Connell, S.; Miltenberger-Miltenyi, G.; Pereira, S. V.; Tavares, A.; Ariens, R. A.; Santos, N. C. Atomic force microscopybased molecular recognition of a fibrinogen receptor on human erythrocytes. ACS Nano 2010, 4 (8), 4609–20. (30) Streuli, C. H. Integrins and cell-fate determination. J. Cell Sci. 2009, 122 (Pt. 2), 171–7. (31) Laubach, J. P.; Fu, P.; Jiang, X.; Salter, K. H.; Potti, A.; Arcasoy, M. O. Polycythemia vera erythroid precursors exhibit increased proliferation and apoptosis resistance associated with abnormal RAS and PI3K pathway activation. Exp. Hematol. 2009, 37 (12), 1411–22. (32) Lang, F.; Gulbins, E.; Lang, P. A.; Zappulla, D.; Foller, M., Ceramide in suicidal death of erythrocytes. Cell. Physiol. Biochem. 2010, 26, (1), 21-8. (33) Foller, M.; Huber, S. M.; Lang, F. Erythrocyte programmed cell death. IUBMB Life 2008, 60 (10), 661–8. (34) Franco, R. S.; Palascak, M.; Thompson, H.; Rucknagel, D. L.; Joiner, C. H. Dehydration of transferrin receptor-positive sickle reticulocytes during continuous or cyclic deoxygenation: role of KCl cotransport and extracellular calcium. Blood 1996, 88 (11), 4359–65. (35) Molli, P. R.; Li, D. Q.; Murray, B. W.; Rayala, S. K.; Kumar, R. PAK signaling in oncogenesis. Oncogene 2009, 28 (28), 2545–55. (36) Jakobi, R.; Moertl, E.; Koeppel, M. A. p21-activated protein kinase gamma-PAK suppresses programmed cell death of BALB3T3 fibroblasts. J. Biol. Chem. 2001, 276 (20), 16624–34. (37) Bokoch, G. M. Biology of the p21-activated kinases. Annu. Rev. Biochem. 2003, 72, 743–81.

1697

dx.doi.org/10.1021/pr101004j |J. Proteome Res. 2011, 10, 1690–1697