Protein Profiling of Active Cysteine Cathepsins in Living Cells Using

Oct 19, 2012 - Fengkai Fan†‡, Si Nie†‡, Eric B. Dammer∥, Duc M. Duong∥, Deng ... Britton Chance Center for Biomedical Photonics, Wuhan Nat...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/jpr

Protein Profiling of Active Cysteine Cathepsins in Living Cells Using an Activity-Based Probe Containing a Cell-Penetrating Peptide Fengkai Fan,†,‡ Si Nie,†,‡ Eric B. Dammer,∥ Duc M. Duong,∥ Deng Pan,†,‡ Lingyan Ping,⊥,§ Linhui Zhai,¶,§ Junzhu Wu,⊥ Xuechuan Hong,¶ Lingsong Qin,†,‡ Ping Xu,*,§,¶,○ and Yu-Hui Zhang*,†,‡,○ †

Britton Chance Center for Biomedical Photonics, Wuhan National Laboratory for Optoelectronics-Huazhong University of Science and Technology (HUST), Wuhan, China ‡ Key Laboratory of Biomedical Photonics of Ministry of Education, Department of Biomedical Engineering, Huazhong University of Science and Technology, Wuhan, China § State Key Laboratory of Proteomics, Beijing Proteome Research Center, Beijing Institute of Radiation Medicine, Beijing, China ∥ Department of Human Genetics, and Center for Neurodegenerative Disease, Emory University School of Medicine, Atlanta, United States ⊥ Department of Biochemistry, School of Medicine, Wuhan University, Wuhan, China ¶ Key Laboratory of Combinatorial Biosynthesis and Drug Discovery, Wuhan University, Ministry of Education, and Wuhan University School of Pharmaceutical Sciences, Wuhan, China S Supporting Information *

ABSTRACT: Cell-permeable activity-based probes (ABPs) are capable of labeling target proteins in living cells, thereby providing a powerful tool for profiling active enzymes in their native environment. In this study, we describe the synthesis and use of a novel trifunctional cell-permeable activity-based probe (TCpABP) for proteomic profiling of active cysteine cathepsins in living cells. We demonstrate that although TCpABP contains cell-impermeable tags, it was able to enter living cells efficiently via the delivery of a cell-penetrating peptide. TCpABP also allowed simultaneous detection and affinity isolation of labeled proteins with a fluorophore and a biotin motif, respectively. We optimized the enrichment protocol to minimize contaminants and identified 7 cathepsins, 2 of which have never been identified using existing ABPs. We also used a label-free quantification approach to quantify the relative abundances of active cathepsins and compared them with their previously published mRNA expression levels. A high degree of correlation between the mRNA expression levels and protein relative activities was observed for most of the identified cathepsins except cathepsin H. The results herein indicate that TCpABP is valuable for the detection of active cathepsins in living cells and provides useful guidelines for designing novel cell-permeable ABPs for in vivo labeling and their applications in in vivo proteomics studies. KEYWORDS: activity-based probe, cell-penetrating peptide, cathepsin, proteomics, living cell

■

INTRODUCTION

In general, ABPs contain two key elements: a reactive “warhead” that covalently reacts with the enzyme’s active site and a reporter tag for purification, identification, and/or visualization of the covalently bound target.4 The most frequently used tags include biotin and various fluorophores. The choice of reporter tag influences the scope of ABP experiments. A biotin motif permits avidin enrichment and molecular identification by mass spectrometry. Fluorescent tags allow for rapid and sensitive target detection by in-gel fluorescence scanning after gel separation, offering a robust and relatively high-throughput platform for ABP experiments.5 However, biotin motifs and most fluorophores are not cellpermeable, thereby limiting most ABPs to in vitro work with

With the completion of the human genome sequencing project, interest in a deeper understanding of protein function has risen.1 Traditional genomic and proteomic methods are mainly based on protein abundance, thus providing only limited information about the functional roles of enzymes because the activities of enzymes are often regulated on a post-translational level, leading to a potentially significant divergence of abundance and activity.2 To address this limitation, various small-molecule activity-based probes (ABPs) have been developed. As a new means for functional proteomics, ABPs can be used to profile enzyme activities in whole proteomes and have been used widely for the study of enzymes implicated in cancer progression and tumorigenesis, such as metalloproteases as well as serine and cysteine proteases.3 © 2012 American Chemical Society

Received: June 26, 2012 Published: October 19, 2012 5763

dx.doi.org/10.1021/pr300575u | J. Proteome Res. 2012, 11, 5763−5772

Journal of Proteome Research

Article

cell lysates.6 In vitro proteomic preparations are usually not capable of maintaining complex intracellular conditions and therefore do not provide accurate information about enzyme activities.7 To address this limitation, several cell-permeable ABPs for intact cell labeling have been developed; the strategies for this purpose can be divided into four main types. One strategy exploits the cell-penetrating ability of several hydrophobic fluorescent tags8 [e.g., boron-dipyrromethene (BODIPY) and cyanine dyes]. However, a long incubation time (approximately 12 h) and pretreatment with a high concentration of dimethyl sulfoxide (DMSO) were required for the efficient cellular uptake of this type of ABP.9 Moreover, membrane-impermeable biotin motifs could not be used in this strategy, which prevents this type of ABP from providing a straightforward means to enrich and identify probe-labeled proteins. Additionally, the commercially available activated ester forms of these cell-permeable fluorophores (e.g., BODIPY and cyanine dyes) are extremely expensive, which would preclude large-scale production of the ABP.8 The second strategy relies on the bio-orthogonal “click” chemistry in which ABPs are divided into two parts and a two-step labeling procedure is employed.10 However, the chemical coupling yields of the “click” reaction fluctuated in different systems,5 and the “click” reaction usually also suffered from high background, reducing the sensitivity of the ABPs and hindering the detection of endogenous low abundance enzymes.7 The third strategy is based on receptor-mediated endocytosis. ABPs modified with a mannose cluster are examples of probes utilizing this strategy for cellular entry;11 they were capable of entering live cells via endocytosis, but their uptake was restricted to several specialized cell types, such as dendritic cells and macrophage cells. The fourth strategy utilizes cellpenetrating peptides (CPPs). CPPs are short peptides that can deliver a range of membrane-impermeable molecules into living cells or tissues via endocytosis or direct membrane translocation.12−15 By covalently conjugating ABPs with CPPs, ABPs containing cell-impermeable motifs (e.g., biotin or fluorophores) have successfully been delivered into various living cells or tissues, thus expanding the range of reporter tags to include otherwise cell-impermeable motifs for intact cell labeling.14−20 Herein, we describe the synthesis of a novel ABP with membrane permeability enabled by a CPP, where the ABP contains both a biotin group and a fluorophore. The probe was successfully used for protein profiling of active cysteine cathepsins in living cells. We chose cysteine cathepsins as the target enzymes for several reasons. First, cysteine cathepsins have been shown to play important roles in tumor growth, angiogenesis, and invasion21 and are emerging as promising targets for diagnosis and therapy. Second, the activities of these proteases are often tightly regulated post-translationally.22

■

Synthesis and Purification of Probes

The activated nitrophenyl ester (compound 3 in Supplementary Figure S1) was synthesized as previously reported.23 The synthesis of probes was carried out using solid-phase peptide synthesis (SPPS) on 2-chlorotrityl-chloride resin.24 For loading of the resin with the first amino acid, Fmoc-Lys(Alloc)-OH (2 equiv, Alloc = allyloxycarbonyl) and N-ethyl diisopropylamine (DIEA, 9 equiv) were dissolved in N,N-dimethylformamide (DMF, anhydrous), and the reaction mixture was added to the resin. The reaction mixture was shaken at room temperature for 3 h. Next, the Fmoc-protecting group of Fmoc-Lys(Alloc)-OH was removed by treatment with 20% piperidine in DMF for 15 min. Couplings of each amino acid (3 equiv) were carried out with N,N′-diisopropylcarbodiimide (DIC; 3 equiv) and 1hydroxy-benzotriazole (HOBt; 3 equiv) in DMF for 2 h at room temperature. Coupling efficiencies were monitored with the Kaiser test. A 5(6)-carboxyfluorescein (FAM) fluorophore was linked via an amide linkage (6-aminohexanoic acid) to the N-terminus of a (ε)-biotin-lysine residue. Then, the Allocprotecting group of the first amino acid was removed with Pd(PPh3)4 and PhSiH3, followed by the coupling of Fmoc-TyrOH and Fmoc-Leu-OH. After the final Fmoc-deprotection, the activated nitrophenyl ester (3 equiv) in DMF was added, and the mixture was allowed to react for 1 h. The DMF was then drained off. The resin was washed 5 times with DMF. To cleave products from the resin, a cleavage cocktail solution containing triisopropylsilane (TIS, 2%) and trifluoroacetic acid (TFA, anhydrous, 98%) was added to the resin. After standing for 2 h, the cleavage mixture was collected. Crude products were washed with dry cold diethyl ether 3 times, neutralized to pH 7 with several drops of ethanediamine, and precipitated with dry cold diethyl ether. Products were purified by preparative high performance liquid chromatography (HPLC). The purification was performed using solvent A (0.05% acetic acid in H2O) and solvent B (0.05% acetic acid in acetonitrile) with a gradient of 5% solvent B to 25% within 20 min. Probe (purity greater than 90%) masses were confirmed by electrospray ionization (ESI) mass spectrometry (Supplementary Figure S2). In Vitro Competitive Labeling Study

RAW264.7 cells (mouse leukemic monocyte macrophage cell line) were seeded onto 100-mm culture dishes (Corning) at a density of 1.0 × 107 cells/well in DMEM (8 mL) containing 10% FBS and grown at 37 °C under CO2 (5%, v/v) overnight. The cells were harvested by trypsin digestion. They were then washed with cold phosphate-buffered saline (PBS, pH 7.4), and the cell pellets were lysed with 50 mM citrate phosphate buffer (pH 4.5) containing 1% 3-[(3-cholamidopropyl) -dimethylammonio]-1-propane sulfonate (CHAPS), 0.5% Triton X-100, and 5 mM 1,4-dithiothreitol (DTT) for 30 min on ice.25 Subsequently, the cell lysates were centrifuged at 17,000g for 15 min at 4 °C to remove cellular debris. The supernatants were diluted in the lysis buffer to reach a final concentration of 2 μg/μL and were aliquoted and stored at −80 °C. The supernatants (8 μL, 2 μg/μL) were treated with the indicated concentrations of E-64 for 1 h at 37 °C and were incubated with TCpABP (final concentration of 10 μM) for an additional 1 h. Subsequently, the lysates were added to the 5× SDS sample buffer, denatured by boiling for 5 min, and resolved on a 13.5% SDS-PAGE gel. The results were visualized using a Typhoon 9410 laser scanning system (Amersham Biosciences).

EXPERIMENTAL SECTION

Materials

N-α-Fluorenylmethyloxycarbonyl (Fmoc)-L-amino acid derivatives, Fmoc-D-amino acid derivatives, 2-chlorotrityl chloride resin, and reagents used for peptide synthesis were purchased from GL Biochem, Ltd. Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine serum (FBS) were purchased from Gibco BRL. LysoTracker Red and RIPA lysis buffer were obtained from Beyotime. Protease Inhibitor Cocktail was purchased from Roche Applied Science. 5764

dx.doi.org/10.1021/pr300575u | J. Proteome Res. 2012, 11, 5763−5772

Journal of Proteome Research

Article

Confocal Laser Scanning Microscopy

for 10 min, and vacuum-dried. Trypsin (12.5 ng/μL unless specified) in 50 mM ammonium bicarbonate was absorbed into the gel during a 45 min incubation at 4 °C. The proteins were then digested into peptides overnight at 37 °C. The peptides were extracted from the digested gel pieces and vacuum-dried for further analysis.

RAW264.7 cells were seeded onto eight-well Lab-Tek chamber slides (Nunc) at a density of 1.0 × 104 cells/well in growth media (DMEM supplemented with 10% (v/v) FBS). Cells were grown at 37 °C under CO2 (5%, v/v) overnight. Cells were washed with PBS (pH 7.4) twice and incubated with the indicated probes (10 μM, 300 μL) in PBS for 30 min. The medium was discarded, and the cells were rinsed with PBS twice. After a post-labeling incubation in DMEM (300 μL) containing 10% FBS for 4 h, the cells were analyzed with a confocal laser scanning microscope. To stain lysosomes, the cells were incubated with 300 μL of 200 μM LysoTracker Red for 30 min before imaging. The fluorescence signals were detected using a Fluoview FV1000 (Olympus) confocal laser scanning microscope equipped with a 60× oil-immersion objective lens (NA 1.4) and a HeNe-G laser (green channel: excitation (EX) 488 nm, emission (EM) 505−555 nm; red channel: EX 543 nm, EM 570−680 nm).

LC−MS/MS Analysis

The dried peptide samples were resuspended in sample buffer (2% formic acid and 5% acetonitrile) and analyzed using capillary reverse phase chromatography combined with tandem mass spectrometry using a LTQ-Orbitrap Velos mass spectrometer (Thermo Finnigan). The instrument was equipped with a Waters NanoAcquity ultra performance liquid chromatography (UPLC, Waters) and a 75 μm i.d. × 15 cm fused-silica capillary column (Sino-America Proteomics) with C18 reversed-phase resins (Magic C18AQ; particle size, 3 μm; pore size, 200 Å; Michrom Bioresources). Samples were loaded onto the column by the autosampler at a flow rate of approximately 600 nL/min. The peptides were eluted over a 70 min gradient (8−45% buffer B; buffer A, 0.1% formic acid, 5% acetonitrile; buffer B, 0.1% formic acid, 95% acetonitrile) at approximately 200 nL/min. The eluted peptides were ionized under high voltage (2 kV) and detected using a LTQ-Orbitrap Velos mass spectrometer in the data-dependent acquisition (DDA) mode. Survey scan MS spectra were acquired in the Orbitrap analyzer with a resolution of 30,000 at the target value of 1,000,000 ions. The 20 most intense ions were sequentially isolated for fragmentation in the linear ion trap using collisioninduced dissociation (CID) mode with 35% normalized collision energy. Dynamic exclusion of precursors within a 0.01 Da window over 30 s was used to suppress repeated fragmentation of peaks.

Direct Labeling of Proteins in Living Cells

RAW264.7 cells were seeded onto 100-mm culture dishes at a density of 1.0 × 107 cells/well in DMEM (8 mL) containing 10% FBS and grown at 37 °C under CO2 (5%, v/v) overnight. Cells were washed with PBS (pH 7.4) twice and then incubated with 5 mL of PBS containing 10 μM of TCpABP for 30 min. Subsequently, all media were replaced with fresh DMEM containing 10% FBS. After 4 h, the cells were harvested and washed three times with cold PBS. The cell pellets were incubated with 500 μL of RIPA buffer containing 1× complete Mini Protease Inhibitor Cocktail (Roche Applied Science) and were sonicated 100× in 1 s pulses (with a 2 s pause between pulses) in an ice-cold water bath. Subsequently, the cell lysates were centrifuged at 40,000g for 20 min at 4 °C to remove cellular debris. The supernatants (20 μg/lane) were then added to 5× SDS sample buffer, denatured by boiling for 5 min, and resolved on a 13.5% SDS-PAGE gel. The gels were visualized using a Typhoon 9410 laser scanning system (Amersham Biosciences). The protein concentration of the supernatant was determined using BCA protein assay kit (Pierce).

Database Search and Data Analysis

MS/MS spectra were searched by SORCERER with TurboSequest v4.0.3 (SageN Research Sorcerer Enterprise)26 against the NCBI-RefSeq Mus musculus database (ftp://ftp.ncbi.nih. gov/refseq/M_musculus/, downloaded on November, 2011). As a control and for the estimation of false discovery,27,28 the decoy components contained pseudoreversed sequences of all target proteins. Parameters included semitryptic restriction, parent ion mass tolerance (50 ppm), fixed modification of Cys (+57.0215 Da, alkylation by iodoacetamide), and dynamic modifications on Met (+15.9949 Da, oxidation). The detailed sequence information for cathepsins (e.g., mature sequence and classification) was obtained from the MEROPS database for peptidases (http://merops.sanger.ac.uk) to provide accurate weight and sequence coverage determination. The peptide matches were grouped by charge state and trypticity and then stringently filtered by (i) mass accuracy of 10 ppm, (ii) minimal peptide length of 6 amino acids, (iii) a maximum of 4 modification sites, (iv) a maximum of 2 miscleavages, and (v) matching scores (cross-correlation score (XCorr) and normalized difference in cross-correlation scores (ΔCn)). In each group, the matching scores were dynamically increased until all decoy matches were discarded, suggesting that the estimated false discovery rate was zero. However, because the predicted false discovery rate has significant theoretical errors when decoy matches are low,27 it is possible that some false matches escaped this filtering process. Thus, we only accepted proteins with at least two peptide matches to further minimize false discoveries, resulting in the identification of 59 proteins in the samples (Supplementary Table S1). The

Affinity Enrichment and Sample Preparation

Cell lysates (500 μL) were prepared as above, and the TCpABP-labeled proteins were captured by adding 50 μL of streptavidin agarose resin (Pierce) and incubated at room temperature for 1 h. The streptavidin beads were collected and washed successively 3 times with 100 μL of PBS (0.001% SDS), 1 time with 100 μL of KCl-NaOH buffer (pH 13, 0.055 M KCl and 0.145 M NaOH), and 3 times with 100 μL of PBS (0.001% SDS). Then, TCpABP-labeled proteins were eluted by boiling in 120 μL 1× SDS loading buffer (80 °C, 5 min), and the supernatants were collected by centrifugation at 17,000g for 5 min. Proteins were resolved on a 10% SDS-PAGE gel. The BenchMark Protein Ladder (Invitrogen) was resolved on the same gel to estimate the molecular weight of the proteins of interest. The fluorescent signals were visualized and recorded using a Typhoon 9410 laser scanning system (Amersham Biosciences). The gel was subsequently stained with Coomassie blue dye, and the stained bands were visualized using a flatbed transmission scanner (HP Scanjet G4050, Hewlett-Packard). According to the fluorescent and Coomassie stain signals, two gel slices in the molecular weight regions of interest were cut and sliced into 1-mm cubes, destained with 50% acetonitrile in 25 mM ammonium bicarbonate, soaked in 100% acetonitrile 5765

dx.doi.org/10.1021/pr300575u | J. Proteome Res. 2012, 11, 5763−5772

Journal of Proteome Research

Article

Figure 1. Structures of the probes TCpABP, CpFABP, and 1.

Figure 2. Analysis of labeled proteins by SDS-PAGE followed by in-gel fluorescence scanning. (A) RAW264.7 cells lysates (pH 5.5) were treated with the indicated concentrations of E-64 for 1 h at 37 °C. The remaining activity of cysteine cathepsins was measured with fluorescent labeling by incubation with TCpABP (10 μM) for 1 h at 37 °C. (B) RAW264.7 cell lysates (pH 5.5) with or without heat denaturation were treated with TCpABP or probe 1 (10 μM). Samples were separated on a 13.5% SDS-PAGE gel and visualized by in-gel fluorescence scanning.

decline, which probably opened the oxirane ring of the probe.30 Therefore, a mixture containing only TIS (2%) and TFA (anhydrous, 98%) rather than a commonly used mixture consisting of TFA/H2O/TIS (95:2.5:2.5) was employed as the cleavage cocktail in the solid-phase synthesis to avoid the ionization of TFA. After the cleavage, the crude products were neutralized to pH 7 with several drops of ethanediamine. During the HPLC purification, acetic acid was used instead of TFA (see Experimental Section). Finally, we obtained TCpABP at greater than 90% purity; pure TCpABP is stable at 4 °C in PBS buffer. We also generated probe 1, which does not contain the reactive epoxysuccinyl scaffold, and CpFABP (cellpermeable fluorescent activity-based probe), which does not contain a biotin motif, as controls for TCpABP (Figure 1).

common contaminant proteins, such as porcine trypsin and human keratins, were removed.

■

RESULTS

Synthesis and Purification of the Trifunctional Cell-Permeable Activity-Based Probe (TCpABP)

We synthesized the trifunctional cell-permeable activity-based probe (TCpABP, Figure 1) according to previous reports23,24 with some modifications (see Experimental Section). TCpABP consists of four components: (a) a Tyr-Leu epoxysuccinyl scaffold that can selectively label cysteine cathepsins by formation of an enzyme activity-dependent covalent bond,10,29 (b) a fluorophore (e.g., 5(6)-carboxyfluorescein (FAM)) for visualizing labeled enzymes, (c) a biotin motif for enrichment as well as purification of labeled enzymes by avidin affinity chromatography, and (d) a cell-penetrating octaarginine peptide (rRrRrRRR, r = D-Arg, R = L-Arg) to engender the membrane permeability of the probe.14 During the synthesis of the probe, we found that if a mixture of TFA with H2O was used in the cleavage of the probe from the solid support or during HPLC separation, the probe was unstable during freeze-drying, possibly owing to the occurrence of nucleophilic oxirane ring opening.30 We noted that the basic arginine residues of the CPP in the probe can bind ionized TFA in H2O, possibly resulting in a difficult evaporation of TFA during freeze-drying and thus leading to a persistent pH

Evaluation of Labeling Specificity of TCpABP toward Cathepsins in Cell Lysates

The Tyr-Leu epoxysuccinyl scaffold (Figure 1) has been reported to selectively label cysteine cathepsins in cell lysates.10,29 To evaluate whether the conjugation with the polycationic CPP affects the labeling specificity of TCpABP, we used the probe to label the cell lysates while in competition with E-64, a natural, potent, and irreversible inhibitor of cysteine cathepsins.31 Figure 2A indicates that the labeling of TCpABP could be specifically out-competed by pretreatment of the lysates with E-64, suggesting that these two compounds target the same subset of proteases. TCpABP failed to label 5766

dx.doi.org/10.1021/pr300575u | J. Proteome Res. 2012, 11, 5763−5772

Journal of Proteome Research

Article

Figure 3. Confocal microscopy of living Raw264.7 cells that were incubated with TCpABP (A) (green, 10 μM) and CpFABP (B) (green, 10 μM) for 30 min and postincubated in DMEM containing 10% FBS for 4 h. Lysosomes were labeled with LysoTracker Red (200 nM) for 30 min before imaging. Scale bar: 10 μm.

charged proteins via electrostatic interactions (Figure 4B). Thus, we performed an additional wash step, in which a basic buffer (pH 13) was used to wash the streptavidin beads before the elution of the TCpABP-labeled proteins to eliminate the positive charges of guanidine groups, thus minimizing nonspecific electrostatic interaction with proteins. Figure 4C indicates that after the basic wash, no obvious change was observed for the fluorescence bands, implying that the basic buffer has little effect on the affinity binding of the TCpABPlabeled proteins. In contrast, silver staining showed much fewer bands on the gel, suggesting that most of the nonspecific proteins binding via electrostatic interactions were washed away, which is in agreement with the observation that there were obvious bands in the base washing solution (Figure 4D, W4*).

heat-denatured proteomes (Figure 2B), implying that the probe cannot bind inactive proteins, which further demonstrated the binding specificity between our probe and the targeted proteins. Probe 1 without an epoxysuccinyl group could not label any proteins, suggesting that the labeling of cathepsins by TCpABP owes to the epoxide electrophile rather than the peptide binding region (Tyr-Leu). In our previous study,15 Western blot analysis using anticathepsin B polyclonal antibody has indicated that one of the labeled proteins by CpFABP is cathespsin B. These results positively imply that TCpABP can label cathepsins. Labeling Enzymes with TCpABP in Living Cells

After incubation with living RAW264.7 cells for 30 min, TCpABP entered the cells (Figure 3), indicating that the CPP module in the probe can efficiently transport the cellimpermeable motifs (e.g., biotin and FAM) into living cells. After a 4 h postincubation in DMEM, TCpABP exhibited diffuse staining in addition to a number of punctate vesicles. A colocalization study employing LysoTracker Red as the lysosome marker (Figure 3) revealed that the punctate vesicles stained by TCpABP were mainly lysosomes, and this result is in good agreement with cysteine cathepsins being typically localized to lysosomal compartments.11 Moreover, under the same conditions, CpFABP containing the Tyr-Leu epoxysuccinyl scaffold but without the biotin motif exhibited staining mainly in punctate vesicles, which colocalized with LysoTracker Red extremely well,15 implying that the diffuse staining of TCpABP was mainly due to binding of the biotin motif with other proteins. The treated cells were lysed, and the labeled proteins were purified using streptavidin resin and subsequently analyzed with SDS-PAGE (Figure 4A). Figure 4C shows that although several distinct bands were detected using in-gel fluorescence scanning, there were many more bands on the gel when stained with silver, indicating that more proteins had been enriched along with the TCpABP-labeled proteins. We noted that the guanidine moieties (pKa ≈ 12.5) of the octa-arginine peptide in TCpABP carry multiple positive charges over a wide physiological pH range32 and therefore possibly bind negatively

Identification and Quantification of the TCpABP-Labeled Enzymes

The TCpABP-labeled proteins were enriched on streptavidin beads, separated using 1D SDS-PAGE, and detected with a laser scanner. Three major protein bands were observed (approximately 25, 30, and 60 kDa) on the SDS-PAGE (Figure 5A). The fluorescence intensity of the band around 60 kDa is nearly 5 times weaker than that of the bands around 25−30 kDa, indicating that the proteins with molecular weights of 25− 30 kDa are the major TCpABP-labeled proteins. Considering that the calculated molecular weights of mature cathepsins are 23−27 kDa, we focused our identification on the bands around 25−30 kDa. The two bands of approximately 25 and 30 kDa were excised from the gel (Figure 5A) and tryptically digested for identification by LC−MS/MS. The generated MS data were searched by Sequest-Sorcerer algorithm (Sage-N-Research) (see Experimental Section). A total of 59 proteins (containing at least two unique, observed peptides per protein) were identified, including 7 cathepsins (e.g., cathepsins B, H, L1, Z, S, F, and O). No peptide derived from the proregions of the zymogen forms of the cathepsins was found, suggesting that the labeling of cathepsins by TCpABP was dependent on enzyme activity, 5767

dx.doi.org/10.1021/pr300575u | J. Proteome Res. 2012, 11, 5763−5772

Journal of Proteome Research

Article

Figure 4. Labeling proteins with TCpABP in living cells. (A) The workflow for identifying active cathepsins in living cells. (B) A schematic representation of the base wash in the enrichment step using streptavidin beads. (C, D) The effect of the base wash on minimizing the nonspecific electrostatic binding proteins. The base wash and control eluent fractions were separated on SDS-PAGE gels and visualized by both in-gel fluorescence scanning and silver stain. W1, W2, W3, and W4 were flow-through fractions obtained by washing with PBS buffer (pH 7.4); W4* was the fraction obtained by washing with a KCl-NaOH buffer (pH 13).

which is consistent with the results in Figure 2B that TCpABP did not label the heat-denatured, inactive proteome. All of the identified peptides of cathepsins and their Sequest XCorr values are shown in Supplementary Table S4. We noted that the XCorr values of the peptides derived from cathepsin O are approximately 2.2, which are relatively low compared with those of other cathepsins (higher than 3.0). Therefore, we further confirmed the identification of cathepsin O by manual interpretation. We totally identified three peptides for cathepsin O (Supplementary Figure S3 and Table S4). The sequence for the first peptide is DFSAYNFR (Supplementary Figure S3A and B), which has full sets of matched product ions and high d C n. T h e se q u e n c e f o r t h e se c o nd p e p t i d e i s HFPQSQAGVSVK, which contains an easily breakable backbone on proline resulting in strong y and b ions surrounding the respective amino acid. The spectrum we identified did show the strongest y10 and b2 ions surrounding proline, which

matched reasonably well as expected besides the other matched ions, suggesting the right identification for the cathepsin O (Supplementary Figure S3C and D). The sequence for the third peptide is TGNTPYWMVR, which contains the proline as in the second peptide. The spectrum (Supplementary Figure S3E) with strongest y6 and b4 product ions suggests the right identification for the peptide sequence. These results further convince us the right identification for cathepsin O. Next, we quantified the relative abundance of each identified protein using spectral counting. Due to the tendency of large proteins to contribute more peptides and spectra than small ones, a simple normalization method based on total spectral counts normalized to protein size (SC/MW*50, SC = spectral count, MW = molecular weight) was applied in the label-free quantification approach.33 We also evaluated the quality of the spectra using a normalization method based on the sum of peptide Sequest XCorr values (ΣXCorr/MW*50).34 The 5768

dx.doi.org/10.1021/pr300575u | J. Proteome Res. 2012, 11, 5763−5772

Journal of Proteome Research

Article

Figure 5. MS analysis of labeled, purified proteins. (A) RAW264.7 cells were incubated with TCpABP (10 μM) for 30 min. After a 4 h postincubation in DMEM with 10% FBS, the cells were lysed, and the biotinylated proteins were enriched and subsequently washed with a base buffer (pH 13). The eluted sample was separated on a 10% SDS-PAGE gel and visualized by in-gel fluorescence scanning and Coomassie brilliant blue (CBB) staining. (B) The relative abundance of each identified protein (x-axis) using the spectral counting method (SC/MW*50) is compared to the quantitative evaluation of the quality of MS spectra (y-axis) using a normalization method based on the sum of peptide Sequest XCorr values (ΣXCorr/MW*50). CTS: cathepsin.

Table 1. Identified Cathepsins and a Comparison of Their Relative Activity Levels and mRNA Expression Levels in RAW264.7 Cells name cathepsin cathepsin cathepsin cathepsin cathepsin cathepsin cathepsin cathepsin cathepsin a

B L Z S C F K H O

NCBI accession

MW (kD)a

transcript valueb

proteomic identification

SC/MW*50c

sequence coverage (%)

NP_031824.1 NP_034114.1 NP_071720.1 NP_067256.2 NP_034112.3 NP_063914.1 NP_031828.2 NP_031827.2 NP_808330.1

27.36 24.21 27.37 23.84 26.20 23.72 23.42 23.94 23.97

3955.1 3458.9 2244.3 2174.8 1727.4 143.2 55.9 23.5

+ + + + − + − + +

140.7 74.4 82.2 31.5

42.46 48.65 30.36 28.77

16.9

32.56

106.5 4.2

45.66 4.57

Calculated molecular weights. bMicroarray measurements of the abundance of a transcript.40 cSpectrum count/molecular weight*50.

explaining their enrichment on avidin columns.35 The second category includes 14 ribosomal proteins (e.g., ribosomal proteins S3, L18, L14, L7, and L7a). Previous studies have indicated that ribosomal proteins are highly abundant and therefore are common contaminants of many organelle proteomes.36 Moreover, we noted that TCpABP contains several cationic guanidine groups, which easily bind the negative phosphate backbones of RNA via electrostatic interaction.37 In eukaryotic ribosomes, rRNA is the scaffold and catalytic heart; therefore, TCpABP most likely bound to ribosomal proteins via formation of TCpABP-RNA-protein complexes. Although careful washing including a base wash was executed, it is likely that we were unable to completely eliminate the highest affinity bulk unwanted contaminants, such as ribosomal proteins and associated rRNA (which would not

results of the quantification (shown in Figure 5B and Supplementary Table S2) reveal that among the 59 identified proteins, cathepsins are the most abundant. Cathepsins B, H, L1, Z, and S are the top 5 most abundant proteins, and cathepsin F is also among the top 10. These results suggest that the labeling of active cathepsins in living cells with TCpABP was highly specific for cathepsins, and that the enrichment and purification of the targeted enzymes were highly efficient. In addition to the 7 cathepsins, we also identified another 52 proteins that can be classified into four categories (Supplementary Table S2). The first category includes 13 proteins, of which 11 proteins have been reported to nonspecifically bind to the affinity matrices (bead proteomes) and the other two [i.e., methylcrotonoyl-CoA carboxylase (MCCC1) and pyruvate carboxylase (Pcx)] are endogenously biotinylated proteins, 5769

dx.doi.org/10.1021/pr300575u | J. Proteome Res. 2012, 11, 5763−5772

Journal of Proteome Research

Article

By exploiting the remarkable ability of CPPs to deliver cellimpermeable compounds into living cells, several CPP-based cell-permeable ABPs have been developed that are capable of entering living cells and efficiently labeling active enzymes in vivo with cell-impermeable reporter tags. However, to date, these studies16−20 are mainly focused on imaging and SDSPAGE analysis. Here, we have described the protein profiling of active cysteine cathepsins in living cells using a novel cell-permeable TCpABP based on CPPs. TCpABP contains both a biotin motif and a fluorescent tag; therefore, it can be used to simultaneously detect and isolate probe-labeled enzymes by ingel fluorescence scanning and avidin chromatography, respectively, thus setting up a direct connection between the visualization of enzyme activities and their subsequent molecular identification. We demonstrated that the cellpenetrating octa-arginine peptide module endowed TCpABP with efficient membrane permeability and had little effect on the labeling specificity of TCpABP toward cathepsins. However, the octa-arginine module is a likely source of copurification of some contaminants that carry through the enrichment process using streptavidin beads owing to the positive charges and high pKa of the module. To minimize this undesirable side effect, a basic wash was conducted and proved to efficiently remove most of the nonspecific electrostatic binding proteins (Figure 4). An alternative to the strong basic wash we used might employ a cleavable linker (e.g., a disulfide bond)46 in a redesigned version of our probe where the octaarginine peptide would be cleaved from the probe after entering cells, thereby dispensing with the need for a base wash. After the labeling with TCpABP, seven cathepsins (i.e., cathepsins B, H, L1, Z, S, F, and O) were identified. The quantification analysis revealed that for cathepsins B, L1, Z, S, and F, there was a high correlation between their mRNA expression levels and relative activities; however, a poor correlation was observed for cathepsin H, suggesting that a high level of mRNA expression does not always indicate high protein activity. Remarkably, we identified cathepsins F and O, which have low transcript levels and were not identified by previous studies employing ABPs, suggesting that TCpABP is a useful tool for the detection of active cathepsins, especially for cathepsins with low abundance, in living cells.

be sequenced by MS/MS in the positive mode). The third category includes 19 membrane-associated proteins involved in biosynthetic and endocytotic pathways of lysosome biogenesis.38,39 We noted that TCpABP-labeled protein complexes tend to bind contaminants via electrostatic interactions. Washing with a basic buffer (pH 13) could largely reduce the nonspecific proteins. However, the harsh basic buffer also possibly disrupted the membrane fraction, leading to the release of integral membrane proteins, most of which are usually soluble only in detergents or organic solvents. These “sticky” membrane-associated proteins could nonspecifically bind the affinity matrices and would be hard to wash away with PBS buffer. Thus, these membrane-associated proteins were possibly enriched together with the targeted proteins. The fourth category includes 6 proteins, most of which are predicted proteins with unknown function. We next compared the relative activities of cathepsins with their mRNA expression levels in RAW264.7 cells.40−42 According to the analysis of mRNA expression,40 there are a total of 17 cathepsins expressed in this cell line. Among them, 10 cathepsins (i.e., cathepsins K, W, R, M, J, Q, 3, 6, 7, and 8) were expressed at very low transcript levels (transcript values