Technical Note pubs.acs.org/jpr
Gold Nanoparticle-Based Immuno Dual Probes for Targeting Proteomics Chan-Hua Chen,† Jing-Xiang Hong,† Chun-Sheng Wu,† and Shu-Hui Chen*,†,‡ †
Department of Chemistry, National Cheng Kung University, Tainan 701, Taiwan Agricultural Biotechnology Center, National Chung Hsing University, Taichung, 40227, Taiwan
‡
S Supporting Information *
ABSTRACT: Immunoprecipitation combined with mass spectrometry (MS) is a promising technique for targeting proteomics in characterizing submicrograms of target protein and interacting proteins in living cells. This method, however, is limited by interference arising from nonspecific binding. We report a novel gold nanoparticle (AuNP)-based immuno probe approach for immunoprecipitation. By cross-linking the antibody Fc domain to protein G covalently modified on AuNPs, the probe was fabricated and characterized to have 60 protein G and 30 immunoglobins per AuNP. We used human immunoglobin against the target and mouse immunoglobin with the same isotype (IgG) to fabricate the target and preclear probe, respectively, and termed it as the dual probe approach. Our results showed that the preclear probe (AuNP-IgG) and the target probe (AuNP-anti-ERα) share a similar panel of nonspecific binders but dramatic different specificity toward the target. Thus, using the dual probe method, we showed major nonspecific binders in the cell lysate could be largely removed without sacrificing the target protein. Compared to the conventional agarose gelchromatography, the AuNP-based probe exhibited less nonspecific interference and higher recovery yield for ERα. Moreover, the AuNP-based probe is more inert than the agarose gel under harsh conditions and does not induce dissociation of the cross-linked IgG that could interfere with target identification. Using AuNP-based dual probes, ERα was shown to be purified from MCF-7 cells with minimum nonspecific binding. Moreover, the identity and phosphorylation sites on the C-terminus of the purified ERα could be positively confirmed by MS using only 1 mg of cellular protein. KEYWORDS: gold nanoparticles, immunoprecipitation, targeting proteomics, nonspecific binding, protein complex, protein identification
■
INTRODUCTION Immunoprecipitation (IP) (or pull-down) is a small-scale affinity purification method using a specific antibody against the target.1 Traditional IP involves capturing the target protein with antibodyimmobilized agarose beads and then recovering the purified target protein and its interacting proteins in sample loading buffer for analysis by gel electrophoresis. The purified proteins are subsequently digested in the gel and identified by mass spectrometry (MS). Alternatively, the eluent can be digested and directly analyzed by MS without gel separation. Microsized magnetic nonporous beads are also used as support material to accelerate processing speed via magnet-assisted separation.2 The IP experiment involves attempting to capture not only the direct target, but also any proteins that interact with the target in a biological system, such as the cellular lysate, by co-immunoprecipitation (Co-IP).3 Thus, combining IP (and Co-IP) and MS-based proteomics techniques, the target protein and its interaction partners can be identified and characterized.4−9 Studying protein complexes and signaling pathways via MS-based IP and Co-IP could reveal potential biomarkers and potential targets for new drug development.10 Although the porous nature of the traditional gel beads used for IP experiments provides high binding capacity, many species © 2012 American Chemical Society
could stick on the porous surface nonspecifically. Moreover, lengthy and continuous rotation is required for gel beads because they precipitate very quickly. Thus, monodispersed nanomaterials such as nonporous material modified nanoparticles are good substitutes for gel beads. The binding capacity of a nanoparticle is comparable to porous gel beads due to a high surface area to volume ratio, and they can be well suspended in the liquid phase to enhance interfacial interactions.11 Superparamagnetic nanoparticles are one of the most popular nanomaterials for affinity purification12 because their separation can be accelerated by using a magnet. The exposed transition metal ion sites, however, could act as a redox center in vivo13 and in vitro. In contrast, gold nanoparticles (AuNPs) are relatively inert and nontoxic.14,15 Compared to superparamagnetic nanoparticles, AuNPs are much easier to be modified uniformly through thiol−Au binding chemistry16 to minimize nonspecific binding. The monodispersed AuNPs prepared from citrate reduction have been demonstrated to be useful for concentrating proteins from a relatively large volume of biological fluid by aggregation, which is superior to the traditional trichloroacetic acid Received: April 1, 2012 Published: May 26, 2012 3921
dx.doi.org/10.1021/pr300315n | J. Proteome Res. 2012, 11, 3921−3928
Journal of Proteome Research
Technical Note
was added to the solution and reacted at 4 °C on a shaker for 12 h. After the reaction, the immunoglobin modified AuNPs were collected and washed at least three times by centrifugation (5100g at 15 °C for 30 min each time), followed by redispersion in 50 μL of sodium tetraborate buffer (200 mM, pH 9.0). For cross-linking, 40 mM dimethyl pimelimidate (DMP, Sigma, St. Louis, MO) solution dissolved in sodium tetraborate buffer was added into the immunoglobin-modified AuNP solution and reacted at room temperature on a shaker for 30 min. The procedure was repeated twice and the last reaction was stopped by adding 100 μL of 0.2 M Tris buffer. The walladsorbed AuNPs were desorbed by adding 1 μL of Tween-20 buffer, and the immunoglobin-modified AuNPs were washed at least three times by centrifugation (5100g at 15 °C for 30 min each time), followed by redispersion in 100 μL of PBS buffer. The number of protein G and IgG bound to AuNP was characterized by titration methods using gel electrophoresis for detection.
precipitation because it is ineffective in a large volume.17 Moreover, a combined approach using stable isotope labeling and a DNA functionalized AuNP probe has been shown to characterize a large transcriptional complex related to estrogen action.11 The quantification measure using stable isotope labeling could be used to differentiate specific versus nonspecific binding and to find out differences between two biological states.7,9,10,18 AuNP-based probe method appears to be promising for targeting proteomics study. Because most antibodies normally have a high affinity (Ka > 107) to their target, target capturing using IP method is less likely to be problematic if a “good” antibody is used. However, nonspecific binders or impurities released from the probe can be incorporated into the sample during the IP process, causing severe interference in identifying target proteins of low abundance. These nonspecific binders can be associated with the solid support, linker, or the bait molecule itself. Tandem affinity purification (TAP) is a popular method used to reduce nonspecific binding by breaking the binding between the bait molecule and the solid support or the linker after pull-down.18−20 The TAP method, however, requires fusion of the TAP tag, either N- or C-terminally, to the target protein of interest and is not applicable for endogenous protein targets. Furthermore, the use of enzyme or UV light to break the binding is likely to cause adverse effects, such as protein degradation. Alternatively, preclearing the sample before IP is a good way to remove potentially reactive components from the sample to prevent nonspecific binding of these components to the IP bead or antibody.21 However, the fabrication of a useful preclear probe that can remove nonspecific binders but have low affinity for the target is challenging. In this study, we report a AuNP-based immuno dual probe approach for efficient sample precleanup and target pulldown to enrich a model endogenous protein, estrogen receptor α (ERα), from MCF-7 cells and compare it with the traditional agarose gel.
■
Cell Culture
Human breast cancer cells (MCF-7) were cultured in phenol redfree Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum, 17.8 mM NaHCO3, and 1% antibiotic−antimycotic at pH 7.2. Cells were grown in a 37 °C humidified incubator containing 5% CO2 until confluence. For phospho-analysis, the cells were starved for 18 h, followed by the addition of 20 μL of 17β-estradiol (Sigma, St. Louis, MO) solution dissolved in DMSO to a final concentration of 10−7 M and then incubated at 37 °C for 5 min. After washing with PBS, the cells were collected using a volume of 350 μL of RIPA buffer containing 50 mM Tris-HCl, 1% NP-40, and 150 mM NaCl at pH 7.4 and one tablet of protease inhibitor cocktail (Complete, Sigma, St. Louis, MO) and lysed for 30 min by gentle rotation at 4 °C. This was followed by passing the sample through a 25 G needle to dissociate the bound DNA. After centrifugation at 14 800g for 10 min at 4 °C, the supernatant, containing the whole cell lysate, was carefully collected and stored at −80 °C until use. The total protein concentration was determined by the Lowry assay as described by the manufacturer.
MATERIALS AND METHODS
Probe Fabrication
AuNPs were prepared by sodium citrate reduction following the procedures reported earlier.17 We used 2-iminothiolane to modify protein G; the resulting thiol linkers were subsequently bound to the surface of AuNPs. The secondary antibody mouse IgGs or specific human antibody (anti-ERα) IgG was then bound to the probe via protein G through their Fc domain to form the AuNP-IgG preclear probe for sample precleanup and the AuNPanti-ERα target probe for target IP, respectively. For protein G modification, a volume of 4.5 μL of 2iminothiolane solution (10 mM, Sigma, St. Louis, MO) was added to the solution containing 100 μg of recombinant protein G (Invitrogen, Carlsbad, CA) dissolved in PBS buffer (2.7 mM potassium chloride, 0.1 M sodium chloride, 10 mM disodium hydrogen phosphate, 1.7 mM monopotassium phosphate, pH 7.4). After 1-h reaction at 37 °C, the remaining 2-iminothiolane was removed by dialysis (10 kDa MW cutoff membrane) against deionized water for 2 h. The thiolated protein G (30 μg) was then incubated with 100 μL of AuNPs solution (10 nM) on a shaker at 4 °C for 12 h. After the reaction, the protein G modified AuNPs were collected and washed at least three times by centrifugation (5100g at 15 °C for 30 min each time), followed by redispersion in 100 μL of PBS buffer. For immunoglobin modification, 10 μg of the secondary antibody (mouse IgG, Jackson Immuno Research, West Grove, PA) or the antibody (mouse IgG) against human ERα (Santa Cruz Biotechnology Inc., Santa Cruz, CA)
IP Assay
A 150-μL volume of the AuNP probe solution was diluted with PBS buffer up to a 1-mL total volume and then incubated with the cell lysate containing 500 μg of total protein for 12 h at 4 °C. The supernatant was collected by centrifugation (10 000g at 4 °C for 2 min) and the pellet was washed three times with 1 mL of PBST buffer containing 10% acetonitrile (ACN) and 0.05% Tween-20 by centrifugation (5100g at 15 °C for 30 min). For elution, the pellet was resuspended with 30 μL of 2% SDS solution and incubated at 100 °C for 5 min. The eluent (supernatant) was collected by centrifugation (10 000g at 4 °C for 2 min) and then loaded into a 12% SDS−PAGE gel for protein separation. The dual probe assay followed the same IP protocol as described but slightly different sequences. As shown in Figure 1, the cell lysate was first incubated with the AuNP-IgG preclear probe. The supernant (precleared lysate) was combined with the wash buffer of the pellet and then incubated with the AuNP-anti-ER probe. The pulled down target protein was eluted with 2% SDS solution and separated by SDS−PAGE. For IP by gel beads, the protein G-coated agarose gel beads were cross-linked with anti-ERα antibody using the same reagents as used for the AuNP anti-ERα antibody probe and processed with the same protocol as described except the Tris buffer was used to dilute the gel instead of PBS buffer which was used for AuNP-based probes. 3922
dx.doi.org/10.1021/pr300315n | J. Proteome Res. 2012, 11, 3921−3928
Journal of Proteome Research
Technical Note
Table 1. Proteins Identified from the Gel Bands Indicated in Figure 4 gel band a
b
b′
c
Figure 1. The workflow for the dual immunoprobe approach using AuNP-IgG preclear probe and AuNP-antitarget target probe.
Gel Band Staining
c′
The gel was placed into 100 mL of buffer containing 50% methanol and 7% acetic acid and shaken for 30 min twice. Next, the gel was stained with 60 mL of Sypro Ruby solution in the dark and shaken for 12 h, then destained twice using 100 mL of the buffer containing 10% methanol and 7% acetic acid. For silver staining, the gel was fixed with 100 mL of the buffer containing 50% methanol and 12% acetic acid, washed three times with 50% ethanol solution, and then treated with 100 mL of the sodium thiosulfate solution (0.8 mM). After washing three times with deionized water, the gel was impregnated with about 100 mL of silver nitrate (2% w/w) solution for 20 min. The stained gel band intensity was digitized using a computerized image analyzer (UVP, Upland, CA).
d
protein name Stress-70 protein, mitochondrial OS = Homo sapiens Heat shock cognate 71 kDa protein OS = Homo sapiens Tubulin alpha-1B chain OS = Homo sapiens Tubulin beta-2C chain OS = Homo sapiens Tubulin beta chain OS = Homo sapiens Ig gamma-2A chain C region, membranebound form OS = Mus musculus Ig gamma-2A chain C region, A allele OS = Mus musculus Tubulin alpha-1B chain OS = Homo sapiens Tubulin beta-2C chain OS = Homo sapiens Tubulin beta chain OS = Homo sapiens Actin, alpha skeletal muscle OS = Homo sapiens Actin, alpha cardiac muscle 1 OS = Homo sapiens Actin, cytoplasmic 2 OS = Homo sapiens Actin, cytoplasmic 1 OS = Homo sapiens Actin, alpha skeletal muscle OS = Homo sapiens Actin, alpha cardiac muscle 1 OS = Homo sapiens Actin, cytoplasmic 2 OS = Homo sapiens Actin, cytoplasmic 1 OS = Homo sapiens Ig kappa chain C region OS = Mus musculus
MW (kDa)
sequence coverage (%)
73.92
46.1
71.08
20.4
50.80 50.26 50.10 44.61
33.3 28.8 38.3 17.8
39.94
21.5
50.80 50.26 50.10 42.37
30.0 32.8 39.0 26.5
42.33
26.5
42.11 42.05 42.37
52.5 52.5 27.9
42.33
27.9
42.11 42.05 11.94
42.4 42.4 50.0
Nano-LC/MS/MS
The dried digest was reconstituted in 25 μL of buffer C (0.1% FA in DI-H2O) and analyzed by LTQ Orbitrap XL (Thermo Fisher Scientific, San Jose, CA). Reverse-phase nano-LC separation was performed on an Agilent 1200 series nanoflow system (Agilent Technologies, Santa Clara, CA). A total of 10 μL of sample was loaded onto an Agilent Zorbax XDB C18 precolumn (0.3 × 5 mm, 5 μm), followed by separation using a C18 column (i.d. 75 μm × 25-cm, 3 μm, Micro Tech, Fontana, CA). Buffer C was 0.1% FA, and buffer solution D was 0.1% FA in 98% ACN. A linear gradient (0−5% D (0−5 min); 5−40% D (5−125 min), 40−90% D (125−145 min); 90−90% D (145−155 min); 90−5% D (155−165 min); 5−5% D (165−180 min)) over a 180-min period at a flow rate of 300 nL/min was applied. The peptides were analyzed in the positive ion mode by electrospray ionization (spray voltage = 1.8 kV). The MS was operated in a datadependent mode, in which one full scan with m/z 300−2000 in the Orbitrap (R = 60 000 at m/z 400) using a rate of 30 ms/scan. The five most intense peaks for fragmentation with a normalized collision energy value of 35% in the LTQ were selected. A repeat duration of 30 s was applied to exclude the same m/z ions from the reselection for fragmentation.
Western Blot
The separated proteins were transferred from the gel to a 0.22μm PVDF membrane (Stratagene, La Jolla, CA). The membrane was first blocked with 5% non-fat milk and then incubated with the primary antibodies, followed by secondary antibodies. The blot was developed using an enhanced chemiluminescence detection reagent (Amersham ECL Plus; GE Healthcare, Piscataway, NJ), and the spot intensity was digitized using the computerized image analyzer described above. In-Gel Digestion
The gel band was cut, divided into small pieces (1 mm3), and then destained several times with buffer containing 40% ACN and 0.1 M ammonium bicarbonate until colorless. After dehydration and drying, the gel pieces were incubated with the reducing agent (1 mL) containing 10 mM dithiothreitol (DTT) in 0.1 M ammonium bicarbonate for 45 min at 56 °C, followed by incubation with 55 mM iodoacetamide dissolved in 0.1 M ammonium bicarbonate for 30 min at room temperature in the dark. After washing and drying, the gel pieces were incubated with 50 μL trypsin (50:1 w/w protein/ trypsin ratio) dissolved in 0.1 M ammonium bicarbonate buffer for 18 h at 37 °C. After digestion, the tryptic peptides were extracted from the gel pieces using a 40−80% ACN solution by sonication and combined with the supernatant, then dried under vacuum.
Database search and protein identification
All MS/MS spectra acquired were converted to peak list using the Mascot Distiller program (version 2.3, Matrix Science Ltd., London, U.K.) with default parameter settings for Orbitrap_ res_MS2. The converted peak list (in the standard ROV format) was searched against the International Protein Index protein databank (IPI_human, 86 845 sequences updated on Nov. 6, 2009). The following settings for Mascot search were used: a mass tolerance of 0.05 Da for precursors and 0.8 Da for product 3923
dx.doi.org/10.1021/pr300315n | J. Proteome Res. 2012, 11, 3921−3928
Journal of Proteome Research
Technical Note
Figure 2. (a) The sequential shift of the UV absorbance maximum through step-by-step modification from bare AuNPs (blue), to protein G-AuNPs (red) and IgG-protein G-AuNPs (green). (b) The amount of bound protein G when the AuNP solution (100 μL) was titrated with various amounts (1− 50 μg) of thiolated protein G. (c) The amount of bound IgG when the protein G-AuNP solution (100 μL) was titrated with various amounts (1−25 μg) of IgG.
scattering and the modified AuNP was estimated to be over 25 nm.11 The number of protein G and antibody bound to AuNPs was characterized by titration methods by adding various amounts of thiolated protein G ranging from 1 to 50 μg to the AuNP solution (100 μL) and then quantifying the silver stained gel spots detected on PAGE (Figure 2b). The gel loading buffer containing reducing agents β-mercaptoethanol and dithiothreitol (DTT) was used to elute the bound protein G and IgG molecules to be detected by PAGE. Protein G was separated as a single band at 36 kDa and the immunoglobin was separated as a heavy chain at 50 kDa and a light chain at 28 kDa on PAGE. We then used the immunoglobin heavy chain (50 kDa) band and the protein G (36 kDa) band for quantification; a calibration curve for each was constructed (Supporting Information Figure S1). As shown (Figure 2b), a plateau was reached as the added protein G exceeded 30 μg; the bound protein G could then be estimated at 1.40 μg based on the calibration curve (Supporting Information Figure S1). Because the concentration of the AuNP solution was approximately 10 nM, the maximum (saturation) number of bound protein G units was estimated to be 60 per AuNP. We used the same approach to characterize the number of immunoglobins bound to the AuNP through protein G. The titration experiment was performed by adding 100 μL of protein G-bound AuNP solution with various amounts of IgG ranging from 5 to 25 μg. As shown (Figure 2c), a plateau was reached as the added IgG exceeded 5 μg corresponding to 2.26 μg of IgG bound on AuNPs from the calibration curve (Supporting Information Figure S1). Thus, the maximum (saturation)
ions; Phos (ST) and Phos (Y) as the variable modifications; carbamidomethyl-cysteine as the fixed modification; and one allowable trypsin miscleavage. Rank 1 sequence (the highest peptide score) was assigned for each MS−MS spectrum and peptides were considered identified if their Mascot individual ion score was higher than 20 (p < 0.05). Only proteins identified by more than one unique peptide and with >15% sequence coverage were reported (Table 1). Details about each protein identification were shown in Supporting Information (Supplement 2 for protein ID).
■
RESULTS AND DISCUSSION
Probe Fabrication and Characterization
As shown in Figure 2a, the sequential red-shift of the maximum absorbance indicates the increase in particle size as a consequence of the step-by-step modifications of bare AuNPs (λmax = 518 nm) to protein G-AuNPs (λmax = 524 nm) and IgGprotein G-AuNPs (λmax = 530 nm). It is noticeable that red-shifts observed upon successive additions of protein G and IgG to AuNPs are accompanied by band broadening which may be due to AuNP aggregation as the particle size increases and may not be proof of protein G and IgG conjugation. Instead of covalent binding, protein G is known to specifically bind to Fc domain of IgG through affinity interaction, making Fab domain of IgG accessible to the antigen. Simpler alternatives using electrostatic adsorption of the antibodies to the AuNP22 may be attempted in the future. The bare AuNP was characterized with an approximate diameter of 19.2 nm in its hydrated form by light 3924
dx.doi.org/10.1021/pr300315n | J. Proteome Res. 2012, 11, 3921−3928
Journal of Proteome Research
Technical Note
number of IgG bound on protein G-bound AuNP was estimated to be 30 per AuNP. Taken together, our probe was estimated to have 60 protein G and 30 IgG molecules bound per AuNP. Because two IgG antibodies were reported to bind to a single protein G in free solution,23 the 0.5/1 for IgG/protein G stoichiometry obtained here indicates that near 3/4 of IgG binding sites of protein G may not be accessible or may have a lower affinity due to the immobilization of protein G on the AuNP surface. Considering the size of the IgG molecule (150 kDa), which is 4−5 times larger than protein G (32 kDa), we believe the surface coverage of IgG on our probe was near saturation due to steric hindrance. For pull-down experiments from the cell lysate, the immobilized antibodies were further cross-linked with protein G by dimethyl pimelimidate to minimize any leakage during washing and elution. As shown in Supporting Information Figure S2, some unknown species were seen as smear bands above and below the ERα band if the probe was not cross-linked. However, these bands disappeared if the probe was cross-linked, indicating that cross-linking could effectively prevent leakage and nonspecific binding.
Figure 3. Comparison of the nonspecific binding using Sypro ruby stain (top) and specific binding using immunoblotting against ERα (bottom). Each lane was loaded with the marker (M, lane 1), 150 μg input cell lysate (lane 2), and eluents under different IP conditions using AuNP-IgG preclear probe (lane 3), AuNP-anti-ERα probe (lane 4), AuNP-IgG/AuNP-anti-ERα probe dual probe (lane 5), gel (Protein G Agarose)-anti-ERα (lane 6), and AuNP-IgG/Gel (Protein G Agarose)-anti-ERα probe dual probe (lane 7) as indicated. The arrow indicates the 65 kDa position in the gel used for immunoblotting shown in the bottom of each lane. The letter indicates the gel band (left to the letter) cut for protein identification shown in Table 1.
IP by Dual AuNP Probes
The effectiveness of the dual probe method was investigated. The probe used for preclearance must have high molecular similarity with the solid support and with the bait molecule, except for the specific binding domain. We used the same solid support, AuNPs, and immunoglobin with the same isotype for fabricating both the target and the preclear probe. Sypro ruby dye, which has a nanogram level of detection sensitivity, was used to detect nonspecific binders because the amount of target protein was expected to be below the detection limit of the Sypro ruby dye. Western blotting/chemiluminescence, which has a picogram level of detection sensitivity, was used to detect the lowabundance target protein (ERα). As shown in Figure 3, compared to the input of the lysate which contains 150 μg of total protein (lane 2), both the AuNP-IgG (lane 3) and AuNPanti-ERα (lane 4) probe could enrich a similar panel of the protein subclass from the cell lysate containing 500 μg of total protein, whereas the AuNP-anti-ERα target probe could enrich a majority of ERα (lane 3 in the ERα blot shown in the bottom of Figure 3) from the cell lysate but the AuNP-IgG preclear probe had poor affinity with ERα (lane 2 in the ERα blot). These results indicate excellent specificity of the anti-ERα antibody and a similar panel of nonspecific bindings shared by two probes. Thus, using the dual probe method, the nonspecific binding could be largely excluded without losing much of the target protein ERα (lane 5 in the ERα blot). As shown in Table 1, the major nonspecific binding proteins were identified to be tubulin (band b′) and actin (band c′) by ingel digestion and they can be removed by the preclear probe. Since there is still significant amount of tubulin (band b′) after preclearance (lane 5), we suspect some tubulin protein could be ERα interacting protein11 which was co-purified by AuNP-antiERα probe. Likewise, band d′ which was suspected to contain 40S ribosomal protein (sequence coverage
