Quantification of Activated NF-κB/RelA Complexes Using ssDNA Aptamer Affinity – Stable Isotope Dilution—Selected Reaction Monitoring—Mass Spectrometry*

Nuclear Factor-κB (NF-κB) is a family of inducible transcription factors regulated by stimulus-induced protein interactions. In the cytoplasm, the NF-κB member RelA transactivator is inactivated by binding inhibitory IκBs, whereas in its activated state, the serine-phosphorylated protein binds the p300 histone acetyltransferase. Here we describe the isolation of a ssDNA aptamer (termed P028F4) that binds to the activated (IκBα-dissociated) form of RelA with a KD of 6.4 × 10−10, and its application in an enrichment-mass spectrometric quantification assay. ssDNA P028F4 competes with cognate duplex high affinity NF-κB binding sites for RelA binding in vitro, binds activated RelA in eukaryotic nuclei and reduces TNFα-stimulated endogenous NF-κB dependent gene expression. Incorporation of P028F4 as an affinity isolation step enriches for serine 536 phosphorylated and p300 coactivator complexed RelA, simultaneously depleting IκBα·RelA complexes. A stable isotope dilution (SID)-selected reaction monitoring (SRM)- mass spectrometry (MS) assay for RelA was developed that produced a linear response over 1,000 fold dilution range of input protein and had a 200 amol lower limit of quantification. This multiplex SID-SRM-MS RelA assay was used to quantify activated endogenous RelA in cytokine-stimulated eukaryotic cells isolated by single-step P028F4 enrichment. The aptamer-SID-SRM-MS assay quantified the fraction of activated RelA in subcellular extracts, detecting the presence of a cytoplasmic RelA reservoir unresponsive to TNFα stimulation. We conclude that aptamer-SID-SRM-MS is a versatile tool for quantification of activated NF-κB/RelA and its associated complexes in response to pathway activation.

The innate immune response is a stereotypic cellular response to alarmins, defensins, and pathogen-associated molecular patterns that initiates inflammation, secretion of antimicrobial proteins and produces adaptive immunity (1,2). A major intracellular effector pathway of the innate immune response involves nuclear factor-B (NF-B) 1 , a ubiquitous family of highly inducible cytoplasmic transcription factors that control the expression of genetic networks that both initiate mucosal inflammation and later terminate the innate immune response (3,4).
In unstimulated cells, the NF-B/RelA transcriptional activator is inhibited by binding to the IB inhibitory proteins, resulting in cytoplasmic retention and inactivation of its DNAbinding activity (5). NF-B/RelA activation is mediated by ligand-bound cell surface cytokine or pattern recognition receptors. Here, receptor-associated adapter proteins are modified by polyubiquitination, a post-translational modification that is recognized by the ubiquitin binding domain of the I kappa B kinase (IKK) regulatory subunit (6), recruiting IKK into the submembranous complex (7,8). This molecular interaction produces serine (Ser) phosphorylation in the IKK activation loop (9). Subsequently IKK dissociates back into the cytoplasm where it phosphorylates the IB␣ NH 2 -terminal regulatory domain on two adjacent Ser residues (10). Phospho-IB␣ is then selectively degraded within the complex (11,12).
Liberated from its IB␣ inhibitor, RelA rapidly enters the nucleus. A parallel pathway, required for NF-B transactivation, produces RelA phosphorylation at serine residue 536 or 276 (13)(14)(15)(16). This event is required for binding the p300 coactivator, stimulating lysine 310 acetylation (17), a post-translational switch that enhances NF-B binding, forming a transcriptionally active enhanceosome, modifying histones of target gene promoters and inducing proteolytic clearance of promoter-bound RelA through the ubiquitin proteasome pathway (13). In this way, subcellular localization, transcriptional activity, and protein turnover of NF-B is controlled by coupled inducible post-translational modifications that affect protein-protein interactions.
A significant technical advance in systematic discovery of protein interaction networks has been the development and refinement of the tandem affinity purification (TAP) technique (18). In this technique, an affinity epitope-tagged "bait" protein is expressed and bait-associated complexes are isolated under native conditions by sequential affinity isolation. Using TAP, the interactome of 32 major regulators of the NF-B pathway was determined, identifying over 680 nonredundant interacting proteins, representing 171 of 241 previously reported protein interactions (19). Bioinformatics filtering to eliminate nonspecific and high abundance contaminants reduced the interacting proteins to 131 "high confidence" interactors (19). Although this work significantly increased our understanding of the NF-B signaling pathway interactome, there were few interactions identified for nuclear coreceptors, such as p300, transcriptional elongation complexes, histone deacetylases, and protein demethylases, representing some of the major protein interactions required for the transcriptional regulatory action of NF-B/RelA.
Methods that selectively isolate activated NF-B/RelA complexes may increase our understanding of the biochemistry of its transcriptional activation. To this end, identification of single-strand oligonucleotides (ssDNA) termed "aptamers" that selectively bind the activated state of target proteins may have an important role in understanding the abundance and interaction network of activated complexes. Here we report an aptamer that stably binds and selectively isolates the activated state of NF-B RelA. ssDNA aptamer P028F4 binds and associates with activated NF-B/RelA in eukaryotic nuclei, and inhibits its association with endogenous promoters within the chromatin environment. Incorporation of aptamer P028F4 as an affinity isolation step after enrichment with FLAG binding enriches for RelA in its Ser 536 phosphorylated and p300-complexed state, while depleting inactive IB␣⅐ RelA complexes. We demonstrate that an aptamer enrichment technique coupled to absolute protein quantification using stable isotope dilution (SID)-selected reaction monitoring (SRM)-mass spectrometry (MS) can be used to quantify the fraction of activated RelA in cytokine-stimulated eukaryotic cells. This method is a versatile tool for quantification of activated NF-B/RelA and analysis of its associated com-plexes that will further extend systems approaches to understand the NF-B interactome.

EXPERIMENTAL PROCEDURES
Materials-Sequence grade modified trypsin was purchased from Promega (Madison, WI). Anti-FLAG M2 Affinity Gel and 3X FLAG peptide were from Sigma-Aldrich (St. Louis, MO), Nanolink Streptavidin Magnetic Beads were from Solulink (San Diego, CA) and recombinant TNF␣ was from PeproTek (Rocky Hill, NJ). Antibodies used were: anti-RelA and anti-␣-tubulin rabbit polyclonal Ab were from Santa Cruz Biotechnology (Santa Cruz, CA); anti-lamin B Ab was from Calbiochem (San Diego, CA); and anti-histone H3 was from Cell Signaling (Danvers, MA).
Recombinant Protein Expression-The cDNA encoding human RelA (aa 1-313) was PCR amplified and cloned into pGEX-KG expression plasmid to encode 67 kDa glutathione-S-transferase (GST)-RelA (1-313) fusion protein. The cDNA for full length human IB␣ was amplified, cloned into pGEX-KG and transformed into Escherichia coli XL1-Blue. Protein expression was induced using 1 mM isopropyl ␤-D-1-thiogalactopyranoside for 4 h at 37°C. GST-tagged proteins were purified to near homogeneity using glutathione agarose affinity chromatography and dialyzed against HBS (20 mM HEPES pH 7.4, 150 mM NaCl). Protein was stored in aliquots at Ϫ20°C in 50% glycerol.
Aptamer Selection-DNA aptamers were selected after genetic enrichment protocols (20,21). A DNA library consisting of a 30-base random sequence flanked by a 25-base common primer binding sites was commercially synthesized (Integrated DNA Technologies, Coralville, IA). Screening was performed by incubating GST-RelA protein with the aptamer pool in aptamer binding buffer (20 mM Tris-HCl pH 7.4, 150 mM NaCl and 1 mM MgCl 2 ) for 30 -60 min at room temperature, followed by separation of the free DNA from the bound DNA by filtering through a 25 mm diameter nitrocellulose filter (Millipore, Billerica, MA). DNA was eluted from the filter with a 72°C incubation in elution buffer (0.25% SDS, 20 mM Tris-HCl, pH 7.5, 1 mM EDTA) for 30 min and ethanol precipitated. The eluted DNA was quantified using quantitative real-time PCR (Q-PCR). The eluted DNA was PCR amplified using a biotinylated (Bt) reverse primer and an unmodified forward primer. The amplified material was bound to Streptavidin (SA)-paramagnetic beads (Solulink, Inc) in bead binding buffer (20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 200 mM NaCl and 0.01% Triton-X), and washed with binding buffer without detergent. The non-Bt single strand was eluted with 0.15N NaOH at room temperature. The DNA was ethanol precipitated, dissolved in TE and quantified using a Nanodrop 1000 instrument (Thermo Scientific, West Palm Beach, FL). At several points during the selection rounds the aptamer pool was incubated with a nitrocellulose filter to remove filter binding aptamers. Sequences of the amplifying primers were Forward, 5Ј-GGTAACC-TTGAGTC ACGAATTCAA-3Ј; Reverse, 5Ј-CAGAAGCTGTAAG TT-GGGTACCTT-3Ј.
Q-PCR assays and filter binding assays were used to determine when the aptamer pool was enriched with RelA binding species. The enriched pool was PCR amplified with unmodified primers and cloned with a TOPO-Blunt cloning kit (Invitrogen, Carlsbad CA). Individual clones were picked and characterized for RelA binding. Positive clones were sequenced by the UTMB Molecular Genomics core and analyzed using Sequencher software (Genecodes, Ann Arbor, MI).

Binding Studies
Nitrocellulose Filter Binding Assay-Filter binding assays were performed as described (22). Initially, 70 mer oligonucleotides were commercially synthesized (IDT) to include the proximal 20 bases of each flanking primer sequence and the 30 base variable region. DNAs were 5Ј end-labeled with 32 P-ATP using T4-PNK (New England Biolabs, Ipswitch, MA) and purified by glass bead binding using a MERMAID kit (MP Biomedicals, Solon, OH). Protein and labeled aptamer were incubated in aptamer binding buffer and filtered through 25 mm nitrocellulose filters. The bound material was counted in a Beckman Liquid Scintillation counter. For competition studies, binding reactions (100 l) in aptamer binding buffer were incubated at RT for 20 min with 100-fold excess (10 picomoles) of unlabeled competitor DNAs, followed by the addition of 0.1 picomole of 32 P-labeled KB55 probe. After 30 min of incubation, reactions were filtered and counted. The KB55 sequence is ATACGGGAATTCCCG and self anneals to form a duplex. Upper and lower strands of IL8KB are ACTAG-GAATTTCCAGTG and CACTGGAAATTCCTAGT, respectively.
Surface Plasmon Resonance-5Ј-Bt aptamer was attached to a CM5-SA chip (GE Healthcare) to a level of 20 -30 RUs in a Biacore T1000. Running buffer was HBS-T (20 mM HEPES ph 7.4, 150 mM NaCl, and 0.01% Triton X) with 100 g/L sheared salmon sperm DNA. GST-RelA was dialyzed against HBS-T and injected onto the chip at the indicated concentrations. Binding kinetics were determined using the Biacore software.
Cell Culture-Human A549 pulmonary epithelial cells were grown in F12K medium as described (3,23,24). Full length human RelA was expressed as a FLAG-EGFP fusion protein in A549 cells using the pCX4-Pur expression vector (25).

Preparation of Cellular Extracts
Cytoplasmic (CE) and Nuclear Extracts (NE)-A549 cells were scraped and subjected to hypotonic buffer and detergent lysis (26). The supernatant CE was saved and the NE was purified by centrifugation through a sucrose cushion followed by extraction in Buffer C (50 mM HEPES, pH 7.9, 10% glycerol, 400 mM KCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, and 0.1 mM PMSF) with protease inhibitor mixture (Sigma Aldrich) (26,27). Protein content was estimated by Coomassie Brilliant Blue staining using bovine serum albumin as a standard (Bio-Rad, Hercules, CA).
Confocal Microscopy-EGFP-A549 cells were stimulated in the absence or presence of TNF␣ (30 ng/ml, 1 h). The cells were then trypsinized and nuclei isolated and counted. 2 ϫ 10 4 nuclei were centrifuged (cytospin) into slides and fixed with 4% paraformaldehyde. After fixation, the slides were washed twice with PEM (0.1 M PIPES, pH 6.9, 1 mM EGTA, and 1 mM MgCl 2 ) buffer (28), and stained with the indicated TEX 615 (IDT)-conjugated aptamer (5 nM) overnight at 4°C. The following day, the nuclei were counterstained with DRAQ5 (2 mM), and washed in PEM buffer. The slides were air-dried, mounted, and images taken by LSM 510 confocal imaging system.
Electrophoretic Mobility Shift Assays (EMSAs)-NE (15 g) were incubated with 40,000 cpm of 32 P-labeled duplex Naf-1 WT probe, oligonucleotide probe, and 2 g of poly (dA-dT) in binding buffer (8% glycerol, 100 mM NaCl, 5 mM MgCl 2 , 5 mM dithiothreitol, and 0.1 g/ml phenylmethylsulfonyl fluoride) in a final volume of 20 l, for 20 min at room temperature. The complexes were fractionated on a 6% native PAG. The duplex DNA corresponding to the NF-B binding site in the Naf-1 gene was used as a probe (30). Competition was performed by the addition of 100-fold molar excess nonradioactive competitor.
Quantitative Reverse Transcriptase PCR (Q-RT-PCR)-One microgram of total RNA was reverse-transcribed using Super Script III in a 20 l reaction mixture. One microliter of cDNA product was amplified in a 20 l reaction mixture containing 10 l of SYBR Green Supermix (Bio-Rad) and 0.4 M each of forward and reverse gene-specific primers (24,31). The plates were denatured for 90 s at 95°C and then subjected to 40 cycles of 15 s at 94°C, 60 s at 60°C, and 1 min at 72°C in an iCycler (BioRAD). Quantification of changes in gene expression was performed using the ⌬⌬Ct method with unstimulated cells as the calibrator (31).
Two-Step Chromatin Immunoprecipitation (ChIP) Assay-Twostep ChIP was performed in intact cells as described (32,33). Decross-linked DNA was determined by quantitative real-time genomic PCR (Q-gPCR) using promoter-specific primers. Standard curves were generated using a dilution series of genomic DNA (from 40 ng-25 g) for each primer pair (33). The fold change of DNA in each immunoprecipitate was determined by normalizing the absolute amount to input DNA reference and calculating the fold change relative to that amount in unstimulated cells.

Aptamer-Tandem Affinity Purification of RelA
Affinity Isolation of FLAG Tagged RelA-Indicated concentrations of WCE from FLAG-EGFP-RelA-expressing A549 cells were incubated with anti-FLAG M2 Affinity Gel 16 h at 4°C. The beads were washed with chilled WCE buffer four times by centrifugation (5K, 2 min, 4°C). The FLAG-EGFP-RelA was eluted by addition of 150 g of 3ϫ FLAG peptide (34) and incubated with mixing for 1 h at 4°C. The mixture containing FLAG-EGFP-RelA protein complex and gel were filtered through empty spin column (BioRad) and the flow-through collected.
Affinity Isolation of Aptamer-bound RelA-Bt-Aptamer was added to final concentration of 20 nM in FLAG-EGFP-RelA for 2 h at 4°C. During this time, a 20-l slurry of SA-Magnetic Beads was preblocked by dilution with 300 l WCE buffer and 200 l acetylated casein (0.2 mg/ml) at room temperature for 2 h, followed by three washes in WCE buffer. Bt-aptamer bound FLAG-EGFP-RelA was added to the SA-Magnetic beads and incubated for 20 min at 4°C. Bt-aptamer bound protein complex was separated from free complex on a magnetic stand and beads washed six times with WCE buffer. 50 l SDS sample loading buffer was added to the washed beads, boiled, and protein complex was separated from SA-beads using a magnetic stand.

Selected Reaction Monitoring (SRM) Assays
Selection of High Responding Signature Peptides-Signature peptides that stoichiometrically represent the protein candidate were selected on the following criteria: uniqueness for the target protein; length of 8 -25 residues; and absence of missed tryptic cleavages, chemically active amino acid residues (such as cysteine or methionine), and no basic amino acids on either cleavage site of the peptide sequence (35). Pilot liquid chromatography-tandem MS (LC-MS-MS) experiments were performed using tryptic digests of affinity-purified RelA or IB␣ to identify the highest MS responding peptides for each protein. The extracted ion chromatograms based on the monoisotopic peak for all charge states and modifications detected from sequence-identified peptides were compared and the peptides with highest MS response were selected. These signature peptides are listed in Table I for each target protein.
Synthesis of Natural and Stable Isotopically Labeled Peptide Standards (SIS)-Natural tryptic peptides were synthesized using N-(9-fluorenyl) methoxycarbonyl chemistry. SIS peptides were commercially synthesized incorporating isotopically labeled [ 13 C 6 15 N 4 ] Arginine to a 98% isotopic enrichment (Sigma Aldrich). Both natural and SIS peptides were high performance liquid chromatography (HPLC) purified to Ͼ98% purity. The molecular weights were measured with electrospray mass spectrometry and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. The specific peptide concentration was determined by amino acid analysis. Individual SIS peptide stocks of 20 pmol/l were made in 80% acetonitrile.
Trypsin Digestion of Subcellular Protein Extracts-Protein extracts were denatured with 8 M guanidinium HCl and a 50 g aliquot of each sample was reduced with 10 mM dithiothreitol for 30 min at room temperature. Protein cysteinyl residues were alkylated with 30 mM of iodoacetamide for 2 h at 37°C. The samples were diluted 10-fold with 100 mM ammonium bicarbonate, and digested with 2 g of trypsin overnight at 37°C. The trypsin digestion was stopped by adding 10% trifluoroacetic acid. The volume of each sample was adjusted with water to 400 l. The tryptic digests were dried and resuspended in 5% formic acid-0.01% trifluoroacetic acid before SID-SRM-MS analysis.
SIS peptide stocks were diluted to a concentration of 10 fmol/l with 0.01% trifluoroacetic acid. Before LC-SRM-MS analysis, 30 l of each tryptic digest was mixed with 10 l of each SIS peptide. These sample solutions were directly used for SID-SRM-MS analysis without further purification or fractionation.

Quantification of Activated Fraction of Endogenous RelA
Single-step Aptamer Isolation of Endogenous RelA-A549 cell extracts were prepared from control or TNF␣ stimulated cells (30 ng/ml, 45 min). Bt-aptamer was added to a final concentration of 20 nM, and incubated for 2 h at 4°C. A 20 l slurry of acetylated casein-preblocked SA-Magnetic Beads was added and binding continued at 4°C, 20 min, followed by three washes in WCE buffer and three washes in phosphate buffered saline.
On-bead Tryptic Digestion-The beads were resuspended in 30 l of 50 mM ammonium hydrogen carbonate (pH 7.8) and 20 l of 0.1 g/l of trypsin was added. The samples were mixed and trypsinized by gentle shaking overnight at 37°C. After digestion, the supernatant was collected. The beads were washed with 50 l of 50% acetonitrile three times and the supernatant was pooled, and dried. The tryptic digests were then reconstituted in 30 l of 5% formic acid-0.01% trifluoroacetic acid. An aliquot of 10 l of diluted SIS peptides were added to each tryptic digest. The final concentration of each SIS peptide was 2.5 fmol/l. These sample solutions were directly used for LC-SRM-MS analysis without further purification or fractionation.
LC-SRM-MS Analysis-LC-SRM-MS analysis was performed with a TSQ Vantage triple quadrupole mass spectrometer equipped with nanospray source (Thermo Scientific, San Jose, CA). The online desalting and chromatography were performed using an Eksigent NanoLC-2D HPLC system (AB SCIEX, Dublin, CA). Aliquots of 10 l of each tryptic digest were injected on a C18 peptide trap (Agilent, Santa Clara, CA), desalted with 0.1% formic acid at a flow rate of 2 l/min for 45 min. Peptides were eluted from the trap and separated on a reverse-phase nano-HPLC column (PicoFrit™, 75 m ϫ 10 cm; tip ID 15 m) packed in house using Zorbax SB-C18 (5-m diameter particles, Agilent, Santa Clara, CA). Separations were performed using a flow rate of 500 nL/min with a 20-min linear gradient from 2-40% mobile phase B (0.1% formic acid, 90% acetonitrile) in mobile phase A (0.1% formic acid), followed by 0.1-min gradient from 40 -90% mobile phase B and 5-min 90% mobile phase B. The TSQ Vantage was operated in high-resolution SRM mode with Q1 and Q3 set to 0.2 and 0.7-Da Full Width Half Maximum. All acquisition methods used the following parameters: 1800 V ion spray voltage, a 275°C ion transferring tube temperature, a collision-activated dissociation pressure at 1.5 mTorr, and the S-lens voltage used the values in S-lens table generated during MS calibration. Each sample was analyzed by LC-SRM-MS twice. The MS raw files are accessible at https://bioinfo.utmb.edu/CPC/Publications/default.jsp .
SRM Data Analysis-All SRM data were processed using Xcalibur® 2.1. The default values for noise percentage and base-line subtraction window were used. All data were manually inspected to ensure peak detection and accurate integration. The chromatographic retention time and the relative product ion intensities of the analyte peptides were compared with those of the SIS peptides. The variation of the retention time between the analyte peptides and their SIS counterparts were within 0.05 min, and no significant differences in the relative product ion intensities of the analyte peptides and SIS peptides were observed. For the SRM-MS analysis of crude protein extracts, ␤-actin was used as loading control. All of the measured natural verse SIS standard peptide ratios were normalized against ␤-actin.  used to screen for high affinity ssDNA aptamers by repeated rounds of in vitro binding and enrichment. After 12 rounds of selection, we found that greater than 50% of the aptamer pool bound to the target RelA protein. Individual clones were obtained, the inserts were amplified by PCR, separated into single-strands, and tested for RelA binding. We sequenced 9 clones that exhibited binding, and synthesized 70-mer oligonucleotides for each. Competition experiments showed that five of the aptamers fall into the same group, competing with each other for binding; from this group, we selected clone P028F4 for further investigation. Fig. 1A shows the sequence of P028F4 including the primer flanking sequences. The variable region contains the sequence GGGAC, matching half of the double-strand NF-B natural binding site GGGACTTTCC. The four other aptamers in this group also contained this sequence or the similar GGGCC. Competition and direct binding studies showed that the 30-base variable region by itself was sufficient for binding (data not shown). Secondary structure analysis using the UNAFold program (36) did not predict significant secondary structure, and there is no complement to the GGGAC sequence in the aptamer to allow it to form double-strand DNA. We performed competition experiments in nitrocellulose binding assays with natural doublestrand binding sites (Table II). Aptamer P028F4 was able to compete effectively for binding to the KB55 NF-B binding site. A nonspecific control aptamer P028A1 did not compete as well for binding. P028F4 binding affinity was determined by surface Plasmon resonance measurement (Fig. 1B). The binding activity fit well to a 1:1 binding model with a K D of 6.4 ϫ 10 Ϫ10 , calculated from a ka of 1.8 ϫ 10 6 1/Ms and kd of 0.0018 1/s. These estimates are similar to those obtained from nitrocellulose filter binding experiments (not shown).

Isolation of
Aptamer P028F4 Binds RelA and Competes With Duplex dsNF-B Sequences for RelA Binding-To confirm if aptamer P028F4 competes with duplex (ds) NF-B binding sites, a competition-EMSA experiment was performed. Nuclear extract (NE) prepared from control or TNF␣ -stimulated A549 cells was incubated with a radiolabeled dsNF-B DNA binding site in the presence of increasing concentration of unlabeled P028A1 or P028F4 and bound complexes fractionated by native gel electrophoresis. Analysis of the DNA-protein complexes showed a migration pattern characteristic of the NF-B complex (3,37). These complexes bound with NF-B binding specificity, being competed with 100-fold molar excess of unlabeled dsNF-B duplex, but not the corresponding site mutation ( Fig. 2A). Although aptamer P028A1 did not disrupt the NF-B complex, aptamer P028F4 competed effectively for dsNF-B binding.
In a converse experiment, NE prepared from unstimulated or TNF␣-stimulated A549 cells was directly bound to radiolabeled P028F4 as a probe in the presence of unlabeled specific or nonspecific dsDNA competitor. In control NE, a complex of nonspecific proteins was detected on P028F4. Upon TNF␣ stimulation, two new nucleoprotein complexes were detected that represent specific NFkB complexes because they competed specifically with a 100-fold excess of unlabeled dsNF-B wild type (WT), but not mutant (Mut) DNA sequences (Fig. 2B). Together, these data indicated that aptamer P028F4 competes for dsNF-B binding to NF-B/RelA.
Aptamer P028F4 Colocalizes with Activated Nuclear RelA-Other studies have shown that aptamers are bound nonspecifically to high abundance ssRNA binding proteins in cells (28); we also found that transfection of intact cells with chemically synthesized fluorescent TEX 615-conjugated P028A1 and P028F4 showed a high amount of nonspecific cytoplasmic staining, despite methods to reduce nonspecific interactions such as zwitterionic detergents [not shown, ref (28)  from control or TNF␣ stimulated EGFP-RelA were stained with TEX 615-conjugated P028A1 or P028F4 and imaged by confocal microscopy. In the absence of TNF␣ stimulation, little EGFP-RelA staining was detected (Fig. 3A). By contrast, upon TNF␣ stimulation, the nuclei were strongly stained with translocated EGFP-RelA. Nuclei treated with P028A1 were TEX 615-negative in the absence or presence of TNF␣ stimulation (Fig. 3A). By contrast, TNF␣ stimulated nuclei treated with P028F4 strongly stained TEX 615-positive, whereas control nuclei were negative (Fig. 3B). The intense yellow staining of the merged images indicated colocalization of the aptamer and EGFP-RelA (Fig. 3B). Together these data indicate that P028F4 colocalizes with activated EGFP-RelA in eukaryotic nuclei.
Aptamer P028F4 Inhibits TNF␣-Induced Endogenous NF-B Dependent Gene Expression-Earlier studies selectively inhibiting NF-B/RelA translocation identified Gro␤ and TNFAIP3/A20 as part of a TNF␣ inducible gene network downstream of NF-B (3,4,38). To determine whether P028F4 interfered with TNF␣-induced NF-B dependent gene expression, A549 cells were transfected with nothing, P028A1, or P028F4 (4 M respectively) and the cells were TNF␣ stimulated. Relative changes in Gro␤ and TNFAIP3/A20 transcripts were determined by Q-RT-PCR. In the absence of aptamer transfection, Gro␤ mRNA was induced fivefold in response to TNF␣ (Fig. 4A, top). In cells transfected with P028A1, we observed an inconsistent effect on unstimulated Gro␤ expression, but did not significantly reduce the Gro␤ expression induced by TNF␣. By contrast, in P028F4-transfected cells, basal and TNF␣-induced Gro␤ expression were significantly reduced to less than twofold induction. Similar qualitative results were observed for the expression of TNFAIP3/A20, where the nearly sevenfold induction observed in control transfected cells was reduced to less than twofold in P028F4-transfected cells (Fig. 4A, bottom). Importantly, each RNA sample was measured for expression of GAPDH as an internal control, where aptamer transfection had no effect on GAPDH mRNA levels. These observations indicated that transfected P028F4 had a consistent inhibitory effect on NF-B dependent gene expression in vivo.
Aptamer P028F4 Blocks RelA Association with Endogenous NF-B Dependent Gene Promoters In Vivo-Previous studies using dynamic live-cell imaging have shown that RelA association with endogenous chromatin is hyperdynamic (39), in contrast to its stable association with DNA targets in vitro. This observation makes the ability of P028F4 to compete for RelA binding seen earlier in EMSA (Fig. 2) of uncertain relevance to the interaction of RelA with endogenous chromatin. To address this issue, aptamer-transfected A549 cells stimulated in the absence or presence of TNF␣ were subjected to two-step ChIP assay for TNFAIP3/A20 (38). Here, specific RelA binding was observed in unstimulated cells (compare RelA immunoprecipitate to IgG control), and relative RelA binding to Gro-␤ increased to 3.5-fold in response to TNF␣. A weak nonspecific reduction of TNF␣-induced RelA was observed in response to P028A1, whereas a greater, significant inhibitory effect of P028F4 was observed in response to TNF␣ (Fig. 4B, top). Similar results were observed for the Gro␤ promoter, where the 3.5-fold increase in RelA binding was reduced to less than twofold (Fig. 4B, bottom). Together these data indicated to us that P028F4 competes for RelA binding to its endogenous chromatin binding sites, explaining its inhibitory activity.
Aptamer P028F4 Binds Free-but Not IB␣-associated RelA-IB␣ binding to the NH 2 terminus of RelA locks it into a closed conformation for DNA binding and masks its nuclear localization sequence. To determine if P028F4 recognizes free RelA or that complexed to IB␣, purified recombinant GST-RelA (1-313) was pre-incubated in the absence or presence of increasing amounts of recombinant full length GST-IB␣. Complex formation was determined by nondenaturing coimmunoprecipitation assay using anti-RelA Ab. In this experiment, RelA was immunoprecipitated under native conditions, and associated IB␣ detected by Western immunoblot (Fig.  5A). IB␣ staining in the immunoprecipitated material indi- cated RelA association with IB␣. The same free RelA and RelA⅐IB␣ complex mixtures were then separately mixed with radiolabeled P028F4 for binding assay in EMSA. Although P028F4 bound to free RelA, no binding was observed for the RelA⅐IB␣ complexes (Fig. 5B). Together, these data indicated P028F4 recognized RelA only in its activated (non IB␣-complexed) form.
Aptamer Tandem Affinity Purification (ATAP) Enriches for IB␣-free RelA Complexes-The ability of P028F4 to recognize non-IB␣-associated RelA indicated that it could be used in affinity binding experiment to selectively enrich for activated RelA complexes. For this purpose, CE or NE from control-or TNF␣-stimulated A549 cells were bound to Bt-P028A1 or Bt-P028F4. After capture on SA-magnetic beads and washing, the presence of RelA was measured by Western immunoblot. In CE, we observed that P028F4 only captured RelA prepared from TNF␣-stimulated cells, whereas P028A1 did not (Fig. 6A). Although P028A1 captured nuclear RelA in TNF␣ stimulated cells, Bt-P028F4 captured significantly more activated RelA. In both CE and NE, aberrantly migrating forms of RelA were also captured by P028F4, suggesting that the P028F4 enriches for post-translationally modified RelA isoforms (indicated by * ; Fig. 6A). These complexes are highly suggestive of inducible RelA-ubiquitin polymers shown previously to migrate at ϳ150 kDa (40). We next specifically assayed whether P028F4-captured RelA was in an activated state by measurement of activating Ser 536 phosphorylation and p300 complex formation. For this purpose, a two-step affinity isolation was performed, initially purifying EGFP-RelA using FLAG-M2 affinity beads, followed by Bt-P028F4 enrichment on SA-Magnetic Beads. Western immunoblot staining with anti-phospho-Ser 536 RelA Ab shows the presence of Ser 536-phosphorylated EGFP-RelA and endogenous phospho-Ser 536 RelA in the P028F4 eluates (Fig. 6B). The abundance of p300 was below the limit of detection in the FLAG eluates, but was strongly detected in the P028F4 eluate (Fig. 6B). Finally, we also monitored for IB␣ association. Although IB␣ was present in the starting material, IB␣ was found only in the flow-through fraction (Fig.  6B, bottom panel). These data indicate that the ATAP technique enriches for a subset of RelA complexes containing activating RelA Ser phosphorylation, and p300 complexes but not those containing inactive RelA⅐ IB␣ complexes.
We next sought to develop quantitative measurements for the P028F4-enriched RelA and its complexes using selected reaction monitoring mass spectrometry (SRM-MS). First, the SRM-MS assays were optimized by selecting high responding quadropole Q1/Q3 transitions, and assay specificity established.
Selection and Optimization of SRM Q1/Q3 Transitions-Because the intensities of individual fragments derived from one precursor ion differ substantially, selection of the most intense fragments for each signature peptide is essential for developing a high-sensitivity SRM assay. In this study, four or five y-ions of the highest intensity observed in previous LC-MS-MS and SRM-triggered MS-MS experiments, and whose mass to charge ratio (m/z) values exceed precursor ion m/z were selected for each peptide. The sensitivity of the SRM assays for each target peptide was further increased by optimizing the collision energy voltage. As shown in Fig. 7, the SIS RelA peptide, TPPYADPSLQAPVR was infused into the TSQ mass spectrometer, and the optimal collision energy for each transition obtained by performing a breakdown curve. Compared with predicted collision energy voltage (given by 0.034 ϫ (precursor ion m/z) ϩ 3.314), the experimentally optimized collision energy voltage yielded slightly better results. We further empirically optimized the S-lens voltage for each signature peptide. The difference between the values in S-lens table generated during MS calibration and empirically derived S-lens value was found to be an average gain of less than 5% of total peak area. Therefore, in this study, we used the values in S-lens table generated during MS calibration for all the experiments. The selected Q1/Q3 transitions and their collision energy voltages are tabulated in Table I.
Specificity of SRM Quantification-Shared precursor-product ion transitions with those from other analytes can lead to false positives in the subsequent peak assignment and quantification inaccuracy. With the use of SIS peptides as internal standards, we used four analyte-specific criteria for identify-ing chromatographic peaks that correspond to the authentic analyte peptide O precursor ion m/z, product ion m/z, chromatographic retention time, and the relative product ion intensities. Analyte peptides and their SIS analogs have identical chromatographic retention time because of their physiochemical similarities. Under the same collision energy and collision-activated dissociation pressure, the analyte and the SIS peptide dissociate to generate the same pattern of product ions, which differ only by the mass difference produced by the incorporation of the stable isotope-labeled amino acids and the difference in their absolute intensity. Importantly, the relative intensities of the complement of product ions formed by each analyte peptide and its SIS analog are nearly identical. These features were used for evaluating the selectivity of SRM assays and the certainty of peak assignment. For this purpose, we chose four to five SRM pairs for each peptide, and manually examined the MS raw data to ensure that the analyte peptides and their SIS analogs had same chromatographic retention time (variance below 0.05 min) and relative product ion intensities (Ϯ20% variance in the relative ratios for each fragment). This inspection ensured that selected SRM ion pairs of both analyte peptides and SIS peptides were free of matrix interference from co-eluting ions. As shown in Fig. 8, RelA natural peptides had nearly identical chromatographic retention time and relative product ion intensities as their SIS analogs, indicating that the SRM assays had high specificity. We believe that the high specificity of the assay was achieved, in part, because the aptamer enrichment significantly reduces the sample complexity and our application of high resolution-SRM mode for data acquisition. Therefore, for quantification of each peptide, we used the sum of the signals from all monitored transitions instead of selecting a single representative SRM ion pair for each peptide. A solution of SIS peptide of RelA, TPPYADPSLQAPVR, was nanoinfused onto a TSQ mass spectrometer, and the optimal collision energy for each transition were obtained by performing a breakdown curve. The vertical dashed lines represent the calculated collision energy voltage using a generic formula (ϭ 0.034 ϫ (precursor ion m/z) ϩ 3.314). Note that the experimentally derived collision energies give increased sensitivity over those determined by the generic formula.

Linear Response and Limit of SRM Quantification of RelA-
The assay dynamic range represents the concentration range where protein abundance is measurable with acceptable levels of accuracy and precision. The assay dynamic range was assessed by the method of standard addition (41)(42)(43). In this method, serial dilutions of target proteins with a fixed amount of SIS peptide are spiked into the similar cellular protein matrix, and a calibration curve is generated for target peptide. However, because a blank cellular protein matrix devoid of activated endogenous RelA was not available, we alternatively used a [ 12 C]-analyte peptide of RelA with its native flanking sequences (VFR-TPPYADPSLQAPVR-VSM), to characterize the assay dynamic range. The [ 12 C]-analyte peptide was digested with trypsin in the same matrix as used for the "onbeads" digestion of aptamer-enriched RelA. We diluted the tryptic digest to generate a range of analyte concentrations spanning a 10,000-fold concentration range (from 200 amol to 20 pmol on the column). These analyte concentrations were then combined with a constant amount of [ 13 C]-analyte peptide internal standard. Two replicate LC-SRM-MS analyses of each sample dilution were performed in the order from the most dilute to the most concentrated. Linear regression analysis was performed on the observed peak area ratios (natural: heavy) versus concentration ratios to generate calibration curves (Fig. 9). The SRM assay yields linear responses over a Ͼ1000-fold concentration range (r Ͼ 0.9971). When the amount of RelA peptide loaded on the column was higher than 10 pmol or lower than 500 amol, the response became nonlinear (Fig. 9). The nonlinearity on the high concentration end of the dilution curve is due to the saturation of both the MS detector and the nano HPLC column, whereas the nonlinearity on the low concentration end is mainly because of matrix interference and the trace amount of [ 12 C]-analyte peptide of RelA contaminating the [ 13 C]-analyte peptide internal standard.
The lower limit of quantification of the SRM assay is defined as the lowest analyte concentration that can be measured with Ͻ20% coefficient of variation (44,45). In our study, the lowest natural RelA peptide that can be quantified is 200 amol (signal-to-noise ratio (S/N) 85, 10.2% coefficient of variation), Aptamer Enrichment Increases Signal-to-Noise (S/N) Ratio for RelA Quantification-Using RelA-specific P028F4 to affinity enrich RelA before SRM-MS analysis dramatically reduces the sample complexity. To determine the effect of pre-enrichment on the S/N ratio, we measured the S/N ratio for endogenous RelA in crude versus P028F4-enriched samples. We found that the S/N ratio of endogenous RelA is only 12 when analyzed directly from crude CE (Fig. 10), whereas P028F4 enrichment resulted in a 36-fold increase in S/N ratio (from 12 to 429, Fig. 10).
Multiplexed LC-SRM-MS Analysis of Single-step Aptamerenriched RelA-We next combined single-step aptamer-enrichment of RelA in eukaryotic cells with the multiplex SRM assay (schematically shown in Fig. 11). Control or TNF␣stimulated A549 cells subcellular fractions were separately incubated with Bt-P028F4 or P028A1 aptamers, and captured on SA-magnetic beads. To first determine whether the aptamer capture on SA-magnetic beads was nonsaturating, RelA from increasing numbers of control-or TNF␣-stimulated A549 cells were subjected to single-step aptamer enrichment, and detected by Western immunoblot. As seen in Fig. 12, increasing amounts of RelA were seen over the input range of 3-12 ϫ 10 6 cells only in the TNF␣-stimulated extracts.
To quantify the amount of RelA in cells, CE and NE from 6 ϫ 10 6 cells were prepared and assayed for cytoplasmic and nuclear markers. As seen in Fig. 13A, CE was highly enriched in ␤-tubulin, whereas ␤-tubulin was undetectable in NE. Similarly, histone H3 was undetectable in CE and highly enriched in NE. The active RelA in CE and NE was then captured by Bt-aptamer-SA magnetic beads. The proteins were trypsinized on-beads, and RelA and IB␣ were quantified with SID-SRM-MS. To determine the fraction of activated RelA complex, total RelA in each subcellular extract was also quantified with SID-SRM-MS. As shown in Fig.  13B, P028F4 demonstrated high specificity toward activated RelA in comparison to P028A1, resulting in 14 -30fold enrichment of activated RelA in CE and NE. With the quantification of SID-SRM-MS, we found that, in unstimulated cells, about 5% of total RelA proteins in the cytoplasm were activated, and only a small amount of activated (and total) RelA was in the nucleus. By contrast, after TNF␣ stimulation, the amount of activated RelA in CE increased about sixfold. We also observed a significant amount of activated RelA translocated into the nucleus, where the amount of activated RelA was the same as the total amount of nuclear RelA. Although it was measured, IB␣ was not observed in any of the aptamer-enriched samples, a finding that is consistent with the lack of RelA-IB␣ complex binding seen in Western blot analysis (Fig. 6B), and in EMSA (Fig. 5). DISCUSSION Mucosal inflammation produced by the innate immune response is dependent on the NF-B pathway, an intracellular signaling network whose activity is regulated by the formation of dynamic protein interactions. Activated pattern recognition-or cytokine receptors binding to their cognate ligands induce Lys 63 linked polyubiquitination of the (receptor-interacting serine/threonine-protein kinase) and (TNF-receptor associated factor) adapter molecules (6). These polyubiquitinmodified adapters recruit IKK to the submembranous receptor complex, producing serine phosphorylation and activation of IKK, an event that controls protein stability of IB␣ and nuclear translocation of activated RelA. In the nucleus, RelA associates with the p300 histone acetyltransferase, whose promoter recruitment activates expression of downstream inflammatory genetic networks. As a result of direct activation of negative feedback networks including IB␣ syn- thesis, RelA undergoes nuclear-cytoplasmic oscillation in response to activation (46). In this study, we describe the selection and characterization of a RelA-directed ssDNA oligonucleotide, P028F4, a ligand that directly binds recombinant RelA in vitro, and binds inducible RelA in isolated NEs. P028F4 enriches for serine phosphorylated and p300-com-plexed RelA. We provide evidence that P028F4 has utility as a method for single step enrichment and quantification of activated RelA complexes using quantitative SID-SRM-MS assays. This method has, for the first time, allowed the determination of the fraction of activated RelA in unstimulated cells and the estimation of activated RelA molecules.
Genetic selection of protein-binding aptamers is a widely exploited method for probing protein function and structure (20,21). Previous efforts isolating RelA ligands have resulted in isolation of ssRNA aptamers and ds-thioate-and dithioate modified ligands that bind NF-B/RelA (47,48). To our knowledge, this work represents the first report of isolation of a ssDNA aptamer that specifically recognizes NF-B/RelA. Our data show that P028F4 binds with a K D of 6.4 ϫ 10 Ϫ10, representing about 10-fold higher affinity than that of the reported ssRNA aptamers (47). Moreover, this binding is more stable than that described for duplex DNA binding to RelA. Although previous work showing that thioate modified duplex aptamers can recognize specific NF-B isoforms, their on and off-rates have not been reported (48). One property of P028F4 is that this ligand binds with a long off-rate, measured by surface plasmon resonance with a kd of 0.0018 1/s. This characteristic lends itself well to its application for affinity isolations of the target protein and its complexes.
The crystallographic structure of RelA complexed to its duplex binding sites shows that RelA binds as a dimer, as-  FIG. 11. Schematic diagram of the SRM assay of aptamer-enriched RelA and its associated proteins. First, biotin-conjugated RelA specific aptamer was incubated with crude cell extract. Second, the Bt-aptamer RelA was captured by SA-coated magnetic beads. The beads were washed six times to remove nonspecifically bound proteins. The washing improved the purity of isolated RelA and reduced the complexity of the sample. Third, the beads were resuspended in trypsin proteolysis buffer and the proteins bound on the beads were digested with trypsin. Fourth, the tryptic peptides were eluted from the beads and the SIS standard peptides were spiked before SID-SRM-MS analysis. Finally, RelA was quantified with SID-SRM-MS. suming an immunoglobulin fold structure with ␤-stranded loops making extensive contract with both DNA base pairs and the sugar-phosphate backbone (49). Our binding characterization shows that ssDNA P028F4 competes with RelA for its cognate dsDNA regulatory sequences. These data may suggest that P028F4 binds RelA on a ␤-stranded loop that is essential for stable DNA binding, or alternatively, P028F4 may induce RelA to assume a conformational change nonpermissive for DNA interaction. In this regard, we note that a high affinity ssRNA aptamer that selectively binds NF-B p50 isoform also competes with duplex DNA binding by inducing an alternative molecular conformation of the protein (50).
Because P028F4 is a competitive antagonist of activated RelA DNA binding activity, we explored its properties as an antagonist of RelA transcriptional activity in intact cells. Although we observed that cellular transfection of P028F4 produced additional inhibition of expression of NF-B dependent target genes over that of the control P028A1 aptamer, our experiments also indicate that at the concentrations needed to overcome nonspecific cellular binding (28), there is a nonspecific aptamer effect on TNF␣inducible transcription (Fig.  4A). In the ChIP experiments, an even stronger nonspecific aptamer effect was also seen (Fig. 4B). We ascribe these nonspecific P028A1 effects as being due to well-recognized low affinity interactions of cellular transcription factors for nonspecific DNA, and is consistent with its weak abilities to compete for RelA binding in nitrocellulose filter binding assays (Table II), and capture RelA in aptamer-affinity binding studies (Fig. 6A). This nonspecific effect will need to be reduced before ssDNA aptamers like P028F4 can be used to modify RelA action in intact cells.
An important feature of P028F4-RelA interaction, exploited in this study, is that IB␣-complexed RelA is not recognized by P028F4. We base this conclusion on the inability of P028F4 to bind to purified recombinant IB␣complexed RelA (Fig. 5), its colocalization with activated EGFP-RelA in isolated nuclei (Fig. 3A), its binding to posttranslationally modified RelA in cellular extracts (Fig. 6B), and its binding enrichment of RelA (over that of IB␣) in affinity isolations measured by quantitative SRM-MS assays. Previous work has shown that IB␣ binds the NH 2 terminus of RelA with a 1:1 stoichiometry on discontinuous sequences through its seven ankyrin repeat domains. X-Ray crystallographic analysis of the IB␣⅐RelA complex shows that first two ankyrin repeats bind (and mask) the RelA nuclear localization signal, and the sixth repeat occludes the RelA DNA-binding cleft (51). Through these interactions, IB␣ is able to both inhibit RelA nuclear translocation as well as its DNA binding activity. More work will be required to map the location of RelA binding by P028F4 and determine its effect, if any, on protein conformation.
Although DNA aptamers have been used as affinity reagents for identifying biomarkers in complex fluids (52), they have not yet been fully exploited in protein quantification. The approach we demonstrate here combines aptamer-enrichment of RelA and SRM-MS quantification using stable isotopic dilution methods. Compared with direct quantification of RelA and associated proteins from crude cell extracts, the aptamer-based enrichment approach has several advantages. First, using a RelA-specific aptamer to enrich RelA and its associated proteins before SRM-MS analysis dramatically reduces the sample complexity and hence improves the S/N ratio of endogenous RelA 36-fold. Second, aptamer enrichment also reduces the likelihood of the interference of other analytes that have isobaric (or very similar m/z values) to the target peptides. Therefore the specificity and accuracy of SRM assay are significantly improved. Third, aptamer enrichment allows RelA to be isolated from a large amount of protein extract so that the amount of endogenous RelA peptide falls in the middle of the linear dynamic range of the calibration curve, which improves the accuracy and reproducibility of quantification. This feature is especially attractive for quantification of low-abundance proteins, such as RelA or p300 transcription factors, in complex protein mixtures. Finally, because P028F4 selectively recognizes the IB␣-free, activated form of RelA, we are able to quantify the abundance of the activated form of RelA. This measurement cannot be accomplished by directly analyzing RelA abundance in crude cellular (or subcellular) extracts.
Described earlier, RelA is a cytoplasmic transcription factor that undergoes constitutive nuclear-cytoplasmic transport. In the unstimulated cell, RelA is sequestered in the cytoplasm predominately by binding the IB␣ inhibitor, an association that masks the nuclear localization sequence, shifting the nuclear-cytoplasmic equilibrium of RelA to the cytoplasmic compartment. Stimulus-induced RelA activation is mediated by IB␣ proteolysis, liberating a fraction of sequestered RelA to translocate into the nucleus, shifting the equilibrium to the nuclear compartment. The theoretical understanding of NF-B regulation has been developed using computational modeling approaches, where the relative contributions of the IB isoforms and the effect of the negative feedback loops have been inferred (53,54). Despite this understanding of pathway connections, the dynamics of negative feedback loops are dependent on the fraction of activated RelA that enters the nucleus. This value is based on assumed values because the precise concentration of activated RelA has not yet been experimentally determined.
This study represents the first quantitative estimate of the number of activated RelA molecules within a cell. The aptamer-SID-SRM estimation of ϳ200,000 total molecules of RelA/cell with 50,000 molecules/cell of activated RelA in the nucleus after TNF␣ stimulation are in good agreement with our previous quantitative Western blot experiments. We also note that the amount of activated RelA is equivalent with the amount of total RelA in both control and TNF␣ stimulated nuclei. These findings indicate that the majority of RelA present in the nucleus is in the activated state, consistent with our previous findings by Western immunoblots that the amount of nuclear IB␣ is very low. The estimate we make here for the number of activated RelA molecules in CE has not been possible without the development of P028F4, and its ability to discriminate between the IB␣-bound and free RelA fractions. We find it interesting that the number of copies of activated RelA in the cytoplasm and nucleus in stimulated cells are similar-a finding that may have implications in understanding the energetics controlling nuclear RelA import.
In conclusion, RelA aptamer-SID-SRM is a versatile tool for quantification of the abundance of activated protein and its complex. Aptamer enrichment enhances assay specificity, S/N ratio, and sensitivity. Further exploitation of this approach will allow for systematic extension of the dynamics of RelA activation in response to diverse ligands, and exploration of protein interaction networks of the molecule in its activated form.