Mass Spectrometry Handbook

SECTION I

BIOTECHNOLOGY/PROTEINS

1

TARGETED PROTEOMICS USING IMMUNOAFFINITY PURIFICATION

KAREN R. JONSCHER, LEI JIN, JOHN C. CAMBIER, SHAIKH M. RAHMAN, AND JACOB E. FRIEDMAN

1.1 INTRODUCTION

Proteins are multimodular and multifunctional, interacting in complex networks that drive cellular function. Pathological alterations in signaling networks are thought to result in a number of diseases, particularly cancer. Understanding the roles and consequences of protein–protein interactions is therefore a fundamental goal in systems biology. The two-hybrid approach [1] emerged in the early 1990s as the first method to assay whether two proteins interact in a pair-wise fashion. A number of bait–target strategies were subsequently developed [2], including techniques exploiting affinity purifications coupled to mass spectrometry (MS) to rapidly identify potentially novel protein interactions. Initial studies were performed in yeast [3,4] and were subsequently expanded to mammalian models [5]. Since MS-based proteomics is not necessarily limited to specific sites or to specific proteins, it represents an unbiased and direct approach to studying cellular processes [6].

As recently described [6,7], immunoaffinity purification has emerged as the most frequently employed method for multiprotein complex purification. Its success is based on the principle that multiple members of a complex may be captured when one complex member is enriched, regardless of whether the complexed proteins are directly bound to the target protein. Additionally, purification of posttranslational modifications has been used extensively to globally profile modified proteins throughout cellular networks [8,9] and provides invaluable insights into signal transduction mechanisms.

A summary of typical steps employed to generate samples using an immunoaffinity-based approach is illustrated in Figure 1.1 and described in detail for two example applications below. Following purification, peptide mixtures resulting from the digestion of bands or eluates are analyzed using tandem mass spectrometry (MS/MS) and proteins are identified by database searching and spectral matching.

FIGURE 1.1 Flowchart of sample preparation steps for immunoaffinity purification workflows. Tissue is homogenized or cells are lysed in an appropriate buffer using protease inhibitors. Whole cell lysates may be used, or subcellular enrichment of organelles, membrane, or cytoplasmic fractions, for example, may be employed for subsequent processing. Primary antibody recognizing the protein or modification of interest is either bound to protein A- or protein G-coated beads or added to the lysate, allowing either target proteins or modifications to bind. Bead-bound protein A or protein G interacts with the primary antibody, enriching the protein complexes through a series of washes. Finally, proteins are eluted from the beads. The protein mixture, also containing the antibody, is either directly digested with an enzyme or separated by 1D gel electrophoresis, with the protein-containing bands of interest subsequently excised and digested.

A gel-based approach may be useful when two conditions are being compared—bands exhibiting visual differences can be excised to yield data most likely to contrast biologically significant results (note that interesting low abundance proteins may be covered up by more abundant nonspecific proteins). Another useful method, gel-enhanced liquid chromatography–tandem mass spectrometry (GeLC-MS/MS), has also emerged for the analysis of complex protein mixtures [10] and can be applied to the separation of immunoaffinity eluates. In this approach, a protein-containing gel lane is chopped into equivalent sections, digested, and peptide mixtures analyzed. When complexed protein levels are extremely low or sample is limited, elution followed by in-solution digestion may provide a better option, as less protein is lost to sample handling. An important caveat to note is that a protein of interest may be covered up by comigrating background or nonspecific proteins. This is a particular concern for proteins that may comigrate with immunoglobulin (Ig) heavy and light chains. Using a cross-linker such as DMP (dimethyl pimelimidate) to bind the primary antibody to Protein A will suppress elution of Ig chains when nonreducing conditions are used for elution. However, cross-linking may result in loss of affinity; an optimization workflow should ideally include both cross-linked and noncross-linked trials.

The primary challenge of immunoaffinity-based workflows lies in the difficulty of separating true low abundance interactors from nonspecifically binding proteins. Use of negative controls, such as preimmune sera or antibodies against other proteins, or, if the model allows, using a knockout or knock down of the protein of interest, can help separate out these background proteins. As described below, cross-linking the antibody, minimizing incubation times and antibody concentrations, optimizing wash buffer stringency, and other approaches may help mitigate the extent of nonspecific binding. Assessing the utility of at least a few of these parameters should be included in the optimization workflow. A good way to begin optimizing the protocol is to immunoprecipitate the protein of interest and probe for a known interactor using Western blot. Begin by titrating the primary antibody and beads to find the minimum amount required to effectively immunodeplete the sample. Then experiment with incubation times. Fewer beads and shorter incubation times will help reduce nonspecific binding. Ultimately, orthogonal techniques such as co-immunoprecipitation (IP) with Western blot should be used to validate a subset of the potential interactors, whenever possible.

In this work we present examples of workflows in which immunoaffinity-purified proteins were either separated using gel electrophoresis and bands exhibiting significant change from control were analyzed, or complexed proteins were eluted from the beads, digested in solution, and analyzed. It is important to note that these protocols provide general guidelines and that several optimization steps with multiple iterations of MS will likely be required for purification of a protein complex of interest.

As discussed in more detail below, data analysis and mining are critical for gleaning relevant information from proteomics studies. Careful extraction of peptide signals, determination of properly stringent search engine parameters [11], and reversed or scrambled database searching leads to an output data set where significance of the identifications may be established with score or probability cutoffs. Although a false discovery rate of ∼1% is often employed in more global approaches [12], establishing criteria for two peptide hits to a protein with peptide probabilities of ∼95% is sufficient to provide a false discovery rate approaching 0% for immunoaffinity purification applications. Following identification, data mining is employed to obtain functional information about the proteins to begin to decipher mechanisms that may be triggered by the interaction. The tools used often include those developed for microarray analysis, where gene ontology information is used to cluster proteins with similar cellular compartments, functions, or pathways [13–16]. Downstream assays based on these results, including IP of proteins identified in the complex, allow investigators to begin to elucidate mechanisms driving cellular function.

1.2 EXPERIMENTAL PROTOCOLS

Materials and Solutions

Cell Lysis Buffer. 0.33% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS); 150 mM NaCl; 10 mM sodium pyrophosphate; 10 mM Tris-HCl pH 7.4; 1 mM phenylmethylsulfonyl fluoride (PMSF); 0.4 mM ethylenediaminetetraacetic acid (EDTA); 1.8 mg/mL iodoacetamide (IAA); 10 mM NaF; 2 mM Na3VO4; and 1 µg/mL each of aprotinin, leupeptin, and pepstatin.

Tissue Lysis Buffer. Buffer A (10 mM HEPES pH 7.9, 1.5 mM KCl, 10 mM MgCl2, 0.5 mM dithiothreitol (DTT), 0.1% IGEPAL CA-630 (Sigma-Aldrich, St. Louis, MO), and 0.5 mM PMSF) and Buffer B (20 mM HEPES pH 7.9, 25% glycerol, 1.5 mM MgCl2, 420 mM NaCl, 0.5 mM DTT, 0.2 mM EDTA, 0.5 mM PMSF, and 4 µM leupeptin). (Note: These buffers were selected for the analysis of liver proteomes, as described below. Other lysis buffers may be more appropriate for the sample/proteins under investigation. A search of recent literature should provide some direction regarding appropriate buffer selection.)

Reagents for preparation of magnetic beads as described below.

Protein A or protein G beads. (Note: Either Sepharose [GE Healthcare, Piscataway, NJ] beads or magnetic beads may be used; however, magnetic beads are preferred as they tend to exhibit less nonspecific binding than Sepharose beads.)

Pipette tips cut 5 mm from the top (will avoid damaging the beads if using Protein A/G Sepharose beads).

Laemmli buffer.

High performance liquid chromatography (HPLC)-grade water (Honeywell Burdick and Jackson [Morristown, NJ] or other high quality liquid chromatography–mass spectrometry [LC/MS]-grade water).

Desalt spin columns.

Reagents for in-gel or in-solution digestion as listed below.

Equipment

Refrigerated centrifuge

Rotary mixer

Vacuum centrifuge

Gel electrophoresis apparatus and 10% polyacrylamide gel

Nanoscale HPLC, tandem mass spectrometer

Lysis

Note that the cell numbers and tissue amounts presented here are a guideline. These numbers should be increased if complexed proteins are of low abundance.

Cell Lysis

1. Lyse ∼5 × 10⁸ cells using 500–1000 µL cell lysis buffer at 4°C overnight. Note that if low abundance or weakly interacting proteins are to be analyzed, increase the cell numbers to as much as 10¹⁰ (as shown by Malovannaya and coworkers [17]).

2. In the morning, centrifuge lysates at 12,000 × g at 4°C for 20 min. Remove supernatants to a clean tube.

Tissue Lysis

1. Homogenize ∼100 mg of tissue using a mortar and pestle over liquid nitrogen. For identification of potentially weakly binding complexes, increase tissue amount to 10–20 g (following Moresco et al. [18]) and increase lysis buffer volume to 5–10 mL.

2. Add homogenized tissue to 0.5 mL of tissue lysis buffer A (ice cold) and ultrasonicate three times, 15 s each. Place sample tubes in an ice bath for at least 1 min between sonications.

3. Incubate samples on ice for 30 min, then centrifuge at 14,000 × g, 4°C for 10 min. Remove the supernatant (cytoplasmic fraction) to clean tube.

4. Resuspend the membrane/organelle fraction pellet in 0.2 mL ice-cold tissue lysis buffer B and incubate on ice for 30 min. Following centrifugation at 14,000 × g for 30 min at 4°C, remove the supernatant to clean tube.

Note: For all subsequent steps, be sure to keep samples on ice or at 4°C, however freezing lysates before the immunoaffinity purification should be avoided [18]. If necessary, store lysates at −80°C prior to use.

Total Protein Quantification

Use either the Bradford or bicinchoninic acid (BCA) method to quantify total protein concentrations following the manufacturer’s instructions for a microwell plate assay. Make sure that the lysis buffer components are compatible with the manufacturer’s stated levels. Try several dilutions to ensure the sample concentration is within the linear range of the assay.

Immunoaffinity Purification

Immunoaffinity purification may be accomplished using either soluble antibodies or antibodies cross-linked to beads. Generally the first step of the optimization should be done using cross-linked antibodies; cross-linking significantly reduces contaminating signals from Ig light and heavy chains. Procedures for both approaches are provided below.

Immunoaffinity Purification Using Magnetic Beads

Reagents for Magnetic Bead Preparation (Dynabeads Are Typically Used)

Citrate Phosphate Buffer, pH 5.0. 25 mM citric acid, 50 mM sodium phosphate (Na2HPO4)

0.2 M Triethanolamine (TEA), pH 8.2. 3.71 g triethanolamine-HCl/100 mL water

20 mM DMP. 5.4 mg DMP-2HCl per milliliter of TEA buffer

50 mM Tris pH 7.5

PBS-T. 0.01% Tween-20 (Thermo Fisher Scientific, Waltham, MA) in phosphate buffered saline

0.1 M glycine pH 2.5–2.7

Storage Solution. PBS-T with 0.02% sodium azide

Dynabeads (Invitrogen Corp., Carlsbad, CA) are packaged as a 5% slurry. Prepare 0.5–1.0 mL of slurry to obtain 25–50 µL of packed beads. A rule of thumb is that 1 mL of slurry binds ∼300 µg of antibody. Incubate with about 400 µg of the primary antibody.

Equilibrate Dynabeads

Centrifuge beads briefly, place tube in a magnetic rack, and remove the supernatant. Add 1 mL citrate phosphate buffer, vortex, spin briefly in a minifuge (1 s to remove bead solution from cap), and place tubes in a magnetic rack. Remove supernatant. Repeat two more times.

Incubate with Primary Antibody

Prepare 400 µg of primary antibody in 1 mL citrate phosphate buffer and add to beads. Reducing the volume may improve binding and may be included in subsequent optimization steps. Rotate tube end over end for 2–3 h at room temperature.

Wash

Centrifuge briefly, place tubes in a magnetic rack, remove supernatant, and add 1 mL citrate phosphate buffer and wash three times as described in the section Equilibrate Dynabeads. Wash two times more with 1 mL 0.2 M triethanolamine-HCl.

Cross-Link

Remove final TEA wash from the beads using the magnet and add 1 mL of DMP solution. Incubate 30 min at room temperature, rotating end over end.

Clean Up

Using the magnet, remove the DMP solution and incubate beads with 50 mM Tris for 15 min to remove free cross-linking reagent.

Wash

Wash beads three times with PBS-T.

Remove Free Antibody

Incubate the beads with 0.1 M glycine for 5 min, rotating end over end at 4°C.

Wash

Wash beads three times with PBS-T.

Store Beads

Bring beads back to original packaged volume in storage solution for storage at 4°C. For a 1 mL stock, use 950 µL of storage solution.

To use, wash beads three times using 1 mL PBS, then three time with 1 mL lysis buffer. Use Western blots and bead titration to determine the minimum amount of beads to use and the minimum amount of time to incubate. A starting point for optimization may be 20 µL of packed beads and 2 h of incubation at 4°C. Elute as described below.

Immunoaffinity Purification Using Protein A/G Sepharose Beads

Incubation with Primary Antibody

Add 3–20 µg of primary antibody to the supernatant and incubate for 2 h to overnight at 4°C. This step should be optimized to use the minimal time and generate the least amount of nonspecific binding. Increasing the incubation time generally increases background and fragmentation due to endogenous protease activity. Increased antibody concentrations are required when using more cells or when the total protein concentration is high.

Note: In order to reduce nonspecific binding of protein aggregates during incubation, ultracentrifuge samples at 100,000 × g for 20 min following incubation and limit incubation time to 2 h. Remove the supernatant to a clean tube. Some groups recommend that the lower 0.1 mL of lysate not be used [17]. Using magnetic beads instead of Sepharose beads may also help.

Prepare Protein A/G Sepharose Beads

Remove 100 µL of bead slurry for each sample (mix the slurry well prior to removing the beads). Centrifuge at 1500 × g at 4°C for 1–2 min, aspirate the supernatant, and wash with 10× bead volume of cell lysis buffer three times.

Bind to Sepharose Beads

Remove sample to the tube containing the washed beads and mix end over end at 4°C for 4 h. Centrifuge at 1500 × g at 4°C for 10 min and remove supernatant to clean tube.

Wash

Wash beads up to five times with 10× bead volume of cell or tissue lysis buffer by adding buffer, centrifuging at 1500 × g at 4°C for 1 min, and removing supernatant. (Note: To retain weakly interacting proteins, use a low stringency wash buffer, such as PBS or 0.5% NP-40, add 10× bead volume, and briefly invert the tube 10 times [17]. Radioimmunoprecipitation assay [RIPA] buffer may also be used to retain more strongly interacting proteins.)

Elution

Proteins may be eluted from the beads either for subsequent gel electrophoresis and in-gel digestion or for in-solution digestion.

Gels

For applications involving gel electrophoresis, add one bead volume of 2× Laemmli buffer to the bead pellet, heat at 60°C for 10–15 min, centrifuge, and remove the supernatant to clean tube. To ensure complete elution, add an additional 10 µL of Laemmli buffer to the bead pellet, repeat, and pool supernatants.

Solution

For in-solution digestion applications, elute the proteins from the antibody-coupled Sepharose beads using one bead volume of a strong acid such as 0.2 M citric acid (pH 2.0), 1% formic acid or 0.5% trifluoroacetic acid. If using citric acid, immediately add an equivalent volume of 2 M Tris to neutralize the pH prior to digestion. Alternatively, a strong and volatile base such as ammonium hydroxide can be used, or even 8 M urea. The urea will not effectively elute the protein target but will disrupt interacting proteins. Perform three similar elutions and combine the eluates. Concentrate to dryness in a vacuum centrifuge. If 8 M urea is used for elution, proceed to the protocol for in-solution digestion.

Note: The elution should be optimized by checking at least two to three different elution solvents. Eluates may be assessed using a one-dimensional (1D) sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gel stained with colloidal Coomassie.

Gel Electrophoresis (for Subsequent in Gel Digestion)

Separate Proteins

Using samples generated in the section Gels, load the proteins onto a 1D 10%–14.5% SDS-PAGE gel and separate them using 120–180 V. A 13.3 × 8.7 cm × 1 mm thick gradient gel generally provides sufficient separation. If more separation is required, a larger gel can be used. A 15% gel should be used for resolving smaller proteins, while a 7%–10% gel is appropriate for investigating larger proteins. All of these types of gels may be useful for the initial stage of optimization for separating the proteins of interest. A lower separating voltage may provide more resolution.

Stain Proteins

Following electrophoresis, carefully wash the gels twice with ultrapure water. Be careful to handle the gels only by the edges. Some commercially available ready-made gels have a thick bottom edge which helps prevent the gels from ripping when they are rinsed. Use either colloidal Coomassie blue (such as EZ Blue, Sigma-Aldrich) or a fluorescent stain and follow the manufacturer’s directions for staining and fixing. Oriole™ fluorescent gel stain (BioRad Corporation, Hercules, CA) is an easy-to-use, sensitive, and highly linear protein gel stain. This stain can be used without prior protein fixation or destaining. For larger protein loads (e.g., 10 mg total protein), use a standard Coomassie blue stain and follow destaining procedure below.

Visualize Proteins

Image the gel using a laser imager capable of exciting and measuring emission at wavelengths appropriate for the stain employed. For example, Oriole has an excitation maximum at 270 nm and emission at 604 nm, thus it is appropriate for UV-based imagers. Coomassie can be imaged using a standard visible light gel scanner.

Excise Protein Bands

Excise gel bands of interest using a scalpel or a cut pipette tip. Dice the gel slice into 1 mm³ cubes. Place into a 1.5 mL microfuge tube and remove liquid. Gels may be stored at −80°C prior to processing.

Sample Preparation for MS

In-Gel Digestion

Reagents

It must be noted, use only high quality HPLC-grade water.

50 mM Ammonium Bicarbonate (NH4HCO3). 4 mg/mL in water.

Destain Reagent. 1:1 acetonitrile (ACN) in 50 mM NH4HCO3.

Reducing Reagent. Prepare a 1.5 mg/mL solution of DTT in 50 mM NH4HCO3. DTT is unstable in solution at room temperature so prepare fresh daily.

Alkylating Reagent. 10 mg/mL solution of IAA in 50 mM NH4HCO3. This buffer should also be prepared fresh daily (while reducing) and kept in the dark.

Trypsin. Use sequencing grade modified trypsin. Promega (Madison, WI) is widely used.

Destain

(This step will remove stain that might interfere with proteolytic digestion. Note that this step is not required if using Oriole stain. Follow manufacturer’s directions for the stain utilized.)

1. Remove any liquid from the microfuge tube. If the gel band was not previously diced, use a sealed pipette tip to cut gel slice into small pieces.

2. Add enough destain reagent to cover gel pieces. Shake or sonicate at room temperature for 10 min. Discard liquid. If gels were stained using a fluorescent stain, go to the section Reduction/Alkylation (for 1D Gel Bands).

3. Repeat the second step until gel pieces are clear. Usually two to three washes are necessary for Coomassie stained gels. If gel pieces still have blue color, rehydrate by adding 50 mM NH4HCO3 and shake for 10 min at room temperature. Discard liquid and repeat.

Reduction/Alkylation (for 1D Gel Bands)

(This step allows for denaturing and separation of complexed proteins)

1. If not already dehydrated, dry gel pieces in vacuum centrifuge. Transfer the dried slices to clean 0.6 mL tubes.

2. Add enough of the reducing reagent to fully cover gel pieces, taking into account that the pieces will swell. Usually 20–50 µL is sufficient.

3. Incubate at 37°C for 1 h. During this time, prepare the alkylating reagent.

4. After incubating, remove excess reducing reagent and cool to room temperature (this can be done by placing samples at −20°C for 5 min). Add enough alkylating reagent to cover gel pieces. Place tubes in the dark and shake gently for 45 min.

5. Discard supernatant. Wash gel pieces with 50 mM NH4HCO3 for 10 min with shaking at room temperature. Discard supernatant.

6. Wash gel slices twice more using destaining reagent (10 min with shaking at room temperature and each time discarding supernatant).

7. Dehydrate gel pieces by vacuum centrifugation. Transfer the dried slices to clean 0.6 mL tubes.

8. Trypsin digestion: Add enough of a 0.2 µg/µL trypsin solution to swell the gels and incubate on ice or at 4°C for 15 min. Usually 20 µL is sufficient. The trypsin solution can be prepared and aliquoted ahead of time (20 microfuge tubes of 10 µL each) and stored in the freezer until ready.

9. Remove excess trypsin solution. Add 50 mM NH4HCO3/10% ACN, enough to cover the gel pieces, but not excess. Place in 37°C incubator.

10. Check pieces in 20 min, adding enough 50 mM NH4HCO3/10% ACN to keep pieces just covered.

11. Digest overnight (∼19 h) at 37°C.

Extracting Tryptic Fragments

1. Following digestion, remove microfuge tubes from incubator. Add 1 µL of formic acid to tubes and sonicate or shake at room temperature for 15 min. Centrifuge and remove supernatant to clean tube.

2. Add enough 1.0% trifluoroacetic acid/60% ACN to cover gel pieces. Shake or sonicate for 10 min at room temperature, centrifuge, and remove supernatant and combine with supernatant from previous step.

3. Vacuum centrifuge the supernatant extract until concentrated to ∼10 µL. If extract goes to dryness (not good), add 5 µL of 1.0% formic acid and vortex vigorously at room temperature.

4. Store digests at −70°C prior to analysis.

In-Solution Digestion

Note that a 100 mM Tris buffer is good to use if samples will be loaded onto a C18 enrichment desalting column prior to separation by nanoscale HPLC. Otherwise, use NH4HCO3.

Reagents

100 mM NH4HCO3. 7.9 mg/mL of HPLC-grade water (or 100 mM Tris, pH 8.5).

8 M Urea. Dilute 480 mg of urea in 1.0 mL of 100 mM NH4HCO3 solution (or 100 mM Tris buffer).

Reducing Reagent. Dissolve 3 mg of DTT in 20 µL of 100 mM NH4HCO3 solution to make 1 M DTT (or use 100 mM Tris buffer).

Alkylating Reagent. Dissolve 3.6 mg in 100 µL of NH4HCO3 or 100 mM Tris solution to make 200 mM IAA.

Trypsin Solution. Make up a 1 mg/mL solution of trypsin in HPLC-grade water or NH4HCO3 (or 100 mM Tris buffer) and add 1–2 µL to each sample.

Digestion Procedure

1. Reconstitute. Reconstitute sample in approximately 20 µL of 8.0 M urea in a 0.5 mL microfuge tube.

2. Reduce. Add 1 µL of reducing reagent and mix the sample by gentle vortex. Reduce the mixture for 1 h at room temperature or in an oven at 37°C. Do not go over 37°C or the urea will react with the sample and generate carbamylated artifacts. Allow the sample to cool to room temperature.

3. Alkylate. Add 20 µL of alkylating reagent and alkylate for 1 h at room temperature in the dark (use aluminum foil to cover the sample). Add 4 µL of reducing reagent to consume any leftover alkylating agent (so the trypsin is not alkylated).

4. Dilute. Add 60 µL of NH4HCO3 or 100 mM Tris solution to dilute the urea before digesting it with trypsin.

5. Digest. Add trypsin solution in appropriate ratio (1:30) to approximate amount of protein by weight. After 1 h, add another microliter of trypsin solution. Digest 4 h to overnight at 37°C.

6. Stop Digestion. In the morning, or following digestion, add 1 µL of 100% formic acid to the sample. Vortex and centrifuge. Freeze sample at −70°C prior to analysis. Use a vacuum centrifuge (Speedvac) for samples with relatively low protein concentrations to maximize signal and minimize loss of sample to the tube.

Mass Spectrometric Analysis

Note that the analyses of the samples presented here were performed using a high capacity quadrupole ion trap (LC/MSD XCT Ultra [Agilent Technologies, Santa Clara, CA]). Although the quadrupole trap was historically the proteomics workhorse, hybrid quadrupole/time-of-flight (QTOF) and Orbitrap instruments are capable of rapid scanning and high resolution of fragment ions, providing a performance advantage over the quadrupole traps [6]. Parameters presented should be viewed as general guidelines that may be modified and incorporated into proteomics workflows using alternative instrumentation.

Loading Samples

Load either 5 µL of in-gel digestion sample or 10 µL of in-solution digestion into a polypropylene autosample vial.

Establishing Flow

Use a 75 µm internal diameter (ID) nanoscale column with C18 packing for separating peptides. Flow rate should be approximately 200–300 nL/min. Adjust following manufacturer’s directions for ultra high performance liquid chromatography (UPLC) pumps.

Desalting

Samples should be desalted prior to loading on the separating column. This can be accomplished using a pipette tip desalting system (Glygen Scientific Products, Columbia, MD) or an online enrichment column with a switching valve.

Peptide Separation

Elute peptides from the separating column using a gradient from 3% to 40% buffer B (90% ACN, 0.1% formic acid). Buffer A is generally 0.1% formic acid. (Note: It is important to purchase high-quality formic acid and other reagents labeled as suitable for LC/MS applications.) For relatively simple mixtures, a gradient length of ∼40 min is generally sufficient. Gradient times may be reduced if UPLC is employed. For more complex mixtures, extending the gradient, as well as performing steps of intact protein fraction, enables the identification of more peptides and proteins. For an in-solution digest, a 2- to 3-h gradient should be sufficient. This step will need to be optimized for different sample types.

Acquiring Spectra

Utilize data-dependent acquisition, where the most abundant peptides are analyzed a limited number of times and then placed in an exclusion list for a given amount of time (e.g., 1 min). The data shown below were acquired by excluding the most abundant three to six peptides in a spectrum from subsequent analysis following three spectral acquisitions. Collect spectra over a mass-to-charge ratio (m/z) range of 350–1500 Da. (Note: The desired acquisition mass range will depend on the instrument employed. This mass range is appropriate when using a quadrupole ion trap.)

Database Searching

Proteomics workflows with MS/MS analysis provide extraordinarily data-rich results. A single digested band can easily result in over 10,000 fragmentation spectra. Therefore, careful analysis of the data must be performed to generate high-quality output protein lists for use in subsequent biological assays. Commonly used [19] database search algorithms include Sequest (Thermo Fisher Scientific), Spectrum Mill (Agilent Technologies), X!Tandem (The Global Proteome Machine Organization, http://www.thegpm.org/TANDEM), Mascot (Matrix Science, Inc., Boston, MA), ProteinLynx Global Server (Waters Corporation, Milford, MA), Phenyx (Geneva Bioinformatics, SA, Geneva, Switzerland), OMSSA (NCBI, http://pubchem.ncbi.nlm.nih.gov/omssa/), PEAKS (Bioinformatic Solutions, Inc., Waterloo, ON, Canada), ProteinPilot (AB Sciex, Framingham, MA), and Sequest Sorcerer (Sage-N-Research, Inc., Milpitas, CA), some of which are freely available (X!Tandem, OMSSA). Although each implementation is different, these search algorithms operate under the same general principles. These include establishment of the database and search space to be used and a statistical method of comparing experimental and theoretically generated fragmentation mass spectra, ultimately outputting a ranked score. Subsequently, users must determine criteria for valid score thresholds and extract potentially interesting results for follow-up.

Determining the Search Space

The search space can be limited by specifying known parameters including sample taxonomy, precursor and fragment ion mass tolerance, enzyme specificity, numbers of allowed missed enzymatic cleavages, and potential amino acid (AA) modifications. Limiting the search space significantly helps reduce false-positive results. Additionally, the selection of a more highly curated database, such as the UniProt Knowledge Base (http://www.uniprot.org/help/uniprotkb), may also help with reducing false positives. In the examples below, taxonomy was mus musculus, precursor ion tolerance was 1.7, and the fragment ion tolerance was 0.6. Trypsin was selected as the enzyme with two missed cleavages allowed. No AA modifications were included. The UniProtKB database was used, updated 04/2010, with 68,507 entries.

Selecting Potential Peptides to Search

First, peptide experimental mass is compared with the theoretical mass generated from an in silico digest of the protein database, and a subset of peptides with mass within the tolerance window selected in the search space are utilized for subsequent processing. Next, the fragment ion masses are compared with in silico fragmentation of the peptides (with masses within the defined mass tolerance) that passed the first criteria. Using a variety of different techniques, the algorithms assign a score or probability to the hit and generate a protein identification.

Interpreting the Score Data

The algorithms generate scores for all of the data within the tolerance window; however, some spectra may be falsely assigned. Generally, users select a score cutoff criterion and only consider data falling within the criterion window. Typically, a minimum of two unique peptides are required for each protein identification, although some groups have validated identifications from just one unique peptide [18]. Using reversed database searching and receiver operating curves may also help determine an appropriate cutoff. The use of data evaluation methods is not yet standardized; see Kapp and Schütz for an excellent discussion of these issues [19]. Due to the large number of spectra generated in a typical experiment it is impossible to manually inspect all of the data. However, search results from proteins that may be used in downstream biological assays should be manually validated.

Mining the Data

Proteomics data output results in long lists of proteins and scores, and it is often quite difficult to extract significant information to use for the development or testing of hypotheses. Initial insights are often obtained by searching the Gene Ontology (GO) annotations [13] for overrepresented terms. A comprehensive list of GO-based tools is available at http://www.geneontology.org/GO.tools.shtml. In the example below, we utilized The Database for Annotation, Visualization and Integrated Discovery (DAVID) v6.7 [20,21] to cluster genes with similar functions, families, or cellular locations to select a group of proteins with characteristics related to our hypothesis that were further investigated. New tools are constantly being developed and a regular review of Web sites and the literature is highly recommended.

1.3 APPLICATIONS OF THE PROTOCOLS

1.3.1 Analysis of the Acetylated Proteome of Mouse Liver in Obesity

Acetylation is now recognized as an emerging mechanism for controlling proteins mediating cellular adaptation to metabolic fuels [22] and is governed, in part, by sirtuins (SIRTs), class III NAD+-dependent histone deacetylases (HDACs) that regulate lipid and glucose metabolism in liver during fasting and aging. However, whether acetylation or SIRTs play a pathogenic role in fuel metabolism under conditions of obesity is unknown.

In the present study, 5- to 6-week-old male C57BL/6 SVJ mice were fed a high-fat diet (45 kcal% fat) or a standard chow diet for 16 weeks, and then were sacrificed by pentobarbital overdose following treatment. Livers were harvested immediately from anesthetized mice and snap frozen at −70°C in liquid nitrogen before analysis. Utilizing the protocol provided for preparation of tissue homogenates (∼200 mg liver tissue, 10 µg primary antibody, 50 µL packed beads), we prepared nuclear/mitochondrial and cytoplasmic fractions and enriched for lysine acetylated proteins using an antiacetyllysine polyclonal antibody. Complexed proteins were eluted from the beads by boiling in Laemmli buffer, separated using a 10% polyacrylamide gel, fluorescently stained with Lava Purple and imaged.

Labeled gel bands in Figure 1.2 that differed between mice fed normal chow (control) and a high-fat diet were excised, digested, and analyzed by MS/MS using a 40-min gradient of acetonitrile. Data were analyzed by LABKEY (http://www.labkey.com), an interface utilizing the X!Tandem search algorithm and elements of the Trans-Proteomics Pipeline, including PeptideProphet and Protein Prophet (these are freely available software packages developed by the Seattle Proteome Center within the Institute for Systems Biology, Seattle, WA; http://www.proteomecenter.org/software.php). Receiver operating characteristic (ROC) curves were generated to determine peptide and protein probability cutoffs providing ∼ 1% false discovery rate (typically ∼0.9–1.0). Following the search, results were exported to Microsoft Excel (cytoplasmic sample, 1703 proteins; nuclear/mitochondrial sample, 307 proteins). Proteins that were observed in multiple samples were assumed to be nonspecifically binding and were excluded from further analysis, as were proteins identified by one unique peptide observed only once. Potential isoforms and putative proteins were excluded as well, reducing the number of identifications to 227 for the cytoplasmic sample and 156 for the nuclear/mitochondrial sample. Proteins were sorted by calculated molecular weight (MW) and MWs differing from the median by more than 20% were assumed to be incorrect identifications. Finally, protein identifications with at least 5% AA coverage of the protein [18] were retained in the analysis.

FIGURE 1.2 Immunoaffinity purification of acetylated proteins reveals increased acetylation in proteins from livers of mice fed a high-fat diet. C57BL/6 SVJ mice were fed either normal chow or a high-fat diet (45% fat) for 16 weeks and fasted overnight before sacrifice (n = 3 per group). Samples were extracted from liver as described in the text. Lysine-acetylated liver proteins were immunoaffinity purified from either cytoplasmic or mitochondrial/nuclear extracts, then separated by 1D gel electrophoresis using a 10% SDS-PAGE gel, visualized with Lava Purple and imaged with a Typhoon 9600 imager (GE Healthcare). Labeled bands with differential staining were excised, digested, and identified by MS/MS and database searching.

As summarized in Tables 1.1 and 1.2, we identified 148 proteins from the cytoplasmic sample and 75 proteins from the nuclear/mitochondrial sample with high confidence. Each band typically contained a number of proteins that were identified. Bands excised from the cytoplasmic samples were larger to incorporate the apparently high number of closely comigrating proteins; therefore, more proteins were typically identified in those bands than in the nuclear/mitochondrial extract bands. An examination of proteins from the latter sample suggests that a high-fat diet led to hyperacetylation of proteins involved in gluconeogenesis, mitochondrial oxidative metabolism, methionine metabolism, liver injury, and the endoplasmic reticulum (ER) stress response. Not shown here, we observed that in mice lacking SIRT3, a sirtuin localized to the mitochondrion, a high-fat diet further increased the acetylation status of liver proteins compared with high-fat diet-fed wild-type mice and was associated with the disruption of mitochondrial oxidative phosphorylation complexes II, III, and V. Our results suggest that hyperacetylation of mitochondrial proteins may play a pivotal role in mechanisms regulating high-fat diet-induced mitochondrial dysfunction in livers of obese mice.

TABLE 1.1 Hyperacetylated Cytoplasmic Proteins from Livers of Mice Fed a High-Fat Diet

a Excised band labeled in Figure 1.2.

b UniProtKB identifier number.

c Number of AAs identified from peptides divided by total number of AAs in the identified protein and multiplied by 100.

d Probability value calculated by the ProteinProphet module of the LABKEY software algorithm.

e Total number of fragmentation spectra identified as belonging to a protein.

f Number of nonredundant fragmentation spectra identified as belonging to a protein.

TABLE 1.2 Hyperacetylated Proteins from a Nuclear/Mitochondrial Extract of Livers from Mice Fed a High-Fat Diet

a Excised band labeled in Figure 1.2.

b UniProtKB identifier number.

c Number of AAs identified from peptides divided by total number of AAs in the identified protein and multiplied by 100.

d Probability value calculated by the ProteinProphet module of the LABKEY software algorithm.

e Total number of fragmentation spectra identified as belonging to a protein.

f Number of nonredundant fragmentation spectra identified as belonging to a protein.

1.3.2 Identification of a Major Histocompatibility Complex Class II (MHC-II)-Complexed Death Transducer

MHC-II is primarily known to function in the presentation of antigenic peptides to T lymphocytes. However, these molecules have also been observed to transduce signals, leading to either cell activation or apoptotic death. The short, cytoplasmic tails of the two transmembrane proteins comprising MHC-II are not required for induction of apoptosis [23], therefore a protein complexing with MHC-II is likely important in mediating death signaling.

In this study, K46 cells (5 × 10⁸) were lysed in cell lysis buffer and immunoaffinity purified, as described above. Shown in Figure 1.3 is a gel separation of proteins from a representative preparation, suggesting the presence of many potentially complexed proteins. Bound proteins were eluted from the beads with citric acid and pH was neutralized with Tris. Following in-solution digestion, ∼30% of the sample was analyzed by nanoscale liquid chromatography–tandem mass spectrometry (nano-LC/MS/MS) using a 145-min gradient of acetonitrile. Data from five separate immunoaffinity purifications were analyzed by LABKEY and proteins with probabilities yielding less than 3% false discovery rates were included. Initially, a total of 2514 protein identifications were obtained from the combined five analyses. Since the data files were obtained from immunoaffinity purification of both MHC-II and a complexed protein (MPYS, subsequently termed TM173), we restricted the list to proteins that were observed in at least three runs, including one run from the MHC-II IP and one run from the MPYS IP. Identifications based on one hit were removed, as were duplicates, and putative uncharacterized proteins, resulting in a list of 237 proteins, of which 83 were isoforms of MHC-II. The list of accession numbers was then clustered using functional annotation tools in DAVID [21] and protein groups present are summarized in Figure 1.4. A recent review of commonly observed protein identifications [24] suggests that overrepresentation of actin, tubulin, and adenosine triphosphate (ATP) synthase isoforms may result from nonspecific binding or represent common cellular stress responses.

FIGURE 1.3 Immunoaffinity purification of MHC-II reveals a number of potentially complexed proteins. MHC-II-associated proteins were enriched from K46 cells lysates and analyzed using nano-LC/MS/MS. Incubation of protein A beads in the absence of primary antibody was employed as a negative control. The gel shown is a representative result from three similar experiments.

FIGURE 1.4 Functional classification of proteins identified in the MHC-II complex. The DAVID algorithm was used to assess overrepresentation of GO terms for the identified proteins. The Similarity Term Overlap was set to 3, the Similarity Threshold was 0.2, the Group Membership was 5, and the Multiple Linkage Threshold value was 0.5. Following data reduction, actins, tubulins, and heat shock proteins (25) were clustered into a group with an enrichment score of 13.7, ATPases (9) were clustered with an enrichment score of 5.7, and transmembrane proteins (17) were clustered with an enrichment score of 1.4. GAPDH was left unclustered.

Since MHC-II is a transmembrane protein, we concentrated further analysis on the group of transmembrane proteins identified in this study, summarized in Table 1.3. Of these proteins, only two were potentially multispanning proteins—CD20 and TM173. Fragmentation mass spectra leading to the identification of these proteins are plotted in Figures 1.5 and 1.6, respectively. CD20 is known to associate with MHC-II but was not thought to be a likely candidate for transducing death signals [23]. As described in the UniProt Knowledge Base, the isotopes of transmembrane emp24 are single-pass transmembrane proteins with cytoplasmic domains not well described, and CDGSH iron sulfur protein, implicated in autophagy, has a short cytoplasmic domain. Ribophorin and solute carrier protein 3 are both single membrane-spanning proteins with longer cytoplasmic domains (150 and 75 AAs, respectively). MHC-II is known to associate with tetraspanins, therefore we selected TM173, a potential tetraspanin with a 204-amino-acid-long cytoplasmic domain, for subsequent study.

TABLE 1.3 Transmembrane Proteins Potentially Complexed with MHC-II

FIGURE 1.5 Fragmentation mass spectrum and sequence analysis of peptide identified from CD20. An almost complete series of singly charged b- and y-type ions were used to unambiguously identify this doubly charged peptide SNVVLLSAGEKNEQTIK from CD20 (MW 31,958 Da). The PeptideProphet probability was 0.9999 and the ProteinProphet probability was 1.000 for this identification.

FIGURE 1.6 Fragmentation mass spectrum and sequence analysis of peptide identified from TM173. An almost complete series of singly charged b- and y-type ions were used to unambiguously identify this doubly charged peptide TLEEILEDVPESR from TM173 (MW 38,036 Da). The PeptideProphet probability was 0.9993 and the ProteinProphet probability was 1.000 for this identification.

As we previously described [23], TM173 is a membrane protein with an immunoreceptor tyrosine-based inhibitory motif (ITIM) contained in its cytoplasmic tail. We demonstrated that TM173 becomes phosphorylated upon cross-linking with MHC-II, recruits the inhibitory signaling effectors Src homology region 2 domain-containing phosphatase-1 (SHP-1) and Phosphatidylinositol-3,4,5-trisphosphate 5-phosphatase 1 (SHIP), reduces calcium mobilization, and negatively regulates cell growth. Confirmation of the important role of TM173 in MHC-II-mediated cell death was obtained in cells with TM173 knocked down.

1.4 CONCLUSION

The use of immunoaffinity purification as a sample preparation method for downstream proteomics applications is emerging for applications involving identification of proteins in functional complexes and global identification of modified proteins. This approach allows investigators to use proteomics approaches to address hypothesis-driven research questions. Although a detailed protocol was presented here, it should be noted that each sample type and antibody will require optimization of the protocol—different buffers may work better for lysis, antibodies may bind more or less strongly—therefore, the information here is presented only as a guideline or starting place.

ACKNOWLEDGMENTS

This work was supported by NIH Grants DK59767 and P30-DK48520 (to J.E.F.), a Pilot & Feasibility Award from the University of Colorado, Center for Human Nutrition P30-DK048520-09 (K.J.), and NIH Grant AI020519-22 (to J.C.C.). K.J. thanks W.H. MacDonald at Vanderbilt University for review of the manuscript and extraordinarily helpful discussions.

REFERENCES

1. Chien, C.T., Bartel, P.L., Sternglanz, R., Fields, S. (1991) The two-hybrid system: a method to identify and clone genes for proteins that interact with a protein of interest. Proceedings of the National Academy of Sciences of the United States of America, 88, 9578–9582.

2. Suter, B., Kittanakom, S., Stagljar, I. (2008) Interactive proteomics: what lies ahead? BioTechniques, 44, 681–691.

3. Gavin, A.-C., Bosche, M., Krause, R., Grandi, P., et al. (2002) Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature, 415, 141–147.

4. Ho, Y., Gruhler, A., Heilbut, A., Bader, G.D., et al. (2002) Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry. Nature, 415, 180–183.

5. Ewing, R.M., Chu, P., Elisma, F., Li, H., et al. (2007) Large-scale mapping of human protein-protein interactions by mass spectrometry. Molecular Systems Biology, 3, 2–17.

6. Choudhary, C., Mann, M. (2010) Decoding signaling networks by mass spectrometry-based proteomics. Nature Reviews Molecular Cell Biology, 11, 427–439.

7. Yang, W., Steen, H., Freeman, M.R. (2008) Proteomic approaches to the analysis of multiprotein signaling complexes. Proteomics, 8, 832–851.

8. Ashman, K., López Villar, E. (2009) Phosphoproteomics and cancer research. Clinical & Translational Oncology, 11, 356–362.

9. Choudhary, C., Kumar, C., Gnad, F., Nielsen, M.L., et al. (2009) Lysine acetylation targets protein complexes and co-regulates major cellular functions. Science, 325, 834–840.

10. Schirle, M., Heurtier, M.-A., Kuster, B. (2003) Profiling core proteomes of human cell lines by one-dimensional PAGE and liquid chromatography-tandem mass spectrometry. Molecular & Cellular Proteomics, 2, 1297–1305.

11. Kumar, C., Mann, M. (2009) Bioinformatics analysis of mass spectrometry-based proteomics data sets. FEBS Letters, 583, 1703–1712.

12. Hubbard, S.J., Jones, A.R., Elias, J.E., Gygi, S.P. (2010) Proteome Bioinformatics. Methods in Molecular Biology, Vol. 604. London: Humana Press.

13. Ashburner, M., Ball, C.A., Blake, J.A., Botstein, D., et al. (2000) Gene Ontology: tool for the unification of biology. The Gene Ontology Consortium. Nature Genetics, 25, 25–29.

14. Okuda, S., Yamada, T., Hamajima, M., Itoh, M., et al. (2008) KEGG Atlas mapping for global analysis of metabolic pathways. Nucleic Acids Research, 36, W423–W426.

15. Carvalho, P., Fischer, J., Chen, E., Domont, G., et al. (2009) GO Explorer: a gene-ontology tool to aid in the interpretation of shotgun proteomics data. Proteome Science, 7, 6.

16. Carvalho, P.C., Fischer, J.S.G., Chen, E.I., Yates, J.R., et al. (2008) PatternLab for proteomics: a tool for differential shotgun proteomics. BMC Bioinformatics, 9, 316.

17. Malovannaya, A., Li, Y., Bulynko, Y., Jung, S.Y., et al. (2010) Streamlined analysis schema for high-throughput identification of endogenous protein complexes. Proceedings of the National Academy of Sciences, 107, 2431–2436.

18. Moresco, J.J., Carvalho, P.C., Yates, J.R. III (2010) Identifying components of protein complexes in C. elegans using co-immunoprecipitation and mass spectrometry. Journal of Proteomics, 73, 2198–2204.

19. Kapp, E., Schütz, F. (2007) Overview of tandem mass spectrometry (MS/MS) database search algorithms. Current Protocols in Protein Science, 49, 25.2.1–25.2.19.

20. Dennis, G., Sherman, B., Hosack, D., Yang, J., et al. (2003) DAVID: database for annotation, visualization, and integrated discovery. Genome Biology, 4, P3.

21. Huang, D.W., Sherman, B.T., Lempicki, R.A. (2008) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature Protocols, 4, 44–57.

22. Bao, J., Sack, M. (2010) Protein deacetylation by sirtuins: delineating a post-translational regulatory program responsive to nutrient and redox stressors. Cellular and Molecular Life Sciences, 67, 3073–3087.

23. Jin, L., Waterman, P.M., Jonscher, K.R., Short, C.M., et al. (2008) MPYS, a novel membrane tetraspanner, is associated with major histocompatibility complex class II and mediates transduction of apoptotic signals. Molecular and Cellular Biology, 28, 5014–5026.

24. Petrak, J., Ivanek, R., Toman, O., Cmejla, R., et al. (2008) Déjà vu in proteomics. A hit parade of repeatedly identified differentially expressed proteins. Proteomics, 8, 1744–1749.

2

MASS SPECTROMETRY-BASED METHODS TO INVESTIGATE POSTTRANSLATIONAL PROTEIN MODIFICATIONS BY LIPID PEROXIDATION PRODUCTS

NAVIN RAUNIYAR AND LASZLO PROKAI

Covalent modifications of proteins are essential posttranslational processing steps that involve linkage of various chemical groups (phosphate, methyl, acetyl, and many others) to specific amino acid residues. These modifications play an important role in rendering the proteins functionally active for numerous cellular functions such as regulating cellular location and dynamic interactions with other proteins, achieving required three-dimensional conformation or catalytic activity. On the other hand, several posttranslational modifications (PTMs) arise from deleterious processes that compromise protein function. Many chemically distinct types of PTMs are known to occur physiologically (A comprehensive list of protein PTMs can be found at http://www.abrf.org/index.cfm/dm.home?AvgMass=all). PTMs of proteins increase the diversity and complexity of a proteome in an organism by providing covalent variation to protein backbones and side chains. It is imperative to identify covalent protein modifications in order to decode their role in biological processes. Mass spectrometry (MS) has emerged as a powerful technology for protein identification and characterization of PTMs. However, comprehensive analysis of posttranslationally modified peptides in complex mixtures is a challenging endeavor because of the necessity to digest proteins to successfully sequence the resulting peptides by MS in bottom-up proteomics approach [1]. The modified peptides occur in substoichiometric levels compared with corresponding native forms in the analytes, a situation analogous to finding a needle in a haystack. Also, the modified peptides may ionize less efficiently by electrospray ionization (ESI) and matrix-assisted laser desorption/ionization (MALDI) than those of the regular, unmodified peptides in a complex mixture, and the low-abundance ions of modified peptides that coelute with intense ions of unmodified peptides may be missed by conventional data-dependent acquisition [2].

The recent advent of a new generation of mass spectrometers with ultra-high mass resolution, exceptionally high mass accuracy and sensitivity, and unique ion manipulation capabilities such as electron capture dissociation (ECD) and electron transfer dissociation (ETD), together with development of various excellent database search algorithms, has proven invaluable for global interrogation of protein modifications from the acquired mass spectra. This has resulted in an unprecedented growth in studies oriented toward comprehensively mapping of protein PTMs. Additionally, the implementation of several enrichment techniques, including covalent coupling and affinity capture, for purifying the subproteome carrying modified forms from native species and simultaneously reducing the sample complexity and peptide coelution prior to mass spectrometric analysis, greatly enhances the chance of detecting those modifications that are substoichiometric in nature or are affecting low-abundance proteins. Thus, the combination of high-performance MS with new computational techniques and experimental methods are facilitating large-scale analysis of PTMs.

PTMs usually are site specific, located at specific amino-acid residues at a particular conserved sequence motif in proteins. However, modifications by 4-hydroxy-2-nonenal (HNE), a reactive carbonyl species (RCS) produced from polyunsaturated fatty acids (PUFAs) during lipid peroxidation, in proteins according to reactions shown in Scheme 2.1, cannot be predicted by computational sequence analysis and, therefore, experimental proteomic techniques have to be developed for their determination. Michael adducts generally represent >99% of HNE protein modifications by HNE, whereas Schiff-base (SB) adduct formation is less prevalent even in the presence of excess HNE [3,4]. A bottom-up protocol for determination of posttranslational protein carbonylation by HNE requires high-efficiency front-end enrichment of modified peptides from biological samples, for which this chapter describes the method that employs solid-phase hydrazide (SPH) chemistry summarized in Scheme 2.2 [5]. In addition, several MS-based approaches have been developed to identify HNE-modified peptides through coupling with online liquid chromatographic separation, data-dependent acquisition, and subsequent database searches from tandem mass spectra [6–9]. From this handbook’s thematic viewpoint, specific highlights of the procedure are data-dependent acquisitions that utilize complementary ion dissociation methods such as collision-induced dissociation (CID) and ECD with tandem mass spectrometry (MS/MS). In case of CID, third-order tandem mass spectrometry (MS/MS/MS or MS³) triggered by modification-specific neutral loss (NL) is also used to interrogate peptides with covalently attached HNE for improved identification and localization of modification sites. These methods can be implemented on modern linear ion trap and, especially, ion trap–Fourier transform hybrid mass spectrometers.

SCHEME 2.1 Reaction of –NH2 groups, imidazole =NH group, and –SH groups of lysine (Lys, K), histidine (His, H), and cysteine (Cys, C) side-chains, respectively, in proteins with 4-hydroxy-2-nonenal (HNE) through MA (a, +156 Da). HNE can also form SB adducts (b, +138 Da) with –NH2 groups.

SCHEME 2.2 Schematic illustration of the procedure used for solid-phase hydrazide (SPH) enrichment of 4-HNE-carbonylated peptides.

Adapted from Reference [9], ©2010 Wiley Interscience, with permission.

2.1 PROCEDURE CONTROL

Although the protocol is aimed at discovery-driven exploration of posttranslational HNE modifications in proteins and, hence, rigorous quality control does not apply, initial demonstration of performance is recommended before application to complex biological samples including cell lysates. Prepare an HNE-modified protein for such demonstration by incubating a protein purchased from a commercial vendor (see the section Reagents and Standards) or, in specific protein-targeted experiments, isolated for the purpose of the study. Operating procedures pertaining to the SPH-based enrichment procedure, as well as the CID-based NL-driven MS³ and NL-driven ECD acquisitions coupled with online nanoscale liquid chromatography (nano-LC) should be verified with the tryptic digest of the modified standard protein(s). Compare the results of performance demonstration to data available from the literature (e.g., [7,9]). For label-free quantitation, consider exploratory validation of spectral counts or extracted ion chromatograms (XICs) versus protein concentrations by spiking the prepared HNE-modified protein into a sample matrix such as a cell lysate [10].

2.2 IDENTIFICATION OF HNE-MODIFIED PEPTIDES IN BIOLOGICAL SAMPLES BY SOLID-PHASE ENRICHMENT AND NANO-LC–ESI-MS/MS

This protocol describes a procedure for the identification of PTM by HNE to reveal protein targets and specific sites of covalent attachment. Identification is performed by combining proteolytic digestion followed by SPH enrichment and nanoscale liquid chromatography–electrospray ionization tandem mass spectrometry (nano-LC–ESI-MS/MS). The modified proteins are subjected to reduction, alkylation, and subsequent digestion by a proteolytic enzyme. The peptides thus obtained are desalted and the substoichiometric quantities of HNE-modified peptides are fractionated from unmodified species using hydrazide-coated glass beads-based enrichment technique that selectively captures peptide carbonyls (as hydrazones) that are subsequently released from the beads in their original forms by acid-catalyzed hydrolysis. The enriched HNE-modified peptides are analyzed by several MS-based approaches, including data-dependent acquisition that utilize CID and ECD as a method of peptide dissociation, as well as NL-driven collision-induced dissociation–mass spectrometry (CID-MS³) and NL-driven electron capture dissociation tandem mass spectrometry (ECD-MS/MS) methods for in-depth profiling of carbonylation sites in the peptides. HNE-modified peptide matches from the database search algorithms are visually inspected to confirm the mass shifts and only peptides with unambiguous sites of modification within their amino acid sequence should be considered correct. The procedure can be verified using HNE-modified synthetic peptide standards such as human angiotensin (AGT) I peptide (DRVYIHPFHL), AGT II octapeptide (DRVYIHPF), and custom synthesized peptides LVLEVAQHLGESTVR and IVYGHLDDPANQEIER (which correspond to tryptic peptides of ATP synthase subunit β and aconitate hydratase, respectively) [7]. HNE-modified apomyoglobin from equine skeletal muscle may also be used for the purpose of verification [9]. Although the characteristic elimination of HNE (156 Da) in CID can serve as a signature tag for the modified peptides, ECD is recommended as a complementary method of ion dissociation to aid the elucidation of primary structural information and assignment of exact carbonylation sites in the protein. A nearly complete fragment ion series is generated with ECD due to efficient peptide backbone cleavage (in most cases over 75%) and is also capable of retaining the labile HNE moiety in the peptides [9].

2.2.1 Materials

Water, high performance liquid chromatography (HPLC) grade (EMD [Philadelphia, PA] or equivalent)

Acetonitrile, HPLC grade (EMD or equivalent)

Acetic acid, 99.99% (Sigma [St. Louis, MO] or equivalent)

Formic acid, ≥98% (Sigma or equivalent)

HNE (5 mg/500 µL in ethanol, Cayman Chemical Company, Ann Arbor, MI)

Apomyoglobin from equine skeletal muscle (Sigma) for modification by HNE (see the section Reagents and Standards), or similar protein commercially available or isolated

Lysis buffer for mammalian cells: 20 mM TrisHCl/50 mM NaCl/6 M urea (pH 8.1)/10 mM sodium pyrophosphate/1 mM sodium fluoride/1 mM β-glycerophosphate/1 mM sodium orthovanadate/1 tablet complete Mini protease inhibitor mixture (Roche, Indianapolis, IN) dissolved per 10 mL buffer

Laboratory centrifuge (Eppendorf [Hauppauge, NY] or equivalent)

Laboratory vortex device (Vortex-Genie [Bohemia, NY] or equivalent)

Speedvac concentrator (Eppendorf Vacufuge or equivalent)

Kit and equipment for spectrophotometric protein concentration measurement by Lowry, Bradford, or bicinchoninic acid (BCA) assay (Thermo [San Jose, CA] or equivalent)

1.5 and 0.5 mL polypropylene centrifuge tubes (Eppendorf)

50 mM Ammonium bicarbonate buffer, pH 7.4

Phosphate-buffered saline (PBS)

Dithiothreitol (DTT, Sigma)

Iodoacetamide (IAA, Sigma)

Sequencing-grade modified porcine trypsin (Promega, Madison, WI)

Sequencing-grade trifluoroacetic acid (TFA) ≥98% (Sigma)

Solid-phase extraction (SPE) C18 column (Sep-Pak, Waters, Milford, MA)

SPE solvents: Binding buffer (1% acetic acid, v/v, in H2O) and elution buffer (80% acetonitrile v/v, plus 1% acetic acid, v/v, in H2O).

SPH reagent (prepared according to Roe et al. [5])

Buffers and solutions for SPH chemistry: Reaction buffer (0.2% acetic acid v/v, 10% acetonitrile v/v, pH 3.6); washing solvent and solutions (distilled water, 1 M aqueous NaCl and 80:20 v/v acetonitrile : H2O); elution solution (10% v/v aqueous formic acid)

250 µL polypropylene autosampler vials with press-on caps and teflon-lined septa (National Scientific Company, Rockwood, TN)

High-performance nanoscale liquid chromatography system (Eksigent [Dublin, CA] nano-LC™ or equivalent), equipped with autosampler

IntegraFrit™ sample trap, 25 mm × 75 µm internal diameter (i.d.) (New Objective [Woburn, MA]) or equivalent preconcentration column

PepMap C18 analytical column, 150 mm × 75 µm i.d., 3 µm particle size, 100 Å pore size (LC Packings, San Francisco, CA) or equivalent nano-LC column packed with octadecylsilica particles.

Picotip emitter, i.d. 10 ± 1 µm (New Objective)

Loading solvent for reversed-phase liquid chromatography (RPLC): 0.1% (v/v) acetic acid and 5% (v/v) acetonitrile in 94.9% (v/v) water

Mobile phase A for reversed-phase (RP) column: 0.1% (v/v) acetic acid and 99.9% (v/v) water

Mobile phase B for RP column: 0.1% (v/v) acetic acid and 99.9% (v/v) acetonitrile

Hybrid linear ion trap–Fourier transform ion cyclotron resonance (FTICR) mass spectrometer, 7T (LTQ-FTICR, Thermo), or equivalent instrument, equipped with an electrospray/nanospray (ESI/NSI) source and ECD accessory, operated under the control of Xcalibur software package (Thermo Fisher Scientific, San Jose, CA; Instrument Configuration, Instrument Setup, Sequence Setup, and Qual Browser modules appropriately created and used) with Bioworks available to preprocess raw data files for database search. ECD may be replaced with ETD and the respective manufacturer’s own version of the above software should be used, when alternative instrument is selected.

International Protein Index database for the species studied (available online at http://www.ebi.ac.uk/IPI) and its sequence-reversed version with the Mascot (Matrix Science, Boston, MA) search algorithm. Equivalent search software (Proteome Discoverer [Thermo Fisher Scientific], Protein Prospector [http://prospector.ucsf.edu/prospector/mshome.htm], Phenyx [GeneBio, Geneva, Switzerland], X!Tandem [http://www.thegpm.org/tandem/], etc.) may also be used.

Additional program(s) to validate the MS/MS-based peptide and protein identifications, to summarize results including MS/MS spectral counts (Scaffold, Proteome Software, Portland, OR), and to create aligned extracted-ion chromatograms (XICs, Sieve®, Thermo), and so on.

2.2.2 Preparation of Cell Lysates

1. Harvest the cells by centrifugation at 2000 × g in 1.5 mL centrifuge tubes for 15 min at 4°C. Pour off the supernatant and discard it.

Note: During lysis, cells and lysates should be kept at 4°C at all times. Temperature has a profound effect on the catalytic activity of most proteases.

2. Carefully wash the cell pellet twice with ice-cold PBS, and then place the washed cell pellet on ice. Resuspend the pellet in 1 mL of lysis buffer previously chilled to 4°C. Incubate the cells for 30 min on ice with occasional vortexing of the tube. Perform freeze/thaw cycles to complete the rupture of the cell membrane and release the contents. Centrifuge the resultant solution at 20,000 × g for 15 min at 4°C to separate the bulk of cell debris from the lysates. Carefully transfer the supernatant to a fresh tube, making sure not to disturb the pellet.

3. Measure protein concentration using Lowry, Bradford, or BCA protein assay and bovine serum albumin (BSA) as a reference.

Note: It is recommended to continue with the trypsin digestion step immediately to minimize protein degradation. If lysates need to be stored at this point, then snap-freeze them using a dry ice/ethanol mixture and store them at −80°C.

2.2.3 Protein Reduction and Alkylation

1. Add DTT, also known as Cleland’s reagent, from a 0.5 M stock solution to reach 10 mM concentration in the cell lysate and incubate for 30 min at 56°C to reduce the disulfide bonds.

Note: Incubation at higher temperature may cause urea-based carbamylation of lysines and protein N-termini.

2. Allow the reduced protein solution to cool to room temperature and add IAA from a 0.5 M stock solution to reach 20 mM concentration in the reduced cell lysate. Incubate the reaction for 30 min at room temperature in the dark to alkylate the cysteine residues. Modification of the protein thiol groups by alkylation prevents reformation of disulfide bonds due to oxidation upon exposure to air.

3. Quench unreacted IAA by increasing DTT content by 10 mM (see step 1 above) and incubating for 15 min at room temperature in the dark.

2.2.4 Trypsin Digestion

1. Dilute the protein mixture in 50 mM ammonium bicarbonate, pH 7.4, to reduce urea concentration (6 M in the lysis buffer) to about 1.6 M.

2. Add trypsin at a final protease-to-protein ratio of 1:50 (w/w) and incubate at 37°C overnight. Longer incubations, up to 24 h, may be required depending on the nature of the protein.

Note: Trypsin is a serine protease that specifically cleaves at the C-terminal side of arginine (R) and lysine