Enabling Real-Time Compensation in Fast Photochemical Oxidations of Proteins for the Determination of Protein Topography Changes

Summary

Fast photochemical oxidation of proteins is an emerging technique for the structural characterization of proteins. Different solvent additives and ligands have varied hydroxyl radical scavenging properties. To compare the protein structure in different conditions, real-time compensation of hydroxyl radicals generated in the reaction is required to normalize reaction conditions.

Abstract

Fast photochemical oxidation of proteins (FPOP) is a mass spectrometry-based structural biology technique that probes the solvent-accessible surface area of proteins. This technique relies on the reaction of amino acid side chains with hydroxyl radicals freely diffusing in solution. FPOP generates these radicals in situ by laser photolysis of hydrogen peroxide, creating a burst of hydroxyl radicals that is depleted on the order of a microsecond. When these hydroxyl radicals react with a solvent-accessible amino acid side chain, the reaction products exhibit a mass shift that can be measured and quantified by mass spectrometry. Since the rate of reaction of an amino acid depends in part on the average solvent accessible surface of that amino acid, measured changes in the amount of oxidation of a given region of a protein can be directly correlated to changes in the solvent accessibility of that region between different conformations (e.g., ligand-bound versus ligand-free, monomer vs. aggregate, etc.) FPOP has been applied in a number of problems in biology, including protein-protein interactions, protein conformational changes, and protein-ligand binding. As the available concentration of hydroxyl radicals varies based on many experimental conditions in the FPOP experiment, it is important to monitor the effective radical dose to which the protein analyte is exposed. This monitoring is efficiently achieved by incorporating an inline dosimeter to measure the signal from the FPOP reaction, with laser fluence adjusted in real-time to achieve the desired amount of oxidation. With this compensation, changes in protein topography reflecting conformational changes, ligand-binding surfaces, and/or protein-protein interaction interfaces can be determined in heterogeneous samples using relatively low sample amounts.

Introduction

Fast photochemical oxidation of proteins (FPOP) is an emerging technique for the determination of protein topographical changes by ultra-fast covalent modification of the solvent-exposed surface area of proteins followed by detection by LC-MS¹. FPOP generates a high concentration of hydroxyl radicals in situ by UV laser flash photolysis of hydrogen peroxide. These hydroxyl radicals are very reactive and short lived, consumed on roughly a microsecond timescale under FPOP conditions². These hydroxyl radicals diffuse through water and oxidize various organic components in solution at kinetic rates generally ranging from fast (~10⁶ M^-1 s^-1) to diffusion-controlled³. When the hydroxyl radical encounters a protein surface, the radical will oxidize the amino acid side chains on the protein surface, resulting in a mass shift of that amino acid (most commonly the net addition of one oxygen atom)⁴. The rate of the oxidation reaction at any amino acid depends on two factors: the inherent reactivity of that amino acid (which depends on the side chain and the sequence context)^4,5 and the accessibility of that side chain to the diffusing hydroxyl radical, which closely correlates to the average solvent accessible surface area^6,7. All of the standard amino acids except glycine have been observed as labeled by these highly reactive hydroxyl radicals in FPOP experiments, albeit at widely differing yields; in practice, Ser, Thr, Asn, and Ala are rarely seen as oxidized in most samples except under high radical doses and identified by careful and sensitive targeted ETD fragmentation^8,9. After oxidation, samples are quenched to remove hydrogen peroxide and secondary oxidants (superoxide, singlet oxygen, peptidyl hydroperoxides, etc.) The quenched samples are then proteolytically digested to generate mixtures of oxidized peptides, where the structural information is frozen as a chemical “snapshot” in the patterns of oxidation products of the various peptides (Figure 1). Liquid chromatography coupled to mass spectrometry (LC-MS) is used to measure the amount of oxidation of amino acids in a given proteolytic peptide based on the relative intensities of the oxidized and unoxidized versions of that peptide. By comparing this oxidative footprint of the same protein obtained under different conformational conditions (e.g., ligand-bound versus ligand-free), differences in the amount of oxidation of a given region of the protein can be directly correlated with differences in the solvent-accessible surface area of that region^6,7. The ability to provide protein topographical information makes FPOP an attractive technology for the higher-order structure determination of proteins, including in protein therapeutic discovery and development^10,11.

Figure 1: Overview of FPOP. The surface of the protein is covalently modified by highly reactive hydroxyl radicals. The hydroxyl radicals will react with amino acid side chains of the protein at a rate that is strongly influenced by the solvent accessibility of the side chain. Topographical changes (for example, due to the binding of a ligand as shown above) will protect amino acids in the region of interaction from reacting with hydroxyl radicals, resulting in a decrease in the intensity of modified peptide in the LC-MS signal. Please click here to view a larger version of this figure.

Different constituents present in the FPOP solution (e.g., ligands, excipients, buffers) have different scavenging activity towards the hydroxyl radicals generated upon the laser photolysis of hydrogen peroxide³. Similarly, a small change in peroxide concentration, laser fluence, and buffer composition may change the effective radical dose, making the reproduction of FPOP data challenging across the samples and between different labs. Therefore, it is important to be able to compare the hydroxyl radical dose available to react with protein in each sample using one of several available hydroxyl radical dosimeters^{12,13,14,15,16}. Hydroxyl radical dosimeters act by competing with the analyte (and with all scavengers in solution) for the pool of hydroxyl radicals; the effective dose of hydroxyl radicals is measured by measuring the amount of oxidation of the dosimeter. Note that “effective hydroxyl radical dose” is a function of both the initial concentration of hydroxyl radical generated and the half-life of the radical. These two parameters are partially dependent on one another, making the theoretical kinetic modeling somewhat complex (Figure 2). Two samples could have wildly different initial radical half-lives while still maintaining the same effective radical dose by changing the initial concentration of hydroxyl radical formed; they will still generate identical footprints¹⁷. Adenine¹³ and Tris¹² are convenient hydroxyl radical dosimeters because their level of oxidation can be measured by UV spectroscopy in real-time, allowing for researchers to quickly identify when there is a problem with effective hydroxyl radical dose and to troubleshoot their problem. To solve this issue, an inline dosimeter located in the flow system directly after the site of irradiation that can monitor the signal from adenine absorbance changes in real-time is important. This helps in carrying out FPOP experiments in buffers or any other excipient with widely differing levels of hydroxyl radical scavenging capacity¹⁷. This radical dosage compensation can be performed in real-time, yielding statistically indistinguishable results for the same conformer by adjusting the effective radical dose.

In this protocol, we have detailed procedures for performing a typical FPOP experiment with radical dosage compensation using adenine as an internal optical radical dosimeter. This method allows investigators to compare footprints across FPOP conditions that have different scavenging capacity by performing compensation in real-time.

Figure 2: Kinetic simulation of dosimetry-based compensation. 1 mM adenine dosimeter response is measured in 5 µM lysozyme analyte with a 1 mM initial hydroxyl radical concentration (▪OH t_1/2=53 ns), and set as a target dosimeter response (black). Upon the addition of 1 mM of the scavenger excipient histidine, the dosimeter response (blue) decreases along with the amount of protein oxidation in a proportional manner (cyan). The half-life of the hydroxyl radical also decreases (▪OH t_1/2=39 ns). When the amount of hydroxyl radical generated is increased to give an equivalent yield of oxidized dosimeter in the sample with 1 mM histidine scavenger as achieved with 1 mM hydroxyl radical in the absence of scavenger (red), the amount of protein oxidation that occurs similarly becomes identical (magenta), while the hydroxyl radical half-life decreases even further (▪OH t_1/2=29 ns). Adapted with permission from Sharp J.S., Am Pharmaceut Rev 22, 50-55, 2019. Please click here to view a larger version of this figure.

Protocol

1. Prepare the Optical Bench and the Capillary for FPOP

CAUTION: KrF excimer lasers are extreme eye hazards, and direct or reflected light can cause permanent eye damage. Always wear appropriate eye protection, avoid the presence of any reflective objects near the beam path when possible, and use engineering controls to prevent unauthorized access to an active laser and to restrain any stray reflections.

1. Prepare the FPOP optical bench.
   1. Turn on the laser to warm up. Set the laser to External Trigger, Constant Energy, No Gas Replacement. Set the laser energy per pulse (typically between 80-120 mJ/pulse).
   2. Set up the optical bench with the plano-convex lens (30 mm Dia. x 120 mm FL uncoated) directly in the path of the laser beam and a non-reflective backstop to absorb the light as shown in Figure 3A.

Figure 3: Optical bench for the FPOP experiment. (A) The sample is mixed with H₂O₂, adenine radical dosimeter, and glutamine scavenger and loaded into the syringe. The sample is pushed through the fused silica capillary through the focused beam path of a KrF excimer UV laser. The UV light photolyzes H₂O₂ into hydroxyl radicals, which oxidizes the protein and adenine dosimeter. The syringe flow pushes the illuminated sample out of the path of the laser before the next laser pulse, with an unilluminated exclusion volume between illuminated regions. Immediately after oxidation, the sample is passed through an inline UV spectrophotometer, which measures the UV absorbance of adenine at 265 nm. The sample is then deposited into a quench buffer to eliminate the remaining H₂O₂ and secondary oxidants. (B) The spot size is measured after irradiating a colored sticky note affixed behind the capillary with the laser at 248 nm. The width of the spot is used for calculating the sample flow rate, and the silhouette of the capillary in the center of the spot is used to align the optical bench. Please click here to view a larger version of this figure.

2. Cut an appropriate length of the fused silica capillary (360 µm outer diameter and 100 µm inner diameter) and using a sleeve, connect to the gas-tight syringe using a low dead volume connector.
3. Gently burn the polyimide coating of the capillary with a butane torch at the place where the inline dosimeter reads the absorbance signal at 265 nm after laser exposure of the samples. Wipe the debris on the capillary gently using methanol on a lint-free wipe. The polyimide coating at the site of laser incidence can either be similarly burned off with the butane torch or burned off with the excimer laser firing at low power.
   NOTE: Wait for the capillary to cool as it is a fire hazard to use the methanol on the hot capillary.
4. Place this capillary through the beam path of the laser and into the inline dosimeter.
   1. Press the lever on the top of the inline dosimeter to open the hinge. Remove the magnetic holders. Place the capillary in the machined groove of the inline dosimeter, using the magnetic holders to keep the capillary in place. Close the dosimeter hinge over the capillary, pressing it until the lever locks in place.
5. Using the dosimetry software, click on the Start Flash button to begin firing the excimer laser. Set the preset laser power between 50-100 mJ/pulse on the laser control software itself, and set the preset repetition rate between 10-20 Hz in the Settings tab of the dosimetry software.
   1. Focus the laser beam using a plano convex lens mounted on a linear motorized stage. Measure the width and the height of the laser spot at the position of the capillary on a sticky note precisely using a caliper to calculate incident fluence (mJ/mm²) as shown in Figure 3B.
6. Place an opaque aperture near the capillary to ensure consistent illuminated width of the capillary regardless of changes in the beam size due to movement of the lens or changing the energy per pulse of the laser¹⁸.
   1. With the laser firing, move the motorized stage through its range of motion. Ensure that the beam stays centered on the aperture and the silhouette of the capillary can be observed throughout. The diameter of the aperture must be smaller than the width of the impinging focused beam at every point in the range of the motorized stage.
7. Run water through the capillary at 20 µL/min for at least one minute to wash the capillary.
   1. Click the Start Data + AutoZero button on the dosimeter software to zero the dosimeter to water and begin data collection.
      NOTE: If the buffer system for FPOP has significant UV absorbance at 265 nm, the FPOP system should be zeroed on the buffer, not water.
8. Set the calculated flow rate on the syringe pump.
   1. The flow rate of the protein sample depends on the irradiated volume per shot (V_Irr), the number of laser shots per second (R), and the desired unirradiated exclusion volume fraction (F_Ex) to correct for laminar flow effects and sample diffusion (0.15-0.30 recommended)^2,19,20. Calculate the V_Irr (in µL) based on the inner diameter of the capillary in mm (d) and the width of the laser spot impinging upon the capillary (i.e., the width of the aperture) in mm (w) using the following equation:
      V_Irr = π(d/2)²w
   2. Calculate the desired flow rate (in µL/min) based on the following equation:
      Flow = 60R[V_Irr (1 + F_Ex)]

2. Preparation of the protein solution for FPOP

1. Prepare the protein in the two or more different conditions to be compared (e.g., ligand-bound and ligand-free; aggregate and monomer; alone and with a protein-protein binding partner; etc.) for detecting the conformation changes.
2. Set the total volume used for FPOP to fit the needs of the experiment. The minimum limit usually depends on the volume of the irradiation capillary and the material required for robust detection and relative quantification, and will vary depending largely on the LC-MS/MS system used and the post-labeling sample processing method. The total volume for FPOP solutions commonly used in our group is 20 µL after the addition of hydrogen peroxide. The final concentration of the protein is commonly 1-10 µM, with 17 mM glutamine (to limit the lifetime of the hydroxyl radical), 1 mM adenine (to act as a radical dosimeter)^13,17 and 10 mM phosphate buffer (a buffer that is a poor scavenger of hydroxyl radicals). Samples are generally prepared with multiple replicates to allow for statistical modeling of results.
   1. For most general purposes, prepare samples in triplicate in both states, plus at least one sample to use as a no-laser control to measure background oxidation. Prepare 18 µL of this FPOP solution mix.
      NOTE: Many buffers and additives commonly used in biochemistry are hydroxyl radical scavengers. These additives and buffers can be used; however, reductions in oxidation due to hydroxyl radical scavenging of the buffer can occur. In general, keep all additives to the minimum required by the biological system to maximize protein oxidation yield. Dimethyl sulfoxide should be avoided due to the propensity to generate secondary radicals; dimethylformamide has been a useful alternative in our hands. When using buffers that are strong hydroxyl radical scavengers, glutamine can often be excluded from the FPOP solution mix.
3. Prepare 1 M hydrogen peroxide immediately before the FPOP experiment.
   NOTE: 30% hydrogen peroxide as commonly sold by vendors includes a stabilizer, which increases the shelf life. Once diluted, hydrogen peroxide should be used quickly, definitely within the same day. Hydrogen peroxide should also be regularly tested for decomposition by FPOP using a hydroxyl radical dosimeter.
4. Prepare microcentrifuge tubes containing 25 µL of quench solution of 0.5 µg/µL of methionine amide and 0.5 µg/µL catalase. If a sample volume greater than 20 µL is used for FPOP, increase the quench solution volume proportionally.

3. Perform the FPOP experiment

1. Add 2 µL of hydrogen peroxide in the 18 µL of the FPOP solution mix. Mix the contents gently with a pipette and quickly spin down the solution to the bottom of the microcentrifuge tubes. Immediately collect with a gastight syringe and load into the syringe pump.
2. Start the flow on the syringe pump with the flow rate as determined in step 1.8.1 (typically between 8-16 µL/min) by clicking the Start Pump button on the dosimeter software.
3. Monitor the real-time adenine reading using the inline dosimeter (see Table of Materials) and collect the sample in waste. Wait for the Abs₂₆₅ signal to stabilize.
4. Click on the Start Flash button in the dosimeter software to start firing the laser at the preset repetition rate and energy.
5. Monitor the real-time adenine reading using an inline dosimeter (see Table of Materials); the difference in Abs₂₆₅ with the laser off and the laser on is the ΔAbs₂₆₅ reading.
   NOTE: The appearance of highly unstable Abs₂₆₅ readings upon firing the laser in the presence of hydrogen peroxide is due to the generation of bubbles in solution. Reduce the fluence of the laser and/or the concentration of hydrogen peroxide to eliminate the bubbles.

4. Perform Compensation

NOTE: Different ligands, buffers, etc. may have different scavenging capacity towards hydroxyl radicals. It is important to ensure that comparable effective hydroxyl radical doses are available to react with protein across different samples. This is accomplished by ensuring equal hydroxyl radical dosimeter response between samples. Using adenine dosimetry, the change in UV absorbance at 265 nm (ΔAbs₂₆₅) reflects the effective hydroxyl radical dose; the larger the ΔAbs₂₆₅, the higher the effective hydroxyl radical dose.

1. Compare the ΔAbs₂₆₅ reading obtained with the inline dosimeter with the desired ΔAbs₂₆₅ reading obtained by prior experiments or controls. A ΔAbs₂₆₅ reading lower than the desired reading indicates an insufficient effective dose of hydroxyl radicals; a ΔAbs₂₆₅ reading indicates an effective radical dose that is too high. If the ΔAbs₂₆₅ reading is at the desired level, collect the sample immediately after laser irradiation in the quench buffer¹⁷.
2. Compensate the effective radical dose to equalize the ΔAbs₂₆₅. This compensation can be performed in three ways: change hydrogen peroxide concentration, increase laser fluence by changing the laser energy per pulse, or increase laser fluence by changing the focal plane of the focusing lens.
   1. To make a large change (>10 mAU) in ΔAbs₂₆₅ reading, remake the sample with more or less hydrogen peroxide and rerun the sample as per Section 3.
   2. To make a small change in ΔAbs₂₆₅ reading in real-time, adjust the focal plane of the incident beam by adjusting the position of the focusing lens using the 50 mm Motorized Stage. Bringing the focal plane closer to the position of the capillary will increase the ΔAbs₂₆₅ reading; bringing the focal plane farther from the position of the capillary will decrease the ΔAbs₂₆₅ reading.
3. Monitor the adenine ΔAbs₂₆₅ to measure the effective amount of hydroxyl radical present in the sample after laser irradiation¹³. Real-time monitoring with an inline UV capillary detector allows for real-time compensation as described in 4.2.2; adjust the lens position using the motorized stage until the ΔAbs₂₆₅ reading is equal to the desired reading. Post-experimental absorbance measurements with a UV spectrophotometer are also accurate, but require new samples to be used for each effective radical dose.

5. Digest the protein samples

NOTE: Trypsin is most commonly used to digest protein samples for FPOP, and is the protease used in this protocol. It is a reliable protease that generates peptides with basic sites both at the N- and C-terminus, promoting multiply charged peptide ions in MS. Moreover, it cleaves after lysine and arginine, two amino acids that are only moderately reactive to hydroxyl radicals; therefore, changes in the digestion pattern due to analyte oxidation is rare. Other proteases have been successfully used with FPOP²¹, but care should be taken to ensure digestion patterns are comparable between unoxidized and oxidized samples.

1. Measure the final volume of quenched FPOP sample. Add 500 mM Tris, pH 8.0 with 10 mM CaCl₂ containing 50 mM dithiothreitol (DTT) to the protein solution after quenching to a final concentration of 50 mM Tris, 1 mM CaCl₂ and 5 mM DTT.
2. Heat the protein sample at 95 °C for 15 minutes.
3. Immediately cool the sample on ice for 2 min.
4. Add the 1:20 trypsin/protein weight ratio to the samples.
5. Digest the protein overnight at 37 °C with mixing.
6. Stop the digestion reaction by the addition of 0.1% formic acid and/or heating the sample to 95 °C for 10 min.
7. Add 2 mM DTT to the samples and heat at 60 °C for 15 min immediately before LC-MS/MS.
   NOTE: While other groups have reported alkylation of thiols in FPOP experiments, in our hands we have noted side products upon alkylation of oxidized proteins (possibly due to reaction with nucleophilic carbonyls formed as a minor oxidation product). Therefore, we choose to avoid alkylation of thiols when possible.

6. Perform liquid chromatography-tandem mass spectrometry (LC-MS/MS)

1. Prepare the mobile phase A consisting of water containing 0.1% formic acid and mobile phase B consisting of acetonitrile with 0.1% formic acid.
2. Load the sample first onto a C18 trap column (300 µm I.D. x 5 mm 100 Å pore size, 5 µm particle size) trapping cartridge and wash with 2% solvent B for 3 minutes at a flow rate of 5.0 µL/min to remove salts and hydrophilic small molecules.
3. Then separate the peptides on C18 nanocolumn (0.75 mm x 150 mm, 2 µm particle size, 100 Å pore size) at a flow rate of 300 nL/min. The gradient consists of a linear increase from 2 to 35% solvent B over 22 min, ramped to 95% solvent B over 5 min and held for 3 min to wash the column, and then returned to 2% B over 3 min and held for 9 min to re-equilibrate the column.
   NOTE: This gradient is sufficient for LC-MS/MS of most one- and two-protein FPOP mixtures seeking to do peptide-level quantification. The percent of solvent B may need to be altered to increase peptide resolution in rare cases where peptides interfere with one another due to similar retention times and m/z values. Proteome-scale FPOP²² or experimental designs seeking to separate peptide oxidation product isomers^1,23,24,25 may require longer LC gradients and are beyond the scope of this report.
4. Elute the peptides directly into the nanospray source of a high-resolution mass spectrometer using a conductive nanospray emitter.
5. Acquire the data in positive ion mode. Set the spray voltage to 2400 V, and the temperature of the ion transfer tube to 300 °C.
6. Acquire the full MS scans from m/z 250 to 2000 at a nominal resolution at m/z 200 of 60,000 followed by eight subsequent data-dependent linear ion trap MS/MS scans on the top eight most abundant peptide ions using collision-induced dissociation at 35% normalized energy to identify the peptides. Fragment the peptides up to five times within 30 s and then transfer to an exclusion list for 60 s.

7. Data processing and calculation of average oxidation of peptides

1. Determine the sequence coverage of the protein, m/z values, and retention times of unoxidized peptides using the MS/MS proteomics search engine.
2. Set the precursor mass tolerance to 10 ppm and allow up to two missed cleavage sites for the trypsin digested samples, using standard trypsin cleavage specificity.
3. Set the peptide mass fragment mass tolerance to 0.4 Daltons.
4. Based on the m/z ratio of the unmodified peptides detected and the known mass shifts of the major oxidation products, calculate the m/z of the various theoretical oxidation products of each peptide^{4,26,27,28,29}.
5. Identify the extracted ion chromatogram of these m/z values using software to view the mass spectrometric run (Figure 4). Identify the peptide oxidation products based on their m/z, their charge state, and the similarity in elution time to the unmodified peptide. In our hands, peptide oxidation products elute between 240 seconds before to 180 seconds after the unmodified peptide using the LC gradient above. As oxidation will often result in multiple isomeric oxidation products, it is common to observe multiple partially resolved peaks in the extracted ion chromatograms of peptide oxidation products, as shown in Figure 4. Peptide oxidation products are quantified based on the area of the peak(s) in the extracted ion chromatograms.

Figure 4: Extracted ion chromatogram of a peptide and its oxidation products after FPOP. The m/z of the peptide oxidation products are calculated based on the m/z of the unoxidized peptide and the known oxidation products; and the areas of these peptide products are determined. The area of the peptide products is then used for the calculation of the average oxidation events per peptide. Please click here to view a larger version of this figure.

6. Calculate the average oxidation of the peptides using the following equation.

where P denotes the average number of oxidation events per peptide molecule, and I represents the peak area of the unoxidized peptide (Iunoxidized) and the peptide with n oxidation events. Note that I(singly oxidized) would include not only additions of a single oxygen atom but also other less-common single oxidation events that the investigator may choose to measure (e.g., oxidative decarboxylation, carbonyl formation, etc.)^{4,26,27,28,29}.

Representative Results

Comparison of the heavy chain peptide footprint of the adalimumab biosimilar in phosphate buffer and when heated at 55 °C for 1 h show interesting results. Student’s t-test is used for the identification of peptides that are significantly changed in these two conditions (p ≤ 0.05). The peptides 20-38, 99-125, 215-222, 223-252, 260-278, 376-413, and 414-420 show significant protection from solvent when the protein is heated to form aggregates (Figure 5)³⁰. This experiment identified the peptide regions that experience topographical changes upon heating and aggregation.

Figure 5: Peptide level footprint of the heavy chain of adalimumab. The peptide average oxidation of adalimumab (blue) at room temperature, and (orange) after adalimumab is heated at 55 °C for 1 hr, then cooled to room temperature. The error bars represent one standard deviation of triplicate measurements. The asterisk represents the peptides that are significantly changed in the two conditions (p ≤ 0.05). Please click here to view a larger version of this figure.

FPOP experiment of myoglobin was performed in the presence of 10 mM phosphate and 10 mM 2-(N-morpholino)ethanesulfonic acid (MES) buffer. MES buffer acts as a good scavenger of the hydroxyl radicals generated upon the photolysis of the hydrogen peroxide after the sample is exposed to the laser irradiation. The difference in the absorbance of the adenine is monitored using an inline dosimeter in real-time. The laser fluence is adjusted in a way to have a comparable change in the adenine absorbance level in MES buffer as compared to the phosphate buffer (Figure 6)¹⁷. The average oxidation of peptides was lower in the presence of MES buffer as compared to the phosphate buffer. However, as the laser fluence was increased to have an equal adenine dosimetry response, the average peptide oxidation values were not significantly different after FPOP in MES buffer and phosphate buffer (Figure 7)¹⁷. This experiment shows the importance of compensation of the signal to be able to compare the footprint with two FPOP conditions that have different scavenging capacity. Similar experiments have successfully used adenine-based compensation to probe structural changes of common excipients in adalimumab preparations³⁰.

Figure 6: Compensation of adenine dosimetry readings. The adenine reading before and after laser irradiation were recorded for FPOP in phosphate buffer at 265 nm with using inline dosimeter. As MES is a good scavenger of the hydroxyl radicals, the difference in adenine readings was lower. Increased the laser fluence of FPOP solution with MES buffer to “compensate” and overcome the effect of MES buffer to have similar adenine reading as phosphate buffer. Please click here to view a larger version of this figure.

Figure 7: Real-time compensation of myoglobin oxidation by inline adenine dosimetry. Myoglobin oxidized in (blue) 10 mM phosphate buffer and (orange) 10 mM MES buffer (orange). As noted, the oxidation of the peptides is lower in the MES buffer. As the laser fluence is increased for the samples in the MES buffer to have almost similar adenine dosimetry level as compared to phosphate buffer (grey), the peptide level oxidation is also similar to the oxidation level seen in samples with phosphate buffer. This figure has been adapted with permission from Analytical Chemistry 2018, 90, 21, 12625-12630. Copyright 2018 American Chemical Society. Please click here to view a larger version of this figure.

Discussion

Mass spectrometry-based structural techniques, including hydrogen-deuterium exchange, chemical cross-linking, covalent labeling, and native spray mass spectrometry and ion mobility have been rapidly growing in popularity due to their flexibility, sensitivity, and ability to handle complex mixtures. FPOP boasts several advantages that has boosted its popularity in the area of mass spectrometry-based structural techniques. Like most covalent labeling strategies, it provides a stable chemical snapshot of protein topography that is compatible with most post-labeling processes (e.g., trypsin digestion, deglycosylation, etc.), avoiding issues of back-exchange and scrambling that hinder hydrogen-deuterium exchange. However, unlike traditional covalent labeling technologies that target specific amino acids, FPOP is able to label a broad array of amino acids in a single experiment. Moreover, FPOP is able to complete the primary hydroxyl radical-protein reaction faster than the proteins are able to unfold to freeze a chemical snapshot of the native conformation¹⁴, although some secondary reactions may occur on a slower timescale^20,31,32. Unlike earlier experiments in hydroxyl radical protein footprinting using X-ray synchrotron beamlines, FPOP allows for this ultra-rapid labeling in a benchtop format^33,34. The major hurdles in FPOP faced by the typical protein mass spectrometry lab is experience in handling samples for FPOP, the laser-based oxidation, and data analysis. The goal of this report is to help new investigators overcome these hurdles to generate valuable and reproducible results.

Proteins may undergo background oxidation that may incorrectly provide the extent of oxidation on the identified peptides. To better understand this, a control sample is prepared in replicates along with FPOP samples in which all steps are carried out except the laser is not triggered (step 3.4). The level of oxidation in the no-laser control reveals the level of background oxidation. In-source oxidation of peptides may contribute towards the observed peptide oxidation, and can be readily determined by the LC elution profile of the unmodified peptide and the modification product. If the elution profiles overlap identically, the oxidation product is almost certainly due to post-column in-source oxidation. In-source oxidation may commonly be reduced by lowering the ionization voltage and/or increasing the distance between the emitter and the ion transfer tube. Other background oxidation may either be present in the protein prior to FPOP treatment, or it can be induced by exposure to hydrogen peroxide. The latter can be minimized by decreasing the concentration of hydrogen peroxide used and/or decreasing the time the protein is in the hydrogen peroxide prior to quenching.

A key problem often encountered by experimenters attempting FPOP for the first time is high background oxidation. This high background oxidation is usually due to the addition of hydrogen peroxide to the sample prior to analysis. While two-electron oxidation by hydrogen peroxide is much slower than one-electron oxidation by hydroxyl radicals, hydrogen peroxide is still easily able to oxidize certain amino acids (most notably methionine and cysteine) on a timescale of minutes. While the other components in the FPOP mixture are much more stable, hydrogen peroxide should be added directly prior to oxidation by FPOP. As a rule of thumb, the time between the addition of hydrogen peroxide and complete deposition of the sample into the quench solution should be kept to under five minutes. It is crucial to always collect a no-laser control (where the protein is handled as normal for FPOP but the laser is not fired) to detect any problems with sample oxidation, either due to prolonged peroxide exposure or due to protein purification and storage. For cases where the protein is particularly sensitive to hydrogen peroxide, on-line mixing with hydrogen peroxide prior to irradiation with the excimer laser can limit exposure to seconds or less^31,35. However, for most proteins, online mixing is unnecessary.

The next difficult hurdle for many novice FPOP experimenters is the setup of the optical path. It is important that the laser impinges squarely upon the capillary in order to get good photolysis of the hydrogen peroxide. While the UV laser light is invisible, it will cause many dyes present in colored paper to fluoresce in the visible range. Therefore, using a piece of colored construction paper on the backstop can help in the alignment of the lens and capillary. The paper can be used to ensure that the laser is squarely striking the center of the focusing lens, and then a piece of colored paper on the laser backstop can help tell when the capillary is successfully positioned in the center of the focused beam, as the diffraction of the capillary will cause a silhouette in the beam profile (Figure 3B).

It is also often not appreciated by novice investigators that the beam cross-sectional area changes based on the pulse energy. Therefore, if an investigator calculates the syringe pump flow rate based on a laser energy of 100 mJ/pulse, and then increases the laser energy to 120 mJ/pulse, the width of the laser beam will similarly increase causing the calculations to be incorrect. In order to prevent this issue, the use of an opaque aperture is recommended. For commercial excimer lasers we have worked with, when the laser energy per pulse is increased the largest change is in the cross-section of the beam, not the laser fluence. Since the concentration of hydroxyl radicals is based in part on the fluence of incident UV light, merely changing the laser energy per pulse is often inefficient at increasing effective hydroxyl radical dosage.

Reproducibility is another common hurdle for novice investigators to overcome. Most commonly, a lack of reproducibility is caused by a failure to generate equivalent hydroxyl radical doses across different replicates. This can be due to improper optical bench alignment, the unwitting use of different levels of hydroxyl radical scavenging agents, or the use of aged hydrogen peroxide. For all of the cases, the use of an internal dosimeter allows for the rapid identification of problems with effective hydroxyl radical dose. Hydroxyl radical dosimeters act by competing with the analyte (and with all scavengers in solution) for the pool of hydroxyl radicals; the effective dose of hydroxyl radicals is measured by measuring the amount of oxidation of the dosimeter. Note that “effective hydroxyl radical dose” is a function of both the initial concentration of hydroxyl radical generated, and the half-life of the radical. These two parameters are partially dependent on one another, making the theoretical kinetic modeling somewhat complex (Figure 2). Two samples could have wildly different initial radical half-lives while still maintaining the same effective radical dose by changing the initial concentration of hydroxyl radical formed; they will still generate identical footprints¹⁷. Adenine¹³ and Tris¹² are convenient hydroxyl radical dosimeters because their level of oxidation can be measured by UV spectroscopy in real-time, allowing for researchers to quickly identify when there is a problem with effective hydroxyl radical dose and to troubleshoot their problem.

Data analysis remains the most time-intensive part of any FPOP experiment. While reports exist using commercial packages to quantify oxidation products, these quantification algorithms have difficulties in properly defining peak areas in the partially resolved, asymmetrical peaks generated from groups of oxidation isomers²¹. In our hands, while current available automated software packages can usually correctly identify changes in oxidation, they often do not correctly quantify the magnitude of those changes (unpublished data), requiring post-analysis auditing and correction. Given the difficulties presented by FPOP data, the current status of available data analysis software able to handle FPOP quantification at all is remarkable; however, continued software development will benefit the field with increases in accuracy and reliability.

The protocol described here generates a peptide-level spatial resolution of hydroxyl radical protein footprints. It is possible to generate spatial resolution up to the amino acid level; however, disagreements in the field remain regarding the absolute accuracy of different methods of generating this high resolution FPOP data. A recent study comparing hydrogen-deuterium exchange and FPOP found that FPOP data can probe solvent accessibility at sub-amino acid level³⁶. One method uses HPLC to resolve oxidation isomers as much as possible, and then to quantify each isomer by peak area^1,23,24,25. However, when a simple mixture of synthetic peptide oxidation isomers was quantified by this method, errors were found in the absolute quantification and previous reports have indicated that collision-induced dissociation MS/MS can misidentify sites of oxidation^37,38. Quantification by electron transfer dissociation (ETD) has been shown to be accurate on synthetic standards and proteins, but direct application of this method requires co-elution of all oxidized peptide isomers which cannot be accomplished using reversed phase HPLC and generally requires size exclusion chromatography or HILIC^7,39,40,41; otherwise, complicated and time-consuming targeted ETD analyses must be used^7,8,9. The current consensus in the field seems to be that LC peak area-based amino acid level quantification seems to at least correctly identify sites of oxidation that change and correctly identify the relative amount of change (i.e., oxidation of amino acid X decreases by Y% in conformation A compared to conformation B), but the accuracy of quantification of the amount of oxidation (i.e., amino acid X is Y% oxidized) remains in dispute.

The strengths of FPOP as a flexible benchtop method for probing protein topography at many sites in a single experiment is driving continued interest in this technology, despite the current hurdles for the novice investigator. Commercial options for performing FPOP are just starting to come on the market; however, it remains quite possible for the interested investigator to develop their own FPOP optical bench and perform experiments using commonly available data analysis software. As the field grows and improvements to the available tools continue, the ease of access to FPOP technology will increase.

Materials

Name	Company	Catalog Number	Comments
Adenine	Acros Organics	147440250	Soluble in water upto 3.5 mM
Aperture	Edmund Optics	39-905	1000 μm Aperture Diameter, Gold-Plated Copper Aperture
Aperture holder	Edmund Optics	53-287	25.8mm Outer Diameter, Precision Pinhole Mount
Catalse	Sigma Aldrich	C-40	Catalase from bovine liver, lyophilized powder, ≥10,000 units/mg protein
COMPex Pro laser	Coherent	1113836	COMPexPRO 102, F-Vversion, KrF laser, No XeCl
Dithiotheitol (DTT)	Promega	V3151	DTT, Molecular Grade (DL-Dithiothreitol)
Fraction collector	GenNext Technologies, Inc.	N/A	Automated fraction collector
Fused silica capillay	Molex	1068150023	Polymicro Flexible Fused Silica Capillary Tubing, Inner Diameter 100 µm, Outer Diameter 375 µm, TSP100375
Glutamine	Acros Organics	119951000	L(+)-Glutamine, 99%
Holder for lens	Edmund Optics	03-668	53 mm Outer Diameter, Three-Screw Adjustable Ring Mount
Hydrogen peroxide	Fisher Scientific	H325-100	Hydrogen Peroxide, 30% (Certified ACS), Fisher Chemical
LC-MS/MS system	Thermo Scientific	IQLAAEGAAPFADBMBCX	Dionex Ultimate 3000 coupled to Orbitap Fusion Tribrid mass spectrometer
Mas spec grade Acetonitrile	Fisher Scientific	A955-1	Acetonitrile, Optima LC/MS Grade, Fisher Chemical
Mass spec grade formic acid	Fisher Scientific	A117-50	Formic Acid, 99.0+%, Optima™ LC/MS Grade, Fisher Chemical
Mass spec grade water	Fisher Scientific	W6-4	Water, Optima LC/MS Grade, Fisher Chemical
MES buffer	Sigma Aldrich	M0164	MES hemisodium salt
Methionine amide	Bachem	4000594.0005	H-met-NH2.HCl
Micro V clamp	Thor Labs	VK250	Micro V-clamp with stainless steel blades
Motorized stage	Edmund Optics	68-638	50mm Travel Motorized Stage System with Manual Control
Nano C18 colum	Thermo Scientific	164534	Acclaim PepMap 100 C18 HPLC Columns
Optical bench	Edmund Optics	56-935	18" x 18" breadboard
Pioneer FPOP Module System	GenNext Technologies, Inc.	N/A	Inline FPOP Radical Dosimetry System
Post holder	Edmund Optics	58-979	3" Length, ¼-20 Thread, Post Holder
Sodium phosphate dibasic	Fisher Scientific	BP331-500	Sodium Phosphate Dibasic Heptahydrate (Colorless-to-White Crystals), Fisher BioReagents
Sodium phosphate monobasic	Fisher Scientific	BP330-500	Sodium Phosphate Monobasic Monohydrate (Colorless-to-white Crystals), Fisher BioReagents
Syringe	Hamilton	81065	100 µL, Model 1710 RN SYR, Small Removable NDL, 22s ga, 2 in, point style 3
Syringe pump	KD Scientific	788101	Legato 101 syringe pump
Trap C18 column	Thermo Scientific	160454	Thermo Scientific Acclaim PepMap 100 C18 HPLC Columns
Tris	Sigma Aldrich	252859	Tris(hydroxymethyl)aminomethane
Trypsin	Promega	V5111	Sequencing Grade Modified Trypsin
UV plano convex lens	Edmund Optics	84-285	30 mm Dia. x 120 mm FL Uncoated, UV Plano-Convex Lens