Combining Chemical Cross-linking and Mass Spectrometry of Intact Protein Complexes to Study the Architecture of Multi-subunit Protein Assemblies

1. Purification of Protein Complexes

1. Prepare protein complex according to optimized standard protocols.
   NOTE: The protocol is demonstrated here with the RvB1/B2 complex from Chaetomium thermophilum and the carbamoyl phosphate synthase (CPS) from Escherichia coli. RvB1/B2 and CPS were purified as described^36,37. Every protein complex requires an individual purification protocol. Adjust the protocol accordingly. Amine-free buffers such as phosphate-buffered saline (PBS) or 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES) should be used for chemical cross-linking. Replace the buffer during purification if possible.
   CAUTION: Do not apply any methods or reagents that disturb the native assembly of the protein complex.

2. Mass Spectrometry-based Protein Identification

1. Gel electrophoresis
   NOTE: There are different gel systems available and every laboratory uses its own setup. Adjust the conditions according to the gel system. Wear gloves and lab coat throughout the protocol as keratins are among the most common contaminants in mass spectrometric analysis.
   1. Apply 7.5 µL 4x sample buffer and 3 µL 10x reducing agent (final concentration 50 mM dithiothreitol (DTT)) to 20 µL protein sample. Use 10 µM CPS or 2.5 µM RvB1/B2. Spin down for 1 min at 16,200 × g and heat for 10 min at 70 °C.
      NOTE: When using a different gel system adjust the required volumes and heating temperature.
   2. Prepare a 4 – 12% gradient gel for electrophoresis. Dilute the running buffer 20 times with water and fill the electrophoresis chamber. Use 2-(N-morpholino)ethanesulfonic acid (MES) buffer for proteins of smaller molecular weight (2 – 200 kDa) and 3-(N-morpholino)propanesulfonic acid (MOPS) buffer for larger proteins (14 – 200 kDa). Add 0.5 mL antioxidant to the inner chamber.
      NOTE: When using a different gel system prepare running buffer and the gel according to manufacturer's protocols.
   3. Load a suitable protein marker (pre-stained, unstained, different molecular sizes) in the first cavity and load the protein samples into the remaining cavities.
   4. Separate proteins for 35 min (MES) or 50 min (MOPS) at 200 V.
      NOTE: Adjust the electrophoresis times and voltage for different gel systems.
   5. To stain the protein bands, transfer the gel into a gel staining box and cover the gel with a water-based Coomassie staining solution. Incubate overnight and at room temperature on a horizontal gel shaker.
   6. Destain the gel by replacing the staining solution with water. Repeat this step several times (approximately 3-5 times) until the gel background appears clear.
2. In-gel digestion
   NOTE: To avoid contamination, use HPLC grade solvents throughout the digestion protocol and for all following steps (i.e., mass spectrometric analysis). Prepare all solutions prior to use and filtrate water and ammonium bicarbonate.
   1. Cut the protein bands that are visualized by blue Coomassie stain from the gel using a scalpel. Carefully cut the protein bands into small pieces of approximately 1 mm × 1 mm. Rinse the scalpel with water between different protein bands. Wash the gel bands with water and acetonitrile (ACN).
   2. Reduce disulphide bonds with DTT, alkylate cysteine residues with iodoacetamide, and digest proteins with trypsin as previously described³⁸; usually an enzyme:protein ratio of 1:20 up to 1:100 is used. Use 100 mM ammonium bicarbonate during the digestion protocol.
   3. Extract peptides in two steps.
      1. First incubate the gel pieces with ammonium bicarbonate and ACN, and collect the peptide-containing supernatant. Second, incubate the gel pieces with 5% (v/v) formic acid and ACN. Incubate for 15 min at each step. Combine both supernatants. Dry the extracted peptides by evaporating the solvents in a vacuum centrifuge.
         NOTE: Dried peptides can be stored at -20 °C for several months.
3. Liquid chromatography-coupled mass spectrometry
   1. Dissolve the dried peptides in 2% ACN/0.1% formic acid. Dissolve the peptides in a sonication bath for 2-3 min and spin down in a centrifuge at 16,200 x g for 30 min. Transfer the samples into autosampler vials.
      NOTE: Adjust the volume according to the protein amount. For well-stained Coomassie bands we use 20 µL.
   2. Inject 5 µL of the sample to the nano-LC-MS/MS system using the autosampler. Load the peptide mixture onto a reversed-phase C18 pre-column (C18, 150 µm I.D., 2 cm, 5 µm pore size) to desalt and concentrate the peptides online.
   3. Use 0.1% (v/v) formic acid as mobile phase A and 80% ACN/0.1% (v/v) formic acid as mobile phase B. Separate the peptides on a reversed-phase C18 analytical column (C18, 75 µm I.D., 50 cm, 3 µm pore size) using a gradient of 4-80% B (containing 0.1% formic acid) at 300 nL/min over 65 min.
      NOTE: A nanospray ion source is used to transfer eluted peptides into the mass spectrometer.
   4. Use (typical) MS conditions: spray voltage of 1.6 kV; capillary temperature of 250 °C; normalized collision energy of 30. Operate the mass spectrometer in data-dependent mode.
   5. Acquire the MS spectra in the mass analyzer (e.g., orbitrap) (m/z 350−1,600) with a resolution of 70,000 and an automatic gain control target of 3 x 10⁶. Select 20 of the most intense ions for HCD fragmentation at an automatic gain control target of 1 x 10⁵. Dynamically exclude previously selected ions for 30 s. Exclude singly charged ions as well as ions with unrecognized charge state.
      NOTE: Internal calibration of the mass spectra was performed using the lock mass option³⁹.
4. Database searching
   NOTE: There is different software available for database searching. MaxQuant⁴⁰, for instance, is freely available.
   1. Convert .raw-files into .mgf-files using a pXtract conversion tool (http://pfind.ict.ac.cn/downloads.html).
   2. Perform database search using typical search parameters: Database, swissprot; peptide mass tolerance, 10 ppm; fragment mass tolerance, 0.5 Da; enzyme, trypsin; missed cleavage sites, 2; variable modifications, carbamidomethylation (cysteine) and oxidation (methionine).
      NOTE: Change the search parameters according to the experimental settings.
   3. Inspect the database search result. Evaluate the protein score, number of peptides identified, peptide scores, and mass accuracy; the sequence coverage returns the protein coverage identified.
      NOTE: Every search engine has its individual scoring algorithm. Evaluate the peptide scoring system by the quality of the tandem mass spectra observed. The main signals of a good spectrum should show (complete) ion series to verify the identified peptide. Usually the peptide score is probability-based, i.e., the peptide score is a measure for how likely the identified peptide sequence matches the obtained spectrum. The protein score is usually derived from peptide scores and is used to rank a protein in a list of identified proteins.

3. Mass Spectrometry of Intact Protein Complexes (Native Mass Spectrometry)

1. Preparation of gold-coated emitters for electrospray ionization
   1. Use a micropipette puller to prepare nanoflow emitters from glass capillaries as described previously^25,41. Use borosilicate capillaries with an inner diameter of 0.78 mm.
      NOTE: By changing the parameters for needle pulling, the tip shape and size can be modified and adjusted for the sample. Capillaries with different inner diameter and wall thickness are available.
   2. Coat the glass capillaries with conductive material (e.g., gold or palladium); the use of a sputter coater generating a gold plasma is common. Follow the manufacturer's instructions to obtain good quality coating.
      NOTE: The coating should be sufficient enough to obtain a stable spray when applying common capillary voltages (see below).
2. Sample preparation for native mass spectrometry
   NOTE: Salts, detergents, or large amounts of glycerol are not compatible with electrospray ionization. Therefore, the purification buffer is exchanged by a volatile, aqueous buffer. 200 mM ammonium acetate is commonly used. Use size exclusion spin columns or ultrafiltration devices for buffer exchange. In some cases, complex stability or activity might be affected by buffer exchange. Evaluate the mass spectra carefully and check the activity of the complex. Add cofactors or additives to the analysis buffer, if necessary.
   1. Use size exclusion spin columns for fast buffer exchange. Remove the storage buffer by centrifugation at 1,000 x g and 4 °C for 1 min. Discard the flow through. Wash three times by adding 500 µL 200 mM ammonium acetate, followed by centrifugation. Load 20 µL of the protein sample onto the column and centrifuge at 1,000 x g and 4 °C for 4 min.
      NOTE: The concentration of the protein complex should be 1-10 µM. The procedure can be repeated if non-volatile components still disturb the analysis.
   2. To concentrate the protein sample and exchange the buffer in the same experiment use centrifugal filters. Use a filtration membrane of pore size 50% smaller than the size of the proteins analyzed.
      1. Transfer the protein sample into the filtration device and add 200 mM ammonium acetate. Spin down at 15,000 x g and discard the flow through. Add 200 mM ammonium acetate and repeat the centrifugation. Repeat this step several times. Follow the manufacturer's instructions for centrifugation speed. Perform the centrifugation at 4 °C.
         Caution: Membrane proteins tend to precipitate on the membrane of the filter device.
3. Mass spectrometric analysis of intact protein complexes
   NOTE: There are different types of mass spectrometers from different manufacturers that can be modified for native mass spectrometry, e.g., quadrupole time-of-flight (Q-ToF) mass spectrometers or orbitrap mass spectrometers. The protocol described below was performed on a Q-ToF instrument.
   1. Place a gold-coated capillary in the capillary holder and fill the capillary with 1-4 µL of protein sample. Open the tip of the needle with a tweezer.
   2. Connect the capillary holder with the nano-electrospray source and adjust the capillary position. Position the tip of the capillary at 0.5-1.5 cm to the cone orifice. Use 80-150 L/h nanoflow gas to initiate the spray and adjust the gas flow to maintain a stable spray.
   3. Adjust parameters in the tune-page of the Q-ToF instrument. Typical starting conditions are: capillary voltage, 1.50 kV; cone voltage, 80 V; RF lens 1 energy, 80 V; collision energy, 20 V; aperture 3, 13.6 V. Modify these parameters to obtain good mass spectra. Start the acquisition by clicking the "acquire" button and combine as many scans as possible to obtain a good mass spectrum.
      NOTE: We recommend combining at least 100 scans.
4. Tandem mass spectrometry of intact protein complexes
   1. Acquire the mass spectrum as described above (protocol section 3.3). Choose a precursor ion of a protein complex.
   2. Change from MS into MS/MS mode in the acquisition file. Set the MS/MS selection to the precursor mass.
   3. Start the acquisition at low collision energy. Combine several scans (approximately 20 scans) to verify selection of the correct precursor mass. Increase the collision energy until the protein complex dissociates. To obtain a good mass spectrum combine at least 500 scans.
      NOTE: Stripped complexes sometimes have low intensity. Combining as many scans as possible might increase the resolution and signal-to-noise ratio.
5. In-solution dissociation of intact protein complexes
   NOTE: To gain additional insight into protein interactions within the protein complexes, in-solution dissociation can be performed.
   1. Prepare the protein sample as described above (section 3.2). Add solvents to the protein sample or change the pH to dissociate intact complexes into sub-complexes. Typical solvents are methanol, isopropanol, and ACN; change the pH of ammonium acetate by the addition of acetic acid or ammonia solution. Acquire mass spectra as described above (sections 3.3 and 3.4).
   2. Vary the amount of solvent or the pH range to generate various sub-complexes. Typically, the amount of solvent is 5-50% (final concentration) and a typical pH range is 4-9. Start with a low concentration of solvent (5%) or slight changes in pH and acquire a mass spectrum (see section 3.3).
   3. Increase the amount of solvent or change the pH stepwise until sub-complexes are generated in solution. Acquire mass spectra to analyze sub-complexes.
6. Calibrate data
   NOTE: Acquired mass spectra are calibrated externally using Cesium iodide (CsI) solution.
   1. Dissolve 100 mg CsI in 1 mL water.
   2. Acquire a mass spectrum of CsI. Vary the collision energy to obtain CsI clusters over the same m/z range as the protein complex analyzed above.
      CAUTION: CsI quickly precipitates at the emitter tip and contaminates the cone. Acquire only as many scans as required to gain a sufficient spectrum. Remove the emitter from the source when finished.
   3. Make a calibration file using the acquired mass spectrum and a CsI reference file.
   4. Apply calibration to acquired mass spectra.
      CAUTION: Calibration of mass spectra might be a permanent change of the raw data. If non-calibrated spectra are needed, make a back-up copy of the file.
7. Data processing and analysis
   NOTE: There are many freely-available software tools for data analysis of native mass spectra; for instance, Massign⁴² or UniDec⁴³. The protocol below describes manual data analysis with the help of instrument software as well as the use of Massign for complex samples. This software is well-suited for the analysis of complex mass spectra. Follow the instructions provided online for use of the program (http://massign.chem.ox.ac.uk/).
   1. For data analysis, smooth spectra by adjusting smoothing parameters. Centroid spectra by adjusting parameters. Calculate complex masses from two adjacent peaks of the protein complex' peak envelope using instrument software tools.
      CAUTION: Too intensive smoothing might cause data loss (e.g., loss of bound ligands).
   2. For analysis with Massign⁴², generate a peak list for the mass spectrum. Linearize data points of the spectrum and smooth. Use the various software tools to assign protein complexes, calculate complex composition, or simulate complex peak envelopes.

4. Chemical cross-linking Coupled with Mass Spectrometry

NOTE: There are numerous cross-linking strategies available. Here, we describe the use of Bis(sulfosuccinimidyl)suberate (BS3), an amine-reactive cross-linker which is commonly used to study protein-protein interactions.

1. Dissolve 1.43 mg BS3 in 100 µL water to prepare a 25 mM stock solution.
   NOTE: Other reagents like disuccinimidyl suberate (DSS) are not water-soluble and are usually dissolved in dimethyl sulfoxide (DMSO). Some cross-linkers are also available in heavy isotopically-labeled forms. Incorporation of heavy stable isotopes generates peak pairs of cross-linked di-peptides in MS spectra, which helps during data evaluation. When using differentially labeled cross-linkers, stock solutions of both variants are prepared and mixed 1:1.
   CAUTION: To avoid dissociation of the protein complexes by DMSO, a highly concentrated stock solution is prepared and diluted with water or buffer prior to the cross-linking reaction.
2. Add BS3 to the protein complex. Use varying amounts of BS3 ranging from 0.5-5 mM to identify optimal cross-linker concentration. Incubate the reaction mixtures at 25 °C for 1 h in a thermomixer. Use 4-12% gradient gels and perform SDS-PAGE to evaluate cross-linking results (see Figure 2 for an example).
   NOTE: Optimal concentration of BS3 is reached when higher molecular weight protein bands, which are not visible in the non-cross-linked control, are obtained during SDS-PAGE while single subunits are still visible (Figure 2).
3. Repeat the cross-linking reaction with the optimal BS3 concentration. Incubate the reaction mixtures at 25 °C for 1 h in a thermomixer.
   NOTE: Some protein complexes are not stable at room temperature. The cross-linking reaction can also be performed on ice; however, the reaction time needs to be adjusted.
4. Quench the cross-linking reaction by adding amine buffer (e.g., 50-100 mM Tris buffer, pH 7.5, final concentration) or perform ethanol precipitation to remove any residual cross-linking reagent.
   1. Add water or buffer to the reaction mixture to reach a final volume of 200 µL. Add 600 µL ice-cold ethanol and 20 µL 3 M sodium acetate, pH 5.3. Mix thoroughly and incubate at -20 °C for 2 h or overnight.
      NOTE: Alternatively, the proteins can be precipitated at 80 °C for 30 min or in liquid nitrogen.
   2. Spin down at 16,200 x g and 4 °C for 30 min. Carefully remove the supernatant.
   3. Wash the pellet with 1 mL ice-cold 80% (v/v) ethanol. Spin down at 16,200 x g and 4 °C for 30 min. Carefully remove the supernatant. Dry the pellet in a vacuum centrifuge.
5. Perform SDS-PAGE of the cross-linked proteins. Cut the gel bands and digest the protein in-gel as described above (sections 2.1 and 2.2).
   NOTE: In-solution digestion can also be performed, however, it usually requires additional separation steps (e.g., size exclusion chromatography of the digested peptides).
6. Perform liquid chromatography-coupled mass spectrometry as described above (section 2.3). As cross-linked peptides are usually low abundant, apply the following variations to the mass spectrometric analysis to increase analytical depth of the cross-linked sample.
   1. Use longer gradients during liquid chromatography separation (e.g., 90 min instead of 65 min, see above).
   2. Exclude the doubly charged peptides from HCD fragmentation.
      NOTE: Doubly charged peptides are usually intra-cross-linked peptides ("loop peptides" or "type 1" cross-links).
   3. When using differentially labeled cross-linking reagents, use the "peak picking" option during analysis, i.e., HCD fragmentation is triggered by the presence of peak pairs of defined mass difference in the mass spectra.
      NOTE: The "peak picking" option might not be available on every mass spectrometer.
7. Use pLink software¹⁸ for identification of cross-linked di-peptides. Use minimized databases for identification. Typical search parameters are: instrument spectra, HCD; enzyme, trypsin; max. missed cleavage sites, 3; variable modifications, oxidation (methionine) and carbamidomethylation (cysteine); cross-linker, BS3; min. peptide length, 4; max. peptide length, 100; min. peptide mass, 400 Da; max. peptide mass, 10,000 Da; FDR, 1%.
   NOTE: When using differentially labeled cross-linkers, the mass increase caused by the linker needs to be configured. There is also other commonly used software for cross-link identification, e.g., xQuest¹⁷, MassMatrix¹⁹, or XlinkX¹⁵.
8. Evaluate the database search results by the quality of fragmentation spectra. Acceptable spectra of cross-links should show ion series of both peptides (at least 4 adjacent ions) at a reasonable signal-to-noise ratio.
   NOTE: When using differentially labeled cross-linking reagents, peak pairs in MS spectra can be used as an additional quality control.
9. If required, visualize the cross-linking results in protein interaction networks using software tools (e.g., XVis, XiNET). Use bar plots or circular plots for visualization of protein interactions.
   NOTE: Both software tools are freely available on a web server. Follow detailed instructions on the respective websites (https://xvis.genzentrum.lmu.de/ and http://crosslinkviewer.org/).