A native mass spectrometry approach to qualitatively elucidate interfacial epitopes of transient protein–protein interactions

Native mass spectrometric analysis of TPR2A and GrpE with unpurified peptides derived from limited proteolysis of their respective PPI partners (HSP90 C-terminus and DnaK) facilitated efficient, qualitative identification of interfacial epitopes involved in transient PPI formation. Application of this approach can assist in elucidating interfaces of currently uncharacterised transient PPIs.

BL21 (DE3) Codon+ strain as fusion proteins with an N-terminal His-SUMO tag from the pCA258 backbone (a kind gift of Matthias Mayer, ZMBH, Heidelberg).His-SUMO-DnaK and His-SUMO-GrpE proteins were purified by Ni-NTA chromatography as described above and buffer exchanged into storage buffer (25 mM NaH 2 PO 4 pH 7.5, 150 mM NaCl, 150 mM KCl).The His-SUMO tag was removed by addition of His-tagged SUMO protease (purified in house) for 4 hours at 4 ⁰C with gentle rotation.Protein mixtures were subjected to a second round of Ni-NTA chromatography to remove the SUMO protease and His-SUMO tag, and then passed through a 10 kDa MWCO filter to obtain untagged DnaK and GrpE proteins.

Sample Preparation a. Protein desalting
TPR2A, HSP90-C, DnaK and GrpE samples were subjected to two rounds of buffer exchange into 100 mM NH 4 OAc using Zeba Spin Desalting Column (Thermo Fisher Scientific).Concentration of all proteins ere adjusted to stock solutions of 250 µM.

b. HSP90-C digestion
100 µL of agarose supported trypsin (Thermo Fischer) was washed three times with 500 µL NH 4 OAc buffer (100 mM), and finally resuspended in 200 µL of the same NH 4 OAc buffer.50 µL of the desalted HSP90-C or DnaK stock solutions (250 µM) were added to this trypsin suspension.The mixture was incubated at 37 °C on a shaking heat block.Digest aliquots were removed at indicated intervals and used directly for native MS analysis.

c. Preparation of Peptide 10 and 11 stocks
Stock solutions (250 µM) of peptides 5 and 6 were prepared in 100 mM NH 4 OAc.

d. Sample preparation for Native MS
TPR2A-HSP90C: 1 µL of the TPR2A stock (250 µM) was utilised for analysis and made up to a final volume of 20 µL (final protein concentration 12.5 µM).HSP90-C peptide binding experiments were prepared by mixing 2 µL aliquots of the digest mixtures with 1 µL of the TPR2A stock, and 17 µL of 100 mM NH 4 OAc.Similarly, binding competition samples were prepared HSP90-C peptide binding experiments were prepared by mixing 2 µL aliquots of the digest mixtures with 1 µL of the peptide stock solution (final conc.12.5 µM) 1 µL of the TPR2A stock, and 16 µL of 100 mM NH 4 OAc.Electronic Supplementary Material (ESI) for ChemComm.This journal is © The Royal Society of Chemistry 2024

Mass Spectrometry
Native MS were obtained on a 12T SolariX 2XR or 7T FTICR (Bruker Daltonics).Ionisation was achieved using a NanoMate nESI infusion robot (TriVersa Biosciences), sampling from a 96-well plate.Typically, a nanoelectrospray voltage of 1.55 kV was used and backing pressure was adjusted to maintain stable electrospray.Instrument conditions were optimised for transmission of native protein ions.Typical source optics were: capillary exit 250V, deflector plate 230V, funnel 1 210V, skimmer 1 50V, funnel2 15 and skimmer 2 8V.Native mass spectra were acquired as the sum of 200 1 MegaWord FID transients.DataAnalysis software (Bruker Daltonics) was used for native MS analysis.LC-MS data was acquired on a Synapt G2 q-TOF mass spectrometer, coupled to an Acquity UPLC (Waters) equipped with a 50 x 2.1 mm C18 reverse phase column (Phenomenex).Chromatography was performed at 298K.Mobile phase A (0.1% TFA in H 2 O) and mobile phase B (0.1% TFA in CH 3 CN) were set at 95% (A) and 5% (B) with a gradient running to 5% (A) and 95% (B) over 20 minutes.LC-MS data were processed using MassLynx v4.0 (Waters).

Fig. S2.
High-resolution ESI FT-ICR MS analysis of the isotope distributions of TPR2A in complex with peptides 1 -4.For each species, the calculated theoretical isotope distribution for the TPR2A-peptide complex is overlaid as a scatterplot and are representative of the following molecular formulae:  10+ .These data facilitated unambiguous Δ m/z assignment and binding peptide identification.7KW7). 1 The HSP90 (pink tube) association with HOP( green van der Waals surface), is highlighted, while the HSP70 and GR are greyed out for clarity.In this structure, the primary interfacial interaction between TPR2A and MEEVD has been resolved.However, the sequence, which connects MEEVD to the reminder of HSP90 is not included in the structure.This omitted region corresponds closely to peptide 2, identified in this study as a TPR2A binder.

Data Analysis
For each spectrum a mass list and accompanying peak areas was generated using the FTMS algorithm (S/N threshold of 5).Protein -Peptide complexes were identified via Δm/z as compared to apo protein at each charge state (Table S1).The area of each peak corresponding to a TPR2Apeptide complex from each of the three native charge states (10+, 9+ and 8+) was combined.TPR2A-peptide association was calculated from the summed peak area of each species as a percentage of the total peak area for each species.TPR2A-peptide associations for each individual charge state was calculated from the individual peak area as a percentage of the total peak area at that charge state (Table S2).Figures are the average of two replicates.
A list of possible tryptic cleavage products was generated using Expasy PeptideMass 3 with missed cleavages set to 5. The crude mixture following 18 hours of tryptic digestion was subjected to UPLC-MS, where tryptic cleavage product were identified by their mass signature (Figure S3).

Supplementary Tables and Figures
-DnaK: GrpE -DnaK PPI formation experiments were conducted by mixing 2 µL of the GrpE stock (250 µM) with 1 µL of the DnaK stock (250 µM) and made up to a final volume of 20 µL in 100 mM NH 4 OAc (final protein concentration 25 µM GrpE and 12.5 µM DnaK).GrpE DnaK peptide binding experiments were prepared by mixing 4 µL aliquots of the digest mixtures with 2 µL of the GrpE stock, and 14 µL of 100 mM NH 4 OAc.All samples were made up in 96-well plates and held at 4 C prior to MS analysis.

.
Fig. S3.Depiction of the sequence overage of peptides derived from tryptic digestion of HSP90-C (A) and DnaK (B) respectively.Peptides coloured red indicate binding to either TPR2A (A) or GrpE (B.).Colours used for non-binding peptides are varied, only for the purposes of figure clarity.A) Residues highlighted in orange, indicate known interfacial associating epitope (MEEVD).Residues highlighted in blue, correlate with previously unreported TPR2A interacting region. 1 B).Residues highlighted in orange, correlate to those mutated by Li and co-workers to assess influence of various contact areas of the DnaK-GrpE PPI. 2

Fig. S4 .
Fig.S4.Cryo-EM structure of the GR-HSP90-HSP70-HOP complex (PDB 7KW7).1The HSP90 (pink tube) association with HOP( green van der Waals surface), is highlighted, while the HSP70 and GR are greyed out for clarity.In this structure, the primary interfacial interaction between TPR2A and MEEVD has been resolved.However, the sequence, which connects MEEVD to the reminder of HSP90 is not included in the structure.This omitted region corresponds closely to peptide 2, identified in this study as a TPR2A binder.

Fig. S7 .
Fig.S7.Cryo-EM structure of DnaK (grey) bound to two subunits of GrpE (blue and purple, PDB 8GB3).2This PPI is formed through three major contact regions as shown.Mutational analysis showed that a dual E236A and Y257A mutation abolished PPI formation and DnaK chaperone activity.

Fig. S8 .
Fig. S8.Zoomed in portion of contact region 1.Overlapping peptides 7 and 8 were found in the alpha helical region highlighted in red, and includes key E236 residue, which forms salt bridges with Arg169 and 180 on the same GrpE monomer (blue)

Fig. S10 .
Fig. S10.High-resolution ESI FT-ICR MS analysis of the isotope distributions of GrpE in complex with peptides 7 and 8.For each species, the calculated theoretical isotope distribution for the GrpE-peptide complex is overlaid as a scatterplot and are representative of the following molecular formulae: 1. [C 1165 H 18174 N 355 O 423 S 4 + 10H] 10+ ; 2. [C 1210 H 1916 N 2360 O 426 S 4 +10H] 10+ .These data facilitated unambiguous Δ m/z assignment and binding peptide identification.