Evolution of Structural Biology through the Lens of Mass Spectrometry

Using Native MS To Study Membrane Proteins.

Understanding membrane protein structures is crucial, because they play essential physiological roles and make up a majority of therapeutic targets. Membrane proteins have been challenging for biophysical studies because of low physiological expression levels, the insoluble nature of biological membranes, and their heterogeneity. Detergents have been used to study membrane protein structure for native MS but may give rise to the destabilization of protein structure, protein–protein interactions, and protein–ligand interactions. Recently, alternative membrane mimetics such as amphipols, lipid nano-structures, liposomes, and intact nanodiscs have been used to create biologically relevant approaches for native MS of membrane proteins.¹² Nanodiscs are nanoscale lipoprotein particles consisting of a lipid bilayer surrounded by two membrane scaffold protein (MSP) belts. Nanodiscs have been shown to have extraordinary gas-phase stability when they are ionized by native ESI.^13,14 To investigate the disassociation of nanodiscs in the gas phase, the collisional-induced dissociation (CID) energy or the multiphoton dissociation energy was increased.¹⁵ A shift of the nanodisc ions to lower m/z values showed that nanodisc complexes lost both mass and charge as they are activated. Lipid composition of heterogeneous nanodiscs was determined by employing lipids of slightly different masses. Nanodiscs were prepared with palmitoyl-oleoyl-phosphatidylcholine (POPC), palmitoyl-oleoyl-phosphatidylglycerol (POPG), and palmitoyl-oleoyl-phosphatidyl-serine (POPS) in different ratios. The nanodiscs displayed similar composition at low collisional energy, but at higher collisional energies, they displayed a polarity dependent depletion of certain lipids, suggesting that the chemistry of the lipid molecules played a crucial role in dissociation mechanisms.¹⁵

The integrity of intact membrane protein nanodiscs was assessed by using two membrane protein oligomers, trimeric AmtB and tetrameric AqpZ, in nanodiscs with different lipid compositions.^15,16 Distinct features of the membrane protein nanodiscs showed variation as a function of collisional energy. At high collisional energy, the nanodisc complex disassociated into the lipids, membrane scaffold proteins (MSPs), and membrane protein monomers (Figure 2). At an intermediate collisional energy, the AmtB trimer was detected with nine lipids bound. At low collisional energy, the majority of the scaffold proteins and lipids were removed, leaving only the membrane protein oligomer and any lipids in contact with the protein surface. A challenge of this method is the m/z overlap between the MSP belts and the lipids. By designing multiple nanodiscs with different lipid compositions and/or MSP belts, their isobaric masses can be distinguished. One problem with this approach is the possible disruption of the protein complex and the time and cost of designing multiple nanodiscs for an experiment. Reid et al. performed a study with mutated MSP belts that resulted in subtle mass shifts to distinguish the MSP belts from the protein-bound lipids.¹⁶ These changes do not disrupt the interaction between protein-bound lipids and/or nanodisc assembly. By adding one to two amino acids to the N-terminus of the MSP belt, the nanodisc structure is preserved, and both the MSP belt and protein-bound lipids can be distinguished. This study demonstrates how native MS with membrane protein nanodiscs can be used to determine the stochiometry of lipid–protein complexes.

Ultrahigh-Mass Range Instrumentation.

High-molecular-weight complexes have been successfully analyzed on time-of-flight mass spectrometers, but recent developments on Orbitrap mass spectrometers have extended the m/z range of these instruments to detect ion signals for biomolecular molecules up to MDa in size. Heck and co-workers modified a Q Exactive (QE) Plus mass spectrometer to a lower RF frequency of the ion optics to allow for better transmission characteristics of high-m/z ions.^9,10 One limitation of high-mass ions is that as they fly from the ion optics through the bent flatapole, their high momentum causes rapid transfer of ions from atmospheric pressure to high vacuum, which causes ion instability and leads to sensitivity loss. The addition of in source trapping within the injection flatapole decreases the ion momentum prior to entering the bent flatapole, therefore increasing sensitivity.¹⁰ Modifications to the stacked ring ion guide (SRIG) exit lens (E1), injection flatapole (E2), and injection flatapole lens (E3) allow for independent control of voltage and two operation modes, normal operation and trapping enabled. In normal operations, the voltage among the E1, E2, and E3 lenses are unmodified. In trapping mode, ions are subjected to a negative potential gradient change between the E1 and the E2, while the injection E3 potential is increased. The difference in voltage among the E1, E2, and E3 enables a trap in E2 for the ions. In addition, ions trapped in the E2 collide with background gas to decrease their momentum. The electrical potential between the ion optics are pulsed back to normal operation, and ions are allowed to exit E2 through E3 reaching the bent flatapole. These modifications allow for transmission of high m/z ions to the C-trap, but transmission of ions from the C-trap into the mass analyzer need to be modified as well. Under standard conditions, when ions are ejected from the C-trap to the analyzer, a deflector electrode turns the ions into the entrance slot of the Orbitrap. The deflector is pulsed with a voltage to minimize field perturbations and “block” any ions from entering the analyzer. The standard pulse is not long enough to allow large m/z ions flight from the C-trap to the analyzer, where the pulse time is increased to allow high m/z ions to enter the analyzer, keeping the Orbitrap “open” for longer. This ultrahigh-mass range (UHMR) instrument has enhanced detection of a homogeneous Cp180-mer virus-like protein, showing well-resolved ion signals up to 70 000 m/z (Figure 3).¹⁰ This mass range has not been achieved by other MS instruments.

Waterbeemd et al., demonstrated the use of the QE-UHMR to study the prokaryotic 30S, 50S, and 70S ribosomal particles, enabling high-resolution MS analysis of large molecules from 0.8 to 2.3 MDa. Native MS of the ribosomal particles from E. coli revealed substochiometric association of the small protein SRA.⁹ This demonstrates the high sensitivity of the instrument since LC/MS/MS studies indicated that only ~22% of the 30S particles are bound to SRA. In addition, the mass of the 30S particle established the presence of the RS1 protein. This protein, which is essential for translation, is highly dynamic and is therefore not present in any high-resolution structure. This indicates the ability of native MS to study dynamic complexes not possible by other structural biology methods. Native MS was able to detect the heterogeneity (i.e., ribosome-interacting proteins that are recruited at different stages of translation) within the ribosomal particles, which has been proven difficult by higher resolution structural methods. This type of high-resolution MS opens doors for studies of ribosomal-binding drugs, antibiotics, and initiation or elongation factors or other large complexes associated with disease states.

Characterizing Antibodies with Ion Mobility.

The developments in structural proteomics have allowed large biomolecules, like antibodies, to be characterized successfully.^19–26 For many years, monoclonal antibodies (mAbs) have been the top therapeutic candidates for various diseases including cancer, autoimmune diseases, and infections. Detection of small structural changes are necessary to monitor the quality of antibodies, compare isoforms, and characterize biosimilar monoclonal antibodies. Current IMS technology does not have the resolving power to distinguish small structural changes in larger proteins.²⁷ It has been shown that heating gas-phase ions with a small amount of collisional energy immediately before IMS has improved separation.²⁸ In 2014, Zhong et al. utilized collision-induced unfolding (CIU) to improve the detection of small structural changes in large biomolecules (Figure 4).²⁷ CIU traps gas-phase native proteins and slowly increases their internal energy with collisional energy of a background gas, like argon.²⁹ By limiting the extent of collisional energy, proteins unfold, but the integrity of the amide backbone is maintained.^30–33 The gas-phase unfolding pattern, or the “fingerprint” plot of the protein, gives insight into protein stability. This has allowed small structural changes, like changes in disulfide bonds, differing glycosylation, or epitope location to be detected. By comparing CIU fingerprints, many different IgG subtypes have been distinguished that differ in either the number of disulfide bonds or locations between the heavy and light chains.²⁵ No other method has shown to distinguish these changes in the interdomain connectivity as quickly and easily by comparing the CIU fingerprints.

CIU has also been used to determine differing extents of mAb glycosylations and structural differences to aid in the comparison of biosimilar mAbs.³⁴ Recently, Remsima was the first biosimilar mAb to get FDA approval for clinical use, and in the next few years, patents of many other mAbs will expire. Developing biosimilar mAbs will provide significant cost savings to patients. To help streamline the comparison of future biosimilar mAbs, Pisupati et al. developed a template using native CIU IM-MS and other MS methods to evaluate and characterize the biosimilar product, Remsima to the original product, Remicade.

The use of CIU fingerprints to characterize epitope changes has recently been performed. Huang et al. observed changes in CIU fingerprints among three malarial antigen–antibody complexes.³⁵ PvDBP is a vaccine to prevent malaria with several inhibitory antibodies, where three of those inhibitory antibodies have known epitopes. The two antibodies, 2D10 and 2H2, have the same epitope, while the third antibody, 2C6, binds to a different epitope. All three antibody/antigen complexes showed similar trends in their CIU fingerprints, but slight changes between 2D10/2H2 and 2C6 were observed. Complexes with 2D10 and 2H2 had identical drift-time transitions, while the 2C6 complex had shorter drift times. Also, the 2C6 complex required less energy to begin unfolding as compared to 2D10 and 2H2 complexes. The change in drift time and decreased stability indicated a more compact but less stable structure because of the different epitope location for the 2C6 complex.

Most recently, Tian and Ruotolo observed slight structural changes in antibody glycoforms with the use of CIU.²⁶ A standard antibody, IgG1, was used with varying glycoforms that were created through various reactions. By comparing each IgG1 glycoform’s CIU fingerprint, they saw subtle decreased stabilities with each sugar molecule addition. This stability represents structural changes taking place in IgG1 because of the altered glycan. Continued work on characterizing mAb is desired to make this a common analytical tool in biotherapeutic development. With continued work, this could become a quick and relatively easy assay capable of characterizing mAb structure and identifying glycosylations that can be used for the analysis of biosimilar mAb and development of new glycoproteins.

IMS has become a valuable tool not only for structural proteomics but lipidomics,³⁶ metabolomics,³⁷ and as a secondary separation of complex systems.^38–47 By pushing the limits of what IMS can do for scientists, developments will continue to grow, providing better conformer separation and CCS calculations. This will make IMS a bioanalytical tool that improves dynamic range with a high throughput, providing quick and descriptive information.

Characterization of Proteoforms by Top-Down MS.

A major advantage of using top-down proteomics is the ability to identify proteoforms. Proteoforms encompass all of the molecular forms of a protein translated from a single gene.⁵¹ These different forms could arise from post-translational modifications, alternative gene splicing, or genetic variations (Figure 5). Studies have demonstrated that proteoforms generated from alternate splicing lead to functional diversity, including differences in interaction networks and subcellular localizations highlighting the importance of identifying different proteoforms.⁵² Top-down provides the advantage of the full characterization of the proteoform, detecting details such as sequence variants and post-translational modifications.⁵³ Other approaches for studying proteoforms and their complexes, which include yeast two hybrid analysis⁵² and proteoform detection through intact mass and lysine count,⁵⁴ are more time-consuming than top-down MS, and in the case of yeast, two hybrid analysis can result in false positives for complex binding.

A challenge, however, in characterizing proteoforms by top-down proteomics is the lack of high-resolution separation techniques for complex protein samples. To address this challenge, Wang et al. recently developed a method to acquire better separation of intact proteins and better proteoform sequence coverage.⁵⁵ A two-dimensional separation approach using “salt-free” high-pH reverse-phase liquid chromatography (RPLC) coupled to low-pH RPLC was compared to separation using conventional one-dimensional low-pH RPLC with top-down MS. For E. coli, a total of 365 proteins and 886 proteoforms were identified with the 2D method compared to 163 proteins and 328 proteoforms identified using 1D separation.⁵⁵ This indicates the efficacy of using 2D separations to increase proteoform coverage.

A recent study by Melania et al. provided the first proteoform-specific look at the rapidly evolving sequences of the venom proteome. Top-down analysis was used to detect the sequence variation and PTMs in Ophiophagus hannah (king cobra) venom. In initial top-down experiments, 17 proteoforms were identified. To improve experimental dynamic range and obtain better coverage of the venom proteome, they then employed off-line fractionation prior to LC/MS/MS analysis. Using two fractionation methods, GELFrEE and solution isoelectric focusing (sIEF), 64 and 113 proteoforms were identified, respectively. Further manual validation to annotate amino acid substitutions, alternative processing cleavage sites, and presence of additional PTMs, resulted in the identification of 53 additional proteoforms, for a total 184 unique proteoforms. The knowledge of sequence variants in snake venom is important for understanding venom lethality. The identifications of larger proteins (>50 kDa) and low-abundant proteins were hindered by this approach, which utilized denaturing conditions for LC/MS/MS. To obtain better coverage of the venom proteome, they employed native MS strategies, which preserved the macromolecular interactions between the toxins and allowed for the characterization of the complexes.⁵⁶

The advantage of coupling top-down with native ESI is that it bridges the gap between proteomics and structural biology for protein complexes. Specifically, many of the fragile noncovalent interactions that are critical in maintaining the three-dimensional structure of proteins are conserved before controlled fragmentation of complexes into the subunits and their backbone fragment ions. The detailed characterization and increased level of coverage achieved with native top-down proteomics allows the study of complexes larger than 100 kDa, representing a powerful middle ground for both targeted and discovery proteomics.⁵³ Native top-down proteomics performs better with larger heterogeneous proteoform distributions because of the spread of the intensity over only a few charge states. This is in contrast to denaturing conditions, where a greater number of charge states are observed leading to signal dilution.

One goal of native top-down proteomics is to directly identify and characterize proteoforms and multiproteoform complexes (MPCs) in a single experiment. MPCs are specific protein complexes formed by different proteoforms (i.e., monomeric proteins and PTMs) from the same or different genes. Using native top-down MS, Skinner et al. identified three related MPCs of the human diphosphate kinase (NDPK) heterohexamer. The identification and characterization of three related MPCs in a single experiment, a feat that is difficult or impossible using other approaches, exemplifies that stoichiometry and precise subunit composition can be quickly determined using a native top-down approach to protein complex analysis.⁵⁷ Beyond sample preparation and separation strategies that retain native protein structure, the key to native top-down proteomics is using a three-tiered tandem MS approach to controllably disassemble protein complexes in the gas phase.⁵⁸ Using this approach, MPCs were identified, characterized, and scored. Also, stoichiometry was calculated using all sources of mass such as cofactors and PTMs.⁵⁷ This study highlights the fact that an unknown complex can potentially be identified by its subunits through native top-down MS. This platform expands the ability of MS to integrate proteomics and structural biology to provide insights into protein structure, functions, and regulation.

To provide even more structural information from top-down, Li et al. used complementary biophysical methods native top-down and ion mobility to elucidate the structural changes of α-synuclein induced by the binding of divalent metal ions (cobalt and manganese).⁵⁹ Mn-/Co-binding interactions can influence protein folding, which may be important factors for neurodegenerative diseases such as Parkinson’s disease. The group was able to locate similar binding sites for both Co and Mn on the C-terminus of α-synuclein using native top-down mass spectrometry with IMS.⁵⁹ This demonstrates the ability of top-down proteomics to identify binding interaction sites.

Advancements in Cross-Linking Reagents.

The development of cross-linkers has generally focused toward charged or polar side chains, as they are more likely to be found to be solvent accessible and therefore more likely to undergo cross-linking. The most widely used and commercially available reagents are the homo-bifunctional N-hydrosuccinimide (NHS) esters, which preferentially react with the amine groups of lysines and the N-terminus. A drawback of using NHS esters is that they mainly target lysine residues, which are highly flexible, where the distance information obtained is not very stringent.⁶⁰ More recently, diazirines have been used for photo-cross-linking of proteins where cross-linking is induced by the formation of a reactive carbene after irradiation with UV-A light. Diazirines do not display a preference for reactivity with amino acids like NHS esters. Photoactivatable diazirine-based amino acids, photoleucine and photomethionine, offer the additional advantage of conducting cross-linking in the cell. The short cross-linker length of the photoactivable diazirines provides more stringent topological information on proteins.

Recently, cleavable cross-linkers under tandem MS conditions have been an area of interest for XL-MS, as they automate the complex data analysis. Protein interaction reporter (PIR)⁶¹ technology developed for XL-MS experiments takes advantage of using a MS-cleavable cross-linker, which has been used to study protein interactions with purified complexes and cell culture and has recently been performed in heart tissue. Iacobucci et al. presented the first MS-cleavable photothiol-diazirine reactive cross-linker for the study of protein structure (Figure 6).⁶² Diazirine-based photo-cross-linkers were shown to generate CID-MS/MS-cleavable cross-linkers that are advantageous to develop fully automated data analysis allowing for proteome-wide cross-linking studies. 1,3-Diallylurea (DAU) was used as the photoreactive cross-linker, where cysteine residues undergo a “click reaction” that yields stable alkyl sulfide products. The central urea bond is then cleaved upon collisional activation during tandem MS, generating characteristic product ions that aid in the automated cross-link identification. The DAU technology was applied to thiol-containing proteins, bMunc 13–2 and GCAP-2, to demonstrate the applicability of DAU in structural proteomics.

Structural Modeling in XL-MS.

XL-MS has specially provided tremendous benefits in the study of increasingly larger macromolecular assemblies. Structural characterization of a large macromolecular machine is usually difficult to attain by standard structural approaches alone. XL-MS is amenable to probing such large complexes and has provided invaluable input into integrative approaches that include cryoEM, crystallography, SAXS, native spray, and hydrogen–deuterium exchange. Such studies have resulted in detailed structural and architectural characterization of the 26s proteasome,⁶³ nuclear pores,⁶⁴ CRISPR⁶⁵ and CRISPR-cas,⁶⁶ ribosomes,^67,68 TFIID,⁶⁹ and pore forming toxins,⁷⁰ among others. The power of XL-MS in probing the architecture of large multi-subunit or multiprotein subunits is particularly evident in conjunction with cryo-EM. Technological advances have permitted dramatic improvement in resolution attained by cryo-EM, and megadalton-sized macromolecular complexes have been reported at near atomic resolution. The reconstruction of a high-resolution model from the cryo-EM-generated 3D volume requires fitting already available subcomponent structures within the larger complex or, alternatively, de novo modeling, if the resolution permits it. For that purpose, XL-MS has proven especially helpful in providing location, spatial orientation, and distance constraints to aid in the reconstruction of the high-resolution model.⁷¹ This is particularly true in the case of the limited resolution (local or global) of the 3D volume generated by cryo-EM, where de novo modeling and unambiguous fitting of known structures is difficult. Studies involving the use of XL-MS to augment cryo-EM structure determination include the structure of the 26S proteasome,⁶³ the Polycomb Repressice Complex 2,⁷² the chromatin remodeling complex,⁷³ chromosome segregation complexes,^74–76 signalosomes,⁷⁷ the membrane embedded receptor,⁷⁸ nuclear pore complexes,^64,79 eukaryotic ribosomal complexes,^67,68,80 the mammalian mitochondrial complex,⁸¹ various RNA polymerase complexes,^82–86 and various spliceosome complexes.^87,88

XL-MS in Cells.

The use of a bottom-up proteomics workflow with XL-MS enables the study of protein structure on the proteome-wide scale. The potential benefit of such expansion is the increased biological significance of reporting structural information within the cellular or tissue context, potentially bridging the gap between structural biology and systems biology. Taking advantage of the advancement in cross-linker technology, studies have characterized the network of interaction around Hsp90 in vivo,⁸⁹ protein communities in cell lysates,⁹⁰ the interactome in mitochondria⁹¹ and in bacteria,⁹² and the perturbation of the interactome upon viral infection of a cell.⁹³ Häupl et al., incorporated photomethionine and photoleucine amino acids into proteins in HeLa cell cultures to study affinity enriched protein complexes of interest.⁹⁴ Photoaffinity-labeled proteins were enriched, and protein complexes were covalently fixed by activation of photo-amino-acid diazirines with UV-A irradiation followed by sample digestion and MS analysis. The study highlighted the advantage of using photo-cross-linkers in cell cultures to identify protein–protein interactions. Finally, the most recent advances have performed XL-MS in tissue,⁹⁵ opening the potential of XL-MS in applications such as clinical diagnostics.

XL-MS in Heart Tissue.

XL-MS was performed in mouse heart tissue to identify protein–protein interactions relevant to organ-level disease states that cannot be obtained with cell culture studies. To obtain structural information, heart tissue was isolated from mice and subjected to chemical cross-linking with the protein interaction reporter⁶¹ cross-linker. Samples were either processed as whole-heart samples, or the mitochondria was isolated by subcellular fractionation. A real time MS3 technique termed ReACT was used to analyze in-tissue cross-linked peptide pairs. ReACT is an MS method that uses low-energy fragmentation to release the cross-linked peptides and the reporter ion. The mass of the reporter ion and the cross-linked peptides in the MS2 spectra are searched in real time, and the two released cross-linked peptides are further fragmented. A total of 572 peptide pairs were detected from all the major sacromeric proteins, where the identified cross-linked pairs were used as probes for protein interactions for future XL-MS studies with diseased heart tissue. The mitochondria were extracted from heart tissue, because cardiomyocytes contain the highest concentration of mitochondria of any cell in the body, which would allow the analysis of oxidative phosphorylation complexes (OXPHOS) (Figure 7). Multiple links were identified between the complexes, suggesting the existence of larger supercomplex assemblies in agreement with single-particle EM results. In total, more than 2000 links were identified between the complexes, which can be used to compare supercomplex assemblies in normal and failing heart tissues to understand their role in mitochondrial function in heart disease.

De Novo Modeling with Covalent Labeling.

Similar to XL-MS, de novo modeling is also an emerging field within covalent labeling. Recent advancements in HRF, including normalization of radical reactivity, have aided the ability to use HRF for molecular modeling. The varying reactivity of residues with hydroxyl radicals has been a challenge in directly correlating HRF data with protein structure, because highly reactive residues that may not be as solvent accessible have a higher extent of modification compared to residues that are solvent accessible but are less reactive. When only comparing different states, i.e., ligand-free vs ligand-bound, these reactivity differences are not a major factor. However, direct correlation of the HRF extent of modification data to protein structure is hampered by these reactivity differences. Huang et al. developed an algorithm that correlates the experimental footprinting rate to a protection factor that is based on the hydroxyl radical reactivity of the modified residue. Normalization of footprinting rates provides more quantitative information on the structure allowing HRF data to be used for molecular modeling.¹¹⁴ Xie et al. used HRF data to distinguish between molecular models that were generated from high- and low-accuracy crystal structures.¹¹⁵ Recently, Aprahamian et al, used HRF-derived protection factors as a score term in Rosetta to obtain atomic resolution of proteins and predict tertiary structures.¹¹⁶ De novo models of four different proteins, calmodulin, cytochrome c, myoglobin, and lysozyme, were generated using Rosetta only and with HRF data as constraints. The incorporation of HRF data increased the RMSD of each structure (Figure 9). Atomic-level resolution was achieved for two structures, cytochrome c and myoglobin, upon the inclusion of HRF data as constraints. This is the first time that de novo structural modeling has been achieved from HRF mass-spectrometry-derived protection factors.

Similar to HRF modeling, other covalent labeling techniques such as DEPC coupled with mass spectrometry have been used to model dynamics of macromolecules. Schmidt et al. generated models from a restraint-based strategy and subjected those models to molecular dynamics simulations to investigate structure and functions of large complexes.¹¹⁷ The solvent accessibility information acquired from covalent labeling MS was converted into modeling restraints using an in-house developed code that estimated the solvent accessible surface area for all models generated using a sampling algorithm. The method successfully reconstructed the 3D assembly structure of three model proteins (tryptophan synthetase, carbonyl phosphate synthetase, and cATPase) with high accuracy and precision. Combining structural modeling and experimental data allows large complexes to be studied that are not feasible by traditional tools used to study protein complexes.

In-Cell FPOP.

Macromolecular crowding in the cell has shown to impact interactions of proteins and protein conformational states, which elucidates the importance of using a whole cell approach to study proteins.¹¹⁸ For HRF MS, FPOP has been successfully applied to African green monkey kidney (Vero) cells to study protein structure and protein–protein interactions directly in the cellular environment.^119,120 Hydrogen peroxide is permeable to cellular and organelle membranes, which enables in-cell FPOP (IC-FPOP) to oxidatively modify proteins in different cellular compartments. Since hydrogen peroxide was added to cells, cell viability assays were performed to demonstrate that FPOP was probing live cells. In the time scale of the FPOP experiment using 20 mM hydrogen peroxide, the majority of the cells (70%) were viable. To increase the number of oxidatively modified proteins, a flow system was developed to allow single-cell flow based on the principles of hydrodynamic focusing. Over 1300 proteins were oxidatively modified by modifications made to the flow system, which allowed equal cell exposure to laser irradiation.¹¹⁹ The modified proteins were present in 27 different compartments of the cell, verifying that hydrogen peroxide was able to penetrate various cellular organelles. The dynamic range of IC-FPOP was evaluated by comparing the oxidized proteins with their expression levels (transcripts per million, TPM) in human kidney cells. High-abundance proteins like actin (3154 TPMs) and low-abundance proteins like Protein Shroom 2 (4 TPMs) were oxidatively modified using the flow system. This shows that IC-FPOP has a large dynamic range.

In Vivo FPOP.

Cell culture studies provide structural information on proteins in their native cellular environment but do not take into consideration interactions at an organ level. Recently, FPOP coupled with MS has been expanded into an animal model, Caenorhabditis elegans (C. elegans), to study protein–protein interactions in vivo. Similar to IC-FPOP, a flow system was used for in vivo FPOP (IV-FPOP) for single-worm flow to avoid clumping of the worms and unequal exposure to laser irradiation.¹²¹ Initial studies indicate that several proteins in varying body systems in the worm were oxidatively modified. The development of IV-FPOP offers the opportunity to study a wide range of biological processes and disease states.