Structure and dynamics of endogenous protein complexes in human heart tissue captured by native nanoproteomics

Protein complexes are highly dynamic entities that display substantial diversity in their assembly, post-translational modifications, and non-covalent interactions, allowing them to play critical roles in various biological processes. The heterogeneity, dynamic nature, and low abundance of protein complexes in their native states present tremendous challenges to study using conventional structural biology techniques. Here we develop a “native nanoproteomics” strategy for the native enrichment and subsequent native top-down mass spectrometry (nTDMS) of low-abundance protein complexes. Specifically, we demonstrate the first comprehensive characterization of the structure and dynamics of cardiac troponin (cTn) complexes directly from human heart tissue. The endogenous cTn complex is effectively enriched and purified using peptide-functionalized superparamagnetic nanoparticles under non-denaturing conditions to enable the isotopic resolution of cTn complexes, revealing their complex structure and assembly. Moreover, nTDMS elucidates the stoichiometry and composition of the heterotrimeric cTn complex, localizes Ca2+ binding domains (II-IV), defines cTn-Ca2+ binding dynamics, and provides high-resolution mapping of the proteoform landscape. This native nanoproteomics strategy opens a new paradigm for structural characterization of low-abundance native protein complexes.


Introduction
The vast majority of proteins within a cell assemble into protein complexes to perform specific functions and play crucial roles in regulating diverse biological processes 1,2 . Thus, comprehensive characterization of the structure and dynamics of endogenous protein complexes is essential for understanding fundamental biological processes and disease mechanisms to develop new therapeutic interventions 1,2 . Protein complexes are highly dynamic entities with substantial diversity in their assembly, post-translational modifications (PTMs), and non-covalent interactions. Moreover, endogenous protein complexes often exist in low abundance in their native states. These present tremendous challenges to studying their structure and dynamics using traditional structural biology techniques such as X-ray crystallography, nuclear magnetic resonance (NMR) spectroscopy, and cryo-electron microscopy (cryo-EM) 2 . Native top-down mass spectrometry (nTDMS), a technique combining native MS [3][4][5] and top-down proteomics [6][7][8][9][10][11] , has emerged as a powerful structural biology tool for characterization of protein complexes [12][13][14][15][16][17] .
In nTDMS, intact proteins are introduced into the mass spectrometer under non-denaturing conditions, preserving the noncovalent interactions between protein subunits and their associated ligands as well as PTMs. The intact proteins are then fragmented in the gas-phase to map the PTMs and ligand binding sites 12 . nTDMS enables the structural characterization of macromolecular protein complexes, subunit stoichiometry, non-covalent interactions, as well as the analysis of their proteoformsthe diverse protein products arising from alternative splice isoforms, genetic variations, and PTMs 7,18,19 . However, so far only over-expressed recombinant or high-abundance proteins and protein complexes have been characterized by nTDMS. Significant challenges remain in the structural characterization of low-abundance protein complexes due to the difficulty in isolating them and sensitivity required to resolve heterogenous complexes 2 .
Here, we have developed a "native nanoproteomics" platform integrating the native enrichment of low-abundance protein complexes directly from tissues using surface functionalized superparamagnetic nanoparticles (NPs) with high-resolution nTDMS to characterize the structure and dynamics of low-abundance endogenous protein complexes for the first time. Specifically, we applied this method to enrich and structurally elucidate the heterotrimeric cardiac troponin (cTn) complex (~77 kDa) directly from human heart tissues. The cTn complex is a master regulator of cardiac contraction and represents the Ca 2+ sensitive switch 20,21 of striated muscles assembled from three molecular subunits: troponin C (TnC), the Ca 2+ -binding subunit; cardiac troponin I (cTnI), the actin-myosin inhibitory subunit; and cardiac troponin T (cTnT), the thin-filament anchoring subunit 22,23 . Both cTnI and cTnT serve as gold standard biomarkers for diagnosing acute coronary syndrome due to their cardiac specificity and their release into the bloodstream following cardiac injury 24 . Moreover, the association of Ca 2+ ions with the TnC subunit along with phosphorylation of the cTnI subunit together initiate a cascade of molecular events on the thin filament and induce actin-myosin cross bridge formation necessary for cardiac contraction 25,26 .
However, only partial structural information has been obtained from conventional X-ray crystallography excluding the intrinsically disordered but functionally critical regions of cTnI and cTnT. Moreover, the cTn structure is highly dynamic due to Ca 2+ binding 23,27,28 and PTMs 9,29,30 that regulates muscle contraction, yet traditional methods have not effectively captured these dynamic conformational changes 31 . Furthermore, recombinantly expressed proteins have been used in previous studies thus important structural features vital to the function of the endogenous cTn complex within the sarcomere were lost 32,33 .
Using the native nanoproteomics approach, we enrich and purify cTn complex in its native state directly from human heart tissue and achieve isotopic resolution of endogenous cTn complexes to reveal cTn complex structure and assembly. Our results elucidate the stoichiometry and composition of the heterotrimeric cTn complex, define the conformational roles of cTn-Ca 2+ binding dynamics, locate the Ca 2+ binding domains (II-IV), and map the proteoform landscape with direct analysis of the stoichiometry of various proteoforms. Overall, this work represents the first nTDMS study to comprehensively characterize the structure and dynamics low-abundance endogenous protein complexes.

A native nanoproteomics platform for the enrichment and comprehensive characterization of low-abundance protein complexes
Our antibody-free "native nanoproteomics" platform integrates native enrichment and purification of low-abundance protein complexes using peptide functionalized superparamagnetic iron oxide (magnetite, Fe3O4) NPs (NP-Pep) followed by comprehensive characterization using nTDMS (Figure 1). Building on our previous denatured NP enrichment study 34 , we sought to establish a method that could enrich and purify native protein complexes directly from human heart tissue. We first optimized the native protein extraction buffers to effectively extract intact protein complexes from human heart tissue using a high ionic strength lithium chloride (LiCl) buffer at physiological pH (Figure 1a, Table S1). Next, we hypothesize that the specifically designed peptide on the NP surface that contains a combination of charged and aromatic residues that leads to the specific binding to the protein complex of interest, would be amenable to a competitive elution strategy using amino acids [35][36][37] . L-Arg and L-Glu are amino acids previously known to improve protein solubility and in-solution stability by forming protonated clusters, 36 and were included to enhance protein complex elution efficacy. Moreover, imidazole, a positively charged aromatic small molecule which functions as the side chain of histidine, 35 was added to disrupt the polar interactions of the peptide functionalized on the NP surface for the native competitive elution of NP-captured protein complexes, without disrupting the intermolecular interactions between the protein subunits. We found that the optimal buffer composition for effective native competitive elution of NP-captured protein complexes was a combination of 750 mM L-Arg, 750 mM imidazole, and 50 mM L-Glu (pH 7.5) as native elution efficiency of amino acid solutions increases with their concentration 38 (Figure 1a, Figure S1).
For a general native proteomics workflow, the NP-Pep was incubated with sarcomeric protein mixtures, magnetically isolated to remove non-specifically bound proteins, and the bound protein complexes were eluted off the NP-Pep using the above optimized native elution buffer cocktail. Non-MS compatible buffers and salts were removed using either online or offline sizeexclusion chromatography (SEC) to transfer the complexes into a MS-compatible ammonium acetate solution. Subsequently, the enriched protein complexes were subjected to various nTDMS techniques to characterize protein complex structure, assembly, and dynamics including: online SEC for rapid screening of protein complexes (Figure 1b)

Native enrichment of endogenous cTn complex from human heart tissue
To enrich cTn complexes from human heart tissues, the Fe3O4 NPs were functionalized with a 13-mer peptide (NP-Pep; HWQIAYNEHQWQC) 34 with high binding affinity (Kd = 270 pM) towards cTnI under non-denaturing condition. Native cTn complex enrichment yielded TnC, cTnI, and cTnT in approximately a 1:1:1 ratio (Figure 2a), reflecting the composition of the heterotrimeric cTn complex in the sarcomere. Moreover, cTn could be reproducibly enriched directly from human heart tissue across multiple NP-Pep synthetic batches and replicates ( Figure   S2). To demonstrate the enrichment and purification of all three cTn subunits while preserving their proteoforms, we further used an online reverse-phase liquid chromatography (RPLC) topdown tandem MS (MS/MS) method comparing the initial sarcomere protein loading mixture (L), the resulting flow through (F), and the final elution mixtures (E) ( Figure S3). Not only were all three subunits of the cTn complex, cTnI, cTnT and TnC, significantly enriched, but also the PTM profiles of endogenous cTnI, cTnT, and TnC were faithfully preserved, without introducing artifactual modifications (Figure S3a-c). The top-down proteomics results provide a bird's eye view of the proteoform landscape of TnC, cTnT, and cTnI for direct analysis of the stoichiometry of their various proteoforms (Figure 2b). Moreover, this native NP enrichment strategy was found to be highly reproducible across three independent donor heart tissues ( Figure S4).
We next investigated whether the enriched endogenous cTn complex could be analyzed under native conditions by SEC-MS using an online buffer exchange (OBE) method 39,40 for rapid analysis of native protein complexes after separation from MS-incompatible buffers ( Figure S5).
Upon analysis of the resulting SEC-MS chromatograms, protein complexes were effectively separated from non-volatile buffer components within 7 minutes ( Figure S5). Additionally, heterotrimeric cTn complexes (z = 16+-21+) were reproducibly enriched across three independent elution mixtures from the same heart tissue sample highlighting the reproducibility of this SEC-OBE native MS method ( Figure S6). (b) 600 ng of E was injected for reverse-phase liquid chromatography (RPLC)-MS analysis of cTn subunits. The deconvoluted mass spectra showing the birds eye view of the proteoform landscape for cTnT, cTnI, and TnC. (c) Native MS spectra of the endogenous cTn complex using ultrahighresolution FTICR-MS to achieve isotopic resolution. Most abundant charge state (z = 19+) is isolated. (d) Zoomed-in mass spectra from c showing accurate determination of isotopically resolved cTn complex proteoforms. Theoretical isotope distributions (red circles) are overlaid on the experimentally obtained mass spectrum to illustrate the high mass accuracy. All individual ion assignments are within 2 ppm from the theoretical mass.

Structural heterogeneity of endogenous cTn complex revealed by nTDMS
To comprehensively characterize the endogenous heterotrimeric cTn complex, we employed a nTDMS approach using an ultrahigh-resolution FTICR mass spectrometer for unequivocal proteoform sequencing and protein complex characterization. Native mass spectra of the enriched cTn complex revealed a charge state distribution of 18+ to 21+ (3800 m/z to 4300 m/z) with the most abundant charge state (z = 19+) detected between 4050 m/z to 4080 m/z (Figure   2c-d, Figure S7). In-depth examination of the endogenous cTn complex revealed four unique proteoforms comprised of both covalent and non-covalent modifications (Figure 2d). All heterotrimer cTn complex proteoforms were identified with high mass accuracy (< 2 ppm) ( Figure   2d, Table S2). Significantly, the predominant forms of the heterotrimeric cTn complex were revealed to comprise of monophosphorylated cTnT, mono-and bis-phosphorylated cTnI, and TnC with three bound Ca 2+ ions (most abundant cTn complex MW = 77136 Da). These results show that the human cTn complex exists in structurally diverse molecular compositions in the sarcomere with highly heterogenous covalent and non-covalent modifications that yield a suite of different intact assemblies.

Sequence-specific structural elucidation of cTn complex quaternary structure through nTDMS
Next, we performed complex-up native MS analysis 41 with collisionally activated dissociation (CAD) 42 to elucidate the stoichiometry and composition of the heterotrimeric cTn complex (Figure 3). First, the intact heterotrimeric cTn complex (z = 18+-20+) was isolated with no dissociated cTn subunits present in the resulting mass spectra (Figure 3a). Subunit ejection of the cTn complex was then observed by CAD (Figure 3b). We detected the intact cTn(I-C) dimer at 2700 m/z to 3600 m/z (z = 12+-16+, MW = 42556 Da), cTnT monomer at 2500 m/z to 3200 m/z (z = 11+-14+, MW = 34580 Da), and TnC monomer at 2600 m/z to 3100 m/z (z = 6+-7+, MW = 18520 Da). The molecular abundance of the dissociated cTn subunits recapitulated the expected composition of the heterotrimeric cTn complex in the sarcomere with an observed subunit stoichiometry of 1:1:1 ( Figure S8).
To gain additional structural and sequence-specific information of the endogenous cTn complex, we performed complex-down analysis 41 of the dissociated cTn subunits (Figure 3c-d).
The isolation of ejected cTnT monomers (z = 14+) revealed multiple proteoforms, including unphosphorylated cTnT, monophosphorylated cTnT (pcTnT), and phosphorylated cTnT with Cterminal Lys truncation (pcTnT [aa 1-286]) (Figure 3c). Further CAD fragmentation revealed fragments which suggest that the C-terminus of cTnT is more solvent exposed than the N-terminus forming the heterotrimer interface (Figure 3c, Figure S9). Next, the isolation of ejected cTn(I-C) dimer (z = 14+) revealed six unique proteoforms with different Ca 2+ occupancy and phosphorylation states: unphosphorylated cTnI associated with TnC and 2 Ca 2+ , unphosphorylated cTnI associated with TnC and 3 Ca 2+ , phosphorylated cTnI (pcTnI) associated with TnC and 2 Ca 2+ , pcTnI associated with TnC and 3 Ca 2+ , bisphosphorylated cTnI (ppcTnI) associated with TnC and 2 Ca 2+ , and ppcTnI associated with TnC and 3 Ca 2+ (Figure 3d). The cTn(I-C) dimer precursor ions (z = 14+) were further subjected to CAD fragmentation to obtain additional structural information (Figure 3d, Figure S10). We observed phosphorylation of dissociated cTnI monomer at Ser22 and Ser23, which are the targets of PKA-mediated phosphorylation 29 . Our nTDMS analysis also suggests that both the intrinsically disordered C-and N-termini of cTnI are more solvent exposed than the stable internal regions that form the subunit-subunit interfaces of the cTn complex.

Direct localization of Ca 2+ binding domains in endogenous cTn Complex
We have characterized the three Ca 2+ binding domains present in endogenous TnC using nTDMS (Figure 4). TnC is an EF-hand Ca 2+ -binding protein that is an essential Ca 2+ sensing molecular subunit in the heterotrimeric cTn complex 27,28 . Human TnC consists of three functional metal-binding motifs (domains II-IV) that can be occupied by Ca 2+ and are responsible for regulating cardiac muscle contraction 33 . The nTDMS isolation spectra of ejected TnC monomer (z = 8+) revealed multiple proteoforms of endogenous TnC in its Ca 2+ -bound states (Figure 4a).
Specifically, we observed TnC with 0, 1, 2, and 3 Ca 2+ ions bound with baseline isotopic resolution and high mass accuracy. The relative proportion of singly, doubly, and triply bound Ca 2+ states was found to be approximately 0.2, 0.5, and 0.1, respectively ( Figure S11).
To localize Ca 2+ binding domains, TnC proteoforms were subjected to CAD to generate extensive backbone fragmentation (Figure 4b-c). Progressive collisional activation ramping revealed TnC domain III to be the least vulnerable Ca 2+ binding to increasing collisional activation, while domain II was found to be the most vulnerable Ca 2+ binding region. To localize the primary region for Ca 2+ binding that is least vulnerable to collisional activation, we isolated the TnC proteoform at 2312 m/z and performed CAD to yield product ions y52 + Ca 2+ , y30, b115 + Ca 2+ , and b109 ( Figure S12). Therefore, the primary Ca 2+ binding domain was localized to 113 DLD 115 in domain III. The next Ca 2+ binding domain was localized to the structural region between 141 DKNND 145 in domain IV by first isolating the TnC proteoform at 2316 m/z and then generating CAD product ions b140, b145 + Ca 2+ , y16, and y22 + Ca 2+ (Figure S13). Finally, the most vulnerable Ca 2+ binding region was localized to regulatory domain II between 73 DFDE 76 by isolating the TnC proteoform at 2321 m/z and generating CAD product ions b65, b91 + Ca 2+ , y85, and y94 + Ca 2+ ( Figure S14). This is the first study to localize endogenous TnC Ca 2+ binding regions to domains II, III, and IV in the cTn complex (Figure 4d). Additionally, a molecular depiction of the divalent association of Ca 2+ to amino acid residues 73 D and 76 E in TnC domain II is illustrated in Figure   4e.

Determination of cTn-Ca 2+ binding and conformational dynamics
Binding of Ca 2+ to TnC has extensive effects to the heterotrimeric cTn complex function and structure in regulating cardiac contraction 23,28,43,44 . Due to the conformational heterogeneity and presence of intrinsically disordered regions along the heterotrimeric cTn complex, it is challenging to obtain crystal structures of cTn in its active and closed states upon Ca 2+ binding using traditional structural biology techniques 43 . Therefore, to assess the intricate relationship First, the native TIMS-MS parameters were optimized using bovine serum albumin (BSA, ~ 66 kDa). Modifying the desolvation parameters proved critical for effective high-resolution ion mobility separation of BSA conformers ( Figure S15). We next used TIMS-MS to separate and analyze the endogenous TnC monomer and cTn(I-C) dimer Ca 2+ -bound proteoforms ( Figure 5).
These results suggest that cTnI and TnC form a more open conformation when saturated with Ca 2+ ions in preparation for cTnT engagement along the thin filament for subsequent cTn complex formation and cardiac muscle contraction 50 .
The structural roles of Ca 2+ in maintaining the intact heterotrimer stability were further profiled by native TIMS-MS analysis (Figure 6). We added EGTA to sequester Ca 2+ from the intact heterotrimer complex and probe the specific contribution of individual Ca 2+ binding regions to cTn complex stability (Figure 6a). First, the heterotrimeric cTn complex (z = 18+ to 21+) was resolved by native TIMS-MS analysis (Figure 6b-c) without EGTA incubation. The CCS of the native cTn complex with all Ca 2+ occupied was determined to be 4880 Å 2 which was comparable to the CCS value calculated using the IMPACT 49 method (4192 Å 2 ) and is in good agreement with the previously reported partial crystal structure 43 that is truncated due to intrinsically disordered regions present within the cTn complex (Table S3)

Discussion
Here we have developed a native nanoproteomics platform for enrichment/purification and structural characterization of low-abundance, endogenous protein complexes and their noncovalent interactions in their native state together with comprehensive proteoform mapping. This approach addresses challenges in the current nTDMS field in the isolation and analysis of lowabundance protein complexes. By avoiding denaturation or digestion steps, nTDMS quickly emerges as a powerful structural biology tool providing insights into the endogenous protein complexes and their functional states 2 . However, nTDMS studies have primarily relied on purification strategies using overexpressed recombinant proteins and/or highly abundant proteins due to the difficulties in isolating low-abundance protein complexes [13][14][15][51][52][53] . To our knowledge this is the first nTDMS study to structurally characterize low-abundance and heterogenous protein complexes directly from tissue samples.
We chose to apply the native nanoproteomics method to characterize the structure and dynamics of the endogenous cTn complex given its high significance in cardiac function and clinical diagnosis. The structure of the cTn complex has been previously investigated by X-ray crystallography, NMR, and cryo-EM, but were limited to resolving only the primary domains of the complex 31,43 . Moreover, it is challenging to characterize the dynamic structural changes of endogenous cTn-Ca 2+ binding events and PTMs directly from human samples by these methods [54][55][56][57]   After washing, the iron oleate was concentrated in vacuo.

BAPTES) and peptide (NP-Pep).
Iron oleate precursor and magnetite nanoparticles were synthesized following published protocol, with minor modifications. 34 Briefly, the NPs were synthesized using 10 mmol (   where μ is the reduced mass of the ion−gas pair (μ = ( + ) , where m and M are the ion and gas particle masses), kb is Boltzmann's constant, T is the drift region temperature, z is the ionic charge, e is the charge of an electron, N0 is the buffer gas density, and K0 is the reduced mobility.
Theoretical CCS values were determined using the IMPACT method. 46

Statistical Analysis
Statistical analysis for group comparison was completed using paired-student t-tests. All p-values at p < 0.01 were considered significant. All error bars indicated in figures represent the mean ± standard error of the mean (SEM).