Native Mass Spectrometry Enables Direct Detection of Endogenous Protein Complexes from Crude Lysates of Cultured Human Cells

Abstract

Native mass spectrometry (MS) enables the direct observation of non-covalent protein interactions while preserving the higher-order structures of protein complexes. We previously developed direct native MS approaches for analyzing overexpressed protein complexes from crude lysates and single human cells, retaining their interactions with small molecules and drugs. Here, we demonstrate that this strategy can be extended to endogenous protein complexes, which are present at substantially lower abundance than the overexpressed protein complexes analyzed in previous direct native MS studies, by directly analyzing crude lysates of cultured human HEK293T cells without prior purification or buffer exchange. Although crude human cell lysates contain salts and matrix compounds that typically cause peak broadening and signal suppression, we successfully observed the endogenous ~145-kDa glyceraldehyde-3-phosphate dehydrogenase (GAPDH) tetramer directly from unpurified human cell lysates, including its NAD⁺-bound forms. These results demonstrate the feasibility of directly detecting an endogenous high-molecular-weight protein complex and its cofactor-binding states from crude human cultured cell lysates, providing a basis for extending this approach to other endogenous protein assemblies and future single-cell native MS analyses.

INTRODUCTION

Native mass spectrometry (MS)^1–4) has emerged as an indispensable platform for probing non-covalent protein interactions while preserving the native topology of protein complexes. Standard native MS workflows generally necessitate rigorous sample isolation, purification, and buffer exchange into volatile solution. However, the actual intracellular environment is highly crowded, containing 50–400 mg/mL of proteins, organelles, and other macromolecules.⁵⁾ Under such crowded conditions, biomolecular interactions often differ from those characterized using isolated proteins in vitro, and drugs optimized under cell-free assays do not always exhibit the expected efficacy within the cellular environment. Consequently, developing direct analytical strategies that bypass extensive purification—thereby preserving the cellular context—is essential for the realistic assessment of biomolecular and therapeutic interactions. To address this challenge, several research groups, including ours, have explored direct native MS analysis of crude cell lysates.^6–8) In 2020, we demonstrated the direct analysis of protein–inhibitor interactions using an overexpressed target protein, dihydrofolate reductase, in Escherichia coli (E. coli) lysates.⁷⁾ Subsequently, Rogawski et al. reported the direct observation of intracellular protein–drug interactions using overexpressed proteins in crude human cell lysates.⁸⁾ We further extended direct native MS to the single-cell level by successfully detecting endogenous hemoglobin tetramers from single red blood cells.⁹⁾ However, red blood cells represent a highly specialized cellular environment dominated by hemoglobin. We subsequently expanded this single-cell approach to the more complex intracellular environment of human cultured cells, where complexes formed between overexpressed target proteins and cell-permeable inhibitors were successfully observed.¹⁰⁾

Despite these advances, previous direct native MS studies have primarily relied on overexpression systems or exceptionally abundant proteins in specialized cell types. Consequently, the direct observation of endogenous high-molecular-weight protein complexes from general human cell types without prior purification remains largely unexplored.

To bridge this gap, this study aimed to directly detect endogenous protein complexes from crude lysates of cultured human cells without prior purification using native MS. Rather than relying on protein overexpression and analyte purification, we sought to determine whether endogenous protein complexes could be directly observed under conditions closer to the native intracellular environment than conventional purified systems. Specifically, we analyzed crude lysates of cultured human HEK293T cells and demonstrated that endogenous high-molecular-weight protein complexes can be directly detected without prior purification or buffer exchange, thereby extending direct native MS beyond the overexpression-based approaches reported previously.

EXPERIMENTAL

Materials

Fetal bovine serum (FBS) was purchased from Corning (Corning, NY, USA). ꞵ-Nicotinamide adenine dinucleotide oxidized form (NAD⁺) was purchased from Oriental Yeast Co., Ltd. (Tokyo, Japan). Other chemical reagents were purchased from Nacalai Tesque Inc. (Kyoto, Japan).

Cell culture and harvesting

HEK293T cells (RIKEN BRC, Tsukuba, Japan) were cultured at 37°C with 5% CO₂ in a Forma^TM Direct Heat CO₂ Incubator (Thermo Fisher Scientific, Waltham, MA, USA), using Dulbecco’s Modified Eagle’s Medium (DMEM, high glucose) supplemented with 10% FBS. For passaging, cells were washed with Dulbecco’s phosphate-buffered saline (D-PBS), detached using trypsin containing 5.3 mM ethylenediaminetetraacetic acid (EDTA), collected, and reseeded.

Sample preparation for direct native MS of crude cell lysates

To prepare samples for direct native MS, HEK293T cells were cultured in 35 mm dishes. The cells were washed with D-PBS and harvested by pipetting. After counting the cell numbers, the suspension was centrifuged at 1,000 × g for 1 min at room temperature. The supernatant was discarded, and the cell pellet was washed twice with 150 mM ammonium acetate (pH 7.4). The washed pellets were rapidly frozen in liquid nitrogen and stored at −80°C for at least 3 h until sample preparation and MS analysis.

Prior to native MS analysis, the frozen cells were thawed on ice. In some experiments, the cells were re-frozen in liquid nitrogen and thawed again on ice to introduce an additional freeze–thaw cycle. Subsequently, 150 mM ammonium acetate (pH 7.4) was added to adjust the cell density to 2 × 10⁴ cells/μL (2 × 10⁷ cells/mL), and the cells were gently resuspended by a few pipetting cycles following the freeze–thaw, thereby facilitating mild cell lysis and the release of intracellular components. The cell lysates were then centrifuged at 1,500 × g for 3 min at room temperature, and the supernatant was collected and used directly for native MS analysis without buffer exchange or further purification (hereafter referred to as crude cell lysate). The crude cell lysate was further diluted 10-, 20-, or 30-fold with 600 mM ammonium acetate (pH 7.4). For NAD⁺ addition experiments, NAD⁺ was supplemented to the diluted lysate at a final concentration of 50 μM. A 4-μL aliquot of each sample was loaded into an in-house-prepared nanoelectrospray ionization (nanoESI) emitter (tip inner diameter: 1.5–2 μm) and subjected to native MS.

nanoESI-MS

Mass spectra were acquired with a prototype multi-turn time-of-flight mass spectrometer (Shimadzu Corporation, Kyoto, Japan) equipped with a nanoESI source, featuring an extended ion flight path of approximately 50 m. The emitter containing the sample solution was mounted on the nanoESI source, and a high voltage was applied through a Pt wire (0.127 mm diameter; Sigma-Aldrich, St. Louis, MO, USA) for ionization. The following parameters were used for obtaining native mass spectra: a capillary voltage of 1.0–1.5 kV, a desolvation line (DL) voltage of 20–50 V, a collision voltage of 10–40 V, a block heater temperature of 100°C, and a collision-induced dissociation (CID) gas pressure of 350 kPa. Mass spectra were acquired over an m/z range of 3000–8000. Each averaged mass spectrum was obtained by signal accumulation for several minutes. Molecular masses were calculated from the observed m/z values, and mass spectra were visualized using Microsoft Excel.

RESULTS AND DISCUSSION

The crude lysate of HEK293T cells was diluted 10-fold with 600 mM ammonium acetate (equivalent to 2 × 10⁶ cells/mL) and directly subjected to native MS without purification. Multiple peaks were observed in the m/z range of 3000–8000, indicating the presence of various endogenous proteins and protein complexes (Fig. 1A).

Fig. 1. Native mass spectra of crude HEK293T cell lysates. Native mass spectra of a 10-fold diluted HEK293T cell lysate in the absence (A) and presence (B) of additional NAD⁺ to a final concentration of 50 μM were acquired over the m/z range of 3000–8000 under native conditions. Panels (C) and (D) correspond to the enlarged views of the m/z 5900–6500 region (yellow-shaded areas in (A) and (B), respectively). Red dashed lines indicate the theoretical m/z values of the GAPDH tetramer (24+ and 23+ charge states) with zero to four bound NAD⁺ molecules. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; NAD⁺, nicotinamide adenine dinucleotide.

Detailed analysis of the peaks in the m/z 5900–6500 region (Fig. 1C) revealed that the signals at m/z 5987.6 (24+) and m/z 6249.4 (23+) corresponded to a species with a mass of 143,696.5 ± 24.2 Da (~145 kDa). Additional peaks at m/z 6015.8 and m/z 6277.8 corresponded to 144,360.2 ± 7.2 Da, while those at m/z 6042.1 and m/z 6306.7 corresponded to 145,013.4 ± 25.7 Da. The mass difference between these species was approximately 660 Da, suggesting binding of a small molecule of approximately this mass.

Considering the abundance of housekeeping proteins in human cultured cells, these observed masses are highly consistent with the theoretical masses of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) tetramer and its NAD⁺-bound forms. GAPDH is a key glycolytic enzyme that catalyzes the reversible oxidation and phosphorylation of glyceraldehyde-3-phosphate to 1,3-bisphosphoglycerate, coupled with the reduction of NAD⁺ to dihydronicotinamide adenine dinucleotide (NADH), and is abundantly expressed in cells. GAPDH is known to exist predominantly as a tetramer in cells, with four NAD⁺ binding sites.¹¹⁾ Specifically, the theoretical molecular mass of GAPDH is 35,922.0 Da for the monomer and 143,688.1 Da for the tetramer, while NAD⁺ has a molecular weight of 663.4 Da. The mass difference between the observed species (~660 Da) is consistent with binding of one NAD⁺ molecule. The results suggest a heterogeneous association of NAD⁺ with the ~145-kDa protein, where distinct peaks likely correspond to tetramers bound to zero, one, or two NAD⁺ molecules (Fig. 1C). The detection of GAPDH tetramers carrying zero to two NAD⁺ molecules may reflect a reduction in NAD⁺ availability upon sample dilution. Tetrameric GAPDH species were still detected after dilution, indicating that the tetrameric form persists in the diluted cell lysate. In addition, the majority of the detected species are likely non-N-terminally acetylated GAPDH.

To confirm that the observed peaks originated from the GAPDH tetramer, NAD⁺ was added to the crude cell lysate to a final concentration of 50 μM, and the binding state was analyzed (Figs. 1B and 1D). Upon NAD⁺ addition, peaks corresponding to the unsaturated states (0–2 NAD⁺-bound forms) were not observed in the mass spectrum, while the spectral features became dominated by signals consistent with the fully NAD⁺-bound GAPDH tetramer. Specifically, in the NAD⁺-supplemented mass spectrum, no peaks were observed in the 23+ charge state region where the 0–3 NAD⁺-bound forms would typically appear (Fig. 1D). For the 24+ charge state, however, a minor peak remained detectable at an m/z slightly above 6000, positioned between the 0 and 1 NAD⁺-bound states. This same peak was present in the untreated spectrum (Fig. 1C). Although the NAD⁺-saturated signals substantially overlapped with this pre-existing background peak in the crude cell lysate, appearing as a major peak with a slight shoulder, the absence of the unsaturated species at the 23+ charge state strongly supports the assignment of these complexes as endogenous GAPDH tetramers directly detected from the crude cell lysate.

The cytoplasmic free NAD⁺ concentration in HEK293T cells has been reported to be approximately 106 μM,¹²⁾ exceeding the reported dissociation constant for the weakest NAD⁺-binding site of GAPDH (~26 μM).¹³⁾ Accordingly, GAPDH tetramers are expected to be highly occupied by NAD⁺ under physiological conditions. The predominance of partially occupied GAPDH tetramers observed in the diluted lysate, together with the shift toward the four-NAD⁺-bound species following supplementation with 50 μM NAD⁺, is consistent with partial dissociation of NAD⁺ during sample preparation.

Although the mass spectrum of the crude cell lysate was dominated by highly complex background signals, careful inspection of specific regions revealed distinct features originating from an additional protein. Upon zooming in on the m/z 3400–4000 range, as shown in Fig. 2, a series of peaks emerged from the baseline, which were tentatively assigned to actin, one of the most abundant intracellular proteins. Actin exists in both filamentous (F-actin) and monomeric (G-actin) forms. G-actin is known to undergo N-terminal acetylation, methylation at His 73, and coordination of ATP and Mg²⁺. The theoretical molecular mass of N-terminally acetylated β-actin lacking the initiator methionine is 41,647.6 Da. When accounting for His 73 methylation, adenosine triphosphate (ATP), and Mg²⁺ binding, the calculated mass of 42,193.1 Da was consistent with the observed mass of 42,194.5 ± 1.4 Da, derived from ions at m/z 3517.1 and 3837.0. Similarly, for γ-actin (theoretical mass of 41,703.7 Da for the acetylated form lacking the initiator methionine), the observed mass of 42,249.1 ± 1.7 Da is consistent with His 73 methylation and binding of ATP and Mg²⁺.

Fig. 2. Native mass spectrum of a 10-fold diluted HEK293T cell lysate in the m/z range of 3400–4000. Peaks corresponding to the 12+ and 11+ charge states of β- and γ-actin were observed, consistent with N-terminal acetylation, His 73 methylation, and binding of ATP and Mg²⁺. ATP, adenosine triphosphate.

Importantly, endogenous protein complexes retaining cofactor binding were directly observed from unpurified human cell lysates. Conventional native MS generally relies on purified samples prepared in volatile buffers because intracellular salts, metabolites, proteins, and other biomacromolecules interfere with ionization and cause peak broadening. Previous direct native MS studies using crude cell lysates have primarily focused on highly overexpressed proteins and their complexes, whereas signals derived from endogenous proteins were rarely observed.^6–8) In such overexpression systems, the target proteins are present at abundances far exceeding those of endogenous proteins. Nevertheless, the present study demonstrates the direct observation of endogenous protein complexes from human cell lysates without purification.

In the present study (Fig. 1; Supplementary Fig. S1), native MS measurements were performed using crude lysates prepared at approximately 0.67–2 × 10⁶ cells/mL (30- to 10-fold dilution). The most abundant endogenous proteins in human cells are estimated to exist at levels of approximately 0.1 fmol per cell,¹⁴⁾ corresponding to a monomeric concentration of only 67–200 nM under our experimental conditions. Because the observed endogenous GAPDH exists as a tetramer, its complex concentration is expected to be approximately one-fourth of that value (16.7–50 nM). In practice, the actual concentration is likely even lower due to factors such as incomplete extraction.

Conventional native MS generally analyzes purified protein complexes at concentrations in the low micromolar range. Previous direct native MS studies using cultured human cell lysates primarily demonstrated the detection of overexpressed protein complexes with molecular masses of up to approximately 35 kDa.^8,10) In contrast, the present study demonstrates that the same general strategy can be extended to the direct detection of an endogenous high-molecular-weight (~145 kDa) protein complex present at substantially lower abundance. The sample preparation workflow used in the present study was essentially similar to that employed in previous direct native MS studies and involved crude cell lysis, removal of insoluble material, and dilution of the resulting lysate prior to nanoESI-MS analysis, without analyte purification or buffer exchange. The methodological advance of the present study therefore lies not in the introduction of a fundamentally new sample preparation procedure, but in demonstrating that the same general direct native MS strategy can detect an endogenous protein complex present at a substantially lower concentration than the overexpressed targets examined previously.

In this context, the direct observation of the endogenous ~145-kDa GAPDH tetramer and its NAD⁺-bound forms at an estimated complex concentration in the tens-of-nanomolar range is particularly notable. These species were detected despite the highly complex lysate matrix containing salts, metabolites, cellular proteins, and other biomacromolecules. Dilution of the crude lysate reduced the amount of matrix introduced into the mass spectrometer while retaining detectable GAPDH signals. Moreover, the continued detection of the GAPDH tetramer at higher dilution factors (Supplementary Fig. S1) demonstrates that the complex remained detectable over the dilution range examined in this study. Taken together, these results demonstrate that direct native MS can be extended from overexpressed target proteins to endogenous protein complexes and their cofactor-binding states without analyte purification or buffer exchange. This capability provides a basis for probing endogenous protein–ligand interactions under conditions that more closely reflect the intracellular environment.

Despite this successful direct observation, several technical limitations remain for further structural confirmation. Although native top-down analysis using an MS/MS system would be highly desirable for definitive identification of GAPDH, sequence analysis of a large protein by MS/MS remains a formidable task, even for its purified forms. Furthermore, selective isolation of the target precursor ion within the highly complex high-m/z region (above m/z 5000) of unpurified crude cell lysates is technically challenging. In particular, overlapping signals and substantial background interference, further compounded by the low endogenous concentration of the target complex, hinder efficient precursor selection and subsequent MS/MS analysis under the present experimental conditions. Future advances in instrumentation enabling precise target ion isolation, highly efficient fragmentation techniques, and optimized MS/MS acquisition in the high-m/z range are expected to further facilitate the direct characterization of endogenous protein complexes and their interaction states in complex biological samples.

In conclusion, the present results demonstrate that endogenous protein complexes can be preserved and directly detected from crude cell lysates under appropriate buffer and dilution conditions. The direct observation of the NAD⁺-bound GAPDH tetramer further indicates that physiologically relevant protein–cofactor interactions can be retained without purification. These findings establish the feasibility of direct native MS analysis of endogenous protein complexes from crude human cell lysates and provide a basis for future development toward single-cell native MS studies under conditions closer to the intracellular environment. It should be noted that the endogenous protein complexes observed in this study mainly correspond to relatively high-abundance cellular proteins, such as GAPDH and actin. Further methodological improvements, including advances in instrumentation and sample preparation strategies, will be required to extend this approach to lower-abundance protein assemblies. This approach may enable direct characterization of endogenous protein assemblies, cofactor-binding states, and drug–target interactions at the single-cell level without extensive purification or overexpression. Furthermore, it may ultimately allow the analysis of dynamic changes in endogenous protein interactions associated with cell cycle progression or cellular stimulation.

SUPPLEMENTARY INFORMATION

Figure S1. Native mass spectra of crude HEK293T cell lysates at higher dilution.

Mass Spectrom (Tokyo) 2026; 15(1): A0198