2-propylphenol Allosterically Modulates COQ8A to Enhance ATPase Activity

Abstract

COQ8A is an atypical kinase-like protein that aids the biosynthesis of coenzyme Q, an essential cellular cofactor and antioxidant. COQ8A’s mode-of-action remains unclear, in part due to the lack of small molecule tools to probe its function. Here, we blend NMR and hydrogen-deuterium exchange mass spectrometry to help determine how a small CoQ precursor mimetic, 2-propylphenol, modulates COQ8A activity. We identify a likely 2-propylphenol binding site and reveal that this compound modulates a conserved COQ8A domain to increase nucleotide affinity and ATPase activity. Our findings promise to aid further investigations into COQ8A’s precise enzymatic function and the design of compounds capable of boosting endogenous CoQ production for therapeutic gain.

Introduction

Coenzyme Q (CoQ) is an essential electron acceptor/donor for diverse cellular processes including oxidative phosphorylation, pyrimidine biosynthesis, and fatty acid oxidation, and also serves as a key lipophilic antioxidant¹. Although this molecule was discovered over 60 years ago², there are still many outstanding questions regarding its biosynthesis and cellular distribution. Both of these processes rely on proteins whose modes-of-action have yet to be determined, including those whose dysfunction is sufficient to cause CoQ deficiency and human disease^3–5.

One such poorly characterized protein is COQ8, an ‘auxiliary’ enzyme that supports the biosynthesis CoQ through unclear means^6,7. COQ8 proteins, which include the paralogs COQ8A and COQ8B in humans, are members of the UbiB family of atypical kinase-like proteins⁸. COQ8A and Coq8p, its corresponding homolog in yeast, are the most well studied members of this protein family. Although COQ8A has a kinase-like fold, it contains a highly conserved region, termed the KxGQ domain, that occludes the protein active site likely preventing canonical protein kinase activity⁹. Instead, evidence to date suggests that COQ8A may utilize an ATPase activity to facilitate CoQ biosynthesis¹⁰. COQ8A also has a well-established connection to human disease, with mutations resulting in a cerebellar ataxia paired with exercise intolerance and a CoQ deficiency^11–14, making it an attractive target for development of molecular tools and therapeutics¹⁵.

We recently investigated COQ8A ATPase activity in detail and, through NMR and activity screens, identified a set of CoQ headgroup-like phenolic compounds, including 2-propylphenol (2-PP) and 2-allylphenol (2-AP), that enhance COQ8A ATPase activity¹⁰. This increase in activity was conserved in other COQ8 homologs and also required the presence of Triton X-100. While this study developed a set of novel tools to study this protein, the activator binding site and mechanism-of-action remained unclear. In this study, we used NMR and hydrogen-deuterium exchange mass spectrometry to reveal that the CoQ precursor mimetic, 2-PP, modulates the quintessential UbiB KxGQ domain to increase COQ8A nucleotide affinity and ATPase activity. We also identify a candidate binding region for this compound. This work expands our understanding of how this protein interacts with CoQ-like molecules and aids in developing the molecular toolkit for this understudied protein family.

Results

2-PP modulates the COQ8A active site and KxGQ domain

Direct activation of the CoQ biosynthetic pathway has great potential for the treatment of primary and secondary CoQ deficiencies. Recently, we discovered that small molecules resembling the CoQ headgroup, including 2-propylphenol (2-PP), enhance COQ8A’s ATPase activity through unclear means¹⁰. A better understanding of how these compounds affect COQ8A could pave the way for more effective CoQ pathway enhancers.

We employed an NMR-based strategy to begin understanding how the CoQ headgroup precursor mimetics stimulate COQ8A ATPase activity. We first validated that COQ8A binds 2-PP by observing reduction of the 2-PP ¹H NMR signal intensity upon addition of COQ8A^NΔ250 (Fig. 1a). Next, based on the size of COQ8A^NΔ250, we opted for a methyl labeling approach amenable to larger proteins to map the protein-ligand interaction. In this scheme, the Ile δ1, Leu δ1/2, and Val γ1/2 methyl groups (collectively ILV) were selectively ¹³C and ¹H labeled in a perdeuterated background. COQ8A^NΔ250 contains 84 ILV residues, providing a significant number of probes to detect 2-PP binding. The ¹H-¹³C HMQC spectrum of apo COQ8A^NΔ250 showed 85% of the expected peaks, with some not observed due to resonance overlap or broadening (full spectra in Supporting Information). Through a combination of NOESY and mutational data we assigned at least one methyl group for 39 residues in COQ8A^NΔ250 (Figure 1b, c, d)(assignment details in Supporting Information). Assigned residues formed four distinct clusters in the N-lobe, nucleotide binding site, KxGQ domain, and part of the C-lobe (Fig. 1b, d). Assignment in the majority of the C-lobe was limited likely because the μs-ms timescale protein dynamics in this region that result in peak broadening. The addition of 1 mM 2-PP to the 219 μM protein sample led to widespread chemical shift perturbations (CSPs) in the COQ8A^NΔ250 spectrum, with 33% of the peaks assigned in both apo and 2-PP-bound states passing our significance threshold (Fig. 1e). Significant shifts occurred in all clusters except for the N-lobe and could suggest that 2-PP is exerting an allosteric effect on the protein structure (Fig. 1f). We compared the four assigned clusters by averaging CSP values for residues in these regions (details in Supporting Information). The N-lobe cluster was essentially unaffected by 2-PP binding, and the most robust changes were observed in the nucleotide binding site and KxGQ domain.

Fig. 1 |. NMR reveals 2-PP induced COQ8A^NΔ250 structural changes.

a, ¹H NMR of the aromatic portion of 2-PP with and without COQ8A^NΔ250. b, Structural overview of COQ8A^NΔ254 (PDB=4ped). c, Mutated ILV residues for methyl assignment mapped onto 4ped. d, Assigned ILV methyls mapped onto 4ped. e, 2-PP induced chemical shift perturbations (CSPs) mapped onto the primary COQ8A sequence. The significance threshold (cyan) represents 1.5x the 10% trimmed mean. f, 2-PP induced CSPs mapped onto the 4ped structure with the color scale ranging from 0–3σ over the 10% trimmed mean. g, Further Triton X-100 induced CSPs mapped onto the 4ped structure with the color scale ranging from 0–3σ over the 10% trimmed mean.

The large CSPs observed in the nucleotide binding site raised the possibility that 2-PP could also bind in this same location. To address this, we performed WaterLOGSY NMR, a sensitive method for observing ligand-protein interaction¹⁶. Spectra were collected for a mixture of COQ8A^NΔ250 and 2-PP before and after addition of ADP. The spectra were phased such that signals from non-binding compounds are negative, while signals arising from binding are positive. Both ADP and 2-PP gave negative WaterLOGSY signals in the absence of protein (purple and blue spectra), and the addition of protein led to either no signal (ADP) or a positive signal (2-PP), both of which are indicative of binding (Fig. 2a). Addition of ADP had no effect on the intensities of the 2-PP WaterLOGSY signals (black and gray spectra), indicating that binding is not competitive. This result demonstrates that 2-PP does not bind in the nucleotide binding site and that observed CSPs in this region are caused by an allosteric effect. Although no precise location of the 2-PP binding site was identified from our NMR experiments, the data suggest the interface of the KxGQ domain and C-lobe as a likely candidate for the binding site since large CSPs were observed. The NMR data also reveal allosteric structural changes induced by 2-PP that extend to the nucleotide binding site.

Fig. 2 |. WaterLOGSY NMR suggests an allosteric effect.

a, WaterLOGSY spectra of 2-PP and ADP with and without COQ8A^NΔ250. Spectra are phased such that signals from non-binding compounds are negative.

Stimulated COQ8A^NΔ250 ATPase activity also requires Triton X-100 (TX-100). We confirmed that TX-100 binds COQ8A^NΔ250 by using ¹H NMR (spectra can be found in Supporting Information). TX-100 (1 mM) was added to the 2-PP-bound COQ8A sample and induced additional CSPs largely localized to the KxGQ domain (Fig. 1g). Based on past molecular dynamics simulations of COQ8A-membrane interactions¹⁰, we suggest that TX-100 could be acting as a membrane mimic to stimulate COQ8A^NΔ250 ATPase activity. The addition of TX-100 at high protein concentrations led to precipitation of COQ8A^NΔ250, preventing further investigations beyond this experiment.

HDX-MS Reveals Largescale Deprotection

To further investigate the interaction between COQ8A and 2-PP, we performed hydrogen deuterium exchange mass spectrometry (HDX-MS). HDX-MS is a powerful tool for evaluating solvent accessibility across regions of a protein and can be used to identify putative ligand binding sites and protein conformational changes based on how this accessibility changes in the presence of a ligand¹⁷. We performed HDX-MS with and without 2-PP (Fig. 3a, b). A total of 92 peptides were identified as statistically different. Regions 251–281, 289–306, 366–381, 397–406 showed strong deprotection upon 2-PP binding and regions 287–288, 331–340, 347–365, 382–396, 407–414, 448–468, 482–494, 529–544, 571–581, 607–618, 629–639 showed intermediate deprotection (Fig. 3c). The observed broad-scale deprotection of the protein upon 2-PP binding is consistent with 2-PP inducing an allosteric effect. Such broad scale deprotection, consistent with increased dynamics or loss of structure of a protein is unusual in our experience. The strongest deprotection was seen in the KxGQ domain, in accord with the significant CSPs observed in this region via NMR.

Fig. 3 |. HDX-MS reveals largescale deprotection.

a, Crystal structure of COQ8A^NΔ254 (PDB=5i35). b, HDX-MS results mapped onto the 5i35 structure and annotated peptides mapped onto the primary COQ8A sequence. c, Representative uptake plots for different peptide classifications showing the effect of 2-PP binding to COQ8A^NΔ250 (n = 3 ± SD).

HDX-MS Nominates a 2-PP Binding Site

In addition to this general deprotection, we also found a 13 amino acid region (region 582–594) whose solvent accessibility was reduced upon 2-PP binding (Fig. 3b, 3c, 4a). These data suggest that a 2-PP binding pocket is in this region, which is near a hydrophobic patch previously shown to open upon nucleotide binding¹⁴. Finally, two peptides located in the region 563–570 showed a composite behavior, showing both protection and deprotection over time. Importantly, there are several false negatives (peptides showing negligible differences located within the regions mentioned above). This false categorization is most probably due to the strict k-means clustering classification or due to high standard deviation in the HDX measurements.

To test whether the protected region is involved in 2-PP binding, we measured ATPase activity ± 2-PP for WT COQ8A^NΔ250, the catalytically inactive D507N mutant, and six individual point mutants (Fig. 4b). Each construct exhibited comparable purity. Although select mutants had slightly decreased basal melting temperatures (Fig. 4c), all mutants had enhanced stability upon nucleotide binding (Fig. 4d). Interestingly, mutations to this region had variable effects, with three mutants exhibiting no stimulation upon 2-PP treatment, and another exhibiting a nearly two-fold increase in ATPase activity compared to WT. Collectively, these results are consistent with this region harboring important residues for 2-PP binding and activation, and suggest that stronger activation may be possible through further ligand modification; however, a deeper structure-function analysis will be required to validate these findings and further elucidate the 2-PP mode-of-action.

Fig. 4 |. HDX-MS nominates a 2-PP binding site.

a, Region protected upon binding as identified by HDX-MS, highlighting introduced mutations. b, ATPase activity of WT or protected region mutants with either DMSO or 1 mM 2-PP (n = 3 ± SD). c, Thermal stability of WT COQ8A^NΔ250 and protected region mutants, as assessed by DSF (n = 4 ± SD). d, Thermal stabilization of WT COQ8A^NΔ250 and protected region mutants by 1 mM ATP or 1 mM ADP, as assessed by DSF (n = 4 ± SD). e, Nucleotide binding curves for COQ8A^NΔ250 with either 1 mM 2-PP or DMSO, as assessed by DSF (n = 4 ± SD).

Last, we determined the K_d,app for ATP and ADP in the presence of 2-PP by using differential scanning fluorimetry (DSF) (Fig. 4e). We found modest but significant increases in nucleotide affinity with the addition of 2-PP, which could contribute to the observed increase in ATPase activity. Overall, our data indicate that 2-PP stimulates COQ8A ATPase activity by inducing an allosteric modulation of the KxGQ domain and nucleotide binding site. We also implicate the interface between the KxGQ domain and the C-lobe as a potential 2-PP binding site.

Discussion

The level of CoQ in the body has a variety of complex implications in human health¹. Owing to its extreme hydrophobicity and limited bioavailability¹⁸, cells are primarily reliant on endogenous production to retain their CoQ pool. Therefore, the ability to manipulate endogenous CoQ biosynthesis presents unique opportunities to affect human health. Nine of the genes required for CoQ production harbor known mutations that result in primary CoQ deficiency, leading to disorders such as ataxias and myopathies¹. Previously, the overexpression of Coq8p in S. cerevisiae has been utilized to overcome deficiencies in several CoQ biosynthetic proteins^19,20, although the mechanism remains unclear. Our investigation into 2-PP activation sets the foundation for understanding the regulation of COQ8 protein activity at a deeper level, and suggests that the specific activation of COQ8 proteins could have therapeutic benefit against diseases associated with primary and secondary CoQ deficiency.

The importance of UbiB proteins in nature extends well beyond CoQ production. Recently, we connected yeast intermembrane space UbiB proteins to cellular CoQ distribution²¹. Other eukaryotic UbiB family members have been implicated in stress responses²², metal homeostasis²³, and phospholipid metabolism²⁴. Here we have begun to reveal how the activity of the archetypal member of this family, COQ8A, is modified by CoQ precursor mimetics at a structural level. Our work promises to inform how these proteins operate mechanistically and guide the design of more efficient small molecules to manipulate their functions.

Methods

COQ8A^NΔ250 General Expression and Purification

The general purification method has been documented previously⁹. COQ8A^NΔ250 constructs were overexpressed in E. coli by autoinduction²⁵. Cells were isolated and frozen at −80 °C until further use. For protein purification, cells were thawed and resuspended in Lysis Buffer [20 mM HEPES (pH 7.2), 300 mM NaCl, 10% glycerol, 5 mM 2-mercaptoethanol (BME), 0.25 mM phenylmethylsulfonyl fluoride (PMSF)] (4 °C). The cells were lysed by sonication (4 °C, 75% amplitude, 20 s × 2). The lysate was clarified by centrifugation (15,000 g, 30 min, 4 °C). The cleared lysate was mixed with cobalt IMAC resin (Talon resin) and incubated (4 °C, 1 h). The resin was pelleted by centrifugation (700 g, 2 min, 4 °C) and washed four times with Wash Buffer [20 mM HEPES (pH 7.2), 300 mM NaCl, 10% glycerol, 5 mM BME, 0.25 mM PMSF, 10 mM imidazole]. His-tagged protein was eluted with Elution Buffer [20 mM HEPES (pH 7.2), 300 mM NaCl, 10% glycerol, 5 mM BME, 100 mM imidazole]. The eluted protein was concentrated with a MW-cutoff spin filter (50 kDa MWCO) and exchanged into Storage Buffer [20 mM HEPES (pH 7.2), 300 mM NaCl, 10% glycerol, 5 mM BME]. The concentration of 8His-MBP-[TEV]-COQ8A^NΔ250 was determined by its absorbance at 280 nm (ε = 96,720 M⁻¹cm⁻¹)(MW=89.9 kDa). The fusion protein was incubated with Δ238TEV protease (1:50, TEV:fusion protein, mass:mass)(1 h, RT). The TEV protease reaction mixture was mixed with cobalt IMAC resin (Talon resin) and incubated (4 °C, 1 h). The resin was pelleted by centrifugation (700 g, 2 min, 4 °C). The unbound COQ8A was collected and concentrated with a MW-cutoff spin filter (30 kDa MWCO) and exchanged into storage buffer. The concentration of COQ8A^NΔ250 was determined by Bio-Rad Protein Assay according to manufacturer protocol. The protein was aliquoted, frozen in N_{2 (l)}, and stored at −80 °C. Fractions from the protein preparation were analyzed by SDS-PAGE. Coq8p^NΔ41 and mutants were also purified using this general method.

Differential Scanning Fluorimetry

The general DSF method has been documented previously²⁶. To determine the nucleotide K_d,app, 20 μL reactions contained 1 μM COQ8A^NΔ250, either 0–4.96 mM ADP (Sigma A2754) or 0–2.48 mM ATP (Sigma A2383), and either 1 mM 2-PP or 0.4% matched DMSO. The other reaction components were 4 mM MgCl₂, 100 mM HEPES pH 7.5, 150 mM NaCl, and 5x Sypro Orange dye. Samples were made in MicroAmp Optical 96-well reaction plates (Thermo N8010560), centrifuged (200 g, RT, 30 sec) and incubated at room temperature in the dark for 10 min. Fluorescence was then monitored with the ROX filter using a QuantStudio 6 Real-Time PCR system (QuantStuido Real-Time pCR v1.2 software) along a temperature gradient from 4–95 °C at a rate of 0.025 °C/s. Protein Thermal Shift software v1.3 (Applied Biosystems) was used to determine T_m values by fitting to a Boltzmann model. Melt curves flagged by the software were manually inspected. ΔT_m values were determined by subtracting the average of buffer only controls from the same plate. ΔT_m values were plotted against [nucleotide] in GraphPad Prism Version 8.4.3 and fit to nonlinear regression (one site specific binding) to determine K_d,app values. Experiments were performed in quadruplicate and error bars represent SD. Error values in K_d,app represent 95% confidence intervals. Nucleotide binding experiments for COQ8A^NΔ250 WT and mutants with buffer only, 1 mM ATP or 1 mM ADP were also performed using this method with no DMSO vehicle (n=4, error bars represent SD).

ATPase Assay

ATPase assays for COQ8A^NΔ250 were performed essentially as described in Reidenbach et al¹⁰. The reaction was initiated by adding a mixture of ATP (Promega V703), 2-propylphenol (Sigma W352209) or matched DMSO, and Triton X-100 (Sigma T9284) in Reaction Buffer [100 mM HEPES pH 7.5, 4 mM MgCl₂, 150 mM NaCl] to purified COQ8A^NΔ250 also in reaction buffer. Final reaction conditions were 100 μM ATP, 0.5 μM COQ8A^NΔ250, 1 mM Triton X-100, and either 1 mM 2-PP or matched DMSO in Reaction Buffer. The final reaction volume was 15 μL. Reactions were performed in clear 384 well plates (Fisher 12565506). Once the reactions were started and mixed, plates were sealed and incubated at 30°C for 45 min. Reactions were then quenched with 35 μL of Cytophos^™ reagent (Cytoskeleton, Inc. BK054), incubated at RT for 10 min, and absorbance was read at 650 nm in a Biotek Cytation 3 plate reader. Absorbance was converted to concentration of inorganic phosphate with a 0–50 μM standard curve run in parallel using the phosphate standard provided in the Cytophos^™ kit diluted in reaction buffer. Experiments were performed in triplicate and error bars on the graph represent SD.

Site Directed Mutagenesis

The COQ8A^NΔ250 in pVP68K vector has been described previously⁹. Point mutations were introduced via PCR-based mutagenesis. Reactions were DpnI digested and transformed into DH5α competent E. coli cells. Plasmids were isolated from transformants and Sanger sequencing was used to identify those harboring the correct mutations. Oligonucleotides were purchased from IDT (Coralville, IA, USA).

1D Ligand-Observed NMR

COQ8A^NΔ250 was purified with the general method described above. After the final buffer exchange the protein was concentrated to less than 1 mL. The sample was centrifuged (20,000 g, 10 min, 4 °C) and the soluble portion was filtered through a 0.22 μm filter. The protein was further purified using a HiLoad 16/600 Superdex 75 pg gel filtration column (GE) with an isocratic elution at 1 mL/min. The protein was exchanged into Analytical Buffer [100 mM NaCl, 20 mM HEPES, pH 7.2] during the FPLC run. Fractions were analyzed by SDS-PAGE and concentrated for NMR analysis. Binding of 2-PP and TX-100 was assessed by ¹H NMR. Samples consisted of 200 μM ligand, 20 μM Sodium trimethylsilylpropanesulfonate (DSS), 8% D₂O, 100 mM NaCl, 20 mM HEPES, pH 7.2. Spectra were collected at 30 °C on a Bruker Avance III spectrometer running Topspin version 3.5pl7 and operating at 500 MHz (¹H). After collecting spectra without protein, a final concentration of 22.8 μM COQ8A^NΔ250 was added to the sample and spectra were collected again.

For WaterLOGSY experiments COQ8A^NΔ250 was expressed and purified using the general method and stored at −80 °C. Protein was thawed on ice and exchanged four times into dPBS (Thermo 14190250) using a 30 kDa MW-cutoff spin filter to remove HEPES and glycerol. It was then further dialyzed for 2 h, then again overnight using a Slide-A-Lyzer Mini Dialysis Device (20 kDa MWCO, 0.5 mL) (Thermo 88402) based on manufacturer protocol. The sample was stored at 4 °C until NMR analysis. Samples were prepared by making a 2X stock solution containing 1 mM 2-PP (from stock in DMSO-d6), 10% D₂O. Half of the stock solution was diluted to a final concentration of 500 μM 2-PP, 5% D₂O. An aliquot of COQ8A^NΔ250 stock was added to the other half, then brought to a final concentration of 25 μM COQ8A, 500 μM 2-PP, 5% D₂O. ADP was titrated into final concentrations of 500 and 750 μM using a 100 mM stock. DSS was added to each sample at the end of each titration for referencing. The final samples contained 0.08% DMSO-d6. WaterLOGSY experiments were recorded at 10 °C on a Bruker Avance III spectrometer operating at 600 MHz (¹H) using Topspin 3.5pl7. Experiments were recorded using the Bruker pulse program ephogsygpno.2 with a mixing time of 1.5 s and 2 s recycle delay. A 40 ms T₁ρ filter was used to suppress the protein signals. All 1D spectra were processed with 1 Hz exponential line broadening using nmrPipe²⁷ and visualized using the nmrglue python package²⁸.

¹³C-Labeled COQ8A^NΔ250 Expression and Purification

The ¹³C labeling schemes have been described previously^29–31. E. coli harboring the COQ8A^NΔ250 PVP68K plasmid were used to inoculate 20 mL LB cultures with 50 mg/L kanamycin (KAN) and 15 mg/L chloramphenicol (CAM). Cultures were grown at 37 °C, 220 RPM until OD₆₀₀=0.8–1. Then 5 mL of the LB culture was used to inoculate a 50 mL culture of M9/H₂O minimal media [12.8 g/L Na₂HPO₄•7H₂O, 3g/L KH₂PO₄, 0.5 g/L NaCl, 1 g/L NH₄Cl, 4 g/L glucose, 2 mM MgSO₄, 0.1 mM CaCl₂, 0.5 mg/L FeSO₄, 5 mg/L thiamine, 0.5X BME vitamin solution (Sigma B6891), KAN, CAM, pH=7.4 in H₂O]. M9/H₂O cultures were grown until OD₆₀₀=0.8–1. Cells were pelleted by centrifugation (4000 g, 15 min, RT) and spent media was decanted. Cell pellets were then resuspended in 100 mL M9/D₂O minimal media [6.8 g/L Na₂HPO₄, 3 g/L KH₂PO₄, 0.5 g/L NaCl, 4 g/L deuterated glucose (Sigma 552003), 0.25 g/L ISOGRO-15N, D Powder (Sigma 608300), 1 g/L ¹⁵NH₄Cl (Sigma 299251), 2 mM MgSO₄, 0.1 mM CaCl₂, KAN, CAM, pH=7.4, in D₂O (Sigma 756822)]. 100 mL M9/D₂O minimal media cultures were grown at 37 °C until OD₆₀₀=0.8–1. The 100 mL cultures were then diluted in 400 mL M9/D₂O minimal media to yield a final 500 mL culture. 500 mL cultures were grown at 37 °C. At 30 min prior to induction (OD₆₀₀=0.75) labeling precursors were added to the cultures. ¹³C-ILV labeling used 150 mg/L 2-Keto-3-(methyl-d₃)-butyric acid-4-¹³C,3-d sodium salt (Sigma 691887) and 70 mg/L 2-Ketobutryic acid-4-¹³C,3,3,d₂ sodium salt hydrate (Sigma 589276) as the precursors. Dimethyl LV labeling for geminal pair determination used 150 mg/L 2-Keto-(3-methyl-¹³C)-butyric-4-¹³C,3-d acid sodium salt (Sigma 589063). Valine only labeling used 150 mg/L 2-Keto-3-(methyl-d₃)-butyric acid-4-¹³C,3-d sodium salt (Sigma 691887) and 150 mg/L sodium 4-methyl-2-oxovalerate (Sigma K0629). Cultures were grown for another 30 min at 37 °C to reach OD₆₀₀=0.8–1. Cultures were cold shocked at 4 °C for 10 min, then induced with a final concentration of 100 μM IPTG. Protein was expressed at 25 °C for 20 h. Cells were then isolated and frozen at −80 °C until further use. Proteins were generally purified using the above method with the following changes. Protein preparations for NMR used no glycerol. After the final COQ8A concentration, the protein was exchanged into Analytical Buffer [100 mM NaCl, 20 mM HEPES, pH=7.2] using the 30 kDa MW-cutoff spin filter. Protein was quantified by its absorbance at 280 nm (ε = 28,880 M⁻¹cm⁻¹)(MW=45.6 kDa) and stored at 4 °C until NMR analysis.

Protein-Observed NMR

All 2D-4D spectra of COQ8A were recorded at 30°C in 20 mM HEPES, 100 mM NaCl, pH 7.2. 5–10% D₂O was added for the lock and 50 μM DSS was added for referencing. Most spectra were recorded on a Bruker Avance III spectrometer operating at 750 MHz (¹H) and equipped with a cryogenic probe. The 4D HMQC-NOESY-HMQC³² was recorded with 8 scans per FID with 36 (¹³C) × 28 (¹H) × 36 (¹³C) × 1024 (¹H, direct) complex points non-uniformly sampled at 19%. The spectrum was reconstructed using the SMILE algorithm implemented in nmrPipe. The NOE mixing time was 300 ms. The total experimental time was ~7 days.

3D NOESY-HMQC were collected on an 800 MHz (¹H) Varian spectrometer running vnmrj version 4.2 and equipped with a cryogenic probe. Two were recorded on a dimethyl-labeled sample of apo COQ8A where both Cγs and Cδs are labeled per V/L to establish geminal methyl pairing. These spectra were recorded with 98 (¹³C) × 98 (¹³C) × 1024 (¹H, direct) complex points and mixing times of 50 and 200 ms. The experiment with a 50 ms mixing time used 16 scans per FID while 32 scans were used for the 200 ms experiment. A third 3D NOESY-HMQC was recorded for COQ8A in the presence of 2 mM adenosine 5’-(β, γ-imido)triphosphate (AMPPNP). This spectrum was recorded in two blocks, both with 68 (¹³C) × 68 (¹³C) × 1024 (¹H, direct) complex points and a mixing time of 350 ms. The first block consisted of 32 scans per FID while the second block consisted of 16 scans per FID. All NOESY spectra were processed in nmrPipe using the SMILE algorithm³³. The two blocks of the NOESY spectrum with AMPPNP were co-added using nmrPipe after NUS reconstruction.

All spectra were analyzed using ccpNmr Analysis version 2.4.2³⁴. NOESY peak picking was performed manually. Chemical shift assignment was performed using the NOESY data in a ‘methyl walk’ approach³⁵, described in more detail in the Supporting Information. Chemical shift perturbations (CSPs) were calculated according to the formula:

$$\text{CSP}\text{=}{\left[\Delta {\delta }_H{}^2+{\left(W*\Delta \delta c\right)}^2\right]}^{0.5}$$

where Δδ is the difference in chemical shift for a given residue between apo and 2-PP-bound states and W a chemical shift weighting factor (W_C = |γ_C/γ_H| = 0.251).

HDX-MS of COQ8A^NΔ250 with 2-PP

The protein was expressed and purified by using the above method and stored at −80 °C. At the time of HDX, the protein was thawed on ice and exchanged into Analytical Buffer [100 mM NaCl, 20 mM HEPES, pH = 7.2] by using a 30 kDa MW-cutoff spin filter. Protein concentration was quantified using the Bio-Rad Protein Assay. HDX experiments were performed similarly as described elsewhere³⁶. Briefly, automated HDX-MS experiments were carried out using a LEAP Technologies HDX PAL DHR robot (Carrboro, NC) with a three-valve configuration system coupled to a Bruker MaXis Q-ToF operating in positive-ion mode. HDX was initiated by diluting 9 μL of purified COQ8A (15 μM in 20 mM HEPES, 100 mM NaCl, pH 7.2) with 81 μL of 20 mM HEPES, 100 mM NaCl, 5% DMSO, pD 7.2 in D₂O (Cambridge Isotope Laboratories, Inc., Tewksbury, MA) with and without 8 mM 2-propylphenol. pD was directly measured with a glass electrode and corrected for the deuterium isotope effect (pD = pH + 0.40)³⁷. Samples were labeled in triplicate for each labeling time (45, 180, 720 and 2880 s) at 25 °C. Immediately following, 85 μL of each labeled solution was quenched with 85 μL precooled quench buffer (6 M Urea, 10 mM PBS, pH 2.4) at 1 °C. Non deuterated controls were prepared identically but using 20 mM HEPES, 100 mM NaCl, 5% DMSO, pH 7.2. Quenched samples were injected into a temperature-controlled chromatography cabinet maintained at 0 °C throughout the whole experiment to reduce back-exchange. In-line digestion was performed by passing the quenched sample through an custom immobilized pepsin column(2.1 × 50 mm) maintained at 10 °C with 0.1% formic acid in H₂O for 180 s at a flow rate of 0.2 mL min⁻¹. Peptic peptides were trapped and desalted on a ZORBAX 300SB-C8 trap (2.1 × 12.5 mm, 5 μm particles) with 0.1% formic acid in H₂O for 60 s at a flow rate of 0.2 mL min⁻¹. Peptides were separated by using reversed phase chromatography with a XSelect CSH C18 column (2.1 × 50 mm, 2.5μm, Waters, Manchester, UK). HDX-MS data was analyzed using HDExaminer (version 3.2.1, Sierra Analytics, Modesto, CA). Adjustment of LC retention times and curation of the data were performed manually on all peptides. Non deuterated COQ8A^NΔ250 peptic peptides were identified using DIA CID MS/MS prior to conducting HDX. Inline digestion of COQ8A^NΔ250 resulted in 100% sequence coverage. Differential HDX data were evaluated for significant differences by using a hybrid significance test³⁸. Normalized HDX differences were clustered³⁶ to define thresholds for strong, moderate, and negligible HDX differences. In Supporting Information is a volcano plot for identification of statistically different peptides and magnitude of the effects. Negligible (black), moderate (yellow) and strong (blue) effects was revealed by k-mean clustering of the data. Visualization of HDX differences was done plotting clustered significant differences into the COQ8A structure (PDB=5i35)¹⁴.

Data Availability

ILV methyl chemical shifts for COQ8A in the apo state, in the presence of 1 mM 2-PP, and in the presence of 1 mM 2-PP and 1 mM Triton X-100 have been deposited in BMRB under accession numbers 51434, 51435, and 51436. Raw NMR data is being deposited in BMRBig.