CIVIL ENGINEERING 365 ALL ABOUT CIVIL ENGINEERING

[ad_1]

The growth of the yeast, the purification of the enzyme, and the preparation of the nanodiscs were described earlier15. TMC207-fumarate was a provided by Janssen Pharmaceutica and from the NIH AIDS Reagent Program. BDQ was dissolved in methanol and aliquoted into microcentrifuge tubes and the methanol was removed under vacuum. The dried BDQ sample was stored desiccated under nitrogen or argon, in a sealed bag at 4 °C. The sample was dissolved either in methanol (ATPase experiments) or DMSO (ATP synthesis experiments) and added to the sample for assay. Any remaining sample was discarded. Oligomycin A (Santa Cruz Biotechnology, Santa Cruz, CA) was dissolved in methanol at a concentration of 1–10 mg/ml and stored at −20 °C. For the inhibition assays, the samples were incubated for 5 min with the inhibitor in the reaction buffer at the stated assay temperature just prior to starting the reaction with substrate.

Statistics and reproducibility

The inhibition data were analyzed and the IC50 calculated using SigmaPlot 11.0 and errors are reported as standard error. For the experiments using isolated mitoplasts from HEK293 cells, only fresh mitoplasts were used with the exception of a control experiment that tested a new batch of BDQ which used a frozen batch. The final results for BDQ inhibition using mitoplasts from HEK293 cells were based on five separated experiments. For mitochondria, SMP, and F1Fo incorporated into bicelles, isolated from yeast, the samples were frozen and defrosted just before use. For these experiments, the results are based on three experiments. While we only used one batch of yeast SMP and F1Fo, we performed similar experiments on oligomycin B, C, and 21-hydroxy derivatives, which gave similar results. We also performed experiments (but with less data points) using new batches of BDQ and enzyme preparations to assess any variations due to BDQ or sample preparations. The relative integrity of the mitochondria and mitoplasts was assessed by measuring the membrane potential prior to each experiment. We used four lots of BDQ for the experiments: two from Janssen and two from NIH AIDS program. While we did not do a full investigation, we did test each lot to determine if the lots had comparable activities—which they did.

Mitochondria isolation and mitoplast preparation

Yeast mitochondria were isolated from W303-1A yeast strain essentially as described39 but using lyticase to digest the cell wall (lyticase 0.4 mg/g yeast; 90 min. at room temperature). Expression and purification of lyticase was done as described40. The mitochondria were frozen and stored in liquid nitrogen. Human mitochondria were isolated from HEK293S GnTI (ATCC CRL-3022) cells. Cells were grown to 90% confluence in Dulbecco’s Modified Eagle Medium supplemented with fetal bovine serum (10%), l-glutamine (4 mM), glucose (4.5 g/L), sodium pyruvate, in the absence of penicillin and streptomycin. HEK293S GnTI cells were incubated at 37 °C with 5% CO2. Mitochondria were isolated as described41.

To prepare mitoplasts from mitochondria isolated from HEK293 cells, the mitochondria (0.6–0.7 ml, 5 mg/ml) were incubated on ice for 20 min., with glyco-diosgenin (0.3 mg/ml, Anatrace, Maumee, OH). To this, ice cold IBc buffer was added (0.2 M sucrose, 10 mM MOPS, 1 mM EGTA, pH 7.4) to final volume of 1.2 ml. The sample is centrifuged at 7000 × g for 10 min at 4 °C and the pellet resuspended in 80% of the initial volume with IBc buffer and the protein concentration is measured.

For freezing, the mitoplasts are centrifuged at 7000 × g for 10 min. and resuspended in freezing buffer42 (10 mM HEPES-KOH pH 7.7, 300 mM trehalose, 10 mM KCl, 0.1% BSA, 1 mM EDTA, 1 mM EGTA) to 10 mg/ml and frozen and stored in liquid nitrogen.

Bicelle reconstitution of the yeast ATP synthase

Saccharomyces cerevisiae strain USY006 (MAT α, ade2-1, his3-11-15, leu2-3-112, trp1-1, ura3-52, can1R-100, atp2::LEU2, TRP1::ATP2-his6) was used for the isolation of the F1Fo ATP synthase. The strain, W303-1A (MATa ade2-1, his3-11,15, leu2-3,112, 112, trp1, ura3-1) was used to isolate coupled yeast mitochondria. The yeast ATP synthase was purified and reconstituted into bicelles as described43. Briefly, to a final protein concentration of 12 mg protein/ml, 0.7% (w/v) bicelle (3:1 molar ratio of 1,2-dimyristoyl-sn-glycero-3-phosphocholine:1,2-dihexanoyl-sn-glycero-3-phosphocholine) was added from 40% stock, mixed gently and incubated on ice overnight. Another 0.55% of the same bicelle was added the next day for a total 1.25% bicelle (w/v). This preparation was dispensed into small volumes, quick frozen in liquid nitrogen, and stored in liquid nitrogen.

ATPase activity assay

The ATPase activity was measured by the coupled enzyme reaction43 at 30 °C in a buffer containing 0.25 M sucrose, 50 mM K-HEPES, pH 8.0, 3 mM MgSO4, 2 mM ATP. In the assay, protein concentration of the yeast SMP and purified ATP synthase was about 0.3 mg/ml and 1 μg/ml, respectively. For these experiments, the experiments were repeated three times with the same preparation of SMP of F1Fo. However, we purified F1Fo on at least five different occasions and did not observe any differences. For inhibition studies, the reaction mix was incubated for 5 min with the inhibitor at 30 °C, without ATP, and then started with the addition of ATP.

ATP synthase activity assay

ATP synthesis was determined essentially as described previously44. Mitochondria (5 μg) was added to buffer (0.65 M mannitol, 20 mM bis-tris-propane, 2 mM phosphate, 0.36 mM EGTA, pH 6.8) and incubated at room temperature for 2 min after which the inhibitor was added and incubation was continued for an additional 5 min and then succinate/ethanol (5 mM/0.8% v/v, for yeast mitochondria) or glutamate/malate (5 mM/2.5 mM for human mitoplasts), and ADP (0.2 mM) was added to start the reaction (note, the ADP must be treated to reduce contaminating ATP. See below). Di-(adenosine-5′) pentaphosphate (40 μM) was added when assaying human mitoplasts to reduce matrix adenylate kinase activity. Di-(adenosine-5′) pentaphosphate increased the background signal and thus was limited to 40 μM. After 60 min, the reaction was stopped with perchloric acid (3.5%) and EDTA (12.5 mM) and the samples were neutralized to pH 6.5 by the addition of NaOH to 0.5 M, and diluted with water by 50–100-fold. ATP was measured with luciferin/luciferase assay in reaction consisting of 25 mM tricine, pH 7.8, 0.2 mM MgSO4, 0.005 mM EDTA, 1 mM DTT, 0.005 mM NaN3, 1 mM d-luciferin and 0.625 μg/ml luciferase and light was measured using a scintillation counter with the channels wide open. A standard curve for ATP was done for all experiments and all reported measurements were within the range of the ATP concentrations used for the standard curve. In some cases, points of the standard curve were repeated if they were outliers. The standard curve was fitted to a polynomial, which was used to calculate the concentration of ATP. Bedaquiline did not affect the activity of luciferase under these conditions.

Substrate ADP was treated with hexokinase and glucose to reduce the level of contaminating ATP. ADP (100 mM) was dissolved in buffer (1 ml, 20 mM Tris-Cl, 0.24 M glucose, 5.5 mM MgCl2, pH 7.6), hexokinase (100 units) was added, and the mixture was incubated at 25 °C overnight. The reaction was stopped by heating the mixture at 65 °C for 20 min, centrifuged at 14,000 × g for 3 min to remove the precipitated protein. This treatment reduced contaminating ATP by 99%.

For yeast mitochondria, the rate of cyanide sensitive ATP synthesis was 51.3 ± 2.6 nmoles/ATP/min/mg protein (n = 3), was inhibited 98% with oligomycin (1 μM) (n = 3, 1 preparation stored in aliquots in liquid nitrogen). Cyanide insensitive ATP synthesis was 0.1 nmoles/min/mg protein. For human mitochondria, the rate of cyanide sensitive ATP synthesis was 9.0 ± 0.6 nmoles/ATP/min/mg protein (n = 5), was inhibited nearly completely with oligomycin (1.25 μM) (n = 5, 5 different fresh preparations). Total ATP synthesis for yeast mitochondria was inhibited by 98.6% with 1.5 mM sodium cyanide and by 59.2% with human mitoplasts. Cyanide insensitive ATP synthesis was likely due to matrix adenylate kinase activity. We did not observe differences in the results using fresh vs. frozen mitochondria or mitoplasts. While the specific activity for ATP synthesis of mitochondria isolated from a human cancer line was not reported in the earlier study24, the reference for the method used reports a specific activity of about 15 nmoles/min/mg protein (oligomycin sensitive) using mitochondria isolated from MRC5 fibroblasts45. Also, for comparison, the specific activity for ATP synthesis using a different method is 0.27 nmoles/min/mg for membranes from Mycobacterium bovis membranes and 0.96 nmoles/min/mg for membranes from Mycobacterium smegmatis46.

Membrane potential measurements

The membrane potential was assessed by measuring fluorescence quenching of Rhodamine 123 at 488 nm excitation and 525 nm emission in a buffer containing 0.65 M mannitol, 20 mM bis-tris-propane, 2 mM tris-phosphate, 0.3 mM EGTA, pH 6.8 with Rhodamine 123 dissolved in methanol (1.0 µM). Yeast mitochondria were energized with ethanol (0.4% v/v), whereas human mitochondria were energized with glutamate/malate (5 mM/2.5 mM).

Cryo-electron microscopy data acquisition

Purified yeast F1Fo reconstituted in nanodiscs at the concentration of 1 mg/ml (2.5 μl, ≈1.7 μM) was applied to a glow-discharged Quantifoil holey carbon grid (1.2/1.3, 400 mesh), and blotted for 3 s with ~91% humidity before plunge-freezing in liquid ethane using a Cryoplunge 3 System (CP3, Gatan). For cryo-EM15, incubation of F1Fo with BDQ was performed in two ways: one is on ice, from which 30 mM stock solution of BDQ in dimethyl sulfoxide was added to a final concentration of 30 μM (twofold higher stoichiometry than of the c-subunit at ten copies/complex), and incubated for 30 min; another one is incubation at room temperature for 10 min, with final concentration of BDQ at 5 μM. The mixture (3.5 μl) from either incubation was applied to a grid, blotted and plunge frozen. Cryo-EM data of sample incubated on ice were recorded on a 300 kV Polara electron microscope (FEI) at Harvard Medical School, data of the sample incubated at room temperature were collected on a 200 kV Talos Arctica microscope (FEI) at University of Massachusetts Medical School. All cryo-EM movies were manually recorded with a K2 Summit direct electron detector (Gatan) in super-resolution counting mode using UCSFImage447. Details of the EM data collection parameters are listed in Table 2.

Table 2 Cryo-EM data collection, refinement, and validation statistics.

EM image processing

EM data were processed as previously described15. Dose-fractionated super-resolution movies collected using the K2 Summit direct electron detector were binned over 2 × 2 pixels, and subjected to motion correction using the program MotionCor2. A sum of all frames of each image stack was calculated following a dose-weighting scheme, and used for all image-processing steps except for defocus determination. CTFFIND4 was used to calculate defocus values of the summed images from all movie frames without dose weighting. Particle picking was performed using a semi-automated procedure with SAMUEL and SamViewer. The images collected from Polara and Talos Arctica were combined together for the final data process. Two- and three-dimensional (2D and 3D) classification and 3D refinement (Supplementary Fig. 5) were carried out using “relion_refine_mpi” in RELION. Masked 3D classification focusing on Fo with residual signal subtraction was performed following a previously described procedure. All refinements followed the gold-standard procedure, in which two-half data sets were refined independently. The overall resolutions (Supplementary Fig. 6) were estimated based on the gold-standard criterion of Fourier shell correlation (FSC) = 0.143. Local resolution variations (Supplementary Fig. 6) were estimated from two-half data maps using ResMap. The amplitude information of the final maps was corrected by applying a negative B factor using the program bfactor.exe.

Model refinement

The initial model of Fo was derived from our previous ATP synthase model (PDBs: 6CP7). Initial model was rigid-body fitted to our cryo-EM map, extensively rebuilt in Coot and refined in Refmac using the script refine_local, and subsequently, using real-space refinement in Phenix essentially as described earlier15. The final model was validated with statistics from Ramachandran plots, MolProbity scores, and EMRinger scores (Table 2). MolProbity and EMRinger scores were calculated as described15.

Molecular dynamics simulations

To guide the interpretation of the experimental cryo-EM data, we carried out a series of molecular dynamics simulations aimed at identifying an energetically favorable binding pose that is also consistent with the density map. As a first step, we simulated BDQ bound to the isolated c-ring, embedded in a POPC lipid bilayer (Fig. 4a). BDQ molecules were modeled on three nonadjacent proton-binding sites, in a tentative initial configuration that appeared compatible with the cryo-EM density data, but differed from that observed previously for the mycobacterial c-ring23. These simulations were based on the high-resolution crystal structure of the yeast mitochondrial c-ring (PDB 3U2F)15. Two possibilities were considered, for each site: one in which cGlu59 in the c-subunit is protonated while BDQ is neutral, and another in which the amine group in BDQ is protonated while cGlu59 is ionized. In the former case, cGlu59 donates a hydrogen bond to BDQ; in the latter, cGlu59 and BDQ form a salt-bridge. The two systems were equilibrated following a staged protocol comprising a series of restrained simulations. The protocol consists of both positional and conformational restraints, gradually weakened over 90 ns, and individually applied to protein side chains and backbone as well as the BDQ molecule. Subsequently, three simulated trajectories of 100 ns each were calculated for each system, free of any restraints. In the simulation where BDQ is neutral and cGlu59 protonated, the three inhibitor molecules were seen to gradually dissociate from the c-ring, spontaneously (Fig. 4b). By contrast, ionized BDQ was found to be stably bound in simulation, for the three binding sites considered and in three independent calculations (Fig. 4b) (note that this ionized configuration is also the most probable given the much greater proton affinity of a methylated amine compared with a carboxylic group; their solution pKa values are 10.6 and 4.0, respectively). Analysis of the simulation data for bound BDQ show the inhibitor to be quite dynamic; nonetheless, a classification of the configurations sampled according to pairwise similarity revealed two major binding modes, referred to as “pose A” and “pose B” and shown in Fig. 4c. The primary difference between these two poses is that in pose A both BDQ and cGlu59 project away from center of the binding site, whereas in pose B the inhibitor packs more closely against the protein and cGlu59 is retracted into the site.

We next evaluated whether these poses are compatible with the interaction site revealed by our cryo-EM data. Superposition of pose A onto the cryo-EM structure of the Fo complex revealed evident steric clashes between the inhibitor and the backbone and side chains of TM3 of subunit-a; therefore, this pose was discarded. Pose B, however, appeared to be sterically compatible, and therefore we set out to examine it further through simulations analogous to those described above, now for the Fo complex, again in a POPC membrane (Fig. 4d). These simulations were based on the high-resolution cryo-EM structure of the yeast ATP synthase (PDB 6B2Z)14. Specifically, the construct considered is a subcomplex that includes the complete c-ring, subunit-a, and the transmembrane regions of subunit-b and subunit-8, i.e., the construct omits the F1 sector. To ensure that the relative arrangement of the four subunits included is preserved during the simulations, despite the absence F1 sector, a weak RMSD restraint was applied to the transmembrane Cα traces of the four subunits, collectively (residues 2–39, 44–73 for subunit-c; residues 28–47, 57–77, 86–104, 115–145, 155–200, 210–246 of subunit-a; residues 63–87 of subunit-b; and residue 11–39 of subunit-8). The force constant of this restraint (k = 4 kcal/mol/Å2) was chosen to impact minimally the conformational variability of the c-ring, relative to what was observed in the unrestrained simulations (average RMSD values of 1.01 ± 0.06 Å vs. 1.03 ± 0.12 Å, respectively). Otherwise, the Fo complex could tumble freely in the lipid membrane, and all protein side chains as well as the BDQ inhibitor itself were also free to reconfigure. For completeness, we again considered two alternative protonation states for BDQ and cGlu59. These systems were prepared and equilibrated using a protocol analogous to that described above for the simulations of BDQ bound to the c-ring. Trajectories of 160 ns were calculated to evaluate the stability and binding mode of BDQ for each protonation state. The outcome of these simulations is discussed in “Results” and summarized in Fig. 4e–g.

All simulations were carried out with NAMD248 using the CHARMM36 force field49,50 periodic boundary conditions and constant temperature (298 K) and semi-isotropic pressure (1 atm). A force field for bedaquiline (BDQ) was developed as described below. Long-range electrostatic interactions were calculated using PME, with a real-space cutoff of 12 Å. Van der Waals interactions were computed with a Lennard–Jones potential, cutoff at 12 Å with a smooth switching function taking effect at 10 Å. The protein–ligand complexes were embedded in pre-equilibrated hydrated bilayers using GRIFFIN51. Clustering analyses based on pairwise similarity used the method of Daura et al.52 with an RMSD threshold of 1 Å.

Electronic structure calculations and force field development for bedaquiline

Force field parameters for BDQ compatible with CHARMM36 were derived in two steps. First, electronic structure calculations were carried out with Gaussian 09 for bedaquiline and for a derivative where the bromine atom is replaced with chlorine (Supplementary Fig. 6). In both cases, the input structure was extracted from PDB 4V1F23. The molecular geometry of each of these two compounds, assumed to be in the protonated form, was optimized using Hartree–Fock theory, initially using the SDD basis set, and subsequently with the 6–311 G** basis set. Natural bonding orbital and natural population analysis were then used to calculate the charge distribution of the geometry-optimized structures, using mPW2PLYP theory and the 6–311 G** basis set. Second, a set of force field parameters was derived for the chlorine derivative of BDQ using the GAAMP server53 on the basis of the same input structure. Briefly, an initial parameter set was deduced from the CHARMM General Force Field, which includes entries for a broad set of chemical compounds. Parameters deemed to be suboptimal for BDQ were then optimized through ab initio calculations at the Hartree–Fock 6–31 G* level and MD simulations. From here, a force field for bedaquiline was derived by adjusting the atomic charges in the chloride derivative to reflect the differences between the two mPW2PLYP calculations mentioned above. Analogously, partial charges for the neutral form of BDQ were obtained by adjusting the amine group. A final MD simulation of 100 ns was then carried out in water to evaluate the resulting force field, resulting in one primary conformer, Supplementary Fig. 6.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

[ad_2]

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *