Cloning, expression, and purification of IF1
The human ATP5IF1 gene encoding residues G26-D106 (1–81 in the numbering scheme of mature IF1, with codons optimized for expression in Escherichia coli), preceded by a 6x-histidine tag and a TEV protease cleavage site, was cloned into a pET-28a vector (Blue Heron Biotechnology, Bothell, WA). The construct was verified by whole-plasmid sequencing using Oxford Nanopore Technology (Plasmidsaurus, Louisville, KY). The plasmid was transformed into E. coli C41 (DE3) cells, and 500 mL cultures were grown at 37°C in M9 minimal media (47.7 mM Na2HPO4, 22 mM KH2PO4, 8.6 mM NaCl, 2 mM MgSO4, 5 µM FeCl3, 5 µM CaCl2) supplemented with 4 g/L 1-13C-D-glucose (4 ml/L glycerol for non-13C-labeled samples) and/or 1 g/L 15NH4Cl. Kanamycin (25mg/L) was included in the growth medium to maintain selection of the pET expression vector. Expression was induced at OD600 = 0.6 with 0.64 mM isopropyl thio-β-D-galactoside (IPTG), and cells were grown for an additional 20 hours at 25°C. The cells were harvested by sedimentation at 3,470 g using a Sorvall Lynx 6000 fixed-angle rotor (Thermo Fisher Scientific, Waltham, MA), and the pellets were flash frozen and stored at -80°C.
All subsequent purification steps were performed at 4°C. Frozen pellets from a 500 mL culture were resuspended in 20 mL of lysis buffer: 25 mM potassium phosphate buffer pH 6.4, 20 mM Imidazole, and 1x cOmplete™ EDTA-free protease inhibitor cocktail (Roche, Basel Switzerland). The pH after imidazole addition was 7.4. Cells were lysed by sonication on a Q700 Sonicator (Qsonica, Newtown, CT) with instrument parameters of 40 s on/20 s off cycles at 70% amplitude, and 3 min on-time total. After sonication, the cell extract was clarified by sedimentation (30,000 g, 45 min). NaCl was added to the supernatant to a concentration of 0.2 M, and the mixture was incubated with 2-ml pre-equilibrated Ni-NTA agarose (QIAGEN GmbH, Hilden, Germany).
All subsequent purification buffers were adjusted to pH 6.4, which significantly reduced IF1 aggregation. The Ni-NTA resin was washed with 7 column volumes of 30 mM imidazole followed by 7 column volumes of 40 mM imidazole in buffer A (25 mM potassium phosphate buffer, pH 6.4, 0.2 M NaCl). IF1 was eluted with 5 column volumes of 500 mM imidazole in buffer A. The eluate was buffer-exchanged using an Amicon® Ultra centrifugal filter, 3 kDa MWCO (MilliporeSigma, Burlington, MA) into buffer A. 1 mM DTT and recombinant His6-TEVsh protease (1:15 (w/w) TEV:IF1) were added to the retentate. The TEVsh protease was prepared in-house as described in the literature (van den Berg, Lofdahl et al. 2006). The TEVsh plasmid, Addgene plasmid # 125194 (Watertown, MA), was a gift from Helena Berglund. TEVsh protease cleavage was done overnight at 4°C with gentle shaking. After sedimentation to remove aggregates (10 min at 10,000 g), the supernatant was incubated with Ni-NTA to capture his6-TEV protease. The flowthrough containing tag-free IF1 was buffer exchanged using Amicon® centrifugal filtration (3 kDa MWCO) into buffer A. Protein concentrations were determined using the Qubit™ Protein Assay Kit on a Qubit™ 4 Fluorometer (Thermo Fisher Scientific, Waltham, MA) following the manufacturer’s instructions. Purified IF1 samples were stored at 4°C and were used for NMR within 24h of purification.
NMR spectroscopy
NMR experiments were performed at 37°C on a Bruker AVANCE NEO 600 MHz spectrometer equipped with a cryogenic probe. Samples contained 0.1–0.4 mM IF1 monomer in 25 mM potassium phosphate buffer and 200 mM sodium chloride. In the first phases of the project, we optimized the sample pH and temperature for NMR experiments. The best NMR spectra were obtained over a narrow range of sample conditions, between about pH 4.7 and 5.6, and temperatures between 37 and 41 degrees. More basic pH conditions led to loss of peaks due to oligomerization into tetramers and higher-order aggregates, while more acidic conditions unfolded the protein. Similarly, temperatures below 37 oC shifted the equilibrium to oligomers, whereas higher temperatures caused losses of amide protons due to hydrogen exchange and eventually protein unfolding. Thus, all NMR experiments were performed at pH 5.3–5.6 and 37°C.
Backbone resonance assignments for the IF1 dimer were determined from u-[15N,13C]-IF1 samples (110 µM, pH 5.3, 10% D2O) using 3D HNCACB, HNCA, HN(CO)CA, HNCO, and HN(CA)CO spectra. The sample was lyophilized and resuspended in D2O (pH 5.6) for 1H-13C experiments: 2D 1H-13C HSQC and 3D HCCH-TOCSY, CCH-TOCSY, 1H-13C NOESY (120 ms mixing time). A separate u-[15N]-IF1 sample (445 µM, pH 5.3, 5% D2O) was used for acquiring 1H-15N TOCSY HSQC (70 ms mixing time) and 1H-15N NOESY-HSQC (120 ms mixing time) spectra. This sample was also used for the “fingerprint” 1H-15N SOFAST-HMQC in Fig. 1. Protein integrity between NMR spectra was checked with serial 1H-15N SOFAST-HMQC spectra. 1H shifts were referenced directly to internal DSS (2,2-dimethyl-2-silapentane-5-sulfonate), whereas 13C and 15N were referenced indirectly (Wishart et al. 1995).
Assignments and data deposition
An annotated 1H-15N SOFAST-HMQC spectrum for dimeric IF1 is shown in Fig. 1. Backbone amide (HN-N) assignments are 98% complete, and assignment coverage is also 98% complete for all backbone atoms. The only residues missing HN-N assignments are Gly1 and Ser2, which experience fast amide proton exchange with solvent. Assignment coverage statistics for all NMR-active atoms are 1H (86%), 13C (77%), 15N (64%). The somewhat low percentages for 13C and 15N are due to the presence of a large fraction of Lys, Arg, and Glu residues in the protein sequence.
Our assigned chemical shifts and preliminary analysis of 1H-15N NOESY-HSQC and 1H-13C NOESY-HSQC spectra, were used to calculate a consensus secondary structure for IF1 (Fig. 2) using the DANGLE subroutine (Cheung et al. 2010) of the CcpNmr Analysis v2.5.2 suite of programs (Cheung et al. 2010). The IF1 dimer consists of an a-helical region between about residues R39-D81, although a few weak NOEs characteristic of a-helical structure extend to about E30. Upstream of E30 towards the N-terminus, the NMR data do not support a-helical structure giving intense NMR signals characteristic of long T2 values and NOEs to solvent, both typical of unstructured conformations. The secondary structure pattern observed for human IF1 in the present study is consistent with the previously described antiparallel coiled-coil structure of bovine IF1 (residues 44–84), and the observation that residues 10–48 corresponding to the inhibitory segment are unstructured in solution (Gordon-Smith et al. 2001).