Protein-protein interaction analysis in crude bacterial lysates using combinational method of 19F site-specific incorporation and 19F NMR

Protein-protein interactions (PPI) are essential for a variety of cellular functions. Many PPI analyses were conducted in vitro, using purified proteins. In this report, the unnatural amino acid tfmF was site-specifically incorporated into several different sites of two Phox-Bem1 (PB1) domains from two mitogen activated protein kinases (MEKK3 and/ or MEK5) in the E. coli cells. Solution NMR 19 F chemical shift and side chain relaxation analysis demonstrated that MEKK3-PB1-I57, MEKK3-PB1-F77, and MEK5-PB1-I70 sites were located in the interaction interface of the MEKK3/ MEK5 complex, which was consistent with the crystal structure of MEKK3-PB1/MEK5-PB1 complex. Furthermore, crude lysates from E. coli cells with co-expressed tfmF incorporated MEKK3-PB1 and MEK5-PB1 were applied for 19 F NMR analysis. The successful implementation of in situ PPI analysis using the combinational method of site-specific tfmF incorporation and 19 F NMR demonstrated that this method could be a valuable general method for conformation and function studies of soluble multi-domain proteins or protein complexes in bacterial crude lysate, without procedures of protein purification. Protein-protein interactions (PPI) play essential roles in cellular functions, such as DNA transcription, signal trans-duction, or cytoskeleton formation. Currently, a variety of techniques, including co-immunoprecipitation, isothermal titration calorimetry, and surface plasma resonance are frequently applied for PPI studies (Syafrizayanti et al., 2014). However, these methods can only provide the overall interaction pattern or internal motion of purified protein complexes , and have many limitations such as low specificity, high background or false positives (Syafrizayanti et al., 2014). Structure determination methods (such as X-ray crystallography and electron cryo-microscopy) can precisely illustrate protein interaction interface, while these structural methods require high concentration of purified proteins. Recently, it has been reported that the cytoplasmic environment might have profound effects in regulating protein–protein and/or protein–ligand interactions that were hardly observed in vitro (Smith et al., 2014). The crucial difference between in vivo and in vitro conditions lies in the high environmental concentration of diverse macro-molecules, which is approximately 200 mg/mL in the eukaryotic cytoplasm and more than 400 mg/mL in prokaryotes (Mika and Poolman, 2011). Traditional in vitro biochemical studies of proteins were conducted in dilute solution with low macromolecular concentration (∼10 mg/ mL), which might not reveal protein-protein interaction or its mechanism in high fidelity. Solution nuclear magnetic resonance (NMR) is powerful for PPI analysis and recently it has been applied to analyze protein conformation changes in living cells (Hansel et al., 2014). However, the traditional …


Supplemental Materials Plasmid construction and site-specific mutagenesis
The amplified MEKK3-PB1 and MEK5-PB1 coding sequence were inserted into the pBAD28 vector (Invitrogen, Thermo Fisher Scientific, USA) between the Nde I and Xho I restriction endonuclease sites. Two residue sites (I57, F77) of MEKK3 and two residue sites (F41, I70) of MEK5 were selected, and codons corresponding to these sites were mutated to the amber stop codon TAG using a standard PCR-based mutagenesis method. Mutagenesis was confirmed by DNA sequencing. The mutated pBAD28 plasmid and another essential plasmid, pDule-tfmf (containing the coding sequence for tRNA CUA and tfmF-specific aminoacyl-tRNA synthetase) (Hammill et al., 2007) were co-transformed in E. coli host Top10 cells for further protein expression.
For the co-expression of MEKK3-PB1 TAG and MEK5-PB1 TAG mutants, another plasmid named pETDuet-1 (Invitrogen, Thermo Fisher Scientific, USA) was used, which contained two multiple cloning sites and a ribosomal binding site. The MEKK3-PB1 DNA fragment was inserted into the pETDuet-1 plasmid using BamH I and Not I restriction enzymes using the standard PCR method. The DNA fragment of MEK5-PB1 was then cloned between the Nde I and Xho I restriction sites. The corresponding TAG mutants of two DNA fragments were performed using the two-step PCR scheme. The constructed plasmid was co-transformed with pDule-tfmF into E. coli host Top10 cells.

Site-specific tfmF incorporation and protein purification
The transformed bacteria were incubated in LB medium containing 1 mM tfmF at 37C (with shaking at 225 rpm) with 100 μg/mL ampicillin and 15 μg/mL tetracycline.
The transformed bacteria were incubated in LB medium overnight, and then transferred into fresh 2x YT medium containing 1 mM tfmF at 37°C in presence of 100 μg/mL ampicillin and 15μg/mL tetracycline. Expression of the target protein was induced using 0.2% arabinose when the OD 600 reached 1.0. After 4 h, bacteria were harvested by centrifugation at 3700g for 20 min at 4°C (Beckman Coulter X-15R).
The pellet was collected and suspended in 40 mL lysis buffer (20 mM Tris, 500 mM NaCl, pH 8.0), and then probe-sonicated (VC500, Sonics and Materials, Danbury, CT) on ice for a total of 10 min (power level, 30%; 3.0 s pulse on and 5.0 s pulse off). The cell lysate was centrifuged at 23,000 g for 20 min at 4°C (Hitachi Himac Centrifuge, CR21GII), after which the supernatant was collected and mixed with preprocessed 3 mL Ni 2+ -NTA resin (QIAgen, Valencia, CA). The mixture was mixed at 4C for 40 min, followed by packing onto a gravity-flow column (BIO-ARD, Hercules, CA). The impurities were washed out by a series of imidazole gradients, and the target protein was then eluted using elution buffer (50 mM Tris, 200 mM NaCl, 250 mM imidazole, pH 8.0).
The concentration of purified target protein was determined using OD 280 measurements, and analyzed by standard SDS-PAGE (15%, w/v). The purified target protein was concentrated to a final volume of 450 μL using an Amicon Ultra-15 device (3000 MWCO, Millipore). Then, 50 μL D 2 O was added to the sample to a 10% (v/v) final concentration before solution NMR analysis.

F chemical shift and side chain relaxation analysis
All one-dimension 19 F NMR spectral measurements were performed at 293 K on an Agilent 500 MHz spectrometer equipped with a HFT probe, and the observation channel was tuned to 19 F (470.2 MHz) with 1024 free induction decay (FID) accumulations in every 4 s recycling delay. One-dimensional 19 F spectra were acquired with a one pulse program with a 90° pulse width of 12.7 μs and power at 57 w. The spectra width (SW) was 60 ppm and the offset was set at -62 ppm. 19 F chemical shifts were referenced to an external standard tfmF (-62.38 ppm), and the chemical shift of tetramethylsilane(TMS) was calibrated to 0 ppm. The data were processed with an exponential window function (LB =15 Hz) and plotted using ACD/NMR Processor Academic Edition software (ACD/Labs). The spectra were collected at 293K.
A pulsed field gradient-based longitudinal eddy delay pulse sequence was applied to measure the relaxation time of proteins. The experiments were conducted at 298 K, in a 500 MHz Agilent spectrometer. Side chain 19 F T 1 relaxation data were collected with eight delay times (50, 100, 200, 400, 600, 800, 1000, 1200, 1500, and 2000 ms) using a standard 1D inverse recovery pulse sequence. The magnetization Mz was inverted by a 180°pulse. Waiting a delay τ, the amount of z-magnetization was measured using a 90°pulse. 64 scans were accumulated for one-dimensional 19 F chemical shift analysis and 256 scans were accumulated for T 1 data acquisition.
According to Bloch function, the dependence of the z-magnetization on τ is given by: M 0 is the equilibrium value of the z-magnetization.
Resonance intensities in relaxation experiments were measured and fit to an exponential function.
The Carr-Purcell-Meiboom-Gill sequence (CPMG) experiment allows measure transverse or spin-spin T 2 relaxation times of any nucleus. Side-chain 19 F T 2 transverse relaxation data were collected with ten delay durations (8,16,32,48,64,80,96,112,144, and 160 ms) using a standard 1D Carr-Purcel-Meiboom-Gill (CPMG) pulse sequence. The intensities of the series of peaks were regressed for transverse relaxation values for tfmF labeled MEKK3-PB1 or MEK5-PB1. Resonance intensities in CPMG experiments were fit to an exponential function. The peak intensity values were analyzed and plotted using OriginPro 8.0.
In the measurement of the T 2 relaxation times, the magnetic field inhomogeneities (T 2 inh ) must be considered: The experimental half-height line-width (d) of a given resonance is directly related to T 2 * by the equation.
d=1/(π*T 2 * ) To estimate the T 2 * value of crude lysate sample the line width was read using ACD/Lab software.

Supplemental References
Hammill