Real-Time Pyruvate Chemical Conversion Monitoring Enabled by PHIP

In recent years, parahydrogen-induced polarization side arm hydrogenation (PHIP-SAH) has been applied to hyperpolarize [1-13C]pyruvate and map its metabolic conversion to [1-13C]lactate in cancer cells. Developing on our recent MINERVA pulse sequence protocol, in which we have achieved 27% [1-13C]pyruvate carbon polarization, we demonstrate the hyperpolarization of [1,2-13C]pyruvate (∼7% polarization on each 13C spin) via PHIP-SAH. By altering a single parameter in the pulse sequence, MINERVA enables the signal enhancement of C1 and/or C2 in [1,2-13C]pyruvate with the opposite phase, which allows for the simultaneous monitoring of different chemical reactions with enhanced spectral contrast or for the same reaction via different carbon sites. We first demonstrate the ability to monitor the same enzymatic pyruvate to lactate conversion at 7T in an aqueous solution, in vitro, and in-cell (HeLa cells) via different carbon sites. In a second set of experiments, we use the C1 and C2 carbon positions as spectral probes for simultaneous chemical reactions: the production of acetate, carbon dioxide, bicarbonate, and carbonate by reacting [1,2-13C]pyruvate with H2O2 at a high temperature (55 °C). Importantly, we detect and characterize the intermediate 2-hydroperoxy-2-hydroxypropanoate in real time and at high temperature.

1. Synthesis of 1,2-13 C pyruvate precursor Vinyl compounds provide an excellent basis to produce signal-enhanced metabolites, for instance pyruvate, via PHIP effect but their synthesis via transesterification were mostly obtained in low yields. Our recent publication (Ding et. al. 2022) reported a procedure to synthesize 1,2-13 C labeled vinyl pyruvate in high yields (80%) as shown in Figure-SI 1. Here briefly, deuteration of the starting material 2-oxopropionic-1,2-13 C-acid (1) was performed in presence of D2SO4/D2O to yield 2, which was then treated with trimethyl orthoformate to obtain dimethyl ketal protected derivative (3) in quantitative yields. Installation of the photo-labile protecting group was performed by coupling of diol (4) with dimethyl ketal protected derivative (3) in toluene to yield 5 with moderate yield. Pd-catalyzed transesterification of 5 with deuterated vinyl acetate (6) was performed to obtain 7 with 81% yield (based on the recovery of 5). Deprotection of 7 containing photo-labile protecting group was carried in presence of LED lamp (365 nm) to obtain desired compound (8) with 12% yield.

LDH solution and Cell preparations
To probe the pyruvate-to-lactate conversion in vitro, L-lactic dehydrogenase (LDH) from rabbit muscle in ammonium sulfate suspension (sigma, USA) was diluted to 250 or 500 units/ml against protonated buffer of 100 mM HEPES (Sigma, UK), 20 mM NADH (Biomedicals, Japan) and 120 mM NaCl (Roth). Around 200 μL LDH solution (50 or 100 units) was used for one PHIP experiment.
For probing intracellular pyruvate-to-lactate conversion, Hela Kyoto wild-type cells was cultured in the Dulbecco's Modified Eagle Medium supplemented with 4.5 g/L D-Glucose, 2 mM glutamine, 1 mM Sodium Pyruvate, 10% (v/v) heat-inactivated fetal bovine serum, 100 U/mL Penicillin and Streptomycin (all the components were purchased from Thermo-Fisher scientific). The cells were grown at 37°C in a humidified 5% CO2 atmosphere to around 70-80% confluency before collection. Before the PHIP experiment, Hela cells were detached from the cell culture flasks using trypsin/EDTA (0.05%/0.02%), centrifuged under 130 × g for 3 min at room temperature, washed once with PBS and counted on a haemocytometer. Around 25 million cells were collected and re-suspended into 120 μL DMEM medium containing 50 mM NADH for one PHIP experiment.
3. Production of rapidly signal-enhanced 1,2-13 C pyruvate A purity of 99% pH2 was generated using a He-cooled pH2-generator (generator T=20 K), equipped with a Cryocooler system (Sumimoto HC-4A helium compressor, Sumimoto Cold Head CH-204 with ph2 reaction chamber by ColdEdge Technologies), temperature controller (Lake Shore Cryotronics, Inc.) and home built valve and tubing system. 5 mM of hydrogenation catalyst ([1,4-bis(diphenylphosphino) butane] (1,5cyclooctadiene)rhodium(I) tetrafluoroborate) was dissolved in acetone-d6. For each sample, the vinyl precursor was added to 0.1 ml of the catalyst solution to obtain a concentration of 5 mM. Samples were placed in a 5mm NMR-tube, degassed by bubbling N2 gas for 1 minute and hydrogenated at 7 bar for 20 s by bubbling using a home-built setup ( Fig. SI 2) inside a probehead of 7 T Bruker spectrometer (Avance III) heated to 328 K. By using the pulse sequence MINERVA 1 ( Figure SI (3), (4) and (5) represents the Rhodium catalyst, which tends to precipitate after acetone removal. Pictures of the different stages of the protocol at the bottom.

(5) Injection of 100-200 uL LDH Enzyme or HeLa cells through a different injection line (extreme right). The orange part at the bottom of the NMR tube in
When polarization transfer was completed, 0.1 ml of 50 mM Na2CO3 in D2O (at room temperature) was added into the NMR tube via a cannula to cleave off the sidearm (Figure-SI 3), followed by 10-12 s of acetone evaporation with a vacuum pump. Then, 0.1 ml of isotonic buffer (0.274 M NaCl, 0.0054 M KCl, 0.0236 mM Na/PO4) was added to adjust the pH of the mixture to around 7.4 followed by addition of around 25 millions Hela cells via a cannula. 13 C spectra were recorded with a flip angle of 45° and a repetition delay of 2 s.

Simulations
The J coupling parameters have been determined via the 1D carbon 13 spectra and proton NMR spectra. The parameters have been used to simulate the effect of the pulse sequence as shown in the following figures-SI 4-8 using Spindynamica 2 package for Mathematica.

4.4.
Transformation amplitudes The following code shows the transformation from the initial state upon hydrogenation to the final target state.

Kinetic analysis
For in-vitro enzymatic experiments, the pyruvate-to-lactate reaction is predominantly shifted towards lactate and in accordance with previously reported studies, we assume an unidirectional conversion rate: A set of two differential equations models the variation of the pyruvate and lactate signals over time: Following the discretized version in reference 3 equations SI.2 and SI.3 become: The index k represent the k th slice in the pseudo 2D experiment consisting in the repeated application of  13C flip angle pulses repeated every TR. The effective decay rate Reff accounts for signal decay due to Boltzmann thermalization, repetitive excitation with  flip angle pulse (20 degree for in-vitro experiments). The pyruvate and lactate integral signal then results in a system of linear equations for kPL and Reff that can be solved by a pseudo matrix inversion:  with β and TR the flip angle and repetition time used is prepended to the solutions in order to account for the experimental conditions used. We have used the Manipulate routine to match the fitting with the experimental data. The code is provided below.
Figure-SI 13: The data points are the integral of the corresponding species at C1 and C2 positions. Product-C1 is the C1 is the integral sum of the signals from HCO3 -, CO3 2and CO2. The product-C2 is the integral value for the acetate resonance. The C1 data points have been fitted with k1 and k2 as free parameters assuming a T1 for Product-C1 of 25s. The C2 data points have been fitted with k1=0.18 s -1 and k2=0.232 s -1 In Figure-SI 13 we report the experimental data points and the corresponding fitting curves for pyruvate, the intermediate and the product at C1 and C2 position according to the chemical reaction in Fig. 5a. The product-C1 is the integral sum of the signals from HCO3 -, CO3 2and CO2. The product-C2 is the integral value for the acetate resonance. Following the kinetic model in Fig. SI-12, assuming a T1=50 s for pyruvate, T1=3.0 s for I and initial conditions I(0)=Q1(0)=0, the fitted values results k1 and k2 values are: 6.1. Pyruvate decarboxylation at different pH The same experiment as in Fig. 5 has been conducted at basic (pH=9) and acid (pH=2)  HCO3 -(it only appears from the spectrum some transients later) follows the 13 CO2 formation (grey box in Figure-SI 14).

Figure-SI 14: H2O2-induced pyruvate decarboxylation reaction monitored at pH=9 and pH=2
We note that the ability to detect via NMR HCO3and 13 CO2 signals enables the direct measurement of the bulk pH in solution via the Henderson-Hasselbalch equation: ⁄ ) (SI.9) We've applied equation (SI.9) to calculate the pH variation during the course of the decarboxylation reaction at around pH 7 by applying the log10 function to the ratio of the integrated signals from HCO3and 13 CO2 at every slice. The HCO3and CO2 signal intensities in Fig. 6 have been normalized. Their similar trend (* and × curves in blue in Figure

Estimation of Lactate hyperpolarization and acetone level
In Figure 4b we report the thermal and hyperpolarized carbon spectra for the experiment with HeLa cells. By taking the integral values of the acetone and lactate peaks in the thermal experiments we can estimate the final substrate and acetone-d6 content.
The enhancement and hyperpolarization level is estimated by the following equations: In the equations SI 10-12, I refers to signal area, rg receiver gain, ns number of scans, h plank constant, K Boltzmann constant, T temperature, B0 magnetic field, ɤ gyromagnetic ratio, flip angle refers to the applied angle before FID detection.
To estimate the level of remaining acetone after cell experiment, the 13 C thermal spectrum of the cell sample was recorded. As an external reference, 350 uL of [2-13 C] sodium acetate in D2O at 298K yields an integral Iref=157000 mol -1 *scan -1 *rg -1 . By integrating the region at 215 ppm we estimate the final acetone concentration and volume.