Real‐Time NMR Recording of Fermentation and Lipid Metabolism Processes in Live Microalgae Cells

Abstract Non‐invasive and real‐time recording of processes in living cells has been limited to detection of small cellular components such as soluble proteins and metabolites. Here we report a multiphase NMR approach using magic‐angle spinning NMR to synchronously follow microbial processes of fermentation, lipid metabolism and structural dynamic changes in live microalgae cells. Chlamydomonas reinhardtii green algae were highly concentrated, introducing dark fermentation and anoxia conditions. Single‐pulse NMR experiments were applied to obtain temperature‐dependent kinetic profiles of the formed fermentation products. Through dynamics‐based spectral editing NMR, simultaneous conversion of galactolipids into TAG and free fatty acids was observed and rapid loss of rigid lipid structures. This suggests that lipolysis under dark and anoxia conditions finally results in the breakdown of cell and organelle membranes, which could be beneficial for recovery of intracellular microbial useful products.


Cell culturing
Wild-type Chlamydomonas reinhardtii (Cr.) cells (strain cc124) and cell wall-deficient mutant (cw15) were grown mixotrophically in tris-acetate-phosphate (TAP) medium at pH 7 in a home-built set up, under continuous illumination with white LEDs (~50 μmol/m 2 s) and constant temperature of 25 °C. For 13 C isotope-label incorporation for NMR experiments, the acetic acid in the TAP medium was replaced by 13 C acetic acid (Cambridge Isotopes, Massachusetts, USA).

Sample preparation for solid-state NMR
The Cr. cells were harvested at exponential phase using centrifugation and washed with high-salt medium [1] to avoid the presence of 13 C acetic acid during NMR experiments. The cells were then spun down into a 4 mm NMR rotor by mild centrifugation (~1000 x g).

Solution NMR experiments
After MAS ssNMR experiments, the cells from NMR rotor were washed two times with phosphate-buffered saline (PBS) buffer to collect the metabolites exerted from the cells. The washed-out cell supernatant were mixed with 10% D2O (final concentration) to enable field locking.
Solution NMR experiments were performed on a Bruker Avance-I 500 MHz spectrometer using 5 mm BBFO Z-gradient high resolution probe head.

Cell viability
To assess cell viability after NMR experiments, cells were resuspended in PBS to an OD750 of 0.9 (~ 1.8 × 10 7 cells/ml) and then 30 times diluted in PBS buffer containing 5 mM ethylenediaminetetraacetic acid (EDTA) to avoid clumping of the cells during flow cytometric experiments. Cell suspensions (~ 6 × 10 5 cells/ml) were subsequently incubated with fluorescein diacetate (FDA) at a final concentration of 2.5 μM for 20 min at room temperature in dark. Two samples were used as negative controls, heat-killed cells (95˚C for 10 min to inactivate esterase activity) and cells from fresh culture without FDA. Fresh cells as a positive control were first washed and resuspend in HSM medium to an OD750 of 0.9 and incubated in dark before staining with FDA in PBS buffer.

Flow cytometry experiments
The flow cytometry experiments were performed on a Guava® easyCyte 12 HT Sampling Flow Cytometer (Luminex Corporation) and  Figure 1).

Lipid isolation
For total lipid extraction, the Cr. cells incubated under high-cell cell density (OD750 ≈ 120), dark and anoxia condition at 23 ˚C for 0, 2.5 and 24 hours were freeze-dried and stored at -80 ˚C until use.

MGDG and DGDG
For assignment of MGDG or DGDG C1' and DGDG C1'' 13-carbon NMR signals of the galactosyl headgroups, we analyzed 2D spectra of Cr. extracted thylakoid membranes, in which the number of carbohydrate components are less redundant than in whole cells. Thylakoids were extracted from the algae according to [6]. Assignment was done using J-coupling based 2D 13 C-13 C INEPT-TOBSY and 1 H-13 C HetCor spectra (Fig. S10B, C) that were recorded at -3 o C (TOBSY mixing time of 6 ms, 14 kHz MAS). A comparison with a DP MAS spectrum of cell-wall deficient Cr. cells (Fig. S11A, blue) further confirms that the galactosyl signals assigned to galactolipids in the spectrum of Cr. cells are not from glycoproteins in the cell wall. In time, the signals of MGDG and DGDG C1' and C1'' vanish and new peaks arise at ~92 and 97 ppm that are attributed to glucose products (see Fig. S11A and B).

Lipid carbonyl
To identify the nature of the peak at 174.2 ppm, we considered the DP and CP MAS spectrum of thylakoid membranes, collected at 17 o C, and compared this to the DP MAS spectrum of Cr. cells collected at 23 o C (Fig. S12). Thylakoid membranes are densely packed with membrane-spanning proteins that make up ~75% of the total surface area [7]. The major contribution to the thylakoid spectrum in the region 170-180 ppm comes from the carbonyl signals of the protein backbones that form a broad band centered at 175 ppm [8] and are visible with DP and CP [6] as demonstrated in Fig. S12, red spectrum. Signals of the protein carboxyl sidechains (glutamate and aspartate residues) typically fall in the range 177-182 ppm depending on their protonation states [8] and are seen as peaks super-imposed on the protein CO band in the thylakoid DP spectrum. Lipid carbonyls have 13 C NMR chemical shifts between 172-175 ppm [9] and are additionally super-imposed. No other thylakoid components are expected to have resonance signals in this region, inferring that the pronounced peak centered at 174.2 ppm that is visible both in spectra of whole cells and of thylakoid membranes, represents the superimposed peak of accumulated lipid CO signals. This is further supported by the fact that the carbonyl signal of the most abundant lipids, MGDG and DGDG, are found here (Fig. S9). Over time, the DP spectra of whole cells show a decrease of the lipid peak at 174.2 ppm and rise of a band at 178.5 ppm, which can be attributed to carboxylic acid carbons.
We attribute the rising carboxyl signal to the accumulation of FFA, supported by the HPTLC analysis in Fig. S8. Note that the intensity of protein CO band, visible in the CP-based spectra, does not change over time (Fig. S13-S16).

TAG
According to literature, the fatty acid carbonyl 13 C chemical shift signals of DAG fall in the region 173.1-173.4 ppm and the carbonyl signals of TAG fall in the region 172.4-172.8 ppm [9]. The DP difference spectra presented in Fig. S11B show the emergence of two small signals at ~172.9 and ~172.4 ppm after 24 h. The thylakoid spectrum does not show any structure there (Fig. S12), i.e. the signals are not from the most abundant lipid types, and they do not overlap with signals of cell-wall components, according to the comparison of wildtype and cell-wall deficient cells (Fig. S11A, blue and red). Supported by the HPTLC analysis shown in Fig.S8, we tentatively assign the emerging small bands to formation of TAG.     B Figure S12. DP 13 C ssNMR spectra of Cr. cells after 1 hr (green) and after 24 hrs (purple) and DP (blue) and CP (red) 13 C ssNMR spectra of Cr. thylakoids. The accumulated signals of lipid carbonyls in whole cells and in thylakoids and the emerging peak of FFA carboxyl atoms are indicated.     Figure S22. Reproducibility of CP and INEPT NMR signal intensities at 30 ppm for two series of experiments at 23 ˚C and 10 ˚C as a function of time.

Supplementary figures
For visual comparison of the kinetic profiles, the intensities of spectra collected at each series are normalized at the first time point.