Biological autoluminescence as a noninvasive monitoring tool for chemical and physical modulation of oxidation in yeast cell culture

Normal or excessive oxidative metabolism in organisms is essential in physiological and pathophysiological processes, respectively. Therefore, monitoring of biological oxidative processes induced by the chemical or physical stimuli is nowadays of extreme importance due to the environment overloaded with various physicochemical factors. Current techniques typically require the addition of chemical labels or light illumination, which perturb the samples to be analyzed. Moreover, the current techniques are very demanding in terms of sample preparation and equipment. To alleviate these limitations, we propose a label-free monitoring tool of oxidation based on biological autoluminescence (BAL). We demonstrate this tool on Saccharomyces cerevisiae cell culture. We showed that BAL can be used to monitor chemical perturbation of yeast due to Fenton reagents initiated oxidation—the BAL intensity changes with hydrogen peroxide concentration in a dose-dependent manner. Furthermore, we also showed that BAL reflects the effects of low-frequency magnetic field on the yeast cell culture, where we observed a disturbance of the BAL kinetics in the exposed vs. control case. Our results contribute to the development of novel techniques for label-free, real-time, noninvasive monitoring of oxidative processes and approaches for their modulation.

. Schematic illustration on the mechanism of BAL under the chemical and physical modulation of yeast cells. The hydroxyl radical (either generated during endogenous metabolism or by addition of external precursors) causes oxidation of biomolecules (RH) and produces secondary radicals (alkyl radical R · , peroxyl radical ROO · , alkoxyl radical RO · produced via nonradical biomolecular hydroperoxide ROOH). Those radicals lead to the formation of unstable intermediates dioxetane (ROOR) and tetraoxide (ROOOOR). These high-energy intermediates can be decomposed to electron-excited species, such as triplet excited carbonyl ( 3 R = O * ) [44][45][46][47] or singlet oxygen ( 1 O 2 ) 48,49 . The transition of these species to the ground state is accompanied by the photon emission (wavy line) manifested as biological autoluminescence. BAL is also produced during cell culture cultivation when oxidative metabolic processes are present. BAL reaction scheme is adopted from 2 . In the case of chemically induced oxidation, hydroxyl radical is formed as a product of Fenton reaction and abstracts hydrogen from biomolecule RH to initiate the cascade of further reactions. The physical stimulus can affect the biological system at multiple points of the scheme: potentially affecting the rate of the endogenous oxidative metabolism or rate of the recombination of radical species.

Results and discussion
BAL for monitoring chemically induced oxidative stimulus. We show the potential use of BAL for monitoring chemically induced oxidation of yeast cells Saccharomyces cerevisiae (Fig. 3). The oxidation is initiated by highly reactive hydroxyl radical, which is formed in the solution from hydrogen peroxide via Fenton reaction. We aim to focus on the BAL that arises from the oxidation of yeast cell material, regardless if the cells are living or not. We present the results of BAL measurements from yeast cell samples with various concentrations of hydrogen peroxide, which is expected to produce various amounts of hydroxyl radical molecules in solution. Figure 3A shows the time dependence of BAL from yeast cell culture with the presence of ferrous ions after adding of three different concentrations of hydrogen peroxide (H 2 O 2 ) . The unit counts per second (counts/s) represents the number of photons detected by a photomultiplier module in 1 s. The length of each measurement is 300 s.
The initial maximum of BAL kinetics (Fig. 3A) is likely caused by fast oxidation of yeast cells by hydroxyl radical created immediately after hydrogen peroxide application. This response of BAL can be conceptually explained by the Fenton reaction kinetics models in biomolecular solutions 50,51 . The decreasing BAL kinetics in all the concentrations probably reflects the progressive consumption of secondary radicals produced in oxidative reactions leading to non-reactive products formation. The higher the hydrogen peroxide concentration, the higher BAL intensity throughout the measurements (Fig. 3A) is observed.
We additionally induce Fenton reaction in purified water (Fig. 3B), in order to detect BAL response to this chemical process in the water without cells. The similar trend of BAL signals as in the case of cells is observed, but the signal levels are lower, as can be seen in subtracted signals (Fig. 3C). Ivanova et al. 51 propose mechanisms, Two identical coils are used, one of them, labeled exposure coil, is connected to the signal generator to generate magnetic field. The second one, labeled control coil, is not connected to the generator and is used to ensure the same air flowing conditions in the chamber with the control sample. The placement of exposure coil with respect to the chambers was randomized. The samples are mechanically stirred to avoid sedimentation of cells. Temperature sensors (T1-T4) for recording the temperature during all experiments are placed at the same locations in both chambers in order to control the potential heating effect by exposure coil: T1 is attached to the Erlenmeyer flask, T2 is free-hanging in the coil cavity, T3 is attached to the outer side of the coil and T4 is attached to the stand of the stirrer to monitor the overall temperature in the chamber. (D) Exposure coil for magnetic field generation (photo). (E) Magnetic flux density distribution within the exposure area (volume of the cell culture) exhibits 95% homogeneity. www.nature.com/scientificreports/ where Fenton reagents in water generate singlet oxygen via a series of reactions. Singlet oxygen then emits light either via monomol or dimol emission manifesting as BAL.
The significantly higher total BAL counts ( Fig. 3D) calculated as the area under the curves in Fig. 3A, compared to the control case ( Fig. 3B) are obtained. With the increasing concentration of hydrogen peroxide applied to the solution, the increasing amount of hydroxyl radical molecules in Fenton reaction is produced. The total amount of electron-excited species responsible for BAL generation is then very likely increased and therefore higher BAL intensity is detected.
To summarize, we propose BAL as a useful and innovative non-invasive method for monitoring of cell culture oxidation. However, in order to develop a reliable assessing method, further studies are necessary to conduct.

BAL signal relationship to physiostructural damage of cells.
To deeper analyze the increase of BAL signal with increasing H 2 O 2 concentration and its relationship to oxidation effects on cell culture, we performed a size exclusion assay. In this assay, we measure the cell concentration and size of untreated cells. The concentration of treated cells was measured, but taking into account that only those cells that fall in the diameter range of untreated cells are considered. The concentration of cells was measured in the range of untreated cells 3-9 µm , which is the typical range for the diameter of a single yeast cell. The results are evaluated as the ratios of cell concentrations in treated and untreated samples. It can be seen that the ratios are decreasing with increasing H 2 O 2 concentration (Fig. 4A), although not significantly. We assume that the higher the concentration of hydrogen peroxide applied to the solution, the more cells are excluded, and therefore the lower concentration in the selected diameter range (3-9 µm ) was measured. This is probably due to the following reasons: either the cells swelled, clustered, or fragmented. We also performed a cell viability test after oxidation using trypan blue stain-  To analyze cell activity after oxidation, we additionally monitored cell growth after oxidation (Fig. 4B) in multiwell plate reader (Tecan Spark). The results indicate the latency in onset of the exponential phase of cell growth curve as well as in reaching saturation (stationary phase) when oxidation is induced, compared to no oxidation treatment (0 mM H 2 O 2 ). The latency tends to increase with the increasing H 2 O 2 concentration with 2.5 and 25 mM conditions giving results similar to each other. The seemingly reverse mutual order of latency might be explained by the fact, that the initial concentrations of cells, which were able to grow and divide, were also in seemingly reverse order for these two particular samples (2.5 and 25 mM). It means the initial concentration of cells able to grow and divide in a sample which underwent oxidation by 25 mM of H 2 O 2 was lower than that one of 2.5 mM. Therefore we observed reverse mutual order of latency in onset of the exponential phase of growth curves. However, the results of the identical experiment 52 yielded the expected order of the latency in the growth curves.
This observation can be linked to the size exclusion test results presented in Fig. 4A. The lowest concentration of cells was detected after the strongest oxidation treatment (Fig. 4A, 250 mM H 2 O 2 ). Similarly, the latest onset of the exponential phase was observed in this condition ( Fig. 4B). At this point, we hypothesized that the cells, which underwent physiostructural damage, either swelling, clustering, or fragmentation, were not able to grow and divide.
To further understand cell behavior under the oxidation treatment, we performed 3D holotomographic microscopy to image our cells (Fig. 5, Fig. S1). It is clearly shown that untreated cells are intact and well suspended (Fig. 5A). Contrary to this control, treated cells tend to form clusters ( Fig. 5B-D), moreover, the refractive index of the suspension was slightly changed (data not shown). Finally, based on our observations, we can assume that increasing concentration of H 2 O 2 leads to stronger oxidation causing an increasing rate of physiostructural damage (Fig. 5B-D), which finally results in increasing of BAL levels (Fig. 3).

BAL for monitoring of physically modulated yeast cells' oxidative metabolism by magnetic field.
Since metabolic processes in cells involve oxidative reactions, shown previously to be detectable by BAL 53 and LF MF was reported to be able to affect yeast Saccharomyces cerevisae cells growth dynamics 37,54,55 , we aim to examine, whether BAL dynamics can reflect LF MF effects on cells. Following our previous results which indicate a proliferative response of cells to LF MF in the frequency range 1-2 kHz 56 we chose for this study similar exposure parameters (800 Hz, 1.5 mT) and we monitor BAL throughout the exposure of yeast cells to LF MF. For long-term BAL measurement (12 h) during our MF experiments, similar BAL dynamics for exposed as well as for control sample is typically observed (Fig. 6A). The distinct maximum of BAL intensity is typical for every measurement (Fig. 6A). Additionally, the fast decrease of BAL intensity is always obtained after reaching this maximum (Fig. 6A). Our results are corroborated by earlier data by Quickenden and Tilbury [57][58][59][60][61] where dynamics of increasing BAL with a sudden drop at certain time point was observed during the growth of the yeast cell culture. In general, since the dynamics of BAL is related to metabolic processes of cells, the specific characteristics of this dynamics could be the result of biochemical shifts during cell metabolism.
Therefore, from paired control-MF experiments (n = 21) we have decided to evaluate time instants when the intensity of BAL reaches the mentioned maximum. The results in a form of the time differences in reaching BAL maxima between control and exposed sample (Fig. 6B, Tab. S1) indicate that the maximum tends to occur www.nature.com/scientificreports/ earlier in exposed samples ( p = 0.039 ). The ratios of BAL maxima intensities (Fig. 6C, Tab. S1) do not indicate any significant differences between the exposed and control samples.
Since there are a number of studies 37,[54][55][56]62,63 investigating the magnetic field effect on growth dynamics of yeast cells, we also hypothesized, the magnetic field in our experimental conditions could affect cell growth rate. Therefore, we decided to measure the cell concentration at a specific time point (6 h after the start of each experiment) when cultivated cell cultures are in the exponential phase of the growth cycle. The results, in this case, do not indicate any significant differences between exposed and control samples (Fig. 6D, Tab. S1). It seems that although the magnetic field in these experimental conditions does not have any observable influence on the cell concentration, it could alter the biochemical reactions during the cell metabolism since the time shift in reaching BAL maxima is observable. To obtain an insight into the biochemical background, we also realized a single measurement of dissolved oxygen concentration in the yeast cell sample simultaneously with BAL (Fig. S2). This measurement reflects the rate of oxygen consumption during cell metabolism and population growth. It can be seen that a mutual relation between the BAL and oxygen level is rather complex. However, especially from the final extreme drop of both BAL and oxygen level we can assume that, at least during certain circumstances, the BAL kinetics reflects the oxidative metabolism of cell culture. This is supported by the investigation of Quickenden and Tilbury 59 , where the absence of oxygen eliminates BAL. www.nature.com/scientificreports/ Since BAL kinetics is tightly connected to the growth of the cell culture, we suggest that there is a slight MF influence on yeast metabolism, not observable by cell concentration measurement, but detectable by using BAL. This could indicate beneficial use of BAL for monitoring of weak effects of MF on cells, which are not detectable by standard techniques of cell culture state characterization. Moreover, it may also contribute to understanding the inconsistency and irreproducibility of experimental results due to often hardly measurable LF MF biological effects. However, further studies are needed to better understand the processes behind the disturbance of the BAL curve.
Limitations and future work. Does the BAL originate in the intracellular or extracellular space? From the current data, it is impossible to disentangle the exact location of BAL emitters, i.e. the location where molecules in the electron excited state emit the photon from. Considering the general scheme in Fig. 1, this question has several levels. At first, the question is reasonable, only when we argue that the presence of cells is inevitable for BAL generation. In general, we believe that the reactions, which generate primary ROS that lead to BAL are taking place only when cells are present in the medium. The secondary ROS can be produced in the cells or outside the cells, hence final electron excited state molecules leading to BAL can be produced in the cells or outside the cells. In short, we believe that without active cell oxidative metabolism (which is definitely present during longterm cell cultivation in YPD medium) there is no ROS formation and following BAL response. There are several experimental observations which support these claims. At first, we know that the absence of cells does not lead to the typical BAL curve with slow increase and sharp decrease, see Supplementary data of the Vahalová et al. 52 . At second, after the BAL signal drops, the addition of glucose (primary carbon source for yeast under given conditions) leads to a partial recovery of the BAL intensity, indicating that cellular metabolism is involved (see again Supplementary data of the Vahalová et al. 52 ). Furthermore, we showed in our earlier work 64 that addition of antioxidants (particularly ascorbic acid) decreases BAL signal. Finally, evidence pointing out towards the intracellular origin of BAL comes from images of yeast cell clusters spread on agar medium 65 . In this case, a solid growth medium in a form of YPD agar is chemically similar to our liquid YPD medium, meaning that a potential for extracellular BAL generation is somewhat similar. The only difference is in diffusion properties of each medium.
In order to address this limitation regarding the spatial origin of the BAL, we propose the procedure that could discriminate whether the BAL origin comes mainly from the cells or from the media solution. For this task, one can exploit a microscopic or macroscopic approach. The former is based on analysis of spatial BAL picture obtained from a sensitive camera, where individual cells or their clusters emit excessive BAL. However, it is close to impossible to detect the image of the BAL signals from individual cells because of the low photon fluxes 65 . On the other hand, the latter approach requires either to treat the cells chemically or to separate them, e.g. by filtration through the semi-permeable membrane or by other physical processes. It seems that the exploitation of sedimentation could be the right experiment for a decision regarding the spatial origin of BAL. Figure 6. BAL monitoring and cell concentration in yeast cell culture exposed to low-frequency magnetic field. (A) The representative shape of BAL dynamics from magnetic field exposed and control sample of yeast cells. (B) Boxplot of time differences of BAL maxima reaching between exposed and control sample (t control − t exposed ) shows that the maximum tends to occur earlier in exposed samples ( p = 0.039 ). (C) Boxplot of ratios of BAL maxima intensities (A control /A exposed ). (D) Boxplot of ratios of cell concentrations measured at the time of 6 h (c control /c exposed ).

Conclusion
In the study, we demonstrated a label-free monitoring of chemical and physical stimuli on yeast cells using BAL. The BAL intensity from yeast cell culture under the chemically induced oxidation is dependent on the hydrogen peroxide concentration applied to the solution. The observed results showed the key role of Fenton reaction, likely due to its hydroxyl radical product, in oxidation processes leading to the BAL from yeast cell culture. Although the exact mechanism of oxidative stress-induced BAL from yeast cell culture is not currently known, the presented results are consistent with theoretical assumptions suggesting the reactive oxygen species as the initiators of reaction pathways leading to BAL origin. Moreover, we showed the physiostructural damage of cells after oxidation by measurement of cell concentration, cell viability, cell culture growth, and microscopy imaging as well. Furthermore, we detected a disturbance of BAL kinetics from yeast cell culture under the influence of lowfrequency magnetic field, whereas cell concentration measurement did not indicate any significant difference between magnetic field exposed and control sample. This finding indicates the potential use of BAL for the detection of weak effects of magnetic field on cells, which is not observable by standard techniques of cell culture characterization. The obtained results contribute to the development of innovative approaches for label-free, real-time, noninvasive monitoring of oxidative processes and procedures for their modulation.
Biological autoluminescence measurement setups. The photomultiplier (PMT) module H7360-01, selected type (Hamamatsu Photonics K.K.) with a spectral sensitivity in the range of 300-650 nm is used to detect biological autoluminescence. The quantum efficiency of the PMT module is displayed in Fig. 2A. Typical dark count (noise) of PMT module H7360-01 is about 15 counts per second. The measurement of the sample takes place in a light-tight chamber (standard black box, Institute of Photonics and Electronics, CZ) specially designed for the purposes of BAL measurements. The PMT module is mounted on the bottom of the chamber, viewing the sample inside the chamber. For monitoring of BAL from yeast cell culture under the chemical stimulus, we performed single-chamber measurements (Fig. 2B). For monitoring the physical stimulation of cells, pair measurements of BAL in two identical chambers were carried out (Fig. 2C). Then the sample is centrifuged twice at 3000 rpm, each time for 5 min, washed with purified water, and stirred at the vortex. For the cell exclusion test, the cell concentration after oxidation in each sample is measured by the cell counter (Beckmann Coulter). The cell size range to be detected by the cell counter is set to 3-9 µm . For the cell viability test after oxidation, we used trypan blue solution (0.4%), stained cells for 5 min and counted the number of viable cells on Burker chamber by using optical microscope (Olympus BX 50). To monitor cell growth, we use Multiwell Plate Reader (Tecan Spark) and measure absorbance in 96-well plate at 600 nm for 24 h at 30 • C. The measurement time step was 10 min. The cell samples after oxidation treatment are diluted into 12 wells for each concentration of H 2 O 2 . The absorbance from untreated samples with the same initial concentration and from pure YPD medium is also measured as control samples. For microscope images capturing of cells treated by oxidation, cell culture is diluted to concentration 5 × 10 6 cells/mL. The microscopy images of oxidized cell sample are captured on Holotomographic Microscope (The 3D cell explorer, Nanolive) to analyze physiostructural damage of cells. www.nature.com/scientificreports/ then diluted into two Erlenmeyer flasks (250 mL) containing 150 mL of liquid cultivation medium YPD to set the same initial concentration ( 5 × 10 6 cells/mL) in both samples. Pair measurements of BAL in two identical light-tight chambers are performed. One sample is exposed to magnetic field and one sample is nonexposed (control sample). Magnetic field is generated by the coil (Fig. 2D), designed previously 66 to achieve 95% homogeneity in the exposure area (Fig. 2E). The coil is fed by a harmonic driving signal with frequency 800 Hz, generated by a signal generator (Agilent E4436B, Agilent Technologies, Inc.) and amplified by a linear amplifier (Hubert A1110-05, Dr. Hubert GmbH). Simulation of magnetic and electric field distribution within the coil for our experimental conditions (frequency 800 Hz, amplitude 1 A) was performed in CST Studio Suite 2018 (Fig. 2E, Fig. S3). In the simulation, the conductivity of cell solution was set to σ = 1 S/m and relative permittivity ǫ = 80 . The simulation was experimentally verified, the magnitude of magnetic flux density varied between 1.50 and 1.54 mT in the exposure area (measured by Gauss/Tesla meter 7010, F.W.Bell). Considering the noise of gaussmeter during measurement was ca. 0.08 mT, values from the measurement and simulation (Fig. 2E) match reasonably well. The schematics of exposure setup for pair measurements can be seen in the Fig. 2C. BAL from both samples is measured for 12 h. The samples are cultivated at 30 • C and are mechanically stirred to avoid sedimentation of cells. The temperature in 4 positions in each chamber (Fig. 2C) is recorded during each measurement by two identical 4-channel thermometers (Voltcraft PL-125-T4).

Magnetic field treatment conditions for
Data analysis. BAL raw data were preprocessed and smoothed using Matlab, version R2019b (MathWorks, Inc.). In the case of magnetic field experiments, BAL curve characteristics were analyzed. Two particular features were extracted from the curve shapes: (1) the time point when BAL intensity reached its maximum, (2) the value of BAL intensity at this time point. For statistical analysis, only those measurements were selected, which passed through the following criteria: (1) temperature difference T < 0.5 • C for mean values between chamber with the exposed and chamber with the control sample, calculated from all four temperature channels (Fig. 2C); (2) BAL curve shape without distortions caused by biological (contamination) and technical (stirring adjustment) issues during experiments. For the statistical analysis, unpaired two-sample t-test was performed for the datasets when normal distribution of data was not refused (BAL and concentration of cells under oxidation treatment) and paired Wilcoxon signed-rank test was performed for the cases when normal distribution of data was refused (BAL from cells under magnetic field treatment).

Data availability
The datasets generated and analysed during the current study are available from the corresponding author on reasonable request.