Noninvasive and safe cell viability assay for Euglena gracilis using natural food pigment

Sample preparation

Sample preparation for each experiment was conducted as follows.

The wild type E. gracilis (Z strain) was cultured in CM medium (Cramer-Myers medium (Cramer & Myers, 1952), pH 3.5), stationary and aerobically under continuous illumination with a cool white fluorescent light at 40 µmol/m²/sec and at constant temperature 21 °C before pigment was mixed. The cells were provided by euglena Co., Ltd. This suspension was injected into a micro test tube containing pigment and thoroughly mixed by pipetting. These samples were kept stationary in the micro test tubes with the cap opened under continuous illumination with a cool white fluorescent light at 100 µmol/m²/sec and at constant temperature 21 °C.

Observation of cells by bright field microscope

The bright field microscopes and objective lenses used for the observation of cells in this manuscript are as follows.

• Measurement of absorbance of single live or dead cells (Figs. 1–5). Inverted microscope (IX71, OLYMPUS) with the 100 × objective lens (NA 0.85, LCPlanFLN, OLYMPUS)

• Experiments other than the above. Digital biological microscope (GR-D8T2, Shodensha, Inc.) with the 40 × objective lens (NA 0.65, Shodensha, Inc.), or the 10 × objective lens (NA 0.25, Shodensha, Inc.)

Confirmation of stained cells

Confirmation of stained cells for each experiment was conducted as follows.

Cells were observed by a bacteria counter (A-161, Sunlead Glass Co., Ltd., Saitama, Japan) and a bright field microscope (GR-D8T2) with the 10 × and 40 × objective lenses once a day for 2 days.

Measurement of absorbance of single live or dead cells in the culture with food pigment or synthetic dye (Figs. 1–5)

Sample preparation for measurement of absorbance spectra of single cells.

Cells were cultured to a cell density 8.1 ×10⁵ cells/mL for 9 days. It was divided into the following three kinds of experiments (Table 4).

• Staining of cells in normal culture environment (Naturally Alive/Dead) Cells were measured by absorbance imaging after incubation for 2 days with pigment.

• Staining of dead cells treated with microwave (MW Dead)

• After microwave treatment of the cell suspension, pigment was added and absorbance imaging was measured.

• Staining of dead cells treated with 0.2%(W/V) benzalkonium chloride (BC Dead)

After BC treatment of the cell suspension, pigment was added and absorbance imaging was measured. Since the BC solution is transparent, it does not itself stain cells.

On measurement of absorbance imaging, each sample was diluted 10-fold with CM medium and injected into a glass bottom dish (Matsunami glass D11130H).

As a control sample, pigment solutions were prepared with fresh CM medium as solvent (Table 4).

• Pigment (CM): pigment solution in which the solvent is CM medium

• Pigment (BC): pigment solution in which the solvent is CM medium with 0.2% BC

These control samples were injected into a glass bottom dish without dilution.

Then, samples were measured by scan-free, non-invasive absorbance spectral imaging A (x, y, λ) microscopy (Isono et al., 2015) (Details are described below).

Table 1:

Staining ability of natural edible pigments for dead cells.

Cells were observed with a bacteria counter (A-161, Sunlead Glass Co., Ltd.) and a bright field microscope (GR-D8T2, Shodensha, Inc.) with the 10×objective lens (NA 0.25, Shodensha, Inc.) and with the 40×objective lens (NA 0.65, Shodensha, Inc.) 25 h after the pigment was mixed. All the pigments are derived from foods and purchased from Watashinodaidokoro Co., Ltd. The names of these pigments were translated into English from Japanese.

Color	Black	Brown	Purple	Blue	Green	Yellow	Pink	Red
Food	Bamboo charcoal	Kaoliang	Purple sweet potato	Spirulina	Mixture of Yellow and blue gardenia	Yellow gardenia	Red beet	Monascus
Staining	×	×	∘	×	×	×	×	∘

DOI: 10.7717/peerj.6636/table-1

Notes:

∘Identifiable staining of dead cells.

×Unidentifiable staining of dead cells.

Table 2:

Composition of samples for staining of dead cells.

Samplea	Pigment concentration %(W/V)	Pigment [mg]	TBb 1% (W/W) [µL]	MBc 1% (W/W) [µL]	Cell suspension (CM medium) [µL]	BCd 10% (W/V) [µL]
CM	0	0	0	0	196	4
MP	1	2	0	0	196	4
AP	1	2	0	0	196	4
TB (0.05%)	0.05	0.1	0	10	186	4
TB (0.3%)	0.3	0.6	60	0	136	4
MB (0.05%)	0.05	0.1	0	10	186	4
MB (0.3%)	0.3	0.6	60	0	136	4

DOI: 10.7717/peerj.6636/table-2

Notes:

aCM, Cell Suspension (CM medium); MP, Monascus pigment; AP, Purple sweet potato; TB, Trypan blue (TOKYO CHYEMICAL INDUSTRY Co.,Ltd.); MB, Methylene blue (KOKUSAN CHEMICAL Co., Ltd.).

bThe solvent of 1% (W/W) TB solution was purified water.

cThe solvent of 1% (W/W) MB solution was purified water.

dBC, 10% (W/V) benzalkonium chloride solution (NIHON PHARMACEUTICAL CO., LTD). Since the BC solution is transparent, it does not itself stain cells.

Table 3:

Results of staining of dead cell treated with microwave or BC.

Cells were observed with a bacteria counter (A-161) and a bright field microscope (GR-D8T2) with the 10 and the 40×objective lens after pigment was mixed (Table 2).

Pigment	T 1		T 2
	Time [min]	Staining	Time [min]	Staining
MP	3	∘	2.5	∘
AP	1	∘	1	∘
TB (0.05%)	∼0.5	△	3.5	∘
TB (0.3%)	∼0.5	∘	3	∘
MB (0.05%)	∼0.5	△	∼0.5	∘
MB (0.3%)	∼0.5	△	∼0.5	∘

DOI: 10.7717/peerj.6636/table-3

Notes:

T1The pigment was mixed after cells were treated with benzalkonium chloride solution.

T2The dye was mixed after the cells were treated with microwaves.

Time minThe time from the mixing of pigment until all the cells in the observation area were stained. The time of “∼0.5” means that staining was completed within 0.5 min.

∘Identifiable staining of dead cells.

△Cells were stained, however, because they were thin, it is difficult to distinguish them from living cells.

Table 4:

Pigment concentrations and condition of samples for absorbance spectral imaging (Figs. 1 to 5).

DOI: 10.7717/peerj.6636/table-4

Notes:

aNaturally Alive/Dead: Staining of cells in normal culture environment.

MW Dead, Dead cells treated with microwave.

BC Dead, Dead cells treated with benzalkonium chloride (The final concentration of benzalkonium chloride was adjusted to 0.2% by stock solution (10%).).

Pigment (CM), Pigment solution in which the solvent is CM medium.

Pigment (BC), Pigment solution in which the solvent is CM medium with 0.2% BC.

bTimes from mixture of pigment with cell suspension to measurement.

cDilution rate at the time of measurement of absorbance imaging by CM medium.

dSince the BC solution is transparent, it does not itself stain cells.

Measurement methods of single-cell absorbance.

Detailed methods of scan-free absorbance spectral imaging were previously described (Isono et al., 2015) and the subsequent improvement was reported recently (Yamashita et al., 2018; Yamashita et al., in press). Here we give only brief specifications.

The system for scan-free (snapshot) absorbance spectral imaging is composed of an inverted microscope (IX71, OLYMPUS), a custom-made fiber bundled array to convert a 2-dimensional image to the 1-dimensional slit image, an imaging spectrometer with a large imaging area of 27 × 27 mm² (f = 32 cm, IsoPlaneSCT-Advance, Roper Scientific, Planegg, Germany), and an electrically-cooled CCD camera (2, 048 × 2, 048 pixels with 13.5 µm pixel size, PIXIS2048BUV, Roper Scientific). For the fiber bundled array (2D-1D converter), a 50-µm core/55-µm clad/60-µm coating silica fibers are assembled to 16 ×16 (0.96 × 0.96 mm²) array on the input side and to 1 ×256 (0.06 × 15.36 mm²) array on the output (slit) side. On the side port of the inverted microscope, the ×100 magnified image of the sample was focused on the 16 ×16 2D array (x × y), which is rearranged in order into the 1D array to fit the entrance slit of the spectrometer. The vertical size of both the imaging area (27 mm) and the CCD (27.65 mm) covers that (15.36 mm) of the 1D array of the 2D-1D converter, so that the whole 3D image A(x, y, λ), a datacube of 16 × 16 × 2, 048, is obtained simultaneously without need for the vertical scan. To be precise, the vertical length in the imaging area where an aberration-corrected image is obtained is limited by 14 mm, so that the images from 22–23 fibers at both ends of the 1D array are not aberration-free. The shortest acquisition time is 0.05 s which is limited by the response time of the mechanical shutter of the spectrometer.

The glass bottom dish with sample was set on the inverted microscope and observed with the 100 × objective lens of NA 0.85 (OLYMPUS, LCPlanFLN) from below. The light source was a 150 W Xenon lamp (Hamamatsu Photonics, Hamamatsu, Japan) to illuminate a region of 2 mm in diameter of the sample from above through a condenser. The intensity (photon flux density) on the sample was 1,800 to 2,100 µmol/m²/s for 0.5 s exposure. The transmitted light was transferred through the objective and a focusing lens to the side port of the microscope. The absorbance of the sample was calculated assuming that the transmitted light intensity in the region where only the culture solution exists is 100%. The spectrometer has automatically exchangeable three gratings, 1,200, 300, and 150 grooves/300 nm blaze. We used the 150 grooves grating with wavelength resolution of 1.1 nm at 546 nm (50 µm slit width, determined by the core size of the fiber) and wavelength span from 390 to 790 nm. All the measurements were performed at room temperature.

Table 5:

Composition of samples for the growth curve and dead cell ratio in the culture mixed with pigment (Figs. 6, 7).

Sample	Pigment concentration% (W/V)	Pigment[mg]	Cell suspensionb(CM medium) [mL]	Glucose [mg]
CM	0	0	1	0
CM+G	0	0	1	12
MP	1	10	1	0
MP+G	1	10	1	12
AP	1	10	1	0
AP+G	1	10	1	12
TB	0.3	3	1	0
TB+G	0.3	3	1	12
MB	0.05	0.5a	0.95	0
MB+G	0.05	0.5a	0.95	12

DOI: 10.7717/peerj.6636/table-5

Notes:

aThe 0.05 mL of 1% (W/W) methylene blue solution with CM medium as a solvent.

b“MB” and “MB+G” are 0.05 mL less than other samples. The difference in this amount can be ignored from the result of 5 times of cell density measurement of the following sample.

Sample 1, Cell suspension approximately equal to the initial cell density in this test 6.39 × 10⁵ ± 2.2 × 10⁴ cells/mL (Average ± Standard deviation).

Sample 2, 95% (V/V) of Sample1 diluted with purified water. 6.29 × 10⁵ ± 2.3 × 10⁴ cells/mL (Average ± Standard deviation).

Principal component analysis (PCA) of absorption spectra.

Absorption spectra obtained from different cells in each sample were principal component analyzed. The spectra of the first principal component (PC1), the eigenvalues of PC1 to PC3, and the number of cells for PCA are shown in Figs. 1–5. For the PCA, web application was used (Easy, 2019).

Searching for an appropriate food pigments in viability assay for E. gracilis (Table 1)

Among the food pigments summarized in Table 1, with the pigments extracted from Monascus (MP) and purple sweet potato (AP), dead cells of E. gracilis were stained as vividly as to be clearly distinguished from live cells. With the other pigments, by contrast, dead cells were not stained clearly enough to be visually distinguished.

Confirmation of reliable staining of dead cells (Table 3)

Table 3 shows the results of staining of dead cells treated with benzalkonium chloride (BC) or microwave. It was confirmed that in all the samples, dead cells were reliably stained. However, in the samples with the synthetic dye mixed and BC treatment (TB (0.05%), MB (0.05%) and MB (0.3%) of T1), dead cells were stained instantly, but since the whole cells became so lightly blue that they were difficult to distinguish from the green color of the chloroplasts of living cells. In TB (0.3%) of T1, dead cells were stained clearly enough to distinguish them from live cells. In all the samples with the synthetic dyes mixed and microwave treatment, dead cells were clearly stained. This suggests that the binding of the synthetic dyes to dead cells was weakened by the presence of BC. TB clearly stained dead cells by increasing the concentration, but there was no noticeable change in MB. On the other hand, in all the samples of the natural pigments, all dead cells were clearly stained. This result suggests that they are less susceptible to BC than the synthetic pigments to have more general versatility.

Measurement of absorbance of single live or dead cells in the culture with natural food pigment or synthetic dye (Table 4, Figs. 1–5)

Absorption spectra of E. gracilis in CM medium mixed with MP (Fig. 2)

Figure 2A shows the PC1 of absorption spectra of E. gracilis cultured in CM medium mixed with MP. Pigment solutions without cells (“Pigment (CM)” and “Pigment (BC)”) are characterized by having two absorption maxima around 425 and 515 nm. Spectra of dead cells are all characterized by an increase in the absorbance towards shorter wavelengths in the range of 490 to 600 nm. This is thought to be due to the MP, indicating that the substances in the cell are bound to it. In “BC Dead”, because there is a Q peak (672 nm) of Chl-a, which is lost in “BC Dead” in Fig. 1A without MP, one can tell that MP suppresses destruction of Chl-a by BC. On the other hand, “Naturally alive” shows almost the same structure as that of Fig. 1.

Since the eigenvalues of each sample are approximately 1, the spectrum of each sample can be represented by PC1 (Fig. 2B). It shows that all the samples of dead cells were stained evenly by MP. In the microscopic observation (100 × objective lens), one can readily distinguish between “Naturally Alive” and other samples of dead cells (Fig. 2C), even at low magnification (10 × objective lens).

Absorption spectra of E. gracilis in CM medium mixed with AP (Fig. 3)

Figure 3A shows the PC1 of absorption spectra of E. gracilis cultured in CM medium mixed with AP. Pigment solutions without cells (“Pigment (CM)” and “Pigment (BC)”) are characterized by having an absorption maximum around 530 nm. It is known that AP extracted from purple sweet potato has absorption maximum at 530 nm (He et al., 2015). Spectra of dead cells are all characterized by an increase in the absorbance towards shorter wavelengths in the range of 540 to 645 nm. This is thought to be due to the AP, indicating that the substances in the cell are bound to it. In “BC Dead”, because there is a Q peak (672 nm) of Chl-a, one can tell that AP suppresses destruction of Chl-a by BC as well as MP. On the other hand, “Natural alive” shows almost the same structure as that of Fig. 1.

Since the eigenvalues of each sample are approximately 1, the spectrum of each sample can be represented by PC1 (Fig. 3B). It shows that all the samples of dead cells were stained evenly by AP. In the microscopic observation (100 × objective lens), one can readily distinguish between “Naturally Alive” and other samples of dead cells (Fig. 3C), even at low magnification (10 × objective lens).

Absorption spectra of E. gracilis in CM medium mixed with TB (Fig. 4)

Figures 4A, 4B shows the PC1 of absorption spectra of E. gracilis cultured in CM medium mixed with TB. Pigment solutions without cells (“Pigment (CM)” and “Pigment (BC)”) are characterized by having an absorption maximum around 585 nm. “Pigment (BC)” has another absorption maximum at 625 nm. The absorption spectra of dead cells have a characteristic broad absorption maximum in the range of 550 to 590 nm not present in living cells. This is thought to be due to the TB, indicating that the substances in the cell are bound to it. There are two kinds of staining colors of dead cells, blue and purple. The spectra of dead cells stained purple have a broadened absorption maximum around 680 nm of Chl-a, whereas the spectra of dead cells stained blue do not. Therefore, when TB coexists with Chl-a within the cells, dead cells are stained purple, and when Chl-a does not exist, they are stained blue. Since the absorption maxima around 420 nm are also found in dead samples having no Chl-a absorption maximum around 680 nm, it is considered to be due to other intracellular pigments. In addition, as blue stained cells increased with time, TB destroys Chl-a although it does not as rapidly as BC. On the other hand, “Naturally alive” shows almost the same structure as that of Fig. 1.

Since the eigenvalues of each sample are approximately 1, the spectrum of each sample can be represented by PC1 (Fig. 4C). It shows that all the samples of dead cells were stained evenly by TB. Note that the above result is obtained by dividing the PCA into the two kinds of colors both in “Naturally Dead” and “BC Dead”. In the microscopic observation (100 × objective lens), one can readily distinguish between “Naturally Alive” and other samples of dead cells (Fig. 4D), even at low magnification (10 × objective lens).

Absorption spectra of E. gracilis in CM medium mixed with MB (Fig. 5)

Figures 5A, 5B shows the PC1 of absorption spectra of E. gracilis cultured in CM medium mixed with TB. Pigment solutions without cells (“Pigment (CM)” and “Pigment (BC)”) are characterized by having two absorption maxima around 610 and 665 nm respectively. The absorption spectra of dead cells have absorption maxima near these two absorption maxima. This is thought to be due to the MB, indicating that the substances in the cell are bound to it. There are two kinds of staining colors of dead cells, blue and blue–green (Fig. 5D). Blue–green is shown as “Green” in the figure. Since one of the absorption maxima (665 nm) is close to the absorption maximum of Chl-a (680 nm), it is difficult to judge the presence of Chl-a, but the spectra of the cells stained green seems to contain more Chl-a due to the higher absorbance in the range of 390 to 500 nm than the blue stained cells. Therefore, when MB coexists with Chl-a, dead cells stained blue–green, and when Chl-a does not exist, they stained blue. In addition, as blue stained cells increased with time, MB destroys Chl-a although it does not as rapidly as BC. On the other hand, “Naturally alive” shows almost the same structure as that of Fig. 1.

Since the eigenvalues of each sample are close to 1, the spectra of each sample can be represented by PC1 (Fig. 5C). Note that the above result is obtained by dividing the PCA into the two kinds of colors both in “Naturally Dead” and “BC Dead”. In the microscopic observation (100 × objective lens), one can readily distinguish between “Naturally Alive” and blue stained cells (Fig. 5D), even at low magnification (10 × objective lens). However, it is difficult to visually distinguish between blue–green stained cells and living cells.

Growth curve and ratio of dead cells of E. gracilis in CM medium mixed with pigment (Figs. 6 and 7)

Appropriate concentration of natural pigments for viability assay (Tables 6 to 8)

As a result of examining an appropriate concentration at which staining and health of cells was maintained over a long period, the concentration of MP was found to be 1.5%(W/V). With this concentration, it was enabled to identify stained dead cells for 6 days while maintaining the health of living cells. However, fading of staining was observed after 4 days. More fading was observed at 0.9 to 1%(W/V), but it was enabled to identify stained dead cells for 6 days. At 2%(W/V) or more, although the number of swimming cells decreased, cells that did not swim were alive because they moved flagella or showed stretching motions, and the ratio of stained cells on the 4th day was roughly 10%.

Because AP tended to increase stained cells in proportion to the concentration, it is difficult to balance the staining of the cells with the health of them, but the balance is relatively good between 0.5 and 1%(W/V). However, even at 0.05%(W/V), since dead cells increased compared with the control sample, it is necessary to use it together with glucose for long-term culture. During the observation period (6 days), since AP contained in dead cells and in the culture did not fade, the stability against light was found to be higher than MP. This is probably due to the fact that AP is a stable pigment under acidic conditions (Khoo et al., 2017). Therefore, it is convenient to use it in combination with CM medium which is acidic (pH 3.5). In addition, from Fig. 3, it is confirmed that the absorption maximum of AP in the solution (not bound to cells) is 529 nm, but its maximum value largely shifts to 540 nm by binding with substances in the cell.

From the above results, it is important to select the pigment properly depending on the purpose and target cells. For example, it may be one of the future tasks to combine MP and AP and to search for the optimum blending ratio for staining dead cells in a shorter time and for longer-term monitoring of cell viability. In addition, since there are many variations of natural pigments, their application range is wide. In particular, it can be applied to cell counting by flow cytometry, since strong staining of dead cells is obtained by selecting an appropriate pigment. It may also contribute to elucidation of the metabolic pathway by interaction between a pigment and a living cell.