Production and high throughput quantification of fucoxanthin and lipids in Tisochrysis lutea using single-cell fluorescence

A R T I C L E I N F O


Introduction
In the past years, Tisochrysis lutea (T. lutea) has obtained increased interests for fucoxanthin (Fx) and lipid production. Fx is of interest due to its biological properties, such as antioxidant, anti-obesity, and antidiabetic (Fung et al., 2013;Maeda et al., 2018). T. lutea contains up to 18.23 mg/g dry weight (DW) Fx and 32% of total fatty acids DW (Balduyck et al., 2016;Kim et al., 2012). The fatty acids are typically classed as polar or neutral lipids. Polar lipids are the main structural fatty acids that can be utilized as functional additive in food or feed, while neutral lipids serve as storage lipids that can be used for biofuels (Mairet et al., 2011). Within the lipids in T. lutea, the long-chain omega-3 fatty acid docosahexaenoic acid (DHA), is associated with numerous health benefits such as promoting fetal development, including neuronal, retinal, and immune function, improving cardiovascular function such as inflammation, peripheral artery disease (Swanson et al., 2012). In addition to health benefits for humans, Fx and long-chain omega-3 fatty acids also play a key role in health and development of aquatic organisms, which makes T. lutea a widely used feed in aquaculture (Bigagli et al., 2018).
Fx and polar/neutral lipids quantification are essential to assess the quality of T. lutea biomass. The traditional methods for pigment and polar/neutral lipids measurements are high-performance liquid chromatography (HPLC) and gas chromatography (GC) respectively (Foo et al., 2017;He et al., 2019;Niemi et al., 2019). These techniques are common methods for the detection of these compounds with acceptable accuracy. However, complex pre-treatments such as freeze drying and cell disruption are required for both methods; several organic chemicals, such as chloroform, methanol, and hexane, are used in the extraction procedure, which are not environmentally-friendly. In addition, a relatively large amount of biomass needs to be used to meet the limit of detection, which limits the number of samples which can be taken. Hence, HPLC and GC are expensive, as well as time and biomass consuming. A rapid method was developed for the determination of Fx in Diatom using spectrophotometry (Li et al., 2018). However, ethanol extraction was still needed for this method. Therefore, a fast method using high-throughput techniques is needed for process control and strain selection.
A promising alternative technique based on FACS offers high throughput and fast analysis of microalgal cells (Park et al., 2019;Pereira et al., 2018). It is a fast, reliable and efficient method that can detect the compounds based on the fluorescence signal from single cells and can also be used for metabolic monitoring (Li and Wilkins, 2020). Previous research studied the possibilities to use FACS for high lipid content microalgal cells selections with Nile red (NR) or BODIPY for neutral lipids staining (Cabanelas et al., 2016;Chen et al., 2009). Monitoring and screening high-polar lipids cells are getting more attention due to more functional applications. However, the ability of FACS to distinguish the fluorescence signals from polar and neutral lipids in T. lutea is still missing. Concerning Fx, no previous studies have reported the use of FACS for determination of this compound. Therefore, the use of FACS to measure polar lipids and fucoxanthin was explored. Moreover, correlations between concentrations of these compounds could make it possible to detect them in simultaneously.
This study aimed to develop a method for high throughput measurement of Fx, polar and neutral lipids in T. lutea using FACS. To do so, correlations were analysed between these compounds with single-cell fluorescence. Batch and continuous experiments were designed to produce T. lutea with varying concentrations of pigments and lipids by varying light and nitrogen concentration. Plate reader and FACS were used to analyse the autofluorescence of pigments and the fluorescence of lipids stained with NR. This study established reliable correlations between single-cell fluorescence and cellular contents of Fx, polar/ neutral lipids, and chlorophyll a (chl-a) consequentially, allowing the use of FACS as a high-throughput tool to monitor these compounds in T. Lutea.

Batch experiment
T. lutea was cultivated in 250 mL flasks containing 100 mL medium at 25°C, light intensity 140 μmol m −2 s −1 , 150 rpm and 2% CO 2 . A 18:6 day:night photoperiod was applied for cultivation under photoautotrophic conditions. All flasks had the same initial biomass concentration with a 0.2 optical density (OD) at 750 nm. The flasks were changed positions in incubators randomly every day to avoid the effect of different light intensities at different positions. A volume of 200 mL culture was harvested every day for Fx and lipid measurement. The batch experiment lasted for 8 days.

Continuous chemostat experiments
T. lutea was inoculated into flat-panel photobioreactors (Algaemist (Breuer et al., 2012), light path 14 mm) containing 400 mL NutriBloom medium at a light intensity of 300 μmol m −2 s −1 with an 18:6 day:night cycle and at 30°C. The cultures were diluted at different specific dilution rates: 0.16, 0.28, 0.35, 0.51, 0.55 d −1 (chemostat experiments ). The overflow culture from the Algaemist was harvested every day for Fx and lipid analysis.

Daily measurements
The OD 750 of cultures was measured with a 1.0 cm light path cuvette in a HACH LANGE DR600 spectrophotometer. Cell number and size were measured using a Beckman Coulter Multisizer. To measure the dry weight (DW), 2 mL of culture was diluted with 50 mL 0.5 M ammonium formate and filtered through a pre-dried Whatman GF/F filter (0.7 μm pore size) and washed twice with 50 mL ammonium formate. Cells on the filters were dried at 100°C in the oven for 24 h (until constant weight) and were subsequently cooled to room temperature in a desiccator for 2 h before weighting. The maximal photosystem II quantum yield (QY) was measured using an AquaPen AP100 after culture dark adaptation for 10 min to determine the ratio of variable fluorescence to maximal fluorescence (Parkhill et al., 2001). The overflow volume in chemostat experiments were measured every day. The harvested culture from batch and continuous experiments was centrifuged in 50 mL Greiner falcon tubes at 4255 × g for 5 min. After discarding the supernatant, the cell pellets were resuspended and combined in 50 mL 0.5 M ammonium formate and were centrifuged using the same centrifugation settings. The washing step with ammonium formate was repeated twice. The cell pellets after washing were flushed with N 2 and stored in a freezer at approximately −20°C. The nitrate in medium was checked using MQuant TM Nitrate Test (Merck KGaA, Germany). To obtain the supernatant, 2 mL culture was centrifuged at 10000 × g for 2 min. The test strip was immersed in the supernatant for 1 s and the excess liquid was shaken off. The nitrate was determined with the colour on the label after 1 min reaction.

Pigments extraction and measurement
The frozen harvest biomass was freeze dried (Sublimator 2 × 3 × 3 5, Zirbus Technology, the Netherlands) for 41 h. Small samples of biomass (1-2 mg) were extracted with 1 mL of ethanol using bead beater tubes (MP Biomedicals, REF 6914-500, composition: 1.4 mm ceramic spheres, 0.1 mm silica spheres, and one 4 mm glass sphere) in a beater (Bertin Technologies). Fx extraction and sample preparation were performed according to the previous reported methods . To quantify the concentration of Fx and chl-a, samples were run in a HPLC (Shimadzu, Nexera UHPLC) according to Grant (2011). The F. Gao, et al. Bioresource Technology 318 (2020) 124104 injected sample volume was 20 µL, the stationary phase was a C18 column (Kinetex C18 5 μm 100 Å 150 × 4.6 mm). The standards of Fx and chl-a were from Sigma-Aldrich and were diluted with methanol to different concentration of 0, 2, 4, 6, 8, 10 μg/mL for concentration calibration.

Fatty acids extraction and measurement
Fatty acids extraction, separation into polar and neutral lipids and quantification were performed as described by Remmers et al. (2017). Approximately 10 mg of freeze-dried biomass was weighted for lipid extraction. Chloroform: methanol (v:v = 2:2.5) solution with added internal standard (pentadecanoic acid (neutral lipid) and decanoic acid (polar lipid)) was used as extraction solvent. After extraction, polar and neutral lipids were separated using an SPE column (Sep Pak Vac Silica cartridge, 6 cc, 1000 mg, Waters). The columns were equilibrated with 10 mL hexane. Concentrated samples were resuspended in 1.5 mL hexane/diethylether (v:v = 7:1) and loaded onto the column. Then, neutral lipids were eluted using 10 mL of hexane/diethylether (v:v = 7:1), which was collected in a fresh glass tube. Subsequently, polar lipids were eluted with 10 mL methanol/acetone/hexane (v:v:v = 2:2:1) and collected in fresh glass tubes. Samples were evaporated with N 2 at 30°C.
To detect fatty acids with GC, the fatty acids were methylated. To every sample, 3 mL of methanol/sulfuric acid (v:v = 95:5) was added, after which the samples were heated at 70°C for 3 h. After cooling to room temperature, 3 mL water and 3 mL hexane was added to the samples. The samples were mixed for 15 min and centrifuged for 5 min (1204 × g, 15°C). 2 mL of the hexane phase was transferred to fresh glass tubes, which was then washed with 2 mL water. The hexane phase was used for GC analysis.
An Agilent 7890 GC system (Santa Clara, CA, USA) was used to determine the lipid contents. GC vials were filled with the sample containing hexane phase and placed in the autosampler. A Supelco Nucol 25357 column (30 m × 530 µm × 1.0 µm) was used with hexane as running solvent and helium as carrier gas (20 mL min −1 ). Data was collected and processed with Openlab software (7890 GC). The total fatty acids was calculated by summing polar and neutral lipids.

Fluorescence measurement 2.7.1. Fluorescence measurement in batch using plate reader
The culture was diluted to OD 750 = 0.2 using fresh medium and an amount of 100 μL was pipetted in triplicate in a black 96-well plate with transparent cover to avoid signal interference from different wells. The cells were stained with NR (final concentration 2 μg/mL) for 5 min in dark (Bertozzini et al., 2011;Sakurai et al., 2016). The NR fluorescence was measured from the top using a Synergy HTX Multi-Mode Microplate Reader at an excitation wavelength of 488 nm with an emission wavelength range from 530 nm to 800 nm. The NR fluorescence was measured every day at the same time for each batch experiment. The autofluorescence was measured using the same procedure without NR staining.

Single-cell fluorescence measurement using FACS in continuous experiment
Samples were collected every day (approximately 1 h before the dark period started) from the Algaemist and were diluted to OD 750 = 0.2 using fresh medium. To measure the autofluorescence of the single cells, 200 μL sample was analysed by FACS (Sony Cell Sorter SH800S) with 100 µm microfluidics sorting chip (LE-C3210, Sony, Japan). A 488 nm laser was used to excite compounds, three emission channels (FL2 585/30, FL3 617/30, FL5 720/60) were applied for the fluorescence detection. Cells for fluorescence measurement were selected from three different density plots. Firstly, the backward scatter area (BSC-A) was plotted against the forward scattered light area (FSC-A) to distinguish bacteria and other impurities from the microalgae. A gate was created to select only microalgal cells. Next, the FSC corresponding to a cells' height was plotted against the FSC area to select single-cells. Finally, a gate was created in a BSC height versus area plot to include all cells. The fluorescence signal was recorded for two minutes.
The fluorescence associated to lipids was measured following the staining method reported in the batch section. The NR fluorescence was measured as the method described for autofluorescence. The survival rate of the single cells was analysed after the sorting procedure. Single cells with or without NR staining were sorted to 96-well plate with 100 μL medium in each well. Each well has one single-cell. The cells were incubated in a climate room at light intensity~40 μmol m −2 s −1 , 25°C. The wells with growing cells were counted after 7-10 days incubation.

Statistical analyses
The results were analysed based on the data from at least three different points during steady state in continuous experiments. Singlecell fluorescence signals were obtained from 100 000 events. Experimental results were expressed as mean value ± SD. Differences between groups were tested for significance by the least significant difference mean comparison using the IBM® SPSS® Statistics software (version 25) software program. The relationship between variables was determined by one-way ANOVA at a significance level of 0.05 using a Duncan Post-Hoc test.

Fucoxanthin and chlorophyll dynamics during growth
During batch cultivation, the fucoxanthin (Fx) content increased rapidly in the first three days of exponential growth, from 6.83 mg/g to 13.09 mg/g, and then declined 52.94% to 6.16 mg/g (Fig. 1a). This decline is due to nitrogen starvation which started at the middle of day 3, when all nitrogen was consumed, and lasted until the end of the experiment (day 8). Chl-a followed a similar trend as Fx, showing an increase from 8.20 mg/g to 18.94 mg/g from day 1 to day 3 and a decrease to 4.87 mg/g at day 8 (Fig. 1b). The Fx and chl-a contents in T. lutea showed a high positive correlation ( Fig. 1b; R 2 = 0.95). In microalgal cells, Fx is bound to protein complexes forming the Fx-chlorophyll protein (FCP), which can be found inside the thylakoid membrane (Pyszniak and Gibbs, 1992). The crystal structure of FCP reported in Phaeodactylum tricornutum reveals the binding of seven chl-a, two chl-c and seven Fx (Wang et al., 2019). Fx is the major light harvesting pigment surrounded by chlorophyll, enabling the energy transfer and quenching via Fx with high efficiency.
Nitrogen plays a key role in pigment accumulation. Similar to our results, the Fx content in P. tricornutum was higher with 14.5 mM nitrogen (6.56 mg/g) than that with 2.9 mM nitrogen (5.53 mg/g) after 12 days cultivation (Gao et al., 2017). Fx is a primary pigment in Isochrysis (Custódio et al., 2014) and, like chlorophyll, is correlated with cellular replication (i.e., formation of new biomass). In the present study, the initial nitrogen concentration was planned to be low (4 mM) and consumed at day 3, which was confirmed by non-detectable concentrations of nitrogen. After that, photosynthesis continues but no new functional (nitrogen-rich) biomass can be produced, which resulted in the declined fucoxanthin content.

Total, polar, and neutral lipids dynamics in batch cultivation
The nitrogen (4 mM NaNO 3 ) in the medium was sufficient for three days cultivation. T. lutea was cultivated at nitrogen starvation (no nitrogen) condition from day 4 to 8. The total fatty acids (TFA) content F. Gao, et al. Bioresource Technology 318 (2020) 124104 remained constant (approximately 250 mg/g) during nitrogen starvation, and were higher than during nitrogen-sufficient period (164.60-200.98 mg/g) (day 1-3) (Fig. 2a). The polar lipids were the main lipid fraction in the first three days (208.39 mg/g) but decreased to 137.41 mg/g from day 4 to day 8 (Fig. 2a). On the contrary, the neutral lipid content increased from 6.81 mg/g to 94.02 mg/g from day 3 to day 8 (Fig. 2a). The percentage of polar lipids in TFA decreased from 95.13% to 59.37%, while the percentage of neutral lipids increased from 4.87% to 40.63% of TFA, under nitrogen starvation. Neutral lipids are accumulated during nitrogen depletion (Benvenuti et al., 2016), which explained their increase after day 3 (Fig. 2a). It is hypothesized that the polar lipids were converted to neutral lipids while TFA stayed constant from day 4 to day 8. A conversion of membrane lipids to storage lipids was observed after changing nitrogen availability in T. lutea (Garnier et al., 2016). In addition, the cellular triacylglycerol (proportion in neutral lipids) content (pg cell −1 ) in T. lutea increased significantly from logarithmic (166-240 fg cell −1 ) to stationary phase (539-566 fg cell −1 ) (da Costa et al., 2017). Moreover, neutral lipids accumulation was reported under nitrogen limitation conditions in T. lutea due to inducing a reorganization of thylakoid membranes .

Correlation between polar/neutral lipids content and Nile red fluorescence
The fluorescence of NR-stained cells showed emission signals within the range of 540-660 nm (Fig. 2b). The fluorescence peak was at 620-640 nm for the first three days when the major lipids were 95% polar (Fig. 2b). However, the fluorescence peak shifted to 590-610 nm when the amount of neutral lipids was above 3.48% (W/W) (Fig. 2a, b). This means that with NR staining, polar and neutral lipids can be distinguished from each other. However, only if cells have < 3.48% (W/ W) neutral lipids, the detected fluorescence can be associated with polar lipids. This limitation is explained by the fact that the fluorescence peaks of polar and neutral lipids are close to each other. Therefore, they are difficult to be differentiated. Similar results were found by Alonzo and Mayzaud (1999) using polar and neutral lipid standards. The NR staining fluorescence emission spectrum of phosphatidylcholine (polar lipid standard) ranged from 530 to 680 nm with a peak at 610 nm with excitation at 530 nm. However, the emission spectrum of triolein (neutral lipid standard) shifted 30 nm towards lower wavelengths ranging from 520 to 660 nm with a peak at 580 nm. (Alonzo and Mayzaud, 1999) The overlay between polar and neutral lipids was also observed. Similar results were obtained in vivo for the first time, as found using standards. The ratio of polar/neutral lipids changes depending on the cultivation stage or metabolic activity of the cells. Consequentially, it is possible to monitor the fluorescence signals associated with these lipid classes as a proxy for gravimetric methods. Thus, fluorescence from 620 and 590 nm can be used for polar and neutral lipids detection in T. lutea. Therefore, the ratio of polar/neutral lipids was positively correlated with the ratio 620/590 nm (R 2 = 0.99), allowing us to estimate the relative lipid composition by assessing single-cell lipid fluorescence. Similar correlations were observed between the cellular contents in polar and neutral lipids and the signal of microalgal cells marked with NR (575 nm signal-neutral lipids, R 2 = 0.87; 620 nm signal-polar  F. Gao, et al. Bioresource Technology 318 (2020) 124104 lipids, R 2 = 0.61) (Guzmán et al., 2011;2010). NR staining is a common method to detect neutral lipids in microalgal cells. For instance, Bertozzini et al. (2011) reported the NR method (excitation wavelength 547 nm; emission wavelength 580 nm) for the absolute quantification of neutral lipids in Skeletonema marinoi, Chaetoceros socialis, and Alexandrium minutum. In the present study, the emission fluorescence signals at 620 nm and 590 nm (excited at 488 nm) can be used for polar and neutral lipids detection.

Growth, fucoxanthin and chlorophyll contents in chemostat experiments
Several chemostat experiments were performed at different dilution rates of 0.16, 0.28, 0.35, 0.51 and 0.55 d −1 with a limiting nitrogen concentration (Fig. 3a). The nitrogen was supplied by dropping from the fresh medium vessel, which was consumed by T. lutea immediately. Therefore, no nitrogen (NO 3 − ) can be detected from the culture. Nitrogen supplementation was higher at higher dilution rates. The biomass concentration was higher at lower dilution rates, with the exception at 0.28 d −1 , which was higher than the biomass at 0.16 d −1 ( Table 1).
The chl-a content was positively correlated to Fx content (R 2 = 0.99) and in line with the results obtained in batch (R 2 = 0.95). The Fx content increased 40.53% from 3.36 mg/g to 5.65 mg/g (p < 0.05) and the chl-a content increased from 3.86 to 8.39 mg/g as the dilution rate increased from 0.16 d −1 to 0.55 d −1 (Fig. 3b). Since the nitrogen supply rate is a function of the dilution rate (i.e., growth), so was chl-a content a function of nitrogen supply rate (Fig. 3b). The reduction in Isochsysis galbana pigmentation was reported before under nitrogen limitation (Falkowski et al., 1989). With decreasing dilution rate from 0.96 to 0.18 d −1 in nitrogen limited chemostat mode, the cellular chl-a decreased 50% in that study (Falkowski et al., 1989).

Correlation between pigment and lipid content
A positive correlation (R 2 = 0.88) was found between Fx content and chl-a content when combining data from chemostat experiments with those from batch experiment (Fig. 4). This correlation is valid at different cultivation settings with different Fx and chl-a concentrations.
The TFA content ranged from 25.59% to 31.43% DW without obvious trends (Table 1). The percentage of polar lipids of TFA increased from 36.98% to 54.78% and the content increased from 116.28 to 157.49 mg/g as increasing dilution rate from 0.16 to 0.55 d −1 (Table 1). In continuous cultivation, the maximum percentage of polar lipids in TFA (54.78%) was lower than the minimum percentage (59.37%) measured in batch, which indicates that a continuous nitrogen limitation leads to a further decrease of the polar lipids. Moreover, the TFA content in continuous nitrogen limited conditions was higher than in batch, reaching 31.34%. The stable TFA content was reported during starvation period in batch. The higher TFA in continuous chemostat experiments was probably related to the different cultivation parameters such as light intensity and temperature.
The correlations between pigment and lipid were analysed both for batch and continuous experiments. Fx content was positively correlated to chl-a content ( Fig. 4; R 2 = 0.88) and polar lipids ( Fig. 4; R 2 = 0.88). Polar lipids are described as membrane-bound while Fx and chl-a are combined in FCP in membrane. Hence, these compounds are all membrane-bound, which explains the positive correlations. To our knowledge, this is the first study reporting the correlations between these three compounds in T. lutea, which is a useful evidence for detecting these compounds simultaneously.  F. Gao, et al. Bioresource Technology 318 (2020)

Correlation between intracellular fucoxanthin or polar/neutral lipids content and single-cell fluorescence
The autofluorescence spectrum showed the distributions of the fluorescence signal from single cells (Fig. 5). The mean intracellular Fx content increased from 0.122 to 0.148 pg/cell, corresponding to the increasing emission autofluorescence signal at 720 nm from FACS (Fig. 5). Intracellular Fx content was positively correlated to single-cell fluorescence (R 2 = 0.90). Therefore, single-cell fluorescence at 720 nm can be used for Fx measurement. The emission bands of extracted Fx from brown algae were around 630, 685 and 750 nm (Katoh et al., 1991). Although the emission brands of extracted Fx can be detected, these bands cannot be distinguished from intracellular Fx due to overlapped signals from multi-compounds and energy transfer from Fx to chlorophyll inside FCP. The emission fluorescence of isolated FCP from diatoms involved in non-photochemical quenching ranged from 680 to 725 nm (Gundermann and Büchel, 2012;Miloslavina et al., 2009). The single-cell fluorescence found at 720 nm in the present study was probably excited from FCP in T. lutea as well. However, the structure and working mechanism of FCP may differ with microalgal species.
The ratio (0.59-1.21) of polar/neutral lipids was positively related to the ratio (1.13-1.52) of single-cell fluorescence at 617 nm/585 nm from FACS (R 2 = 0.89); similar to the results in batch using a plate reader. In addition, intracellular polar lipids content was positively correlated to chl-a autofluorescence (R 2 = 0.98). Similar relationships between lipid and NR fluorescence was found in Crypthecodinium cohnii (De La Jara et al., 2003). The fluorescence signal ratio of FL3/FL2 (FL3 > 650 nm; FL2 560-640 nm) was linearly correlated with the neutral and polar lipid content (R 2 = 0.96). Although the emission lasers they used were different in this study, the relationships were similar. FACS showed a strong ability to analyse pigment and lipid contents by measuring single-cell fluorescence.
In the present study, the positive correlation between Fx and chl-a autofluorescence was found in both plate reader measurements from batch experiment and FACS measurements from continuous chemostat experiments. The relative neutral and polar lipids contents can be determined by NR fluorescence signal at 585 nm and 617 nm using FACS. Although many previous studies reported the ability of FACS for neutral lipid measurement (Cabanelas et al., 2015;De La Jara et al., 2003), this is the first work reporting Fx measurement using FACS.

Single-cell sorting and viability assay
In recent years, FACS has been widely used for high throughput measurement and cell selection (Hyka et al., 2013;Montero et al., 2011). The direct sorting onto solid medium upon FACS could save about 3 weeks during the scale-up process as compared to the growth of the same cultures in liquid medium (Pereira et al., 2011). The ability of FACS for high throughput measurement of pigments and lipids can be also used for the selection of high-Fx or high-lipids cells of T. lutea. Several previous studies reported protocols for viable microalgal cell sorting after staining with BODIPY or NR (Cabanelas et al., 2016(Cabanelas et al., , 2015Montero et al., 2011;Yen Doan and Obbard, 2011). Nevertheless, some strains have been reported to be sensitive to the stress of the selection procedure (Elliott et al., 2012;Lee et al., 2013;Li et al., 2019). Therefore, the suitability of our method was evaluated in staining and sorting viable cells of T. lutea.
Using FACS as a selection tool in T. lutea strain improvement processes depends on cell survival after sorting. T. lutea is a sensitive microalgal species without cell wall. The single-cell viability of T. lutea was analysed after sorting with or without NR staining. Surprisingly, 77.5% single-cells sorted from the gate of autofluorescence at 720 nm survived while 80.0% single cells selected from the gate of NR staining fluorescence at 585 nm survived after sorting. Most of the cells can survive after sorting (Cabanelas et al., 2015). In this work, survival rates of 77.5% and 80.0% were observed for respectively non-stained and stained cells. This finding demonstrated FACS could be a promising tool in monitoring Fx and polar/neutral lipids and cell sorting of T. lutea. Cells with high autofluorescence at 720 nm are expected to have high Fx, chl-a and polar lipids contents.
The correlations between pigments and lipids make it possible to quantify these compounds in one single measurement. In the present study, it was shown that polar lipids can only be determined by NR fluorescence when its content is higher than 96.52% (W/W) of TFA. The positive correlations between these three compounds evidenced the use of autofluorescence from chl-a for both Fx and polar lipids determination. Compared to NR staining, autofluorescence without dye pre-treatment at 720 nm is more convenient for pigments and lipids monitoring and cell sorting due to the positive correlations between these compounds.    Gao, et al. Bioresource Technology 318 (2020) 124104

Conclusions
Here, batch and continuous experiments were conducted to produce biomass with varying concentrations of pigments and lipids. Fx was always coupled with chl-a and was positively correlated to polar lipids. Fx can be estimated by measuring chl-a autofluorescence emission at 720 nm. The polar/neutral lipids ratio can be monitored by NR fluorescence at 617 nm and 585 nm. High throughput measurement of pigments and polar/neutral lipids content in T. lutea were shown from single-cell fluorescence. FACS technique can be used in high throughput measurement of pigments and lipids and can be further applied in high-Fx/lipids cell selection for strain improvement.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.