A Rapid Method for the Determination of Fucoxanthin in Diatom

Fucoxanthin is a natural pigment found in microalgae, especially diatoms and Chrysophyta. Recently, it has been shown to have anti-inflammatory, anti-tumor, and anti-obesityactivity in humans. Phaeodactylum tricornutum is a diatom with high economic potential due to its high content of fucoxanthin and eicosapentaenoic acid. In order to improve fucoxanthin production, physical and chemical mutagenesis could be applied to generate mutants. An accurate and rapid method to assess the fucoxanthin content is a prerequisite for a high-throughput screen of mutants. In this work, the content of fucoxanthin in P. tricornutum was determined using spectrophotometry instead of high performance liquid chromatography (HPLC). This spectrophotometric method is easier and faster than liquid chromatography and the standard error was less than 5% when compared to the HPLC results. Also, this method can be applied to other diatoms, with standard errors of 3–14.6%. It provides a high throughput screening method for microalgae strains producing fucoxanthin.


Introduction
Fucoxanthin is a carotenoid belonging to the xanthophyll class of carotenoids [1]. In recent years, much work has focused on studying the effect of dietary fucoxanthin and on demonstrating that fucoxanthin can be used as a safe and effective dietary supplement. It has anti-inflammatory, anti-tumor, anti-obesity, anti-diabetes, anti-malarial, and other physiological activities [2][3][4]. Clinical studies have shown that taking fucoxanthin can speed up the metabolism, but will not stimulate the central nervous system [5]. Fucoxanthin is widely found in brown algae and diatoms [6]. In macroalgae, the content of fucoxanthin is about 0.1-1 mg·g −1 (dry cell weight). Health products containing fucoxanthin derived from brown algae are already sold on markets. However, the production price is too high to meet the expectations of markets, due to the long growth cycle of macroalgae and the low extraction efficiency [7,8]. In microalgae, especially diatoms, fucoxanthin is one of the main pigments in cells, and accounts for about 1-2.5% of the dry cell weight, which is several fold higher than in macroalgae [9,10]. Fucoxanthin is a major component in the Fucoxanthin-Chlorophyll Protein (FCP) complex [11]. The FCP has the function of trapping light energy and light protection, and plays an important role in photosynthesis in diatoms. In the extraction procedures, the high content of lipid is beneficial to the extraction of fucoxanthin. At present, all over the global market, the

Results
When the total pigments of P. tricornutum at different culture stages were quantified by HPLC, there are mainly five product peaks: chlorophyll c, fucoxanthin, diatoxanthin, chlorophyll a, and β-carotene. Chl c levels varied greatly at different culture stages, but Chl a and fucoxanthin were relatively stable over the culture (Figure 1). The visible spectrum scan of the total pigment extract contained two main peaks. A broad peak centered at 445 nm and a sharper peak centered at 663 nm. But the purified fucoxanthin exhibits little absorption at 663 nm. However, at 445 nm, the absorption peak has contributions from both fucoxanthin and Chl a ( Figure 2). Based on the Lambert-Beer law, the content of Chl a could be calculated by the absorbance values at 445 nm and 663 nm, respectively. Chl a of P. tricornutum, Nannochloropsis oceanica, Mychonastes afer, and tobacco leaf were obtained by thin-layer chromatography (TLC). The Chl a concentration was calculated by the absorbance at 663 nm based on Arnon's method (ε 663 nm = 82.04; Formula (2)) [27]. In addition, the light absorption of the samples was measured at 445 nm and a calibration curve was established between the results for the two wavelengths ( Figure 3). The extinction coefficient of Chl a is 66.8 at 445 nm with the concentration range from 1 to 7 mg·L −1 .          According to the literature [28,29], the extinction coefficient of fucoxanthin at 449 nm is 1600, with a concentration of 1% (w/v, g/100 mL). The extinction coefficient at 445 nm was corrected to 156.54 (with a concentration of 1 g·L −1 ) using the purified standard as a sample over the concentration range from 0.005 to 0.05 mg·L −1 .
According to the additivity law of absorbance, the A 445 of the pigment extract is the absorbance of fucoxanthin and Chl a at 445 nm (Formula (1); a 1 is the extinction coefficient for fucoxanthin; and, a 2 is the extinction coefficient for Chl a), because they are the main absorbing components at 445 nm in P. tricornutum.
The concentration of Chl a is the absorption value at 663 nm divided by the extinction coefficient. The Chl a absorption value at 445 nm is the extinction coefficient at 445 nm multiplied by the Chl a concentration (Formula (2)). Because we desire to calculate the content of fucoxanthin, Formula (2) was rearranged into the form of Formula (3). When the numerical values for the extinction coefficients were placed into Formula (3), we could further simplify the formula to Formula (4). In the process, we transformed C fuc to C fuc so the results of the equation would be in mg·L −1 .
Using Formula (4), we calculated the fucoxanthin content of the culture using culture extracted by ethanol and cell debris removed by centrifugation prior to reading the samples in the spectrophotometer. However, this process was not accurate enough (data not shown), so we needed to optimize the formula further.
We tried to calculate the fucoxanthin concentration in whole cells using the absorbance value of cell culture at 445 nm and 663 nm. However, fucoxanthin is insoluble in the culture medium, and its extinction coefficients in water and ethanol are different. The A 445 and A 663 of algal cells suspension in f/2 medium or in ethanol showed no strong correlation coefficients ( Figure 4). The absorbance values of algae suspension in culture medium (ASC) and algae suspension in ethanol (ASE) were collected. The coefficient of determination of the absorbance (R 2 ) at 750 nm is 0.366 between ASC and ASE, they also showed significant difference (p = 0.004; Figure 4c). On the other hand, as mentioned before, the A 445 and A 663 of the ASC also exhibited significant differences from ASE (R 2 445 nm = 0.618; p 445 nm = 0.003; R 2 663 nm = 0.733; p 663 nm = 0.041; Figure 4a,b). Therefore, if we simplify this method by measuring the data in one situation, we are unable to obtain a correct result.
The previous results suggested that we had to determine the absorbance of algal cells in ethanol at 445 nm and 663 nm with some correction for the interference from other pigments and/or cell debris. Therefore, we needed to remove the "background noise" present in the whole-cell suspensions at 445 nm and 663 nm, which were represented by n1 and n2 in the Formula (5), respectively. We took an experimental approach to determine the values of n1 and n2 in formula 5. In theory, the "background noise" would be a function of the number of cells used to make the measurement. We tested this by determining whether there was a correlation between the number of cells (using A750 as a proxy) and cell debris with all of the pigments, except fucoxanthin and Chl a. We extracted the pigments from cells, saving the cell debris. After Chl a and fucoxanthin were separated out, all of the other pigments (including Chl c, diatoxanthin, and β-carotene) and the saved cell debris were resuspended with ethanol and measured at 445 nm and 663 nm. With the absorption value at 750 nm as ordinate, and the absorption value of other pigments and cell debris at 445 nm and 663 nm as abscissa, respectively, the regression curves shown in Figure 5 were obtained. We took an experimental approach to determine the values of n1 and n2 in formula 5. In theory, the "background noise" would be a function of the number of cells used to make the measurement. We tested this by determining whether there was a correlation between the number of cells (using A 750 as a proxy) and cell debris with all of the pigments, except fucoxanthin and Chl a. We extracted the pigments from cells, saving the cell debris. After Chl a and fucoxanthin were separated out, all of the other pigments (including Chl c, diatoxanthin, and β-carotene) and the saved cell debris were The equations for the regression lines shown in Figure 5 represent the "background noise" interfering at 445 nm and 663 nm. We can substitute the equations into Formula (5), which results in Formula (6). This can be further simplified to Formula (7). Using Formula (7), the concentration of fucoxanthin in algal cells could be determined through measuring the absorbance of cell culture at 750 nm and algal cell suspension in ethanol at 445 nm and 663 nm ( Figure 6A).
Cfuc' = 6.39 × A445 − 5.18 × A663 + 0.312 × A750 − 5.27 (7) In order to verify the correctness of Formula (7), P. tricornutum cultures at a cell density of 2 × 10 7 -1 × 10 8 cells·mL −1 at different growth stages were selected for fucoxanthin assay by HPLC and spectrophotometry. Regression analysis and paired t-test of these two results were carried out. The statistical analysis of these data derived from above two methods showed in early stage the standard errors were 0.80-4.56%, in middle stage the standard errors were 1.38-4.73%, in late stage the standard errors were 0.23-4.70% ( Figure 6B). Therefore, the spectroscopic method has enough accuracy to calculate the fucoxanthin concentration of P. tricornutum culture throughout its growth cycle. The equations for the regression lines shown in Figure 5 represent the "background noise" interfering at 445 nm and 663 nm. We can substitute the equations into Formula (5), which results in Formula (6). This can be further simplified to Formula (7). Using Formula (7), the concentration of fucoxanthin in algal cells could be determined through measuring the absorbance of cell culture at 750 nm and algal cell suspension in ethanol at 445 nm and 663 nm ( Figure 6A).
In order to verify the correctness of Formula (7), P. tricornutum cultures at a cell density of 2 × 10 7 -1 × 10 8 cells·mL −1 at different growth stages were selected for fucoxanthin assay by HPLC and spectrophotometry. Regression analysis and paired t-test of these two results were carried out. The statistical analysis of these data derived from above two methods showed in early stage the standard errors were 0.80-4.56%, in middle stage the standard errors were 1.38-4.73%, in late stage the standard errors were 0.23-4.70% ( Figure 6B). Therefore, the spectroscopic method has enough accuracy to calculate the fucoxanthin concentration of P. tricornutum culture throughout its growth cycle. The x-axis is fucoxanthin production determined by HPLC, the y-axis is fucoxanthin production determined by spectrophotometry. Different colored squares represent different growth stages of the culture. The green area represents the early stage (from 1st day to 3rd day); the blue area represents the middle stage (from 4th day to 6th day); and the orange area represents the late stage (from 7th day to 9th day). The black, green, blue and orange lines represent the curve fit for the total data set, the early stage data set, middle stage data set and late stage data set, respectively. R 2 is the coefficient of determination. The p value is calculated by paired t-test. All of the experiments were repeated three times and expressed as mean ± standard deviation (SD).
In order to further improve the efficiency of testing fucoxanthin with our spectroscopic method, a multi-mode microplate reader (Synergy™ HT, BioTek, Winooski, VT, USA) was used, in which 96 samples could be analyzed in 10 min for one batch. It was noted that the volume of samples in wells of 96-well plate was 200 μL and the light path was about 0.5 cm. The absorption should be revised by the calibration coefficient of 0.504, 0.497, and 0.444 at A445, A663, and A750, respectively (Figure 7). The x-axis is fucoxanthin production determined by HPLC, the y-axis is fucoxanthin production determined by spectrophotometry. Different colored squares represent different growth stages of the culture. The green area represents the early stage (from 1st day to 3rd day); the blue area represents the middle stage (from 4th day to 6th day); and the orange area represents the late stage (from 7th day to 9th day). The black, green, blue and orange lines represent the curve fit for the total data set, the early stage data set, middle stage data set and late stage data set, respectively. R 2 is the coefficient of determination. The p value is calculated by paired t-test. All of the experiments were repeated three times and expressed as mean ± standard deviation (SD).
In order to further improve the efficiency of testing fucoxanthin with our spectroscopic method, a multi-mode microplate reader (Synergy™ HT, BioTek, Winooski, VT, USA) was used, in which 96 samples could be analyzed in 10 min for one batch. It was noted that the volume of samples in wells of 96-well plate was 200 µL and the light path was about 0.5 cm. The absorption should be revised by the calibration coefficient of 0.504, 0.497, and 0.444 at A 445 , A 663 , and A 750 , respectively (Figure 7).

Discussion
Diatoms are a potentially important resource for the production of fucoxanthin. In order to screen for high-fucoxanthin producing strains, a high throughput method for measuring the concentration of fucoxanthin is required. Based on the properties of pigments and statistical analysis, Formula (7) was obtained, which can be efficiently used to calculate the content of fucoxanthin in cultures of P. tricornutum.
In developing this method, we wanted to use ethanol to extract the pigment, but the extinction coefficient used for calculating Chl a is based on 80% acetone [27,30]. In order to ensure that the extinction coefficient determined in 80% acetone could be used with our method, we compared the absorbance value (at 663 nm) of Chl a isolated from different plant or algal material in either 80% acetone or ethanol. The results for the samples in different solvents were analyzed by paired t-test, and all of the samples show no significant differences (p = 0.063). Based on that result, we were convinced that we could use the classic extinction coefficient in our formulas.
When detecting the content of pigment using the HPLC, we found that the extraction efficiency was dependent on whether the cells are washed or not, as well as the wash solution used. We tested three methods to wash the cells after centrifugation prior to extracting the pigments: the first consisted in ddH2O, the second in f/2 culture medium, and the third was a control group that was not washed (simply resuspended in the supernatant in the tube after centrifugation). Our results show that washing with ddH2O yielded the highest extraction efficiency (Figure 8). This is probably because washing with ddH2O started lysing the cells, improving pigment extraction. Based on these results, we used ddH2O washed cells to compare the spectrophotometric method against the HPLC method.

Discussion
Diatoms are a potentially important resource for the production of fucoxanthin. In order to screen for high-fucoxanthin producing strains, a high throughput method for measuring the concentration of fucoxanthin is required. Based on the properties of pigments and statistical analysis, Formula (7) was obtained, which can be efficiently used to calculate the content of fucoxanthin in cultures of P. tricornutum.
In developing this method, we wanted to use ethanol to extract the pigment, but the extinction coefficient used for calculating Chl a is based on 80% acetone [27,30]. In order to ensure that the extinction coefficient determined in 80% acetone could be used with our method, we compared the absorbance value (at 663 nm) of Chl a isolated from different plant or algal material in either 80% acetone or ethanol. The results for the samples in different solvents were analyzed by paired t-test, and all of the samples show no significant differences (p = 0.063). Based on that result, we were convinced that we could use the classic extinction coefficient in our formulas.
When detecting the content of pigment using the HPLC, we found that the extraction efficiency was dependent on whether the cells are washed or not, as well as the wash solution used. We tested three methods to wash the cells after centrifugation prior to extracting the pigments: the first consisted in ddH 2 O, the second in f/2 culture medium, and the third was a control group that was not washed (simply resuspended in the supernatant in the tube after centrifugation). Our results show that washing with ddH 2 O yielded the highest extraction efficiency (Figure 8). This is probably because washing with ddH 2 O started lysing the cells, improving pigment extraction. Based on these results, we used ddH 2 O washed cells to compare the spectrophotometric method against the HPLC method. Fucoxanthin is widely found in diatoms so we tested our method with two other diatoms: C. muelleri and T. pseudonana (both centric diatoms). We did a direct test without modifying Formula (7). The results demonstrate that the formula can measure the concentration of fucoxanthin in these two diatom strains with good confidence ( Table 1). The standard error for the measurement is larger than what we saw for P. tricornutum, presumably due to a lower fucoxanthin content. Therefore, in a rough comparison, this spectroscopic method could be applied to screen other diatom species, as well as being used to quickly detect the fucoxanthin content in production. Due to the differences in pigment distribution, cell size, and other cell components at different grow stages (as well as between pennate and centric diatoms), we expected that there would be significant differences within the growth cycle as well as between different species of diatoms that would impact the accuracy of Formula (7). However, we did not see a significant difference among different growth stages of P. tricornutum ( Figure 6). We did see an increase in the error associated with the centric diatoms. It is possible that developing species-specific measurements of the "background noise" and substituting them in for the P. tricornutum values could improve the measurements. In comparison with the HPLC method of fucoxanthin detection, the spectrophotometric method described here simplified the operation process and saved time without markedly affecting accuracy. Fucoxanthin is widely found in diatoms so we tested our method with two other diatoms: C. muelleri and T. pseudonana (both centric diatoms). We did a direct test without modifying Formula (7). The results demonstrate that the formula can measure the concentration of fucoxanthin in these two diatom strains with good confidence ( Table 1). The standard error for the measurement is larger than what we saw for P. tricornutum, presumably due to a lower fucoxanthin content. Therefore, in a rough comparison, this spectroscopic method could be applied to screen other diatom species, as well as being used to quickly detect the fucoxanthin content in production. Due to the differences in pigment distribution, cell size, and other cell components at different grow stages (as well as between pennate and centric diatoms), we expected that there would be significant differences within the growth cycle as well as between different species of diatoms that would impact the accuracy of Formula (7). However, we did not see a significant difference among different growth stages of P. tricornutum ( Figure 6). We did see an increase in the error associated with the centric diatoms. It is possible that developing species-specific measurements of the "background noise" and substituting them in for the P. tricornutum values could improve the measurements. In comparison with the HPLC method of fucoxanthin detection, the spectrophotometric method described here simplified the operation process and saved time without markedly affecting accuracy.

Microalgae and Cultivation
Algal strains P. tricornutum, Mychonastes afer, and Nannochloropsis oceanica IMET1 were provided by the center of BioEnergy Culture Collection (CBECC) in Qingdao Institute of Bioenergy and Bioprocess Technology. Chaetoceros muelleri (CCMP 1316) and Thalassiosira pseudonana (CCMP 1335) were purchased from National Center for Marine Algae and Microbiota (NCMA, East Boothbay, ME, USA). P. tricornutum and N. oceanica IMET1 were grown photoautotrophically in modified f/2 medium [31], with increased concentration of sodium nitrate (1 g·L −1 ). M. afer was culture in BG-11 medium. Cells were cultured at 23 • C in 100 mL bubble columns (30 cm in height, 4 cm in diameter, 100 mL medium) at a photon flux density of 80 µmol photons·m −2 ·s −1 with a 12:12-h light/dark photoperiod.

Extraction and Purification of the Fucoxanthin
The P. tricornutum cells in different culture stages were centrifuged at 4000× g, then rinsed with ddH 2 O, and recollected by centrifugation. Pellets were suspended in ethanol for pigment extraction (ethanol:algae culture volume = 1:1; v/v). In agreement with others, we found ethanol to be the most effective solvent in the extraction of fucoxanthin, with the extraction yield being ethanol > acetone > ethyl acetate [20]. The extraction system was incubated at 45 • C for 2 h, and mixed by vortex mixer every half an hour. Finally, the pigment solution was separated by centrifugation at 4000× g. The visible spectrum of the pigment solution was obtained by scanning from 400 to 800 nm with a spectrometer (Perkin Elmer UV-VIS Spectrometer Lambda 25, Waltham, MA, USA).
Purification of the fucoxanthin was carried out by using solid phase extraction (SPE) columns (Agilent Bond Elut HF Mega BE-SI, 5 mg 20 mL, Santa Clara, CA, USA). The pigment extracts were dried under nitrogen and resuspended in the mobile phase (n-hexane:acetone = 6:4) [6]. The total pigments were loaded on the SPE columns, then eluted by the mobile phase. Chl a eluted first, followed by fucoxanthin.
Pigments purity was checked using silica plates (Merck TLC Silica gel 60, Darmstadt, Germany) using hexane:acetone = 6:4 as the mobile phase. The pigment spots were detected visually, scraped from the plate, and then resuspended in ethanol for spectrophotometric analysis.
Quantification of fucoxanthin by HPLC was accomplished using a HITACHI Primaide HPLC system (HITACHI, Tokyo, Japan) with a C 18 reverse phase column (2.7 µm particle size, 100 × 4.6 mm). The mobile phase consisted of acetonitrile and water with a flow rate of 1 mL·min −1 . After loading the column with the fucoxanthin extract in ethanol, the mobile phase was an acetonitrile:water solution with the ratio increasing from 80:20 to 100:0 over 8 min, maintained at 100:0 for 3 min, and then decreased back to 80:20 over 5 min. The chromatogram was recorded at 445 nm. Fucoxanthin standards (ChromaDex, fucoxanthin (P), ASB-00006296-010, Irvine, CA, USA) were used for the construction of standard curve in the concentration range of 0.01-1 mg·mL −1 .

Spectrophotometric Assay
The extinction coefficient of the Chl a at 445 nm was calculated using a published extinction coefficient at 663 nm (ε 663 nm = 82.04) [30] using Chl a prepared from various sources, including tobacco leaf, N. oceanica, and P. tricornutum. The absorbance values at 445 nm and 663 nm were measured and a calibration curve was established. This allowed for us to estimate the extinction coefficient of Chl a at 445 nm.
For detecting the concentration of fucoxanthin in P. tricornutum using a spectrophotometry, the cultures were diluted with culture medium, and the absorbance measured at 750 nm (A 750 ranges from 0.1 to 0.8). In parallel, a volume of culture was centrifuged and the cells resuspended in an equal volume of ethanol, then the A 445 and A 663 -values were detected after dilution with ethanol (A 445 & A 663 range from 0.2 to 1). Samples were protected from light exposure as much as possible using foil.
The cells were suspended in ethanol and analyzed at A 445 and A 663 within 5 min. With this data, the concentration of fucoxanthin could be calculated using our formula.
A multi-mode microplate reader (Synergy TM HT, BioTek, Winooski, VT, USA) with 96-well plates was used to study the feasibility of high-throughput analysis following the method described above with the following modifications. The volume of samples in each well was 200 µL. We corrected to a 1 cm path-length using the software provided with the plate reader.

Statistical Analysis
Statistical significance of the values obtained from each experiment was evaluated by regression analysis and paired t-test using the software SPSS (version 19.0, IBM, Chicago, IL, USA). All of the experiments were repeated three times. Unless otherwise stated, all data were expressed as mean ± standard deviation (SD). The p values of less than 0.05 were considered statistically significant.

Conclusions
Fucoxanthin is a bioactive substances from marine with high economic value. Using microalgae for the fucoxanthin production enjoy great development and market potential. In this study, we developed an accurate and convenient method to test the concentration of fucoxanthin in diatoms by spectrophotometer. This method not only improve screening efficiency of diatom mutants, but also increase the monitoring efficiency of fucoxanthin content in the cultivation process.