Stereoisomers of Colourless Carotenoids from the Marine Microalga Dunaliella salina.

Carotenoids comprise a diverse range of naturally occurring stereoisomers, which differ in their physico-chemical properties. Their biosynthesis begins with phytoene, which is a rarity among carotenoids because it is colourless. Phytoene is sought after as a skin protectant against harmful UV range B (290–320 nm) and C (100–290 nm) light, and as a natural skin-whitening agent and is synthesized from geranylgeranyl diphosphate. Geranylgeranyl diphosphate is catalysed by phytoene synthase and phytoene desaturase to phytoene and phytofluene, respectively. The subsequent steps involve desaturation, isomerisation and cyclisation reactions to form α- and β-carotene stereoisomers, via all-trans lycopene. The marine microalga Dunaliella salina is the richest source of β-carotene, but it can accumulate phytoene and phytofluene as well. In the present study, different analytical tools including High-Performance Liquid Chromatography (HPLC), Ultra-Performance Convergence Chromatography (UPC2-MS) and Nuclear Magnetic Resonance (NMR) were used to characterize and quantify the phytoene isomeric configurations in D. salina in order to explore both the feasibility of D. salina as a cell factory for phytoene production and to gain new insight into the carotenoid synthesis pathway in D. salina. D. salina, similar to tomato, produced predominantly 15-cis phytoene isomer (>98%) and a trace amount of all-trans phytoene (<2%). High light stress, red light stress, or use of a phytoene desaturase inhibitor or a mitotic disrupter herbicide led to the accumulation of 15-cis phytoene but not all-trans phytoene. 9-cis phytoene was not detected in any of the extracts of D. salina biomass. Our main findings suggest that 15-cis phytoene is the most abundant isomer in D. salina and that it is subject to a series of isomerisation and desaturation reactions to form all-trans and 9-cis β-carotene.

The halotolerant green microalga Dunaliella salina is one of the richest sources of carotenoids, especially all-transand 9-cis β-carotene. However, it will also accumulate phytoene and phytofluene in the presence of a phytoene desaturase (PDS) inhibitor [13,14]. D. salina is easy to culture, as it requires light and a few other nutrients (nitrogen, phosphate and salts) to grow. For these reasons, it is considered a promising species to produce carotenoids [15].
Bleaching herbicides such as norflurazon are known to boost phytoene accumulation in the cells of D. salina and other higher plants, by inhibiting PDS, which prevents the conversion of phytoene to phytofluene [14,16,17]. However, little information has been provided on the isomer composition of the accumulated phytoene and available data to date are contradictory. Ebenezer [18] for example showed by using Nuclear Magnetic Resonance (NMR) methods that the main phytoene isomer, which accumulated in norflurazon-treated D. bardawil cultures (Dunaliella bardawil is a variant of Dunaliella salina), was the 15-cis isomer, whilst Werman et al. [6] and Ben-Amotz et al. [13] using High Performance Liquid Chromatography (HPLC) techniques, reported that the phytoene-rich D. bardawil powder used in their study contained the 9-cis phytoene isomer together with all-trans phytoene. This leads to an unresolved point in regard to the carotenoid biosynthetic pathway of D. salina: whether the high quantity of 9-cis β-carotene derives directly from 9-cis phytoene, as proposed in [13], or from the isomerization of all-trans β-carotene, as proposed by Davidi et al. [19] and by Xu and Harvey [14,20]. Resolving the identity of the phytoene isomers produced by D. salina could provide new insight into the carotenoid synthesis pathway and clarify this point.
The ratio of absorption peak heights at defined wavelengths, which defines fine structure, is useful for distinguishing carotenoids and their isomers and HPLC coupled with UV-Vis detection appears to be the most common method adopted for their study [21]. However, most studies have examined phytoene extracted from tomato. In ripening tomato, all-trans phytoene usually has a higher fine structure than 15-cis phytoene [22]. NMR analysis is necessary to discriminate at a structural level between different isomers. The values of specific chemical shifts for the cisand transisomeric forms of phytoene have been assigned in previous studies [18,23,24].
The main objective of this work was to resolve the phytoene isomers in D. salina and to gain new insight into the carotenoid synthesis pathway of D. salina. Methods of sample preparation from microalgal biomass and NMR methods were developed to characterize the phytoene isomers. The effects of cultivation conditions, light wavelength and light intensity, and herbicide treatment with two different classes of herbicides, chlorpropham and norflurazon, on carotenoid synthesis and phytoene isomer accumulation were also investigated.

Identification of Phytoene Isomers in D. salina with HPLC
In the present study, HPLC analysis was carried out to identify the phytoene isomers in D. salina cultures. Method parameters were tuned to reduce the shifts in retention time of the phytoene peaks among runs. The phytoene standard purchased from Sigma is composed of a mixture of 6 isomers ( Figure 2). Peak 3 of the standard matched with the phytoene isomer from tomato extract, which is widely considered as 15-cis phytoene. Peak 3 is also the closest peak to that from D. salina extracts, suggesting that phytoene accumulated in D. salina is most likely to be 15-cis phytoene as well. A flow rate of 0.5 mL/min under isocratic solvent conditions (MeOH/MTBE-88/10) was used to elute stereo-mutated tomato extract and phytoene standard to reveal the details of the elution peaks ( Figure 3). The tomato extract stereo-mutated showed the presence of a second peak eluting after the 15-cis isomer, which was identified as all-trans phytoene, in accordance with previous work by Melendez-Martinez et al. [22]. This peak matched with peak 5 of the phytoene standard and peak 2 of D. salina extract shown previously in Figure 2, suggesting they are all-trans phytoene isomers. Although both 15-cis and all-trans phytoene isomers gave absorption maxima at 275, 285 and 297 nm, they had slightly different absorption spectra ( Figure 4). 15-cis phytoene showed a loss in fine structure in comparison with the all-trans phytoene in all the extracts, which agrees with the work by Melendez-Martinez et al. [22] and Britton et al. [25]. A flow rate of 0.5 mL/min under isocratic solvent conditions (MeOH/MTBE-88/10) was used to elute stereo-mutated tomato extract and phytoene standard to reveal the details of the elution peaks ( Figure 3). The tomato extract stereo-mutated showed the presence of a second peak eluting after the 15-cis isomer, which was identified as all-trans phytoene, in accordance with previous work by Melendez-Martinez et al. [22]. This peak matched with peak 5 of the phytoene standard and peak 2 of D. salina extract shown previously in Figure 2, suggesting they are all-trans phytoene isomers. A flow rate of 0.5 mL/min under isocratic solvent conditions (MeOH/MTBE-88/10) was used to elute stereo-mutated tomato extract and phytoene standard to reveal the details of the elution peaks ( Figure 3). The tomato extract stereo-mutated showed the presence of a second peak eluting after the 15-cis isomer, which was identified as all-trans phytoene, in accordance with previous work by Melendez-Martinez et al. [22]. This peak matched with peak 5 of the phytoene standard and peak 2 of D. salina extract shown previously in Figure 2, suggesting they are all-trans phytoene isomers. Although both 15-cis and all-trans phytoene isomers gave absorption maxima at 275, 285 and 297 nm, they had slightly different absorption spectra ( Figure 4). 15-cis phytoene showed a loss in fine structure in comparison with the all-trans phytoene in all the extracts, which agrees with the work by Melendez-Martinez et al. [22] and Britton et al. [25]. Although both 15-cis and all-trans phytoene isomers gave absorption maxima at 275, 285 and 297 nm, they had slightly different absorption spectra ( Figure 4). 15-cis phytoene showed a loss in fine structure in comparison with the all-trans phytoene in all the extracts, which agrees with the work by Melendez-Martinez et al. [22] and Britton et al. [25].

Phytoene Accumulation in D. salina Treated by Chlorpropham or Norflurazon
Use of either chlorpropham or norflurazon leads to over production of phytoene in D. salina [13,14]. Here, we treated the cultures with 20 µM chlorpropham or 5 µM norflurazon and found that the percentage abundance of the 15-cis phytoene in total phytoene was 99.80 ±0.08, 99.80 ± 0.07 and 99.78 ± 0.07 for the control cultures, chlorpropham-treated cultures and norflurazon-treated cultures respectively. 15-cis phytoene was therefore the predominant isomer that accumulated with herbicide treatment ( Figure 5).

Analysis of Phytoene with UPC 2 -MS
The presence of the 15-cis phytoene in the norflurazon-and chlorpropham-treated D. salina extracts was evaluated with UPC 2 analysis coupled with mass spectrometry, using 15-cis phytoene extracted from tomato as a standard. Figure 6 shows the UPC 2 chromatograms of phytoene extracted from tomato and D. salina cultures treated with either chlorpropham or norflurazon. The chromatograms could be superimposed on one another (RT chlorpropham extract = 2.50 min; RT norflurazon

Phytoene Accumulation in D. salina Treated by Chlorpropham or Norflurazon
Use of either chlorpropham or norflurazon leads to over production of phytoene in D. salina [13,14]. Here, we treated the cultures with 20 µM chlorpropham or 5 µM norflurazon and found that the percentage abundance of the 15-cis phytoene in total phytoene was 99.80 ±0.08, 99.80 ± 0.07 and 99.78 ± 0.07 for the control cultures, chlorpropham-treated cultures and norflurazon-treated cultures respectively. 15-cis phytoene was therefore the predominant isomer that accumulated with herbicide treatment ( Figure 5).

Phytoene Accumulation in D. salina Treated by Chlorpropham or Norflurazon
Use of either chlorpropham or norflurazon leads to over production of phytoene in D. salina [13,14]. Here, we treated the cultures with 20 µM chlorpropham or 5 µM norflurazon and found that the percentage abundance of the 15-cis phytoene in total phytoene was 99.80 ±0.08, 99.80 ± 0.07 and 99.78 ± 0.07 for the control cultures, chlorpropham-treated cultures and norflurazon-treated cultures respectively. 15-cis phytoene was therefore the predominant isomer that accumulated with herbicide treatment ( Figure 5).

Analysis of Phytoene with UPC 2 -MS
The presence of the 15-cis phytoene in the norflurazon-and chlorpropham-treated D. salina extracts was evaluated with UPC 2 analysis coupled with mass spectrometry, using 15-cis phytoene extracted from tomato as a standard. Figure 6 shows the UPC 2 chromatograms of phytoene extracted from tomato and D. salina cultures treated with either chlorpropham or norflurazon. The chromatograms could be superimposed on one another (RT chlorpropham extract = 2.50 min; RT norflurazon

Analysis of Phytoene with UPC 2 -MS
The presence of the 15-cis phytoene in the norflurazon-and chlorpropham-treated D. salina extracts was evaluated with UPC 2 analysis coupled with mass spectrometry, using 15-cis phytoene extracted from tomato as a standard. Figure 6 shows the UPC 2 chromatograms of phytoene extracted from tomato

NMR Analysis of Phytoene Extracts from D. salina
In the present study, NMR analysis was carried out to prove that the main peak from D. salina detected with HPLC was 15-cis phytoene. Phytoene in a carotenoid-lipophilic mixture extracted from tomato was analysed with NMR, and chemical shifts values were assigned and compared to those identified in the mixture extracted from D. salina biomass. Figure 8a shows the results of 2D-HSCQ obtained from the tomato extract. Proton Chemical shifts of the conjugated central part C 14 , C 15 , C 15 , C 14 were as follow: δ h14,14 = 6.31 ppm, δ h15, 15 = 6.100. Carbon chemical shifts were H-C (14,14 ) 120.00, H-C (15, 15 ) 123.18. Chemical shifts values assigned were in accordance with the literature [18,23,24]. 2D-HSCQ NMR spectra of the alga biomass are shown in Figure 8b,c. The spectra of the tomato and the alga mixtures are superimposable. Proton chemical shifts were δ h14,14 = 6.31 ppm and δ h15, 15 = 6.10, and δ h14,14 = 6.31 ppm and δ h15, 15 = 6.10 ppm for the chlorpropham and norflurazon treated alga respectively. Carbon chemical shifts were H-C (14,14 ) 120.42, H-C (15,15 ) 123.60 for both the treated algae. These results show that in the presence of either chlorpropham or norflurazon, the alga Dunaliella salina accumulates 15-cis phytoene as the main isomer.

NMR Analysis of Phytoene Extracts from D. salina
In the present study, NMR analysis was carried out to prove that the main peak from D. salina detected with HPLC was 15-cis phytoene. Phytoene in a carotenoid-lipophilic mixture extracted from tomato was analysed with NMR, and chemical shifts values were assigned and compared to those identified in the mixture extracted from D. salina biomass. Figure 8a shows the results of 2D-HSCQ obtained from the tomato extract. Proton Chemical shifts of the conjugated central part C14, C15, C15′, C14′ were as follow: δ h14,14′ = 6.31 ppm, δ h15, 15′ = 6.100. Carbon chemical shifts were H-C (14, 14′) 120.00, H-C (15, 15′) 123.18. Chemical shifts values assigned were in accordance with the literature [18,23,24]. 2D-HSCQ NMR spectra of the alga biomass are shown in Figure 8b,c. The spectra of the tomato and the alga mixtures are superimposable. Proton chemical shifts were δ h14,14′ = 6.31 ppm and δ h15, 15′ = 6.10, and δ h14,14′ = 6.31 ppm and δ h15, 15′ = 6.10 ppm for the chlorpropham and norflurazon treated alga respectively. Carbon chemical shifts were H-C (14, 14′) 120.42, H-C (15, 15′) 123.60 for both the treated algae. These results show that in the presence of either chlorpropham or norflurazon, the alga Dunaliella salina accumulates 15-cis phytoene as the main isomer.

Colourless Phytofluene Isomers Produced by D. salina
Phytofluene isomeric forms were detected at the wavelength of 350 nm in HPLC-diode array analysis (Figure 9). Two phytofluene peaks were detected in tomato extracts. Peak 1 from tomato extracts matched with peak 1 of D. salina extracts (RT: (a) 16

Phytoene and Phytofluene Production in D. salina under Different Light Conditions
Previously, we have shown that light wavelengths affect the synthesis of carotenoids in D. salina. We found that the red light increased the cellular content of all major carotenoids and total carotenoids [14]. Figure 10 shows the concentration of phytoene isomers in D. salina cultures grown under red or blue light at different light intensities (200, 500 or 1000 µmol·m −2 ·s −1 ) for 48 h. The major phytoene isomer 15-cis phytoene in D. salina cultures changed with light wavelengths and light intensities, while no significant difference was found in the concentration of all-trans phytoene under all light conditions.

Phytoene and Phytofluene Production in D. salina under Different Light Conditions
Previously, we have shown that light wavelengths affect the synthesis of carotenoids in D. salina. We found that the red light increased the cellular content of all major carotenoids and total carotenoids [14]. Figure 10 shows the concentration of phytoene isomers in D. salina cultures grown under red or blue light at different light intensities (200, 500 or 1000 µmol·m −2 ·s −1 ) for 48 h. The major phytoene isomer 15-cis phytoene in D. salina cultures changed with light wavelengths and light intensities, while no significant difference was found in the concentration of all-trans phytoene under all light conditions.
Concentrations of 15-cis phytoene in the cultures increased with light intensity under both red and blue light, and generally showed higher concentrations under red than blue light. The ratio of 15-cis/all-trans phytoene increased with light intensity from 2.2 ± 0.1 to 5.9 ± 0.1 under red light and increased from 1.4 ± 0.2 to 3.9 ± 0.4 under blue light. A Pearson correlation analysis was carried out between the two phytoene isomers, and no significant correlation was found (p = 0.34). Phytofluene was measured at 355 nm using HPLC. Only one phytofluene isomer was detected in the carotenoid extracts from D. salina biomass, and its RT corresponded to that for phytofluene from tomato extracts. The concentration of phytofluene increased with light intensity under both red Concentrations of 15-cis phytoene in the cultures increased with light intensity under both red and blue light, and generally showed higher concentrations under red than blue light. The ratio of 15-cis/all-trans phytoene increased with light intensity from 2.2 ± 0.1 to 5.9 ± 0.1 under red light and increased from 1.4 ± 0.2 to 3.9 ± 0.4 under blue light. A Pearson correlation analysis was carried out between the two phytoene isomers, and no significant correlation was found (p = 0.34).
Phytofluene was measured at 355 nm using HPLC. Only one phytofluene isomer was detected in the carotenoid extracts from D. salina biomass, and its RT corresponded to that for phytofluene from tomato extracts. The concentration of phytofluene increased with light intensity under both red and blue light, with concentrations under red light more than double the concentrations under blue light ( Figure 11).

Discussion
Phytoene demand has increased recently due to its potential health benefits. In higher plants and fruits, such as tomato, phytoene is found predominantly in the 15-cis form. However, little work has been done to characterize the phytoene isomers in microalgae, such as D. salina, which is a rich source of natural carotenoids, especially β-carotene with a high ratio of the valuable 9-cis β-carotene. This work characterized phytoene isomers in D. salina and compared the production of phytoene in D. salina cultures treated with the two different classes of herbicides norflurazon and chlorpropham. After both treatments, the concentration of 15-cis phytoene increased. Moreover, the colourless phytofluene isomers over-accumulated in chlorpropham, but not norflurazon-treated D. salina cultures.
Results of HPLC analysis, UPC 2 analysis and NMR analysis each showed that 15-cis phytoene was the main phytoene isomer that accumulated in D. salina (>98% of total phytoene). Furthermore, the 15-cis phytoene increased with the treatment of both the herbicides norflurazon and chlorpropham and with increasing light intensity. Identification of 15-cis phytoene with only trace amounts of all-trans phytoene is in concord with findings for ripe tomato and higher plants [8,22] and with results obtained by Ebenezer et al. for D. bardawil [18], but not with the results obtained by Ben-Amotz et al. [13] and Werman et al. [6], who showed that both 9-cis phytoene and all-trans phytoene accumulated in D. bardawil treated with herbicides.
15-cis phytofluene has been considered by many to be the predominant phytofluene isomer (for a review see [4]). However, Koschmieder et al. [26] compared synthetic 15-cis phytofluene and alltrans phytofluene with phytofluene isomers from tomato and norflurazon-treated Dunaliella bardawil, and on the basis of RT obtained after HPLC analysis as well as absorption maxima at 331nm, 348 nm and 366 nm, concluded that phytofluene in tomato extract was predominately 9,15-di-cis phytofluene, which also exists in Dunaliella bardawil in a small proportion. They also determined that the main phytofluene isomer in norflurazon treated D. bardawil was 9-cis phytofluene but did not present the isomeric configurations in untreated D. bardawil cultures. In contrast, our results showed that the main phytofluene isomer in both untreated and norflurazon in chlorpropham treated D. salina was the same as in tomato extracts, which is 9,15-di-cis phytofluene. The 9-cis phytofluene shown by Koschmieder et al. [26] is likely not a phytofluene peak because norflurazon inhibits PDS function and blocks the desaturation of phytoene to subsequent phytofluene products.
The fact that only 15-cis phytoene accumulated in D. salina when treated with phytoene desaturase inhibitor, while no accumulation of all-trans phytoene was found, is consistent with 15-cis phytoene as the most likely precursor of phytofluene. All-trans phytoene present in D. salina is likely to be formed by the isomerisation of the predominant 15-cis phytoene and this reaction step along with the conversion of 15-cis phytoene to phytofluene might be fast enough to occur prior to the

Discussion
Phytoene demand has increased recently due to its potential health benefits. In higher plants and fruits, such as tomato, phytoene is found predominantly in the 15-cis form. However, little work has been done to characterize the phytoene isomers in microalgae, such as D. salina, which is a rich source of natural carotenoids, especially β-carotene with a high ratio of the valuable 9-cis β-carotene. This work characterized phytoene isomers in D. salina and compared the production of phytoene in D. salina cultures treated with the two different classes of herbicides norflurazon and chlorpropham. After both treatments, the concentration of 15-cis phytoene increased. Moreover, the colourless phytofluene isomers over-accumulated in chlorpropham, but not norflurazon-treated D. salina cultures.
Results of HPLC analysis, UPC 2 analysis and NMR analysis each showed that 15-cis phytoene was the main phytoene isomer that accumulated in D. salina (>98% of total phytoene). Furthermore, the 15-cis phytoene increased with the treatment of both the herbicides norflurazon and chlorpropham and with increasing light intensity. Identification of 15-cis phytoene with only trace amounts of all-trans phytoene is in concord with findings for ripe tomato and higher plants [8,22] and with results obtained by Ebenezer et al. for D. bardawil [18], but not with the results obtained by Ben-Amotz et al. [13] and Werman et al. [6], who showed that both 9-cis phytoene and all-trans phytoene accumulated in D. bardawil treated with herbicides.
15-cis phytofluene has been considered by many to be the predominant phytofluene isomer (for a review see [4]). However, Koschmieder et al. [26] compared synthetic 15-cis phytofluene and all-trans phytofluene with phytofluene isomers from tomato and norflurazon-treated Dunaliella bardawil, and on the basis of RT obtained after HPLC analysis as well as absorption maxima at 331nm, 348 nm and 366 nm, concluded that phytofluene in tomato extract was predominately 9,15-di-cis phytofluene, which also exists in Dunaliella bardawil in a small proportion. They also determined that the main phytofluene isomer in norflurazon treated D. bardawil was 9-cis phytofluene but did not present the isomeric configurations in untreated D. bardawil cultures. In contrast, our results showed that the main phytofluene isomer in both untreated and norflurazon in chlorpropham treated D. salina was the same as in tomato extracts, which is 9,15-di-cis phytofluene. The 9-cis phytofluene shown by Koschmieder et al. [26] is likely not a phytofluene peak because norflurazon inhibits PDS function and blocks the desaturation of phytoene to subsequent phytofluene products.
The fact that only 15-cis phytoene accumulated in D. salina when treated with phytoene desaturase inhibitor, while no accumulation of all-trans phytoene was found, is consistent with 15-cis phytoene as the most likely precursor of phytofluene. All-trans phytoene present in D. salina is likely to be formed by the isomerisation of the predominant 15-cis phytoene and this reaction step along with the conversion of 15-cis phytoene to phytofluene might be fast enough to occur prior to the interference of the chlorpropham herbicide action. As a second hypothesis, all-trans phytoene may also be formed directly from the precursor of geranylgeranyl pyrophosphate as suggested by Gregonis and Rilling [27]. An early study on a green microalga Scenedesmus obliquus also supports the idea that 15-cis phytoene is converted via 15-cis phytofluene and 15-cis ζ-carotene into all-trans-gz-carotene and trans-bicyclic carotenoids such as α-carotene and β-carotene [28]. Our results show that 15-cis phytoene is the main stereoisomer produced by D. salina and that it is over-accumulated with red light and with the use of herbicides. Our results also support the recent finding that 9-cis β-carotene is formed by isomerisation of all-trans β-carotene with the function of a 9-cis/all-trans β-carotene isomerase [19,20], rather than synthesised from the precursor of 9-cis phytoene.

Algal Cultivation
D. salina strain CCAP 19/41 (PLY DF15) was obtained from the Marine Biological Association, Plymouth, UK (MBA). Algae were cultured as described in previous work [14]. Inoculum was prepared by adding 10 mL of the culture stock of D. salina to 250 mL of fresh media until the culture reached the exponential phase. Algae were cultured in 500 mL Modified Johnsons Medium containing 1.5 M NaCl and 10 mM NaHCO 3 in an illuminated incubator (Varicon Aqua, Worcester, UK) under white light of 500 µmol·m −2 ·s −1 at 25 • C. To study the effect of light wavelength, algae were cultivated in Algem photobioreactor under red or blue light of 500 µmol·m −2 ·s −1 at 25 • C. To study the effect of herbicides, cultures were treated with either 5 µM of norflurazon or 20 µM of chlorpropham for 48 h before carotenoid extraction. All the experiments were performed in triplicate.

Carotenoids Extraction
The carotenoids were extracted from wet alga biomass as follows: the samples were centrifuged at 3000× g for 5 min at 5 • C; 10 mL MeOH-MtBE (80:20) were added to the pellets and sonicated and vortexed for 20 s; extracts were clarified at the centrifuge then filtered (0.20 µm filter) into amber HPLC vials before HPLC analysis.
15-cis phytoene was also extracted from ripe tomatoes (purchased from Waitrose and Partners, UK) to be used as control material because it favourably accumulates the compound at high concentration.

HPLC Analysis
High-Performance Liquid Chromatography equipped with Diode-Array Detection (HPLC-DAD; Agilent Technologies 1200 series, Agilent, Santa Clara, CA, United States), on-line degasser, a quaternary pump system, a YMC30 250 × 4.9 mm I.D S-5µ column (YMC, Europe GmbH) was used to resolve the phytoene and the phytofluene from the carotenoid extracts. The column temperature was set at 15 • C, the gradient solvent system was MeOH (A): MtBE (B) running at 80% A for the first 10 min at a flow rate of 1 mL/min, then increased to 100% A for the next 10 min at a flow rate of 0.5 mL/min before going back to the initial conditions at 20 min. Total run time was 45 min. The absorbance at four wavelengths (276, 282, 293 and 350) was monitored. All data were acquired and analysed with Chemstation for LC System software. Phytoene extracted from ripe tomatoes was stereo-mutated similarly to Melendez-Martinez et al. [29]. The sample extracted was heated at 75 • C for a total of 240 min and it was then injected in the HPLC system for analysis. Methods parameters were as follow: the column was set at 15 • C, the isocratic solvent system was MeOH (A): MtBE (B) running at 88% A at a flow rate of 0.5 mL/min.

UPC 2 Analysis
Supercritical, or near-critical, CO 2 was used as the primary eluent in the mobile phase with a MeOH cosolvent gradient to resolve the phytoene from the total-tomato and D. salina extracts using a seven-minute gradient. Supercritical fluid chromatography is very powerful when it comes to identifying compounds based on their retention time, as it allows shorter run times than HPLC and therefore a decrease in shifts in the retention time among runs. The UPC 2 (Waters Limited, Herts, UK) was equipped with a binary solvent delivery pump, auto-sampler, column oven, photodiode array detector and back-pressure regulator. Mobile phase A was CO 2 and mobile phase B was MeOH with 0.1% formic acid (v/v). The gradient was held at 5% B for 0.5 min, then increased to 50% B at 4.9 min, held at 50% for 0.1 min, ramped to 5% B in 0.1 min, and equilibrated for 1.9 min to give a total run time of 7 min. The binary solvent flow rate was 1.0 mL/min, the isocratic make-up solvent flow was 0.3 mL/min, and the UPC 2 was operated with a back-pressure of 2000 psi. The column oven was maintained at a temperature of 50 • C and an Acquity UPC 2 HSS C18 SB (1.8 µm, 3.0 mm × 100 mm) column was used. A sample injection volume of 3.0 µL was used with all samples. The UPC 2 was fitted with an Acquity PDA detector (190-800 nm) and operated in the wavelength range 210-600 nm with a 1.2 nm resolution. In addition, the absorbance at four wavelengths (276 nm, 282 nm, 293 nm and 450 nm) was monitored with a resolution of 1.2 nm.

Mass Spectrometry Analysis
All mass spectrometry experiments were collected on a Waters Synapt G2 Q-ToF mass spectrometer (Manchester, UK). The instrument was operated in positive ion mode electrospray with a capillary voltage of 2.0 kV and sampling cone voltage was 30 V. Nitrogen at a flow rate of 650 L/h was used as the desolvation gas with a constant desolvation temperature of 350 • C, a cone gas flow rate of 50 L/h, and a source temperature of 130 • C. Data were acquired over the m/z range 50-800. An integral LockSpray unit infusing Leucine-Enkephalin peptide into the electrospray sample stream was used to collect reference scans. Scans were performed every 10 s with the reference calibrant introduced at a flow rate of 10 µL/min using the fluidics system of the instrument. Single point lock-mass correction was used for the protonated pseudo-molecular ion at m/z 556.2771 (+ve). All data were acquired and analysed with Waters MassLynx v4.1 software (Manchester, UK).

NMR Sample Preparation
Fresh tomato samples and harvested D. salina cultures were freeze-dried and a total of 20 g of dried tomato powder and dried algae samples was used for the carotenoid extraction. The procedure described above was modified as follow: MtBE was chosen as solvent for extraction due to its lower polarity index (2.4), hence its higher selectivity for only carotenoids (and lipids) in comparison to the MeOH and ethanol solvents; moreover, its high volatility value allowed shorter evaporation times and minimisation of sample losses. The extracts were injected in the HPLC system to confirm the presence of phytoene. The MtBE extracts were evaporated for one and a half hour, flushed with nitrogen and suspended with deuterated chloroform prior to analysis.