Metabolomic profiling of the astaxanthin accumulation process induced by high light in Haematococcus pluvialis

Algal Research 20 (2016) 35–43

Contents lists available at ScienceDirect

Algal Research journal homepage: www.elsevier.com/locate/algal

Metabolomic profiling of the astaxanthin accumulation process induced by high light in Haematococcus pluvialis Hexin Lv ⁎, Feng Xia, Miao Liu, Xianggan Cui, Fazli Wahid, Shiru Jia ⁎ Key Laboratory of Industrial Fermentation Microbiology, Ministry of Education, Tianjin Key Lab of Industrial Microbiology, Tianjin University of Science and Technology, Tianjin 300457, PR China

a r t i c l e

i n f o

Article history: Received 18 December 2015 Received in revised form 13 September 2016 Accepted 19 September 2016 Available online xxxx Keywords: Haematococcus pluvialis Astaxanthin metabolism Gas chromatography–mass spectrometry Metabolomics Amino acid Fatty acid

a b s t r a c t The growth of Haematococcus pluvialis exposed to high light was divided into three pigmentation phases: a green phase, a yellow phase and a red phase. Classification was based on astaxanthin and chlorophyll content and cell morphology. Metabolic profiling of the three pigmentation phases was performed. A total of 81 metabolites were identified and quantified by gas chromatography–mass spectrometry, including 23 amino acids, 19 sugars, 15 organic acids, 8 alcohols, 7 amines, 4 nucleic acids and 5 others. These were subdivided into four groups according to their changes during the three phases. The clustering of metabolites was revealed, and potential biomarkers were identified by principal component analysis, partial least squares analysis and hierarchical clustering analysis, suggesting details of the metabolic pathways of cells. The contents of some cytoprotective metabolites were increased in the yellow phase, such as sucrose, proline and glutamic acid. The precursors of these metabolites are the intermediates of the Calvin cycle and the TCA cycle, indicated those two cycles provided more precursors for the synthesis of the cytoprotective metabolites. A hypothetical metabolic regulation model of H. pluvialis exposed to high light was proposed. Our study provides the first metabolomics view of the astaxanthin accumulation process that is induced by high light in H. pluvialis. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The xanthophyll carotenoid astaxanthin is a type of antioxidant that is mainly found in marine animals and in many microorganisms, including microalgae [1]. Astaxanthin is called ‘super vitamin E’ because of its strong antioxidant properties. It is estimated that astaxanthin shows ten times greater antioxidant activity than β-carotene, which is another commercialized antioxidant product [2]. Nutraceutical and pharmacological studies have shown that astaxanthin has anti-aging effects [3], pro-immunity effects [4], anti-cancer effects [5], and effects against cardiovascular disease [6] and diabetes [7]. The global market size of astaxanthin is estimated to reach $253 million by 2015 [8]. The unicellular green alga Haematococcus pluvialis can accumulate large quantities of astaxanthin under various stress conditions, including high light, nutrient starvation and osmotic stress. H. pluvialis cells can be divided into four types that occur during its life cycle, including stationary microzooids, large flagellated swarming macrozooids, nonmotile palmella forms and resting hematocysts [9]. Astaxanthin biosynthesis in H. pluvialis occurs in two steps: β-carotene synthesis occurs first, followed by production of astaxanthin through oxidation and hydroxylation reactions of β-carotene [10]. The key genes involved in astaxanthin synthesis have been cloned and metabolically engineered successfully into tobacco and other species to produce astaxanthin ⁎ Corresponding authors. E-mail addresses: [email protected] (H. Lv), [email protected] (S. Jia).

http://dx.doi.org/10.1016/j.algal.2016.09.019 2211-9264/© 2016 Elsevier B.V. All rights reserved.

[11,12]. However, the astaxanthin levels in metabolically engineered plants are too low for them to be used for industrial scale production. Therefore, further exploration and exploitation of the underlying regulatory mechanisms and components of the astaxanthin biosynthesis pathway in H. pluvialis are needed. Metabolomics is a powerful tool for systematically analyzing the dynamic responses to stimulation and small molecule metabolite changes in many organisms [13]. Recently, many studies on microalgae using metabolomics have been performed. For example, a metabolomics approach was used to study the metabolite responses and phytotoxic effects of the herbicide prometryn on the growth of the green alga Scenedesmus vacuolatus [14]. The diatom Skeletonema marinoi was studied in different growth phases by using comparative metabolomics [15]. Metabolomics analysis of H. pluvialis was used to study responses to stress factors including acetate (Ac), Fe2 + and high light stress [16]. Until now, metabolomic analysis of astaxanthin accumulation process in H. pluvialis has not been performed. In the present study of astaxanthin and chlorophyll contents, growth and astaxanthin accumulation in H. pluvialis were divided into three phases, i.e., a green phase with higher chlorophyll content and lower astaxanthin content (contents equal to those of microzooids and macrozooids), a yellow phase with intermediate levels of chlorophyll and astaxanthin (equal to those of the non-motile palmella forms), and a red phase with lower chlorophyll and higher astaxanthin (equal to the contents found in hematocysts). The metabolomics of the three phases induced by high light in H. pluvialis were profiled.

36

H. Lv et al. / Algal Research 20 (2016) 35–43

2. Materials and methods 2.1. Algal culture and stress conditions H. pluvialis were cultured in liquid Bold's basal medium (BBM) [17] in 500 mL Erlenmeyer flasks under a light intensity of 24 μmol/m2 s for 16 h/8 h light/dark cycle at 24 °C without aeration. Log phase cells (OD669 = 0.4) were cultured under a light intensity of 24 μmol/m2 s for 7 days (OD669 = 0.7) and collected as green phase cells, Yellow phase cells were generated by culturing green phase cells under a light intensity of 80 μmol/m2 s for 14 days, Red phase cells were generated by culturing green phase cells under a light intensity of 80 μmol/ m2 s for 28 days. All experiments were carried out in four replicates and repeated at least thrice in this study. 2.2. Analytical methods 2.2.1. Pigment analysis H. pluvialis cells were harvested by centrifuge at 4000 rpm for 5 min and washed with Milli-Q water. The dry cell weight of the algal biomass was measured repeatedly until a constant weight was obtained by drying at 90 °C in a hot-air oven. Pigments of algal biomass used for astaxanthin HPLC analyses were extracted by acetone. For astaxanthin saponification, 1 mL NaOH-CH3OH (0.1 M) was added into 5 mL pigment extract and kept at 55 °C for 12 h. The astaxanthin content was analyzed by HPLC using a reverse-phase BDS HYPERSIL C18 (250 × 4.6 mm, 5.0 μm). Samples were eluted using acetonitrile/methanol (75/25, v/v) as mobile phases. Elution was carried out at 1 mL/ min with a 20 μL injection volume loop with micro-syringe with a detection wavelength and a column temperature of 476 nm and 25 °C, respectively. Pigments of algal biomass used for chlorophyll and total carotenoids analyses were extracted with 5 mL DMSO for 12 h at room temperature in the dark. The contents of chlorophyll and total carotenoids were determined as follows [18]: Chl a (mg/L) = 12.19 A665– 3.45A649, Chl b (mg/L) = 21.99 A649–5.32A665, Total carotenoids (mg/ L) = (1000A480 −2.14Ca −70.16Cb) / 220. An analytical electronic balance (Mettler-Toledo, ME204, Greifensee, Switzerland) with an accuracy of ±0.0001 g was used for all weightings in the experiments.

2.2.4. Derivatization and gas chromatography–mass spectrometry (GC– MS) analysis Prior to GC–MS analysis, a two-stage chemical derivatization was performed. First, 50 μL of methoxamine hydrochloride (20 mg/mL in pyridine, Sigma) was added to the lyophilizate, and the mixture was incubated at 40 °C for 80 min. Then, 80 μL of N-methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA, Sigma) was added to the samples, which were subsequently incubated at 40 °C for 80 min [23,24]. GC–MS analysis was performed by using a GC–MS system (Agilent Technologies, Palo Alto, CA, USA) equipped with a HP-5 capillary column (60 m × 320 μm i.d., 0.25 μm film thickness; Agilent J&W Scientific, Folsom, CA, USA). 1 μL of sample was injected without a split ratio. Helium was used as the carrier gas at a constant flow rate of 1 mL/ min. The electron impact ionization (70 eV) was set to full scan mode (m/z: 50–800). The GC oven temperature for metabonomics was set to 70 °C for 2 min, then raised to 290 °C at a rate of 5 °C/min, and maintained at 290 °C for 6 min. The GC oven temperature for fatty acid methyl esters analysis was set to 70 °C for 2 min, followed by an increasing rate of 8 °C/min to 200 °C, which was held for 2 min, then increased at a rate 3 °C/min to 245 °C and maintained at 245 °C for 3 min. 2.2.5. Fatty acid methyl esters extract The fatty acid of three different phases were extracted according to a modified method [25]. Briefly, 10 mg algae powder was suspended in 2 mL 2 M NaOH-CH3OH solution and the 40 μL 2 mg/mL nonadecanoic acid (internal standard solution) was added and mixed thoroughly.

2.2.2. Microscopy H. pluvialis cells were observed under an optical microscope (OM) (Olympus BX53, Tokyo, Japan) with an Olympus DP72 digital color camera (Olympus, Tokyo, Japan). The morphology of the cell surface was investigated using a Hitachi SU1510 scanning electronic microscope (SEM) (Hitachi, Tokyo, Japan). For SEM observation, H. pluvialis cells were harvested by centrifugation at 4000 rpm for 5 min and washed with Milli-Q water. Cell fixation, dehydration and coating were performed according to previously reported methods [19]. 2.2.3. Extraction of intracellular metabolites The quenching and extraction processes were performed according to previously reported methods [20]. In detail, samples were immediately quenched with pre-chilled − 40 °C 60% (v/v, methanol/water) methanol solution for 5 min. Cells were harvested by centrifugation at 8000 rpm for 5 min at 4 °C and washed with PBS (pH 7.0,0.1 mM) twice, then washed with ultrapure water once. The cell pellets were immediately frozen in liquid nitrogen and ground into powder. The intracellular metabolites were extracted according to previously reported methods [21,22]. Briefly, 50 mg cell powder was suspended in 1 mL of pre-chilled − 20 °C extraction buffer (1:1, v/v, methanol/water) and 5 μL of internal standard solution (succinic acid, 2,2,3,3-d4, 1.5 mg/mL, Sigma), then thoroughly mixed by vortexing. The mixture was frozen in liquid nitrogen for 2 min and then thawed, and the cycle was repeated five times. The supernatant was collected by centrifugation at 10,000 rpm for 5 min and was then used for lyophilization.

Fig. 1. The contents of chlorophylls, total carotenoids, astaxanthin and microscopy of cells in three phases. The data shown are the averages ± SE of three replicates. **P b 0.01 compared with green phase. Changes in culture color of three pigmentation phases (b). Changes in cell color of three pigmentation phases (c). Changes in cell surface of three pigmentation phases, irregular concave features are marked by arrows (d). OM, optical microscope; SEM, scanning electron microscope. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

H. Lv et al. / Algal Research 20 (2016) 35–43

37

Table 1 Relative content of 81 metabolites and the ratio of the three pigment phases Metabolites contents are calculated from the comparisons of GC peak areas between metabolites and internal standard. aY/G: The ratio of yellow phase with green phase. bR/Y: The ratio of red phase with yellow phase. L-Valine 2 is the tautomeric form of L-Valine, likewise other metabolites with a number followed. ND: Not Detected. The values are means (±SE) of four biological replicates. Type

Name

Green Phase

Yellow Phase

Red Phase

Y/Ga

R/Yb

Amino acids

L-Valine

10.9 ± 1.1

25.0 ± 0.7

3.3 ± 0.6

2.3

0.1

L-Alanine

25.7 ± 2.7

97.5 ± 3.4

5.2 ± 0.9

3.8

0.1

L-Leucine

11.1 ± 1.7

20.5 ± 4.3

2.4 ± 0.3

1.9

0.1

L-Isoleucine

6.2 ± 1.4

12.5 ± 1.0

ND

2.0

ND

L-Norvaline

3.8 ± 0.6

ND

1.4 ± 0.4

ND

ND

L-Valine 2 Serine

5.4 ± 1.5

18.1 ± 4.9

ND

3.4

ND

20.6 ± 2.6 14.9 ± 0.6

44.6 ± 0.7 34.1 ± 2.5

20.2 ± 0.3 8.0 ± 0.3

2.2 2.3

0.5 0.2

4.9 ± 2.7 5.9 ± 2.3 9.1 ± 0.5

ND 13.5 ± 1.0 17.9 ± 1.2

5.0 ± 2.0 2.9 ± 0.9 3.7 ± 0.4

ND 2.3 2.0

ND 0.2 0.2

3.7 ± 0.5 2.6 ± 1.1

8.4 ± 2.5 8.0 ± 2.7

ND ND

2.3 3.0

ND ND

46.4 ± 2.9

87.7 ± 6.2

34.7 ± 2.9

1.9

0.4

L-Lysine

6.5 ± 0.8 5.9 ± 1.6 36.3 ± 6.0 12.9 ± 2.3 34.0 ± 2.3 5.6 ± 2.2 18.7 ± 0.8

ND 18.7 ± 0.4 46.5 ± 3.3 ND 25.4 ± 0.4 16.8 ± 4.8 63.1 ± 3.5

1.2 ± 0.2 ND 14.5 ± 1.7 ND ND ND 2.3 ± 0.4

ND 3.2 1.3 ND 0.7 3.0 3.4

ND ND 0.3 ND ND ND 0.1

L-Tyrosine

9.0 ± 1.7

21.3 ± 2.3

1.6 ± 0.2

2.4

0.1

L-Cystine

5.2 ± 0.2

25.9 ± 1.6

10.7 ± 0.9

5.0

0.4

DL-Arabinose

2.1 ± 0.2 3.4 ± 0.1

6.6 ± 0.4 19.8 ± 2.8

2.7 ± 0.0 4.0 ± 0.2

3.1 5.8

0.4 0.2

D-(+)-Galactopyranose

16.0 ± 1.0 ND

24.2 ± 2.4 23.4 ± 1.5

4.3 ± 0.1 ND

1.5 ND

0.2 ND

D-(−)-Fructose

29.6 ± 0.1

69.0 ± 2.9

13.6 ± 0.2

2.3

0.2

23.4 ± 0.4

60.6 ± 3.5

9.1 ± 0.4

2.6

0.2

D-Galactose

22.3 ± 1.0

29.4 ± 1.2

ND

1.3

ND

D-Glucose

59.0 ± 0.6

169 ± 7.9

24.7 ± 1.0

2.9

0.2 0.1

L-Threonine

Serine 2 N,O,O-Tris(trimethylsilyl)-L-threonine L-Aspartic acid β-Alanine L-Methionine L-Proline β-Alanine Phenylalanine Glutamic acid DL-Ornithine DL-Ornithine 2 N-α-Acetyl-L-Lysine

Sugars

D-(−)-Tagatofuranose Glycoside

D-(−)-Fructose

2

5.9 ± 0.2

9.9 ± 1.5

1.2 ± 0.1

1.7

D-Allose

7.9 ± 0.2

28.6 ± 2.3

4.5 ± 0.7

3.6

0.2

Glyceryl-glycoside TMS ether

57.6 ± 3.0 5.5 ± 0.7

106 ± 3.7 21.3 ± 0.2

5.7 ± 0.5 ND

1.8 3.9

0.1 ND

ND

ND

14.8 ± 0.7

ND

ND

3710 ± 750 12.0 ± 0.1

10,900 ± 790 10.8 ± 0.4

2580 ± 190 1.3 ± 0.2

2.9 0.9

0.2 0.1

15.8 ± 1.3 10.7 ± 0.9

52.2 ± 1.8 32.5 ± 1.0

36.3 ± 1.7 17.4 ± 1.8

3.3 3.0

0.7 0.5

D-Galactose

2

D-Galactofuranose D-Mannose Sucrose D-(+)-Turanose 3-α-Mannobiose D-(+)-Trehalose

Organic acids

D-(+)-Cellobiose Galactinol Propanoic acid Phosphoric acid 2-Butenedioic acid Butanoic acid Pentanedioic acid 2-Propenoic acid L-Threonic acid 1,4-Benzenedicarboxylic acid 3,5-Dioxa-4-phospha-2-silaoctan-8-oic acid 1,2,3-Propanetricarboxylic acid L-Ascorbic

acid

D-Gluconic

acid Hexadecanoic acid 9,12-Octadecadienoic acid Octadecanoic acid Alcohols

L-Threitol Xylitol L-(−)-Arabitol

Ribitol D-Mannitol D-Sorbitol Myo-Inositol D-Myo-Inositol

5.2 ± 0.1

39.5 ± 0.8

5.4 ± 0.4

7.6

0.1

ND 22.8 ± 9.8 1090 ± 13 4.1 ± 0.2 ND 3.2 ± 0.7 ND ND

42.8 ± 1.6 101 ± 1.2 3420 ± 340 7.5 ± 0.1 84.9 ± 2.7 5.9 ± 0.1 ND 5.4 ± 0.3

ND 20.5 ± 0.6 387 ± 14 2.6 ± 0.1 ND 3.2 ± 0.5 4.7 ± 0.8 ND

ND 4.4 3.1 1.8 ND 1.9 ND ND

ND 0.2 0.1 0.4 ND 0.5 ND ND

23.5 ± 2.0 ND ND 10.8 ± 1.0

81.3 ± 0.6 ND 21.8 ± 2.2 ND

13.7 ± 0.2 18.5 ± 1.6 113 ± 7.3 ND

3.5 ND ND ND

0.2 ND 5.2 ND

1.3 ± 0.1

4.8 ± 0.5

ND

3.6

ND

99.5 ± 2.9 25.7 ± 0.6 96.5 ± 19 6.6 ± 0.3

273 ± 9.0 42.4 ± 1.4 199 ± 28 20.6 ± 1.2

67.0 ± 2.4 4.2 ± 0.4 66.7 ± 2.7 3.2 ± 0.1

2.8 1.7 2.1 3.1

0.3 0.1 0.3 0.2

2.8 ± 0.2 3.0 ± 0.3

11.5 ± 0.2 14.0 ± 0.8

2.6 ± 0.3 4.2 ± 0.8

4.1 4.7

0.2 0.3

ND 5.9 ± 0.1

ND 12.3 ± 1.1

1.7 ± 0.3 7.0 ± 0.0

ND 2.1

ND 0.6

ND

10.4 ± 0.4

3.6 ± 0.4

ND

0.4

25.9 ± 4.2 5.5 ± 0.1

87.9 ± 16 21.7 ± 1.7

18.8 ± 0.5 3.9 ± 0.5

3.4 4.0

0.2 0.2

(continued on next page)

38

H. Lv et al. / Algal Research 20 (2016) 35–43

Table 1 (continued) Type

Name

Green Phase

Yellow Phase

Red Phase

Y/Ga

R/Yb

Amines

Urea Tris(hydroxymethyl) aminomethane N-{4-[Bis(trimethylsilyl)amino] butyl}acetamide

3.1 ± 0.6 2.3 ± 0.2 8.6 ± 1.0 5.0 ± 0.6

9.8 ± 1.1 7.7 ± 0.1 28.4 ± 1.2 10.2 ± 0.6

ND ND 19.0 ± 0.2 1.2 ± 0.1

3.1 3.3 3.3 2.1

ND ND 0.7 0.1

8.0 ± 1.0

20.2 ± 0.7

ND

2.5

ND

2.9 ± 0.1 2.2 ± 0.2 1.2 ± 0.1 1.2 ± 0.3 50.1 ± 2.9 16.0 ± 1.7 22.7 ± 0.3 22.4 ± 0.6 36.9 ± 2.3 2.6 ± 0.1 ND

5.7 ± 0.9 12.1 ± 1.3 3.2 ± 0.1 ND 125 ± 11 75.2 ± 4.9 68.3 ± 2.1 64.4 ± 4.5 111 ± 13 26.6 ± 1.3 ND

ND ND 0.8 ± 0.0 0.7 ± 0.2 3.4 ± 0.3 15.1 ± 3.7 16.6 ± 0.1 13.5 ± 1.6 6.7 ± 0.2 2.4 ± 0.1 3.2 ± 0.4

2.0 5.6 2.7 ND 2.5 4.7 3.0 2.9 3.0 10.3 ND

ND ND 0.3 ND 0.1 0.2 0.2 0.2 0.1 0.1 ND

L-Asparagine

Nucleic acids

Others

L-Asparagine 2 N-Acetyl glucosamine methoxime N-Acetyl-D-glucosamine Pyrimidine Pyrazine N,9-bis(Trimethylsilyl)-6-[(trimethylsilyl)oxy]-9H–purin-2-amine Adenosine 3,7-Dioxa-2,8-disilanonane Pentasiloxane Silanamine Hexopyranose 2-Monopalmitoylglycerol trimethylsilyl ether

The mixture was incubated at 75 °C for 30 min. Two mL 12 M HCl was added into the mixture (pH b 2.0) and spiked with 2 mL chloroformmethanol solution (0.01% BHT, 2:1, v/v). The extract was ultrasonicated for 30 min. The supernant was discarded and lower solution was collected. 1 mL of 10% HCl-CH3OH was added into the mixture and then incubated at 75 °C for 40 min. Fatty acid methyl esters were extracted with 1 mL hexane and vortexed for 2 min. After centrifuge at 4000 rpm for 2 min, the hexane phase was used for further GC–MS analysis. 2.2.6. Data analysis and statistical analysis Typical total ion chromatograms (TIC) of the H. pluvialis cells were obtained by GC–MS. The MSD Productivity ChemStation software (version E.02.01.1177 Agilent) was used for data analyses. To screen for valid metabolites, the initial GC peak width and initial threshold were set to 0.1 and 15.0, respectively, in our study. The National Institute of Standards and Technology (NIST) mass spectral library 2011 (version 2.0 g) were used for metabolite identification, the match degrees were set at 700. Unsupervised principal component analysis (PCA) was used for initial analysis. Supervised partial least-squares (PLS) analysis was used to further verify the differences and identify the metabolites responsible for distinguishing the green, yellow and red growth phases. Both supervised PLS and unsupervised PCA were performed by SIMCA package (ver. 11.5) (Umetrics, Umea, Sweden). Score plots of PLS and PCA were used to previewing the clustering effect. Loading plots of PCA and PLS were used for finding biomarkers. HCE 3.5 software (Human-computer interaction lab, University of Maryland) was used for unsupervised hierarchical cluster analysis (HCA), the metabolite levels of three pigmentation phase were normalized by default firstly, then average linkage method were adopted for hierarchical clustering. Data were statistically analyzed with SPSS20.0 statistical software packages (IBM). One-way ANOVA followed by the Least Significant Difference Test in the post hoc analysis was used in this study. Differences with a P value of b0.05 were considered statistically significant. The relative contents of metabolites and fatty acids were calculated according to a previous method [26].

decreased gradually from 16.7 mg/g to 0.9 mg/g as the culture time progressed. Compared to the green phase, astaxanthin and total carotenoids both increased gradually during the transition to the red phase, increasing by 14.6-fold and 8.5-fold, respectively. The changes in photosynthetic and accessory pigments indicate that the metabolism of cells during astaxanthin accumulation shifted from active photosynthesis to a resting state. In addition, along with the transition from the green to the red phase, the color of cultures turned to yellow and red from green gradually when subjected to high light (Fig. 1b), cell volumes also increased gradually (Fig. 1c). In the green phase, cell diameter was approximately 18.8 μm, and the cell diameters of the yellow phase and red phase were very similar (statistics from 100 cells: 25.3 μm and 29.8 μm, respectively). The surfaces of yellow and red phase cells were more irregular and concave than those of green phase cells (Fig. 1d). This indicates that the cell structures also change during astaxanthin accumulation. 3.2. Metabolomic characterization of the three pigmentation phases of H. pluvialis 3.2.1. Multivariate data analysis In the present study, GC–MS and a two-stage chemical derivatization were used for metabolomics analysis of the three pigmentation phases [23,24]. Thermally stable and volatile compounds with molecular weight generally b1000 Da could be detected, mainly including amino acids, organic acids, fatty acids, sugars and so on, but not biological macromolecules such as proteins and carotenoids. In the total ion chromatogram, N 260 chromatographic peaks were detected in each sample. To screen for valid metabolites, the initial GC peak width and initial threshold were set to 0.1 and 15.0, respectively. By comparing with the NIST library and filtering for match degrees N 700, we identified a total of 81 metabolites, which included 23 amino acids, 19 sugars, 15 organic acids, 8 alcohols, 7 amines, 4 nucleic acids and 5 other compounds (Table 1). To study changes in the intracellular metabolites across the three pigmentation phases, multivariate statistical analysis was performed by using PCA and PLS. The results show that both models were well constructed, with excellent fit and satisfactory predictive ability (Table 2). The unsupervised clustering method PCA was used to

3. Results and discussion 3.1. Identification of three growth phases of H. pluvialis The collected cells from three pigmentation phases were analyzed for changes in the contents of chlorophylls, total carotenoids and astaxanthin (Fig. 1a). The content of chlorophyll a in the green phase, yellow phase and red phase were 17.3 mg/g, 6.2 mg/g and 10 mg/g, respectively. A similar trend was observed for chlorophyll b, which

Table 2 Statistical data from PCA and PLS at the three pigment phases. PCA

PLS

R2X

Q2

R2X

R2Y

Q2

0.992

0.505

0.992

0.991

0.987

H. Lv et al. / Algal Research 20 (2016) 35–43

identify and rank major sources of variance within the three data sets. Based on similarities and differences in the measured parameters, PCA was able to cluster biological samples into both expected and unexpected groups. Samples in different phases were separated clearly on the PCA score plot (Fig. 2a). The first principal component (PC1) accounted for 92.5% of the total variance among the three experimental groups. A PCA plot of PC1 vs. PC2 (these two components account for 99.2% of the variance) clearly separates the three phases. These results show that the intracellular metabolic profiles of H. pluvialis cells differ significantly between pigmentation phases. The supervised clustering method PLS was used to further validate the differences between the three pigmentation phases. The PLS results also support clear differences between the three pigmentation phases (Fig. 2c), Taken together, the PCA and PLS analyses indicate that there is systemic variation in intracellular metabolism during the astaxanthin accumulation process.

39

To identify the metabolites responsible for distinguishing different pigmentation phases, PCA loading plots were used to analyze the contribution of each metabolite to the principal components. The contributions of data points were evaluated based on their distances from the origin point [27]. As shown in Fig. 2b, the potential biomarkers identified by PCA loading plots were sucrose, phosphoric acid, D-glucose, hexadecanoic acid, 1, 2, 3-propanetricarboxylic acid, octadecanoic acid and glyceryl-glycoside. Like the PCA loading plots, the PLS loading plots showed the main metabolites responsible for distinguishing cells in different pigmentation phases (Fig. 2d). These metabolites were sucrose, phosphoric acid, octadecanoic acid, D-glucose, D-(−)-fructose, 1,2,3-propanetricarboxylic acid and hexadecanoic acid. The VIP coefficients reflect the contribution of each metabolite to the PLS models; a higher VIP value indicates that the metabolite has a larger contribution. The results from VIP coefficients analyses show that four potential

Fig. 2. PCA score plot (a), PCA loading plot (b), PLS score plot (c), PCA loading plot (d) and VIP results of samples from H. pluvialis in the three pigmentation phases. In the score plots, the confidence interval is defined by Hotelling's T2 ellipse (95% confidence interval), and observations outside the confidence ellipse are considered outliers. The relative intensity of each metabolite peak is expressed as the ratio of the peak area to that of the internal standard. Normalized peak areas were imported into SIMCA-P for multivariate statistical analysis.

40

H. Lv et al. / Algal Research 20 (2016) 35–43

biomarkers made a relatively high contribution to both the PCA and PLS, which were identified as sucrose, phosphoric acid, 1,2,3propanetricarboxylic acid and hexadecanoic acid (Fig. 2e). 3.2.2. Hierarchical cluster analysis Cluster Analysis, also called data segmentation, is used to group or segment a collection of objects into subsets or “clusters”. In this work, the levels of metabolites contents of cells were normalized firstly and hierarchical clustering was adopted to reveal the degree of dissimilarity between the three groups from the PCA and PLS analysis. The results showed that the three sets of samples are very distinct (Fig. 3). Furthermore, the green phase and red phase can be classified into one category because their metabolic profiles are more similar to each other than to the yellow phase. This result implicated that the levels of most metabolites in cells of yellow phase were changed when subjected to high light and returned to similar levels of green phase when cells entered red phase. Additionally, as revealed by cluster analysis, all 81 metabolites

could be divided into multiple clusters. However, they all belonged to the 4 largest clusters (C1, C2, C3 and C4). Cluster 1 includes 12 amino acids, 7 sugars, 5 organic acids, 3 alcohols, 2 amines, 2 nucleic acids and 1 other compound, and it includes the biomarkers sucrose and phosphoric acid. Cluster 2 includes 7 amino acids, 10 sugars, 7 organic acids, 4 alcohols, 4 amines, 1 nucleic acid and 3 other compounds, including the biomarker hexadecanoic acid. In both C1 and C2, the metabolite contents increase before becoming reduced in the yellow phase. The initial contents of C1 compounds are higher than those of C2 compounds. Clusters 3 include 4 amino acids, 1 organic acid, 1 amine, and 1 nucleic acid. Cluster 4 includes 1 sugar, 3 organic acids, 1 alcohol and 1 other compound, and it includes 1,2,3-propanetricarboxylic acid. The trend observed in cluster C3 differs from the trends of C1 and C2. In cluster C4, the contents of metabolites increase gradually. The coordinated changes of metabolites in the same cluster implicated that the underlying relationships among the metabolic pathways. The results of HCA also validate the predictive accuracy of the PCA models.

Fig. 3. HCA results of samples from H. pluvialis in the three pigmentation phases. Four replicates of each pigmentation phase were hierarchical clustered by HCE software. The color of clusters indicates the normalized levels of metabolites contents of cells from three pigmentation phases. Cluster 1 includes 12 amino acids, 7 sugars, 5 organic acids, 3 alcohols, 2 amines, 2 nucleic acids, 1 other compound. Cluster 2 includes 7 amino acids, 10 sugars, 7 organic acids, 4 alcohols, 4 amines, 1 nucleic acid, and 3 other compounds. Cluster 3 includes 4 amino acids, 1 organic acid, 1 amine, and 1 nucleic acid. Cluster 4 includes 1 sugar, 3 organic acids, 1 alcohol, and 1 other compound. The change trends of each cluster were indicated on the left. The Y-axis indicates the normalized levels of metabolites contents of cells. GP, Green phase; YP, Yellow phase; RP, Red phase. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

H. Lv et al. / Algal Research 20 (2016) 35–43

3.2.3. Changes related to the intracellular metabolites in different pigmentation phases From the green phase to the yellow phase, i.e., when cells were exposed to stress conditions, the contents of most metabolites increased significantly (Fig. 4). Sucrose was the most important of these metabolites. It contributed more than other metabolites to defining the three pigmentation phases of H pluvialis. Sucrose is the major product of photosynthesis in plants, and it has important roles in carbohydrate storage and stress responses [28]. For example, Krasensky and Jonak reported that plants accumulate large amounts of carbohydrates, including sucrose, under stress conditions [29]. Similar to sucrose, some amino acids in microbial cells also play important roles in stress tolerance; they accumulate when cells are exposed to stresses such as oxidative environments, freezing injury, and unfavorable temperature [30–32]. In this work, we detected all 23 amino acids, except L-norvaline, serine 2, alanine, L-ornithine and L-ornithine 2, the other 18 amino acids were all increased when cells entered the yellow phase (Fig. 5a). Proline and glutamic acid are common cytoprotective metabolites. Proline is synthesized from glutamic acid, and it can also be converted to glutamic acid [33]. These amino acids are involved in cell metabolism under a variety of stress conditions, and they enhance the stability of the cell membrane [34–36]. However, there are many metabolites that protect cells from stress, and we found that the cell surface exhibits irregular, concave features. This is a phenomenon that occurs after culture under stress conditions. The flagellum of the cell disappears, rendering the cell unable to move, and cell volume increases gradually. The cell

41

turns red as it gradually produces astaxanthin. These results are consistent with other studies [37]. It has been reported that the contents of amino acids are directly related to their metabolic precursors [38]. In the present study, synthetic precursors of most of the detected amino acids are the intermediates of the Calvin cycle and the TCA cycle (Fig. 4). For instance, glyceraldehyde 3-phosphate (G3P) is the precursor of serine and threonine. G3P is an important intermediate of the Calvin cycle [39]. The precursor of leucine, isoleucine and valine is pyruvic acid [40]. Alanine and pyruvic acid can be interconverted. Pyruvic acid is an important intermediate of the TCA cycle. The precursor of proline and glutamic acid is αketoglutarate, which is also an important intermediate of the TCA cycle. Therefore, increased levels of sucrose and amino acids indicate that Calvin cycle and TCA cycle provided more precursors for other pathways. The ATP and NADPH required for the synthesis of these metabolites come from light captured by photosynthesis [41]. Therefore, we inferred the increased light exposure before cells enter the yellow phase is the major reason for increased supply of precursors from the Calvin cycle and TCA cycle, and for the accumulation of metabolites for cell protection. A hypothetical metabolic regulation model that occurs during astaxanthin accumulation in H. pluvialis was proposed based on the present metabolomics data (Fig. 6), i.e., high light lead to the accumulation of the intermediates of Calvin cycle and TCA cycle, G3P, pyruvic acid and α-ketoglutarate. This further promotes the accumulation of intracellular cytoprotective metabolites, such as sucrose, proline and glutamic acid.

Fig. 4. Diagram showing changes in metabolite abundance mapped onto the metabolic network, including the Calvin cycle, the TCA cycle, amino acid metabolic pathway, fatty acid metabolic pathway and other pathways. Undetected metabolites are shown in boxes. G3P: Glyceraldehyde 3 Phosphate; PEP, phosphoenolpyruvate; OAA, oxaloacetate; α-KG, αketoglutarate. The data shown are the averages ± SE of four replicates.

42

H. Lv et al. / Algal Research 20 (2016) 35–43

Fig. 5. Amino acids and fatty acids identified in H. pluvialis in the three pigmentation phases. SFA: saturated fatty acid; UFA: unsaturated fatty acid. F1: Tridecanoic acid (C13:0); F2: Pentadecanoic acid (C15:0); F3: Hexadecanoic acid (C16:0); F4: Heptadecanoic acid (C17:0); F5: Octadecanoic acid (C18:0); F6: Eicosanoic acid (C20:0); F7: Docosanoic acid (C22:0); F8: 4,7,10-Hexadecatrienoic acid (C16:3); F9: 7,10-Hexadecadienoic acid (C16:2); F10: 7,10,13-Hexadecatrienoic acid (C16:3); F11: 9-Hexadecenoic acid (C16:1); F12: 6,9,12Octadecatrienoic acid (C18:3); F13: 9,12-Octadecadienoic acid (C18:2); F14: 9,12,15-Octadecatrienoic acid (C18:3); F15: 11-Octadecenoic acid (C18:1); F16: Cyclopropaneoctanoic acid (C17:1); F17: 5,8,11,14-Eicosatetraenoic acid (C20:4); F18: 7,10,13-Eicosatrienoic acid (C20:3); F19: 10,13-Eicosadienoic acid (C22:2). TFA: Total fatty acids. The data shown are the averages ± SE of four replicates. *P b 0.05 compared with green phase;**P b 0.01 compared with green phase. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

When cells enter the red phase from the yellow phase, astaxanthin gradually accumulates. When this happens, the contents of sucrose and amino acids decrease. The contents of total fatty acids in the red phase sharply increase by 3.4 and 4.3-fold compared to the green and yellow phases, respectively (Fig. 5b). The highest content of astaxanthin occurs in the red phase. In H. pluvialis, fatty acid synthesis and accumulation esterify astaxanthin, and oil droplets can accommodate more astaxanthin esters [42]. In fatty acid metabolism, we have detected 19 fatty acids (Fig. 5b). In the green phase, although astaxanthin yield was just 0.8 mg/g, fatty acids were still synthesized. Thus, under stress conditions, astaxanthin accumulation and fatty acid synthesis become synchronized. Fatty acid synthesis for astaxanthin accumulation seems to be necessary, but fatty acid accumulation is not dependent on astaxanthin synthesis [43]. Another study showed that astaxanthin and fatty acid biosynthesis were feedback-coordinated at the metabolite level. In vivo and in vitro experiments indicated that astaxanthin

esterification drove the formation and accumulation of astaxanthin [44]. In the latter stress culture, cells increased their volume drastically and entered a resting state. The cell surface showed more irregular concave features than in the initial stress culture. This suggests that astaxanthin is a product of H. pluvialis that protects it against a stressful environment. Our results are consistent with those of S. Boussiba, who reported that the accumulation of astaxanthin is accompanied by increased cell volume and a resting state [37]. 4. Conclusion In this study, GC–MS based metabolomics was used to analyze astaxanthin accumulation process induced by high light exposure in H. pluvialis, and a total of 81 metabolites were identified and quantified. In different pigmentation phases, the intracellular metabolism of H. pluvialis change significantly. When light intensity increased, the Calvin cycle and TCA cycle provided more precursors for other pathways, and the contents of various metabolites increased significantly. When incubation time and the consumption of nutrients were increased, the primary metabolite was used for the synthesis of other metabolites, and those metabolites were significantly reduced. In fatty acid metabolism, the fatty acid contents of the red phase were the highest among the three phases, and the astaxanthin content in the red phase was also the highest of the three phases. Authors' contributions SJ and HL conceived and designed the project. HL and FX analyzed the data and wrote the paper. ML performed the cultures materials preparation and revised the paper. XC and FW participated in metabolism analysis. All authors have read and approved the final manuscript. Acknowledgments

Fig. 6. The possible metabolic regulation model of H. pluvialis Undetected metabolites are shown in boxes. G3P: glyceraldehyde 3 Phosphate; α-KG: α-ketoglutarate; Ast: astaxanthin.

This work was supported by the National Natural Science Foundation of China (No. 31401029) and the Foundation (No. 2015IM101) of Key Laboratory of Industrial Fermentation Microbiology of Ministry of Education and Tianjin Key Lab of Industrial Microbiology (Tianjin University of Science & Technology). There are no conflicts of interest to declare.

H. Lv et al. / Algal Research 20 (2016) 35–43

References [1] I. Higuera-Ciapara, L. Felix-Valenzuela, F. Goycoolea, Astaxanthin: a review of its chemistry and applications, Crit. Rev. Food Sci. Nutr. 46 (2006) 185–196. [2] R.T. Lorenz, G.R. Cysewski, Commercial potential for Haematococcus microalgae as a natural source of astaxanthin, Trends Biotechnol. 18 (2000) 160–167. [3] K. Suganuma, H. Nakajima, M. Ohtsuki, G. Imokawa, Astaxanthin attenuates the UVA-induced up-regulation of matrix-metalloproteinase-1 and skin fibroblast elastase in human dermal fibroblasts, J. Dermatol. Sci. 58 (2010) 136–142. [4] B.P. Chew, J.S. Park, Carotenoid action on the immune response, J. Nutr. 134 (2004) 257S–261S. [5] D.N. Tripathi, G.B. Jena, Astaxanthin intervention ameliorates cyclophosphamide-induced oxidative stress, DNA damage and early hepatocarcinogenesis in rat: role of Nrf2, p53, p38 and phase-II enzymes, Mutat. Res. 696 (2010) 69–80. [6] F.J. Pashkow, D.G. Watumull, C.L. Campbell, Astaxanthin: a novel potential treatment for oxidative stress and inflammation in cardiovascular disease, Am. J. Cardiol. 101 (2008) S58–S68. [7] R. Otton, D.P. Marin, A.P. Bolin, C. Santos Rde, T.G. Polotow, S.C. Sampaio, M.P. de Barros, Astaxanthin ameliorates the redox imbalance in lymphocytes of experimental diabetic rats, Chem. Biol. Interact. 186 (2010) 306–315. [8] I. Schmidt, H. Schewe, S. Gassel, C. Jin, J. Buckingham, M. Hümbelin, G. Sandmann, J. Schrader, Biotechnological production of astaxanthin with Phaffia rhodozyma/ Xanthophyllomyces dendrorhous, Appl. Microbiol. Biotechnol. 89 (2011) 555–571. [9] R.T. Lorenz, A technical review of Haematococcus algae, Naturose Technical Bulletin, 60, 1999. [10] N. Misawa, H. Shimada, Metabolic engineering for the production of carotenoids in non-carotenogenic bacteria and yeasts, J. Biotechnol. 59 (1998) 169–181. [11] T. Gerjets, G. Sandmann, Ketocarotenoid formation in transgenic potato, J. Exp. Bot. 57 (2006) 3639–3645. [12] V. Mann, M. Harker, I. Pecker, J. Hirschberg, Metabolic engineering of astaxanthin production in tobacco flowers, Nat. Biotechnol. 18 (2000) 888–892. [13] O. Fiehn, Metabolomics-the link between genotypes and phenotypes, Plant Mol. Biol. 48 (2002) 155–171. [14] C. Kluender, F. Sans-Piché, J. Riedl, R. Altenburger, C. Härtig, G. Laue, M. SchmittJansen, A metabolomics approach to assessing phytotoxic effects on the green alga Scenedesmus vacuolatus, Metabolomics 5 (2009) 59–71. [15] C. Vidoudez, G. Pohnert, Comparative metabolomics of the diatom Skeletonema marinoi in different growth phases, Metabolomics 8 (2012) 654–669. [16] Y. Su, J. Wang, M. Shi, X. Niu, X. Yu, L. Gao, X. Zhang, L. Chen, W. Zhang, Metabolomic and network analysis of astaxanthin-producing Haematococcus pluvialis under various stress conditions, Bioresour. Technol. 170 (2014) 522–529. [17] U. Tripathi, R. Sarada, S.R. Rao, G. Ravishankar, Production of astaxanthin in Haematococcus pluvialis cultured in various media, Bioresour. Technol. 68 (1999) 197–199. [18] A.R. Wellburn, The spectral determination of chlorophylls a and b, as well as total carotenoids, using various solvents with spectrophotometers of different resolution, J. Plant Physiol. 144 (1994) 307–313. [19] Z. Gao, C. Meng, X. Zhang, D. Xu, X. Miao, Y. Wang, L. Yang, H. Lv, L. Chen, N. Ye, Induction of salicylic acid (SA) on transcriptional expression of eight carotenoid genes and astaxanthin accumulation in Haematococcus pluvialis, Enzym. Microb. Technol. 51 (2012) 225–230. [20] S.G. Villas-Bôas, J. Højer-Pedersen, M. Akesson, J. Smedsgaard, J. Nielsen, Global metabolite analysis of yeast: evaluation of sample preparation methods, Yeast 22 (2005) 1155–1169. [21] S. Lu, J. Wang, Y. Niu, J. Yang, J. Zhou, Y. Yuan, Metabolic profiling reveals growth related FAME productivity and quality of Chlorella sorokiniana with different inoculum sizes, Biotechnol. Bioeng. 109 (2012) 1651–1662. [22] W. Weckwerth, K. Wenzel, O. Fiehn, Process for the integrated extraction, identification and quantification of metabolites, proteins and RNA to reveal their co-regulation in biochemical networks, Proteomics 4 (2004) 78–83.

43

[23] S.G. Villas-Boas, M. Akesson, J. Nielsen, Biosynthesis of glyoxylate from glycine in Saccharomyces cerevisiae, FEMS Yeast Res. 5 (2005) 703–709. [24] W. de Koning, K. van Dam, A method for the determination of changes of glycolytic metabolites in yeast on a subsecond time scale using extraction at neutral pH, Anal. Biochem. 204 (1992) 118–123. [25] A. Lei, H. Chen, G. Shen, Z. Hu, L. Chen, J. Wang, Expression of fatty acid synthesis genes and fatty acid accumulation in Haematococcus pluvialis under different stressors, Biotechnol. Biofuels 5 (2012) 18. [26] M. Liu, C. Zhong, X.Y. Wu, Y.Q. Wei, T. Bo, P.P. Han, S.R. Jia, Metabolomic profiling coupled with metabolic network reveals differences in Gluconacetobacter xylinus from static and agitated cultures, Biochem. Eng. J. 101 (2015) 85–98. [27] M. Jalali-Heravi, S. Masoum, P. Shahbazikhah, Simulation of 13 C nuclear magnetic resonance spectra of lignin compounds using principal component analysis and artificial neural networks, J. Magn. Reson. 171 (2004) 176–185. [28] M.W. Vaughn, G.N. Harrington, D.R. Bush, Sucrose-mediated transcriptional regulation of sucrose symporter activity in the phloem, Proc. Natl. Acad. Sci. 99 (2002) 10876–10880. [29] J. Krasensky, C. Jonak, Drought, salt, and temperature stress-induced metabolic rearrangements and regulatory networks, J. Exp. Bot. 63 (2012) 1593–1608. [30] H. Li, M.L. Ma, S. Luo, R.M. Zhang, P. Han, W. Hu, Metabolic responses to ethanol in Saccharomyces cerevisiae using a gas chromatography tandem mass spectrometrybased metabolomics approach, Int. J. Biochem. Cell Biol. 44 (2012) 1087–1096. [31] S. Jozefczuk, S. Klie, G. Catchpole, J. Szymanski, A. Cuadros-Inostroza, D. Steinhauser, J. Selbig, L. Willmitzer, Metabolomic and transcriptomic stress response of Escherichia coli, Mol. Syst. Biol. 6 (2010) 364. [32] H.J. Kang, H.J. Yang, M.J. Kim, E.S. Han, H.J. Kim, D.Y. Kwon, Metabolomic analysis of meju during fermentation by ultra performance liquid chromatography-quadrupole-time of flight mass spectrometry (UPLC-Q-TOF MS), Food Chem. 127 (2011) 1056–1064. [33] M.C. Brandriss, D.A. Falvey, Proline biosynthesis in Saccharomyces cerevisiae: analysis of the PRO3 gene, which encodes delta 1-pyrroline-5-carboxylate reductase, J. Bacteriol. 174 (1992) 3782–3788. [34] H. Takagi, K. Sakai, K. Morida, S. Nakamori, Proline accumulation by mutation or disruption of the proline oxidase gene improves resistance to freezing and desiccation stresses in Saccharomyces cerevisiae, FEMS Microbiol. Lett. 184 (2000) 103–108. [35] H. Takagi, Proline as a stress protectant in yeast: physiological functions, metabolic regulations, and biotechnological applications, Appl. Microbiol. Biotechnol. 81 (2008) 211–223. [36] X.J. He, K.E. Mulford, J.S. Fassler, Oxidative stress function of the Saccharomyces cerevisiae Skn7 receiver domain, Eukaryot. Cell 8 (2009) 768–778. [37] S. Boussiba, Carotenogenesis in the green alga Haematococcus pluvialis: cellular physiology and stress response, Physiol. Plant. 108 (2000) 111–117. [38] C.A. Rodrigues-Pousada, T. Nevitt, R. Menezes, D. Azevedo, J. Pereira, C. Amaral, Yeast activator proteins and stress response: an overview, FEBS Lett. 567 (2004) 80–85. [39] C.A. Raines, The Calvin cycle revisited, Photosynth. Res. 75 (2003) 1–10. [40] R. Devantier, B. Scheithauer, S.G. Villas-Bôas, S. Pedersen, L. Olsson, Metabolite profiling for analysis of yeast stress response during very high gravity ethanol fermentations, Biotechnol. Bioeng. 90 (2005) 703–714. [41] B.B. Buchanan, W. Gruissem, R.L. Jones, Biochemistry & Molecular Biology of Plants, American Society of Plant Physiologists, Rockville, MD, 2000. [42] M. Zhekisheva, S. Boussiba, I. Khozin-Goldberg, A. Zarka, Z. Cohen, Accumulation of oleic acid in Haematococcus pluvialis (Chlorophyceae) under nitrogen starvation or high light is correlated with that of astaxanthin esters, J. Phycol. 38 (2002) 325–331. [43] M. Zhekisheva, A. Zarka, I. Khozin-Goldberg, Z. Cohen, S. Boussiba, Inhibition of astaxanthin synthesis under high irradiance does not abolish triacylglycerol accumulation in the green alga Haematococcus pluvialis (Chlorophyceae), J. Phycol. 41 (2005) 819–826. [44] G. Chen, B. Wang, D. Han, M. Sommerfeld, Y. Lu, F. Chen, Q. Hu, Molecular mechanisms of the coordination between astaxanthin and fatty acid biosynthesis in Haematococcus pluvialis (Chlorophyceae), Plant J. 81 (2015) 95–107.