Three different microalgal species Haematococcus pluvialis (H), Desmodesmus subspicatus (D) and Chlorella variabilis (C) were selected due to their different cell wall and pigment composition. The cell wall of C. variabilis is comprised predominantly of chitin, glycosaminoglycans and sialic acid (39, 40), H. pluvialis has cellulose and non-hydrolysable sporopollenin-like material (41) and D. subspicatus has a pecto-cellulosic and glycoproteic cell wall (42). Typically, the pigment composition of H. pluvialis consists of high concentrations of astaxanthin, canthaxanthin, β-carotene and lutein (43) . Both C. variabilis and D. subspicatus also have similar pigment composition that varies depending on the culture conditions but both lack astaxanthin (44, 45).
An optimised LC-MS method based on (11, 46) was initially developed, followed by an examination of different cell lysis techniques and finally a comparison of five commonly used solvent compositions was carried out to test their extraction efficiency across a range of pigments. (Table 1). The five solvent extraction systems tested were 100 % methanol (method 1), 90% acetone (method 2), ethanol:hexane (2:1) (method 3), chloroform:methanol:water 8: 4: 3 (method 4) and ethanol:hexane (1:1) (method 5).
The optimised procedure for the three key steps in LC-MS analysis of algal pigments included cellular lysis, solvent extraction and chromatographic separation is presented with a particular focus on the impact that these extraction solvents have on the overall pigment profile. The method can also be extended to include lipid analysis so data presented on lipid extraction efficiency has also been included.
S.No.
|
Extraction solvent
|
Run time (min)
|
Column used
|
Detector used
|
Pigments detected
|
Reference
|
1
|
Methanol
|
15
|
C8
|
UHPLC-MS/MS
|
14
|
(11)
|
2
|
Acetone
|
42
|
C18 PFP
|
HPLC-PDA
|
70
|
(46)
|
3
|
Ethanol/hexane (2:1)
|
20
|
C18
|
UPLC-UV-MS
|
37
|
(18)
|
4
|
Chloroform/methanol/water*
|
-
|
-
|
-
|
-
|
(47)
|
5
|
Ethanol/hexane (1:1) +10% Na2SO4
|
50
|
C30
|
HPLC/APCI-MS/MS
|
7
|
(10)
|
Table 1. Comparison of extraction techniques used in this study; *classical Folch method that has no chromatography involved
Eighteen pigments were selected for measurement using the multiple reaction monitoring (MRM) described in a method reported elsewhere (11) . The MRM collision energies were optimised using standards of each compound (Table 2). Standards for the following pigments were used for quantification (astaxanthin, canthaxanthin, zeaxanthin, lutein, chl-a and chl-b) while relative values were used for the other pigments. Two key lipids classes were also chosen phosphatidylcholine (PC) and triacylglycerol (TAG). Using only 1 µl and 25 mins run time separation of 18 pigments was achieved (Fig 1). Some peaks are not fully visible as the concentration is very low compared to others.
PIGMENT
|
RETENTION
TIME
|
PRECURSOR ION (m/z)
|
PRODUCT ION (m/z)
|
COLLISION ENERGY
|
Peridinin
|
4.8
|
631.4
|
553.4
|
10
|
Nile Red (IS)
|
6.7
|
319
|
275
|
35
|
Neoxanthin
|
8.2
|
601.4
|
221.0
|
18
|
Fucoxanthin
|
8.3
|
659.5
|
109
|
26
|
Violaxanthin
|
8.4
|
601.4
|
221.0
|
18
|
Prasinoxanthin
|
8.5
|
601.4
|
583.4
|
10
|
Astaxanthin
|
8.6
|
597.4
|
147.1
|
25
|
Alloxanthin
|
9.3
|
565.5
|
157.1
|
23
|
Zeaxanthin
|
9.9
|
569.4
|
145
|
29
|
Lutein
|
10.1
|
569.4
|
339.1
|
15
|
Canthaxanthin
|
10.5
|
565.4
|
203.1
|
25
|
Diadinoxanthin
|
10.7
|
583.5
|
157.1
|
34
|
Chlorophyll b
|
12.1
|
907.5
|
569.1
|
38
|
19-but-fucoxanthin
|
12.7
|
745.5
|
109
|
31
|
Chlorophyll a
|
13.1
|
893.5
|
555.0
|
38
|
Cryptoxanthin
|
13.2
|
552.4
|
460.4
|
25
|
DV Chlorophyll a
|
13.3
|
891.5
|
553.0
|
38
|
Lycopene
|
15.3
|
537.4
|
444.4
|
25
|
19-hex-fucoxanthin
|
15.7
|
773.5
|
109
|
30
|
beta-carotene
|
15.8
|
536.4
|
444.4
|
25
|
Table 2. Phytoplankton pigments and internal standard with their MRMs for analysis
Optimised chromatographic separation for tandem mass spectrometry analysis
Reversed phase columns ranging from C8 – C30 bonded phases have been employed in pigment analysis (48). The C30 columns provide the highest selectivity, however, the higher hydrophobicity results in the requirement for stronger non-polar mobile phase solvents and longer run times to facilitate elution (49). Both Fraser et al. (50) and Soares et al. (10) reported efficient separation of pigments in a single run but with analysis time of 42 minutes and 50 minutes, respectively. In their study, Gupta et al. resolved the issue of longer analysis time by separating 15 major carotenoids in 20 minutes (51). Using C18 column, Chauveau-Duriot et al. provided method for separation of carotenoids with a 45-minute run (52). The issue was again the longer run time, whereas Fu et al. could separate 37 pigments in 20 minutes using C18 column (18).
In this study, various mobile phases were tested starting with 0.1% formic acid in water (A) and 0.1% formic acid in methanol (B). Solvent system consisting of isopropanol (50%) and methanol (50%) (B) was also tested. Finally, the mobile phase consisting of 10 mM ammonium formate in water (A) and 75% isopropanol, 20% acetonitrile and 5% water (10 mM ammonium formate) (B) was found to be the most suitable as it facilitated the ionisation of pigments and non-polar lipids (triacylglycerols) as ammonium adducts (Fig. 1). The details of the final method used are in the method section below with runtimes of less than 30 mins. The use of a mass selective detection provides a second dimension for separation therefore complete resolution of each pigment is not as necessary.
Cell lysis using bead-beating and ultrasonication
Many cell lysis techniques are utilised in pigment analysis including maceration, ultrasonication, high-pressure assisted extraction and bead beating, depending on the microalgal species. In this study, two extraction techniques were tested: straight solvent extraction with sonication and\ solvent extraction with bead-beating using silica beads. The pigment recoveries with the different cell lysis techniques were then compared (Fig. 2). Bead-beating recovered significantly higher concentrations for all pigments: chl-b was 380-fold higher than ultrasonication followed by chl-a which was 9.5-fold higher, and both zeaxanthin and lutein were ~3.4-fold higher in C. variabilis (Fig 2). This clearly shows that complete cell lysis using mechanical maceration is an important first step in the overall extraction protocol and that solvents will not efficiently extract pigments through non-lysed cells. The same results were observed for the other two species (Fig 2). Similar findings have been reported in microalgal species Scenedesmus obliquus (53). It was also reported that bead-beating pre-treatment with ethanol led to higher carotenoid extraction than ultrasound and freeze-thawing (54). Some contradicting results have been reported for astaxanthin extraction, an 81% astaxanthin extraction efficiency from H. pluvialis biomass was reported that was pre-treated by sonication at 750 W coupled with chemical lysis using 2 M NaOH solution followed by methanol extraction (55). However, in contrast a relatively low astaxanthin recovery of 19.8 mg/g cell from dried H. pluvialis biomass by two-step treatment with 4 M HCl at 70 °C and ultrasonic extraction for 20 min in acetone was documented (56). Similarly, a very low astaxanthin extraction efficiency (12%) from intact H. pluvialis cysts, despite the use of a high-power 600 W horn-type sonicator (1 s on and 3 s off for 30 min) with methanol was also reported (57).
Selection of optimum extraction solvent for separation of pigments
Another important parameter for extraction efficiency is the solvent used (58). The goal of this study was to provide a universal extraction protocol from the many existing solvent systems in literature depending on the microalgae species and target pigments (59, 60).
The following data is based on the quantitative estimates for 18 pigments and 2 lipid classes. The ethanol:hexane (1:1) (method 5) solvent mixture provided the highest extraction efficiency for most of the pigments (zeaxanthin, lutein, chl-b, astaxanthin, Div Chl a, canthaxanthin, alloxanthin, β-carotene) in H. pluvialis. Chl-a made an exception that was extracted by ethanol:hexane (2:1) (method 3) Folch method is routinely used for lipidomics extractions, here it is shown that PCs have a greater extraction efficiency with the folch (method 4) that corroborates with the literature (Fig 3). PCs were extracted best through folch whereas TAGs through ethanol:hexane (2:1) (method 3) in C. variabilis and D. subspicatus. Similarly, the PCs were extracted most efficiently using the Folch mixture in H. pluvialis closely followed by acetone (method 2). Pigments from D. subspicatus showed a maximum recovery using the ethanol:hexane solvent system (zeaxanthin, lutein, Div Chl a, β-carotene, cryptoxanthin and lycopene). Chlorophylls (chl-a and chl-b) were extracted best by ethanol:hexane solvent systems in all the three species. Ethanol:hexane solvent system worked best for extraction of pigments in C. variabilis as well. The pigments neoxanthin, prasinoxanthin, violaxanthin, diadinoxanthin were maximally extracted through Folch (method 4) in H. pluvialis. Neoxanthin, prasinoxanthin, violaxanthin were also equally extracted through methanol (method 1) in D. subspicatus and through acetone (method 2) in C. variabilis (Supplementary Fig 1).
Ethanol:hexane (method 3 and 5) system gave the highest amount for most of the pigments and also for TAGs except for the pigments neoxanthin, prasinoxanthin, violaxanthin and diadinoxanthin. PCs overall were maximised using Folch (method 4) and acetone (method 2). Acetone and ethanol are generally regarded as safe solvents (GRAS) (61) and therefore provides a better option than the ethanol/hexane combination due to the toxicity of hexane. Furthermore, it has been shown that methanolic pigments are degraded more quickly than other solvents (62, 63). Several reports have also shown that methanol promotes formation of allomers of chlorophyll (64, 65). Acetone therefore provides a more stable system in comparison to methanol. For these reasons, 90% acetone (method 2) was selected as the system which provides the best compromise as it’s less volatile so easier to handle, less toxic moderately polar, simple solvent system that provides a broad spectrum of pigments. However overall, it was shown that no single solvent system is optimal for all pigments, this is due to the relative polarities of the solute and solvents. The key to selection therefore is the ease of handling, toxicity and target pigments of interest.
Overall Pigment profile analysis
The whole data set contained 18 pigments and 2 lipids. To visualise the impact of solvent composition on the whole profile, principal components analysis (PCA) was used. A PCA described 83% of the variance in the first two components. Clear separation of each of the species (C, D and H) can be observed along PC1 based on the overall pigment profiling where D and H are slightly closer to each other (Fig 4(a)). Most variation can be seen across PC1 (47.2%) and therefore, C, H and D have been grouped separately and interestingly is based on different pigments rather than lipids. Across PC2 (36%), clustering among different species can be seen based on the extraction solvents and the compounds extracted. A clear gradient from high polarity to low polarity solvent can been seen in PC2. The loading plot also show this (Fig4(b)). Therefore, C3 -C5 and C1- C4 are clustered together wherein C2 separates out as an intermediate group that justifies the use of acetone as an extraction solvent with intermediate polarity. Similarly, D1 -D4 and D3- D5 are clustered together and D2 seem to separate out as a subgroup closer to D3- D5. Same pattern can be observed for H where H3 -H5 are close to each other and so as H2-H1 and H4 separate out as a different cluster. The species of the extraction methods 3 and 5 cluster together, this is due to the similar composition of the extraction solvent with different ratios of ethanol: hexane. The compound group mainly responsible for the separation from the other extraction methods are the non-polar TAGs. The PCs are responsible for clustering 1,2,4 in C and D together. Canthaxanthin, astaxanthin, alloxanthin and diadinoxanthin cause separation between H (Fig 4(b)).