Enhancing Astaxanthin Biosynthesis by Rhodosporidium toruloides Mutants and Optimization of Medium Compositions Using Response Surface Methodology

Abstract

Astaxanthin is a valuable carotenoid, which has been approved as a food coloring by the US Food and Drug Administration and is considered as a food dye by the European Union (European Commission). This work aimed to attain Rhodosporidium toruloides mutants for enhanced astaxanthin accumulation using ultraviolet (UV) and gamma irradiation mutagenesis. Gamma irradiation was shown to be more efficient than UV for producing astaxanthin-overproducer. Among the screened mutants, G17, a gamma-induced mutant, exhibited the highest astaxanthin production, which was significantly higher than that of the wild strain. Response surface methodology was then applied to optimize the medium compositions for maximizing astaxanthin production by the mutant G17. The optimal medium compositions for the cultivation of G17 were determined as a peptone concentration of 19.75 g/L, malt extract concentration of 13.56 g/L, and glucose concentration of 19.92 g/L, with the maximum astaxanthin yield of 3021.34 µg/L ± 16.49 µg/L. This study suggests that the R. toruloides mutant (G17) is a potential candidate for astaxanthin production.

1. Introduction

Astaxanthin (3,3′-dihydroxy-β-carotene-4,4′-dione) is a keto-carotenoid with extensive use as a feed supplement in aquaculture industry due to its contributions to the attractive coloration, growth, and survival of marine animals such as shrimps, salmon, and trout [1,2]. Astaxanthin has been approved as a color additive for specific uses in fish and animal foods by the Food and Drug Administration (FDA, USA) [3]. The European Union also considers astaxanthin as a food dye within the E number system (E161j) [3]. Furthermore, astaxanthin has a strong antioxidant activity and reveals a potential therapeutic agent against various diseases such as cardiovascular disease, skin-related illness, atherosclerotic, and cancer [4,5]. Therefore, astaxanthin has various applications in pharmaceutical, nutraceutical, and cosmetic industries [3,6]. Because of such benefits, there is an increasing demand for astaxanthin, which consequently requires an efficient and safe approach for astaxanthin production.

Astaxanthin is commonly synthesized by using a chemical process [6]. Although this process can produce astaxanthin in large quantities, it has a negative effect on environment [7]. Remarkably, concerns remain about the food safety and biological function of the synthetic astaxanthin [8,9]. To prevent these obstacles, using microorganisms for natural astaxanthin production has increasingly gained considerable interest as microorganisms can produce astaxanthin through their biosynthesis pathway [10,11]. Bacteria (e.g., Brevibacterium sp. [12] and Escherichia coli [13]) and green microalgae (e.g., Chromochloris zofingiensis (formerly Chlorella zofingiensis) [14] and Haematococcus lacustris (formerly Haematococcus pluvialis) [15]) have been mainly studied for astaxanthin synthesis. Among them, H. lacustris is a potential source of astaxanthin because of its high content of astaxanthin [16,17]. Nevertheless, the use of H. lacustris for astaxanthin production retains several limitations due to high cost of cultivation technique, and its low cell concentration and slow growth (specific growth rate of 0.31/d) [18,19]. Therefore, great efforts have been made to search for more candidates as new astaxanthin producers.

In recent years, yeast has received increasing attention as a substitute source of astaxanthin due to its high biomass productivity (up to 56.4 g/L d) and short cultivation time (3–4 d) compared to microalgae (which require more than 10 d for cultivation and have biomass productivity of about 0.028–0.7 g/L d) [20,21,22]. In addition, the use of yeast for astaxanthin production eliminates the negative impact of the chemical process on the environment. Particularly, Rhodosporidium toruloides (current name: Rhodotorula toruloides), an oleaginous yeast (Fungi, Basidiomycota), has been reported as a promising source of carotenoids [23,24]. This yeast can use various low-cost carbon sources and grow to very high density (100 g/L dry cell mass) [25]. Recently, our previous study has shown a potential use of R. toruloides to facilitate a low-cost approach for astaxanthin production [2]. R. toruloides can reach high biomass yield within a short cultivation time (6.84 g/L dry cell mass within four-day cultivation). However, the wild strains of R. toruloides have low astaxanthin yield (927.11 µg/L) [2], thus limiting its industrial application. Therefore, increasing efficiency of astaxanthin production by R. toruloides can greatly expand the utilization of R. toruloides.

The astaxanthin accumulation in microorganisms is generally facilitated in response to the stress conditions such as salinity, nutrients-starvation, pH, temperature, or unfavorable light [26]. Studies have reported that astaxanthin synthesis by microorganisms can also be enhanced by mutagenization [26,27]. Different approaches, including chemical mutagenesis and physical mutagenesis methods, have been developed to generate mutant strains of microorganisms with higher astaxanthin content [26,27,28]. Among these approaches, the physically treated methods (ultraviolet (UV) and gamma-ray) have been successfully used as a random mutagenesis strategy to generate genetic mutants with higher astaxanthin accumulation in yeasts [29] and algae [26]. In addition, the physical mutagenesis methods are simple, compared to the chemical mutagenesis methods [26]. Therefore, this study proposed, for the first time, the use of R. toruloides mutants for enhanced astaxanthin production.

In this work, UV light and gamma irradiation were used as a mutagenesis agent to produce R. toruloides mutants. The R. toruloides mutants were screened to obtain astaxanthin-overproducing strains. Response surface methodology (RSM) was then employed to optimize the medium compositions for the cultivation of the selected R. toruloides mutant to obtain the maximum astaxanthin yield.

2. Materials and Methods

2.1. Yeast Strains and Culture Conditions

Wide-type R. toruloides was provided by University of Science, Vietnam National University Ho Chi Minh City, Vietnam. R. toruloides strain was grown in the Hansen broth medium (pH 6) containing 10 g/L peptone, 50 g/L sucrose, 3 g/L KH₂PO₄, and 3 g/L MgSO₄. The seed culture was prepared by inoculating R. toruloides cells (5%, v/v) into 250-mL Erlenmeyer flasks containing 100 mL of the Hansen broth and incubated at 30 °C with shaking at 200 rpm for 96 h. The prepared cultures were then used for next studies. In all experiments, the flasks and media were sterilized at 121 °C for 20 min.

2.2. Random UV Mutagenesis

The wild-type R. toruloides cells cultivated in the Hansen broth medium in the logarithmic phase (cell density of 10⁸ cells/mL) were used for UV mutagenesis as described previously [26]. The cells were obtained from the culture by centrifugation at 13,000× g rpm for 10 s. The cell pellets were then rinsed twice with distilled water before being resuspended in distilled water to 10⁸ cells/mL. The cell suspension (1 mL) was then illuminated at room temperature for different exposure times (5–30 s) using 248 nm UV Display lamp (Mini UV sterilisation cabinet, Cleaver scientific, England). The treated cells (100 µL) were spread on Hansen agar plate and incubated at 30 °C for 7 d in the dark. The survival rates of the UV-treated cells were subsequently determined by counting the colonies on agar plates. The UV-induced mutants were selected and cultivated in the Hansen broth medium for 4 d at 30 °C with shaking to determine the biomass and astaxanthin yield.

2.3. Random Gamma Irradiation Mutagenesis

The wild-type R. toruloides cultivated in the Hansen broth medium in the logarithmic phase (cell density of 10⁸ cells/mL) were used for ⁶⁰Co-gamma mutagenesis as described previously [30]. The cells were obtained from the culture by centrifugation at 13,000× g rpm for 10 s. The cell pellets were then rinsed twice with distilled water before being resuspended in distilled water to 10⁸ cells/mL. The cell suspension (1 mL) was then irradiated using a ⁶⁰Co-gamma ray irradiator (GC-5000, India) as a source with 0.5, 1.0, 1.5, 2.0, 2.4, and 3.0 kGy doses. The irradiated cells (100 µL) were plated on Hansen agar medium and incubated at 30 °C for 7 d in the dark. The survival rates of the gamma-treated cells were then determined by counting the colonies on agar plates. The gamma-induced mutants were selected for further cultivation in the Hansen broth medium at 30 °C with shaking for 4 d. The culture was subsequently obtained for biomass and astaxanthin yield determination.

2.4. Optimization of Medium Compositions

The influences of medium compositions on the astaxanthin production by R. toruloides mutant were studied using a five-level and three-factorial central composite design.

Table 1. Coded values of the variables for the central composite design.

Variables	Symbols	Levels
		−1.68 (−α)	−1	0	1	1.68 (+α)
Peptone concentration (g/L)	X1	11.6	15	20	25	28.4
Malt extract concentration (g/L)	X2	6.6	10	15	20	23.4
Glucose concentration (g/L)	X3	11.6	15	20	25	28.4

illustrates the coded and actual values of the input variables (nutritional parameters). R. toruloides mutant were cultivated in the medium containing various peptone concentrations (11.6–28.4 g/L), malt extract concentrations (6.6–23.4 g/L), and glucose concentrations (11.6–28.4 g/L) at 30 °C for 4 d with shaking at 200 rpm. The cells were then collected from the culture to determine the astaxanthin yield. The relation of the determined astaxanthin yield (measured response) to the nutritional parameters was expressed as follows:

$$Y={\beta }_0+{\beta }_1{X}_1+{\beta }_2{X}_2+{\beta }_3{X}_3+{\beta }_{11}{X}_1^2{+\beta }_{22}{X}_2^2{+\beta }_{33}{X}_3^2{+\beta }_{12}{X}_1{X}_2{+\beta }_{13}{X}_1{X}_3{+\beta }_{23}{X}_2{X}_3$$

(1)

where Y is the astaxanthin yield (µg/L); X₁ is the peptone concentration (g/L); X₂ is the malt extract concentration (g/L); X₃ is the glucose concentration (g/L); β₀ is the intercept coefficient; β₁–β₃ are the linear coefficients; β₁₁, β₂₂, and β₃₃ are the quadratic coefficients; and β₁₂, β₁₃, and β₂₃ are the interaction coefficients. These model coefficients were determined using the least-squares method [31]. The experimental design, an analysis of variance (ANOVA), and regression analysis were also conducted using Minitab 16 (Minitab Inc., State College, PA, USA). The established model was subsequently applied to examine the combined effect of the input variables and determine the optimal medium compositions for maximizing astaxanthin yield by solving the regression equation [31].

2.5. Analysis

2.5.1. Determination of Survival Rate

The survival rate of R. toruloides cells was calculated as follows [8]:

$$Survival rate (\%)=\frac{C_s}{C_c}\times 100$$

(2)

where C_S and C_c are the total colony count after treatment (by UV or gamma irradiation) and the control (sample without treatment), respectively.

2.5.2. Biomass Determination

The yeast cells were obtained from the culture by centrifugation at 4000× g rpm for 5 min. The cell pellets were subsequently rinsed twice with distilled water and dried at 60 °C to a constant weight. Yeast biomass was calculated and expressed as grams in dry weight per liter of culture medium (g/L). The dried biomass was then used to determine astaxanthin content.

2.5.3. Astaxanthin Content Determination

Astaxanthin content was determined based on the methods of Fang and Cheng [32] and An et al. [33] with modifications. Dried biomass (0.2 g) was ground in 3 mL dimethyl sulfoxide (DMSO) in a glass homogenizer for 30 min. The mixture was centrifuged at 4000× g rpm for 5 min to obtain the supernatant. The pellet containing residual pigments was well mixed with acetone (5 mL) and centrifuged at 4000× g rpm for 5 min to obtain the supernatant. This acetone-extraction step was repeatedly performed 2–3 times to extract all pigment from the solid residue. The DMSO and acetone extracts were then mixed with petroleum ether (1:2, v/v), 10 mL of distilled water, and 5 mL of 20% NaCl solution and placed at room temperature for phase separation. The upper layer containing distilled water, DMSO, and acetone was then removed to obtain the petroleum ether extract. The extract was measured the absorbance at 468 nm against the blank (petroleum ether) using a Genesis 10S UV-Vis spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Astaxanthin content was subsequently calculated as follows [34]:

$$\mathrm{Astaxanthin}\mathrm{content}\left(\mathsf{\mu }\mathrm{g}/\mathrm{g}\right)=\frac{{\mathrm{A}}_{468}\times \mathrm{V}\times {10}^4}{{\mathrm{E}}_{1\mathrm{cm}}\%\times \mathrm{G}}$$

(3)

$$\mathrm{Astaxanthin}\mathrm{yield}\left(\mathsf{\mu }\mathrm{g}/\mathrm{L}\right)=\frac{\mathrm{Astaxanthin}\mathrm{content} \left(\mathsf{\mu }\mathrm{g}/\mathrm{g}\right)\times \mathrm{G}}{\mathrm{Volume}\mathrm{of}\mathrm{culture}\left(\mathrm{L}\right)}$$

(4)

where A₄₆₈ is the absorbance of sample at 468 nm; G is the total dry weight of yeast biomass (g); V is the total volume of petroleum ether extract (mL); E_1cm% is the absorbance of 1% (w/v) astaxanthin solution in petroleum ether (cuvette width of 1 cm, E_1cm% = 2100).

2.5.4. Statistical Analysis

Data were determined in triplicate and shown as the mean ± standard deviation (SD). The data analysis was performed by an ANOVA with the Duncan’s multiple range test at p values ≤ 0.05 using SAS software ver 8.2 (SAS Institute, Cary, NC, USA).

3. Results and Discussion

3.1. Astaxanthin Production by UV-Induced Mutants

In this study, the wild-type R. toruloides was mutated by using UV and screened for astaxanthin overproducers. After being treated with UV for different exposure times (0–30 s) and cultivated on the solid medium for 7 d, the R. toruloides cells were efficiently induced with a rapid color change from slightly pink to red and orange, indicating that UV acts as an inducer for carotenogenesis in R. toruloides. The color change in the yeast cells may be due to the respiratory effect of UV light that induces the formation of reactive oxygen species (ROS)-sensitive mutant [35]. Under the UV treatment, the survival of R. toruloides cells was declined but the astaxanthin production was significantly promoted. As shown in Figure 1, the survival rate of the R. toruloides was dramatically decreased to below 7% with the treatment of UV for longer than 25 s. Although a long UV treatment time caused a low survival rate, it resulted in higher astaxanthin accumulation in the cells. The enhanced astaxanthin production may be due to several protective mechanisms. The carotenoid synthesis in yeast is a complex process, which could be regulated at multiple levels, involving the flow of metabolic intermediates among different pathways, the concentration of carotenogenic gene mRNAs, and the amount of the carotenogenic proteins and their enzymatic activities [36]. When the cells are exposed to oxidative stress (such as UV radiation and gamma irradiation), reactive oxygen species (such as H₂O₂ and ¹O₂) are produced at the cellular level and cause DNA damage and oxidized proteins [29,35]. To resist the stress, carotenogenesis, which is a photoprotection mechanism in yeast, is subsequently activated to produce carotenoids [35,37]. The produced carotenoids protect the cells by removal of ROS generated by light stress [37]. Studies have reported the presence of singlet oxygen (¹O₂) induces the synthesis of astaxanthin by activating enzymes involved in carotenoid biosynthesis [35,38]. Consequently, the UV-induced mutant cells enhanced their astaxanthin production. In addition, mutagenesis process may occur and causes the changes in the nucleotide sequence of several genes (required for carotenoid biosynthesis) to enhance the biosynthesis of astaxanthin [36]. Studies have reported the increase in the mRNA levels of two genes involved in the astaxanthin biosynthesis in yeast, thus enhancing the astaxanthin production [36,39]. For further biomass and astaxanthin analysis, more than 150 mutant cells were selected for cultivation to screen astaxanthin overproducer. Among 150 mutants, 16 strains were proved as the real astaxanthin-overproducer (

Table 2. Biomass and astaxanthin production by UV-induced mutants.

Strain	Biomass Yield (g/L)	Astaxanthin Content (µg/g)	Astaxanthin Yield (µg/L)
UV1	2.847 ± 0.0039 dfe	187.641 ± 0.026 g	534.28 ± 7.22 j
UV2	3.461 ± 0.0007 dc	166.927 ± 0.018 l	577.68 ± 1.14 h
UV3	2.684 ± 0.0004 fe	253.334 ± 0.004 c	679.95 ± 1.00 e
UV4	2.931 ± 0.0005 dfe	194.522 ± 0.003 f	570.21 ± 0.96 i
UV5	4.149 ± 0.0004 ab	144.290 ± 0.002 n	598.70 ± 0.54 g
UV6	4.629 ± 0.0004 a	183.345 ± 0.017 h	848.77 ± 0.77 a
UV7	2.931 ± 0.0004 dfe	274.293 ± 0.002 b	803.86 ± 1.21 b
UV8	2.538 ± 0.0002 dfe	275.241 ± 0.003 a	698.56 ± 0.55 c
UV9	3.849 ± 0.0005 bc	172.380 ± 0.001 j	663.55 ± 0.87 f
UV10	1.754 ± 0.0004 g	232.629 ± 0.007 e	407.95 ± 1.03 o
UV11	2.595 ± 0.0003 f	182.855 ± 0.010 i	474.51 ± 0.63 m
UV12	3.016 ± 0.0002 dfe	171.430 ± 0.002 k	517.09 ± 0.27 k
UV13	3.734 ± 0.0002 bc	130.476 ± 0.007 q	487.15 ± 0.26 l
UV14	2.753 ± 0.0003 fe	250.245 ± 0.003 d	688.84 ± 0.72 d
UV15	3.131 ± 0.0003 dfe	155.485 ± 0.004 m	486.82 ± 0.40 l
UV16	3.249 ± 0.0006dce	142.641 ± 0.026 o	463.49 ± 0.80 n
WT *	2.777 ± 0.0003 fe	131.664 ± 0.010 p	365.63 ± 0.42 p

* WT: wild-type strain (control). Data are mean ± SD. Means within a column with different letters significantly differ by Duncan’s test at p < 0.05.

). Those strains exhibited a significantly higher astaxanthin yield than the wild-type strain (365.63 µg/L ± 0.42 µg/L). Among the improved mutant strains, the UV6 strain exhibited the highest biomass (4.629 g/L ± 0.0004 g/L) and astaxanthin yield (848.77 µg/L ± 0.77 µg/L). The astaxanthin yield of the UV6 strain was 2.3-fold higher than that of the wild-type strain. This result is comparable to a study reported by Stachowiak [35] that astaxanthin yield was enhanced 1.7-fold by a light-induced mutant strain of Phaffia rhodozyma (Xanthophyllomyces dendrorhous) 34B compared to the parental strain.

3.2. Astaxanthin Production by Gamma-Induced Mutants

In this work, gamma irradiation was employed as another method to generate R. toruloides mutants. The wild-type R. toruloides was treated with ⁶⁰Co-gamma irradiation under different doses and subsequently cultivated on a solid medium for 7 d to screen for astaxanthin overproducers. The R. toruloides cells were efficiently induced with a rapid color change from slightly pink to red and orange, indicating that the gamma irradiation induces carotenogenesis in R. toruloides. Similar to UV treatment, gamma irradiation treatment also declined survival rates of R. toruloides. As shown in Figure 2, the survival rate of the wild-type R. toruloides was dramatically decreased with increasing gamma irradiation dose but enhanced the astaxanthin accumulation in the cells. This result indicated that the enhanced astaxanthin production protects the cells against gamma irradiation. Studies have reported that gamma irradiation produces oxygen radicals, which modify the astaxanthin biosynthesis pathway in yeast [29,30]. The oxygen radicals may change enzyme activity or protein properties and may also induce the mutation of yeast through a chromosomal rearrangement [29,30]. Consequently, the mutant strains increase their astaxanthin biosynthesis [29,30]. To screen astaxanthin overproducer, more than 150 mutant cells were selected for cultivation. Among them, 20 strains were proved as the real astaxanthin-overproducer (

Table 3. Biomass and astaxanthin production by gamma-induced mutants.

Strain	Biomass Yield (g/L)	Astaxanthin Content (µg/g)	Astaxanthin Yield (µg/L)
G1	4.737 ± 0.0007 e	206.602 ± 0.023 o	978.67 ± 1.0 j
G2	3.710 ± 0.0002 l	212.161 ± 0.030 l	787.12 ± 0.4 n
G3	4.556 ± 0.0004 f	210.909 ± 0.019 m	960.90 ± 0.6 k
G4	5.113 ± 0.0015 c	196.645 ± 0.030 r	1005.45 ± 2.0 h
G5	5.403 ± 0.0003 a	220.475 ± 0.045 i	1191.23 ± 0.6 c
G6	3.855 ± 0.0004 k	229.481 ± 0.035 f	884.65 ± 0.8 l
G7	4.236 ± 0.0006 i	185.002 ± 0.040 s	783.67 ± 0.9 o
G8	3.845 ± 0.0004 k	230.779 ± 0.030 e	887.34 ± 0.7 l
G9	4.071 ± 0.0005 j	218.515 ± 0.025 j	889.57 ± 0.8 l
G10	5.104 ± 0.0004 c	214.647 ± 0.035 k	1095.56 ± 0.7 f
G11	5.098 ± 0.0005 c	228.490 ± 0.010 g	1164.84 ± 0.7 d
G12	4.943 ± 0.0004 d	200.579 ± 0.041 q	991.46 ± 0.7 i
G13	4.474 ± 0.0004 h	233.815 ± 0.050 d	1046.09 ± 0.9 g
G14	4.724 ±0.0005 e	208.408 ± 0.037 n	984.52 ± 0.8 j
G15	4.254 ± 0.0005 i	204.344 ± 0.077 P	869.28 ± 1.0 m
G16	4.516 ± 0.0003 g	253.835 ± 0.025 c	1146.32 ± 0.5 e
G17	3.297 ± 0.0408 m	603.659 ± 0.025 a	1990.26 ± 1.68 a
G18	5.245 ± 0.0024 b	222.142 ± 0.497 h	1165.14± 6.3 d
G19	4.729 ± 0.0004 e	337.306 ± 0.035 b	1595.12 ± 1.0 b
G20	4.563 ± 0.0003 f	173.364 ± 0.040 t	791.06 ± 0.4 n
WT *	2.777 ± 0.0003 l	131.664 ± 0.010 u	365.63 ± 0.42 p

* WT: wild-type strain (control). Data are mean ± SD. Means within a column with different letters significantly differ by Duncan’s test at p < 0.05.

). Those strains showed a significantly higher biomass yield, astaxanthin content, and astaxanthin yield than the wild-type strain. Notably, the gamma-induced mutant strains showed higher astaxanthin accumulation than the UV-induced mutant strains, indicating the gamma irradiation was more efficient than UV treatment to produce astaxanthin-overproducing R. toruloides mutants. Among all mutant strains (UV-induced and gamma irradiation-induced mutants), G17 strain exhibited the highest astaxanthin content (603.659 µg/g ± 0.025 µg/g) and astaxanthin yield (1990.26 µg/L ± 1.68 µg/L). The astaxanthin yield of the G17 strain was 5.4-fold higher than that of the wild strain. This result is in agreement with another study that a gamma-induced mutant strain of P. rhodozyma YZUXHONG686 exhibited 17-fold higher astaxanthin yield than the parental strain [40]. Therefore, G17 strain was selected for further experiments.

3.3. Optimization of Medium Compositions Using RSM

3.3.1. Development of RSM Model

The nutrients in culture medium are crucial factors affecting the growth and metabolism of microorganisms [2,41]. Studies have reported that astaxanthin accumulation is strongly affected by the medium compositions [2,42]. To maximize the astaxanthin production by the R. toruloides, the mutant G17 was cultivated in the medium containing different concentrations of peptone, malt extract, and glucose (

Table 4. Central composite design matrix for the effect of the nutritional parameters on the astaxanthin yield in coded values and experimental results.

Run	Variables			Response, Y (µg/L)
	X1	X2	X3	Actual Value	Predicted Value
1	−1	−1	−1	2514.50	2467.02
2	−1	−1	1	2425.02	2382.75
3	1	−1	−1	2401.92	2367.29
4	1	−1	1	2451.20	2375.71
5	−1	1	−1	2270.44	2249.67
6	−1	1	1	2297.28	2235.65
7	1	1	−1	2144.48	2090.50
8	1	1	1	2217.93	2169.16
9	0	0	−1.68	2253.40	2300.27
10	0	0	1.68	2206.29	2295.55
11	−1.68	0	0	2224.83	2280.79
12	1.68	0	0	2060.85	2141.02
13	0	−1.68	0	2594.37	2666.81
14	0	1.68	0	2246.67	2310.36
15	0	0	0	3014.95	2995.97
16	0	0	0	2981.78	2995.97
17	0	0	0	3016.95	2995.97
18	0	0	0	2988.78	2995.97
19	0	0	0	3004.95	2995.97
20	0	0	0	2991.78	2995.97

). The relationship between three nutritional parameters (peptone, malt extract, and glucose concentrations) and astaxanthin yield was then evaluated using RSM model to obtain the optimal medium compositions. As can be seen from Table 4, the central experiments (runs 15–20) presented a low coefficient of variance (CV = 0.49%), demonstrating the reproducibility of the experiments. Therefore, the relationship between astaxanthin yield and three nutritional parameters were modeled as follows:

$$Y=2995.97-41.55X_1-105.97X_2-1.4X_3-277.56X_1^2-179.39X_2^2-246.8X_3^2-14.86X_1{X}_2+23.17X_1{X}_3+17.56X_2{X}_3$$

(5)

where X_1×3 and X_2×3 have positive influences on the measured response, whereas the other parameters have negative influences.

The model (Equation (5)) was then evaluated using F test for ANOVA. As can be seen from

Table 5. ANOVA for the empirical model.

Source	DF b	SS b	MS b	F Value	Probability (P) > F
Model a	9	2,251,169	250,130	49.39	<0.0001
Residual (error)	10	50,642	5064
Lack-of-fit	5	49,582	9916	46.75	<0.0001
Total	19	2,301,811

^a Coefficient of determination (R²) = 0.98; adjusted R² = 0.96.^b DF, degree of freedom; SS, sum of squares; MS, mean square. Coefficient of variance (CV) = 0.49%; adequate precision = 17.99.

, p value of the model was 0.0001, signifying that the established model was statistically significant at the 95% confidence level. Furthermore, the coefficient of determination (R²) and adjusted R² were 0.98 and 0.96, respectively, demonstrating the precision of the model for predicting the astaxanthin yield. The adequate precision was 17.99, which was greater than four, signifying that the signal was adequate, and the developed model could be efficiently used to navigate the design space. As can be seen from Table 4, a good agreement between the actual and predicted values of the astaxanthin concentration was observed; therefore, the developed model could provide reliable results. In addition, the t test was used to evaluate the overall influences of the nutritional parameters on the astaxanthin yield. As shown in

Table 6. Significance of the coefficients in the empirical model.

Model Term	Parameter Estimate	Standard Error	t Value a	p Value
β0	2995.97	29.02	103.23	0.000 b
β1	−41.55	19.26	−2.16	0.056
β2	−105.97	19.26	−5.503	0.000 b
β3	−1.40	19.26	−0.07	0.943
β11	−277.56	18.75	−14.81	0.000 b
β22	−179.39	18.75	−9.57	0.000 b
β33	−246.80	18.75	−13.17	0.000 b
β12	−14.86	25.16	−0.59	0.568
β13	23.17	25.16	0.92	0.379
β23	17.56	25.16	0.70	0.501

^a t_α/2,n-p = t_0.025,10 = 2.23. ^b p < 0.05 indicates that the model terms are significant.

, small p values (<0.05) were observed for the constant, a linear term X₂, and all quadratic terms, indicating that they were crucial parameters in the medium compositions. The established model was subsequently applied to predict the optimal medium compositions for maximizing astaxanthin yield.

3.3.2. Combined Effect of Nutritional Parameters

Figure 3 presents the combined effects of the concentrations of peptone and malt extract on the astaxanthin production while keeping glucose concentration at a central level. At a given malt extract concentration, the astaxanthin yield enhanced with increasing the amount of peptone. This is because peptone supported the biosynthesis of astaxanthin in R. toruloides [32]. Nevertheless, a further increase in the peptone amount caused a decline in astaxanthin accumulation. This could be explained that high peptone concentration resulted in a decrease in ratio of carbon to nitrogen, which consequently causes a suppression of the cell growth and the biosynthesis of secondary pigments [43]. Initial ratio of carbon to nitrogen in the medium has reported to significantly influence the astaxanthin synthesis in yeast including R. toruloides [2]. A high ratio of carbon to nitrogen suppresses the synthesis of protein, which consequently reduces NADPH consumption for primary metabolism. As a result, more NADPH is available for the astaxanthin synthesis, thus increasing the astaxanthin production [43,44]. This result is similar to the astaxanthin production in microalgae (C. zofingiensis) [44] and yeast (Phaffia rhodozyma) [45], which requires a high ratio of carbon to nitrogen for the biosynthesis of astaxanthin.

Figure 4 illustrates the combined effects of concentrations of peptone and glucose on the astaxanthin accumulation. At any amount of peptone, astaxanthin yield increased with increasing glucose concentration. This could be explained that glucose enhances metabolism and cell growth [46]. Studies have demonstrated that a sufficient amount of glucose promoted astaxanthin accumulation in yeast through the carotenogenesis pathway [47,48]. During the fermentation, yeast cells use glucose as a substrate to synthesize intermediates such as ethanol and acetic acid. The synthesized ethanol is subsequently converted to acetyl CoA (a precursor for the synthesis of carotenoids) by alcohol dehydrogenase, thus facilitating the biosynthesis of carotenoids [48]. Nevertheless, a higher glucose amount caused a decrease in astaxanthin accumulation. This could be because the cell growth is suppressed by the excess glucose, and high sugar osmotic pressure induces the production of energetic substances such as protein to resist the environmental stress [46], thus reducing the astaxanthin production. This finding is consistent with the astaxanthin accumulation in P. rhodozyma [48] and microalgae C. zofingiensis [47].

Figure 5 presents the combined effect of malt extract and glucose concentrations on the astaxanthin yield when maintaining peptone concentration at a central level. At any glucose concentration, the astaxanthin yield considerably increased with increasing amount of malt extract. This might be because malt extract plays an important role as a complex of nitrogen and carbon source, which promotes the growth of yeast, resulting in higher astaxanthin yield [49,50,51]. In this work, malt extract was used as an additional nitrogen source for the cultivation of R. toruloides. The result showed that malt extract exhibited a significant effect on astaxanthin synthesis (p < 0.05) while peptone had insignificant effect (p > 0.05) in the evaluated range (Table 6). Studies have reported that the nitrogen level of malt extract may differ from 1.4% to 1.8% which may be insufficient in itself [49], but when mixed with another nitrogen source (peptone in this study), it was observed to be essential for R. toruloides to synthesize astaxanthin. However, a further increase in malt extract resulted in a decline in astaxanthin accumulation due to the negative effect of high levels of nitrogen on the astaxanthin biosynthesis [2].

3.3.3. Obtaining Optimal Medium Compositions

By solving the RSM model (Equation (5)), the optimal medium compositions were predicted to be peptone concentration of 19.75 g/L, malt extract concentration of 13.56 g/L, and glucose concentration of 19.92 g/L, with the corresponding astaxanthin yield of 3012.68 µg/L. To verify the prediction, an experiment was performed under the optimized medium compositions. An astaxanthin yield of 3021.34 µg/L ± 16.49 µg/L was obtained, signifying a good agreement with the model prediction. Consequently, the established RSM model can be used to describe the relationship between astaxanthin accumulation and nutritional parameters in the cultivation of R. toruloides. R. toruloides mutant exhibited a higher astaxanthin yield than Spirulina platensis (38 µg/L) [8] and P. rhodozyma (639.6 µg/L) [40]. Although the astaxanthin accumulation in R. toruloides mutant was lower than that in Thraustochytrium striatum (6.2 mg/L) [52] and microalgae H. lacustris (84.8 mg/L) [53], the R. toruloides is more efficient than these microorganisms in term of fast growth rate and short cultivation time (

Table 7. Astaxanthin production by several microorganisms.

Strains	Cultivation Time (d)	Astaxanthin Yield (μg/L)	References
Spirulina platensis	10	38	[8]
Phaffia rhodozyma	8	639.6	[40]
Thraustochytrium striatum	15	6200	[52]
Haematococcus lacustris	30	84,800	[53]
R. toruloides (wild strain)	4	927.1	[2]
R. toruloides (mutant)	4	3021.3	This study

). Notably, R. toruloides mutant exhibited 3.3-fold higher astaxanthin yield than the wild-type R. toruloides (927.1 µg/L) reported in a previous study [2], indicating the gamma irradiation efficiently enhanced the astaxanthin production in R. toruloides. Those results suggested that R. toruloides mutant could be a promising source for the astaxanthin production. Further studies are still required to enhance the astaxanthin production of R. toruloides.

4. Conclusions

This paper reports the generation of R. toruloides mutants for enhanced astaxanthin production using UV and gamma irradiation mutagenesis. The UV- and gamma-induced mutants exhibited a higher astaxanthin production than that of a wild-type strain. Among the screened mutants, the gamma-induced mutant G17 showed the highest astaxanthin content (603.70 µg/g) and astaxanthin yield (1990.26 µg/L). To further increase the astaxanthin production, the medium compositions for the cultivation of G17 strain were then optimized for maximizing the astaxanthin yield. Through RSM, the optimal medium compositions were determined and a maximum astaxanthin yield of 3021.34 µg/L was obtained. This study suggests that the R. toruloides mutant strain of G17 is a potential microbial source for astaxanthin production.