Purple LED light and crude glycerol synergistically enhance astaxanthin production in Aurantiochytrium limacinum

Abstract

Aurantiochytrium limacinum is a heterotrophic marine protist known for its ability to co-produce docosahexaenoic acid (DHA) and carotenoids such as astaxanthin. Despite its biotechnological potential, industrial application is hindered by high production costs. In this study, we investigated the effects of different light wavelengths and carbon sources on the production of carotenoids and neutral lipids in A. limacinum. Purple LED light (410–420 nm) significantly enhanced the biosynthesis of β-carotene, canthaxanthin, and astaxanthin compared to white light or darkness. Additionally, the use of glycerol or crude glycerol, a byproduct of biodiesel production, affected carotenoid profiles. β-Carotene and canthaxanthin were more abundant in cultures supplemented with glycerol or crude glycerol, whereas astaxanthin was predominantly produced in cultures supplemented with glucose. Notably, crude glycerol treated via acidification and activated carbon enabled cell growth and metabolite production comparable to that observed with glucose. These findings suggest that light quality and carbon source synergistically influence carotenoid biosynthesis in A. limacinum. The use of industrial byproducts combined with light optimization represents a sustainable and cost-effective strategy for enhancing microbial production of high-value compounds.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-37313-4.

Introduction

Thraustochytrids (family Thraustochytriaceae, order Thraustochytriales) are widely distributed in marine and brackish water environments^1–4. They have a complex life cycle that may vary among genera². One of their ecological characteristics is their high biomass yield during cultivation, as well as their ability to produce and accumulate n-3 fatty acids, primarily docosahexaenoic acid (DHA)^5–7.

Among them, Aurantiochytrium limacinum has attracted particular attention for its ability to produce DHA and carotenoids, including β-carotene, canthaxanthin, and astaxanthin^1,6,8–10. These compounds have significant physiological roles, such as anti-inflammatory and antioxidant effects, and are in high demand as functional ingredients^11–15. These valuable compounds have traditionally been obtained from fish oil or through chemical synthesis; however, these sources present challenges related to sustainability, safety, and social demand underscoring the need for microbial alternatives^16,17. Compared to other DHA-producing microorganisms such as Schizochytrium, or astaxanthin-producing algae Haematococcus pluvialis, A. limacinum is notable for its ability to simultaneously synthesize both types of high-value compounds under heterotrophic conditions^18–20. This unique capacity to co-produce multiple valuable metabolites makes A. limacinum a promising microbial platform for biotechnological applications.

Various cultivation strategies have been investigated to enhance the productivity of high-value metabolites such as carotenoids^21–24. Among these, white light irradiation has been reported to significantly increase astaxanthin accumulation in A. limacinum²⁵. This strategy offers a controllable and potentially low-cost means of enhancing yields without altering medium composition. While blue or purple light–induced enhancement of astaxanthin production has been observed in algae^26–28 and fungi²⁹, the wavelength-dependent regulation of carotenoid production in A. limacinum remains poorly understood. Most previous studies have relied on white light, and systematic evaluations of specific wavelengths—including purple light (410–420 nm)—have not been performed. As a result, the optimal irradiation conditions for maximizing carotenoid biosynthesis in this species remain undefined.

In contrast, A. limacinum is a heterotrophic organism that requires organic carbon sources such as glucose for growth, and the high cost of cultivation has been a major barrier to its industrial application^30,31. To address this issue, the exploration of alternative carbon sources has been actively pursued for thraustochytrids, and their growth has been demonstrated using various low-cost, waste-derived byproducts such as spent yeast from beer factories³², residues from shochu distillation³³, and sorghum stalks³⁴. Crude glycerol, a major byproduct of biodiesel fuel production, has been overproduced due to increasing biodiesel demand, creating a disposal issue^35–38. Recent studies have demonstrated robust growth of A. limacinum when cultivated with crude glycerol^35,39,40. However, the effects of crude glycerol on carotenoid biosynthesis and the associated metabolic adaptations in A. limacinum remain largely unknown, as previous studies have focused primarily on growth and lipid accumulation rather than carotenoid metabolism.

Importantly, no previous study has simultaneously examined the effects of specific light wavelengths, particularly purple LED light and treated crude glycerol on A. limacinum physiology. This integrated approach distinguishes our work from prior studies that evaluated these factors independently. Furthermore, the combined use of wavelength-specific irradiation and industrial waste–derived carbon sources provides new insight into both metabolic regulation and the development of cost-effective, sustainable production processes.

In this study, we investigated the effects of specific light wavelengths on the production of neutral lipids and carotenoids in A. limacinum, as well as its production characteristics when cultured with crude glycerol, an industrial byproduct, as a carbon source. By clarifying how purple light (410–420 nm) and treated crude glycerol interact to modulate metabolite production, our study contributes both to basic understanding of A. limacinum metabolism and to the development of sustainable bioprocesses for carotenoid production.

Materials and methods

Cell growth under different light irradiation and carbon source conditions

Aurantiochytrium limacinum ATCC MYA-1381^T was obtained from the American Type Culture Collection (Rockville, MD, USA). Cells were precultured in By + V/2 medium (ATCC Medium 790, Viesalt (Nihonkaisui Co., Ltd., Tokyo, Japan) 17.5 g/L) or in By + V/2 medium in which glucose was replaced with 0.5% (w/v) glycerol (referred to as By + V/2-Gly) under dark conditions at 25 °C with stirring at 150 rpm for 48 h. The preculture medium, adjusted to an OD₆₀₀ of 2.0, was inoculated into fresh medium (total volume, 40 mL). The cultures were maintained under three different light conditions: dark, white fluorescent light (light intensity: 50 µmol photons m^− 2 s^− 1), and purple LED light (410–420 nm, 50 µmol photons m^− 2 s^− 1) at 25 °C with shaking at 150 rpm. For cultures under white fluorescent or purple LED light, the cells were incubated under a repeated cycle of 14 h light and 10 h dark. The light intensity (50 µmol photons m⁻² s⁻¹) and wavelength range (410–420 nm) of the purple LED were selected based on our previous optimization and the specifications of the LED equipment. These parameters were newly and specifically measured in the present study. Cell growth was evaluated by measuring OD₆₀₀ every 24 h using a spectrophotometer UV‐1800 (SHIMADZU, Kyoto, Japan).

Measurement of neutral lipid in A. limacinum cells

After culturing for 72 h under each culture condition, 1 mL of the culture medium was centrifuged at 4 °C and 4,000 ×g for 8 min. The resulting suspension was diluted 10-fold. 750 µL of the diluted suspension (10-fold dilution), 240 µL of DMSO, and 10 µL of DMSO containing Nile Red (50 µg/mL, FUJIFILM Wako Pure Chemical, Japan) were mixed and incubated at 40 °C for 10 min. The fluorescence intensity at an excitation wavelength of 530–575 nm was measured using a spectrofluorometer F-2700 (Hitachi High-Tech, Japan). The fluorescence intensity values were normalized to the cell density based on the OD₆₀₀ measurements.

Carotenoid extraction and analysis from A. limacinum cells

The culture broth (20 mL) was centrifuged at 4 °C, 4,000 ×g for 8 min after 72 h of cultivation under three light conditions (dark, white, and purple LED) in By + V/2 or By + V/2-Gly media. For cultures under white fluorescent or purple LED light, the cells were incubated under a repeated cycle of 14 h light and 10 h dark. The supernatant was discarded, and the cells were sequentially washed with 5 mL of 50 mM sucrose solution and distilled water, followed by centrifugation under the same conditions as above. The cell pellet was resuspended in 1 mL of distilled water containing 3% (w/v) protease (Protin SD-NY10, Amano Enzyme, Japan) and incubated at 50 °C for 1 h. The cell suspensions were transferred to a screw-cap tube with 0.2 g of sterilized glass beads (φ = 0.1 mm and φ = 1.0 mm, Tomy Seiko, Tokyo, Japan) and disrupted using Micro Smash MS-100 (Tomy Seiko) at 4,000 rpm for five cycles, each lasting 15 s. Subsequently, 2 mL of chloroform/methanol solution (2:1, v/v) was added to the processed solution, vortexed for 1 min, and then shaken in the dark at 10 °C for 1 h. The mixture was centrifuged at 3,000 ×g, 4 °C for 5 min, and the extracted carotenoid layer was collected into glass vials.

The carotenoid pigments (β-carotene, canthaxanthin, and astaxanthin) were quantified using a high-performance liquid chromatograph (HPLC) MD-2018 (JASCO, Osaka, Japan) equipped with a JASCO PU-2089 pump and CrestPak C18T-5 column, as described previously⁴¹. Each carotenoid was quantified by comparing its retention time and peak area with those of the corresponding standard. The carotenoid content was calculated per milliliter of culture medium and normalized to cell density (OD₆₀₀ = 1.0).

Identification of carotenoids

The remaining carotenoid extract after HPLC analysis was evaporated and then redissolved in methanol for further analysis. LC-MS was performed using a Prominence LC system (SHIMADZU) equipped with a PDA detector SPM-M20A (SHIMADZU) and an ion-trap time-of-flight mass spectrometer (LCMS-IT-TOF, SHIMADZU). The extract was separated using a TSK gel ODS-80TsQA column (2.0 × 250 mm, Tosoh) heated to 40 °C. The mobile phases consisted of solution A (90% methanol, 10% deionized water, and 0.1% (w/v) ammonium acetate) and solution B (30% ethyl acetate, 70% methanol, and 0.1% (w/v) ammonium acetate). The gradient elution was programmed as follows: 100% A and 0% B for the first 5 min, followed by a linear transition to 0% A and 100% B over 15 min, which was maintained for 15 min. The composition was then returned to 100% A and 0% B for 5 min and maintained for another 5 min. Mass spectra were recorded in positive ion mode under the following conditions: detection voltage of 1.7 kV, APCI temperature of 400 °C, CDL temperature of 200 °C, and block temperature of 200 °C. Carotenoids were identified by comparing their m/z values and absorption spectra with those of known carotenoid molecular weights and reference spectra.

Purification of crude glycerol

Crude glycerol was obtained from Aburatou Shoji Co., Ltd. in Shiga, Japan, and several treatments were performed before it was used as a carbon source. Step 1 (Diluted): crude glycerol and distilled water were mixed in a 1:4 (v/v) ratio. Step 2 (Acid treated): The pH of the glycerol mixed solution was adjusted to 1 using HCl. The solution was then centrifuged at 5,000 ×g for 5 min at room temperature. After removing the supernatant, the solution was filtered through a GF/C filter (particle retention capacity: 1.2 μm, Whatman). The pH was then adjusted to 7 using 3 M NaOH. Step 3 (Activated carbon treated): to the filtered glycerol mixed solution, 50 g of dried activated carbon powder (Fuji Film Wako Pure Chemical) was added per liter, and the mixture was stirred at 150 rpm for 3 h. Afterward, it was centrifuged at 5,000 ×g for 5 min at room temperature, and the activated carbon powder was removed by filtration through a GF/C filter. The glycerol content was measured using GC-MS by Sumika Chemical Analysis Service in Osaka, Japan. The purity of the glycerol was evaluated using a spectrophotometer to measure the absorbance at a wavelength of 457 nm, as described previously⁴².

Cell growth and analysis of total neutral lipids and carotenoids with crude glycerol addition

Cells were precultured in By + V/2-Gly medium under dark conditions at 25 °C with stirring at 150 rpm for 48 h. One mL of precultured cells, adjusted to a cell density of OD₆₀₀ = 2.0, was added to 39 mL of By + V/2-Gly medium, which contained 75 g L^− 1 of crude glycerol treated at three stages (step 1, step 2, step 3), bringing the total volume to 40 mL. The three types of culture media were incubated at 25 °C with shaking at 150 rpm under purple LED light (410–420 nm, 50 µmol photons m^− 2 s^− 1), and growth (OD₆₀₀) was measured using a spectrophotometer every 24 h. For the two media (crude glycerol processed up to step 2 and up to step 3 had been added) that yielded sufficient biomass, carotenoid content was measured at 24-h intervals up to 120 h using the method described above, and neutral lipid content was measured at 72 h using the method described above.

Cell culture and library preparation for RNA-seq

The cells of A. limacinum ATCC-MYA-1381^T were precultured in By + V/2 medium (ATCC Medium, 790 By + Medium) or By + V/2-glycerol under dark conditions at 25 °C with stirring at 250 rpm for 48 h. The preculture medium adjusted to an OD₆₀₀ of 4.0 was inoculated into fresh medium (total volume 30 mL), and cultured under two light conditions: dark and purple LED PF5-75T (Nihon Ikakiki Seisakusho, the purple LED: 410 ~ 420 nm, 50 µmol photons m^{− 2} s^{− 1}) at 25 °C and 150 rpm for 24 h in an artificial climate chamber (LH-410SCPF-S, Nihon Ikakiki Seisakusho). One mL of culture medium was collected from each condition after incubation for 24 h, and total RNA was extracted following the method reported in a previous study²⁵. Residual DNA was removed using TURBO DNase (Thermo Fisher Scientific, Waltham, MA, USA) following standard protocol. For samples that still contained DNA, additional DNase treatment was performed using RNA Clean & Concentrator-5 with DNase I (ZYMO Research, CA, USA). The quality of the RNA samples was evaluated using 5200 Fragment Analyzer System and Agilent HS RNA Kit (Agilent Technologies). Library preparation and sequencing (2 × 150 bp; DNBSEQ-T7 (MGI Tech)) were performed by Seibutsu Giken Inc. (Kanagawa, Japan).

RNA-seq analysis and identification and annotation of expressed genes

Low reads acquired from each library were trimmed using fastp version 0.23.4 36 and mapped to the reference genome of A. limacinum ATCC MYA-1381^T (Joint Genome Institute, Project ID: 402021) using HISAT2 version 2.2.1 37⁴³ with default setting. Quantification for each gene from trimmed reads were using StringTie version 2.2.1 38⁴⁴, and the expression value defined by transcripts per million (TPM).

Comparison of transcript levels

Genes involved in carotenoid biosynthesis and lipid metabolism pathways were identified based on the annotation information provided in the KEGG pathway map registered at the JGI. Transcript levels of these genes were compared across the different culture conditions. Transcript abundance was evaluated using TPM and fold change, the latter calculated by normalizing the TPM of each gene under each condition to the TPM of the same gene in the dark–glucose condition, which served as the reference.

For metabolic pathway genes lacking an Enzyme Commission (EC) number in the KEGG map, gene function was inferred based on previously published studies. Specifically, we either consulted annotation information for this species reported in studies of a closely related thraustochytrid, Hondaea fermentalgiana CCAP 4062/3⁴⁵, or Aurantiochytrium sp. T66⁴⁶.

Statistical analyses

Neutral lipids and carotenoids were assessed using the Shapiro-Wilk test, and homogeneity of variance was evaluated using the F-test. Based on these tests, either the Student’s t-test or Mann-Whitney U test was applied as appropriate. All experiments were performed in biological triplicates (n = 3) to ensure reproducibility.

Results

Effects of different light conditions on cell growth, neutral lipids, and carotenoid contents

First, no significant changes in growth rate and cell density were observed under different carbon sources and light conditions (Student’s t-test, p > 0.05; Fig. 1a). No significant differences in neutral lipid content were observed at 72 h after culture initiation, regardless of the carbon source (glucose or glycerol) or light conditions (Student’s t-test, p > 0.05; Fig. 1b). In contrast, after 72 h of cultivation, carotenoid biosynthesis exhibited light-dependent variation, with the highest production observed under purple light, followed by white light. Specifically, the amount of β-carotene after 72 h increased in the order of dark, white light, and purple light for both carbon sources (glucose and glycerol) (p < 0.01; Fig. 2a). Under all three light conditions, glycerol significantly enhanced β-carotene production compared with glucose (dark: 5.50 ± 0.67 vs. 9.80 ± 0.93 ng/mL, white light: 9.18 ± 0.98 vs. 17.42 ± 1.08 ng/mL, purple light: 16.80 ± 1.17 vs. 32.30 ± 3.47 ng/mL; p < 0.05 for all comparisons). The amount of canthaxanthin after 72 h also increased in the order of dark, white light, and purple light for both carbon sources (glucose and glycerol) (p < 0.01; Fig. 2b). Under all light conditions, glycerol significantly enhanced canthaxanthin production compared with glucose (p < 0.01). Notably, under purple light, the amount of canthaxanthin synthesized with glycerol was approximately twice that with glucose (glucose: 27.53 ± 0.26 ng/mL, glycerol: 48.38 ± 6.18 ng/mL; p < 0.01). In contrast, although both β-carotene and canthaxanthin biosynthesis were consistently higher with glycerol as the carbon source, astaxanthin production showed the opposite trend, being greater with glucose. For both carbon sources (glucose and glycerol), the amount of astaxanthin after 72 h increased in the order of purple light, white light, and dark conditions (p < 0.05; Fig. 2c). Although astaxanthin production under dark and white light conditions was higher with glucose than with glycerol, no significant difference was observed between the carbon sources (p > 0.05; Fig. 2c). In contrast, astaxanthin synthesis was significantly higher with glucose than with glycerol under purple light (glucose: 33.20 ± 1.98 ng/mL, glycerol: 16.14 ± 2.42 ng/mL, p < 0.01; Fig. 2c). These results collectively demonstrate that light quality and carbon source distinctly affect carotenoid composition without altering growth or neutral lipid accumulation.

Fig. 1.

Growth and neutral lipid production of A. limacinum under different light and carbon source conditions. (a) Growth curves based on OD₆₀₀ values. Each point represents the OD₆₀₀ value under specific culture conditions, as indicated in the legend. (b) Neutral lipid content per cell. Bars are grouped by culture medium: the first three bars correspond to cultures in By + V/2 medium (glucose) under dark (black), white light (white), and purple light (purple) conditions, and the following three bars represent cultures in By + V/2 + Gly medium (glycerol) under the same light conditions. All error bars indicate the standard deviation of the mean (n = 3).

Fig. 2.

Carotenoid contents of A. limacinum cultured under different light and carbon source conditions. (a) β-Carotene, (b) canthaxanthin, and (c) astaxanthin contents were measured after 72 h of cultivation under three different light conditions and with two different carbon sources. All values were normalized per unit cell volume based on the OD₆₀₀ measurements. Data are presented as standard deviation (n = 3). Statistical significance between groups was assessed using Student’s t-test (*p < 0.01). All error bars indicate the standard deviation of the mean (n = 3).

Identification of carotenoids

Quantitative analysis by HPLC revealed eight peaks that were identified as carotenoids based on their mass-to-charge ratio (m/z), absorption spectra and retention times. The All-E carotenoid forms were identified by comparing their retention times with those of authentic standards, whereas the 13-Z or 13’-Z isomers were assigned based on their retention times and characteristic shapes of their absorption spectra. In order of increasing retention time, these compounds were tentatively assigned as All-E astaxanthin (Rt = 15.491 min), 13-Z astaxanthin (Rt = 17.129 min), All-E adonirubin (phenicoxanthin, Rt = 18.075 min), 13-Z or 13’-Z adonirubin (phenicoxanthin, Rt = 18.988 min), All-E canthaxanthin (Rt = 20.221 min), 13-Z canthaxanthin (Rt = 20.764 min), echinenone (Rt = 24.884 min), and β-carotene (Rt = 30.081 min) (Supplementary Fig. 1a and b). Additionally, for astaxanthin, adonirubin, and canthaxanthin, approximately 25% of the All-E form was detected as the 13-Z (13’-Z) isomer, based on the peak area (Supplementary Fig. 1c). The same set of carotenoids was detected in samples derived from cultures grown with glycerol as their carbon source.

Evaluation of crude glycerol as a sustainable carbon source for biomass and metabolite production in A. limacinum

The absorbance of crude glycerol, a byproduct of biodiesel fuel production, decreased as its purity increased after three treatments (Fig. 3a). The addition of crude glycerol treated with step 1 (diluted) resulted in cell growth; however, the growth rate was slower than that of cultures with glucose or glycerol, and the maximum cell density was significantly lower (Diluted: OD₆₀₀ = 2.323, glucose: 4.704, glycerol: 4.842, Student’s t-test, p < 0.01) (Fig. 3b). The growth suppression observed with step 1-treated glycerol suggests that dilution alone was insufficient to remove the inhibitory compounds. Using crude glycerol processed up to step 2 (acid treated) as the sole carbon source resulted in a longer growth period (72 h); however, the maximum cell density was not significantly different from that observed in glucose and glycerol cultures (Student’s t-test, p > 0.05) (Fig. 3b). Moreover, cultures with step 2 (acid treated) glycerol maintained the maximum cell density for a longer period beyond 96 h after inoculation (Fig. 3b). These results indicate that the growth-inhibitory effects of crude glycerol were progressively mitigated as the degree of purification increased. After using crude glycerol processed through step 3 (activated carbon treated) as the carbon source, it took 96 h to reach the maximum cell density, which was comparable to the maximum cell densities observed in glucose and glycerol cultures (Student’s t-test, p > 0.05) (Fig. 3b). While both step 2 and step 3 treatments enabled comparable maximum cell densities, the time required to reach peak growth, and its duration differed between the two. Thus, crude glycerol requires at least acid and activated carbon treatments to serve as an effective carbon source for sustaining cell growth.

Fig. 3.

Effects of crude glycerol purification on its physical properties and on the growth of A. limacinum using treated glycerol as a carbon source. (a) Photographs and absorbance (OD₄₅₇) of crude glycerol at various processing stages. From left to right: 5× diluted crude glycerol, acid-treated crude glycerol, activated carbon (AC)-treated crude glycerol, and the reagent glycerol. (b) Growth curves of A. limacinum cultured with glucose, glycerol, acid-treated crude glycerol, or AC-treated crude glycerol. OD₆₀₀ values were measured every 24 h from 24 to 120 h. (c) Neutral lipid contents per cell after 72 h of cultivation. All error bars indicate the standard deviation of the mean (n = 3).

Changes in neutral lipid amount due to the addition of treated crude glycerol

The amounts of neutral lipids per cell in cultures using crude glycerol as a carbon source treated up to step 2 (acid treated) and step 3 (AC treated) were comparable to those in cultures using glucose and glycerol, with no significant differences observed (glucose: 0.90 ± 0.08, glycerol: 1.08 ± 0.11, step2: 0.91 ± 0.08, step 3: 0.91 ± 0.06, Student’s t-test, p > 0.05, Fig. 3c).

Time-course of carotenoid biosynthesis with different carbon sources under purple light condition

In cultures using crude glycerol processed up to steps 2 and 3, the amount of β-carotene was significantly higher at all time points from 48 to 120 h than in cultures using glucose or glycerol (Student’s t-test, p < 0.05, Fig. 4a). In the culture using the medium supplemented with crude glycerol processed up to step 2 (acid-treated), the maximum amount of β-carotene (41.5 ng/mL·OD₆₀₀) was reached after 72 h (Fig. 4a). Additionally, in the medium supplemented with crude glycerol processed up to step 3, the β-carotene content was lower than that in cultures with step 2 at 72 h. However, it exceeded the maximum β-carotene value observed in step 2 (acid treated) at 96 h and reached 61.3 ng/mL·OD₆₀₀ at 120 h (Fig. 4a). The amounts of canthaxanthin per cell reached its maximum at 96 h in both step 2 and step 3, with values of 30.4 ng/mL·OD₆₀₀ and 34 ng/mL·OD₆₀₀, respectively (Fig. 4b). These values were slightly higher than those in cultures with glycerol, but lower than those in cultures with glucose. The amount of astaxanthin per cell was highest at 120 h when the cells were cultured in a medium supplemented with crude glycerol processed in steps 2 and 3 (Fig. 4c). These values were not significantly different from those obtained when the cells were cultured with glucose or glycerol (Student’s t-test, p > 0.05). In cultures using crude glycerol processed up to step 2, the amount of astaxanthin at 96 and 120 h were equivalent to that in cultures with glucose and glycerol. In contrast, in cultures using crude glycerol processed up to step 3, the amount of astaxanthin was only approximately 4.50 ng/mL·OD₆₀₀ at 96 h but increased to 77.32 ng/mL·OD₆₀₀ at 120 h, reaching a level comparable to that observed in cultures with other carbon sources (Fig. 4c).

Fig. 4.

Carotenoid contents of A. limacinum cultured using treated crude glycerol as the carbon source. (a) β-Carotene, (b) canthaxanthin, and (c) astaxanthin content per cell were measured every 24 h from 24 to 120 h during cultivation with four different carbon sources: glucose, glycerol, acid-treated crude glycerol, and activated carbon (AC)-treated crude glycerol. All values were normalized per unit cell volume based on the OD₆₀₀ measurements. Statistical significance at each time point was assessed using Student’s t-test (*p < 0.05, **p < 0.01). All error bars indicate the standard deviation of the mean (n = 3).

Expression of genes involved in lipid and carotenoid biosynthesis under different light and carbon-source culture conditions

A total of 24.7–24.9 million reads were obtained per sample, providing sufficient sequencing depth for descriptive transcriptomic analyses. Overall transcript distributions were highly similar among samples, and all major genes involved in glycerol metabolism, fatty acid biosynthesis, the PKS pathway, the mevalonate pathway, and carotenoid biosynthesis were detected under every condition, enabling pathway-level comparison of transcriptional trends. Global transcriptional profiles showed no large-scale shifts among conditions, but pathway-specific differences were evident and are described below.

Glycerol uptake gene expression

The transcript level of the glycerol uptake gene GUP (estExt_fgenesh1_pg.C_120144) increased markedly during glycerol cultivation under dark conditions, reaching nearly four times the level observed during glucose cultivation (Table 1). Similarly, genes involved in the conversion of imported glycerol to glycerol-3-phosphate, such as glycerol kinase GK-1 (fgenesh1_kg.16_#59#_isotig01436), and those involved in the subsequent oxidation of glycerol-3-phosphate to dihydroxyacetone phosphate (DHAP), including G3PDH-3 (gm1.7656_g) and G3PDH-4 (estExt_fgenesh1_kg.C_1_t10471), also showed substantial increases in transcript levels during glycerol cultivation (Fig. S2). Under purple light conditions, transcript levels generally decreased compared with dark conditions for each carbon source, except for a few genes such as GUP (Table 1).

Table 1.

Transcript levels of genes involved in carotenoid and lipid metabolism under different light and carbon-source culture conditions. Glu_D and Gly_D indicate cultures grown with glucose or glycerol, respectively, under dark conditions; Glu_P and Gly_P indicate cultures grown with glucose or glycerol, respectively, under purple LED light conditions.

Gene name (abbr)	Protein_ID	Model Name	Annotation	Glu_D	Gly_D	Glu_P	Gly_P
Glycerol metabolism pathway
GUP	132,254	estExt_fgenesh1_pg.C_120144	Glycerol facilitator	195.748	769.26	181.094	1708.427
GK-1	46,627	fgenesh1_kg.16_#_59_#_isotig01436	Glycerol kinase	147.696	711.414	171.97	853.217
GK-2	8349	gm1.8349_g	Glycerol kinase	18.501	12.405	17.146	19.223
G3PDH-1	71,806	e_gw1.3.902.1	glycerol-3-phosphate dehydrogenase (NAD+)	88.509	81.714	43.848	38.089
G3PDH-2	110,754	estExt_Genewise1Plus.C_190339	glycerol-3-phosphate dehydrogenase (NAD+)	30.437	23.463	29.415	17.408
G3PDH-3	7656	gm1.7656_g	Glycerol-3-phosphate dehydrogenase	129.334	504.581	70.573	174.333
G3PDH-4	136,473	estExt_fgenesh1_kg.C_1_t10471	Glycerol-3-phosphate dehydrogenase	32.852	211.506	28.024	219.085
TIM-1	42,275	fgenesh1_kg.2_#_3_#_isotig06918	Triosephosphate isomerase	515.905	399.975	417.142	270.225
TIM-2	45,035	fgenesh1_kg.10_#_22_#_isotig07893	Triosephosphate isomerase	275.476	145.056	266.37	170.463
Mevalonate pathway
ACAT-1	46,177	fgenesh1_kg.14_#_41_#_isotig05280	Acetyl-CoA C-acetyltransferase	88.348	49.231	69.05	131.412
ACAT-2	123,015	estExt_fgenesh1_pm.C_70068	Acetyl-CoA C-acetyltransferase	5.959	5.331	8.391	3.52
ACAT-3	137,176	estExt_fgenesh1_kg.C_30031	Acetyl-CoA C-acetyltransferase	316.375	119.002	320.359	57.664
HMGCS	48,616	fgenesh1_kg.29_#_70_#_isotig05214	Hydroxymethylglutaryl-CoA synthase	1715.248	350.705	2129.93	128.928
HMGCR-1	124,781	estExt_fgenesh1_pm.C_130228	Hydroxymethylglutaryl-CoA reductase (NADPH)	13.936	3.206	12.876	1.007
HMGCR-2	141,519	estExt_fgenesh1_kg.C_190076	Hydroxymethylglutaryl-CoA reductase (NADPH)	370.813	222.611	286.02	119.513
MVK	142,637	estExt_fgenesh1_kg.C_280013	Mevalonate kinase	237.359	114.127	229.392	82.861
MVAK2	76,700	e_gw1.9.328.1	Phosphomevalonate kinase	62.107	36.886	54.955	20.174
MVAD	46,997	fgenesh1_kg.18_#_18_#_isotig06081	Diphosphomevalonate decarboxylase	215.722	132.779	209.796	74.903
IDI	139,555	estExt_fgenesh1_kg.C_100226	Isopentenyl-diphosphate Delta-isomerase	230.606	163.139	212.054	139.156
Carotenoid biosynthesis pathway
GGPPS	140,061	estExt_fgenesh1_kg.C_120121	Geranylgeranyl pyrophosphate synthase/Polyprenyl synthetase	61.574	90.287	39.89	22.49
CrtIBY	150,841	estExt_Genemark1.C_230099	Squalene/phytoene synthase	14.801	15.997	4.164	3.451
CrtO	73,472	e_gw1.5.336.1	Phytoene desaturase	4.014	2.82	3.327	1.368
CrtZ	107,058	estExt_Genewise1Plus.C_10_t20199	Unknown	5.34	8.753	3.818	0.908
Fatty acid biosynthesis pathway
ACC	63,582	gw1.10.632.1	Acetyl-CoA carboxylase	115.385	131.26	139.337	48.262
FAS1	83,486	e_gw1.21.366.1	Fatty acid synthase subunit beta/Fas1-like, helical	120.303	80.656	150.449	34.39
FabD-1	62,164	gw1.18.381.1	[acyl-carrier-protein] S-malonyltransferase	19.553	11.372	13.547	6.431
FabD-2	72,672	e_gw1.4.94.1	[acyl-carrier-protein] S-malonyltransferase	4.758	5.098	5.832	6.943
FabD-3	131,064	estExt_fgenesh1_pg.C_9_t10004	[acyl-carrier-protein] S-malonyltransferase	25.426	12.36	24.32	13.626
FabF	44,432	fgenesh1_kg.8_#_26_#_isotig05580	Beta-ketoacyl-[acyl-carrier-protein] synthase I	57.451	36.996	41.733	25.7
FabG-1	11,390	gm1.11390_g	3-oxoacyl-[acyl-carrier-protein] reductase	109.059	27.385	59.226	35.612
FabG-2	81,423	e_gw1.17.462.1	3-oxoacyl-[acyl-carrier-protein] reductase	49.204	13.271	64.016	16.835
FabG-3	137,211	estExt_fgenesh1_kg.C_30066	3-oxoacyl-[acyl-carrier-protein] reductase	68.488	47.304	73.725	27.949
MECR	44,953	fgenesh1_kg.9_#_261_#_isotig05282	Trans-2-enoyl-CoA reductase (NADPH)	104.612	51.858	102.267	64.683
Δ9ΦΑΔ	77,652	e_gw1.10.221.1	Predicted E3 ubiquitin ligase	60.732	72.177	60.27	249.343
Δ6/Δ8ΦΑΔ	77,887	e_gw1.11.243.1	Acyl-CoA 6-desaturase	174.644	156.261	84.77	126.751
Δ4ΦΑΔ	140,026	estExt_fgenesh1_kg.C_120085	Cytochrome b5	59.766	40.3	94.919	80.284
ELO-1	73,494	e_gw1.5.731.1	Fatty acyl-CoA elongase/Polyunsaturated fatty acid specific elongation enzyme	47.266	38.29	37.228	24.783
ELO-2	138,338	estExt_fgenesh1_kg.C_60194	Fatty acyl-CoA elongase/Polyunsaturated fatty acid specific elongation enzyme	72.574	48.588	81.741	101.739
PUFA-A	36,775	fgenesh1_pg.14_#_251	Animal-type fatty acid synthase and related proteins	326.054	310.672	265.918	118.385
PUFA-B	140,546	estExt_fgenesh1_kg.C_140136	Animal-type fatty acid synthase and related proteins	68.927	77.734	67.097	36.34
PUFA-C	141,473	estExt_fgenesh1_kg.C_190026	FabA-like domain	172.935	218.298	140.192	162.484
Cryptochrome related genes
CRY-1	99,951	estExt_Genewise1.C_360044	Deoxyribodipyrimidine photolyase/cryptochrome	41.508	68.2	18.819	12.83
CRY-DASH	141,295	estExt_fgenesh1_kg.C_180070	Deoxyribodipyrimidine photolyase/cryptochrome	60.199	91.111	15.853	15.506

Fatty acid–related gene expression

Genes in the FAS pathway, particularly those related to saturated fatty acid synthesis (Table 1; ACC to MECR), generally exhibited decreased transcript levels during glycerol cultivation. The only gene showing an increased transcription trend was S-malonyl-CoA transferase FabD-2 (e_gw1.4.94.1). Genes in the PKS pathway (PUFA-A, PUFA-B and PUFA-C) showed a different pattern (Table 1). Under dark conditions, PUFA-B (estExt_fgenesh1_kg.C_140136) tended to exhibit higher transcription during glycerol cultivation. PUFA-C (estExt_fgenesh1_kg.C_190026) showed increased transcription during glycerol cultivation under both dark and purple light conditions. Under purple light conditions, genes in the FAS elongase/desaturase (elo/des) pathway (Table 1; Δ9FAD to ELO-2) tended to show higher transcript levels—particularly Δ4 desaturase FAD4 (estExt_fgenesh1_kg.C_120085) and elongase ELO-2 (estExt_fgenesh1_kg.C_60194) under both carbon sources. Genes related to fatty acid synthesis and their TPM values and fold changes are summarized in Table 1 and Figure S3.

Carotenoid-related gene expression

Genes related to the mevalonate pathway generally showed lower transcript levels during glycerol cultivation compared with glucose cultivation. Notably, hydroxymethylglutaryl-CoA synthase HMGCS (fgenesh1_kg.29_#70#_isotig05214) exhibited more than a four-fold lower transcript level under glycerol than under glucose cultivation (Table 1). Within the carotenoid biosynthesis pathway, under dark conditions, carotenoid hydroxylase CrtO (e_gw1.5.336.1) showed a slight trend toward higher expression during glucose cultivation. In contrast, the multifunctional carotenoid synthase CrtIBY (estExt_Genemark1.C_230099) and carotenoid hydroxylase CrtZ (estExt_Genewise1Plus.C_10_t20199) showed higher transcript levels during glycerol cultivation. Under purple light conditions, CrtIBY, CrtO, and CrtZ all exhibited an apparent decrease during glycerol cultivation (Table 1). No genes showed a trend toward increased transcription under purple light for either carbon source. To investigate the potential effects of purple light on photoreceptor signaling, we searched the JGI database for cryptochrome-related genes, which function as purple light receptors. Four candidate genes were identified (estExt_Genewise1.C_360044, estExt_fgenesh1_kg.C_180070, estExt_fgenesh1_kg.C_50119, and estExt_fgenesh1_kg.C_150066). All four genes tended to show higher transcript levels under dark conditions. Genes related to the mevalonate and carotenoid pathways, along with their TPM values and fold changes, are summarized in Table 1 and Figure S4.

Discussion

No differences were observed in the growth rate or maximum cell density due to variations in the carbon sources or light conditions (Fig. 1a). Previous studies culturing Aurantiochytrium limacinum under the same light intensity as in this study reported a reduction in light stress due to irradiation²⁵. In the present study, both white and purple light irradiation were examined, but neither showed significant effects on growth. Moreover, in the closely related species Aurantiochytrium sp. TZ209, culturing under higher light intensities (60 or 120 µmol photons m⁻² s⁻¹) than those used in this study (50 µmol photons m⁻² s⁻¹) led to a decrease in biomass, which correlated with light intensity⁴⁷. Taken together, these observations suggest that the light intensity used in this study did not exert adverse effects on growth and is appropriate for supporting carotenoid biosynthesis.

Furthermore, no significant differences in fatty acid biosynthesis were observed across the different culture conditions. In contrast, the levels of carotenoids—astaxanthin, canthaxanthin, and β-carotene—increased substantially under white light compared to that under darkness and were further increased under purple light. Because carotenoid biosynthesis proceeds via the mevalonate pathway, which shares acetyl-CoA with fatty acid biosynthesis^46,48, we initially hypothesized potential competition for acetyl-CoA. However, neutral lipid levels remained unchanged, likely because carotenoids constitute only a minor fraction relative to fatty acids.

In addition to these findings, previous studies have reported enhanced astaxanthin production in Aurantiochytrium through a variety of metabolic and environmental strategies. For example, overexpression of carotenoid biosynthetic genes markedly increased astaxanthin accumulation in A. limacinum⁴¹, while chemical mutagenesis has yielded high-carotenoid–producing strains with improved productivity⁴⁹. More recently, modulation of key enzymatic steps has been shown to alter carotenoid and lipid profiles in related species^21,50,51. Although these studies differ in strain background, cultivation conditions, and the use of genetically modified or mutant strains, the astaxanthin levels obtained in the present study under purple LED irradiation fall within the range of high productivity reported previously. This indicates that environmental optimization such as the strategic use of specific light wavelengths combined with an inexpensive carbon source can achieve competitively high carotenoid yields without the need for genetic manipulation.

Crude glycerol discharged from a biodiesel production facility was subjected to a three-step purification process consisting of dilution, acid treatment, and activated carbon (AC) treatment to make it suitable for the growth of A. limacinum. As a result, crude glycerol that had undergone acid and activated carbon (AC) treatments supported sufficient growth of A. limacinum. Although the growth rate was slightly slower than that observed when glucose or glycerol was used as the carbon source, the maximum cell density reached comparable levels (Fig. 3b). This finding is consistent with previous studies using glycerol or crude glycerol as a carbon source, suggesting that crude glycerol can serve as an alternative to glucose or pure glycerol under properly treated conditions³⁵. Carbon source concentration is an important factor in the growth of this species^39,40,52. Given that the nitrogen concentration was constant across all media, the high carbon content in crude glycerol may have contributed to the maintenance of A. limacinum growth. The variation in the time required to reach the maximum cell density depending on the treatment step may be due to the removal of certain growth-promoting substances during the activated carbon treatment. A previous study reported that small amounts of soap may enhance the growth of A. limacinum⁴⁰. Moreover, neutral lipid production was comparable between cultures using acid- or AC-treated crude glycerol and those using glucose or glycerol as carbon sources (Fig. 3c). These results are consistent with those of previous studies showing no significant difference in total lipid content when A. limacinum was cultured with glycerol or crude glycerol⁴⁰. Collectively, these results demonstrate that even simple purification such as pH adjustment by acid treatment is sufficient to support biomass accumulation and lipid production, underscoring the usefulness of treated crude glycerol as a sustainable and cost-effective carbon source.

In A. limacinum, glycerol is initially converted to glycerol-3-phosphate by glycerol kinase and then metabolized to glyceraldehyde-3-phosphate via dihydroxyacetone phosphate (DHAP), after which it enters the same glycolytic pathway as glucose^31,46. Based on this pathway and previous findings showing that total lipid production derived from acetyl-CoA does not significantly differ between cultures using glucose and glycerol as carbon sources⁴⁰, we hypothesized that carotenoid production, which also derives from acetyl-CoA, would exhibit a similar trend. However, our results revealed that although the maximum levels of astaxanthin and canthaxanthin were comparable among the tested carbon sources, β-carotene accumulated at significantly higher levels in cultures supplemented with glycerol and crude glycerol. In addition, the astaxanthin accumulation rate differed between glucose- and glycerol-based cultures, suggesting that carbon sources affect not only the amount but also the kinetics of carotenoid biosynthesis. The differences in accumulation kinetics may be partially explained by the varying efficiencies of carbon uptake and metabolic processing. It has been reported that glucose is more rapidly and efficiently utilized than other carbon sources in this species⁵³, possibly enabling a quicker satisfaction of the carbon demand required for carotenoid synthesis. Although the biological significance of enhanced β-carotene accumulation remains unclear, our results suggest that this pattern is not merely a consequence of the overall metabolic flux. Rather, it may reflect regulatory or enzymatic bottlenecks specific to carotenoid biosynthetic pathways.

We propose that in glucose-based cultures, the downstream conversion of β-carotene to astaxanthin via hydroxylation and ketolation occurs more efficiently than in other cultures^41,54. This results in the rapid accumulation of astaxanthin and relatively lower levels of β-carotene. In contrast, in cultures utilizing glycerol or crude glycerol, the rate of these enzymatic modifications may be slower, allowing β-carotene to accumulate more substantially before being converted to downstream products. This differential regulation may involve carbon source–dependent expression or activity of key enzymes such as β-carotene ketolase (CrtW) and hydroxylase (CrtZ)⁸. Regarding the biosynthetic pathway from β-carotene to astaxanthin, LC-MS analysis identified echinenone, canthaxanthin, and adonirubin as intermediate metabolites. Echinenone and canthaxanthin are products of CrtW, whereas adonirubin is generated by CrtZ.

To the best of our knowledge, this is the first report to demonstrate distinct patterns of carotenoid production depending on the use of glucose, glycerol, or crude glycerol in A. limacinum, providing potentially valuable insights for industrial applications. Although a detailed cost analysis was not performed in this study, the use of crude glycerol—a low-cost byproduct of biodiesel production—and energy-efficient LED lighting suggests that this strategy could reduce production costs compared to conventional glucose-based and continuous-light cultivation systems. Elucidating whether this effect is due to transcriptional regulation, post-translational modification, or differences in cofactor availability will be a critical next step. Further transcriptomic or proteomic analyses under different carbon source conditions may help to clarify the regulatory mechanisms underlying carbon flux partitioning in carotenoid biosynthesis. A deeper understanding of these dynamics will not only enhance the physiological insight into A. limacinum metabolism but also support targeted metabolic engineering for improved yields of specific carotenoids under cost-effective cultivation conditions. Taken together, our findings strongly suggest that purple light irradiation, in combination with glycerol or crude glycerol as a carbon source, can effectively enhance the production of valuable compounds by A. limacinum. However, the detailed metabolic changes that occur during purple light exposure and glycerol-based cultivation remain unclear. Elucidating these mechanisms will be essential for optimizing the practical use of this organism in industrial applications.

To complement the physiological measurements and clarify the underlying metabolic responses, RNA-seq analysis was carried out under each experimental condition. Under purple light irradiation, nearly all investigated genes exhibited lower transcript levels compared with the dark condition, regardless of the carbon source. This trend is consistent with a previous transcriptomic analysis comparing dark and white light conditions in A. limacinum, in which only 53 genes were upregulated upon light exposure, whereas 909 genes were downregulated²⁵. These results suggest that purple light, similar to white light, broadly suppresses transcription at the early growth stage.

Despite increased carotenoid accumulation at later stages, all carotenoid-biosynthesis genes showed lower transcript levels under purple light at 24 h. This contradiction may reflect temporal dynamics: early suppression followed by later induction, as shown previously for CrtIBY^25,55,56. A similar temporal pattern may apply to downstream genes such as CrtO and CrtZ. Interestingly, Cryptochrome-related genes also showed higher transcript levels in darkness. This may reflect negative photoregulation, where light suppresses photoreceptor-gene expression to avoid excessive signaling^29,55,56, or a protective response to photooxidative stress under short-wavelength light^25,57–59. Although statistical evaluation was not possible without replicates, these trends imply that cryptochrome regulation in A. limacinum may differ from classical photoreceptor models.

During glycerol cultivation, TAG- and PKS-related genes were upregulated, consistent with DHA being synthesized exclusively through the PKS pathway⁵⁷ and its higher proportion during glycerol cultivation^53,58. However, neutral lipid levels did not differ significantly, indicating that transcriptional upregulation alone was insufficient to increase lipid accumulation.

This discrepancy can be explained by the reduced expression of upstream pathways. Genes involved in glycolysis, the FAS pathway, the mevalonate pathway, and carotenoid synthesis were all downregulated under glycerol cultivation. Because glucose is the most efficiently utilized carbon source⁵³, glycerol metabolism may reduce energy availability. Downregulation of the TCA cycle may also limit acetyl-CoA supply for fatty acid synthesis, creating a metabolic bottleneck⁵⁹.

Although both glycerol and purple light suppressed early transcription of lipid- and carotenoid-related genes, final carotenoid levels were comparable between glucose and glycerol cultures, suggesting that slower early metabolism may delay—but not prevent—carotenoid accumulation. Purple light may further enhance carotenoid production at later stages, despite early transcriptional suppression.

Future work should incorporate biological replicates, time-series transcriptomics, RT-qPCR validation, multi-omics analyses, and functional assays to clarify how light and carbon sources regulate metabolic flux in A. limacinum. A deeper understanding of these regulatory layers will support the development of cost-effective strategies for carotenoid and DHA production.

Conclusions

Aurantiochytrium limacinum has emerged as a promising source of DHA and carotenoids; however, high cultivation costs remain a major limitation for its industrial application. In this study, we investigated the effects of light quality and carbon sources on the production of valuable compounds to identify cost-effective cultivation strategies. Purple light stimulated the carotenoid biosynthetic pathway, particularly enhancing astaxanthin production. Cultivation indicated that although astaxanthin and neutral lipid biosynthesis partially influence each other, there is no significant metabolic competition between these two pathways. Furthermore, crude glycerol, an inexpensive carbon source, supported cell growth and lipid and carotenoid production comparable to that observed under glucose-based conditions. Overall, our findings provide valuable insights into enhancing both the productivity and cost efficiency of A. limacinum cultivation, laying the groundwork for its future applications in the biotechnological production of DHA and carotenoids.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

K.Y., K.K. and S.K. designed this study and S.S. supervised the project. K.Y. performed cell culture and measured the neutral lipids and carotenoids. K.Y. and K.K. analyzed and interpreted the data. Y.M. and S.T. performed carotenoid identification using LC-MS. K.Y., K.K. and S.S. prepared the manuscript with substantial input from all the authors. All authors have read and approved the manuscript.

Funding

This work was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Numbers 22H02237 and 23K23504 awarded to S.S.

Data availability

The RNA sequencing reads were deposited in the National Center for Biotechnology Information under accession numbers SRR36180814 to SRR36180818. The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.