Spatial Engineering of Yarrowia lipolytica for the Biosynthesis of Capsanthin

ABSTRACT

Capsanthin is a high‐value tetraterpenoid widely utilized in the food, cosmetic, and pharmaceutical industries. However, its traditional production via plant extraction and chemical synthesis faces severe sustainability and cost constraints. The oleaginous yeast Yarrowia lipolytica (Y. lipolytica) offers a superior chassis due to its robust mevalonate pathway and abundant lipid droplets for pigment sequestration. In this study, we systematically engineered Y. lipolytica for the de novo biosynthesis of capsanthin. By heterologously co‐expressing of Capsicum annuum capsanthin/capsorubin synthase (CaCCS) and Arabidopsis thaliana zeaxanthin epoxidase (AtZEP), we achieved the first reported biosynthesis of capsanthin in Y. lipolytica. Combinatorial screening revealed that full‐length sequences of CaCCS and AtZEP yielded optimal production of 6.92 mg/L. To further enhance the productivity of capsanthin, we evaluated spatial compartmentalization strategies. The results demonstrated that targeting the biosynthetic enzymes to the peroxisome via Ser–Lys–Leu (SKL) tags successfully increased the titer to 8.8 mg/L by leveraging its specialized lipophilic and redox microenvironment. Finally, a 5‐L fed‐batch fermentation was conducted, achieving a maximum capsanthin titer of 48.15 mg/L. This work provides a sustainable and scalable alternative to plant extraction and chemical synthesis for capsanthin production.

Graphical Abstract and Lay Summary

For the first time, capsanthin was biosynthesized in Yarrowia lipolytica. Peroxisome targeting and fermentation optimization greatly boosted product yield.

Abbreviations

AtZEP

Arabidopsis thalianazeaxanthin epoxidase

CaCCS

Capsicum annuum capsanthin/capsorubin synthase

CaZEP

Capsicum annuum zeaxanthin epoxidase

CCS

capsanthin/capsorubin synthase

CrtY

lycopene cyclase

CrtZ

β‐carotene hydroxylase

DCW

dry cell weight

DMAPP

dimethylallyl pyrophosphate

ER

endoplasmic reticulum

ESI‐TOF MS

electrospray ionization time‐of‐flight mass spectrometry

FADH₂

reduced flavin adenine dinucleotide

FPP

farnesyl pyrophosphate

GGPP

geranylgeranyl pyrophosphate

GGPPS

geranylgeranyl pyrophosphate synthase

GRAS

generally recognized as safe

HMG‐CoA

3‐hydroxy‐3‐methylglutaryl coenzyme A

HPLC

high performance liquid chromatography

IPP

isopentenyl pyrophosphate

LDs

lipid droplets

MEP

methylerythritol 4‐phosphate

MVA

mevalonate

NADPH

nicotinamide adenine dinucleotide phosphate

OD₆₀₀

optical density at 600 nm

PTS1

peroxisomal targeting signal type 1

rDNA

ribosomal DNA

ROS

reactive oxygen species

SKL

serine‐lysine‐leucine

TP

transit peptide

ZEP

zeaxanthin epoxidase

1. Introduction

Capsanthin is a high‐value tetraterpenoid and the predominant carotenoid in red peppers (Capsicum annuum) [1]. Endowed with a unique 3‐hydroxy‐𝜅‐end group [2], it exhibits exceptional tinctorial strength and a vibrant hue, making it a staple ingredient in the food, cosmetic, and nutraceutical industries [3]. Beyond its role as a natural colorant, capsanthin has gained prominence for its potent biological activities. It displays superior antioxidant properties, effectively scavenging free radicals [4] and mitigating oxidative stress [5]. Recent pharmacological studies have further highlighted its potential in anti‐inflammatory treatments, obesity prevention [1], cardiovascular health [6], and antitumor therapies [7]. As global consumer preference shifts toward “natural and clean‐label” ingredients, the market demand for capsanthin continues to expand rapidly [8].

Despite its industrial significance, the sustainable supply of capsanthin remains a formidable challenge. Conventional production relies primarily on natural plant extraction and chemical synthesis. Extraction from botanical sources is constrained by the seasonality of pepper cultivation, fluctuating yields caused by climate change, and the intensive consumption of land and water resources. Furthermore, downstream processing requires significant quantities of organic solvents and complex purification, which often results in low recovery rates and concerns regarding pesticide residues [9]. Conversely, while chemical synthesis allows for large‐scale production, the complex molecular architecture of capsanthin, defined by multiple chiral centers and conjugated double bonds, renders the process technically arduous and cost‐prohibitive [10]. Moreover, synthetic pigments frequently fail to satisfy stringent “natural” certification requirements for food and pharmaceutical applications. Thus, developing a sustainable, eco‐friendly, and efficient microbial alternative is imperative [11].

In recent years, microbial biosynthesis has emerged as a promising frontier, with Escherichia coli (E. coli) serving as the inaugural heterologous host for capsanthin production. Although initial efforts in 2021 yielded only 0.5 mg/L [12], recent breakthroughs utilizing multi‐layered metabolic engineering have advanced titers in engineered E. coli to 6.77 mg/g dry cell weight (DCW) [13]. These milestones were achieved through integrated strategies such as the construction of chloroplast‐like thioredoxin systems to enhance the reducing power supply for zeaxanthin epoxidase (ZEP), the co‐expression of plant‐derived molecular chaperones to facilitate the functional folding of capsanthin/capsorubin synthase (CCS), and the implementation of SpyTag/SpyCatcher‐based enzyme scaffolding and membrane targeting to optimize metabolic flux and FADH₂ regeneration [13]. Despite these advancements, E. coli faces inherent physiological bottlenecks that constrain its production potential. As a highly hydrophobic molecule, capsanthin partitions extensively into the phospholipid bilayer of the cytoplasmic membrane. As E. coli lacks specialized subcellular organelles like lipid droplets for effective sequestration, excessive accumulation leads to membrane overloading, disrupted fluidity, and localized ion leakage, a phenomenon termed hydrophobic stress [14]. This imposes a significant metabolic burden, inhibits biomass accumulation, and establishes a rigid physiological ceiling on biosynthetic titers. Furthermore, the prokaryotic redox environment is often incompatible with plant‐derived enzymes involved in capsanthin synthesis, resulting in catalytic uncoupling [15] and the accumulation of reactive oxygen species (ROS). Additionally, its endogenous methylerythritol 4‐phosphate (MEP) pathway is subject to stringent feedback inhibition [16] and competes with primary metabolism for precursors, thus hindering the simultaneous attainment of high yield and cellular robustness [17].

In contrast, the naturally oleaginous yeast Yarrowia lipolytica (Y. lipolytica) offers a superior platform for hydrophobic pigment production [18]. Boasting a well‐developed endomembrane system and abundant lipid droplets (LDs) [19], Y. lipolytica has been observed to accumulate carotenoids within intracellular lipid droplets, which may help to mitigate membrane‐associated stress and product toxicity. Additionally, it natively employs the high‐flux mevalonate (MVA) pathway, which confers superior metabolic flexibility and sufficient precursor supply for terpenoid biosynthesis [20, 21]. As a eukaryotic host, its intracellular environment, including redox systems, chaperone networks, and membrane‐anchoring sites, is highly compatible with the functional expression of plant‐derived enzymes [22]. Moreover, Y. lipolytica exhibits excellent genetic tractability, supported by a versatile synthetic biology toolkit (e.g., CRISPR/Cas9 systems, strong promoters, and multi‐copy integration) [23]. Its ability to utilize low‐cost substrates like crude glycerol or waste oils [24], coupled with its “Generally Recognized as Safe” (GRAS) status [18], underscores its suitability for large‐scale industrial biomanufacturing. Given the aforementioned advantages of Y. lipolytica and the unmet need for a robust capsanthin production platform, the objective of this study is to systematically engineer Y. lipolytica for the de novo biosynthesis of capsanthin. As illustrated in Figure 1, the pathway involves a series of sequential enzymatic reactions: the cyclization of lycopene to β‐carotene by CrtY, followed by the hydroxylation to zeaxanthin by CrtZ [25]. The subsequent tailoring steps involve the epoxidation of zeaxanthin to antheraxanthin and violaxanthin, mediated by ZEP [26], and finally, the crucial conversion to capsanthin and capsorubin catalyzed by CCS [12].

FIGURE 1.

Metabolic engineering of Y. lipolytica for the synthesis of capsanthin. The complete biosynthetic pathway is partitioned into four distinct functional modules to facilitate systematic optimization. The endogenous mevalonate (MVA) pathway provides the fundamental precursors, isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP), from glucose via acetyl‐CoA. Module 1 (Lycopene Synthesis): Heterologous expression of CrtE/xdGGPPS, CrtB, and CrtI/CarB enables the conversion of farnesyl pyrophosphate (FPP) to lycopene via the intermediates. Module 2 (β‐carotene Synthesis): Cyclization of lycopene into (Zeaxanthin Synthesis): Hydroxylation of β‐carotene by β‐carotene hydroxylase (CrtZ) leads to the formation of zeaxanthin. Module 4 (Ketocarotenoid Tailoring): The final tailoring stage involves the epoxidation of zeaxanthin to antheraxanthin and violaxanthin by ZEP followed by the crucial cyclopentane ring formation catalyzed by CCS to produce the high‐value pigments capsanthin and capsorubin. Dashed arrows indicate multiple enzymatic steps; solid arrows represent single catalytic. ERG10, acetyl‐CoA C‐acetyltransferase; ERG13, HMG‐CoA synthase; HMG1, HMG‐CoA reductase; ERG12, mevalonate kinase; ERG8, phosphomevalonate kinase; ERG19, mevalonate pyrophosphate decarboxylase; IDI, IPP isomerase; ERG20, FPP synthase; GGPPS, geranylgeranyl pyrophosphate synthase; CCS, capsanthin/capsorubin synthase; ZEP, zeaxanthin epoxidase.

Recent studies have demonstrated the effectiveness of spatial compartmentalization strategies in Y. lipolytica for enhancing terpenoid production. For instance, peroxisomal engineering has been successfully applied to produce α‐humulene [27], limonene [28], and trans‐nerolidol [29], highlighting the potential of this organelle to serve as a subcellular factory for various classes of terpenoids.

In this study, we successfully established a synthetic pathway for the heterologous production of capsanthin in Y. lipolytica (Figure 1). By co‐expressing Capsicum annuum CCS (CaCCS) and Arabidopsis thaliana ZEP (AtZEP), we achieved the first reported biosynthesis of capsanthin in Y. lipolytica. To maximize production, we screened CCS and ZEP orthologs from different sources and individually truncated their transit signal peptides to identify the optimal enzyme pair. Furthermore, the titer was significantly enhanced through the implementation of a RIAD–RIDD protein scaffold and organelle‐targeted localization strategies. Finally, through fed‐batch fermentation, the capsanthin titer reached a maximum of 48.15 mg/L. This study reports the first successful synthesis of capsanthin in Y. lipolytica, thereby providing a feasible strategy for its green and scalable production.

2. Materials and Methods

2.1. Culture Conditions

Y. lipolytica strains were cultivated in yeast peptone dextrose (YPD) medium at 28°C with agitation at 200 rpm. E. coli strains were grown in super optimal broth (SOB) medium supplemented with ampicillin (Amp) or Kanamycin (Kan) at 37°C and 250 rpm. Agar plates for YPD and SOB were made by incorporating 2% agar into the respective media. Selection of Y. lipolytica transformants was conducted using synthetic complete medium plates deficient in uracil (SC‐Ura) or leucine (SC‐Leu). The Cre‐loxP site‐specific recombination system was used to recover the URA3 marker. YPD medium was composed of 10 g/L yeast extract, 20 g/L peptone, and 20 g/L glucose. For solid plates, 20 g/L agar was added before autoclaving. The medium was autoclaved at 115°C for 20 min and cooled to approximately 55°C before use. SOB medium contained 20 g/L tryptone, 5 g/L yeast extract, 0.5 g/L NaCl, 0.186 g/L KCl, 2.03 g/L MgCl₂·6H₂O, and 2.46 g/L MgSO₄·7H₂O. For solid SOB agar plates, 20 g/L agar was included prior to autoclaving. After autoclaving and cooling to 55°C, filter‐sterilized antibiotics were added to final concentrations of 100 µg/mL Amp and 50 µg/mL Kan.

The shake flask fermentation of Y. lipolytica began by inoculating a single colony from a YPD plate and inoculating it into 15 mL of YPD liquid medium. The culture was incubated in the shaking incubator at 28°C and 200 rpm overnight for 16 h. Subsequently, this seed culture was transferred to 35 mL of fermentation medium, ensuring an initial optical density at 600 nm (OD₆₀₀) of 0.5. The shake flask was then placed back in the shaking incubator and incubated at 28°C and 200 rpm for 144 h. Besides, the fermentation medium (YPD, YPD40, YPD60, andYPD80) contains 20 g/L peptone, 10 g/L yeast extract, and 20, 40, 60, or 80 g/L glucose.

2.2. Plasmid Construction

Plasmids use LEU2 as the selection marker and include a JMP62 backbone, upstream homologous arm of the integration site, loxP‐LEU2‐loxP fragment, and downstream homologous arm of the integration site. To express genes, one to three gene expression cassettes were inserted between the homologous arms and the loxP‐LEU2‐loxP fragment via single‐step cloning.

The plasmids pJMP62‐LEU‐18S‐28S rDNA were stored in the laboratory and used to integrate endogenous or exogenous gene expression cassettes into the 18S‐28S sites. Primers in this study were designed using SnapGene and synthesized by GenScript (Nanjing, China). The gene elements were amplified using polymerase chain reaction (PCR) with 2 × Phanta Max Master Mix (Vazyme, Nanjing, China) and the primers. The resulting plasmid fragments were then ligated into plasmids with homologous sequences targeting different integration sites using the Hieff Clone Universal II One Step Cloning Kit (10923ES) (Yeasen Biotechnology, Shanghai, China). The plasmids all carried a kanamycin resistance marker and were transformed into E. coli XL‐10, followed by selection on SOB plates containing kanamycin. The successful integration of the target fragments into the plasmids was verified by colony PCR. Positive clones were extracted using a plasmid extraction reagent (Sangon Biotech, Shanghai, China) and sent to Sangon Biotech (Shanghai, China) for sequencing. The successfully sequenced plasmids were then linearized for transformation. The codon‐optimized genes and primers are shown in Tables S1 and S2, respectively.

2.3. Strain Construction

The plasmids (Table 2) were linearized using NotI, and transformed into the auxotrophic strain Po1f‐Δku70 (Leu⁻, Ura⁻ ) using the Frozen‐EZ Yeast Transformation II Kit (ZYMO RESEARCH). Transformants were selected on SC‐Leu plates and confirmed via PCR to yield the engineered Y. lipolytica strains listed in Table 1.

TABLE 2.

Plasmids used in this study.

Plasmids	Description	Sources
pJMP62’1	Kan, Leu2 marker, TEF promoter and LIP2 terminator	Our laboratory
pJMP62’2	Kan, Leu2 marker, GPD2 promoter and XPR2 terminator	Our laboratory
pS1	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐TLIP2‐PGPD2‐AtZEP‐TXPR2	This study
pS2	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐TLIP2‐PGPD2‐CaZEP‐TXPR2	This study
pS3	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐TLIP2‐PGPD2‐Tr59 AtZEP‐TXPR2	This study
pS4	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐TLIP2‐PGPD2‐Tr49 CaZEP‐TXPR2	This study
pS5	pJMP62’1,18S‐28S rDNA‐PTEF‐Tr52 CaCCS‐TLIP2‐PGPD2‐AtZEP‐TXPR2	This study
pS6	pJMP62’1,18S‐28S rDNA‐PTEF‐Tr52 CaCCS ‐TLIP2‐PGPD2‐CaZEP‐TXPR2	This study
pS7	pJMP62’1,18S‐28S rDNA‐PTEF‐Tr52 CaCCS ‐TLIP2‐PGPD2‐Tr59 AtZEP‐TXPR2	This study
pS8	pJMP62’1,18S‐28S rDNA‐PTEF‐Tr52 CaCCS ‐TLIP2‐PGPD2‐Tr49 CaZEP‐TXPR2	This study
pZ1	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐RIAD‐TLIP2‐PGPD2‐AtZEP‐TXPR2	This study
pZ2	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐RIDD‐TLIP2‐PGPD2‐AtZEP‐TXPR2	This study
pZ3	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐TLIP2‐PGPD2‐AtZEP‐RIAD‐TXPR2	This study
pZ4	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐TLIP2‐PGPD2‐AtZEP‐RIDD‐TXPR2	This study
pZ5	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐RIAD‐TLIP2‐PGPD2‐AtZEP‐RIDD‐TXPR2	This study
pZ6	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐RIDD‐TLIP2‐PGPD2‐AtZEP‐RIAD‐TXPR2	This study
pSKL	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐SKL‐TLIP2‐PGPD2‐AtZEP‐SKL‐TXPR2	This study
pER	pJMP62’1,18S‐28S rDNA‐PTEF‐CaCCS‐ER‐TLIP2‐PGPD2‐AtZEP‐ER‐TXPR2	This study

TABLE 1.

Strains used in this study.

Strain	Characteristics	Source
E. coli XL‐10	endA1, glnV44, recA1, thi‐1, gyrA96, relA1, lac, Hte, Δ(mcrA)183, Δ(mcrCB‐hsdSMR‐mrr)173, Tn10 (Tetʳ), Amy, Camʳ.	This study
Y. lipolytica Po1f	ATCC MYA‐2613, ura3‐302, leu2‐270, xpr2‐322, axp2­ΔNU49	ATCC1
Y. lipolytica Po1f‐Δku70	Derived from Po1f, KU70 deleted	This study
Y. lipolytica Zea‐1	ΔKu70, E2::PTEF‐CarRPY27R‐TXPR2‐ PTEF‐CarB‐TXPR2, A3:: PMNDH2‐tHMG1‐TXPR2‐ PTEF ‐GGPPSa‐TXPR2, F1‐3:: PTDH‐MVAE‐TICLt‐ PMNDH2 ‐MVAS‐ TXPR2, E3:: PTEF‐IDI‐TXPR2‐ PTDH –ERG20MTF88S‐ TICLt, B1:: PTEF‐CarRPY27R‐TXPR2‐PTEF‐CarB‐TXPR2, SQS1‐50 bp, A2:: PMNDH2‐tHMG1‐TXPR2‐ PTEF ‐GGPPSa‐TXPR2, B3:: PTEF‐ERG12‐TXPR2‐ PTEF‐IDI‐TXPR2, F4:: PTEF‐crtYB‐TXPR2, E53:: PTEF‐DGA1‐TXPR2, RAS2:: PTEFin‐PaCrtZ‐Tlip2t‐PGPD‐Trp1‐XPR2t	Our laboratory
Y. lipolytica S1	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐lip2t, PGPD2‐AtZEP‐xpr2t	This study
Y. lipolytica S2	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐lip2t, PGPD2‐CaZEP‐xpr2t	This study
Y. lipolytica S3	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐lip2t, PGPD2‐Tr59 AtZEP‐xpr2t	This study
Y. lipolytica S4	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐lip2t, PGPD2‐Tr49 CaZEP‐xpr2t	This study
Y. lipolytica S5	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐lip2t, PGPD2‐AtZEP‐xpr2t	This study
Y. lipolytica S6	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐Tr52 CaCCS‐lip2t, PGPD2‐CaZEP‐xpr2t	This study
Y. lipolytica S7	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐Tr52 CaCCS‐lip2t, PGPD2‐Tr59 AtZEP‐xpr2t	This study
Y. lipolytica S8	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐Tr52 CaCCS‐lip2t, PGPD2‐Tr49 CaZEP‐xpr2t	This study
Y. lipolytica Z1	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐RIAD‐lip2t, PGPD2‐AtZEP‐xpr2t	This study
Y. lipolytica Z2	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐RIDD‐lip2t, PGPD2‐AtZEP‐xpr2t	This study
Y. lipolytica Z3	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐lip2t, PGPD2‐AtZEP‐RIAD‐xpr2t	This study
Y. lipolytica Z4	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐lip2t, PGPD2‐AtZEP‐RIDD‐xpr2t	This study
Y. lipolytica Z5	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐RIAD‐lip2t, PGPD2‐AtZEP‐RIDD‐xpr2t	This study
Y. lipolytica Z6	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐RIDD‐lip2t, PGPD2‐AtZEP‐RIAD‐xpr2t	This study
Y. lipolytica Cap1‐SKL	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐SKL‐lip2t, PGPD2‐AtZEP‐SKL‐xpr2t	This study
Y. lipolytica Cap2‐ER	Zea‐1, Δ18S‐28S rDNA::LEU2, PTEFin‐CaCCS‐ER‐lip2t, PGPD2‐AtZEP‐ER‐xpr2t	This study

¹ ATCC, American Type Culture Collection (Manassas, VA, USA).

2.4. Quantification of Carotenoids and Intermediates

For extraction of intracellular carotenoids, 200 µL of culture broth was transferred into a 1.5 mL centrifuge tube and centrifuged at 12,000 rpm for 1 min. After discarding the supernatant, the cell pellet was resuspended in 1 mL of acetone containing a small amount of glass beads, followed by vigorous shaking for 30 min to ensure complete extraction of intracellular metabolites. The acetone extraction efficiency was assessed, and the cell pellets turned white after extraction, confirming the complete recovery of pigments into the acetone phase. The organic phase was collected and subjected to subsequent quantitative analysis. Blank control experiments confirmed that the culture medium was devoid of endogenous carotenoids. Under identical extraction and chromatographic conditions, no carotenoid peaks were detected in the uninoculated medium, ruling out any potential matrix interference and ensuring that all subsequent quantitative measurements reflected only carotenoids produced by the cultured organisms. Prior to HPLC analysis, all extracts were filtered through 0.22 µm nylon syringe filters. Chromatographic analysis was performed using an HPLC system (SHIMADZU LC‐20AT) equipped with a C18 reversed‐phase column (250 mm × 4.6 mm, 5 µm; Thermo Fisher Scientific, Waltham, MA, USA). Isocratic elution was carried out with a mobile phase consisting of acetonitrile, methanol, and isopropanol at a volume ratio of 5:3:2. The flow rate was maintained at 1.0 mL/min, the detection wavelength was set at 450 nm, and the column temperature was held at 35°C. Under these conditions, the retention times of capsanthin and zeaxanthin were 3.4 min and 4.0 min, respectively. Target compounds were identified by comparison of retention times and spectral profiles with authentic carotenoid standards. Quantification was performed by peak area integration based on pre‐established calibration curves for each analyte.

2.5. Fed‐Batch Fermentation in a 5 L Bioreactor

Fed‐batch fermentation was conducted in a 5‐L bioreactor with an initial working volume of 2.5 L. A single colony was isolated from a YPD agar plate and inoculated into a 50‐mL sterile flask containing 15 mL of YPD medium. The culture was grown overnight at 28°C and 200 rpm until the OD₆₀₀ reached approximately 8. This primary seed culture was then used to inoculate eight 250‐mL flasks, each containing 35 mL of YPD medium, at an initial OD₆₀₀ of 0.05. These secondary seed cultures were incubated at 28°C and 200 rpm for 24 h to reach an OD₆₀₀ of approximately 12. The secondary culture was then inoculated into the 5‐L bioreactor containing 2.5 L of sterile fermentation medium, with the starting OD₆₀₀ adjusted to 0.2. Fermentation was performed at 25°C with a constant aeration rate of 2.0 VVM and an initial agitation speed of 400 rpm. The pH was maintained at 5.5 by the 4.4 mol/L NH₄OH. A DO‐stat feeding strategy was employed for glucose feedback control in this study. This strategy uses the dissolved oxygen (DO) concentration as an indirect indicator of glucose depletion: when glucose in the fermentation broth is exhausted, microbial metabolic activity drops sharply, leading to a decrease in the oxygen uptake rate and a consequent rapid rise in DO concentration. In the present experiment, DO was initially set at 100% and DO‐stat feeding operation commenced when the DO dropped to the 15% threshold. Once the DO concentration spontaneously rose above 15%, a peristaltic feeding pump was automatically triggered to pulse a glucose solution into the bioreactor. Upon glucose addition, microbial metabolism resumed, causing the DO to fall back below the setpoint. Through the alternating cycle of DO oscillating between the lower limit and the 15% threshold, intermittent glucose supply was achieved. The total fermentation lasted 180 h. Samples were collected every 12 h to monitor OD₆₀₀ and residual glucose concentration.

3. Results

3.1. Construction of the Capsanthin Biosynthetic Pathway in Y. lipolytica

To achieve the heterologous production of capsanthin in Y. ipolytica, an artificial metabolic pathway was constructed using a laboratory‐maintained zeaxanthin‐producing strain Y. lipolytica Zea‐1 as the parental strain. For the genetic implementation, we designed a multigene expression cassette (Figure 2b) harboring the CaCCS and AtZEP genes (Figure 2a). These genes were placed under the control of the strong constitutive promoters pTEF and pGPD2 [30, 31], respectively. To achieve robust and high‐fidelity heterologous expression at the genetic level, the engineered biosynthetic cassette was site‐specifically integrated into the ribosomal DNA (rDNA) locus of the Y. lipolytica genome. Episomal plasmid systems are frequently plagued by segregational instability. In contrast, this rDNA‐based strategy relies on homologous recombination to exploit the highly repetitive characteristics of rDNA clusters. As a result, this approach enables the efficient multi‐copy integration of target pathway genes [32, 33]. The elevated gene dosage significantly enhances the expression levels of key enzymes, which effectively relieves rate‐limiting reactions, redirects carbon metabolic flux, and guarantees the high‐yield biosynthesis of end products. Furthermore, the structural identity and successful biosynthesis of capsanthin in the engineered strains were rigorously confirmed through comprehensive analytical characterization. Electrospray ionization (ESI) time‐of‐flight (TOF) mass spectrometry (MS) [34] analysis was performed in positive electrospray ionization mode. As shown in Figure 2c, both the capsanthin standard and the yeast‐derived product displayed a dominant molecular ion peak at m/z 585, corresponding to the [M+H]⁺ ion of capsanthin. The MS/MS fragmentation patterns of the product were consistent with those of the standard, with characteristic fragment ions at m/z 109, 493 and 567 [35]. This comparative analysis unequivocally confirms the successful biosynthesis of capsanthin in the engineered Y. lipolytica strain.

FIGURE 2.

Construction and analytical validation of the capsanthin biosynthetic pathway in engineered Y. lipolytica. Schematic of the heterologous biosynthetic route for capsanthin and capsorubin. The pathway utilizes lycopene as the starting substrate, sequentially converted to β‐carotene, zeaxanthin, antheraxanthin, and violaxanthin. (a) Architecture of the multigene expression cassette for rDNA‐mediated genomic integration. The expression of CaCCS and AtZEP is controlled by the constitutive pTEF and pGPD2 promoters, respectively. The cassette is flanked by 18S and 28S rDNA targeting sequences to facilitate high‐copy‐number integration into the Y. lipolytica genome, with the Leu2 gene serving as the auxotrophic selection marker. (b) Mass spectrometry (MS) characterization of the target products. The MS spectra (positive mode) of the authentic standard (left) and the fermentation product (right) both display a predominant molecular ion peak at m/z 585 [M+H] ⁺, providing definitive structural verification of the heterologous capsanthin. (c) MS, mass spectrometry.

3.2. Screening of Optimal Enzyme Combinations

To systematically identify the most catalytically efficient enzyme combination for capsanthin biosynthesis, a series of combinatorial expression cassettes were constructed, encompassing CCS and ZEP genes derived from different plant origins along with their N‐terminal transit peptide (TP) truncated variants. Specifically, the CCS gene was sourced from Capsicum annuum (CaCCS), while the ZEP genes were obtained from both Arabidopsis thaliana (AtZEP) and Capsicum annuum (CaZEP). In plants, both CCS and ZEP are chloroplast‐localized proteins, each harboring an N‐terminal transit peptide sequence that directs their transmembrane translocation into the chloroplast. However, these transit peptides are typically not recognized or correctly processed in prokaryotic hosts, and may also interfere with proper protein folding and subcellular localization in eukaryotic heterologous hosts. Therefore, whether to retain or remove the N‐terminal TP sequence has a significant impact on the functional expression of these heterologous enzymes [12]. Based on this rationale, both the full‐length sequences and the truncated versions lacking the predicted TP‐coding regions were constructed for the above genes, thereby generating a library of enzyme variants encompassing different sequence forms. A total of eight distinct recombinant strains (designated S1 to S8) were generated by systematically pairing the full‐length and truncated variants of CaCCS, AtZEP, and CaZEP (Figure 3a). All expression cassettes employed uniform strong constitutive promoters and the rDNA locus integration strategy to minimize potential confounding effects arising from variations in expression strength and copy number. Among these, strain S1, which co‐expressed the full length CaCCS and the full length AtZEP, exhibited the highest capsanthin titer, reaching 6.92 mg/L (Figure 3a). In contrast, strains expressing any truncated version of either enzyme consistently showed lower production levels, irrespective of the origin or pairing combination.

FIGURE 3.

Eight different engineered strain combinations. The bar graph presents the capsanthin yields (mg/L) across eight different engineered strain combinations (S1–S8) after fermentation. The left panel details the specific gene expression cassettes for each strain, involving CCS from Capsicum annuum (CaCCS) and ZEP from either Arabidopsis thaliana (AtZEP) or Capsicum annuum (CaZEP). “Tr” denotes a truncated version of the N‐terminal targeting signal peptide. (a) Schematic of gene expression cassettes and genomic integration: Illustration of the DNA constructs used for transformation. (b) The fragments are integrated into the host genome via homologous recombination mediated by 18S rDNA and 28S rDNA sequences. The constructs include a Leu selection marker and are driven by strong constitutive promoters (pTEF and pGPD2).

3.3. RIAD/RIDD‐Based Protein Scaffolding System

To mitigate the diffusion of unstable intermediates (antheraxanthin and violaxanthin), we first employed the RIAD/RIDD‐based protein scaffolding system [36]. The RIAD–RIDD system is a scaffold‐free modular enzyme assembly strategy based on a pair of short, high‐affinity peptide tags, which was originally developed to enhance cascade bio‐catalysis and metabolic flux by constructing artificial multienzyme complexes in microbial cell factories. Specifically, RIDD is a 44‐amino‐acid peptide derived from cAMP‐dependent protein kinase A, which spontaneously forms stable dimers under physiological conditions; RIAD, by contrast, is an 18‐amino‐acid amphiphilic peptide derived from kinase A anchoring proteins, capable of specifically binding to the RIDD dimer to form a stable trimeric structure.

By fusing the target enzymes with RIAD and RIDD tags respectively, the two enzymes can be efficiently assembled into a physical complex in vivo, thereby shortening the spatial distance between them, reducing the diffusion loss of reaction intermediates, and improving the efficiency of substrate transfer between cascade enzymes—an effect that has been well verified in both prokaryotic and eukaryotic hosts. In our study, we constructed four strains (strains Z1–Z4) to test whether the peptide scaffolds RIAD and RIDD have any effect on the expression of CCS and ZEP. To test whether the RIAD–RIDD peptide scaffolds could promote the cascade reaction between ZEP and CCS by enhancing their physical proximity, we constructed two strains (strains Z5 and Z6) with different combinations of RIAD–RIDD‐tagged CCS and ZEP, aiming to verify the effect of the scaffold system on the expression and catalytic coordination of these two key enzymes in capsanthin synthesis. As shown in Figure 4b, the control strain S1 produced 6.92 mg/L of capsanthin and 25.53 mg/L of zeaxanthin. The capsanthin titers of the other six strains showed no statistically significant difference compared with the un‐scaffolded control.

FIGURE 4.

Construction of RIAD‐RIDD‐based synthetic protein scaffolds and their impact on capsanthin production. Schematic of the RIAD–RIDD‐based enzyme co‐localization assembly strategy. The diagram illustrates the engineered synthetic protein scaffolds (strains Z1–Z6) constructed by fusing polypeptide affinity tags. CaCCS and AtZEP are terminal‐fused with either a RIAD (blue wavy line) or RIDD (green double‐ring) polypeptide tag. Z1–Z4 serve as single‐tag fusion controls or configurations unable to form complete multimers. Z5 and Z6 are designed to exploit the high‐affinity interaction between RIAD and RIDD to drive the formation of stable multienzyme complexes of CaCCS and AtZEP within the cytoplasm, aiming to achieve spatial physical co‐localization. (a) Fermentation metabolite production and cell growth analysis of engineered strains. (b)The chart presents the fermentation results for the optimal non‐scaffolded control strain (S1) alongside the engineered strains carrying the synthetic scaffolds (Z1–Z6). Left Axis (Production, mg/L): The purple bars represent the production of the final target product, capsanthin; the blue bars represent the accumulation of the upstream intermediate precursor, zeaxanthin. Right Axis (OD600): The yellow line represents the cell biomass at the end of fermentation.

3.4. Subcellular Localization

Peroxisomes and the endoplasmic reticulum (ER) were selected as the primary subcellular targeting sites for heterologous capsanthin biosynthesis in Y. lipolytica based on their respective advantages: the peroxisome provides a rich pool of acetyl‐CoA and NADPH, and the ER offers a hydrophobic microenvironment conducive to carotenoid accumulation. The SKL and KDEL signal sequences employed in this study are not speculative elements; SKL is the well‐established classical motif of the peroxisomal targeting signal type 1 (PTS1) [37, 38], while KDEL is the recognized signal for endoplasmic reticulum retention/retrieval [39, 40]. Compartmentalizing pathway enzymes within peroxisomes has proven effective in enhancing the production of various terpenes and terpenoids [41]. Peroxisomes, ubiquitous single‐membrane‐bound organelles in eukaryotic cells, represent an ideal subcellular compartment for heterologous capsanthin biosynthesis in Y. lipolytica. As a naturally oleaginous yeast, Y. lipolytica possesses abundant and metabolically active peroxisomes. These organelles are tightly coupled with fatty acid β‐oxidation and the mevalonate (MVA) pathway, two key metabolic processes that supply the requisite acetyl‐CoA and terpenoid precursors for carotenoid synthesis. Notably, heterologous enzymes can be directed to peroxisomes via specific targeting signals, among which the C‐terminal PTS1 motif, featuring a consensus SKL sequence, is the most highly efficient and widely employed. In this study, the terminal of CCS and ZEP were fused with PTS1 tags to achieve precise peroxisomal localization [42, 43]. As a parallel control for the spatial compartmentalization strategy, this study also systematically evaluated the feasibility and potential of the ER as an alternative subcellular targeting site. The ER, the most extensively distributed and largest single‐membrane endomembrane system in eukaryotic cells provides an exceptionally vast membrane surface area through its highly folded sheet‐like and tubular network architecture [44]. In this study, the C‐termini of CCS and ZEP were fused with KDEL tags to achieve precise endoplasmic reticulum localization. As illustrated in Figure 5b, distinct subcellular targeting strategies exerted markedly different effects on capsanthin biosynthesis. Relative to the control strain S1, the peroxisome‐targeted strain Cap1‐SKL achieved a 29% increase in capsanthin titer, reaching 8.8 mg/L, along with a concomitant 15% elevation in the precursor zeaxanthin pool. Conversely, the ER‐targeted strain Cap2‐ER exhibited only a marginal improvement, yielding a capsanthin titer of 7.21 mg/L without a concomitant increase in the zeaxanthin pool. Notably, the three strains exhibited comparable OD₆₀₀ values, thereby excluding cell growth or biomass accumulation as contributing factors to the enhanced capsanthin production. Quantitative analysis further revealed that the peroxisome‐targeted strain (Cap1‐SKL) exhibited only a marginally higher conversion efficiency of zeaxanthin to capsanthin (22.7%) compared to the cytosolic control (21.3%). However, the most notable change was the substantial increase in the total zeaxanthin pool (from 25.5 to 30.0 mg/L), which increased concomitantly with the capsanthin titer (Figure 5b). Collectively, these observations support the conclusion that the enhanced capsanthin production in Cap1‐SKL is primarily driven by improved precursor supply that pulls upstream flux, rather than by a direct enhancement of CCS/ZEP catalytic activity.

FIGURE 5.

Impact of subcellular compartmentalization on capsanthin biosynthesis and cell growth. (a) Schematic diagram of the biosynthetic pathway and subcellular localization of key enzymes (ER, endoplasmic reticulum; CaCCS, capsanthin/capsorubin synthase from Capsicum annuum; AtZEP, zeaxanthin epoxidase from Arabidopsis thaliana; SKL, peroxisomal targeting signal). (b) The bar chart illustrates the production of capsanthin (orange) and the accumulation of the precursor zeaxanthin (yellow) in engineered Y. lipolytica strains. The biosynthetic enzymes were targeted to either the peroxisome or the ER. The blue line represents the corresponding cell biomass measured at OD₆₀₀.

3.5. Fed‐Batch Fermentation Dynamics in a 5‐L Bioreactor

In the 5‐L bioreactor scale‐up experiment, the engineered strain demonstrated robust production potential, as evidenced by the deep orange–red color of the fermentation broth (Figure 6a). According to the fermentation kinetics (Figure 6b, c), cell growth (OD₆₀₀) increased steadily during the first 120 h before reaching a plateau, while glucose was rapidly consumed to support both biomass accumulation and carotenoid biosynthesis. Zeaxanthin, the primary precursor, reached a peak titer of approximately 337.73 mg/L at 120 h. While capsanthin production exhibited a characteristic lag phase, with significant accumulation beginning after 72 h and reaching approximately 48.15 mg/L by the end of the 168‐h fermentation period. The observed decline in zeaxanthin levels after 144 h, coupled with the simultaneous rise in capsanthin titer, clearly indicates the effective in vivo conversion of zeaxanthin into capsanthin via the heterologous pathway. These results confirm that the controlled bioreactor environment successfully enhances the production of capsanthin, although further optimization of the conversion efficiency from zeaxanthin remains a key target for future yield improvements.

FIGURE 6.

Fermentation kinetics of capsanthin and zeaxanthin production in a 5‐L bioreactor. Visual appearance of the fermentation broth in the 5‐L bioreactor after 180 h of cultivation, showing a distinct deep orange–red color. (a) Time‐course accumulation of capsanthin (red bars) and zeaxanthin (yellow bars) production. The stacked bar chart represents the total carotenoid composition over time. (b) Detailed fermentation profiles including cell growth (OD₆₀₀, green circles), glucose concentration (Glu, blue squares), zeaxanthin titer (yellow inverted triangles, left y‐axis), and capsanthin titer (red triangles, right y‐axis) (c).

4. Discussion

This study reports the first successful synthesis of capsanthin in Y. lipolytica. Notably, while the targeted capsanthin was successfully accumulated, capsorubin was not detected in the intracellular extracts. The exclusive accumulation of capsanthin, coupled with the complete absence of capsorubin, can primarily be attributed to substrate competition and metabolic bottlenecks at the antheraxanthin node. In this engineered pathway, the epoxidation catalyzed by ZEP occurs in two sequential steps, with the second step (antheraxanthin to violaxanthin) typically acting as a rate‐limiting barrier in heterologous hosts [45]. Concurrently, the introduced CaCCS strongly competes with ZEP for their shared intermediate, antheraxanthin. Given the potent catalytic pull and potential higher substrate affinity of CaCCS toward antheraxanthin, the intracellular metabolic flux is rapidly and preferentially redirected into the capsanthin branch. Consequently, this competitive consumption rapidly depletes the antheraxanthin pool, restricting the accumulation of violaxanthin, which is the indispensable precursor for capsorubin, thereby precluding its downstream synthesis [46]. These analytical results provide definitive evidence for the successful de novo biosynthesis of capsanthin in the Y. lipolytica chassis.

In heterologous yeast systems, truncation of plant derived N‐terminal transit peptides is widely adopted as a standard strategy to avoid misfolding, cytosolic aggregation, or aberrant mitochondrial targeting, issues commonly triggered by plastid specific signal sequences. Contrary to this conventional paradigm, our results demonstrate that retention of the full‐length enzyme sequences (as in strain S1) leads to superior capsanthin biosynthesis. This unexpected advantage of the native sequences may be because the N‐terminal regions of these plant enzymes may harbor hitherto unrecognized motifs that are integral to the conformational stability of the catalytic core; removal of these regions could inadvertently disrupt intramolecular interactions essential for proper folding or allosteric regulation, thereby impairing enzymatic turnover. Our observation that truncated variants consistently underperform supports the hypothesis that these domains contribute to structural integrity beyond simple subcellular targeting. Furthermore, the marked superiority of AtZEP over CaZEP in pairing with CaCCS highlights that optimizing enzyme origin is as critical as structural tuning. Even though both are functional ZEP orthologs, AtZEP apparently exhibits better compatibility with the Y. lipolytica intracellular environment, including its redox balance, chaperone network, and membrane anchoring systems, leading to higher epoxidation efficiency and a more favorable flux toward capsanthin. Taken together, the synergistic pairing of full‐length CaCCS and full‐length AtZEP establishes the most robust biosynthetic scaffold for capsanthin production in Y. lipolytica. Consequently, all subsequent strain optimization efforts should leverage this highly efficient genetic configuration as the foundation for further pathway engineering.

After establishing the optimal enzyme combination, we further explored the possibility of overcoming the pathway bottleneck from the perspective of spatial engineering. We first employed the RIAD–RIDD scaffolding system to enhance the spatial proximity between ZEP and CCS at the molecular level, aiming to facilitate their cascade reaction [36, 46]. A panel of tagged strains was generated to assess the impact of this system on enzyme expression and catalytic coordination. We constructed four strains (Z1–Z4) to evaluate whether the RIAD and RIDD peptide scaffolds affect the expression of CCS and ZEP. The results showed that the capsanthin titers of strains Z1–Z4 did not differ significantly from those of the control strain S1, indicating that the short peptides RIAD and RIDD had no discernible impact on the expression of either CCS or ZEP. Nevertheless, the capsanthin titers obtained from the scaffolded strains did not differ significantly from those of the scaffold‐free control. This finding indicates that although the RIAD–RIDD‐mediated spatial proximity between AtZEP and CaCCS may have facilitated the directional transfer of reaction intermediates, physical proximity alone is insufficient to overcome the central rate‐limiting bottleneck of this pathway. In other words, within this cascade reaction system, the primary constraint on metabolic flux toward capsanthin is not the diffusion efficiency of intermediates, but rather the inherent catalytic capacity of CCS itself. Therefore, the RIAD–RIDD scaffolding strategy alone can hardly resolve this metabolic node at a fundamental level. Furthermore, the overall inefficacy of the RIAD–RIDD system in Y. lipolytica may also be attributed to fundamental physiological incompatibilities between the engineered heterologous pathway and the host chassis. Given that these affinity tags were originally developed and optimized for model organisms such as E. coli and Saccharomyces cerevisiae, their expression levels, folding fidelity, and structural stability may all be significantly compromised when deployed in this nonconventional yeast host. Meanwhile, the forced trimeric assembly induced by the RIAD/RIDD system may drive the enzyme complex into a catalytically unfavorable conformation, rendering the assembled structure functionally suboptimal. This non‐preferred conformation may further impair the efficient substrate channeling between ZEP and CCS, such that even if the two enzymes are successfully brought into close physical proximity, their catalytic synergy cannot be substantially improved.

Given the limited efficacy of the molecular‐level spatial proximity strategy, we shifted the focus of spatial engineering from nanoscale enzyme proximity to a more macroscopic subcellular compartmentalization approach, investigating the impact of distinct organellar microenvironments on the terminal biosynthetic efficiency of capsanthin. Specifically, we systematically compared the efficiency of peroxisome‐targeting versus ER‐targeting strategies for the terminal biosynthetic enzymes CCS and ZEP in Y. lipolytica. The superior performance of the peroxisome‐targeted strain over its ER‐targeted counterpart highlights the unique advantages of peroxisomes as a metabolic platform for heterologous carotenoid production in Y. lipolytica. This peroxisomal compartmentalization strategy offers multiple distinct advantages for metabolic optimization. First, peroxisomes function as specialized microcompartments that effectively sequester the engineered enzymes and their lipophilic substrates. This spatial confinement minimizes the diffusion loss of intermediates and elevates the local concentration of reactants, thereby enhancing overall catalytic efficiency [43]. Second, the unique internal environment of peroxisomes, characterized by a specific balance of reducing equivalents, is highly conducive to the correct folding and functional activation of plant enzymes like CCS and ZEP [47]. These heterologous enzymes typically exhibit suboptimal stability and activity within the highly divergent redox landscape of the bulk yeast cytoplasm. Third, peroxisomal sequestration mitigates potential product toxicity and feedback inhibition [48]. By isolating highly hydrophobic molecules like capsanthin from the delicate plasma membrane, this strategy prevents lipid bilayer disruption and preserves overall cellular robustness. Finally, the intrinsic metabolic coupling between peroxisomes and the MVA pathway in Y. lipolytica guarantees a continuous supply of terpenoid precursors [43], directly fueling the metabolic flux toward capsanthin. Although the peroxisomal approach proved empirically superior, the ER targeting strategy was investigated based on a robust and distinct mechanistic rationale. As an extensive eukaryotic membrane network, the ER supplies abundant anchoring sites critical for the proper structural assembly and sustained catalytic activity of membrane‐associated enzymes. However, our experimental data indicate that, within this specific metabolic framework, the ER compartmentalization is ultimately less efficient at driving capsanthin accumulation than the highly specialized, precursor‐rich microenvironment of the peroxisome.

In the 5‐L bioreactor scale‐up experiment, the engineered strain exhibited robust production potential, as visually evidenced by the deep orange‐red coloration of the fermentation broth (Figure 6a). Analysis of the fermentation kinetics (Figure 6c) revealed that cell growth (OD₆₀₀) increased steadily over the first 120 h before reaching a stationary phase, while glucose was rapidly consumed to sustain both biomass accumulation and carotenoid biosynthesis. Zeaxanthin, the major upstream precursor, reached a peak titer of approximately 337.73 mg/L at 120 h. In marked contrast, capsanthin production displayed a pronounced lag phase, with appreciable accumulation initiating only after 72 h and reaching a final titer of approximately 48.15 mg/L by the end of the 168‐h fermentation cycle. Notably, the progressive decline in zeaxanthin levels after 144 h, which coincided with a corresponding rise in capsanthin titer, clearly signifies the efficient in vivo conversion of zeaxanthin to capsanthin via the engineered heterologous pathway. Collectively, these results demonstrate that a controlled bioreactor environment can substantially enhance capsanthin production. Nevertheless, the conversion efficiency from zeaxanthin to capsanthin represents a critical bottleneck, and its further optimization remains a key objective for future yield improvement.

In conclusion, this study establishes Y. lipolytica as a highly capable platform for the heterologous production of capsanthin, effectively overcoming the physiological limitations typically encountered in prokaryotic hosts. Our systematic engineering approach revealed that preserving the native N‐terminal sequences of CaCCS and AtZEP is crucial for achieving optimal enzymatic activity in Y. lipolytica. Furthermore, we demonstrated that targeting the biosynthetic pathway to the peroxisome significantly outperforms the RIAD–RIDD scaffolding strategy, as the peroxisomal microenvironment naturally sequesters toxic intermediates and furnishes a highly compatible redox landscape for plant‐derived enzymes. Finally, scale‐up of the optimal engineered strain in a 5‐L bioreactor achieved a maximum capsanthin titer of 48.15 mg/L. Although the substantial accumulation of the zeaxanthin precursor underscores the need for future directed evolution of CCS to enhance its catalytic turnover, this study provides a vital, environmentally sustainable framework for the large‐scale biomanufacturing of high‐value natural pigments.

Author Contributions

Jialei Zhang: conceptualization, methodology, investigation, formal analysis, writing – original draft, writing – review and editing, visualization, software, and data curation. Ling Sun: investigation, validation. Liang Zhang: writing – review and editing, validation. Chongyang Din: writing – review and editing, investigation. Zhenghua Gu: writing – review and editing, investigation. Lei Chen: writing – review and editing, validation. Sha Xu: conceptualization, supervision, funding acquisition, project administration, and resources.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: biot70291‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article.