Advances in microbial biosynthesis of indirubin

Abstract

Indirubin (IND), a bisindole alkaloid with remarkable pharmacological activities, has attracted significant attention in the pharmaceutical field due to its antileukemic, anti-inflammatory, and immunomodulatory properties. Currently, indirubin primarily relies on plant extraction and chemical synthesis, which are hindered by complex processes, low yields, and poor environmental compatibility. These challenges pose serious obstacles to clinical supply and sustainable industrial development. In recent years, microbial synthesis technology, which is based on synthetic biology and metabolic engineering, has provided a novel approach for the efficient production of indirubin. This method offers several advantages, including high efficiency, environmental sustainability, and eco-friendliness. Hence, this manuscript systematically summarizes the biosynthesis mechanisms of indirubin, the catalytic characteristics of key enzymes, the construction strategies of engineered bacteria, and the progress in fermentation condition. To address challenges such as the difficult separation of indirubin isomers, by-product inhibition, and industrialization bottlenecks, potential solutions are proposed, aiming to promote the green biomanufacturing of indirubin.

1. Introduction

Indirubin is a principal bioactive component found in the traditional Chinese medicine (TCM) Indigo Naturalis (Qingdai in Chinese), classified as a bisindole alkaloid. It exhibits a range of biological activities, including anti-inflammatory (Huang et al., 2017), antibacterial (Yang et al., 2022), antiviral (Chang, Chang, Lu, Tsou, & Lin, 2012), and immunomodulatory (Shao et al., 2019) properties. Researchers discovered in the 1970s that indirubin is the active component with anti-leukemic effects in “Danggui Longhui Wan”—a traditional Chinese patent medicine (Hoessel et al., 1999). Clinical studies have shown that indirubin exerts a significant inhibitory effect on chronic myeloid leukemia (CML), with reliable clinical efficacy and minimal toxic side effects. Among 86 cases, the overall effective rate of indirubin in treating chronic myeloid leukemia was 82.6%, of which 25.6% of patients achieved complete remission, and no severe toxic side effects were observed (Wang et al., 2017).

However, indirubin faces severe challenges in clinical supply. In the early stage, indirubin was mainly obtained through extraction from indigo-containing plants such as Baphicacanthus cusia (Nees) Bremek., Isatis indigotica Fort., and Polygonum tinctorium Ait.. Due to its extremely low content in plants (usually less than 0.1%) and poor water solubility, the extraction and separation process is costly with extremely low yield, which cannot meet large-scale clinical needs. Although the subsequently developed chemical synthesis method has improved the yield to a certain extent, it still has two major problems (Qu et al., 2019, Huccetogullari et al., 2019, Chen et al., 2012). First, the synthesis process produces a large number of by-products that are difficult to separate, such as indigo and indirubin isomers, leading to difficulties in final product purification and reduced efficacy. Second, it relies on toxic reagents and harsh reaction conditions, resulting in severe environmental pressure and economic costs (Wang et al., 2025). The limitations of these traditional production models have seriously hindered the further clinical development and wide application of indirubin as a high-quality anti-cancer drug. Therefore, how to achieve green and efficient production of indirubin has become the focus of researchers' attention.

In recent years, biosynthesis has gradually become a key technical route for compound synthesis, owing to its core advantages of environmental friendliness, product specificity, and resource sustainability. Compared with chemical synthesis processes that rely on high temperatures, high pressures, and toxic catalysts, modern biosynthesis technology, supported by advancements in genetic engineering, metabolic regulation, and synthetic biology, enables the directed synthesis of target active molecules under mild reaction conditions through the construction of efficient microbial cell factories. This not only significantly reduces energy consumption and pollutant emissions during production but also effectively avoids the by-product formation commonly seen in chemical synthesis, thereby greatly improving product purity and preparation safety (Sun, Liu, Zhao, & Ang, 2015).

Looking back at the research on indirubin biosynthesis, since the elucidation of the indole hydroxylation pathway catalyzed by tryptophan monooxygenase in 1928, researchers have progressively uncovered the indirubin biosynthetic network in microorganisms. This includes precursor supply via the shikimic acid/tryptophan pathway, key oxidase systems such as cytochrome P450 and laccase, as well as heterodimerization regulatory mechanisms (Gray, 1928). Studies have demonstrated that through the construction of engineered Escherichia coli strains, targeted modification of key enzymes, and optimization of fermentation conditions, it is possible to significantly enhance indirubin yield—achieving levels up to 860.7 mg/L (Sun et al., 2022). Accordingly, this review presents a systematic compilation of high-yield indirubin-producing microbes and the critical enzymes involved in the microbes’ indirubin conversion and synthesis processes. It also addresses challenges related to isomer separation, by-product inhibition, and industrialization bottlenecks while proposing corresponding solutions. The aim is to provide theoretical support for the green biomanufacturing of indirubin and to establish a methodological framework for the intelligent biosynthesis of complex natural products. This endeavor aims to facilitate a transition in TCM active ingredient production towards low-carbon, precise, and intelligent methodologies.

2. Biosynthetic pathway of indirubin

The biosynthetic pathway of indirubin was illustrated in Fig. 1. Its primary substrates include glucose and tryptophan. Glucose undergoes a series of transformations through the glycolytic pathway, ultimately being converted into indole-3-glycerol phosphate (IGP). Subsequently, tryptophan is synthesized under the catalysis of tryptophan synthase (trpA/trpB) (Du, Yang, Luo, & Lee, 2018). Indole serves as the central compound within the entire indirubin biosynthetic pathway, primarily produced from IGP via the action of tryptophan synthase (trpA) or derived from tryptophan through endogenous tryptophanase (tnaA) (Ma, Zhang, & Qu, 2018).

Fig. 1.

Biosynthetic pathways of indirubin and indigo. Both CYP71B102 and P450 BM-3 belong to cytochrome P450 enzyme. PAMO: phenylacetone monooxygenase; FMO: flavin-containing monooxygenase; IGP: indole-3-glycerol phosphate; trpA, trpB: tryptophan synthase subunits; tnaA: tryptophanase; IsH: isatin hydrolase.

There are three principal pathways for converting indole to indirubin: 1) Indole is catalyzed by oxidase such as (e.g., FMO (Sun et al., 2022), CYP71B102 (Sagwan-Barkdoll, Kim, Berim, Gang, & Anterola, 2025), PAMO (Núñez-Navarro et al., 2022), P450 BM-3 (Li et al., 2000, Li et al., 2008) to form 2-hydroxyindole and 3-hydroxyindole, which subsequently dimerized to form indirubin; 2) Both 2-hydroxyindole and 3-hydroxyindole are oxidized to produce 2-oxoindole and 3-oxoindole respectively, which can then condense to generate indirubin. Alternatively, 2-oxoindole may react with isatin to form indirubin (Sagwan-Barkdoll, Kim, Berim, Gang, & Anterola, 2025); 3) Isatin is synthesized from indole through oxidation by specific enzymes, after which it reacts with either 3-oxoindole or 3-hydroxyindole to produce indirubin (Sagwan-Barkdoll et al., 2025, Wongsaroj et al., 2015).

It is noteworthy that during the reaction between isatin and 3-hydroxyindole, as well as during the dimerization process involving both forms of hydroxyindoles, a by-product known as indigo may form. Indigo acts as an isomer of indirubin and shares similar chemical properties with it.

3. Key enzymes in indirubin biosynthesis

Enzymes serve as the fundamental catalysts driving complex metabolic pathways. The key enzymes involved in the biosynthesis of indirubin can be primarily categorized into monooxygenase and dioxygenase systems.

3.1. Monooxygenase system

Monooxygenases are responsible for incorporating a single oxygen atom from molecular oxygen into the substrate, while reducing the other oxygen atom to water. This enzymatic process requires reducing equivalents such as NADH or NADPH, and typically depends on cytochrome P450, flavoproteins, or metal ions (including copper and iron) as catalytic centers. Currently, the identified monooxygenases involved in indirubin biosynthesis include flavin-containing monooxygenase (FMO) (Sun et al., 2022), toluene monooxygenase (TMO) (Wongsaroj et al., 2015, McClay et al., 2005, Rui et al., 2005), phenylacetone monooxygenase (PAMO) (Núñez-Navarro et al., 2022), and the cytochrome P450 enzyme system (CYP450) (Sagwan-Barkdoll, Kim, Berim, Gang, & Anterola, 2025).

3.1.1. Flavin-containing monooxygenase

FMO is a type of redox enzyme that is widely distributed in nature, primarily derived from Corynebacterium glutamicum and Methylobacterium extorquens. This NADPH-dependent enzyme specifically catalyzes the hydroxylation of indole at the 2- and 3-positions under aerobic conditions. Research has demonstrated that the catalytic efficiency of wild-type FMO exhibits relatively low catalytic efficiency, with a Kcat/Km value of only (0.22 ± 0.02)[L/(s·mmol)].

However, engineering modifications, such as alterations to the active site, flexible loop regions, and substrate tunnel structures, have successfully enhanced enzyme-substrate interactions, thereby enhancing catalytic efficiency and increasing indirubin yield. For instance, when examining bFMO from Methylobacterium extorquens, its N291T mutant exhibits a mutation in the flexible loop that widens the diameter of the substrate tunnel while improving both substrate affinity and NADPH binding, doubling catalytic efficiency. Additionally, the K223R mutant modifies the α4-β11 connecting loop, which strengthens π-π stacking interactions between indole and FAD (flavin adenine dinucleotide) and increases catalytic efficiency 2.5-fold. Notably, the double mutant K223R/D317S not only expands the diameter of the substrate tunnel, thereby reducing spatial hindrance for indole entry into the active pocket, but also forms new hydrogen bonds that further enhance interactions between indole and the FAD cofactor. This significant improvement raises catalytic efficiency to (1.45 ± 0.30)[L/(s·mmol)] —a remarkable increase of 6.6 times compared to wild-type levels—and achieves an impressive indirubin yield of 860.7 mg/L, marking it as one of the highest yields reported in literature thus far (Sun et al., 2022).

Meanwhile, a study by Xiamen University demonstrated that the FMO5B4 mutant obtained via directed evolution could enhance catalytic selectivity by 45-fold. When combined with oxygen-limited conditions, the proportion of indirubin increased from 10% to 35%, with a yield of 56 mg/L—representing a 2.5-fold improvement compared to the single-stage process (Shao et al, 2024). However, despite the fact that different FMO mutants have achieved enhanced indirubin yield and catalytic efficiency through structural modification or directed evolution, the yield of indirubin remains significantly lower than that of indigo (0.3 mg/L vs 1.3 mg/L). The primary reason lies in the fact that indirubin synthesis relies on the selectivity of free radical coupling: during the free radical coupling of indole-based precursors, the coupling pathway tends to favor indigo synthesis due to differences in thermodynamic energy barriers. This limits the efficiency of directed indirubin production, and this selectivity bottleneck urgently requires resolution.

3.1.2. Toluene monooxygenase

TMO is a multi-component monooxygenase that operates through dual iron active centers and is commonly found in bacteria capable of degrading aromatic hydrocarbons. In the biosynthesis of indirubin, the isozymes of TMO (T4MO and TOM) exhibit distinct catalytic properties. For instance, T4MO not only efficiently catalyzes the conversion of toluene to p-cresol with a regioselectivity reaching up to 96%, but it also oxidizes indole into 3-hydroxyindole and 2-hydroxyindole, further transforming these compounds into indirubin (Wongsaroj et al., 2015, McClay et al., 2005). The active site of the α-hydroxylation subunit (TomA) within TOM features a binuclear iron cluster (Fe-O-Fe), which plays a crucial role in mediating oxygen activation and regulating substrate binding orientation via key residues V106 and A113. For example, the TomA3 mutant (V106A/A113V) modifies product distribution by enlarging the hydrophobic cavity of the substrate binding pocket, thereby shifting indole's hydroxylation pathway from traditional 3-position to both the 2-position and 7-position. This phenomenon mainly occurs because the diiron center (Fe-O-Fe) in the TomA subunit binds to O₂ via Fe²⁺ to form a Fe³⁺-O-O-Fe³⁺ intermediate. The V106A mutation reduces the steric hindrance around the iron center, accelerates electron transfer, and ultimately leads to the shift of hydroxylation sites. The reductase component (TomB) of TOM contains an NADH-bound Rossmann fold (GxGxxG) and a [2Fe-2S] cluster, primarily responsible for facilitating electron transfer to the hydroxylation subunit. When utilizing either toluene or indole as substrates, this enzyme depends on NADH as a reducing cofactor and preferentially catalyzes ortho-hydroxylation of toluene to produce 3-methylcatechol in an oxygen-rich environment. Simultaneously, it can also oxidize indole to generate various pigment precursors, such as indigo, indirubin, and isatin (Rui, Reardon, & Wood, 2005).

The research demonstrated that in a single-phase system using Luria-Bertani medium, high-concentration indole significantly inhibits the growth of engineered bacteria, suppresses T4MO activity, and results in low indirubin production. In contrast, the water/dioctyl phthalate (DOP) biphasic system can reduce the toxicity of indole to bacterial strains and alleviate the issue of growth inhibition. Taking the biphasic system with 50% (volume percentage) DOP as an example, when 5 mmol/L indole is added, T4MO can oxidize indole more efficiently to generate higher amounts of 3-hydroxyindole and 2-hydroxyindole, with the indirubin yield reaching a maximum of 11.8 mg/L. However, in the tryptophan medium supplemented with cysteine, 5 mmol/L indole, and 1 mmol/L isatin, the indirubin yield can be increased to 102.4 mg/L, while preventing the conversion of indole into indigo—a byproduct with lower value (Wongsaroj, Sallabhan, Dubbs, Mongkolsuk, & Loprasert, 2015). This indicates that the biphasic system enables efficient conversion by dissolving high-concentration indole and reducing cytotoxicity. In the future, further integration of enzyme modification and metabolic engineering could facilitate the construction of a TMO-based biphasic system, thereby achieving high-yield and high-purity production of indirubin.

3.1.3. Phenylacetone monooxygenase

PAMO, a member of the Baeyer-Villiger monooxygenase family, is also crucial for indirubin biosynthesis. This enzyme is predominantly produced by Thermobifida fusca. The wild-type PAMO enzyme cannot use indole as a substrate to produce indigoid compounds such as indirubin. However, iterative saturation mutagenesis can be employed to expand the substrate scope of this enzyme. Specifically, using plasmid Ppamo_PAC as the template (this plasmid contains pre-existing mutations P440F, Q93N, and P94D), the QuikChange PCR method was applied to introduce random mutations of 2 amino acids at residues 441−444. Eventually, two mutants, PAMO_HPCD and PAMO_HPED, were obtained. These two variants are capable of using indole and its derivatives as substrates to catalyze the production of indigoid compounds including indigo and indirubin. Among them, in the presence of added L-tryptophan, PAMOHPCD can produce approximately 130.0 mg/L of indirubin (Núñez-Navarro et al., 2022).

It can be seen from this that this achievement can be combined with the strategy of T4MO’s biphasic system to alleviate indole toxicity and improve conversion efficiency. In the future, through multi-dimensional synergistic optimization involving “modification of enzyme catalytic properties-regulation of substrate toxicity-optimization of precursor supply”, it is expected to further break through the current yield bottleneck in indirubin biosynthesis and provide a more efficient technical pathway for its large-scale production.

3.1.4. Cytochrome P450 enzyme

CYP450 is a class of heme-dependent monooxygenases that are widely distributed across various organisms and are capable of catalyzing diverse oxidation reactions. Research has demonstrated that indole, serving as the substrate for these reactions, undergoes P450-mediated hydroxylation to form 2-hydroxyindole and 3-hydroxyindole. However, significant variations exist in the catalytic efficiency of P450 enzymes derived from different origins when hydroxylating indole. For instance, the P450 BM-3 enzyme isolated from Bacillus megaterium necessitates mutations to effectively catalyze the hydroxylation of indole; furthermore, catalytic efficiency improves with an increasing number of mutation sites. Notably, the Phe87Val mutant exhibits a catalytic efficiency of 119 L/(s·mol) for indole hydroxylation, while the double mutant Phe87Leu188 enhances this value to 543 L/(s·mol). The triple mutant Phe87Leu188Ala74 achieves an impressive rate of 1 365 L/(s·mol) (with a Kcat value of 2.73/s and a Km value of 2.0 mmol/L) (Li, Schwaneberg, Fischer, & Schmid, 2000). In contrast, mammalian P450 enzymes (such as CYP2A6) demonstrate significantly lower catalytic efficiencies compared to those observed with the triple mutant form of P450 BM-3 (Gillam, Notley, Cai, De Voss, & Guengerich, 2000). This is because the active site of P450 BM-3 is a funnel-shaped channel with a larger diameter (approximately 10 Å, 1 Å = 0.1 nm) and specific amino acid residues that bind to the substrate indole. Meanwhile, the active site of CYP71B102 exhibits a specific spatial positional relationship when binding to indole (Sagwan-Barkdoll et al., 2025, Li et al., 2008). In contrast, the Phe209 residue of CYP2A6 blocks its active site, resulting in difficulty in indole binding and a reduction in catalytic efficiency (Li, Schwaneberg, Fischer, & Schmid, 2000).

Additionally, although plant-derived CYP71B102 can enhance indirubin production through the exogenous supplementation of isatin and 2-oxoindole substrates, it is important to note that excessive levels of precursor substances may inhibit enzymatic activity and potentially reduce indirubin synthesis efficiency. Therefore, future strategies should encompass enzyme engineering modifications alongside optimization techniques for expression systems and dynamic regulation of precursor concentrations to facilitate efficient synthesis pathways for indirubin.

3.2. Dioxygenase system

Dioxygenases catalyze the direct incorporation of both oxygen atoms from molecular oxygen (O₂) into substrates, resulting in dihydroxylation or cleavage reactions. This reaction process does not necessitate additional reducing equivalents, such as NADH or NADPH; although, certain enzymes require metal ions, such as Fe²⁺, Fe³⁺, or Mn²⁺—as cofactors. Currently, naphthalene dioxygenase (NDO) is primarily responsible for driving the biosynthesis of indirubin (Zhang et al., 2014).

3.2.1. Naphthalene dioxygenase

NDO is a multi-component dioxygenase that plays a crucial role in aromatic hydrocarbon degradation (e.g., naphthalene) in bacteria. In the biosynthesis of indirubin, NDO catalyzes the conversion of indole into several intermediate products, such as 3-hydroxyindole, isatin, and 2-oxoindole. However, the expression of NDO genes is regulated by induction conditions, suitable induction conditions can enhance NDO expression and subsequently facilitate indirubin synthesis. For example, in a recombinant Escherichia coli expression system harboring the NDO gene, when cell growth reaches an optical density at 600 nm of approximately 1.20, the addition of 0.5 mmol/L isopropyl-β-D-thiogalactoside (IPTG) at 30 °C significantly boosts indirubin production (Zhang et al., 2014).

3.2.2. Other dioxygenases

Research has demonstrated that tetrahydronaphthalene dioxygenase and toluene dioxygenase are capable of catalyzing the conversion of indole into indigo; however, it remains unconfirmed whether these two enzymes can directly participate in the biosynthesis of indirubin (Royo et al., 2005, da Silva and Alvarez, 2010). Considering that indigo and indirubin are isomers with similar chemical properties, it is theoretically plausible that these enzymes could facilitate the transformation of indole into indirubin. Nonetheless, their specific mechanisms of action and catalytic efficiencies require further validation through experimental studies.

3.3. Other enzymes

The flavin-dependent terpenoid cyclase (XiaI) catalyzes the production of indigo and indirubin by utilizing tryptophan precursors through intermediate compounds such as indole. The enzymatic activity of XiaI exhibits significant variations at different temperatures: 25 °C is optimal for the synthesis of indirubin and indigo, whereas 37 °C is less conducive to product formation. Furthermore, the expression of XiaI is influenced by induction conditions; a low concentration of 0.02 mmol/L IPTG promotes both protein expression and product yield, with indirubin production reaching its maximum efficiency under these conditions. Conversely, elevated concentrations of IPTG inhibit bacterial growth and adversely affect product formation. Research has demonstrated that introducing auxiliary enzymes such as Fre, TnaA, TnaB, and KatE while optimizing fermentation parameters, including temperature, IPTG concentration, and tryptophan precursor levels, can significantly enhance the efficiency of this enzyme in indirubin synthesis (Yin et al., 2021).

Additionally, another flavin-dependent terpenoid cyclase known as XiaF plays a role in the biosynthesis of indirubin in conjunction with the flavoenzyme reductase XiaP. Initially, XiaF binds to reduced flavin (FADH₂), after which it specifically recognizes indole and introduces a hydroxyl group at the C-3 position to produce 3-hydroxyindole. In an oxygen-rich environment, some 3-hydroxyindole is converted into isatin, which subsequently undergoes a condensation reaction with additional 3-hydroxyindole to form indirubin (Kugel et al., 2017).

It is important to highlight that the current modification of key enzymes for indirubin biosynthesis still primarily relies on traditional protein engineering approaches. These include screening-based directed evolution (e.g., FMO double mutant K223R/D317S) and single-point mutations (e.g., TOM mutant V106A/A113V). While these methods have achieved improved catalytic efficiency, the application of structure biology-guided rational design strategies remains absent. In modern enzyme engineering, AI-assisted protein structure prediction tools such as AlphaFold2 can accurately construct three-dimensional models of key enzymes. Combined with molecular docking simulations to analyze enzyme-substrate and enzyme-cofactor (NADPH) interactions, these tools are expected to address the issues of “high trial-and-error rates and vague targets” associated with traditional modification approaches. For instance, predicting the active pocket conformation of bFMO through AlphaFold2 can identify critical amino acid residues at the entrance of the substrate tunnel, enabling site-directed mutagenesis to expand tunnel diameter and enhance substrate binding capacity. Similarly, using docking simulations to analyze the binding mode between indole and the active center of P450 BM-3 can guide the modification of hydrophobic cavity residues, improving regioselectivity for 2-position hydroxylation and reducing byproduct indigo formation.

Furthermore, structure-based enzyme design can address core bottlenecks in current modifications: first, for the low assembly efficiency of the diiron center in dioxygenases (e.g., NDO), structural simulations can optimize the amino acid composition of Fe²⁺ binding sites to enhance cofactor binding stability. Second, for the narrow substrate spectrum of enzymes such as PAMO, analyzing the hydrogen bond network in the enzyme's active center allows modification of key residues to broaden substrate compatibility. Future research can leverage structural biology and AI simulation technologies to establish an efficient modification system encompassing “structure prediction-target design-functional verification”. This will provide new pathways for the precise optimization of key enzymes in indirubin synthesis, promoting a transition in enzyme engineering technology from “empirical-driven” to “rationally designed” and further breaking through current bottlenecks in yield and selectivity.

A systematic comparison in Table 1 shows that monooxygenases outperform dioxygenases in terms of catalytic efficiency, product selectivity, and the reported indirubin yield. Therefore, monooxygenases represent a technical route with greater industrial prospects and economic feasibility at the current stage. However, due to the limited existing research on dioxygenase systems, their catalytic mechanisms have not yet been clarified, and their potential advantages remain to be explored. Based on this, future research can be advanced in parallel along the following two paths:

Table 1.

Multi-dimensional comparison of monooxygenase and dioxygenase systems in indirubin biosynthesis.

Key performance indicators	Monooxygenase system	Dioxygenase system	References
Catalytic mechanism	Incorporation of one oxygen atom requires NAD(P)H as a reducing cofactor	Incorporation of two oxygen atoms does not require an additional reducing cofactor.	−
Catalytic efficiency	It becomes even higher after protein engineering optimization [e.g., the Kcat/Km value of the bacterial flavin-containing monooxygenase (bFMO) mutant can reach 1.45 L/(s·mmol), which enables high indirubin production up to 860.7 mg/L]	Relatively low. It has been improved after optimization of induction conditions (e.g., the yield of recombinant E. coli containing NDO gene was increased from 9.37 ± 1.01 mg/L to 57.98 ± 2.62 mg/L after optimization).	Sun et al., 2022, Zhang et al., 2014
Product selectivity	Natural product regulation: Can be directionally regulated via mutation (e.g., TOM A113V mutant produces indigo at 89%, and TOM A113I mutant produces indirubin at 64%); Regulation method: Active site mutation (e.g., K223R mutation enhancing π-π stacking in bFMO; Indirubin proportion: Up to 90% (achieved with bFMO combined with cysteine)	Natural product regulation: Predominantly produces indigo; Regulation method: Adding precursors (2-oxindole/isatin); Indirubin proportion: Up to 48%.	Rui et al., 2005, Sun et al., 2022, Royo et al., 2005, da Silva and Alvarez, 2010, Zhang et al., 2014
Metabolic burden	Advantages: High enzyme activity (PAMO kcat = 6.8/s); Disadvantages: Plasmid is prone to inactivation; a biphasic system is required to alleviate substrate toxicity	Advantages: High stability of chromosomal integration; Disadvantages: The assembly of the diiron center consumes Fe2+, requiring supplementation in the culture medium.	Núñez-Navarro et al., 2022, Sagwan-Barkdoll et al., 2025, Royo et al., 2005
Industrialization potential	Broad application prospect. The implementation strategy for the current maximum yield record (860.7 mg/L)	Potential to be developed. Currently, both the yield and productivity are significantly low, and a breakthrough in process technology is required to achieve cost advantages.	Sun et al., 2022

1) For the monooxygenase system: Continue to improve catalytic activity and product selectivity through directed protein evolution. Meanwhile, use synthetic biology to construct dedicated chassis cells or develop in vitro cofactor recycling systems to address its cost bottlenecks.

2) For the dioxygenase system: Conduct in-depth analysis of its substrate recognition and oxygen activation mechanisms to provide theoretical guidance for the rational design of highly selective mutants. Additionally, develop two-phase fermentation or in-situ extraction fermentation processes that are highly compatible with it to achieve real-time product separation, alleviate by-product inhibition, and simplify downstream purification processes.

4. Indirubin biosynthetic engineered bacteria

4.1. Selection of engineering bacteria

The selection of engineering bacteria requires comprehensive consideration of multiple factors, including product characteristics, production efficiency, and economic viability. In the biological synthesis pathways, three microorganisms such as Escherichia coli, Saccharomyces cerevisiae, and Bacillus subtilis are commonly employed as engineering bacteria (Tous Mohedano et al., 2023, Zhang et al., 2023, Wan et al., 2023). E. coli has become the preferred choice for rapid validation of the indirubin synthesis pathway at the laboratory stage due to its well-characterized genetic background, ease of genetic manipulation, and short doubling time. However, as a precursor for indirubin synthesis, indole is toxic to E. coli, which inhibits cell growth and affects indirubin production, thus limiting the application of E. coli in indirubin production (Pontrelli et al., 2018). S. cerevisiae presents several advantages, including rapid growth rates, a clear genetic framework, established gene editing technologies, and a comprehensive post-translational modification pathway that facilitates the expression of key enzymes requiring modifications during synthesis. Nevertheless, its low recombinant protein yield coupled with high medium costs constrains its feasibility for large-scale industrial production. Furthermore, certain strains generate ethanol as a by-product during fermentation processes, which can inhibit their own growth and product synthesis (Yang et al., 2023), thereby impacting the cost-effectiveness of S. cerevisiae as an engineering bacterium. In contrast, B. subtilis exhibits significant potential for industrial production owing to its natural secretion capacity and FDA-certified safety. This characteristic streamlines the purification process and effectively mitigates the risk of endotoxin contamination, rendering it particularly suitable for drug-grade manufacturing. Furthermore, its aerobic characteristic allows for enhanced adaptability to oxidation reaction pathways. However, there is currently a paucity of research and applications involving B. subtilis in indirubin biosynthesis, several metabolic pathways and regulatory mechanisms remain poorly understood, which constrains its further engineering and application (Kang et al., 2014, Liu et al., 2019). In addition, Pseudomonas putida is gradually emerging as a crucial industrial microbial chassis cell, attributed to its low nutrient requirements, rapid proliferation, and strong tolerance to oxidation and reactions involving toxic substances (Weimer, Kohlstedt, Volke, Nikel, & Wittmann, 2020). Early studies revealed that even in the context where only m-toluic acid dioxygenase or p-cumate dioxygenase was expressed, with the absence of diol dehydrogenase, recombinant Pseudomonas putida could still convert indolecarboxylic acid into colored products (Eaton & Chapman, 1995). This result not only clarifies the high redundancy and plasticity of its metabolic network but also provides a highly promising alternative host for the green biomanufacturing of bisindole alkaloids such as indirubin.

As indicated in Table 2, during the laboratory validation phase, E. coli can be prioritized for the rapid construction of synthetic pathways. In the pilot-scale scale-up phase, if cost advantage is the primary goal, emphasis should be placed on optimizing the indole tolerance of E. coli, if safety is the focus, exploration can be directed toward the secretion pathway of B. subtilis. For pharmaceutical-grade industrial production, B. subtilis is more competitive due to its safety and secretion advantages, while Pseudomonas putida—after enhancing its stability via genetic engineering modification—can serve as a supplementary option for scenarios involving specific substrates (e.g., indolecarboxylic acid). In the future, with the continuous development and advancement of gene editing technologies, the selection of engineered bacteria will increasingly rely on the dynamic balance between technological breakthroughs and cost estimation.

Table 2.

Comparative analysis of different chassis microorganisms in biosynthesis of indirubin.

Characteristics	E. coli	S. cerevisiae	B. subtilis	P. putida
Genetic operability	High (clear background, mature tools)	Medium (eukaryotic system, more complex but mature operation)	Medium (tools gradually improving)	Low (relatively limited genetic tools)
Growth rate	Fast (doubling time-20 min)	Medium (doubling time-90 min; fast in eukaryotes)	Fast (doubling time-30 min; fast in prokaryotes)	Medium (doubling time-60 min)
Indole tolerance	Low (toxicity significantly inhibits growth)	Medium	Medium	High (strong tolerance to toxic substances)
Protein secretion ability	Low	Medium	High (strong natural secretion ability)	Medium
Safety evaluation	Medium (contains endotoxins)	High (GRAS-recognized)	High (GRAS-recognized)	Medium (most not GRAS-certified, only environmental strains approved)
Industrial maturity	High (widely used in biomanufacturing)	High (commonly used in fermentation industry)	Medium	Low
Production cost	Low	Medium (high medium cost)	Low	Low (low nutrient requirements)

4.2. Genetic engineering modification methods

Currently, the synthesis of indirubin is primarily achieved through the construction of engineered E. coli strains. In 1992, Hart's team successfully cloned the Rhodobacter sphaeroides gene into E. coli, utilizing the endogenous tryptophanase present in E. coli to convert tryptophan into indole. This indole was subsequently oxidized to indirubin by an enzyme encoded by the Rhodobacter sphaeroides gene, marking the first instance of heterologous synthesis of indirubin (Hart, Koch, & Woods, 1992). In follow-up studies, cases of yield improvement through gene cloning and plasmid construction have continued to emerge. For example, the novel flavin reductase gene MoxB was cloned into the plasmid pTre-TA, which overexpresses tnaA and encodes flavin monooxygenase MoxA; this led to the creation of a new plasmid designated pTre-TAB. When expressed in E. coli, this modification resulted in a remarkable 15.12-fold increase in indirubin yield compared to that obtained with the original plasmid (Du et al., 2024). Furthermore, key enzyme genes such as NDO gene (Yin et al., 2021), FMO gene (Ameria et al., 2015, Han et al., 2012), CYP71B102 (Sagwan-Barkdoll, Kim, Berim, Gang, & Anterola, 2025), and toluene 4-monooxygenase gene (tmoABCDEF) (Wongsaroj, Sallabhan, Dubbs, Mongkolsuk, & Loprasert, 2015) were sourced from various origins and introduced into appropriate vectors for recombinant plasmid construction and expression in E. coli, each significantly enhancing indirubin production.

In terms of co-expression and cooperative regulation of multiple genes, there are mainly two strategies, as shown in Fig. 2. The first strategy involves the construction of polycistronic expression vectors. For instance, Sagwan-Barkdol et al. successfully cloned the P450 gene (CYP71B102) from Isatis indigotica and concurrently obtained the P450 reductase gene (AtR2) from Arabidopsis thaliana. They subsequently cloned these two genes in tandem into a bicistronic plasmid. Additionally, the isatin hydrolase gene (IsH) from Pseudomonas putida was cloned into the recombinant plasmid. After constructing the new plasmid, it was transformed into E. coli cells respectively (Sagwan-Barkdoll, Kim, Berim, Gang, & Anterola, 2025). The second strategy involves co-transforming host strains with multiple single-gene expression vectors. For example, Yin et al. introduced Fre, TnaA, TnaB, and KatE while optimizing fermentation parameters, this approach resulted in a 60-fold increase in the total production of indigo and indirubin in the recombinant E. coli strain compared to its original counterpart. Following an expansion of the fermentation system and the addition of specific substances, indirubin production reached 250.7 mg/L (Yin et al., 2021). Furthermore, CRISPR-Cas9 technology can be employed for precise editing of the host strain's genome by knocking out genes associated with competing pathways in engineered bacteria, such as trpR, pykA, pykF, and ppc in E. coli, to minimize metabolic diversion and enhance metabolic pathway optimization. Sun et al. constructed a classic strain designated W1 that expressed the bFMO gene under the control of the promoter P tac. They then transformed this relevant plasmid into W1 to generate engineered strains W2 and W3. Notably, strain W3 achieved an impressive indirubin production level of 860.7 mg/L—the highest reported yield to date (Sun et al., 2022). Meanwhile, Du et al. constructed a series of multi-gene co-expression plasmids and employed CRISPR-Cas9 technology to knock out the trpR gene, thereby alleviating its negative regulation on L-tryptophan synthesis. They also modified the AroG and TrpE genes to eliminate feedback inhibition and enhance metabolic flux. Additionally, they screened for and overexpressed the aroL gene to optimize metabolic flow, while removing the pykF and pykA genes in conjunction with overexpressing the tktA gene to increase the supply of precursors PEP (phosphoenolpyruvate) and E4P (erythrose 4-phosphate). Ultimately, these efforts culminated in the direct production of indirubin from glucose, paving a new pathway for indirubin biosynthesis (Du, Yang, Luo, & Lee, 2018).

Fig. 2.

Genetic engineering strategy for indirubin production by engineered E. coli strains (Created with BioRender.com).

Genetic engineering technology has successfully reconstructed the metabolic network of engineered E. coli through the cloning and expression of key enzyme genes, co-expression of multiple genes, and coordinated regulation, in conjunction with CRISPR-Cas9 technology. This advancement has surpassed the natural yield limitations and enabled stable, large-scale synthesis of indirubin. This innovative approach effectively addresses the bottlenecks associated with traditional plant extraction methods, which are characterized by low efficiency and prolonged processing times. Furthermore, it provides a valuable technical framework for the biosynthesis of other natural products while simultaneously fostering advancements in synthetic biology.

4.3. Metabolic engineering optimization

Although genetic engineering has the potential to enhance indirubin production, current yields remain insufficient to meet market demands. Consequently, precise modulation of host strain metabolism has become a critical step in the synthetic pathway of indirubin.

Firstly, overexpression of key enzyme genes can enhance the supply of precursor metabolites. Du highlighted that the insufficient supply of reduced flavin is a key factor restricting synthesis efficiency. In response, the research team introduced the novel flavin reductase gene MoxB and simultaneously overexpressed TnaA and TnaB to facilitate the efficient conversion of tryptophan to indole, thereby providing ample precursor metabolites for indirubin synthesis and significantly increasing its yield (Du et al., 2024). Furthermore, overexpression of the aroL gene optimizes the aromatic amino acid pathway, further enhancing the availability of precursor metabolites and promoting increased indirubin production. Secondly, knocking out genes associated with by-product formation can lead to enhanced indirubin yields. For instance, in E. coli, deletion of the trpR gene removes its negative regulation on L-tryptophan biosynthesis pathways. This redirection allows more metabolic flux towards tryptophan synthesis while minimizing unnecessary metabolic branches. Additionally, knockout of the pykA and pykF genes reduces conversion rates from phosphoenolpyruvate to pyruvate, promoting accumulation of phosphoenolpyruvate and provides additional raw materials for indirubin synthesis while inhibiting competitive side reactions (Du, Yang, Luo, & Lee, 2018). Thirdly, enhancing the supply of cofactors can elevate the concentration of reaction substrates. The addition of substances such as cysteine can influence the catalytic activity of flavin-containing monooxygenase, thereby altering the generation ratio of 2-hydroxyindole to 3-hydroxyindole and facilitating indirubin synthesis. Furthermore, optimizing fermentation conditions to augment the availability of intracellular cofactors like NADPH creates more favorable environments for related enzymatic reactions, thus promoting the biosynthesis of indirubin (Yin et al., 2021, Han et al., 2012).

Three major challenges persist in current metabolic engineering research. First, the co-expression of multiple genes frequently imposes an increased metabolic burden on bacterial strains. A typical example is that the simultaneous overexpression of FMO, TnaA, and MoxB leads to a reduced growth rate of E. coli. Second, the precision of metabolic flux regulation remains inadequate. Partial knockout of genes involved in competitive pathways (e.g., ppc) often triggers the accumulation of intermediate metabolites such as succinic acid, which subsequently exerts an inhibitory effect on the activity of key enzymes. Third, an imbalance between cofactor supply and demand represents another critical issue: the consumption rate of NADPH usually exceeds its regeneration rate, thereby resulting in the obstruction of hydroxylation reactions—a pivotal enzyme-catalyzed step in target metabolite synthesis. To address these challenges, further efforts should focus on integrating dynamic regulation technologies (e.g., CRISPRi) with cofactor recycling systems.

As illustrated in Table 3, the engineered E. coli bacteria produce indigo as a by-product during the synthesis of indirubin, which significantly compromises both the purity and conversion rate of the final product. Currently, most engineered bacteria continue to depend on exogenous tryptophan, 2-oxoindole, or 2-hydroxyindole as precursor substances. This reliance leads to a considerable increase in fermentation costs and associated toxicity. Furthermore, gene modification techniques, such as multi-gene knockout, overexpression, or the introduction of heterologous enzyme genes often result in decreased strain stability. At present, the yield of indirubin remains insufficient to meet industrial production demands.

Table 3.

Engineered strain for indirubin biosynthesis and their applications in production.

Engineered strain	Source of key gene	Substrate	Indirubin yield (mg/L)	By-product (indigo) yield (mg/L)	Advantages	References
Escherichia coli DH5α	fmo gene from Methylophaga aminisulfidivorans MPT	2 000 mg/L tryptophan 360 mg/L cysteine	223.6	6.8	User-friendly operation Elevated production efficiency Minimal by-product generation	Han, Gim, Kim, Seo, & Kim, 2012
Escherichia coli BL21 (DE3) (named ND_IND)	Fragment of nag gene (encodes naphthalene dioxygenase, NDO) from Comamonas sp. MQ	3 280 mg/L tryptophan	9.37 ± 1.01	−	Precursor addition for higher indirubin proportion	Zhang et al., 2014
		+ 500 mg/L 2-oxindole (1 h post-induction)	57.98 ± 2.62
		+ 100 mg/L isatin (5 h post-induction)	The output has increased by twice the base amount, accounting for 15 %.
Escherichia coli Top10, W3110	cFMO gene from C. glutamicum ATCC13032	2 500 mg/L L-tryptophan	103	685	Enhanced raw material utilization (higher tryptophan conversion)	Ameria et al., 2015
Escherichia coli DH5α, W3110	Introduction of Methylophaga FMO and E. coli TnaA, knockout of trpR, pykF, pykA, overexpression of tktA, aroL	Glucose	56	640	Cost-effective	Du, Yang, Luo, & Lee, 2018
Escherichia coli BL21 (DE3)	Optimization of novel terpenoid cyclase XiaI Introduction of fre, tnaA, tnaB, and katE genes	1 021.15 mg/L tryptophan 1 331.5 mg/L 2-hydroxyindole	250.7	26	High indirubin yield	Yin et al., 2021
Escherichia coli W3	Structure-guided bFMO mutations (K223R/D317S) for enhancing flavin monooxygenase activity	Tryptophan	860.7	−	Highest indirubin production Significantly enhanced enzymatic activity	Sun et al., 2022

To address these challenges, future research should prioritize four key directions. Firstly, it should focus on optimizing glucose-based direct biosynthesis pathways, with the goal of reducing tryptophan dependence and overall costs. Secondly, it is necessary to enhance the catalytic activity and substrate specificity of key enzymes via directed evolution which can help minimize the production of by-products. Thirdly, efforts should be made of refine metabolic pathways, thereby improving the synthesis efficiency of indirubin. Finally, in the context of large-scale production, it is required to conduct large-scale performance testing of engineered bacteria, carry out technical and economic cost breakdown, and perform pilot case benchmarking to identify current industrialization gaps in indirubin biosynthesis and provide quantitative targets for subsequent process optimization.

5. Biosynthesis process and downstream processing of indirubin

With the advancements of green chemistry and biotechnology, optimizing the biosynthesis process of indirubin has emerged as a crucial focus for enhancing production efficiency and reducing costs. Through microbial fermentation, enzyme engineering, and metabolic engineering, researchers are continuously investigating methods to produce indirubin in an efficient and environmentally friendly manner while simultaneously improving its purity and stability.

5.1. Optimization of biosynthesis conditions

5.1.1. Regulation of precursor supply

In both the plant extraction and microbial synthesis of indirubin, the regulation of precursors is a core strategy for improving yield. Tryptophan (Trp), as the most critical precursor, directly determines the output of the final product. Precursor regulation can be primarily achieved through two approaches.

The first is the direct exogenous addition of tryptophan to supplement raw materials for metabolic flux. Studies have shown that adding an appropriate concentration of tryptophan to Isatis indigotica Fort. root cultures can significantly promote the synthesis of indigo and indirubin. More importantly, tryptophan also exhibits a gene-regulating function: it can upregulate the expression of genes encoding key enzymes in the synthetic pathway, such as It-TSA and CYP79B2, thereby systematically enhancing the biosynthetic pathway. It should be noted that the added concentration must be controlled within an optimal range; excessive addition may conversely lead to reduced efficacy due to feedback inhibition (Cessur, Tuğlu, & Albayrak, 2025).

The second strategy involves indirect regulation using signaling molecules. For instance, hydrogen sulfide (H₂S) donors (e.g., NaHS) can significantly upregulate the expression of multiple upstream key enzyme genes (such as BcSK and BcDXR) in I. indigotica under specific treatment conditions, optimizing precursor supply at a more fundamental level (Jia, Liu, & Liu, 2020).

5.1.2. Regulation of biological synthesis parameters

Key parameters in the biosynthetic process, such as pH, temperature, dissolved oxygen levels, and induction timing, exert a decisive influence on the yield and selectivity of indirubin. The dynamic regulation of these parameters constitutes the core of fermentation process optimization.

The synergistic regulation of pH and temperature is recognized as a critical factor influencing microbial growth and enzyme activity. Generally, an increase in temperature accelerates enzymatic reactions, enhances the metabolic activity of microorganisms, and consequently increases the growth rate. However, when temperatures exceed a certain threshold, enzyme activity may be compromised, resulting in a decrease in product yield (Du et al., 2024). Therefore, controlling both temperature and pH during fermentation is crucial for maintaining enzyme activity and product stability while enhancing the selectivity of indirubin.

Some studies have implemented a temperature gradient strategy: starting at 30 °C during the initial phase of fermentation to promote microbial proliferation before raising the temperature to 37 °C in later stages to activate monooxygenase activity. The final results indicated that this increase in temperature could expedite the indole hydroxylation reaction rate and shorten the overall fermentation cycle (Pan, He, Xuan, & Kong, 2003). In addition to employing temperature gradients, pH gradients can also facilitate reactions effectively. In experiments involving TfdB-JLU engineered bacteria during fermentation, an initial pH of 7.0 was established to support microbial growth; this was subsequently adjusted to 6.5 during production phases to minimize by-product formation (the indigo/indirubin ratio decreased from 10:1 to 3:1) (Jilin University, 2024). Furthermore, Du et al. utilized pH gradient regulation (from 7.0 down to 6.5) to stabilize enzyme activity while preventing the accumulation of acidic by-products (Du et al., 2024).

Dissolved oxygen levels and feeding strategies exert a pronounced influence on fermentation efficiency and product composition. A low concentration of dissolved oxygen can inhibit oxidative side reactions, such as the formation of indigo, whereas a high concentration is advantageous for monooxygenase activity. Maintaining an appropriate dissolved oxygen content is essential for ensuring the normal growth and metabolism of microorganisms, which in turn affects the synthesis of indirubin. During fermentation, the dissolved oxygen levels can be regulated by adjusting stirring speed and aeration rates to optimize enzyme activity and product stability. In research focused on synthesizing indirubin using glucose as a carbon source, researchers successfully maintained the dissolved oxygen levels between 30% and 50% during a 5 L tank amplification experiment. This balance addressed both the respiratory needs of microorganisms and the catalytic requirements of enzymes effectively. Based on signals from dissolved oxygen measurements, glucose was fed at a controlled rate to maintain substrate concentrations within 2−5 g/L, thereby preventing acetic acid accumulation.

Following optimization efforts, the yield of indirubin reached 56 mg/L, representing a remarkable enhancement by a factor of 24 compared to that obtained with initial strains, with a carbon conversion rate recorded at 0.82%. Additionally, the ratio of indigo to indirubin decreased from 10:1 to 3:1 (Du et al., 2024). This study not only substituted traditional aniline-based raw materials with glucose but also achieved significant reductions in production costs, ranging from 60% to 70%. Furthermore, it validated process stability through large-scale amplification in a 5 L tank setting, thereby providing a technical model for future industrial applications [such as demonstrated by Pham et al., who reported producing 965 mg/L indigo in a separate study utilizing a 3 L tank (Pham et al., 2024)].

The precise selection of induction timing is crucial for the synthetic efficiency and product specificity of indirubin. Too early or too late induction may reduce product yield, so precise regulation of induction timing is key to indirubin biosynthesis efficiency and needs optimization based on cell growth and metabolic characteristics in different systems. In microbial fermentation, induction is usually scheduled in the mid-exponential growth phase to balance cell biomass and enzyme activity expression.

In the optimization of fermentation conditions by Tianjin University after obtaining the flavin monooxygenase mutant L701 and S321T via directed evolution, induction at a 600 nm optical density value of 1 significantly improved indirubin synthesis efficiency. Indirubin accumulated rapidly within 16 h post-induction, with the yield peaking at 40.1 mg/L at 60 h (Feng, 2020). Notably, indigo, a competitive byproduct, typically starts forming 10 h post-induction, earlier than indirubin’s synthesis peak. This time lag suggests that delaying induction or adopting a staged induction strategy could inhibit the indigo pathway, extend indirubin’s synthesis window, and thereby enhance product specificity and final yield. In plant systems (e.g., I. indigotica hairy root culture systems), the selection of induction timing focuses on stimulating secondary metabolic activity. Inducers are usually added 14−21 d after cultivation, when biomass enters the stationary phase—at this stage, indirubin synthesis capacity is strongest, with accumulation peaking 7−10 d later. Studies also show that combining a 16 h light/8 h dark cycle regulation can simultaneously promote hairy root growth and indirubin synthesis, demonstrating the synergistic regulatory role of light signals in plant hosts (Speranza, Miceli, Taviano, De Feo, & Menichini, 2020).

5.2. Product enrichment and purification

5.2.1. Separation process

Indirubin, recognized as a natural drug active ingredient, has traditionally been purified using column chromatography and solvent crystallization as the primary techniques. Alumina column chromatography, noted for its high adsorption capacity and cost-effectiveness, has emerged as the predominant technology for laboratory-scale purification of indirubin. Research indicates that employing dry column chromatography with Al₂O₃ and utilizing a chloroform-ethyl acetate (10:3) gradient can achieve indirubin purity levels ranging from 97.5% to 98%. In wet column chromatography, the dynamic adsorption equilibrium established between basic alumina and crude indirubin extracts further reduces separation time to within 4 hours. However, the high regeneration frequency of alumina packing (requiring activation treatment after every three uses) limits its potential for industrial-scale application (Xie, Zhang, Qiu, & Jia, 2006). To address industrial-scale demands, type 12⁻¹ macroporous resin demonstrates significant promise due to its hydrophobic framework and π-π interactions with indirubin, exhibiting a selective adsorption capacity of 0.256 mg/g. By optimizing the eluent (for instance, using a methanol-acetone mixture), it is possible to enhance the purification factor of indirubin up to 9.2 times. Importantly, the pore size distribution of this resin (ranging from 20 to 60 nm) plays a crucial role in influencing mass transfer efficiency for indirubin molecules (with a molecular weight of approximately 262 Da), necessitating careful regulation of pore sizes to improve adsorption kinetics. Additionally, liquid-liquid extraction methods, including multiple-step extractions using ether or chloroform, are frequently employed for impurity removal. Studies have demonstrated that optimal recovery rates are attained following three ether extractions. Thin layer chromatography can be utilized concurrently to verify separation efficacy in real-time (Han, Sun, & Wang, 2002). It is important to highlight that traditional processes create industrial-scale solutions by leveraging the advantages of multiple sequential steps. For instance, an optimized procedure initially eliminates water-soluble impurities through decoction with water, followed by extraction using a 75% ethanol-HCl solution. This is then succeeded by alkalinization with ammonia and ultimately purification via ethyl acetate extraction and column chromatography. The resulting product achieves purity levels that comply with HPLC detection standards (Li, 2014).

In addition to the aforementioned methods, membrane separation technology has been employed in the production of indirubin. During the extraction process of Isatidis Radix, an ultrafiltration membrane with a molecular weight cut-off of 30 000 effectively retains indirubin in the retentate through the size exclusion effect. When combined with multiple ethanol extractions, this approach enhances the transfer rate to 32.86%, which is 13.91% higher than that achieved by traditional pharmacopoeial processes (Xu, Liu, & Huang, 2004). This technology mitigates the degradation of indirubin associated with high-temperature treatments (with a critical temperature for thermal stability around 80 °C). However, challenges related to membrane fouling must still be addressed through periodic backwashing every two hours. The integrated process utilizing ultrafiltration and vacuum concentration significantly reduces energy consumption by approximately 40% and increases the indirubin content from 0.12% to 0.35% through optimization of drying temperatures (45 °C compared to 60 ℃) (Gao & Yang, 2006). Nevertheless, the cost of membrane materials (for instance, polyethersulfone membranes priced at approximately $200 per square meter) continues to pose a significant barrier to large-scale applications.

Currently, the technologies for the separation and purification of indirubin are evolving in diverse ways. Traditional methods possess distinct advantages and limitations. While alumina column chromatography is cost-effective and offers high purification efficiency, its frequent need for regeneration poses constraints on industrial applications. Macroporous resins achieve highly selective adsorption through π-π interactions. However, optimizing pore size is essential to enhance mass transfer efficiency. Membrane separation technology significantly improves product yield due to its low-temperature operation and energy-saving characteristics, yet challenges such as membrane fouling and material costs persist for large-scale implementation.

These bottlenecks have driven the iteration of separation technologies toward higher efficiency and greener processes. Foam separation technology, a separation method based on differences in surface activity, has been successfully applied to the separation and purification of indirubin in recent years. The innovation of this technology lies in combining the traditional Indigo Naturalis water-floating process with modern foam separation principles: it leverages the hydrophobic properties of active components such as indirubin and the excellent foaming properties of proteins and peptides to achieve efficient separation. Researchers have designed a Indigo Naturalis water-floating separation device with independent intellectual property rights; by optimizing key parameters including liquid level difference (the position between the foam collector and the crude indigo slurry surface) and air pressure, efficient separation of indirubin has been realized. The most notable advantage of foam separation technology is that it increases the yield of indirubin from less than 0.1% (of traditional processes) to 3%, solving the problems of low yield, low efficiency, and unstable quality associated with the traditional water-floating process. Additionally, this technology can effectively reduce the skin irritation of indirubin products, improve safety for external use, enhance its ability to resist oxidative stress, and alleviate inflammatory damage caused by lipopolysaccharides. Foam separation technology is not only suitable for laboratory-scale applications but has also been successfully used in pilot-scale production, providing reliable technical support for the market circulation of high-quality indirubin (Yang et al., 2023).

Macroporous adsorption resin technology, a separation method based on adsorption, has shown promising application prospects in recent years for the separation and purification of water-insoluble active components from Chinese herbal medicines, such as indirubin. This technology achieves the separation of indirubin from impurities through the large specific surface area and selective adsorption of resins. The typical process of separating indirubin using macroporous adsorption resins includes three steps: adsorption, impurity washing, and desorption. Under the separation conditions for indirubin in the compound Polygala japonica Houtt. (Fufang Guazijin) preparation, the adsorption capacity of 12-1 resin for indirubin is 0.256 mg/g, with a purification factor of up to 9.2 (Dong, Xu, & Liu, 2005). The core advantages of macroporous adsorption resin technology include large processing capacity, simple operation, recyclable resins, and relatively low cost, making it particularly suitable for industrial-scale production. However, factors such as resin type selection, regeneration performance, and service life directly affect separation efficiency and economic viability, requiring optimization based on specific application scenarios.

5.2.2. Two-Stage fermentation technology

The two-stage fermentation technology significantly enhances the yield and selectivity of indirubin and indigoid compounds by separating the microbial growth phase, which is primarily focused on biomass accumulation, from the product synthesis phase, which emphasizes the production of target metabolites, as shown in Fig. 3. This approach allows for optimization under varying environmental conditions. Although traditional indigo production processes (such as Indigo Naturalis processing) did not explicitly articulate the modern concept of “two-stage fermentation”, they inherently established a prototype for staged regulation within their production methodology. By controlling immersion and hydrolysis conditions—facilitating the release of indigo precursors—and managing indigo oxidation to promote the formation of both indigo and indirubin in distinct stages, these processes effectively embody the principle of staged optimization in biosynthesis.

Fig. 3.

Illustration of two-stage fermentation using Polygoni Tinctorii Folium for indirubin preparation.

In this method, bluegrass leaves are soaked and fermented in water at a pH range of 6.5–7.5 and temperatures between 25 °C to 35 °C to generate indophenol precursors. Subsequently, lime is added incrementally to adjust the pH above 10.5, thereby promoting oxidation and condensation into indigo and indirubin (Chen et al., 2022). While this physical staging operation relies heavily on experiential knowledge, it offers valuable insights for contemporary technological applications; specifically, staged chemical regulation can strategically enhance product yields. For example, a Box-Behnken experimental design achieved an indirubin yield of 0.758 mg/g through systematic adjustments in hydrochloric acid concentration and temperature (Liu, Su, Yang, & Zou, 2010).

Modern microbial engineering has further refined the two-stage strategy. In the growth stage, engineered strains rapidly accumulate biomass by optimizing carbon sources and metabolic pathways. For instance, E. coli was genetically modified to develop the optimal indirubin-producing strain W3. During the fermentation stage, by optimizing fermentation conditions, strain W3 can efficiently convert substrates into indirubin, ultimately achieving an indirubin yield of 860.7 mg/L and a production rate of 18 mg/(L·h)—an impressive increase of 3.43 times compared to the previously reported highest yield (Sun et al., 2022). The engineered E. coli can reach a biomass density of OD₆₀₀ = 15 at 37 °C under high dissolved oxygen conditions. Upon entering the production phase, the expression of flavin monooxygenase (FMO) is activated by an inducer (such as IPTG), facilitating the conversion of indole to 2-hydroxyindole (the precursor for indirubin), while simultaneously reducing dissolved oxygen levels (< 10%) to inhibit indigo formation (Du, Yang, Luo, & Lee, 2018).

The introduction of dynamic regulation technology has revitalized the two-stage process. For example, temperature-responsive systems, such as the SIMTeGES system, facilitate bacterial growth at 30 °C while activating product synthesis genes at 25 °C, thereby eliminating the need for exogenous inducers. The metabolic burden balancing strategy employs repressors, such as TetR, to inhibit synthesis gene expression during the growth phase and alleviates this expression pressure once sufficient biomass is achieved (Gou et al., 2024). Xiamen University enhanced the selectivity of indirubin from 15% to 60% and increased the total yield to 120 mg/L through a two-stage dissolved oxygen control approach (maintaining levels at 50% during the growth stage and reducing them to 10% during production) (Shao et al, 2024). Furthermore, cell cycle synchronization techniques, such as DNA synthesis inhibitors, can synchronize microbial populations to enter the production phase simultaneously, resulting in a 20% increase in synthesis efficiency (Chen, Reddy, & Suzuki, 2025).

Although the two-stage technology offers significant advantages, such as mitigating metabolic resource competition, enhancing product purity (for example, the foam flotation method increases the recovery rate of indirubin to 75%), and promoting environmentally friendly production by reducing lime usage—it still faces challenges in terms of industrialization. These challenges include dependence on high-sensitivity sensors (such as hyphenated near-infrared spectroscopy) for accurate dynamic control, potential strain stability issues arising from gene element leakage, and mass transfer limitations due to uneven mixing at scale. Future research could concentrate on synthetic biology tools (such as division of labor in co-culture of two strains), intelligent fermentation systems that utilize real-time regulation through machine learning, and the development of specialized chassis organisms (for instance, using yeast instead of E. coli). By integrating processes across different scales and innovating green production models, it is anticipated that two-stage technology will catalyze technological breakthroughs in indigo-based biomanufacturing, thereby paving efficient and sustainable pathways for both the natural medicine and dye industries.

6. Summary and outlook

Despite the promising potential of indirubin biosynthesis, it faces critical technical and industrialization bottlenecks, primarily in four areas: by-product control, metabolic efficiency, industry-academia-research disconnection, and pharmaceutical compliance.

6.1. Indigo separation difficulty

As structural isomers, indigo and indirubin coexist in fermentation broths. Traditional separation techniques have low resolution and high costs, and current industrial purification relies on expensive preparative chromatography, leading to low efficiency.

6.2. Low metabolic pathway efficiency

Existing engineered strains exhibit inherent defects (e.g., high NADPH dependence, metabolic imbalance from exogenous pathways, membrane permeability limitations, and product toxicity to hosts), resulting in a large gap between the maximum current yield (40.1 mg/L) and the industrial economic threshold (>1 g/L).

6.3. Industry-academia-research disconnection

A significant disconnect exists between upstream basic research (e.g., enzyme mechanism analysis, pathway design) and downstream process scaling (e.g., fermentation engineering, separation/ purification), hindering the translation of laboratory results into profitable industrial solutions.

6.4. Pharmaceutical compliance requirements

As a pharmaceutical ingredient, biosynthetic indirubin must meet strict pharmaceutical standards (impurity profile control, residue detection, bioequivalence certification), but relevant research remains insufficient.

To address these challenges, interdisciplinary innovative strategies are essential for systematic optimization of synthesis efficiency, selectivity, and economy:

1) Strengthen metabolic regulation and pathway design: Apply dynamic regulation (e.g., CRISPRi) to design intelligent systems that sense intermediates (e.g., indole) for real-time regulation of competitive pathway gene expression, balancing metabolic flux and reducing host burden. Combine directed evolution of key enzymes (e.g., P450 enzymes, glycosyltransferases) to enhance catalytic efficiency and specificity.

2) Develop novel biosynthetic systems: Cell-free biosynthetic platforms (encapsulating key enzymes/cofactors in liposomes for compartmentalized catalysis) overcome membrane permeability limits (substrate concentration up to 50× intracellular levels), eliminate product inhibition via continuous flow reactors, decouple cell growth from product synthesis, and expand biosynthesis chemical space using non-natural substrates/catalysts.

3) Promote green low-carbon integrated processes: Develop integrated “fermentation-separation-crystallization” processes (e.g., two-phase fermentation for in-situ product extraction, membrane separation coupling) to reduce downstream costs and environmental impact.

4) Coordinate pharmaceutical standards with industrialization: Use green efficient synthesis to ensure high purity, yield, and batch stability of active ingredients, and optimize purification processes and refine quality control systems per international pharmaceutical standards to meet safety, efficacy, and quality uniformity requirements.

Future research must prioritize meeting drug regulatory requirements (systematic evaluation of impurity profiles, safety, bioequivalence) and strengthen industry-academia-research collaboration to integrate basic research, process optimization, and large-scale production, building a complete innovation chain from laboratory to factory.

In conclusion, indirubin biomanufacturing is in a critical transition stage from laboratory to industrialization. In-depth integration of synthetic biology, enzyme engineering, and process engineering is expected to establish an efficient, green, and compliant production process, providing high-purity, low-cost indirubin for clinical practice and promoting the upgrading and development of the biomanufacturing industry.

CRediT authorship contribution statement

Keqian Li: Writing – original draft, Data curation, Supervision, Investigation. Fangyu Xiang: Writing – original draft, Investigation. Xin Yang: Writing – review & editing, Data curation. Mengqi Liu: Writing – review & editing. Rui Long: Writing – review & editing. Dingkun Zhang: Writing – review & editing, Conceptualization. Li Han: Writing – review & editing, Conceptualization. Yanan He: Writing – review & editing, Conceptualization, Project administration.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.