Engineered Bacteria Transform Tyrosine Into Danshensu Using Cascade Biotransformation

A research team has established a multi-enzyme cascade in engineered Escherichia coli capable of synthesizing the cardioprotective compound Danshensu directly from L-tyrosine. Danshensu, also known as DSS, is a water-soluble bioactive molecule recognized for its antioxidant, anti-inflammatory, cardioprotective, and neuroprotective properties, making it widely utilized in therapeutic applications targeting cardiovascular and cerebrovascular diseases. Beyond direct clinical interest, Danshensu serves as a foundational building block for advanced synthetic derivatives, including Danshensu Bingpian Zhi, which is produced with borneol to enhance therapeutic effects, and isopropyl dihydroxyl hydroxy propionate, synthesized using isopropyl alcohol to bolster myocardial protection. The study demonstrates a sustainable, high-efficiency whole-cell biotransformation platform designed to circumvent the strict raw material limitations, toxic organic solvents, complex downstream purifications, and unwanted chemical byproducts that historically impede commercial extraction from the roots of Salvia miltiorrhiza or multi-step chemical synthesis routes starting from benzaldehyde.

To overcome these production bottlenecks through synthetic biology and green chemistry principles, the investigators constructed an artificial biosynthetic cascade inside E. coli BL21(DE3). The metabolic architecture relies on the coordinated heterologous expression of four key enzymes: 4-hydroxyphenylacetate 3-monooxygenase, designated as HpaBC, L-amino acid deaminase, designated as aadL, glucose dehydrogenase, designated as gdh, and D-lactate dehydrogenase, designated as ldhD. The catalytic mechanism accommodates two parallel metabolic trajectories because of the promiscuous substrate selectivity exhibited by HpaBC and ldhD. In the first pathway, ldhD reduces 4-hydroxyphenylpyruvic acid to generate 4-phenyllactic acid, which HpaBC subsequently hydroxylates into Danshensu. In the alternative second pathway, HpaBC first hydroxylates 4-hydroxyphenylpyruvic acid into 3,4-dihydroxyphenylpyruvic acid, which ldhD subsequently reduces to produce Danshensu. Across both functional routes, gdh efficiently converts oxidized NAD+ into reduced NADH, driving crucial cofactor regeneration to sustain the continuous reduction steps without depleting internal coenzyme pools.

The enzymatic construction was refined by incorporating precisely engineered genetic sequences verified by sequencing. The HpaBC component was amplified from E. coli BL21(DE3) and altered to harbor the specific 23F9-M4 mutation profile, which incorporates targeted amino acid substitutions at positions T15P, S210F, A211K, Q212F, and D284E to elevate enzymatic performance based on established screening strategies. In parallel, the aadL sequence was derived to ensure the efficient conversion of the intermediate amine into its corresponding keto acid, while ldhD was deployed specifically to reduce 3,4-dihydroxyphenylpyruvic acid into Danshensu. In initial baseline evaluations, the investigators coupled the medium-copy-number plasmid pETDuet-1 carrying gdh and ldhD with the high-copy-number plasmid pRSFDuet-1 carrying HpaBC and aadL, successfully creating strain 1. During preliminary whole-cell biotransformation tests using 25 mM L-tyrosine and 50 mM glucose at 30 degrees Celsius and neutral pH over a 6-hour incubation, strain 1 achieved an initial Danshensu yield of 5.1 mM, confirmed by mass spectrometry and liquid chromatography matching verified Danshensu reference standards.

Because metabolic bottlenecks often arise from imbalanced enzyme levels, the investigators performed systematic transcriptional fine-tuning by evaluating seven distinct recombinant strains featuring variable plasmid copy numbers. Measuring relative messenger RNA abundance through one-step quantitative reverse transcription PCR with 16S ribosomal RNA as an internal standard revealed substantial shifts in catalytic capacity. For instance, shifting HpaBC from the high-copy-number plasmid in strain 1 to the medium-copy-number plasmid in strain 2 caused a 2.4-fold drop in HpaBC transcription, precipitating a steep decline in Danshensu generation down to 2.26 mM. Similarly, reconfiguring aadL and ldhD onto high-copy plasmids while lowering HpaBC expression in strain 4 cut output to 2.45 mM, or 48.1 percent of strain 1 levels. These comparative experiments confirmed that HpaBC-mediated hydroxylation represents the rate-limiting step within the cascade, necessitating strong high-copy-number expression.

The optimization of plasmid configurations culminated in the design of strain 6, which paired high-copy pRSFDuet-1 expressing HpaBC and aadL with the low-copy plasmid pCDFDuet-1 harboring gdh and ldhD. This specific architectural pairing yielded 6.53 mM Danshensu, representing a 28.1 percent increase over strain 1 and reaching 1.3 times its initial productivity. A related construct, strain 7, distributed the pathway across three separate vectors—pRSFDuet-1 for HpaBC and aadL, pETDuet-1 for ldhD, and pCDFDuet-1 for gdh—producing 6.18 mM Danshensu. The superior output of strain 6 indicated that low-level expression of gdh is entirely sufficient to maintain vital NADH cofactor regeneration without overburdening host translational resources, allowing cellular energy to concentrate around the demanding rate-limiting hydroxylation catalyzed by HpaBC.

Following strain selection, the authors examined single-factor induction parameters in shake-flask cultures to identify optimal protein expression settings for strain 6 under the control of the inducible T7 promoter. Systematic testing of bacterial optical densities at 600 nanometers spanning 0.32 to 1.91 demonstrated that inducing cells at an OD600 of 0.93 yielded optimal synthetic capacity. Titrating isopropyl-beta-D-thiogalactoside concentrations across 0.1 to 1.22 mM identified 0.5 mM as the most effective inducer level, successfully avoiding the metabolic toxicity and impaired heterologous synthesis associated with excessive chemical induction. Examining induction temperatures from 15 to 40 degrees Celsius established 25 degrees Celsius as the optimum for balanced folding and soluble expression, while time-course evaluations over 6 to 21 hours showed that an induction period of 12 hours generated peak enzymatic activity.

The physical and biochemical conditions governing the resting whole-cell biotransformation were subsequently optimized in single-factor trials tracking product concentration across differing operational environments. Examining reaction temperatures from 20 to 45 degrees Celsius revealed an initial upward trajectory in Danshensu synthesis that peaked sharply at 35 degrees Celsius before undergoing thermal decline, boosting product recovery to 8.66 mM. Evaluation of reaction pH values spanning 6.0 to 8.5 in phosphate and Tris-HCl buffers established 7.5 as the ideal operational threshold, with deviations above this point significantly lowering yields. Assessing wet cell biomass densities between 20 and 70 grams per liter demonstrated that while overall product output rose with increasing biocatalyst concentration, the specific synthetic activity plateaued at 50 grams per liter, establishing this density as the most cost-effective and operationally stable biocatalyst loading.

To evaluate commercial viability and process scalability, the researchers transferred the biotransformation into a 5-liter stirred-tank bioreactor operating a 1.5-liter working volume. Seed cultures grown in Terrific Broth at 35 degrees Celsius were introduced into the vessel, where agitation was dynamically regulated between 100 and 500 revolutions per minute alongside a ventilation rate of 5 liters per minute to maintain dissolved oxygen at 20 percent. Following induction with 0.5 mM IPTG at an OD600 of 0.93 and subsequent incubation at 25 degrees Celsius for 12 hours, wet cells were collected and resuspended in the biotransformation system comprising 25 mM L-tyrosine, 50 mM glucose, and 50 grams per liter wet cells maintained strictly at 35 degrees Celsius and pH 7.5.

Within the monitored bioreactor environment, strain 6 produced 19.47 mM Danshensu within 32 hours from the initial 25 mM L-tyrosine supply. Tracking intermediate and product accumulation over hourly sampling intervals revealed only minor accumulation of metabolic intermediates, supporting the conclusion that HpaBC and aadL govern overall cascade throughput. In comparison, running strain 7 under identical bioreactor settings converted 25 mM L-tyrosine into 17.44 mM Danshensu over the same operational window. Both strains exhibited continuous substrate conversion that closely tracked the progressive depletion of L-tyrosine, demonstrating that the single-pot cascade substantially simplifies manufacturing compared to previous three-step enzymatic approaches or complex multi-stage chemical purifications.

Despite these substantial biosynthetic gains, the researchers highlighted several technical constraints and outlined future experimental refinements. The current single-factor optimization methodology, while clarifying isolated operational thresholds, cannot fully evaluate complex multi-variable interactions between induction temperature, IPTG levels, biomass density, and solution pH. The team noted that applying multivariate statistical modeling will be necessary in future work to systematically resolve these dynamic interactions and further maximize carbon conversion efficiency. Nevertheless, the successful demonstration of a 19.47 mM yield in a bench-scale bioreactor confirms that whole-cell biotransformation by engineered E. coli provides an effective, mild, and green synthetic alternative for transforming inexpensive natural L-tyrosine into high-value pharmaceutical-grade Danshensu.

Subject of Research: Biosynthesis of Danshensu from L-tyrosine using engineered Escherichia coli cascade biotransformation

Article Title: Multienzyme cascade biotransformation of L-Tyrosine to Danshensu by engineered Escherichia coli

Article References: Xiong, T., Li, S., Li, X., Li, X., Wu, M., Liu, J., Guo, J., & Li, X. (2026). Multienzyme cascade biotransformation of L-Tyrosine to Danshensu by engineered Escherichia coli. PLOS One, 21(10), e0361484. https://doi.org/10.1371/journal.pone.0361484

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0361484

Keywords: Danshensu, L-Tyrosine, Escherichia coli, Biotransformation, Metabolic Engineering, Enzyme Cascade, HpaBC, aadL, ldhD, gdh, Bioreactor, Biosynthesis, Synthetic Biology, Green Chemistry, Whole-Cell Biocatalysis, Cardiovascular Therapeutics