Specialized metabolites in plants play crucial roles in communication and defense, and many find applications as flavors, dyes, and pharmaceuticals. While plant-based manufacturing enables commercial-scale production of some metabolites, most high-value products are synthesized at low concentrations. Chemical synthesis can be challenging due to the structural complexity and stereocontrol issues of plant metabolites. Microbial biosynthesis offers a promising alternative to chemical and pharmaceutical manufacturing. Significant achievements include the industrial-scale production of artemisinic acid and β-farnesene. Tetrahydroisoquinoline (THIQ) alkaloids comprise a large class of plant natural products (>3000), including benzylisoquinolines (BIAs), phenethylisoquinolines, ipecac alkaloids, and Amaryllidaceae alkaloids. These alkaloids possess a THIQ substructure that confers diverse bioactivities. Several THIQ derivatives are commercial drugs (e.g., galantamine, colchicine, apomorphine, trabectedin, cisatracurium, praziquantel, tetrabenazine). BIAs, the largest class of THIQs, include important medicines such as morphine and codeine. Current methods rely on extraction from opium poppy, but most BIAs do not accumulate sufficiently in plant tissues. While plant pathways for several BIAs have been reconstructed in yeast, titers remain very low (<2 mg L⁻¹), hindering industrial viability. A key objective in synthetic biology is to expand natural product diversity by engineering novel scaffolds and modifications not found in nature. The Pictet-Spengler condensation, central to THIQ biosynthesis, is known for its broad substrate specificity, yet natural pathways utilize a limited set of aldehydes. This study aimed to establish an efficient biosynthetic route to diverse THIQ alkaloids in vivo from simple substrates by focusing on overproduction of BIAs, specifically (S)-reticuline, a key intermediate in BIA biosynthesis, and expanding the synthesis of substituted THIQs using the Ehrlich pathway and diverse amino acid substrates.

Numerous studies have explored the biosynthesis of THIQ alkaloids in various organisms. Early work focused on elucidating the pathways in plants, identifying key enzymes like norcoclaurine synthase (NCS) and various methyltransferases and cytochrome P450 enzymes. Several research groups have successfully reconstructed parts of the BIA pathways in microorganisms, most notably in yeast (*Saccharomyces cerevisiae*) and *E. coli*. However, these efforts have been limited by low production titers, often in the microgram-per-liter range, primarily due to bottlenecks in the supply of key precursors like 4-hydroxyphenylacetaldehyde (4-HPAA) and dopamine. Efforts to improve these pathways have focused on metabolic engineering strategies, including the optimization of precursor supply, the introduction of more efficient enzymes, and the reduction of competing metabolic pathways. Several studies have demonstrated the use of enzyme engineering and pathway optimization to improve the production of specific BIAs, however, the production levels remain far below what is required for industrial applications. In vitro studies have highlighted the broad substrate specificity of NCS, suggesting the potential for generating a much wider range of THIQ structures than are naturally produced. This opened the possibility of synthesizing non-canonical THIQs by feeding the engineered system various aldehydes, but efficient in vivo synthesis remained a challenge.

This study employed a combination of metabolic engineering and synthetic biology techniques to create a high-yielding yeast platform for THIQ alkaloid production. The researchers began with a previously engineered yeast strain capable of producing dopamine, a key precursor to (S)-norcoclaurine. The initial focus was on improving (S)-norcoclaurine production by systematically modifying the shikimate, Ehrlich, and L-tyrosine metabolic pathways. This involved multiple rounds of genetic engineering using CRISPR-Cas9 technology and in vivo DNA assembly, including gene deletions, overexpressions, and the introduction of feedback-resistant enzyme variants. A significant challenge was the efficient conversion of L-tyrosine to 4-HPAA, which is rapidly metabolized to tyrosol or 4-hydroxyphenylacetic acid (4-HPAC) via the Ehrlich pathway. To overcome this, the researchers performed a gene deletion screen to identify and remove functionally redundant oxidoreductases responsible for this unwanted metabolism. The process involved systematic deletion of multiple genes encoding NADPH-dependent reductases and dehydrogenases. The researchers then focused on improving the supply of dopamine through the introduction of a more efficient tyrosine hydroxylase ortholog (CYP76AD5). The production of (S)-reticuline was achieved by introducing a multi-enzyme pathway module consisting of methyltransferases and cytochrome P450 enzymes from opium poppy and California poppy. This process also involved optimizing the expression of these heterologous genes and addressing bottlenecks in the pathway through further gene deletions. In addition to optimizing the canonical BIA pathway, the researchers explored the promiscuity of NCS by analyzing the LC-MS spectra of culture supernatants for peaks indicative of substituted THIQ products. A series of assays were conducted to explore the production of THIQs from various endogenous and exogenous amino acids by cultivating the engineered strains on different amino acids as the major nitrogen source. The identification of these THIQs was confirmed by LC-MS analysis including targeted fragmentation of the products and comparing the data with that of authentic standards. Finally, the researchers investigated the ability of BIA tailoring enzymes to modify the newly synthesized substituted THIQs, using various LC-MS techniques to analyze and confirm the compounds' identities.

The researchers achieved a significant 57,000-fold increase in (S)-reticuline production compared to their previous work, reaching titers of 4.6 g L⁻¹ in a simple mineral medium using a fed-batch fermentation process. This high titer was obtained via extensive metabolic engineering of the yeast, including optimization of the L-tyrosine and Ehrlich pathways, the introduction of improved enzyme variants, and the deletion of a consortium of functionally redundant oxidoreductases responsible for the conversion of 4-HPAA to tyrosol or 4-HPAC. The high (S)-reticuline titers coincided with the production of several substituted THIQs derived from amino acid catabolism. By inactivating specific enzymes involved in fusel alcohol and acid formation (Aril, Adh6, Ypr1, Ydr541c, Aad3, Gre2, and Hfd1), they successfully rerouted carbon flux towards the desired THIQ products. The use of a more efficient tyrosine hydroxylase ortholog (CYP76AD5) further improved dopamine supply, which proved crucial for enhancing BIA production. The study revealed that the Ehrlich pathway in yeast can be repurposed for the biosynthesis of a diverse array of THIQs. The engineered yeast strain produced various substituted THIQs from both endogenous and fed amino acids. The researchers synthesized over ten diverse THIQ scaffolds, demonstrating the broad substrate specificity of NCS and the ability to incorporate both aromatic and aliphatic substituents onto the THIQ core structure. The identities of many of these THIQs were confirmed by LC-MS and comparison to authentic standards. The study also showed that BIA tailoring enzymes (Ps6OMT and PsCNMT) could efficiently modify the newly synthesized substituted THIQs, demonstrating the potential for generating a vast array of modified THIQ structures.

This work makes significant progress towards industrial-scale production of THIQ-based pharmaceuticals, particularly opioid analgesics. The exceptionally high (S)-reticuline titer (4.6 g L⁻¹) achieved represents a substantial advancement in microbial biosynthesis of complex plant metabolites. Many bioprocesses require gram-per-liter titers for commercial viability, and this study demonstrates that this is achievable for a key intermediate in opioid biosynthesis. The repurposing of the yeast Ehrlich pathway expands the chemical diversity of accessible THIQ compounds significantly, generating structures not observed in nature. This significantly expands the scope for discovering and producing novel drugs and pharmaceuticals with potential bioactivity, derived from the highly versatile THIQ scaffold. The identification and inactivation of multiple functionally redundant oxidoreductases responsible for fusel product formation provide valuable insights into yeast metabolism and highlight the power of systematic metabolic engineering for pathway optimization. The potential applications extend beyond THIQs; the engineered yeast's high production of aldehydes makes it a valuable platform for producing other carbonyl-derived compounds, like biofuels or fragrances. The broad substrate specificity of NCS and the successful incorporation of both aromatic and aliphatic substituents onto the THIQ core open doors to creating highly diverse THIQ libraries.

This study demonstrates a highly efficient yeast platform for producing gram per liter titers of the key BIA intermediate (S)-reticuline, providing a foundation for industrial-scale microbial production of natural and semi-synthetic opioids. The successful redirection of the Ehrlich pathway enables the synthesis of diverse THIQ scaffolds with significant potential for drug discovery. Future research could focus on further optimizing the downstream steps in morphine biosynthesis within this high-producing yeast strain, exploring the full potential of THIQ diversity through combinatorial biosynthesis approaches, and investigating the biological activities of the newly synthesized THIQ analogs.

While the study achieved exceptionally high (S)-reticuline titers, the overall production of morphine and other downstream alkaloids remains to be fully optimized. The identification of substituted THIQs is primarily based on LC-MS data; future work could utilize NMR spectroscopy to definitively confirm the structures of all newly synthesized compounds. The study focuses on a specific yeast strain; the generalizability of these findings to other microbial hosts needs further investigation. Finally, the scale-up of this process to industrial levels will require further optimization and process engineering.