Metabolic engineering of Synechocystis sp. PCC 6803 for olivetolic acid production

Abstract

Olivetolic acid (OLA), the critical aromatic precursor for plant-derived cannabinoids, conventionally relies on resource-intensive plant extraction or complex chemical synthesis. This study establishes an environmentally friendly biosynthetic platform by genetically engineering the model cyanobacterium Synechocystis sp. PCC 6803. We constructed the heterologous OLA pathway by co-expressing plant-derived tetraketide synthase (TKS), olivetolic acid cyclase (OAC), and acyl-activating enzyme 1 (AAE1). Through cultivation supplemented with sodium hexanoate, successful biosynthesis of OLA was achieved within this photosynthetic host. To overcome metabolic bottlenecks, the pathway was further optimized using the robust P_cpc560 promoter coupled with carbon sink redirection strategy. Disruption of the primary glycogen storage sink (ΔglgC) combined with the overexpression of the anaplerotic malic enzyme gene (maeB) significantly enhanced the specific OLA yield. Notably, while continuous aeration with 5% CO₂ drastically accelerated cell growth and prolonged the cultivation period, it simultaneously led to a severe decrease in the specific OLA yield. This work demonstrates the feasibility of utilizing a cyanobacterial chassis for precursor-directed OLA biosynthesis, establishing a foundational blueprint for the sustainable manufacturing of cannabinoid-related metabolites.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12934-026-03058-y.

Highlights

Initial synthesis of olivetolic acid in Synechocystis sp. PCC 6803.

Production titer increased after knocking out the glycogen synthesis pathway.

Overexpression assays optimized the metabolic pathway and increased production.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12934-026-03058-y.

Introduction

Phytocannabinoids, such as cannabidiol (CBD) and Δ⁹-tetrahydrocannabinol (THC), represent a class of highly valued meroterpenoids produced by Cannabis sativa (C. sativa) [1]. They exert therapeutic effects through modulation of the endocannabinoid system (CB1/CB2 receptors) [2], evidenced by FDA-approved drugs for epilepsy (Epidiolex^® - CBD) and chemotherapy-induced nausea (Marinol^®/Syndros^® - synthetic THC) [3, 4]. Typical production relies on plant extraction or chemical synthesis [1]. However, the sustainable supply of cannabinoid drugs via traditional plant-based extraction is severely bottlenecked. On one hand, production is hindered by the protracted growth cycles of C. sativa (typically 3–4 months) and stringent legal frameworks (generally restricting cultivation to low-THC varieties) [5]. On the other hand, low specialized metabolite titers and resource-intensive farming render the process both economically and ecologically inefficient [6]. While chemical synthesis provides a potential route, it remains challenging due to the intricate molecular architecture, necessitating the use of hazardous organic solvents and metal catalysts that not only raise environmental concerns but also pose significant hurdles for the downstream purification of pharmaceutical-grade products [7]. Consequently, there is an urgent imperative to develop sustainable, carbon-neutral production platforms that leverage green manufacturing principles to circumvent the multifaceted ecological and technical bottlenecks inherent in traditional cannabinoid supply chains.

As the major specialized metabolites in C. sativa, phytocannabinoids are predominantly biosynthesized in glandular trichomes and reach their maximum accumulation on the bracts of female inflorescences [8]. While over 150 different cannabinoids have been identified to date, the biosynthetic pathways of the most extensively studied analogs—including CBD, THC, cannabichromene (CBC), and cannabigerol (CBG)—all converge at a common core precursor, olivetolic acid (OLA) [9, 10]. In the native C. sativa plant, the formation of this pivotal aromatic intermediate is governed by a three-step enzymatic sequence: The pathway initiates with the activation of the short-chain fatty acid hexanoate into hexanoyl-CoA by the acyl-activating enzyme (AAE1) in an ATP-dependent manner. This hexanoyl-CoA then serves as the starter unit for the tetraketide synthase (TKS), a type Ⅲ polyketide synthase, which performs the iterative decarboxylative condensation of three molecules of malonyl-CoA to generate a linear triketide intermediate. To prevent the spontaneous formation of metabolic side-products, the OAC acts as a regiospecific scaffold to direct the C2-C7 aldol condensation and subsequent cyclization of the intermediate, ultimately yielding the pivotal aromatic intermediate, OLA [11].

The elucidation of this native enzymatic logic has paved the way for synthetic biology to transcend the limitations of traditional plant-based production. Since the production of OLA is a fundamental prerequisite for cannabinoid synthesis, researchers have successfully reconstituted this metabolic pathway in various heterologous microbial hosts. To date, research has predominantly focused on heterotrophic platforms, underscoring the feasibility of OLA production via advanced metabolic engineering strategies. Escherichia coli (E. coli) was among the earliest chassis employed for the biosynthesis of OLA. Tan et al. [12] reconstituted this pathway by integrating C. sativa OLS (polyketide synthase) and OAC with a reversed β-oxidation cycle for de novo hexanoyl-CoA synthesis. Through integrated metabolic optimization and controlled bioreactor cultivation, this strategy ultimately achieved an OLA titer of 80 mg·L^{− 1}. In the Yarrowia lipolytica chassis, Ma et al. [13] achieved an 83-fold increase in OLA titers to 9.18 mg·L^{− 1} by optimizing precursor supply, augmenting cofactor regeneration, and integrating metabolic pathways to effectively redirect carbon flux toward polyketide biosynthesis. Similarly, Schmidt et al. [14] optimized Saccharomyces cerevisiae by introducing CsAAE1, CsOLS, and CsOAC for OLA synthesis, alongside a soluble NphB and ERG20^WW fusion to bypass the limitations of membrane-bound prenyltransferases. By further overexpressing the transcription factor HAC1s to boost enzyme activity, they achieved 117 mg·L^{− 1} of OLA and successfully synthesized cannabigerolic acid (CBGA) and cannabichromenic acid (CBCA). However, cannabinoid biosynthetic enzymes of plant origin are highly susceptible to forming inactive inclusion bodies within E. coli, which limits the yield of soluble active proteins [15]. While yeast is considered energy-intensive in production due to its reliance on refined organic carbon sources and the requirement for high-density cultivation [16]. Furthermore, the heavy reliance of these conventional hosts on organic carbon sources necessitates the exploration of more sustainable and cost-efficient production platforms.

The model cyanobacterium Synechocystis sp. PCC 6803 (Synechocystis) represents a versatile photosynthetic chassis exceptionally suited for the sustainable biosynthesis of plant-derived metabolites, including terpenoids, flavonoids, and phenolics [17–21]. Unlike conventional heterotrophic hosts that require costly organic feedstocks, Synechocystis operates photoautotrophically—utilizing ambient CO₂ and solar energy to enable carbon-negative manufacturing [22]. Benefiting from a rapid doubling time (7–10 h) and a sophisticated carbon-concentrating mechanism (CCM), the organism achieves superior photosynthetic performance and highly efficient nitrogen utilization [23, 24]. Structurally, cyanobacteria share with plant cells thylakoid membranes containing galactolipids, sulfolipids, various photosystem components, and photosynthetic pigments [24–26]. Metabolically, the cyanobacterial photosynthetic apparatus yields substantial reducing equivalents and critical precursor pools (e.g., acetyl-CoA). This dual congruence renders cyanobacteria inherently compatible with the biosynthesis of plant-derived metabolites, such as terpenoids and phenolics [19, 27]. Collectively, these unique features, complemented by the facile genetic engineering of cyanobacteria, position this organism as an optimal photosynthetic platform for the sustainable biosynthesis of cannabinoids and their respective precursors.

As a proof of concept, we established a biosynthesis platform for the synthesis of the pivotal precursor, OLA, in Synechocystis. This was achieved through an integrated metabolic engineering strategy that coupled the heterologous expression of cannabinoid biosynthetic genes (TKS, OAC, and AAE1) with host-specific modifications, including the disruption of carbon storage and the reinforcement of the malonyl-CoA pool. This work presents a novel and sustainable alternative for microbial-based cannabinoid biomanufacturing platforms.

Materials and methods

Strains and growth conditions

The wild type and genetically engineered Synechocystis strains were cultivated in BG-11 medium using an illuminated incubator (GDN-260 A, YANGHUI, China) under a continuous light intensity of 28 µmol m^− 2 s^− 1 and a temperature of 30℃. The cell density of strains was measured at an optical density (OD) of 730 nm using a visible spectrophotometer (YOKE, Shanghai, China). Except for the aeration experiments, all cultivations were conducted under ambient air conditions.

The DH5α Chemically Competent Cell (Vazyme, Nanjing, China) was used for plasmid construction and amplification. The DH5α strain was cultured in LB medium at 37℃ with shaking at 220 rpm (Zhichu, Shanghai, China).

Different antibiotic concentrations were employed for various strains during cultivation. For genetically engineered cyanobacteria, spectinomycin was added at a final concentration of 50 µg/mL, while chloramphenicol was added at a final concentration of 34 µg/mL. During the initial selection after cyanobacterial transformation, antibiotics were applied at one-fifth of the regular cultivation concentrations and gradually increased to the required levels for routine.

Plasmids construction

This study constructed a series of integration plasmids based on the pMD18-T backbone (Sangon Biotech Co Ltd., Shanghai, China) to integrate optimized synthetic exogenous genes (Sangon Biotech Co Ltd., Shanghai, China) into the cyanobacterial chromosome. The plasmid construction involved restriction digestion using endonucleases (NEB, Beijing, China), PCR amplification with high-fidelity polymerase (Vazyme, Nanjing, China), and assembly using homologous recombinase (NEB, Beijing, China). DH5α was used for plasmid amplification. Single colonies were screened via colony PCR, and verified plasmids were sent for sequencing. Subsequent experiments proceeded after confirmation of correct sequences. The characteristics of all plasmids used for transformation in this study are summarized in Table 1. Plasmids construction process and primers are provided in the Appendix Table S1 and S2.

Table 1.

Plasmids used for transformation in this study

Plasmids	Characteristic	Resistance
pST407-TKS/OAC	slr0168up-SpecR-PpsbA2s-TKS-OAC-TrbcL-slr0168dw	AmpR+SpecR
pST02	slr0168up-SpecR-PpsbA2s-TKS-OAC-AAE1-TrbcL-slr0168dw	AmpR+SpecR
pST03	slr0168up-SpecR-Pcpc560-TKS-OAC-AAE1-TrbcL-slr0168dw	AmpR+SpecR
pST04	slr1176up-CmR-slr1176dw	AmpR+ CmR
pST05	slr1176up-CmR-Pcpc560-maeB-TrbcL-slr1176dw	AmpR+ CmR

slr0168up: upstream homology arm of the slr0168 locus, slr0168dw: downstream homology arm of the slr0168 locus, slr1176up: upstream homology arm of the slr1176 locus, slr1176dw: downstream homology arm of the slr1176 locus, Amp^R: ampicillin resistance gene, Spec^R: spectinomycin resistance gene, Cm^R: chloramphenicol resistance gene, P_psbA2s: a derivative of P_psbA2 promoter with shorter sequence, P_cpc560: a cyanobacterial native super strong promoter, T_rbcL: a cyanobacterial native terminator, TKS: tetraketide synthase, OAC: olivetolic acid cyclase, AAE1: hexanoyl-CoA synthase, maeB: NADP-dependent malic enzyme

Transformation of Synechocystis

Transformation of Synechocystis was performed using a method described by Williams et al. [28–30].

Cultures in the exponential phase (30 mL, OD₇₃₀ = 0.6–0.8, cultured at 30℃ with a light intensity of 28 µmol m^− 2 s^− 1) were harvested by centrifugation, washed twice with fresh BG-11 medium, and resuspended in 1 mL of antibiotic-free BG-11 medium. To the suspension, 10 µg of the target plasmid was added and mixed by gentle pipetting. The mixture was incubated for 18 h in a low light incubator, with gentle shaking 3–4 times during this period. Following incubation, the mixture was spread onto five antibiotic-free BG-11 agar plates, each containing a 0.45 μm hydrophilic cellulose membrane, pretreated by boiling and autoclaving. The plates were incubated for 24 h, after which the membranes were transferred to BG-11 plates supplemented with appropriate antibiotic concentration.

Transformants appeared after 2–3 weeks. Individual colonies were isolated and streaked onto BG11 agar plates containing antibiotics for further purification. One colony from each plate was then inoculated into liquid BG-11 medium with a low concentration of antibiotics, with the antibiotic concentration gradually increased during serial passaging. After five passages (approximately one month), cells in the exponential phase were harvested for genomic DNA extraction and PCR-based validation. The construction process for all cyanobacterial strains in this study are presented in Appendix Fig. S1.

The genetic modifications in this study involve loci slr0168 and slr1176. The locus slr0168 is a neutral site on the Synechocystis genome, encoding a hypothetical glycoprotein and forming part of the slr0338-slr0168-slr0169 operon. The functions of this operon and the genes within it remain unclear [31]. Typically, modifications at neutral sites do not significantly affect cellular functions [32], which is why this locus is frequently chosen for introducing exogenous genes when Synechocystis is used as a chassis organism. The locus slr1176 ([EC:2.7.7.27]) encodes ADP-glucose pyrophosphorylase, referred to as glgC. This gene is involved in the first step of glycogen and starch biosynthesis in Synechocystis, and the activity of GlgC protein regulates the overall rate of glycogen accumulation. If this gene is knocked out, glycogen cannot be synthesized in Synechocystis [33].

OLA extraction and measurement

All engineered strains were cultured in BG–11 medium supplemented with sodium hexanoate at a light intensity of 28 µmol m^{− 2} s^{− 1} and incubated at 30℃. Cyanobacterial cells were first disrupted using an ultrasonic cell disruptor, and the resulting mixture was centrifuged at 8000 rpm for 10 min. The supernatant was then transferred and extracted with an equal volume of ethyl acetate [34]. After combining the extracts, the solution was concentrated to 1 mL by rotary evaporation and subsequently dried under a flow of nitrogen gas. The residue was re-suspended in 1 mL methanol and filtered through a 0.22 μm syringe filter before HPLC (high-performance liquid chromatography) analysis. HPLC analysis was performed using a Shimadzu LC-20AT system with a Thermo Scientific Hypersil ODS-2 C18 column (4.6 mm × 250 mm, 5 μm) and detection at 262 nm.

Improvement of culture conditions

Since cyanobacterial growth and OLA production rely on CO₂ consumption, continuous aeration was employed to enhance CO₂ supply. Logarithmically growing strains were inoculated to an initial OD₇₃₀ of 0.06. The CO₂ was mixed with air at a 1:19 ratio (5% CO₂, 95% air), filtered through a 0.22 μm membrane and continuously supplied to the culture flasks. Cultures were maintained at 30℃ and illuminated at 28 µmol m^− 2 s^− 1 with continuous aeration for 30 days.

Statistics and reproducibility

Data are presented as mean ± standard deviation (SD) from three independent biological replicates. Statistical analysis was performed with GraphPad Prism (version 10.4.1) using ordinary one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test. Statistically significant differences (p < 0.05) among groups are indicated in the figures using the compact letter display (CLD) method, where sharing the same letter represents no significant difference.

Results

Construction of synthetic OLA pathway

To establish a biosynthesis platform for OLA production, five engineered cyanobacterial strains were constructed in this study (Appendix Fig. S1). The gene clusters of the engineered cyanobacterial strains are shown in Table 2, and the genomic PCR verification are shown in Appendix Fig. S2. The metabolic pathways involved in the genetic modification are shown in Fig. 1.

Table 2.

The gene cluster of engineered cyanobacterial strains

Fig. 1.

Modifications of metabolic pathway in engineered strains. In the engineered cyanobacteria constructed in this study, the heterologous OLA synthesis pathway (Blue part) begins with exogenously supplied sodium hexanoate and endogenously derived malonyl-CoA. Hexanoate is converted to hexanoyl-CoA by the enzyme AAE1. Subsequently, TKS catalyzes three successive two-carbon extensions using malonyl-CoA as the donor via decarboxylative condensation. The intermediate is finally cyclized at the C2-C7 position through an alcohol-aldehyde cyclization catalyzed by OAC, forming OLA. To redirect carbon flux toward product synthesis, the glgC gene was knocked out to block glycogen synthesis (Red part), and the maeB gene was introduced to adjust cellular metabolism (Orange part)

In our initial design, the TKS and OAC genes were organized into an operon under the control of the constitutive promoter P_psbA2s and integrated into the neutral site slr0168 of the Synechocystis genome (Strain SCY01). However, HPLC analysis revealed that SCY01 failed to produce detectable OLA under routine photoautotrophic conditions (Appendix Fig. S3). We hypothesized that the primary bottleneck was the insufficient endogenous pool of hexanoyl-CoA.

To address the precursor limitation and verify the enzymatic functionality, we introduced the AAE1 gene to activate exogenous hexanoic acid into its CoA-thioester form (Strain SCY02). As shown in Fig. 2A, initial assessments showed that in the absence of hexanoate supplementation, OLA was still barely detectable, confirming that hexanoic acid is not naturally accumulated in Synechocystis. To optimize the precursor supply, a range of sodium hexanoate concentrations (2, 4, 6, 8, 10, and 12 mM) was supplemented.

Fig. 2.

Effects of hexanoic concentration on the production and growth of the engineered strain SCY02. OLA titer A and cell growth B of SCY02 strain at day 14 with hexanoic acid under ambient air conditions. The initial cell concentration OD₇₃₀ = 0.06. Values are presented as mean ± SD (n = 3). Data with different superscript letters differ significantly (p < 0.05, one-way ANOVA with Tukey’s HSD test), whereas sharing the same letter indicates no significant difference

The volumetric OLA titers increased dose-dependently with exogenous hexanoate in the range of 0–8 mM, peaking at 0.870 mg·L^− 1 with the 8 mM supplement. Beyond this threshold (10–12 mM), severe inhibition of cyanobacterial growth led to a sharp decline to 0.137 mg·L^− 1 and 0.097 mg·L^− 1, respectively (Fig. 2B). The effects of different concentrations of sodium hexanoate on the growth of strain SCY02 over 14 days are shown in Figure S7. Since no significant growth inhibition was observed at a concentration of 8 mM, this concentration was selected for all subsequent experiments.

Effect of promoter on OLA production

Given the functional validation of the OLA pathway in SCY02, we next sought to enhance metabolic flux by optimizing the transcriptional strength of the heterologous genes. We replaced the P_psbA2s promoter in SCY02 with the super-strong constitutive promoter P_cpc560 to drive the TKS-OAC-AAE1 operon, resulting in strain SCY03.

Under the optimized 8 mM sodium hexanoate supplementation, the volumetric OLA titer in SCY03 reached 1.71 mg·L^− 1 by day 14, representing a significant improvement compared to SCY02 (Fig. 3). Strain SCY03 exhibited a continuous accumulation of OLA throughout the cultivation period (Appendix Fig. S5). To determine whether the enhancement in OLA accumulation stemmed from increased biomass or improved intracellular carbon partitioning, OLA titers were normalized to growth density (expressed as mg·OD₇₃₀⁻¹·L^− 1). Interestingly, the specific yield provided distinct insights across two dimensions. On the one hand, within the same strain (SCY03), the comparable specific yield from day 3 to 14 indicates that the temporal increase in volumetric OLA titers was primarily driven by biomass expansion. On the other hand, when comparing different strains, SCY03 exhibited an 89.89% increase in specific yield over SCY02, reaching 0.883 mg·OD₇₃₀⁻¹·L^− 1 compared to 0.465 mg·OD₇₃₀⁻¹·L^− 1 for the latter (Fig. 4), demonstrating an enhanced overall OLA biosynthetic capacity consistent with the higher transcription-initiation frequency typically associated with the P_cpc560 promoter in Synechocystis. The results indicate that increasing the expression levels of the TKS-OAC-AAE1 operon effectively bolstered the metabolic flux toward OLA. However, the overall yield remained limited, suggesting that the supply of the endogenous precursor, malonyl-CoA, might have become the next bottleneck.

Fig. 3.

Different metabolic modification strategies to promote OLA production. A OLA titer of cyanobacterial strains at day 14 with 8 mM sodium hexanoate and cultivated under ambient air conditions. Values are presented as mean ± SD (n = 3). Data with different superscript letters differ significantly (p < 0.05, one-way ANOVA with Tukey’s HSD test), whereas sharing the same letter indicates no significant difference. B Extracts were analyzed by HPLC at a wavelength of 262 nm and signals were compared to authentic OLA standards (Appendix Fig. S4). The retention time of OLA was approximately 18.1 min. Two subsequent minor peaks were likely derived from endogenous lipid backgrounds or other interfering metabolites in the cyanobacterial extracts, posing no obstruction to the identification and integration of the target peak

Fig. 4.

OLA specific yield and cell growth of engineered cyanobacterial strains. Times courses of OLA specific yield (bars) and cell growth (circles) in the (A/B)SCY01, (C/D)SCY02, (E/F)SCY03, (G/H)SCY04 and (I/J)SCY05. Values are presented as mean ± SD (n = 3). Statistics for normalized yields (mg·OD₇₃₀⁻¹·L^− 1) were derived via the law of error propagation based on the independent measurements of titer and OD₇₃₀. The initial cell concentration OD₇₃₀ = 0.06. Data with different superscript letters differ significantly (p < 0.05, one-way ANOVA with Tukey’s HSD test), whereas sharing the same letter indicates no significant difference. All cyanobacterial strains were supplemented with 8 mM sodium hexanoate and cultivated under ambient air conditions for a continuous period of 14 days

Deletion of the glgC from the Synechocystis genome

While the supply of hexanoyl-CoA and the efficiency of the metabolic pathway were optimized, the limited availability of malonyl-CoA was identified as the next bottleneck restricting further increases in OLA biosynthetic capacity. In Synechocystis, a substantial portion of fixed carbon is diverted toward glycogen, which serves as a major intracellular carbon sink and energy reserve. To alleviate this competition for carbon precursors, we disrupted the glycogen biosynthetic pathway by knocking out the glgC gene in the SCY03 background, generating strain SCY04.

To evaluate the physiological impact of glycogen synthesis disruption, the cell growth, volumetric OLA titer, and specific yield of strain SCY04 were systematically monitored over a 14-day cultivation period. Intriguingly, the specific yield varied depending on the cultivation stage (Fig. 4G). Following an initial rise, the specific yield decreased to a temporary nadir on day 7 (marked by letter BC). This decline in value occurred precisely when the cell density was undergoing a relatively fast increase. After day 7, as the growth rate decelerated and the culture progressively transitioned into the stationary phase (days 10–14), the specific yield flattened out and maintained a relatively constant level. Under the 8 mM sodium hexanoate supplementation, SCY04 achieved a final specific yield of 1.358 mg·OD₇₃₀⁻¹·L^− 1 by day 14 (Fig. 5), representing an approximately 53.79% increase compared to its parental strain SCY03 (0.883 mg·OD₇₃₀⁻¹·L^− 1). This improvement confirms that blocking the primary carbon sink (glycogen) effectively increases the intracellular malonyl-CoA pool, thereby providing more building blocks for polyketide synthesis.

Fig. 5.

OLA specific yield of engineered cyanobacterial strains. Cultivation conditions are defined as LC (ambient air, day 14) and HC (5% [v/v] CO₂ bubbling, day-30), with 8 mM sodium hexanoate supplemented in all treatments. Values are presented as mean ± SD (n = 3). Statistics for normalized yields (mg·OD₇₃₀⁻¹·L^− 1) were derived via the law of error propagation based on the final-day independent measurements of titer and OD₇₃₀. Data with different superscript letters differ significantly (p < 0.05, one-way ANOVA with Tukey’s HSD test), whereas sharing the same letter indicates no significant difference

As shown in Fig. 4H, the disruption of the glgC gene did not significantly alter the overall growth trend or cell density compared to the parental strain SCY03. Both SCY03 and SCY04 strains exhibited continuous cell growth during the first 10 days before progressively stabilizing, with final OD₇₃₀ values reaching 1.937 and 1.789 on day 14, respectively (Fig. 6). This absence of macroscopic growth inhibition demonstrates the robust nature of the Synechocystis chassis under core storage pathway manipulations. Despite the stable growth profile, the fluctuations in the specific yield indicated that the coordination between biomass synthesis and product allocation varied across different cultivation stages. Nevertheless, the high chassis robustness under altered carbon flux conditions implies a great potential for heterologous pathway efficiency, prompting us to explore further strategies to balance carbon fixation and metabolic drainage to achieve even higher titers.

Fig. 6.

Growth density of cyanobacterial strains. The initial cell concentration OD₇₃₀ = 0.06. Values are presented as mean ± SD (n = 3). Data with different superscript letters differ significantly (p < 0.05, one-way ANOVA with Tukey’s HSD test), whereas sharing the same letter indicates no significant difference. All strains were supplemented with 8 mM sodium hexanoate and cultivated under ambient air conditions for a continuous period of 14 days

Effect of maeB overexpression on OLA production

To further enhance OLA productivity, we developed strain SCY05 by integrating the E. coli maeB gene into the slr1176 (glgC) locus of SCY03. This site-specific integration was designed to simultaneously block the major carbon sink (glycogen) and introduce a heterologous malic enzyme.

At the end of the 14-day cultivation, strain SCY05 achieved the highest growth density with an OD₇₃₀ of 2.410, which was markedly superior to that of SCY03 (OD₇₃₀ = 1.937) and SCY04 (OD₇₃₀ = 1.789). This enhanced growth performance indicates that the site-specific integration of the maeB gene into the glgC locus did not impair the viability or robustness of the host; instead, it supported more efficient biomass accumulation. Marked variations were also observed for specific OLA yieid (Fig. 4I). While strain SCY03 exhibited slow and limited specific OLA accumulation, peaking at only 0.883 mg·OD₇₃₀⁻¹·L^− 1 by day 14, strains SCY04 and SCY05 demonstrated rapid accumulation shortly after inoculation and maintained higher rates throughout the time course. By day 14, the specific yield of SCY05 reached 1.722 mg·OD₇₃₀⁻¹·L^− 1, representing a substantial increase of 95.02% and 26.80% compared to SCY03 (0.883 mg·OD₇₃₀⁻¹·L^− 1) and SCY04 (1.358 mg·OD₇₃₀⁻¹·L^− 1), respectively. The volumetric titers further confirmed the superior metabolic efficiency of the SCY05 strain (Fig. 3). All strains showed a time-dependent increase in volumetric OLA concentration. In alignment with the growth and specific yield trends, strain SCY05 consistently outperformed the other strains across the entire cultivation period. By day 14, SCY05 achieved a final OLA volumetric titer of 4.15 mg·L^− 1, demonstrating the highest OLA biosynthetic capability among all the engineered strains evaluated in this study.

Influence of aeration on OLA production performance

To evaluate the long-term OLA production potential of the engineered strain SCY05, we monitored its growth and biosynthetic performance under two distinct cultivation regimes: ambient air for 14 days (designated as the LC condition) and 5% CO₂ aeration for an extended 30-day period (designated as the HC condition).

Under LC conditions, cell growth of SCY05 entered the stationary phase at approximately day 12, reaching a maximum OD₇₃₀ of 2.410 by day 14 (Fig. 4J). The specific OLA yield rapidly increased to 1.422 mg·OD₇₃₀⁻¹·L^− 1 on day 1, and then gradually plateaued, fluctuating between 1.511 and 1.780 mg·OD₇₃₀⁻¹·L^− 1 from day 5 to day 14 (Fig. 4I). Consequently, the volumetric OLA titer accumulated progressively with biomass growth, reaching a maximum of 4.150 mg·L^− 1 on day 14 (Appendix Fig. S5E).

Under HC conditions, the biomass accumulation of SCY05 was substantially enhanced; the OD₇₃₀ reached 2.460 on day 3, which already exceeded the maximum cell density achieved under the LC condition over the entire 14-day period (Fig. 7B). The culture then entered the stationary phase at around day 21, stabilizing at 14.074 through day 30. However, despite this remarkable enhancement in biomass, the specific OLA yield was significantly compromised under HC conditions, with a final value of 0.669 mg·OD₇₃₀⁻¹·L^− 1 on day 30(Fig. 7A). The volumetric OLA titer under HC conditions on day 15 was 5.285 mg·L^− 1, which was comparable to the 14-day titer of 4.150 mg·L^− 1 under LC conditions (Appendix Fig. S5E and S6). Even by the end of the extended cultivation on day 30, the total volumetric titer was limited to 9.410 mg·L^− 1. Overall, while aeration vastly accelerated the growth rate of the strain and prolonged the time required to reach the stationary phase, it led to a substantial reduction in the specific biosynthetic yield.

Fig. 7.

Effects of aeration conditions on the engineered cyanobacterial strain SCY05. OLA specific yield A and cell growth B of strain SCY05 in 30 days with bubbling 5% CO₂-air (v/v) and adding 8 mM hexanoic. The initial cell concentration OD₇₃₀ = 0.06. Values are presented as mean ± SD (n = 3). Data with different superscript letters differ significantly (p < 0.05, one-way ANOVA with Tukey’s HSD test), whereas sharing the same letter indicates no significant difference

Discussion

Over the past years, OLA has been successfully synthesized in several heterotrophic microorganisms, including E. coli [12], Yarrowia lipolytica [13] and Saccharomyces cerevisiae [14]. However, these platforms necessitate the consumption of refined organic carbon sources, posing challenges for large-scale sustainable manufacturing. In this study, we transitioned OLA production to the photoautotrophic chassis Synechocystis. Since free fatty acids (FFAs) can be released and recycled during membrane lipid turnover in Synechocystis [35], and medium-chain free fatty acids are detectable during the cultivation of Synechococcus 7942 [36], we hypothesized that Synechocystis could produce C6:0 FFA and possesses the enzymes required to convert them into their corresponding CoA forms.

While the heterologous TKS and OAC genes were successfully integrated into the genome, the initial strain SCY01 failed to accumulate OLA under standard conditions (Appendix Fig S3). To determine whether this bottleneck stemmed from a shortage of precursor free hexanoic acid or a deficiency in acyl-activation, we engineered strain SCY02 to co-express AAE1. Comparative cultivation of these strains with and without sodium hexanoate supplementation confirmed that both a sustained precursor supply and active acyl-activation are indispensable for driving OLA biosynthesis. Native Synechocystis prioritizes FFA flux toward long-chain acyl-ACPs to sustain membrane biogenesis rather than accumulating short-chain CoA thioesters [36]; rendering endogenous hexanoic acid pools negligible. Furthermore, directly engineering early lipid pathways risks disrupting thylakoid membrane architecture, phycobilisome assembly, and the function of membrane-bound photosynthetic complexes [37]. To bypass these physiological constraints and isolate pathway functionality, we supplied the six-carbon precursor (C6:0) via exogenous sodium hexanoate supplementation.

The uptake of medium- and short-chain fatty acids by cyanobacterial cells usually occurs through energy-independent mechanisms, including porins or passive diffusion [38]. When determining the optimal sodium hexanoate concentration, cyanobacterial cell growth and volumetric OLA titers were both enhanced in cultures supplemented with less than 8 mM sodium hexanoate (Fig. 2B). This enhancement could be attributed to a higher intracellular-extracellular concentration gradient that facilitated greater hexanoate influx. Alternatively, imported hexanoate may enter native catabolism as an auxiliary carbon source, indirectly supporting biomass accumulation and boosting productivity. Conversely, concentrations exceeding 10 mM triggered severe cytotoxicity and arrested cell growth. Because the pH shift (~ 8.4) induced by 10 mM hexanoate remains exceptionally close to the growth optimum for Synechocystis (~ 8.5) [39], this growth inhibition is likely driven by lipophilicity-mediated disruption of plasma and thylakoid membrane integrity.

Beyond the initial functional assembly of the heterologous pathway, optimization of genetic parts is essential to align the host’s metabolic flux with non-native synthesis. In this study, replacing the P_psbA2s promoter with the robust, constitutive P_cpc560 promoter led to a substantial increase in volumetric OLA titers (Strain SCY03 vs. SCY02). To account for variations in cell growth, volumetric OLA titers from all strains were normalized against biomass and expressed in mg·OD₇₃₀⁻¹·L^{− 1}. The resulting specific productivity data further verified that the elevation in OLA accumulation was driven by augmented enzyme activity into the engineered pathway instead of an increase in biomass. This performance disparity reflects the enzymatic demand required to sustain heterologous TKS activity in Synechocystis host. TKS must orchestrate the iterative condensation of three malonyl-CoA molecules onto a hexanoyl-CoA starter unit—a complex catalytic sequence characterized by inherently slow kinetics that must directly compete with primary fatty acid biosynthesis for the shared intracellular malonyl-CoA pool [40]. Although the light-induced P_psbA2s promoter is a standard choice, its transcriptional output appears insufficient to maintain the high steady-state TKS abundance. Consequently, the superior performance of P_cpc560 highlights that ensuring a high-level, consistent transcriptional output of biosynthetic genes is a prerequisite for driving carbon flux toward this kinetically challenging, non-native polyketide pathway in Synechocystis.

Maximizing carbon flux through this engineered pathway requires global metabolic realignment to complement transcriptional optimization. While the strong P_cpc560 promoter exerts a robust downstream “pull,” upstream efficiency remains constrained by the intensive diversion of fixed carbon into endogenous storage reserves. Eliminating the primary glycogen sink via ΔglgC successfully blocks this major carbon trap to “push” the fixed intracellular carbon pool toward the heterologous OLA biosynthetic pathway [41].

From the perspective of carbon allocation, the 53.79% increase in the final specific yield of SCY04 (1.358 mg·OD₇₃₀⁻¹·L^{− 1}) compared to SCY03 (0.883 mg·OD₇₃₀⁻¹·L^{− 1}) solidifies the hypothesis that glycogen inactivation successful redirects the intracellular carbon pool. In wild-type Synechocystis, glycogen synthesis acts as a massive alternative sink that competes directly with central metabolism for glucose-1-phosphate, draining precursors that would otherwise flow into glycolysis and the lower tricarboxylic acid (TCA) cycle. By disrupting glgC, this major competing sink was eliminated, effectively expanding the intracellular malonyl-CoA pool and providing abundant building blocks for the OLA pathway. Beyond direct carbon rerouting, the elevated specific yield may be augmented by the unique energetic profile of glycogen-deficient mutants, which frequently exhibit higher intracellular ATP pools than wild-type strains [42]. Given that both the priming of hexanoate by AAE1 and the carboxylation of acetyl-CoA by ACC (Acetyl-CoA Carboxylase) are strictly ATP-dependent, this elevated energy charge directly enhances the thermodynamic driving force and catalytic throughput of both upstream pathway nodes.

Another key finding is that the removal of the glycogen synthesis pathway did not inflict a severe physiological penalty on the cyanobacterial chassis (Fig. 6). Glycogen has long been regarded as an indispensable metabolic and energy buffer in cyanobacteria. Under conditions of imbalanced nutrient and energy availability, the amount of newly fixed carbon compounds may exceed both growth requirements and cellular storage capacity, leading to the release of these metabolites into the culture medium [43]. The final biomass density of SCY04 (OD₇₃₀ = 1.789) showed no statistically significant reduction compared to SCY03 (OD₇₃₀ = 1.937), indicating the absence of a severe growth penalty. Under continuous illumination and nutrient-replete conditions, the engineered strains are spared from environmental fluctuations, minimizing their physiological reliance on glycogen as an emergency energy buffer [44]. Furthermore, any potential metabolic congestion or intermediate accumulation caused by the glgC block might be relieved. The heterologous OLA pathway acted as an artificial metabolic valve to drain excess acetyl-CoA and malonyl-CoA pools, while native overflow pathways simultaneously funneled surplus fixed carbon into secondary metabolites such as glutamate [45]. Consequently, this multi-channel carbon redistribution suggests that intracellular carbon flux might be transitionally redirected, potentially preventing the metabolic stagnation typically expected in glycogen-deficient mutants under standard conditions.

The metabolic performance of strain SCY05 underscores the importance of coordinating carbon sink disruption with the reinforcement of key node fluxes in central metabolism. Although blocking glycogen biosynthesis via glgC deletion creates a theoretical surplus of fixed carbon, previous studies have demonstrated that this strategy often leads to an accumulation of upstream phosphorylated intermediates and a kinetic bottleneck in the Calvin-Benson-Bassham (CBB) cycle due to the loss of a major carbon and energetic sink [41]. In this study, the heterologous expression of maeB at the glgC locus effectively mitigated these limitations, resulting in simultaneous increases in both biomass density (OD₇₃₀ = 2.410) and specific OLA yield (1.722 mg·OD₇₃₀⁻¹·L^{− 1}). This synergistic improvement can be rationalized by the biochemical functionality of MaeB, which catalyzes the oxidative decarboxylation of malate to pyruvate with the simultaneous reduction of NADP⁺ to NADPH. The native TCA cycle in Synechocystis operates as a non-cyclic, low-flux branched pathway during photoautotrophic growth [46], primarily serving to generate biosynthetic precursors rather than driving oxidative phosphorylation [13]. Although the host possesses an endogenous malic enzyme, its basal expression cannot accommodate large-scale carbon realignments. Targeted integration of the heterologous maeB significantly amplifies this native anaplerotic node, accelerating flux from malate to pyruvate and generating NADPH to alleviate oxidative stress while sustaining robust central metabolism [47]. The stoichiometry of the MaeB reaction generates NADPH, which, while not directly consumed by the non-reducing Type III polyketide synthase, plays a pivotal role in maintaining the cellular pool of reducing equivalents and alleviating the photo-oxidative stress associated with glycogen deficiency [13]. Consequently, the enhancement of this pathway in strain SCY05 ensures a more dynamic and balanced turnover of central carbon intermediates, thereby accelerating central metabolism and promoting biomass accumulation.

The availability of inorganic carbon (Ci) is a critical determinant governing photosynthetic efficiency and biomass accumulation in cyanobacterial cell factories [48], and introducing the OLA pathway into cyanobacteria aims to leverage their photosynthetic carbon-fixation capacity for high-value compound production. In this study, continuous aeration with 5% CO₂ drastically accelerated the growth rate and extended the growth period of the engineered strain SCY05 (Fig. 7). Elevating the CO₂ concentration facilitates intracellular Ci enrichment, which effectively suppresses the competitive oxygenase activity of RuBisCO and minimizes energy-wasteful photorespiration [49]. This enhances photosynthetic carbon fixation through the Calvin-Benson-Bassham (CBB) cycle, thereby promoting accelerated cell growth and substantial biomass accumulation [50]. Furthermore, the prolonged time required to reach the stationary phase demonstrates that a sustained carbon supply preserves the physiological vigor of the photosynthetic apparatus, enabling long-term autotrophic cultivation. A distinct kinetic mismatch between biomass accumulation and OLA synthesis emerged under HC conditions, where the total specific OLA yield dropped significantly. This phenomenon reflects a profound source-sink imbalance and carbon flux overflow. Although the ΔglgC and maeB modifications successfully closed the glycogen sink, the massive influx of photosynthetically fixed carbon under high CO₂ saturation is inevitably partitioned into primary metabolic pathways to sustain rapid cellular replication. This metabolic diversion is seemingly governed by energetic and kinetic constraints during rapid growth phases. Mechanistically, rapid macromolecular polymerization for cell division creates a massive ATP drain that deprives the OLA pathway of required energy equivalents. Concurrently, rapid cell proliferation upregulates endogenous FabD (malonyl-CoA: ACP transacylase) to meet the intense demand for membrane lipid synthesis, effectively outcompeting TKS for the intracellular malonyl-CoA pool and starving the non-native plant-drived C. sativa polyketide pathway. In summary, although redirecting carbon flux toward malonyl-CoA successfully initiated cannabinoid precursor production, the high-rate cell proliferation shifted the metabolic equilibrium away from target chemical synthesis.

Conclusions

As a proof of concept, this study demonstrates the viability of Synechocystis sp. PCC 6803 as a photosynthetic chassis for the biosynthesis of the cannabinoid precursor olivetolic acid (OLA). By heterologously expressing the plant-derived Cannabis sativa OLA biosynthetic pathway (tetraketide synthase, olivetolic acid cyclase, and acyl-activating enzyme) and utilizing exogenous sodium hexanoate supplementation, we successfully achieved production of this compound. The specific yield was further optimized by disrupting the primary glycogen storage sink and overexpressing the anaplerotic malic enzyme gene to redirect central carbon flux toward polyketide synthesis. Ultimately, this sustainable cyanobacterial platform offers a highly promising, environmentally friendly alternative to traditional industrial C. sativa, cultivation.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

E-Bin Gao, Junhua Wu designed the experiment.E-Bin Gao, Wenrui Ji and Yangjie Zhu performed the experiments and collected the data.E-Bin Gao, Wenrui Ji, and Yangjie Zhu analyzed the data and drafted the article.E-Bin Gao and Wenrui Ji critically revised the article.All authors approved the final draft for submission.

Funding

This work was supported by the National Natural Science Foundation of China (31200019), Open Fund of Key Laboratory of Tropical Marine Bio-resources and Ecology, Chinese Academy of Sciences (LMB131001), Ningbo Clinical Research Center for Children’s Health and Diseases (2019A21002), Ningbo Top Medical and Health Research Program (No. 2022020405), Ningbo Leading Medical &Health Discipline (No.2026-A34); Ningbo Municipal Bureau of Science and Technology Project (No.2025Z153).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding authors on reasonable request.

Declarations

Ethics approval and consent to participate

No applicable.

Consent for publication

All authors have participated in the review study and manuscript preparation.

Competing interests

The authors declare no competing interests.