A dynamic and multilocus metabolic regulation strategy using quorum-sensing-controlled bacterial small RNA

This in vitro study (2021) developed a novel method of transcriptional gene repression within E. coli that increased the yield of psilocybin biosynthesis by 302.9% without affecting cell growth.

Abstract

“Introduction: Metabolic regulation strategies have been developed to redirect metabolic fluxes to production pathways. However, it is difficult to screen out target genes that, when repressed, improve yield without affecting cell growth.

Methods: Here, we report a strategy using a quorum-sensing system to control small RNA transcription, allowing cell-density-dependent repression of target genes. This strategy is shown with convenient operation, dynamic repression, and availability for simultaneous regulation of multiple genes.

Results/Discussion: The parameters Ai, Am, and RA (3-oxohexanoyl-homoserine lactone [AHL] concentrations at which half of the maximum repression and the maximum repression were reached and value of the maximum repression when AHL was added manually, respectively) are defined and introduced to characterize repression curves, and the variant LuxRI58N is identified as the most suitable tuning factor for shake flask culture. Moreover, it is shown that dynamic overexpression of the Hfq chaperone is the key to combinatorial repression without disruptions on cell growth. To show a broad applicability, the production titers of pinene, pentalenene, and psilocybin are improved by 365.3%, 79.5%, and 302.9%, respectively, by applying combinatorial dynamic repression.”

Authors: Shao-Heng Bao, Hui Jiang, Ling-Yun Zhu, Ge Yao, Peng-Gang Han, Xiu-Kun Wan, Kang Wang, Tian-Yu Song, Chang-Jun Liu, Shan Wang, Zhe-Yang Zhang, Dong-Yi Zhang & Er Meng

Notes

The demand for psilocybin is on the rise, yet the cost of pharmaceutical-grade psilocybin is estimated at around $2,000 USD per gram (Fricke et al., 2019) while researchers might pay up to $7,000-$10,000 per gram. This raises the incentive to develop a more cost-effective means of producing the substance. And while the traditional route of chemical synthesis is being pursued most vehemently by COMPASS pathways and the Usona Institute, a number of new companies, such as Psybio Therapeutics, Psygen, Octarine Bio, and CB Therapeutics, are exploring the alternative route of biosynthesis. This involves genetically altering the metabolic networks of yeast or bacteria to convert sugar molecules into the desired compound.

A previous paper that described the first successful attempt of producing psilocybin on the gram scale with the bacteria E. coli, yielded a 32-fold improvement over earlier techniques. But the costs per serving (before any other costs) would then be $40-$50 at 20-30mg per 70kg. This was still higher than the costs of commercially available magic mushrooms or truffles (at $/€10-15). However, the current paper applied a novel gene repression method, which enabled them to selectively inhibit metabolic enzymes that interfere with psilocybin biosynthesis. This effectively improved the yield by an additional 302.9%.

What is novel about this technique?

• The previous attempt to synthesize psilocybin still required expensive precursor materials, given the difficulty of facilitating trypthophan synthesis within E. coli. To solve this problem, they used a more efficient mutation of the tryptophan synthase enzyme and inserted it into E. coli
• The efficiency of psilocybin synthesis was previously limited by the efficiency of methyltranferase enzyme (psiM), so the investigators inserted additional RNA strains into E. coli to overexpress this enzyme
• The investigators identified native pathways in E. coli that interfere with the biosynthesis of psilocybin, for instance transporter enzymes that diminish the cellular concentration of its unfinished precursors. The selective repression of this drug transport enzyme alone increased the production of psilocybin by 107.2%
• The investigators developed a novel RNA-based method that selectively repressess certain genes in a cell-density-dependent manner. Interfering with the genetic make up of the organism itself may otherwise severely impair its natural growth cycle of the host organism and lower its productivity. However, this technique repressess genes on a transcriptional (RNA) level and they were able to repress interference with psilocybin synthesis without impeding on the cellular growth of E. coli

The combined result of these innovations resulted in a 302.9% improvement in their biosynthesis of psilocybin. In reference to previous cost efficiency estimates, this method puts the cost of psilocybin at around $13 – $16.50 at 20-30mg per 70kg. Hence the current study significantly closes the cost gap between psilocybin biosynthesis and growing psilocybin mushrooms/truffles. And although it raises the competitive potential of the biosynthesis method, it is still questionable whether it can match the synthetic method developed by the non-profit Usona Institute that can produce psilocybin even on a large (1kg) scale.

Related Articles:

• Enzymatic synthesis of psilocybin
• Production Options for Psilocybin: Making of the Magic
• In vivo production of psilocybin in E. coli
• Direct Phosphorylation of Psilocin Enables Optimized cGMP Kilogram-Scale Manufacture of Psilocybin

Summary

INTRODUCTION

Novel regulation strategies to understand and manipulate microbial metabolism have been developed to produce natural products, accumulate desired intermediates, and reduce the production of toxic byproducts.

Bacterial small RNA (sRNA) can be used to activate or repress genes at the posttranscriptional level. It can be used to control metabolic flux by downregulating competing branch pathways.

Recent advances in sRNA-based regulation have been applied in cyanobacterium, Helicobacter pylori, Caulobacter crescentus, and Vibrio cholerae to achieve diverse goals, including delayed repression of gene expressions at steady-state phase and combined activating RNA and antisense RNA control of gene expressions.

Repressive sRNA regulation systems may cause slow cell growth if strong repression occurs on essential genes at early stages of growth.

The use of stress-driven promoters and negative autoregulation systems to relieve stress during lignocellulosic ethanol fermentation have been successful in increasing ethanol productivity and naringenin production. Bacillus subtilis uses quorum sensing to balance growth and production by automatically redirecting metabolic flux toward the production pathway when high cell density is reached. The QS mechanism controls cell density-dependent bacterial behaviors, and is used to balance competing essential pathways without weakening cellular conditions and productivity. Several substrate-responsive regulation strategies have been developed to achieve autonomous regulation of pathways or selective utilization of nutrients.

The Lux QS system from Vibrio fischeri has been widely applied in genetic circuit design and autonomous regulation of product synthesis. Here, we combined the Lux system with bacterial repressive sRNA to achieve fast and dynamic control of metabolic flux without modifying the genome.

Pinene, pentalenene, and psilocybin were synthesized using this strategy, and the production titer of pinene was improved by enzymatic fusion, directed evolution of pinene synthase, eliminating plasmid recombination, and establishing an E. coli-E. coli modular co-culture system. Psilocybin, an indole-alkaloid drug, can be synthesized in E. coli and S. cerevisiae by adding 4-hydroxyindole as a starting substrate. The highest titers were obtained by using combinatorial dynamic repression of multiple key genes.

Characterization of dynamic repression

The QS-controlled dynamic sRNA regulation mechanism involves the transcription factor LuxR, the signal molecule 3-oxohexanoyl-homoserine lactone (AHL), and the Hfq chaperone. AHL accumulates during cultivation, resulting in an increasing level of transcription and the degradation of target mRNA.

We placed bacterialMicC sRNA between thePlux promoter and the BBa_B0025 terminator and inserted the Plux-sRNA units into the pUC57 plasmid with a high-copy replication origin pMB1. The sRNA targeted superfolder green fluorescent protein (sfGFP) and the protein degradation tag ssrA was added at the C terminus.

We tested several LuxR variants for AHL sensitivity. The most sensitive LuxR variant was LuxR1, which had a threshold of 150 nM and a maximum threshold of 500 nM, and the least sensitive LuxR variant was LuxR2, which had a threshold of 300 nM and a maximum threshold of 800 nM.

The results revealed growth-stage-dependent repression throughout cultivation, with the maximum repression occurring at the stationary phase. A more sensitive LuxR variant enabled earlier and higher repression at the stationary state, while the maximum repression of both variants remained at approximately 80%.

Improving pinene titer through dynamic pathway regulation

We used QS-controlled sRNA regulation to enhance the efficiency of the pinene-producing pathway by redirecting metabolic fluxes. However, repressive sRNA regulation may cause a delayed logarithmic phase and disrupt cell states at the stationary phase, resulting in a reduction in production titer.

The pinene-producing pathway is heterologous, and its intermediate substrates cannot be consumed by the host. Therefore, the GPP-consuming branch pathway was chosen as our first testing subject, and a series of Plux-sRNA units were designed to target genes associated with the vitamin K synthetic pathway. Repression of the expression levels of octaprenyl diphosphate synthase (OPPS) and farnesyl diphosphate synthase (FPPS) could increase pinene titers by causing the accumulation of farnesyl diphosphate (FPP) and GPP, respectively.

Plux-sRNA and PR-sRNA units designed to repress the expression of OPPS were compared to evaluate dynamic repression. The results showed that both LuxR1- and LuxR2-tuned sRNA led to attenuated levels of OPPS with cell growth.

The LuxR2 variant was determined to control repression in shake flask fermentation, and the LuxR3 variant slightly improved pinene titers. The LuxR3 variant imposed no metabolic burden upon the host, and the LuxR3 variant had no effect on the expression of OPPS at steady state.

Characterization of combinatorial QS-sRNA repression

The repression of ispB, a single gene, redirected flux to the synthesis of pinene. However, combinatorial sRNA repression of multiple genes is expected to be more effective.

We hybridized eight different sRNA structures with the MicC structure and substituted the downstream terminator to avoid unwanted plasmid recombination. This method successfully avoided recombination among eight sRNA units. To test the feasibility and efficiency of the strategy, an increasing number of non-target Plux-sRNA units were added to co-work with an anti-GFP Plux-sRNA unit. The repression of GFP gradually decreased as more non-target Plux-sRNA units were added, and the repression remained stable at approximately 85%.

We applied combinatorial repression of ispB, sthA, and serA to strain BS1101, and found that the logarithmic phase was delayed as the applied sRNA units increased. By removing the native hfq gene and introducing double aHfq into the sRNA plasmid, the additive effect on cell growth was effectively overcome.

Application of combinatorial repression in pinene production

We used single-gene repression to screen out gene candidates of which repression can improve the pinene titers, and then applied combinatorial repression of candidates to achieve optimal improvement.

Based on the metabolic-flux analysis of the mevalonate pathway in E. coli, 48 genomic genes upstream of the mevalonate pathway were selected for screening. The repression of most of these genes resulted in an increase in pinene titers, but the increase was disproportionate to the increase in ACCoA levels.

To further indicate the essentiality of NADPH for pinene synthesis, we screened key genes related to the source of NADPH. Repressing pntA and sthA resulted in a significant increase in pinene titers, whereas repressing pntA failed to improve pinene titers.

17 genes were repressed in combination with sthA and ispB, which increased the pinene titer to 135.8 mg/L. Additional repression of serA, fbp, gcd, poxB, and ilvD further improved the pinene titer to 165.1 mg/L, a 365.3% improvement compared with the non-sRNA control.

Pentalenene production optimization

Plasmid pAC-6409PENTS was transformed into strain BS1101 and used to produce pentalenene, a sesquiterpene with a similar production pathway to pinene. The genes for vitamin K synthesis were screened by using single-gene repression. Repressing the gene menA increased the production titer of pentalenene by 46%, while the production titers were not significantly changed when the genes ispA and ispB were repressed. The genes pntA, pgi, and sthA were further screened for pentalenene production, and their effects were consistent with pinene production.

Improving biosynthetic titers of psilocybin

Psilocybin is synthesized by four heterologous enzymes using tryptophan and S-adenosyl-L-methionine as substrates. A mutant tryptophan synthase b-chain TrpBM149T, N171D derived from E. coli was used to facilitate the synthesis of tryptophan. Analysis of the native metabolic network of E. coli indicated that by repressing the expression level of the drug-export enzyme encoded by acrF, the production titer of psilocybin significantly increased by 107.2%, from 6.9 to 14.3 mg/L.

Repressing the expression of mtn led to a significant reduction of psilocybin titer, and selecting the genes acrF, tynA, and speD for combinatorial dynamic repression led to the highest psilocybin titer of 27.7 mg/L.

DISCUSSION

Before optimizing metabolic networks, it is necessary to identify competing branch pathways from the extremely massive genome. A plasmid system consisting of bio-bricks can effectively tune the expression levels of genomic genes and can be used to stably and specifically target genes in the entire genome.

The QS-based sRNA regulation system can be adopted in high-throughput gene screening for metabolic studies. It is expected that more genes can be repressed simultaneously without interrupting cell growth, and different variants of the transcription factor LuxR can be used to enable effective repression at different growth stages.

Endogenous bacterial metabolites are constrained at low levels. The dynamic sRNA regulation strategy improves pinene synthesis by redirecting metabolic fluxes. After two rounds of screening, increasing levels of NADPH, ACCoA, and GPP resulted in a 365.3% improvement in pinene titer, and a 6.9 mg/L improvement in psilocybin titer by repressing an indole-efflux enzyme and tryptophan and SAM-consuming pathways.

STAR+METHODS

Detailed methods are provided in the online version of this paper, including a key resources table, resource availability, experimental model and subject details, and quantitative and statistical analysis.

Data and code availability

Plasmids pTD103luxI_sfGFP and pJBEI-6409 were purchased from Addgene, and plasmid pAC-6409PS for pinene production was constructed by replacing the gene for limonene synthase with the gene for pinene synthase by using the OEPR cloning technique.

The psilocybin-producing plasmid was constructed by deleting the p15A origin of replication and the chloramphenicol resistance operon from plasmid pJBEI-6409 and inserting the genes for sfGFP and the constitutive promoter Ptrc* upstream of sfGFP by using the OPER cloning method.

The hfq fragment was amplified, then inserted into the 50 end of the gene for sfGFP. The Plux promoter and the BCD RBS series were then inserted upstream of the hfq fragment.

Western blot experiments

Colonies were grown overnight at 37 C in LB medium, and then harvested and lysed in PBS. Protein concentrations were measured by BCA Protein Assay Kit, and transferred onto a PVDF transfer membrane and blocked by 5% nonfat milk powder dissolved in PBST buffer. Primary antibodies against OPPS, HA tag, and GroEL were hybridized on a membrane overnight at 4 C. The expression profiles of GroEL were obtained by a C-DiGit Blot Scanner.

Quantitative assays of products and key intermediates

Fermentation cultures were collected, centrifuged, and the supernatant was diluted in ethyl acetate and spiked with limonene and b-caryophyllene for pinene and pentalenene quantification, respectively. The samples were analyzed by gas chromatography-mass spectrometry (GC-MS) using a 7890A GC and a 5975C MS detector, and the oven temperature was set to 300 C with a flow rate of 1 mL/min. Psilocybin was quantified using the area of the peaks at retention times of 3.0 min, 3.6 min and 8.3 min, respectively.

Harvested fermentation cultures were centrifuged at 12,000 rpm for 10 min, the supernatants were filtered and transferred to new vias, and a HPLC-HRMS was used to determine prepared samples. The data were acquired using Xcalibur 4.1 software and a spray voltage of 3.5 kV (+). The Q Exactive detector was operated in full scan and data-dependent MS2 modes. The resolution was set at 70,000, the AGC target was 2 3 105 ions capacity, and the maximum injection time was 50 ms.

Mathematical models QS tuned sRNA level

The mathematical models for the simulations below were all performed in MATLAB. The growth rate of the total biomass (N) is related to the level of LuxR (R) and LuxR-AHL dimer complex (RA) and the level of LuxI produced AHL (A). The differential equations for the target mRNA, Hfq protein and target protein are quoted from Sagawa et al. supplemental methods (Sagawa et al., 2015), and the differential equations for the Plux controlled sRNA are set to estimate the variation of substrates associated with the pinene-producing pathway and the vitamin K synthesis pathway.

The repression of UPPS has negligible effect on the production of pinene, so vUPPS = 0. The reaction rates of induced GPPS and PS are correlated with the levels of corresponding enzymes.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data analysis was performed using GraphPad Prism 6.0. A two-tailed unpaired t test was used to compare the two groups.