Gene order proves as crucial as enzyme choice for rhamnosylated flavonoid yields in E. coli

Engineered Escherichia coli can now pump out rhamnosylated flavonoids at titers that rival the best microbial flavonoid glycoside systems, and the decisive factor turned out to be not the enzymes themselves but the order in which their supporting genes are strung together. Researchers at a South Korean laboratory systematically reshuffled the four-gene rfb operon that builds the sugar donor dTDP-rhamnose inside E. coli, then measured how each arrangement affected the output of two plant-derived flavonoid rhamnosides. Their winning configuration, a dual-promoter construct named pA-rfbBD-rfbAC, produced 81.9 milligrams per liter of genistein 7-O-rhamnoside and 241.1 milligrams per liter of quercetin 3-O-rhamnoside, outperforming the native bacterial operon that the cell has maintained over evolutionary time.

The study, published in Applied Biological Chemistry, tackles a long-standing measurement problem in metabolic engineering. Intracellular dTDP-rhamnose is technically difficult to quantify directly, so the team instead deployed two Arabidopsis thaliana glycosyltransferases, AtUGT89C1 and AtUGT78D1, as sensitive metabolic reporters. These enzymes attach rhamnose to flavonoid acceptors supplied in the culture medium, and because the acceptors were fed in excess, the rate of glycosylation depended strictly on how much dTDP-rhamnose the engineered rfb pathway supplied. The final flavonoid rhamnoside yield therefore served as a faithful proxy for pathway flux, converting an intractable analytical question into a straightforward HPLC readout.

In E. coli, the rfbB, rfbD, rfbA, and rfbC genes convert glucose 1-phosphate into dTDP-rhamnose through a four-step enzymatic sequence. Genomic inspection of the native cluster revealed telling structural details: the stop codon of rfbB overlaps the start codon of rfbD, indicating tight translational coupling, while only 4 base pairs separate rfbA and rfbC. Between rfbD and rfbA, however, lies a conspicuous 58-base-pair gap, a distance large enough to break translational coupling and potentially house a weak promoter or an mRNA stability element that fine-tunes the expression ratio between the rfbB-rfbD and rfbA-rfbC halves of the pathway.

Armed with that observation, the researchers built four competing constructs. The first, pA-rfbBDAC, preserved the endogenous gene order under a single promoter and a single ribosome binding site. The second, pA-rfbB-rfbD-rfbA-rfbC, kept the same order but gave every gene its own promoter and ribosome binding site, completely fragmenting the operon. The third, pA-rfbDCBA, reversed the gene order relative to the metabolic sequence, an arrangement that certain metabolic engineering principles had suggested could pull flux forward. The fourth, pA-rfbBD-rfbAC, split the pathway into two transcriptional units, each with its own promoter and ribosome binding site, directly mimicking the natural bifurcation implied by the 58-base-pair interval.

When each construct was co-transformed into E. coli BL21(DE3) alongside the reporter plasmids, the dual-promoter design emerged as the clear winner, delivering 77.5 milligrams per liter of genistein 7-O-rhamnoside and 197.8 milligrams per liter of quercetin 3-O-rhamnoside in the initial screen. The native arrangement trailed at 71.1 and 186.6 milligrams per liter respectively, the fully fragmented independent-promoter construct ranked third, and the reversed pA-rfbDCBA configuration performed worst for both products. The authors attribute the poor showing of the reversed order to disruption of natural translational coupling, which likely created a metabolic imbalance or allowed intermediates to pile up inside the cell.

Extended biotransformation runs with the optimized construct pushed the titers even higher. Genistein 7-O-rhamnoside production climbed rapidly during the first 12 hours and peaked at 81.9 milligrams per liter by 24 hours, while quercetin 3-O-rhamnoside rose sharply for 9 hours and then accumulated steadily until 48 hours, finishing at 241.1 milligrams per liter. Cell density remained constant throughout the cultivation, indicating that the redirected metabolic flux did not severely impair growth or trigger cell lysis, an encouraging sign for any future scale-up effort.

The most conceptually striking result concerns where the bottleneck sits. By supplementing each individual rfb gene into the three main plasmid configurations, the team discovered that the rate-limiting step shifts depending on gene arrangement. In the unbalanced reversed construct pA-rfbDCBA, adding an extra copy of rfbC produced the largest gain, lifting genistein 7-O-rhamnoside output from a control of 42.8 milligrams per liter to 57.2 milligrams per liter, with extra rfbB, rfbD, and rfbA each delivering smaller improvements. In the native pA-rfbBDAC configuration, however, the bottleneck moved to rfbB, whose supplementary copy drove the highest titer at 79.6 milligrams per liter.

For the already balanced pA-rfbBD-rfbAC design, single-gene reinforcement yielded only marginal returns. An additional rfbC copy nudged production to 83.4 milligrams per liter, but overexpressing rfbA, rfbB, or rfbD actually lowered output below the 78.5 milligrams per liter control. The researchers interpret this asymmetry as evidence that the dual-promoter construct had reached a state of metabolic harmony, leaving little room for simple overexpression to help. Reinforcing individual genes can partially rescue a poorly organized cluster, they conclude, but strategic rearrangement of the whole cluster matters more for maximizing final titers than amplifying any single bottleneck enzyme.

The mechanistic explanation rests on transcriptional polarity, the well-known tendency of mRNA abundance to decline with distance from the promoter in bacterial operons. In the reversed construct, the initial pathway enzymes rfbB and rfbA sat farthest from the promoter, and because these first steps are highly sensitive to expression levels, the flux stalled at the very beginning of the pathway. Splitting the pathway at the natural 58-base-pair boundary and giving rfbAC its own promoter mitigates this distal attenuation, ensuring the second half of the pathway retains enough enzymatic activity to prevent accumulation of intermediates such as dTDP-4-keto-6-deoxy-D-glucose. The authors argue that the intergenic gap is not genomic noise but an evolved modularity that partitions the pathway into two functionally balanced units.

Beyond the specific case of dTDP-rhamnose, the work offers a template for engineering other nucleotide-sugar pathways, which feed the glycosylation reactions that modify a vast range of plant secondary metabolites, including alkaloids, glucosinolates, phenylpropanoids, and terpenoids. Previous efforts to make flavonoid rhamnosides in microbes relied on introducing plant UDP-rhamnose synthase genes, a mechanism distinct from the native E. coli dTDP-rhamnose route. By demonstrating that the endogenous four-gene pathway can be retooled for high-yield production, and that preserving a modular operon structure beats both complete fragmentation and wholesale reversal, the study underscores that endogenous gene organization is optimized for cellular fitness rather than maximal metabolite output, and that industrial bio-foundry optimization often demands a deliberate departure from natural evolutionary paths.

Subject of Research: How the gene arrangement of the E. coli rfb operon controls metabolic flux toward dTDP-rhamnose and flavonoid O-rhamnoside production

Article Title: Impact of operon architecture on metabolic flux: highlighting gene arrangement as a critical factor for flavonoid O-rhamnoside synthesis

Article References: Jo, J.-H., & Ahn, J.-H. (2026). Impact of operon architecture on metabolic flux: highlighting gene arrangement as a critical factor for flavonoid O-rhamnoside synthesis. Applied Biological Chemistry, 69(1), Article 60. https://doi.org/10.1186/s13765-026-01120-2

Image Credits: AI Generated

DOI: 10.1186/s13765-026-01120-2

Keywords: rfb operon, dTDP-rhamnose, metabolic flux, flavonoid rhamnosides, operon architecture, Escherichia coli, glycosyltransferases, transcriptional polarity, dual-promoter construct, metabolic engineering, genistein, quercetin