Synthetic Biology Approaches for the Optimization and Improvement of a Live Bacterial Therapeutic for the Treatment of Phenylketonuria (PKU)

Phenylketonuria (PKU) is a human metabolic disease characterized by an inability to degrade phenylalanine (Phe), causing neurotoxicity. A novel therapeutic treatment to consume Phe in the GI tract, Escherichia coli Nissle (EcN) SYNB1618 was engineered to express two independent enzymes, PAL and LAAD, which metabolize Phe into the respective metabolites, trans-cinnamic acid (TCA) and phenylpyruvate (PP). Uptake of Phe is facilitated by a high efficiency transporter, PheP. In a Phase 1 trial, SYNB1618 led to dose-dependent increases in SYNB1618-specific biomarkers in healthy humans.

To further develop our Synthetic Bioticâ„¢ platform, a more potent PKU strain was sought through optimization of whole cell PAL activity. In vitro analyses suggested the rate-limiting step was feedback inhibition of PAL by TCA. Increasing PAL expression did not increase whole cell activity but did increase activity in lysate, while salicylate, an efflux pump inducer, increased whole-cell PAL activity.

Using Zymergen’s proprietary sensor engineering platform, an allosteric transcription factor biosensor controlling GFP expression was developed which responded specifically to TCA in a dose-dependent manner. Biosensor intensity correlated with PAL activity as measured by GFP and TCA production. Utilizing a directed evolution approach, PAL variants were designed for improved activity using phylogenetic analysis combined with computational protein design algorithms. Iterative screening of oligo-based combinatorial libraries resulted in progressive activity improvement over multiple rounds of enrichment. Unique PAL variants with increased activity were identified and characterized. The best-performing variants displayed increases in K_m and V_max.

SYNB1934 was constructed by integrating four copies of a lead PAL variant into the chromosome of EcN modified with key features of SYNB1618. In vitro, SYNB1934 exhibited a 2-fold increase in TCA production rate compared to SYNB1618. In healthy non-human primates, SYNB1934 treatment resulted in increases of strain activity biomarkers in vivo. These results define a strategy for optimizing live bacterial therapeutics.