Engineering a Fast-Responding Bacterial Test for Zinc Deficiency

Zinc deficiency is responsible for over one hundred thousand childhood deaths each year, but the high cost and logistical challenges associated with current zinc diagnostic tools prevent adequate surveillance and treatment of zinc deficiency. To be used in nutrition surveillance programs, a zinc diagnostic must be inexpensive, easy to administer, nearly equipment-free, and fast-responding. A bacterial biosensor has the potential to meet all of these requirements. Our group recently developed an E. coli biosensor that produces different visible outputs based on the concentration of zinc in which the cells grow. However, these sensor cells always produce some pigment, which necessitates that they be grown from a very small inoculum so that initial coloration is minimal – thus requiring overnight assay times. To make an easy-to-use, field-deployable assay, we engineered cells that completely repress pigmentation during growth and quickly produce the appropriate pigment upon induction. This enables sample addition to concentrated pre-cultured cells, decreasing assay time from about 24 hours to less than 4 hours.

We first focused on eliminating unwanted pigment production. A sensor that produces pigmented metabolites (instead of more commonly used fluorescent proteins) is advantageous in resource-limited settings, since sensor readout can be interpreted without advanced equipment. However, using pigments as reporters is challenging, primarily because small amounts of unwanted enzyme can produce visible amounts of pigment. To combat unwanted pigment expression, we engineered three commonly used inducible promoter systems—pLac, pBad, and pT7—to be more repressible. Despite orders of magnitude decreases in uninduced protein expression, even the most repressible systems could not fully eliminate production of the red pigment lycopene. Translational modifications proved much more effective. When the original ribosomal binding sites for the lycopene-producing proteins were replaced with weak RBSs, all systems fully repressed lycopene, and upon induction produced visible lycopene within three hours. Supplementation of metabolic precursor pathways further reduced the time required for visible pigmentation to 1.5 hours.

We next incorporated zinc-responsive transcriptional elements into these systems to engineer cells that respond to different zinc concentrations by producing one of multiple pigments when induced. We added operator sites for the repressor Zur to inducible promoters to create hybrid promoters that respond both to an exogenous inducer (IPTG or arabinose) and to zinc. The location and number of operator sites was optimized, and the best-responding hybrid promoters only produce visible pigment in the presence of inducer and in low zinc concentrations. The zinc-responsive activator ZntR and its cognate promoter were then added to the circuit to control production of the protein that converts lycopene to beta-carotene. The resulting cells are colorless during the uninduced growth stage, and upon addition of inducer, they produce violacein (purple), lycopene (red), or beta-carotene (orange) to indicate low, medium, or high zinc concentrations, respectively.

The resulting biosensor is a significant step towards a field-friendly zinc diagnostic tool. More generally, this work demonstrates ways that synthetic biology and metabolic engineering approaches can be used to engineer systems that precisely respond to changes in their environment.