Integrating artificial with natural cells to translate chemical messages that direct E. coli behaviour

Abstract

Previous efforts to control cellular behaviour have largely relied upon various forms of genetic engineering. Once the genetic content of a living cell is modified, the behaviour of that cell typically changes as well. However, other methods of cellular control are possible. All cells sense and respond to their environment. Therefore, artificial, non-living cellular mimics could be engineered to activate or repress already existing natural sensory pathways of living cells through chemical communication. Here we describe the construction of such a system. The artificial cells expand the senses of Escherichia coli by translating a chemical message that E. coli cannot sense on its own to a molecule that activates a natural cellular response. This methodology could open new opportunities in engineering cellular behaviour without exploiting genetically modified organisms.

Figure 1. Artificial cells translate chemical signals for E. coli.

(a) In the absence of artificial cells (circles), E. coli (oblong) cannot sense theophylline. (b) Artificial cells can be engineered to detect theophylline and in response release IPTG, a chemical signal that induces a response in E. coli.

Figure 2. In vitro characterization of the theophylline-sensing device and αHL.

(a) The cell-free expression of αHL–GFP behind a theophylline riboswitch gives rise to similar levels of fluorescence both in the presence (+theo) and absence (−theo) of theophylline at 37 °C. (b) The removal of the theophylline riboswitch and thus, the RBS preceding the start codon of αHL–GFP shows production of a fluorescent protein product when incubated with transcription–translation machinery (−RBS). The removal of a putative internal RBS within the αHL coding portion of the fusion construct significantly decreases the production of the fluorescent protein product (−RBS mutant). (c) The activity of the theophylline-sensing device is observable by fluorescence when an internal RBS is removed. The top and middle curves are the in vitro expression of αHL–GFP behind the theophylline riboswitch in the presence (+theo) and absence of theophylline (−theo), respectively. Background fluorescent protein production is shown with the same construct lacking the theophylline riboswitch (−RBS mutant) used in b. (d) The cell-free expression of theophylline riboswitch-controlled αHL-degraded red blood cells (RBCs) in the presence (+theo) but not the absence of theophylline (−theo). Control reactions include the expression of an αHL construct lacking the theophylline riboswitch (αHL) and RBCs alone (negative control). RBC degradation was monitored by attenuance at 22 °C. The exploited constructs were SP011A for panel A, SP002A and AS014A for panel B, RL069A and AS014A for c, and RL067A and JF001A for d (Supplementary Table 1). Data are averages of three independent reactions.

Figure 3. The artificial translator cells are functional.

(a) Artificial cells can induce the expression of a plasmid encoded gene within E. coli in response to a molecule that E. coli cannot naturally sense. BL21(DE3) pLysS carrying a plasmid encoding GFP behind a lac operator sequence was incubated with the following components at 37 °C for 3 h: theophylline (theo), artificial cells (AC), artificial cells plus theophylline (AC+theo), IPTG encapsulated inside of vesicles (encapsulated IPTG), and unencapsulated IPTG (IPTG). E. coli fluorescence was quantified by flow cytometry. The reported averages and s.e.m. were calculated from three separate reactions run on three different days from independently assembled artificial cells. (b) A histogram of a subset of the FACS data used in panel a shows a clear shift in the E. coli population in the presence of artificial cells plus theophylline. (c) Artificial cells can induce the expression of chromosomally encoded genes of E. coli. After 4 h of incubation of artificial cells with E. coli at 37 °C, the messenger RNA encoding lacZ, lacY and lacA was quantified by RT–qPCR. Data are reported as averages of three measurements and error bars represent s.e.m.