Heterologous Expression of Pseudomonas putida Methyl-Accepting Chemotaxis Proteins Yields Escherichia coli Cells Chemotactic to Aromatic Compounds

Abstract

Escherichia coli, commonly used in chemotaxis studies, is attracted mostly by amino acids, sugars, and peptides. We envisioned modifying the chemotaxis specificity of E. coli by expressing heterologous chemoreceptors from Pseudomonas putida enabling attraction either to toluene or benzoate. The mcpT gene encoding the type 40-helical bundle (40H) methyl-accepting chemoreceptor for toluene from Pseudomonas putida MT53 and the pcaY gene for the type 40H receptor for benzoate and related molecules from P. putida F1 were expressed from the trg promoter on a plasmid in motile wild-type E. coli MG1655. E. coli cells expressing McpT accumulated in chemoattraction assays to sources with 60 to 200 μM toluene, although less strongly than the response to 100 μM serine, but statistically significantly stronger than that to sources without any added attractant. An McpT-mCherry fusion protein was detectably expressed in E. coli and yielded weak but distinguishable membranes and polar foci in 1% of cells. E. coli cells expressing PcaY showed weak attraction to 0.1 to 1 mM benzoate, but 50 to 70% of cells localized the PcaY-mCherry fusion to their membrane. We conclude that implementing heterologous receptors in the E. coli chemotaxis network is possible and, upon improvement of the compatibility of the type 40H chemoreceptors, may bear interest for biosensing.IMPORTANCE Bacterial chemotaxis might be harnessed for the development of rapid biosensors, in which chemical availability is deduced from cell accumulation to chemoattractants over time. Chemotaxis of Escherichia coli has been well studied, but the bacterium is not attracted to chemicals of environmental concern, such as aromatic solvents. We show here that heterologous chemoreceptors for aromatic compounds from Pseudomonas putida at least partly functionally complement the E. coli chemotaxis network, yielding cells attracted to toluene or benzoate. Complementation was still inferior to native chemoattractants, like serine, but our study demonstrates the potential for obtaining selective sensing for aromatic compounds in E. coli.

Keywords: biosensing; chemotaxis.

FIG 1

Chemotaxis response of E. coli MG1655 toward various common attractants in agarose plug assays. (A) Average cell accumulation of E. coli MG1655 as a function of distance from the source edge with 100 μM serine, aspartate, methylaspartate, ribose, galactose, or a no-attractant control. Ribbon traces show the average of triplicates (central line) ± one standard deviation (bordering lines). (B) Same as in panel A but with source concentration of 10 μM of the different attractants. (C) Average gray values across the three zones closest to the source edge (7.5-μm width) summarized for the different attractants and concentrations. Asterisks indicate significantly different values at a P value of <0.0001 in one-way ANOVA, followed by Tukey post hoc multiple-comparison test. Dashed boxes indicate the distance zones used for calculating cell accumulation.

FIG 2

Chemotaxis of E. coli expressing mcpT of P. putida toward toluene. (A) Cropped 100-fold magnification phase-contrast images of one agarose plug replicate experiment with sources containing no attractant (Ctl), 60 μM toluene (Tol), or 100 μM serine (Ser). Yellow curves represent the measured cell accumulation. Note the agarose sources localized on the left of the images, with the source edge typically resulting in a dark-light band. (B) Average cell accumulation (as image average gray values [AGV]) as a function of distance from the source edge averaged from four biological replicates imaged on both sides of the agarose plug with toluene (6 μM [TOL 6] and 60 μM [TOL 60]), serine (SER; 100 μM), or a no-added-attractant control for E. coli MG1566(pCRO20) expressing the McpT receptor from P. putida MT53 (no A). Ribbon traces show the average of four replicates ± one standard deviation. Inset shows the average gray value across the three zones closest to the source edge (7.5-μm width). Letters indicate significance groups in a one-way ANOVA, followed by post hoc Tukey multiple-comparison test. (C) Same as in panel B but with E. coli MG1655(pSTV) (empty plasmid). (D) Same as in panel B but with E. coli MG1655(pCRO35), which contains a frameshift mutation in mcpT causing a premature translation stop. Dashed boxes indicate the distance zones used for calculating cell accumulation.

FIG 3

Chemotaxis response of PcaY expressing E. coli MG1655. (A) Cell accumulation as a function of distance to an agarose plug with benzoate (1 or 0.1 mM), serine (100 μM), or no attractant of E. coli MG1566(pCRO33) expressing the PcaY receptor for benzoate of P. putida F1. (B) Same as in panel A for E. coli MG1566(pSTV) (empty plasmid). Cell accumulation, ribbon traces, and dashed boxes and inset are the same as described in the legend to Fig. 2. The benzoate source concentration is 1 mM for the data shown in the inset.

FIG 4

E. coli cell accumulation in wells of an in situ chemotaxis microfabricated chip. (A to E) E. coli MG1655 wild type (strain 4498) (A), E. coli MG1655(pCRO20) expressing McpT (strain 5197) (B), E. coli MG1655(pCRO33) expressing PcaY (strain 5447) (C), E. coli MG1655 Δtsr(pCRO33) expressing PcaY (strain 6068) (D), E. coli MG1655 Δtsr(pCRO20) expressing McpT (strain 6085) (E). Bars show average cell accumulation plus standard deviation (SD; error bars) to the indicated chemoattractants measured by absolute flow cytometric counting across 5-fold replicate cavities, normalized to that of cavities filled with motility buffer (MB) alone. Note that panels may be composed of different independent experiments, which are normalized to the respective cell accumulation in MB as a control for every individual chemotaxis assay. SER, serine; BEN, benzoate; TOL, toluene. Concentrations are in micromolar or millimolar, as indicated. Asterisks and daggers denote significantly increased and decreased responses, respectively, compared to motility buffer at P values of <0.05 in pairwise t tests.

FIG 5

Characterization of MCP receptor expression in E. coli by fluorescent protein fusions. (A to F) Phase-contrast (PhC) and mCherry (mCHE) epifluorescence images of E. coli MG1655(pCRO34), expressing the Tsr receptor (TSR) (A), MG1655(pCRO38), expressing a Tsr-mCherry fusion protein (TSR-MCHERRY) (B), MG1655(pCRO36), expressing a fusion protein of McpT and mCherry (MCPT-MCHERRY) (C), MG1655(pCRO37), expressing an mCherry fusion protein but with a frameshift mutation in mcpT coding sequence (MCPT^FS-MCHERRY) (D), MG1655(pCRO20) expressing McpT (MCPT) (E), and MG1655(pCRO33-mCHE), expressing the PcaY-mCherry fusion protein (F). (C and F) Arrows show visible membrane foci of McpT-mCherry and PcaY-mCherry. Images were recorded and autoscaled in ImageJ, saved as 8-bit grayscale for reproduction, opened and cropped to their final size in Adobe Photoshop (version CC2017), and finally saved as .TIF with 300 dpi resolution for display. Numbers in fluorescence images indicate the absolute intensity scaling (minimum to maximum) for reproduction.

FIG 6

Localization and quantification of chemoreceptor-mCherry fluorescent protein fusions in E. coli. (A) Positions of fluorescent foci (black dots) in n individual cells extracted by SuperSegger from image series in the different strains (as indicated), superposed, and plotted on a standardized E. coli cell by a MatLab custom subroutine. px, pixels. (B) Heatmap of fluorescent pixel intensity extracted from 1,000 E. coli cells showing the position of Tsr-mCherry fluorescence normalized to a standardized cell length and width. (C) Average top 10% pixel intensity per cell among n cells as from panels A to H, normalized to the mean fluorescence intensity of all cells of that strain. Error bars show SD of 10 images. Note the different intensity scales between strains expressing McpT derivatives, PcaY-mCherry, and Tsr-mCherry. Letters above bars indicate statistically significantly different categories in ANOVA, followed by Tukey post hoc testing (P < 0.005).