Reliable measurement of E. coli single cell fluorescence distribution using a standard microscope set-up

Abstract

Background: Quantifying gene expression at single cell level is fundamental for the complete characterization of synthetic gene circuits, due to the significant impact of noise and inter-cellular variability on the system's functionality. Commercial set-ups that allow the acquisition of fluorescent signal at single cell level (flow cytometers or quantitative microscopes) are expensive apparatuses that are hardly affordable by small laboratories.

Methods: A protocol that makes a standard optical microscope able to acquire quantitative, single cell, fluorescent data from a bacterial population transformed with synthetic gene circuitry is presented. Single cell fluorescence values, acquired with a microscope set-up and processed with custom-made software, are compared with results that were obtained with a flow cytometer in a bacterial population transformed with the same gene circuitry.

Results: The high correlation between data from the two experimental set-ups, with a correlation coefficient computed over the tested dynamic range > 0.99, proves that a standard optical microscope- when coupled with appropriate software for image processing- might be used for quantitative single-cell fluorescence measurements. The calibration of the set-up, together with its validation, is described.

Conclusions: The experimental protocol described in this paper makes quantitative measurement of single cell fluorescence accessible to laboratories equipped with standard optical microscope set-ups. Our method allows for an affordable measurement/quantification of intercellular variability, whose better understanding of this phenomenon will improve our comprehension of cellular behaviors and the design of synthetic gene circuits. All the required software is freely available to the synthetic biology community (MUSIQ Microscope flUorescence SIngle cell Quantification).

Keywords: Fluorescence microscopy; Phenotypic noise; Single-cell fluorescence; Synthetic biology.

Fig. 1

Flowchart representing the image analysis pipeline implemented in our algorithm. The first section of the software (yellow box) is responsible for segmenting the images and extracting the fluorescence intensity emitted by each cell. Beside the segmentation routine, it includes the pre-elaboration phase and its input are the raw images (in 8-bit RGB format). The outputs of this section are i) a pdf file, containing the images at different stages of elaboration, ii) a text file containing each image cell density, computed as number of segmented cells divided by the volume used to prepare the slide and iii) another text file, in which the fluorescence intensities emitted by each segmented cell are reported. This last text file is the input of the second section of the algorithm that is responsible for the analysis of the fluorescence data and for the production of the outputs of our software. It mainly consists in the computation of the statistical modes of the fluorescence distribution (average, standard deviation, CV) and their graphical representation. However it also includes a data analysis step in which the fluorescence intensities are normalized and the cardinality of the populations is equalized

Fig. 2

Schematic representation of the synthetic gene circuit used in this work. The expression of the green fluorescent protein (GFP) can be modulated transcriptionally, through the addition of IPTG, due to the presence of the Lac Operator O₁. The circuit was cloned in a pSC101 low copy number plasmid and transformed in E. coli cells of the strain TOP 10 F’ overexpressing the Lac repressor

Fig. 3

Camera Response Function calibration experiment. In a is reported one of the sets of brightfield images of fixed eukaryotic cells used during this preliminary test. The same field of view was captured at different exposure times (between 1 and 4 ms). The third degree polynomial function representing the CRF (blue line) is plotted together with the data on which it was fitted (red, green and yellow dots in b). The different colors identify the pixels belonging to each image pair

Fig. 4

Exemplification of the main steps of the segmentation algorithm. The leftmost image results from the pre-elaboration phase. The application of the zero-crossing edge detection method leads to the central image, in which only the outline of the cells are visible. The rightmost image represents the result of the segmentation procedure, in which all the pixel belonging to a cell, and only those pixels, are different from 0

Fig. 5

Study of the correlation between the data acquired with the microscope set-up and those evaluated with a flow cytometer. The former dataset was divided in groups of equal cardinality and without common elements, and then the Pearson’s correlation coefficient (R²) between each sub-population and the entire flow cytometer dataset was calculated. In a the fluorescence data are considered while the CV² data are shown in (b)

Fig. 6

Comparison between the fluorescent signal acquired with the microscope setup and the one obtained with the flow cytometer. In a the dose response curve for the tested gene circuit is reported. There is a very good agreement between the results obtained with our set-up (red dots) and those of the flow cytometer (blue dots), with only minor discrepancies at the lower induction levels. In b the accordance between the data acquired with the two instruments is analyzed through a correlation graph. The relation is linear (green line, R ² > 0.99) and the SEs are comparable, demonstrating the ability of our set-up in capturing, at the single cell level, the average fluorescence intensity emitted by a population of bacteria. A quadratic function fits better the relation between the two datasets at the beginning of the dynamic range, likely due to the higher sensitivity of the flow cytometer (MSE = 7.6x 10^-5). The equations of the green line and of the parabola that fit the data points are reported below. Fluo_micro = 1.15 Fluo_cyto -0.16, Fluo_micro = 0.44 Fluo_cyto ² + 0.55 Fluo_cyto + 0.02

Fig. 7

Analysis of the biological variability within an isogenic bacterial population as determined with the flow cytometer and our microscope set-up. In a the dependence of the CV² [computed as (σ/μ)2] on the concentration of IPTG is investigated through a dose response curve. The agreement between the data acquired with the alternative set-ups is very good, with only a slight tendency of the microscope setup (red dots) of underestimating the variable of interest, due to the higher dynamic range of the flow cytometer (blue dots). b The correlation graph supports these considerations, since the relation between the data acquired with the two instruments has a linear trend (green line R ² > 0.99). Only the data point with the highest CV², corresponding to the lowest induction level, deviates from linearity (as presented in Fig. 6b). Again a quadratic function fits better the relation between the two datasets at the rightmost end of the dynamic range, likely due to the higher sensitivity of the flow cytometer (cyan line, MSE = 1.8 x 10^-4). The linear equation and the parabola that best fit the data points are reported below. CV² _Micro = 1.28 CV² _cyto -0.13. CV² _Micro = -0.61(CV² _cyto)² + 1.58 CV² _cyto -0.13