Opportunities in optical and electrical single-cell technologies to study microbial ecosystems

Abstract

New techniques are revolutionizing single-cell research, allowing us to study microbes at unprecedented scales and in unparalleled depth. This review highlights the state-of-the-art technologies in single-cell analysis in microbial ecology applications, with particular attention to both optical tools, i.e., specialized use of flow cytometry and Raman spectroscopy and emerging electrical techniques. The objectives of this review include showcasing the diversity of single-cell optical approaches for studying microbiological phenomena, highlighting successful applications in understanding microbial systems, discussing emerging techniques, and encouraging the combination of established and novel approaches to address research questions. The review aims to answer key questions such as how single-cell approaches have advanced our understanding of individual and interacting cells, how they have been used to study uncultured microbes, which new analysis tools will become widespread, and how they contribute to our knowledge of ecological interactions.

Keywords: CMOS-MEA; Raman spectroscopy; electrical techniques; flow cytometry; impedance flow cytometry; microbial ecology; optical techniques; single-cell.

Figure 1

Workflow for single-cell analysis. The left column explains different methods used for sample preparation (gray). The two squares in the middle are showing optical (orange) and electrical (blue) methods to observe (green line on the left), manipulate and isolate (purple line) and identify (orange line) the single cells.

Figure 2

Schematic overview of microbial cytometric fingerprinting and its most common uses. Raw flow cytometry data are obtained from the measurement of the sample (left), and are often displayed in two-dimensional density plots. For each cell, scatter and fluorescence can be measured, leading to multi-parameter data for each individual cell. Following, the cytometric space is divided in bins (middle) and the density of cells in each bin is determined. In this schematic, equal size binning in two dimensions is displayed, but alternative binning approaches in multiple dimensions can be considered. Obtained discretized data (data in bins) can be used for further statistical analysis (right). Distribution parameters such as richness, evenness, and diversity can be calculated as well as between diversity [for example non-metric multidimensional scaling (NMDS) or Principal Coordinate Analysis (PCoA)]. Data can also be used to train classification algorithms and regression models.

Figure 3

Schematic representation of a flow cytometer. The fluidic system compromises the sample line with sheath fluid. The laser, dichroic mirrors, and filters make up the optical system, the PMT detectors (FCS, SSC, FL1, FL2, and FL3), and the computer make up the electronic system. Figure adapted from Rubbens and Props (2021).

Figure 4

Conceptual figure on FlowFISH. The main difference between regular staining and FlowFISH is that in the latter case, cells are differently fluorescently labeled according to their taxonomy and thus contain taxonomic information in their fluorescent scattering.

Figure 5

The pipeline followed when performing FISH-flow cytometry fingerprinting of microbial communities.

Figure 6

Conceptual figure of a suggested FACS workflow, showing how it can allow to use phenotypic data as a basis for taxonomic data extraction.

Figure 7

Raw data preprocessing of Raman spectra & data analysis. Left: First, the spectra are baseline corrected and normalized. Smoothing and alignment steps can be included. However, smoothing can erase potentially relevant information and should be carefully considered. Similarly, alignment can produce faulty spectra by displacing the signal and thus needs to be used with care. Right: Information that can be obtained with single-cell Raman spectra of cells: (A) The spectrum of individual cells can be plotted using clustering and/or dimensionality reduction techniques. (B) The peaks of the Raman spectra correspond to a different metabolite or a combination of metabolites, called here components (x). The intensity of the signal of each component can be normalized by the sum of all intensities, and this information can be then used in the Hill equation. The order of diversity (q) can be 0, 1 or 2, meaning that the richness, evenness or both richness and evenness are taken into consideration in the metric. (C) The information from the spectral peaks correspond to one or multiple molecules, and can be used (semi)quantitatively.

Figure 8

Schematic representation of IFC experiment and data acquisition. (A) An AC voltage is applied to the top electrodes in the microfluidic channel at a high and low frequency. The electric field around a cell between electrodes illustrates the frequency dependency. (B) A differential current is measured when a cell passes between the electrode pairs. (C) From this measurement, the impedance at the two measured frequencies Z(f_high), Z(f_low) of each cell is extracted and presented on a scatter plot where gating of cell populations is possible.

Figure 9

Illustration of the principle of a CMOS MEA device. A purpose-designed Complementary Metal-Oxide-Semiconductor (CMOS) integrated circuit (IC) is post-processed to encompass an array of microelectrodes (MEA) on its top surface. Each electrode is connected to the inner circuitry of the chip. The biological sample is placed on the electrode surface for fine-grained electrochemical characterization.