Single-molecule and super-resolution imaging of transcription in living bacteria

Abstract

In vivo single-molecule and super-resolution techniques are transforming our ability to study transcription as it takes place in its native environment in living cells. This review will detail the methods for imaging single molecules in cells, and the data-analysis tools which can be used to extract quantitative information on the spatial organization, mobility, and kinetics of the transcription machinery from these experiments. Furthermore, we will highlight studies which have applied these techniques to shed new light on bacterial transcription.

Keywords: Live cells; PALM; Quantitative microscopy; Single-particle tracking; Super-resolution imaging; Transcription.

Fig. 1

The transcription cycle. RNAP associates with a sigma factor before binding to a promoter site. After initial binding, the enzyme opens a bubble in the duplex DNA to form an ‘open complex’. From here, it can initiate transcription; however, on many promoters, the polymerase makes several attempts to start transcribing, generating short abortive RNAs . Once past the ∼10th nucleotide, the RNAP breaks its interactions with promoter DNA and enters into processive synthesis of RNA as an ‘elongation complex’. At some point during elongation, the sigma factor usually dissociates from the core enzyme . Finally, RNAP reaches the end of the gene, and the RNA transcript and the core enzyme dissociate from DNA.

Fig. 2

In vivo single-molecule fluorescence microscopy. A) Schematic of an example microscope setup for single-molecule microscopy. Photoactivation and excitation lasers are coupled into an optical fiber. Light from the fiber output is collimated and focused on the back focal plane of the objective. Translation of the fiber output, collimation and focusing lenses allows for control of the incident angle of the beam at the coverslip. The emission signal is filtered from the excitation light with a polychroic mirror and focused onto an EMCCD camera. Transmission light is provided by an LED above the sample, and autofocus is provided by an infrared LED and a position-sensitive photo detector. B) An example transmission image of a live E. coli cell. C) A single frame of a PALM movie showing the fluorescence image from a single labelled RNA polymerase molecule. D) A super-resolved image of RNAP generated from imaging and localizing all available RNAP molecules over 20,000 frames. E) Trajectories of RNAP; each color corresponds to a single track. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Generating and analyzing super-resolved images of live cells. A) An example field of view with a high density of photoactivated PAFPs. Localizations are identified with a crowded-field algorithm. B) Rapid-acquisition (15 s) PALM images RNAP in live E. coli analyzed with a crowded-field localization algorithm. Comparing slow (top) and fast (bottom) growth conditions, highlights increased clustering of RNAP in fast growth conditions. This can be quantified using a clustering algorithm. C) Super-resolved images of RNAP (red) and DNA (blue) imaged with 3D SIM. D) Pair correlation analysis of RNAP localizations in panel B. Panel A adapted from , panel B-D adapted from . (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Analyzing single-particle tracking PALM data. A) Plotting the mean squared displacement of the sptPALM trajectories against time lag can provide information about the mobility of the labelled protein, and establish if motion is Brownian (where MSD increases linearly with increasing time lag) or sub-diffusive. B) Cumulative distribution of the squared displacements. This distribution can be fitted with Eqs. (2), (3), to extract information about the mobility of the proteins and the number of diffusive species. C) Distribution of apparent diffusion coefficients (D_app) calculated for each single-molecule trajectory. A threshold can be used to sort individual trajectories based on their D_app value, as shown in the example cell with slow trajectories colored red and fast trajectories colored blue (right). D) Examples of long trajectories (ten or more localizations) classified according to their D_app transitions: a fast diffusing molecule, with a high average D_app value over the whole trajectory (blue), a slow-moving molecule, with a low average D_app value (red), and a molecule undergoing transition from fast (high D_app) to slow (low D_app) (purple). E) The D_app distribution for DNA polymerase 1 treated with a DNA damaging agent to recruit molecules to DNA. The distribution shows two clearly resolvable peaks, which can be fitted with a two-species model (using Eq. (6)) to extract fractions of molecules in the low-mobility DNA-bound state, and the mobile state. F) The distribution of RNAP D_app values can also be fitted with a two-species model. Treatment with rifampicin blocks transcription, causing a large drop in the fraction of DNA-bound RNAPs (inset). Panels A–C adapted from . Panel D adapted from Ref. . Panels E,F adapted from . (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 5

Spatial organization of transcription and non-specific DNA interactions. A) Transmission image of an example cell, and an image of DNA stained with an intercalating fluorescent dye. The distribution of sorted mobile RNAP trajectories (blue lines/bars) closely matched the distribution of DNA (green line). The distribution of bound RNAPs in the same example cell shows a more clustered distribution which does not closely follow the distribution of DNA. B) Spatial distribution of sorted RNAP trajectories averaged over ∼200 cells between 1.6 and 2.5 µm long. Transcribing RNAPs show a bias towards the periphery of the nucleoid region, which is lost after blocking elongating RNAPs with rifampicin. C) Example ‘minimal-DNA’ cell (top); temperature-sensitive DnaC mutant cells are grown at a non-permissive temperature to give long cells with a single centrally located chromosome. Tracking RNAPs only in the DNA-free cell endcaps (green dashed region) allows the free 3D diffusion to be determined. The mean squared displacement (bottom) shows that the diffusion of RNAP in DNA-free cell endcaps (green line) is much faster than the average diffusion of RNAP molecules in normal unperturbed cells (grey line). Panels A–C adapted from . (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)