Limits...
Large Scale Bacterial Colony Screening of Diversified FRET Biosensors.

Litzlbauer J, Schifferer M, Ng D, Fabritius A, Thestrup T, Griesbeck O - PLoS ONE (2015)

Bottom Line: Biosensors based on Förster Resonance Energy Transfer (FRET) between fluorescent protein mutants have started to revolutionize physiology and biochemistry.Thus, a major effort in the field currently lies in designing new optimization strategies for these types of sensors.We describe optimization of biosensor expression, permeabilization of bacteria, software tools for analysis, and screening conditions.

View Article: PubMed Central - PubMed

Affiliation: Max-Planck-Institut für Neurobiologie, Am Klopferspitz 18, Martinsried, Germany.

ABSTRACT
Biosensors based on Förster Resonance Energy Transfer (FRET) between fluorescent protein mutants have started to revolutionize physiology and biochemistry. However, many types of FRET biosensors show relatively small FRET changes, making measurements with these probes challenging when used under sub-optimal experimental conditions. Thus, a major effort in the field currently lies in designing new optimization strategies for these types of sensors. Here we describe procedures for optimizing FRET changes by large scale screening of mutant biosensor libraries in bacterial colonies. We describe optimization of biosensor expression, permeabilization of bacteria, software tools for analysis, and screening conditions. The procedures reported here may help in improving FRET changes in multiple suitable classes of biosensors.

No MeSH data available.


Imaging bacterial colonies expressing a FRET biosensor library.(A) Basal ratio R0 plotted against calcium-induced ratio change ΔR/R for every colony of one agar plate blotted onto white paper. Each colony is supposed to express a different diversified FRET calcium sensor variant. The 10 best performing colonies of this experiment are highlighted in yellow and numbered. (B) Single FRET traces of the 10 best colonies as identified in A. The arrow indicates when calcium was applied (C) Scheme of a screening plate with the positions of the 10 best performing colonies highlighted.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4464885&req=5

pone.0119860.g004: Imaging bacterial colonies expressing a FRET biosensor library.(A) Basal ratio R0 plotted against calcium-induced ratio change ΔR/R for every colony of one agar plate blotted onto white paper. Each colony is supposed to express a different diversified FRET calcium sensor variant. The 10 best performing colonies of this experiment are highlighted in yellow and numbered. (B) Single FRET traces of the 10 best colonies as identified in A. The arrow indicates when calcium was applied (C) Scheme of a screening plate with the positions of the 10 best performing colonies highlighted.

Mentions: Analysis of FRET responses from a large number of bacterial colonies provided another challenge. We developed software in the programming language Python, used with the software package μManager [31] to control the camera, filter wheels and shutters and to guide the user through the screening process step by step, prompting certain actions at the appropriate time. Furthermore the program identified all colonies on a plate, recorded fluorescence intensities at all wavelengths of interest for those colonies, calculated FRET ratios, analyzed the data and identified the best performing colonies and sensors (Fig 4). In detail, fluorescent colonies were first imaged for YFP fluorescence, and the threshold for colony identification was set at 3σ of pixel intensity. This clearly found all the brightly fluorescent bacterial colonies on the non-fluorescent plate, generating a binary map of colony locations that was further processed with binary erosion and opening operations to clean up colony boundaries. Automated image segmentation was performed on the binary image, and the identified location and areas of the colonies were used to measure fluorescence intensities and calculate the FRET ratios for the colonies in real-time from subsequent images. The colonies were then ranked by ΔR/R0 divided by R00.5, resulting in the selection of colonies according to their responses with a bias towards higher fluorescent changes rather than lower starting ratio. Following an experiment, the user was presented with three plots; a landscape of all sensors on the plate in respect to their R0 and ΔR/R, with a pre-selected number of well performing colonies in both criteria highlighted and labeled according to their ranking, single traces of those colonies, and lastly a scheme of the plate indicating the positions of the best colonies (Fig 4A–4C). This scheme could then be used to pick these colonies from the plate and retrieve the DNA coding for the expressed biosensor. The commented code was deposited at: https://github.com/GriesbeckLab/ColonySelection


Large Scale Bacterial Colony Screening of Diversified FRET Biosensors.

Litzlbauer J, Schifferer M, Ng D, Fabritius A, Thestrup T, Griesbeck O - PLoS ONE (2015)

Imaging bacterial colonies expressing a FRET biosensor library.(A) Basal ratio R0 plotted against calcium-induced ratio change ΔR/R for every colony of one agar plate blotted onto white paper. Each colony is supposed to express a different diversified FRET calcium sensor variant. The 10 best performing colonies of this experiment are highlighted in yellow and numbered. (B) Single FRET traces of the 10 best colonies as identified in A. The arrow indicates when calcium was applied (C) Scheme of a screening plate with the positions of the 10 best performing colonies highlighted.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4464885&req=5

pone.0119860.g004: Imaging bacterial colonies expressing a FRET biosensor library.(A) Basal ratio R0 plotted against calcium-induced ratio change ΔR/R for every colony of one agar plate blotted onto white paper. Each colony is supposed to express a different diversified FRET calcium sensor variant. The 10 best performing colonies of this experiment are highlighted in yellow and numbered. (B) Single FRET traces of the 10 best colonies as identified in A. The arrow indicates when calcium was applied (C) Scheme of a screening plate with the positions of the 10 best performing colonies highlighted.
Mentions: Analysis of FRET responses from a large number of bacterial colonies provided another challenge. We developed software in the programming language Python, used with the software package μManager [31] to control the camera, filter wheels and shutters and to guide the user through the screening process step by step, prompting certain actions at the appropriate time. Furthermore the program identified all colonies on a plate, recorded fluorescence intensities at all wavelengths of interest for those colonies, calculated FRET ratios, analyzed the data and identified the best performing colonies and sensors (Fig 4). In detail, fluorescent colonies were first imaged for YFP fluorescence, and the threshold for colony identification was set at 3σ of pixel intensity. This clearly found all the brightly fluorescent bacterial colonies on the non-fluorescent plate, generating a binary map of colony locations that was further processed with binary erosion and opening operations to clean up colony boundaries. Automated image segmentation was performed on the binary image, and the identified location and areas of the colonies were used to measure fluorescence intensities and calculate the FRET ratios for the colonies in real-time from subsequent images. The colonies were then ranked by ΔR/R0 divided by R00.5, resulting in the selection of colonies according to their responses with a bias towards higher fluorescent changes rather than lower starting ratio. Following an experiment, the user was presented with three plots; a landscape of all sensors on the plate in respect to their R0 and ΔR/R, with a pre-selected number of well performing colonies in both criteria highlighted and labeled according to their ranking, single traces of those colonies, and lastly a scheme of the plate indicating the positions of the best colonies (Fig 4A–4C). This scheme could then be used to pick these colonies from the plate and retrieve the DNA coding for the expressed biosensor. The commented code was deposited at: https://github.com/GriesbeckLab/ColonySelection

Bottom Line: Biosensors based on Förster Resonance Energy Transfer (FRET) between fluorescent protein mutants have started to revolutionize physiology and biochemistry.Thus, a major effort in the field currently lies in designing new optimization strategies for these types of sensors.We describe optimization of biosensor expression, permeabilization of bacteria, software tools for analysis, and screening conditions.

View Article: PubMed Central - PubMed

Affiliation: Max-Planck-Institut für Neurobiologie, Am Klopferspitz 18, Martinsried, Germany.

ABSTRACT
Biosensors based on Förster Resonance Energy Transfer (FRET) between fluorescent protein mutants have started to revolutionize physiology and biochemistry. However, many types of FRET biosensors show relatively small FRET changes, making measurements with these probes challenging when used under sub-optimal experimental conditions. Thus, a major effort in the field currently lies in designing new optimization strategies for these types of sensors. Here we describe procedures for optimizing FRET changes by large scale screening of mutant biosensor libraries in bacterial colonies. We describe optimization of biosensor expression, permeabilization of bacteria, software tools for analysis, and screening conditions. The procedures reported here may help in improving FRET changes in multiple suitable classes of biosensors.

No MeSH data available.