Summary Fluorescent proteins (FPs) are of vital importance to biomedical research. Many of the currently available fluorescent proteins do not fluoresce when expressed in non‐native environments, such as the bacterial periplasm. This strongly limits the options for applications that employ multiple FPs, such as multiplex imaging and Förster resonance energy transfer (FRET). To address this issue, we have engineered a new cyan fluorescent protein based on mTurquoise2 (mTq2). The new variant is dubbed superfolder turquoise2ox (sfTq2ox) and is able to withstand challenging, oxidizing environments. sfTq2ox has improved folding capabilities and can be expressed in the periplasm at higher concentrations without toxicity. This was tied to the replacement of native cysteines that may otherwise form promiscuous disulfide bonds. The improved sfTq2ox has the same spectroscopic properties as mTq2, that is, high fluorescence lifetime and quantum yield. The sfTq2ox‐mNeongreen FRET pair allows the detection of periplasmic protein‐protein interactions with energy transfer rates exceeding 40%. Employing the new FRET pair, we show the direct interaction of two essential periplasmic cell division proteins FtsL and FtsB and disrupt it by mutations, paving the way for in vivo antibiotic screening. Significance The periplasmic space of Gram‐negative bacteria contains many regulatory, transport and cell wall‐maintaining proteins. A preferred method to investigate these proteins in vivo is by the detection of fluorescent protein fusions. This is challenging since most fluorescent proteins do not fluoresce in the oxidative environment of the periplasm. We assayed popular fluorescent proteins for periplasmic functionality and describe key factors responsible for periplasmic fluorescence. Using this knowledge, we engineered superfolder mTurquoise2ox (sfTq2ox), a new cyan fluorescent protein, capable of bright fluorescence in the periplasm. We show that our improvements come without a trade‐off from its parent mTurquoise2. Employing sfTq2ox as FRET donor, we show the direct in vivo interaction and disruption of unique periplasmic antibiotic targets FtsB and FtsL.

1 the bacterial periplasm allows high 2 efficiency in vivo FRET of cell division 3 antibiotic targets. Abstract 8 Fluorescent proteins (FP)s are of vital importance to biomedical research. Many of the currently 9available fluorescent proteins do not fluoresce when expressed in non-native environments, such as 10 the bacterial periplasm. This strongly limits the options for applications that employ multiple FPs, such 11 as multiplex imaging or FRET. To address this issue, we have engineered a new cyan fluorescent protein 12 based on mTurquoise2 (mTq2). The new variant is dubbed superfolder turquoise 2 ox (sfTq2 ox ) and is 13 able to withstand challenging, oxidizing environments. sfTq2 ox has improved folding capabilities and 14 can be expressed in the periplasm at higher concentrations without toxicity. This was tied to the 15 replacement of native cysteines that may otherwise form promiscuous disulfide-bonds. The improved 16 sfTq2 ox has the same spectroscopic properties as mTq2, i.e. high fluorescence lifetime and quantum 17 yield. The sfTq2 ox -mNeongreen FRET pair allows the detection of periplasmic protein-protein 18 interactions with energy transfer rates exceeding 40 %. Employing the new FRET pair, we show the 19 direct interaction of two essential periplasmic cell division proteins FtsL and FtsB and disrupt it by 20 mutations, paving the way for in vivo antibiotic screening. 21 22

This protocol was developed to qualitatively and quantitatively detect protein-protein interactions in all compartments of Escherichia coli by Förster Resonance Energy Transfer (FRET) using the Superfolder mTurquoise2 ox-mNeonGreen FRET pair (sfTq2 ox -mNG). This FRET pair has more than twice the detection range for FRET interaction studies in the cytoplasm or periplasm of E. coli compared to other pairs to date. These protein-interaction studies can be performed in vivo because fluorescent proteins can be genetically encoded as fusions to proteins of interest and expressed in the cell. sfTq2 ox and mNG fluorescent protein fusions are co-expressed in bacterial cells and the fluorescence emission spectra are measured. By also measuring reference spectra for the background, sfTq2 ox -only and mNG-only samples, expected emission spectra can be calculated. Sensitized emission for mNG above the expected spectrum can be attributed to FRET and quantified by spectral unmixing. This bio-protocol discusses the sfTq2 ox -mNG FRET pair and provides a practical guide in preparing the protein fusions, setting up and running the FRET experiments, measuring fluorescence spectra and gives the tools to analyze the collected data.

The fluorescent proteins superfolder mTurquoise2ox (sfTq2ox) and mNeonGreen function excellently in mammalian cells, but are not well expressed in E. coli when forming the N-terminus of constructs. Expression was increased by decreasing structures at the start of their coding sequences in the mRNA. Unfortunately, the expression of mNeonGreen started from methionine at position ten as optimisation introduced an alternative RBS in the GFP N-terminus of mNeonGreen. The original start-codon was not deleted, which caused the expression of isomers starting at the original start-codon and at the start-codon at the beginning of the GFP N-terminus. By omitting the GFP N-terminus of mNeonGreen and optimising the structure of its mRNA, the expression of a mixture of isomers was avoided, and up to ~ 50-fold higher expression rates were achieved for mNeonGreen. Both fluorescent proteins can now be expressed at readily detectable levels in E. coli and can be used for various purposes. One explored application is the detection of in-cell protein interactions by FRET. mNeonGreen and sfTq2ox form a FRET pair with a larger dynamic range than commonly used donor–acceptor pairs, allowing for an excellent signal-to-noise ratio, which supports the detection of conformational changes that affect the distance between the interacting proteins.