Bacteria and growth conditions
E. coli strain DH5α  was used for DNA cloning of the expression vectors encoding the fluorescent reporter proteins FbFP and YFP, as well as the biosensor FluBO. E. coli strain BL21(DE3) (Novagen, distributed by Merck KGaA, Darmstadt, Germany) was used for expression of the fluorescent proteins and FRET-based biosensor FluBO. For protein expression and purification, bacterial cells were grown either in autoinduction TB medium consisting of 5 g/l glycerol, 12 g/l tryptone, 24 g/l yeast extract, 2.32 g/l KH2PO4, 12.5 g/l K2HPO4 (pH 7.2), lactose 2 g/l, glucose 0.5 g/l (for FbFP expression and purification) or in LB medium consisting of 10 g/l tryptone, 5 g/l yeast extract, 10 g/l NaCl (for YFP expression and purification). BioLector cultivation experiments were conducted in Overnight Express™ Instant TB medium (Novagen) as described below. All media were supplemented with 50 μg/l kanamycin to maintain expression vector pRhotHi-2.
Construction of FP expression vectors
The construction of expression vector pRhotHi-2-YFP encoding the enhanced yellow fluorescent protein YFP (available from Clontech-Takara Bio Europe, Saint-en-laye, France) and pRhotHi-2-FbFP encoding EcFbFP (GenBank number: ABN71355; commercially available as evoglow-Bs2, evocatal GmbH, Düsseldorf, Germany) was described previously . The YFP variant used in this study contains four amino acid substitutions previously published as GFP-10C , as well as the substitution H231L and the insertion of valine behind the first amino acid. The expression vector pRhotHi-2-FluBO harboring the FRET-based oxygen biosensor encoding gene was constructed using a synthetic FluBO gene (Eurofins MWG Operon, Ebersberg, Germany) whose DNA sequence is shown in Additional file 1. The biosensor gene was cloned into the NdeI and XhoI restriction sites of pRhotHi-2  and recombinant FP genes were confirmed by DNA sequencing. The pRhotHi-2-FluBO vector will soon be commercially available (evocatal GmbH, Düsseldorf, Germany).
Protein expression and purification
E. coli BL21(DE3) cells carrying the expression plasmids pRhotHi-2-FbFP, pRhotHi-2-YFP and pRhotHi-2-FluBO, respectively, were grown until the cultures reached an optical density of approximately 0.5 at 580 nm. Protein expression was then induced by adding 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) to the culture medium. Recombinant oxygen biosensor FluBO, YFP and FbFP were purified as His6-tagged proteins from E. coli BL21(DE3) after 16 h of expression at 37°C. Purification was carried out by using HisTag Superflow Hightrap gravityFlow cartridges (Qiagen, Hilden, Germany) and peristaltic pump P-1 (GE Healthcare Europe GmbH, Munich, Germany) under standard operation conditions as described by the manufacturers. The purified proteins were stored at 4°C in protein storage solution containing 10 mM NaCl, 10 mM NaH2PO4, pH 8.0.
Spectral analysis and fluorometry
The absorption and fluorescence properties of FbFP, YFP and FluBO were determined in 10 mM NaCl and 10 mM sodium phosphate buffer pH 8.0 on a UV-2450 absorption spectrophotometer (Shimadzu Europa GmbH, Duisburg, Germany) and a QM-4 fluorescence spectrophotometer (PTI, Ford, West Sussex, UK), the latter with spectral correction for detector and emission monochromator transmission. The absorption spectra were recorded by scanning from 330 nm to 620 nm; the emission spectra were recorded at excitation wavelengths of 380 nm or 440 nm. When emission spectra were measured the maximum absorption of the samples did not exceed 0.15. In the case of excitation/emission (2D) scans, fluorescence excitation was determined from 300 nm to 600 nm in 5 nm steps by using the PerkinElmer LS50B Luminescence Spectrometer (Wellesley, MA, USA) at 22°C. Here, the fluorescence emission spectra were detected from 450 nm to 595 nm wavelength.
Thrombin protease cleavage
For the thrombin protease mediated cleavage of the FluBO fusion protein, 10 μg of purified protein and one unit of thrombin protease (Novagen, distributed by Merck KGaA, Darmstadt, Germany) were diluted in 50 μl thrombin cleavage buffer. After incubation at room temperature for 16 hours an aliquot of 10 μl was used for SDS-PAGE analysis to corroborate the cleavage of the YFP-FbFP fusion by thrombin protease. SDS-PAGE gels were prepared using standard methods (NuPAGE, 4 to 12% Bis-Tris gel 1 mm, Invitrogen, Paisley, UK). After SDS-PAGE separation, proteins were transferred to a PVDF membrane (Bio-Rad Laboratories GmbH, Munich, Germany) by Western blotting (NuPAGE, Invitrogen, Paisley, UK). Subsequently, YFP and FbFP proteins were detected using the respective antibodies (rabbit anti-GFP, BD Biosciences Erembodegem, Belgium; YtvA-specific antiserum). To detect the YFP-FbFP fusion protein FluBO, a His-tag antibody (Anti-His-HRP antibody, Invitrogen, Paisley, UK) was used. Immunodetected proteins were visualized using a CCD camera-based chemiluminescence detection system (Stella imaging system with AIDA advanced image data analyzer, Raytest, Straubenhardt, Germany). Fluorescence emission of FPs was determined by using the PerkinElmer LS50B Luminescence Spectrometer at 22°C as described above.
Fluorescence lifetime imaging and analysis
The E. coli cells expressing FbFP, YFP or FluBO were placed in suspension between two glass slides on the stage of an upright fluorescence microscope (BX50WI; Olympus Optical, Tokyo, Japan) and observed through a 60 × water-immersion objective (NA = 0.9; Olympus Optical). Fluorescence was excited with ca. 100 fs light pulses (λexc = 760 nm) applied at sufficient intensity to generate two-photon excitation. Light pulses were generated at a frequency of 80 MHz by a mode-locked Titan-Sapphire laser (MaiTai DeepSee; output power > 2W; Newport, Spectra Physics Irvine, CA, USA). The laser light was directed through the lens onto the E. coli at reduced power (ca. 5 mW) and scanned over the sample using a beam scanner (TILL Photonics, Munich, Germany). Fluorescence was recorded by a photomultiplier (PMT-100, Becker & Hickl, Berlin, Germany) using appropriate filters for detecting the FbFP fluorescence of FbFP and FluBO (short pass filter 500 nm; λobs < 500 nm; Omega Optical, Brattleboro, VT, USA) and the YFP fluorescence of YFP and FluBO (band pass filter 535 +/- 13 nm; Omega Optical). Fluorescence lifetime imaging was performed using electronics for time-correlated single photon counting (Simple-Tau 152; Becker & Hickl) and acquisition software (SPCM8.95; Becker & Hickl) for time-correlated single-photon counting as described before [50, 51]. Lifetime images were analyzed using SPCImage 2.97 (Becker & Hickl) by fitting mono- or biexponential model functions to the fluorescence decays in every pixel of the image. The program uses an iterative reconvolution of the exponential function with an instrument response function and a least square algorithm for finding parameters of the exponential function for a satisfactory fit.
Characterization of the oxygen biosensor using the BioLector microbioreactor system
All cultivations were carried out in sterile black 48-well microtiter plates (Flowerplate, m2p-labs, Aachen, Germany) in the microcultivation and on-line monitoring system BioLector (m2p-labs, Aachen, Germany) [36, 37]. These microtiter plates allow higher oxygen transfer rates as compared to conventional microtiter plates . Furthermore, each well is equipped with an optode for the determination of DOT (% air saturation) via fluorescence measurement. The microtiter plates were sealed with sterile gas-permeable adhesive seals (Thermo Scientific, Waltham, MA, USA), ensuring sterile conditions in the wells and allowing good ventilation. The following conditions were applied for all cultivations in the BioLector: temperature 30°C, total filling volume per well: 600 μl, shaking diameter: 3 mm, relative humidity in the incubation chamber: 80%. For all cultivations Overnight Express™ Instant TB medium (Novagen, distributed by Merck KGaA, Darmstadt, Germany) was used. Cultures were inoculated from pre-cultures to an initial OD600 of 0.1. Pre-cultures were made in 250 ml flasks under the following conditions: inoculation from a cryoculture to yield an initial OD600 of 0.1, temperature 37°C, total filling volume 10 ml of TB medium, shaking diameter 50 mm, shaking frequency 350 rpm, growth overnight to stationary phase. For the batch cultivation in the BioLector the shaking frequency was 800 rpm. For DOT shift experiments the initial shaking frequency was 600 rpm, which was increased to 700 rpm after 11.2 h and 800 rpm after 12.2 h. The biomass (I-Io) was measured via scattered light intensity at an excitation wavelength of 620 nm. Donor (FbFP) and acceptor (YFP) fluorescence were excited at 380 nm and emission was detected at 492 nm and 532 nm, respectively. DOT was measured at an excitation wavelength of 505 nm and an emission wavelength of 590 nm. The measuring cycle was 10 minutes for the batch cultivation and 4.5 minutes for the DOT shift experiment. Parallel fermentation experiments in the microwells were performed in triplicates. These results were in excellent agreement. Maximum time shifts of 10 minutes occurred due to unavoidable slight differences of cell density at the beginning of the cultivation. Therefore, a representative set of data from these three parallel independent measurements is shown instead of the corresponding mean values.
For the calibration of the optode-derived DOT signal and the FRET signal we considered the derivation of the YFP-FbFP ratio over the time (d (ratio)/d (t)). In this way, it is possible to also follow decreasing DOT values. To reduce the influence of measuring errors for the calculation of d (ratio)/d (t) the curves for the FluBO fluorescence ratio were fitted to the raw data. This was done in MATLAB 7.11 (The MathWorks, Inc., Natick, MA, USA) applying a smoothing spline with a smoothing parameter of 0.999. In the used medium the value of 100% DOT corresponds to an oxygen concentration of 0.24 mmol/l at 30°C and an atmospheric pressure of 100 kPa.