Quenched Substrates For Background-free Fluorescence .

Supporting InformationQuenched substrates for background-free fluorescence labellingof hAGT (proteins) in living cellsKatharina Stöhr, Daniel Siegberg, Tanja Erhard, Konstantinos Lymperopoulos, Simin Öz, SonjaSchulmeister, Andrea C. Pfeifer, Julie Bachmann, Ursula Klingmüller, Victor Sourjik, Dirk-PeterHertenEXPERIMENTAL PROCEDURESLabelling of Benzylguanine. For our experiments we created a library of different BG-dye conjugatesusing the fluorescent dyes ATTO 425, ATTO 495, ATTO 647N, ATTO 488, ATTO 520, ATTO 532,ATTO 550, ATTO 565, ATTO 620, ATTO 633, ATTO 655, ATTO 680, ATTO 700 (ATTO-TECGmbH, Siegen, Germany), MR 121 and DY 113 (provided by Prof. K.-H. Drexhage, University Siegen),Alexa 532, Alexa 568, Alexa 594, (Invitrogen Deutschland, Karlsruhe, Germany), Cy 3b and Cy 5 (GEHealthcare, Munich, Germany), as well as Dyomics 630 (Dyomics GmbH, Jena, Germany). BG-NH2was labelled with the N-hydroxysuccinimide (NHS) activated fluorescent dyes using standardprocedures1. 60µM of a 2·10-3M solution of BG-NH2 in dimethylformamide was labelled at its freeamino-group with the NHS-ester of the respective dye (2mg/ml dimethylformamide). The solution wasincubated for 4 hours at room temperature in the dark. The product was purified by reversed-phase(RP18 column) HPLC (Agilent, Waldbronn, Germany) using triethylammonium acetate. Purified BGDye conjugates were dissolved in 20µl DMF.Spectroscopic characterisation. Steady state absorption and emission spectra have been recorded on aUV/VIS/NIR spectrometer (Cary 500 Scan, Varian, Darmstadt, Germany). Fluorescence emissionspectra of the substrates were measured in standard quartz cuvettes with a Cary Eclipse fluorescence1

spectrometer with temperature controller (Varian). Relative fluorescence intensities and quantum yieldswere determined with respect to the optical density of the sample. The optical density of the solutionwas chosen to be below 0.1 µmol/L to avoid inner filter effects.Engineering and purification of His-tagged SNAP protein. The construct pSS7 expressing Cterminal fusion of 6xHis tag to SNAP-tag was made by cloning SNAP-tag into pQE60 expression vector(Qiagen, Hilden, Germany) carrying an ampicillin resistance and T5 promoter inducible by isopropyl-βD-thiogalactopyranoside (IPTG). SNAP26b gene was amplified from pSS26b (Covalys, Switzerland) byPCR, introducing NcoI and BamHI restriction sites with the 5’ –primer GAT ATC GAA TTC CCCATG GAC AAA GAT TGC and the 3’ – primer GCT CTA GAA CTA GTG GATC CCC GTG CAGGT, respectively.For preparative purification of 6xHis-tag fusion to SNAP-tag, pSS7 was co-transformed in E.colistrain M15 together with pREP4 plasmid (both from Qiagen, Hilden, Germany) that encodes akanamycin resistance and the lacI repressor for T5 promoter. Cells were grown and proteins werepurified according to the manufacturer specifications using Protino Ni-IDA-Kit 2000 packed columns(MACHEREY – NAGEL, Düren, Germany).Engineering of Tar-SNAP fusion protein. The construct pTE1 expressing C-terminal fusion of SNAPtag to Tar was made by replacing EYFP gene in the previously described Tar-YFP expression constructpDK582 using BamHI and NotI restriction sites. Restriction sites were introduced by amplifyingSNAP26b gene from pSS26b (Covalys, Switzerland) using PCR with the 5’ - primer ATA TAG GATCCG TGG ACA AAG ATT GCG and the 3’ – primer TAT ATG CGG CCG CTC ACC CTG CAGGTC CCA G.Engineering of NIH 3T3 cells stably expressing STAT5b-SNAP. The retroviral expression vectorpMOWS containing the cDNA for murine HA-EpoR was introduced into NIH3T3 cells (ATCC) and asingle cell clone stably expressing HA-EpoR was obtained by selection with G418. pMOWSIN-TREtStat5B- SNAP26m was cotransduced into NIH3T3-EpoR cells together with the cDNA for thetransactivator protein (H. Bujard) contained in pMOWS-rtTAM2. A single cell clone stably expressing2

murine Stat5B- SNAP26m was obtained by selection with puromycin. Expression of Stat5B-SNAP26mwas regulated by a Tet-inducible promoter included in pMOWSIN-TREt. pMOWSIN-TREt wasgenerated by digesting pTRE-tight (Clontech) and inserting TREt into the self-inactivating (SIN)retroviral vector pMOWSIN.Cell culturing and sample preparation. For SNAP-tag labelling in E.coli, liquid cultures of RP437strain (Parkinson and Houts 1982) transformed with pSS26b and pREP4 or control RP437 straintransformed with pQE60 and pREP4 were grown in TB medium with 50 µg/ml kanamycin and 100µg/ml ampicillin over night at 30 C at 150 rpm. 100 µl of the overnight culture were mixed with 10 mlTB supplemented with kanamycin and ampicillin and expression of SNAP26b was induced using60 µM isopropyl-β-D-thiogalactopyranosid (IPTG). The culture was grown at 275 rpm and 34 C up toan optical densities of OD600 0.5. Labelling with BG-dye-substrate was performed with 5 µM BGsubstrate. After 2 h, bacteria were centrifuged and the pellet was washed twice with PBS. For labellingof Tar-SNAP fusion, culture of VS102 strain3 transformed with pTE1 was grown in TB with ampicillinand induced overnight with 100 µM IPTG at 30 C and then the bacteria were incubated with a 10 µM ofBG-MR 121 conjugate for 15 min at room temperature. After one washing step, we analyzed thebacteria on our home-build confocal raster-scanning microscope. For imaging purposes the cells wereimmobilized on poly-L-lysin coated slides.NIH-3T3-EpoR.-Stat5B-SNAP cells were cultured in DMEM (Invitrogen, Carlsbad, USA)supplemented with 10% Newborn Calf Serum (PAN-Biotech, Aidenbach, Germany), 100U/mlPenicillin (Sigma Aldrich, St. Louis, USA) and 100µg/ml Streptomycin (Sigma Aldrich) at 37 C in acontrolled humidified atmosphere supplemented with 5% CO2. Cells were subcultured usingTrypsin/EDTA (Sigma Aldrich) before they reached 80% confluency, counted using a Neubauerhaemocytometer and 5x104 cells per well were seeded in a 8-well Ibidi slide (Ibidi GmbH, Martinsried,Germany). Expression of STAT5b-SNAP26m was induced by 50 ng/ml doxycycline (Sigma Aldrich)overnight. Labelling with BG-conjugated dyes was performed at 5 µM final concentration for 1hr at3

37 C after which cells were washed three times with complete media and imaged on a confocalmicroscope.Fluorescence microscopy. Fluorescence microscopy for E.coli expressing SNAP26b was performed onan inverted Axiovert S-100-Microscope (Carl Zeiss, Jena, Germany) with the mercury-arc lamp houseHAL 100 (Carl Zeiss, Jena, Germany) and an Ixon EMCCD-camera (Andor, Belfast, UK.). Filters andbeam splitters have been chosen to get the highest possible detection efficiencies with respect to theinvestigated dye conjugates. Imaging of E.coli expressing Tar-SNAP was performed on a customconfocal raster-scanning microscope. Briefly, a pulsed diode laser emitting at 635 nm (PicoQuantGmbH, Berlin, Germany) was driven by a pulsed laser driver (PDL 808 “Sepia”, PicoQuant) at arepetition rate of 10 MHz and was coupled into an inverted microscope (Axiovert 100TV, Zeiss, Jena,Germany) equipped with a xyz-piezoscanning table (PI Physik Instrumente, Karlsruhe, Germany). Thecollimated laser beam was directed into a water immersion objective (60x, NA 1.2, Olympus, Japan).Fluorescence emission was collected by the same microscope lens and was detected by focusing on anavalanche photo diode (SPCM-AQR 15, Perkin-Elmer). For data recording and analysis a TCSPC PCcard (SPC-630 Becker&Hickl, Berlin, Germany) and custom written LabView-based software wereused. Live mammalian cell imaging was performed on a Leica TCS SP5 X microscope system (LeicaMicrosystems, Wetzlar, Germany) equipped with a supercontinuum light source 470-670 nm (KoherasA/S, Copenhagen, Denmark) and acousto-optical beamsplitter (AOBS) (Leica). Signals were detectedwith cooled Photon Multiplier Tubes (PMTs – Leica). Samples were imaged using a 63x planarapochromatic water immersion objective (Leica) and similar conditions (excitation intensities anddetection sensitivity), visualised using the LAS AF software (Leica) and further processed usingCorelDraw Graphics Suite X4 software (Corel, Ottawa, Canada).Analysis of microscopic images. For image processing of the E.coli samples we used ImageJ(W.Rasband, NIH, Bethesda, USA). To account for systematic variations due to different excitation andemission filters we estimated the total excitation and emission efficiencies in the microscopic images bymultiplying the overlap integral of the dye’s absorption and emission spectra with the respective4

transmission curves of the used filters. According to eq. 1, the detected normalized fluorescenceintensity IN is proportional to the relative quantum yield ΦF of the dye, which we assumed to beindependent of the wavelength λ. The molar extinction ε(λ) and the normalized emission F(λ) spectra ofthe dye are multiplied by the respective transmission spectra TExc(λ) / TEm(λ) of the excitation filter andemission filter.(1)I N φ F TExc (λ )ε (λ ) dλ TEm (λ ) F (λ ) dλThe calculated efficiencies were multiplied with the measured average fluorescence intensities of thebacteria. The average fluorescence intensities were estimated by using a particle auto recognition whichautomatically detected every fluorescent E.coli. The fluorescence intensities of all bacteria in an imagewere averaged over a series of images yielding an average intensity and a standard deviation. Afterwardsthe efficiency was calculated with the fluorescence intensity, leading to an average fluorescenceintensity which corresponds to an excitation and emission of 100%. This value was multiplied with theabsolute quantum yield. The obtained values of the mean fluorescence intensity and the standarddeviation allow a comparison of dyes measured with the same filters.Fluorescence lifetime measurements. All experiments have been carried out in standard quartzcuvettes (Hellma, Müllheim Germany) with a maximum volume of 1.5 ml. For UV/Vis andfluorescence spectroscopy the concentrations were chosen well below an optical density of 0.1 to avoidinner filter effects. Optical density at excitation wavelength has been checked with a Cary Scan 500UV/Vis Photo spectrometer (Varian, Darmstadt, Germany). Ensemble fluorescence lifetimemeasurements were performed on a FluoTime 100 (PicoQuant, Berlin, Germany) using time-correlatedsingle-photon counting (TCSPC). For excitation we used LEDs emitting at 370 nm, 450 nm, 500 nmand 600 nm with a pulse width 600 ps (fwhm) operated at 10 MHz. The measurements were done instandard quartz cuvettes (d 0.3 cm). To exclude polarization effects fluorescence was observed undermagic angle conditions (54.7 ). Typically 3,000 – 5,000 photons were collected in the maximumchannel of a total of 4096 channels. The lifetime was determined with the software provided by the5