Genetic Code Expansion Enables Live-Cell and Super-Resolution Imaging of Site-Specifically Labeled Cellular Proteins

Methods to site-specifically and densely label proteins in cellular ultrastructures with small, bright, and photostable fluorophores would substantially advance super-resolution imaging. Recent advances in genetic code expansion and bioorthogonal chemistry have enabled the site-specific labeling of proteins. However, the efficient incorporation of unnatural amino acids into proteins and the specific, fluorescent labeling of the intracellular ultrastructures they form for subdiffraction imaging has not been accomplished. Two challenges have limited progress in this area: (i) the low efficiency of unnatural amino acid incorporation that limits labeling density and therefore spatial resolution and (ii) the uncharacterized specificity of intracellular labeling that will define signal-to-noise, and ultimately resolution, in imaging. Here we demonstrate the efficient production of cystoskeletal proteins (β-actin and vimentin) containing bicyclo[6.1.0]nonyne-lysine at genetically defined sites. We demonstrate their selective fluorescent labeling with respect to the proteome of living cells using tetrazine-fluorophore conjugates, creating densely labeled cytoskeletal ultrastructures. STORM imaging of these densely labeled ultrastructures reveals subdiffraction features, including nuclear actin filaments. This work enables the site-specific, live-cell, fluorescent labeling of intracellular proteins at high density for super-resolution imaging of ultrastructural features within cells.

(18) Mukai, T.; Kobayashi, T.; Hino, N.; Yanagisawa, T.; Sakamoto, K.; Yokoyama, S. Biochem. Biophys. Res. Commun. 2008, 371, 818. (19) Greiss, S.; Chin, J. W. J. Am. Chem. Soc. 2011, 133, 14196. (20) Figure S1. BCNK is more cleanly removed from mammalian cells than TCOK by washing. COS-7 cells were incubated with 30 µM -1 mM of TCOK or BCNK for 24 hours. Thereafter, the media containing TCOK or BCNK was replaced by fresh growth media 5-6 S5 times over 2 hours. Cells were then fixed and blocked with 3% bovine serum albumin (BSA) in DPBS. To detect residual TCOK or BCNK in cells, cells were treated with 400 nM Alexa Fluor 647-3,6-dipyridyl-1,2,4,5-tetrazine conjugate in 3% BSA in DPBS for 30 min, the duration of which should allow Diels-Alder derivatization on both TCOK and BCNK to go to completion. Cells were then washed 3 times with 0.1% Tween-20 in DPBS and imaged. A) Representative images of fluorescent signals from residual amino acids. B) Quantification of fluorescent signals from residual amino acids. >15 single cells were used for quantification for each condition. Error bars, ± s.d. Thereafter, actin-BCNK was detected via anti-HA immunoblotting, while vimentin-BCNK was detected via anti-c-myc blotting. To quantify relative amount of actin/vimentin produced, a standard curve was generated from band densitometries from serial dilutions of known amount of lysates containing vimentin N116TAG (for the vimentin standard curve) or actin K118TAG (for the actin standard curve). Using the standard curves, band densitometries obtained from 3 expression replicates of each actin/vimentin variant are converted to actin/vimentin amount (in arbitrary units) in the lysates. The relative expression levels given are normalized with respect to the higher expressing variant for each protein (actin K118TAG or vimentin N116TAG ). Figure S3. An optimized tRNA synthetase/tRNA expression system improves unnatural amino acid incorporation into mammalian proteins. HEK293T cells were transfected with either of the two plasmid configurations: 1) pBCNKRS-actin TAG and p4CMVE-U6-PylT; or 2) (U6-PylT) 4 /EF1α-BCNKRS and (U6-PylT) 4 /EF1α-actin TAG . Cells were grown in the presence of 1 mM BCNK for 48 hours before they were lysed. Actin expression was visualized by western blot using anti-HA antibody against the N-terminal HA-tag on actin.    4 /vimentin N116TAG and grown in the presence of 1 mM BCNK for 48 hours. After removal of excess BCNK, cells were treated with 400 nM tet1-CFDA for 20 min, then washed for 2 hours before fixation. Vimentin expression was visualized via anti-c-myc immunostaining, while Pyl tRNA CUA expression was detected via fluorescence in situ hybridization with an oligonucleotide-Alexa Fluor 647 conjugate specific to Pyl tRNA CUA . Three different fields of view are shown. White arrows in the fluorescein channel highlight a few cells with clear Pyl tRNA CUAdependent labeling background in the cell nuclei. Note that immunostaining of vimentin shows that the protein is exclusively cytosolic, and thus does not contribute to labeling observed in the nuclei. In order to highlight subcellular features in different samples, images in this Figure were not normalized to the same intensity range. Figure S8. Labeling specificity with various tetrazine-fluorophore conjugates. A) Cellsurface protein labeling with tet2-ATTO488 and tet2-Alexa Fluor 647. For ATTO488 labeling, HEK293T cells were transfected with pTCOKRS-EGFR(128TAG) and (U6-PylT U25C ) 4 /EF1α-BCNKRS and grown in the presence of 1 mM BCNK. After removal of excess BCNK, cells were treated with tet2-ATTO488, then washed before fixation. Anti-FLAG immunostaining of BCNKRS was used as a marker for transfection. ATTO488 fluorescence is shown as green and anti-FLAG staining as red in the merged image. For Alexa Fluor 647 labeling, HEK293T cells were transfected with pTCOKRS-EGFR(128TAG)-GFP and (U6-PylT U25C ) 4 and grown in the presence of 1mM TCOK. After removal of excess TCOK, cells were treated with tet2-Alexa Fluor 647, then washed before fixation. Alexa Fluor 647 fluorescence is shown as red and GFP fluorescence as green in the merged image. B) Intracellular protein labeling with tet1-SiR. COS-7 cells were transfected with plasmids encoding (Pyl tRNA CUA ) 4 /BCNKRS and (Pyl tRNA CUA ) 4 /POI (POI = protein of interest = actin K118TAG or vimentin N116TAG ) and grown in the presence of 1 mM BCNK. After removal of excess BCNK, cells were treated with tet1-SiR, then washed for 2 hours before fixation. Expression of actin-BCNK and vimentin-BCNK was further confirmed via anti-HA and anti-c-myc immunofluorescence staining, respectively. Note that the nuclear background level is higher in COS-7 cells than HEK cells and cannot be easily removed through long amino acid washout, reflecting potential differences in PylT turnover capabilities in different cell types. DAPI is a nuclear marker. DIC: differential interference contrast image. Scale bar, 5 µm. S14 Figure S9. The vimentin filament in the boxed region was used for cross-sectional analysis shown in Figure 3B.   Figure S13. Imaging of nuclear actin filaments. Widefield fluorescence images of a COS-7 cell expressing actin K118TAG labeled with BCNK and tet1-SiR show nuclear filamentous actin (first column), and at higher contrast, cytosolic actin (second column). DAPI is a nuclear marker and was imaged at an identical incident angle as the actin images (third column). Scale bar, 5 µm.
Tetrazine-fluorophore conjugates were synthesized using the following generic procedure. To a solution of the succinimidyl ester (NHS) of the fluorophore (5 µmol) in anhydrous N,Ndimethylformamide was added the tetrazine-HCl salt (10 µmol) and N,Ndiisopropylethylamine (15 µmol). After stirring at ambient temperature in the dark, the reaction progress to completion (judged by consumption of the starting fluorophore-NHS) was monitored by LC-MS. Thereafter, DMF was evaporated, and the resulting residue was dissolved in water/acetonitrile (9/1) and the product was purified by preparative reverse phase HPLC using a gradient from 20% to 85% of buffer B in buffer A (buffer A: H 2 O, 0.2% TFA; buffer B: acetonitrile, 0.2% TFA). The fractions containing product were evaporated, the resulting concentrated solution flash-frozen and lyophilized to yield the desired tetrazinefluorophore conjugate. The tet1-SiR conjugate was synthesized as previously described 5 .
ESI-MS characterizations of various tetrazine-fluorophore conjugates were as follows:

Mammalian cell culture and transfection
Human embryonic kidney 293T (HEK293T) and COS-7 cells were cultured in Dulbecco's modified Eagle Medium (DMEM, Life Technologies) supplemented with 10% v/v fetal bovine serum (FBS) and non-essential amino acids (Life Technologies). Cells were maintained at 37°C in a 5% CO 2 atmosphere. For imaging, cells were plated as a monolayer on an 8-well Lab-Tek II chambered coverglass (Thermo Scientific) pre-coated with 50 µg/mL human fibronectin (Millipore). For western blotting and gel-based analyses, cells were plated in 24-well plates.
Transient transfection was performed using 25 kDa polyethylenimine (PEI, Polysciences) in a 5:1 PEI:DNA ratio (1.25 µL of 1 µg/µL PEI and ~250 ng total DNA plasmid per well of an 8-well chambered coverglass). DNA was diluted in 150 mM NaCl prior to addition of PEI.

Intracellular labeling of actin and vimentin in mammalian cells
HEK293T or COS-7 cells were transfected at ~80% confluency with PEI with 125 ng each for (U6-PylT U25C ) 4 /EF1α-BCNKRS and (U6-PylT U25C ) 4 /EF1α-POI (POI = protein of interest = actin D4TAG , actin K118TAG , vimentin N116TAG , vimentin E187TAG , vimentin N116TAG -mApple, or vimentin E187TAG -mApple). 4 hours after transfection, the media was exchanged to normal cell growth media (with 10% FBS), and 1 mM BCNK was added. After 36-48 hours of incubation with the media containing BCNK, cells were rinsed with fresh growth media 3-4 times, over the period of 6-9 hours. 400-800 nM of a tet1-fluorophore conjugate in DMEM (without serum) was then added to cells for 20-30 minutes at ambient temperature. For experiments in COS-7 cells, pluronic F-127 (Life Technologies) was added to the final concentration of 0.1% w/v and mixed with the tet1-fluorophore before DMEM was added, to improve cellular dye loading capacity. Thereafter, cells were rinsed 3-4 times over 2 hours with fresh growth media at 37°C. Cells were either imaged live, fixed before imaging using specific fixation protocols to preserve actin or vimentin ultrastructure, or lysed for gel electrophoretic analyses.
To generate cell lysates, cells were lysed in RIPA buffer (Sigma) supplemented with complete protease inhibitor (Roche). Lysates were denatured and run on a 4-12% Bis-Tris protein gel (Life Technologies). In-gel fluorescence of fluorescein was visualized on a Typhoon Trio phosphoimager (GE Life Sciences).

Western blotting of mammalian cell lysate
HEK293T cells were grown in 24-well plates, transfected as previously described (using twice the amount of PEI and DNA compared to experiments in an 8-well Lab-TEK II well), and incubated with 1 mM BCNK or TCOK for 36-48 hours. Thereafter, excess amino acid was removed via quick rinsing (2-3 times) with DPBS, and cells were lysed in RIPA buffer supplemented with complete protease inhibitor. Lysates were denatured, run on a 4-12% Bis-Tris gel, and transferred onto a nitrocellulose membrane via an iBlot dry transfer system (Life Technologies). The membrane was blocked with 5% milk in PBS + 0.1% Tween-20 (PBS-T) for 1 hour at ambient temperature. For anti-HA blotting, the membrane was incubated with rat monoclonal anti-HA antibody (clone 3F10, Roche) in 5% milk in PBS-T at 1:2000 dilution overnight at 4°C, followed by 5-6 washes with PBS-T over one hour. The membrane was then incubated with anti-rat horseradish peroxidase conjugate (Santa Cruz Biotechnology) in 5% milk in PBS-T at 1:5000 dilution for one hour at ambient temperature, followed by 5-6 washes with PBS-T. For anti c-myc blotting, the membrane was incubated with mouse monoclonal anti c-myc antibody (clone 9B11, Cell Signaling) in 5% milk in PBS-T at 1:1000 dilution overnight at 4°C, followed by 5-6 washes with PBS-T over one hour. The membrane was then incubated with anti-mouse horseradish peroxidase conjugate (Santa Cruz Biotechnology) in 5% milk in PBS-T at 1:10,000 dilution for one hour at ambient temperature, followed by 5-6 washes with PBS-T. The blots were developed using Supersignal West Femto substrate (Pierce) and visualized on a ChemiDoc MP imaging system (Bio-Rad).

Immunofluorescence staining
For anti-FLAG staining of HEK cells, cells were fixed with 4% formaldehyde in PBS, then blocked in 10% FBS in PBS for 30 minutes. FLAG tag immunofluorescence staining was performed with mouse monoclonal anti-FLAG antibody (clone M2, Sigma) in 10% FBS in PBS at 1:200 dilution for 1 hour at room temperature. After several washes with PBS, cells were incubated with goat anti-mouse antibody conjugated to Alexa Fluor 405 (Life Technologies) at 1:100 dilution for 1 hour at room temperature, then washed and imaged.
For anti-HA staining of COS-7 cells, cells were fixed with 4% formaldehyde in DPBS for 10 minutes at room temperature, permeabilized in 0.5% v/v Triton X-100 in DPBS for 10 minutes at room temperature, then blocked in 3% BSA in DPBS for 1 hour at room temperature. HA tag immunofluorescence staining was performed with rat monoclonal anti-HA antibody (clone 3F10, Roche) in 3% BSA in DPBS at 1:300 dilution for 1 hour at room temperature. After washing with 0.1% Tween-20 in DPBS, cells were incubated with goat anti-rat antibody conjugated to Alexa Fluor 488 (Life Technologies) at 1:300 dilution for 1 hour at room temperature, then washed and imaged. localizations yielded a standard deviation (σ) of the distribution, which is proportional to the filament/membrane width (full width at half maximum, FWHM; FWHM = 2.35σ). The matlab code provided by http://www.diplib.org/add-ons/ 2 was used to calculate the Fourier ring correlation (FRC) curve and the STORM image resolution with a 1/7 threshold criterion. The spurious correlation parameter Q was not used in the FRC calculations. Separate software from Banterle et al J Struct Biol 2013, 183, 363 was used to calculate a FRC-based resolution value with a 2σ threshold. The local resolution map was generated by calculating FRC-based resolution values for square subregions of 12.8 x 12.8 pixels, each subregion translated by 9.6 pixels across the image, so that each pair of neighbouring regions overlapped by 25% of their area. The displayed resolution map was generated by linear interpolation.

STimulated Emission Depletion (STED) microscopy
Images were acquired on a Leica TCS SP8 STED confocal system with time-gated HyD GaAsP detectors using a 100x/1.4 NA oil immersion objective lens. Excitation was achieved with a NKT Super K pulsed white light (470-670 nm) laser tuned to 488 nm and STED was induced with a 592 nm laser (~40 MW/cm 2 ). A time gate window of 2-6 ns was used to maximise STED resolution. 20 nm pixels were used to ensure adequate spatial sampling to support maximal expected resolution. Data were processed and analysed using LAS AF software.

Structured Illumination Microscopy (SIM)
Images were acquired on a Zeiss ELYRA S.1 system with a pco.edge 5.5 scientific CMOS camera using a 63x/1.4 NA oil immersion objective lens. For optimal lateral spatial sampling a 1.6x intermediate magnification lens was used, resulting in a pixel size of 64 nm. Excitation was achieved with a 488 nm laser and emission was filtered with a 495-550 nm band pass filter. Modulation of the illumination light was achieved using a physical grating with 28 µm spacing. This patterned illumination was shifted through 5 phases at each of 3 rotational angles per z plane. A z-interval of 110 nm was used to ensure adequate spatial sampling in the axial dimension. Raw data were processed using ZEN software and analysed in ImageJ.