Smallest near-infrared fluorescent protein evolved from cyanobacteriochrome as versatile tag for spectral multiplexing

Abstract

From a single domain of cyanobacteriochrome (CBCR) we developed a near-infrared (NIR) fluorescent protein (FP), termed miRFP670nano, with excitation at 645 nm and emission at 670 nm. This is the first CBCR-derived NIR FP evolved to efficiently bind endogenous biliverdin chromophore and brightly fluoresce in mammalian cells. miRFP670nano is a monomer with molecular weight of 17 kDa that is 2-fold smaller than bacterial phytochrome (BphP)-based NIR FPs and 1.6-fold smaller than GFP-like FPs. Crystal structure of the CBCR-based NIR FP with biliverdin reveals a molecular basis of its spectral and biochemical properties. Unlike BphP-derived NIR FPs, miRFP670nano is highly stable to denaturation and degradation and can be used as an internal protein tag. miRFP670nano is an effective FRET donor for red-shifted NIR FPs, enabling engineering NIR FRET biosensors spectrally compatible with GFP-like FPs and blue-green optogenetic tools. miRFP670nano unlocks a new source of diverse CBCR templates for NIR FPs.

Fig. 1

Molecular engineering of miRFP670nano. a–n Comparison of clones selected on each round of selection in HeLa cells. The main mutations are indicated. o Quantification of the data represented in a–n. Mean NIR fluorescence intensity was normalized to mean green fluorescence intensity of co-expressed EGFP and to mean fluorescence intensity of mock-transfected cells. Error bars, s.d. (n = 3; transfection experiments)

Fig. 2

Characterization of miRFP670nano. a Fluorescence excitation and emission spectra of miRFP670nano. b Size-exclusion chromatography of miRFP670nano at concentration 10 mg ml⁻¹ and indicated molecular weight standards. miRFP670nano with polyhistidine tag and linker runs as a monomer with the apparent molecular weight of 18.8 kDa. c pH dependencies of NIR fluorescence for miRFP670nano and miRFP670. d Kinetics of miRFP670nano and miRFP670 maturation. Time “0” corresponds to the beginning of the 1-h-long pulse-chase induction of the protein expression in bacteria. e Effective (cellular) brightness of miRFP670nano, miRFP703, and miRFP670 in mammalian cells. Live HeLa, U87, U-2 OS, PC6-3, and NIH3T3 cells were transiently transfected with miRFP670nano, miRFP703, or miRFP670. Fluorescence was analyzed by flow cytometry 72 h after transfection. NIR fluorescence intensity was normalized to that of co-transfected EGFP (to account for differences in transfection efficiency), to excitation efficiency of each NIR FP by 640 nm laser, and to emission spectrum of each FP in the emission filter. Effective brightness of miRFP670 was assumed to 100% for each cell line. Error bars, s.d. (n = 3; transfection experiments). f Photobleaching of miRFP670nano and miRFP670 in live HeLa cells. g Mean fluorescence intensity of HeLa cells transiently transfected with miRFP670nano, miRFP703, miRFP670, and EGFP before and after 4 h of incubation with 20 µg ml⁻¹ cycloheximide. Error bars, s.d. (n = 5; transfection experiments). h Mean fluorescence intensity of HeLa cells transiently transfected with miRFP670nano, miRFP703, miRFP670, and EGFP before and after 4 h of incubation with 10 µM bortezomib. Error bars, s.d. (n = 5; transfection experiments). i Tolerance of miRFP670nano to fixation in paraformaldehyde. HeLa cells transfected with miRFP670nano, miRFP670, and miRFP703 were incubated with 4% paraformaldehyde for 10–60 min. The fluorescence of cells treated with paraformaldehyde was normalized to fluorescence of non-fixed cells. Error bars, s.d. (n = 3; transfection experiments)

Fig. 3

Comparison of miRFP670nano, miRFP670, BphP1-FP, and AnPixJ structures and chromophores. a–c Overall structures of a miRFP670nano, b miRFP670 (PDB ID: 5VIV), BphP1-FP (PDB ID: 4XTQ), and c AnPixJ (PDB ID: 3W2Z). The BV and PCB chromophores are in magenta. α1-Helix removed in miRFP670nano is indicated in AnPixJ structure. The PAS and GAF domains of miRFP670 are in cyan and yellow, respectively, and the figure-of-eight knot is indicated. Because of the very similar structures of miRFP670 and BphP1-FP, only the former one is shown. d–g Chromophores (rings A and B only) bound to Cys residues in d miRFP670nano, e miRFP670, f BphP1-FP, and g AnPixJ. Carbon, nitrogen, oxygen, and sulfur atoms are in white, blue, red, and yellow, respectively. Single chromophore species are observed in miRFP670nano and AnPixJ only. Two BV chromophore species are observed in miRFP670 and BphP1-FP. h–m Chemical formulas of the chromophores in h miRFP670nano, i, j miRFP670, k, l BphP1-FP, and m AnPixJ. In miRFP670nano, the BV chromophore (h) is bound to the Cys86 residue via the C3¹ atom. In miRFP670 the BV chromophore (i) is bound via the C3² atom to the Cys253 in the GAF domain, and the BV chromophore (j) is bound via the C3¹ atom to Cys253 in the GAF domain and also via the C3² atom to Cys20 in the PAS domain. In BphP1-FP the BV chromophore (k) is bound via the C3¹ atom to Cys253 in the GAF domain, and the BV chromophore (l) is bound via the C3² atom to Cys253 residue in the GAF domain. In AnPixJ the PCB chromophore (m) is bound to the Cys201 residues via the C3¹ atom

Fig. 4

miRFP670nano fusions imaged using epifluorescence microscopy. Live HeLa cells transfected with the miRFP670nano N- and C-terminal fusion constructs. The C-terminal fusions are a actin; b vesicular protein clathrin; c myosin; d α-tubulin. The N-terminal fusions are e α-actinin; f microtubules-binding EB3; g keratin; h actin-binding LifeAct; i lysosomal membrane glycoprotein LAMP1; j vimentin; k histone H2B. l Cells expressing untagged miRFP670nano. m Dissociated rat cortical neurons transfected with miRFP670nano encoding plasmid at 3 days in vitro (DIV 3). Neurons were imaged 48 h after transfection. Left images are zoom-in of the indicated areas of the right images. n Two-color images of cells co-expressing α-tubulin tagged with miRFP670nano and H2B tagged with miRFP720. o Two-color images of cells co-expressing LAMP1 tagged with miRFP720 and H2B tagged with miRFP670nano. p miRFP670nano internally inserted between the helical and GTPase domains of the G-protein α subunit (Gαs). r miRFP670nano internally inserted into the intracellular loop 3 of the β2 adrenergic receptor (β2AR). mVenus with membrane targeting CAAX motif was used for membrane visualization. Scale bars, 10 μm

Fig. 5

NIR biosensors for detection of PKA and JNK kinase activities. a Schematic representation of miRFP670nano-miRFP720-based NIR FRET biosensor for kinase activity. b Time-lapse FRET/miRFP670nano ratio images of HeLa cell expressing NIR PKA biosensor stimulated with 1 mM dbcAMP and visualized using pseudocolor. c FRET/miRFP670nano ratio time courses of HeLa cells expressing PKA biosensor stimulated with dbcAMP in the presence (red) and absence (black) of chemical PKA inhibitor, AT13148 (n = 3 independent experiments). d Time-lapse FRET/miRFP670nano ratio images of HeLa cell expressing NIR JNK biosensor stimulated with 1 μg ml⁻¹ anisomycin and visualized using pseudocolor. e FRET/miRFP670nano ratio time courses of HeLa cells expressing JNK biosensor stimulated with anisomycin in the presence (red) and absence (black) of chemical JNK inhibitor, SP600125 (n = 3 independent experiments). In b–e the miRFP670nano and FRET fluorescence signals were detected at 667 and 725 nm, respectively. Scale bars, 10 μm

Fig. 6

HeLa cell stably expressing NIR JNK biosensor co-transfected with p38 kinase translocation reporter (p38 KTR). a p38 KTR-EGFP translocation (top row) and FRET/miRFP670nano ratio changes (bottom row) upon stimulation with 1 μg ml⁻¹ anisomycin. Dashed line marks the region used for profile plotting. FRET/miRFP670nano ratio images are visualized using intensity pseudocolor. Scale bar, 10 µm. b Intensity profiles of p38 KTR-EGFP fluorescence before and after stimulation with anisomycin. c Kinetics of FRET/miRFP670nano ratio upon stimulation with anisomycin. The miRFP670nano and FRET fluorescence signals were detected at 667 and 725 nm, respectively

Fig. 7

Multiplexing of NIR PKA and JNK biosensors with optogenetic kinase inhibitors. a Schematic representation of LOV2-domain-based blue-light-regulatable kinase inhibitor in combination with respective fully-NIR kinase biosensor. Upon illumination with blue light, the Jα helix of LOV2 unfolds, resulting in uncaging of a peptide, which inhibits kinase. b HeLa cells stably expressing NIR PKA biosensor co-transfected with optogenetic PKA inhibitor, PA-PKI, tagged with mVenus (top row). Upon simultaneous 460 nm illumination and stimulation with 1 mM dbcAMP, the changes in FRET/miRFP670nano ratio are shown in pseudocolor (bottom row). c FRET/miRFP670nano ratio time courses of HeLa cells expressing NIR PKA biosensor only (red) or NIR PKA biosensor with PA-PKI (green) upon simultaneous 460 nm illumination and stimulation with 1 mM dbcAMP (n = 3 independent experiments). d HeLa cells stably expressing JNK biosensor co-transfected with optogenetic JNK inhibitor, optoJNKi, tagged with EGFP (top row). Upon simultaneous 460 nm illumination and stimulation with 1 μg ml⁻¹ anisomycin, the changes in FRET/miRFP670nano ratio are shown in pseudocolor (bottom row). e FRET/miRFP670nano ratio time courses of HeLa cells expressing NIR JNK biosensor only (red) or NIR JNK biosensor with optoJNKi (green) upon simultaneous 460 nm illumination and stimulation with anisomycin (n = 3 independent experiments). White arrows indicate cells expressing optogenetic regulators. In b–e the miRFP670nano and FRET fluorescence signals were detected at 667 and 725 nm, respectively. Scale bars, 10 μm

Fig. 8

Characterization of miRFP670nano in vivo. a Comparison of miRFP670nano with miRFP670 in vivo. Fluorescence (top row) and bioluminescence (bottom row) images of living mice injected with 3 × 10⁶ HeLa cells expressing miRFP670 (left) and miRFP670nano (right). Cells were co-transfected with Rluc8 (miRFPs:Rluc8 plasmid ratio is 10:1). The fluorescence images were obtained with excitation at 640 nm and emission at 680 nm using IVIS Spectrum instrument 72 h after cell transfection. b Brightness of injected HeLa cells expressing miRFP670 or miRFP670nano as shown in a. Mean fluorescence intensity was normalized to mean bioluminescence intensity. Error bars, s.d. (n = 3 experiments). c Minimal amount of detectable miRFP670nano cells. Fluorescence (top row) and bioluminescence (bottom row) images of living mice injected with various quantity of HeLa cells expressing miRFP670nano. Left mouse was injected with 3 × 10⁶ (left) and 10⁶ (right) cells; middle mouse was injected with 10⁶ (left) and 3 × 10⁵ (right) cells; right mouse was injected with 3 × 10⁵ (left) and 10⁵ (right) cells. Cells were co-transfected with Rluc8 (miRFPs:Rluc8 plasmid ratio is 10:1). The fluorescence images were obtained with excitation at 640 nm and emission at 680 nm 72 h after cell transfection. d Transfection efficiency of injected HeLa cells obtained by FACS analysis. e Two-color imaging of miRFP670nano and miRFP720 in vivo. Fluorescence images of living mice injected with 3 × 10⁶ HeLa cells expressing miRFP670nano (top row) and miRFP720 (middle raw) and its overlay (bottom raw) are shown. The fluorescence images were obtained with excitation at 640 nm and emission at 680 nm for miRFP670nano and with excitation at 675 nm and emission at 720 nm for miRFP720 72 h after cell transfection