Genetically-encoded photosensitizer KillerRed

- Light-induced production of reactive oxygen species
- Direct expression in cells, easy targeting to various subcellular compartments
- No exogenous chemical compounds required for chromophore maturation
- Not toxic before activation by green or orange light irradiation
- Recommended for selective light-induced protein inactivation and cell killing

Performance and use

KillerRed can be used for in vivo killing of selected cells and CALI applications. It can be expressed and induced in various experimental systems, including bacteria, Xenopus, zebrafish, and mammalian cells.

KillerRed is 2000-fold more phototoxic than EGFP; however, it is not as effective as chemical probes. KillerRed-mediated ROS production is accompanied by profound KillerRed photobleaching.

Despite it dimerization capacity, KillerRed demonstrates successful performance in many fusions including that with mitochondrial targeted signal, cytoplasmic β-actin, fibrillarin, dopamine transporter, Tau34, etc.

KillerRed suitability to generate stably transfected cells has been proven by Marinpharm company.

KillerRed expression in mammalian cells. (A) Stably transfected ARPE-19 cells expressing membrane-targeted KillerRed. Image was kindly provided by Dr. Christian Petzelt (Marinpharm); (B) Transiently transfected HeLa cells expressing KillerRed fusion with Tau34.

KillerRed's suitability for light-induced killing of prokaryotic cells has been demonstrated using E. coli XL1-Blue strain. A single E. coli colony expressing KillerRed was picked out, diluted into 1 ml of PBS buffer and divided into two equal portions. One of them was irradiated with white light (1W/cm², light source Fiber-Light from Dolan-Jenner Industries, Inc) for different periods of time, whereas the other was kept in darkness. Both sample aliquots were then plated onto Petri dishes at different dilutions. The number of growing colonies corresponded to the number of bacteria cells surviving after irradiation (i.e. colony forming units, CFU). CFU number for the irradiated E. coli portion was compared with non-irradiated one, thus allowing estimation of the relative phototoxic effect for KillerRed. In control experiments, E. coli cells expressing different fluorescent and non-fluorescent proteins were used.

KillerRed killed 96% of bacterial cells after 10 min and almost 100% of cells after 20 min of irradiation with white light.

		KillerRed-mediated killing of bacteria cells upon irradiation with different wavelength. Left graph: Blue and green rectangles show relative phototoxic effect from irradiation with blue (460-490 nm) and green (540-580 nm) light of 35 mW/cm2. Numbers above the rectangles represent decrease in viable E. coli cells after 30-min irradiation (folds). Right graph: Time-course of white light-induced (1 W/cm2) killing of E. coli expressing KillerRed.

KillerRed-mediated killing of eukaryotic cells: The following two ways have been found to be effective for killing the eukaryotic cells using KillerRed: (1) via an apoptotic pathway using KillerRed targeted to mitochondria, and (2) via membrane lipid oxidation using membrane-localized KillerRed. Irradiation of KillerRed localized in cell cytosol has a weak effect on cell survival.

Effects of KillerRed localized in mitochondria: Mitochondrial localization increases the phototoxic effect of photosensitizers (primarily by provoking the apoptosis). Use of KillerRed targeted to mitochondria allows effective cell killing through an apoptotic pathway as has been demonstrated in the following experiments:
HeLa cells expressing cytoplasmic TurboGFP and mitochondria-localized KillerRed-dMito were generated. 10-min (515-560 nm excitation filter, 7 W/cm²) irradiation of selected cells with green light resulted in profound KillerRed photobleaching. 60 min after irradiation, cells had an abnormal shape and "bubbles" typical of apoptotic pathway. These cells disrupted within the next 30-60 min.

Confocal image of HeLa cells expressing KillerRed in mitochondria (red) and TurboGFP in cytosol (green). Lower left cells were pre-irradiated with green light (515-560 nm, 7W/cm2) for ten minutes. This led to profound KillerRed photobleaching. Cells are shown in 60 minutes after irradiation. It is clearly visible that irradiated cells have abnormal shape and "bubbles", characteristic for apoptotic cell-death pathway.

In another experiment, nearly 100% of B16 melanoma cells expressing KillerRed targeted to mitochondria died within 45 min after 15-min of irradiation (40x objective, 535-575 nm excitation filter, 3.3 W/cm²). It is noticeable that when preincubated with the pancaspase inhibitor zVAD-fmk (10 μM), the cells were resistant to the same light exposure and held their shape for at least 1.5 hours after illumination.

Apart from the immediate phototoxic effect, photosensitizers can mediate postponed cellular responses such as cell growth arrest or cell death via long-term apoptotic mechanism:
B16 melanoma cells expressing mitochondria-targeted KillerRed or EGFP were mixed together and irradiated by green light of low intensity (3.7x objective, 535-575 nm excitation filter, 115 mW/cm²) for 45 min [Bulina et al., 2006a]. No red fluorescent cells were observed in 16 hrs after irradiation, whereas green fluorescent cells remained viable. It confirms that mitochondria-localized KillerRed can mediate cell death through long-term mechanisms in response to low-intensity illumination. This effect can be used in different applications.

Effects of KillerRed localized in membrane: Comparing with the mitochondria-targeted KillerRed, irradiation of membrane-localized KillerRed causes a more effective and fast cell death within 10-30 min, presumably because of lipid oxidation [Bulina et al., 2006b].

Light-induced killing of HeLa cell using membrane-targeted KillerRed (movie). First image: intact cell features red fluorescent cellular membrane and green fluorescent cytosol due to membrane-targeted KillerRed and TurboGFP expression, respectively. Further images: 40 minute time-course of cell fragmentation induced by 10-min green light (laser 543 nm, 300 mW/cm2) irradiation. Movie from Bulina et al., 2006b.

Moreover, membrane-targeted KillerRed was shown to be an effective tool for the light-induced cell killing within a developing zebrafish. Zebrafish embryo was microinjected with a mixture of vectors driving expression of membrane-targeted KillerRed and a green fluorescent protein at the single- cell stage. A muscle cell expressing both proteins was irradiated with green light (40x objective, TRITC filter set, 10 min) at 48 hrs after fertilization. By the end of 10-min irradiation, the cell already started to change its shape. Within 20 min after irradiation was stopped, the cell was disrupted completely. Mitochondria targeted KillerRed was shown to be of low efficiency in similar experiments.

Light-induced killing of a muscle cell within a developing zebrafish embryo. Left image: A region expressing membrane-targeted KillerRed and green fluorescent marker; right image: time course of light-induced killing of a muscle cell within a developing zebrafish. Fluorescence was collected using standard FITC and TRITC filter sets.

KillerRed use for CALI applications: KillerRed use for CALI application has been demonstrated on the model of β-galactosidase inactivation in bacterial cells and inactivation of pleckstrin homology (PH) domain of phospholipase C delta-1 (PLC delta-1) in mammalian cells. In the first experiment, KillerRed was fused to β-galactosidase (β-gal) enzyme and expressed in E. coli. Upon green light irradiation (540-580 nm, 30 min, 360 mW/cm²) β-gal activity was effectively suppressed in living E. coli streaks. On the contrary, no effect of green light on the enzyme activity was detected in control cells containing unmodified β-gal gene. In vitro test showed that in E. coli cell extract β-gal fused to KillerRed lost 99.4% of enzymatic activity within 25 min of white light exposure (1W/cm²), with half inactivation time of about 5 min. Irradiation of E. coli extracts containing unfused β-gal protein alone or β-gal mixed with KillerRed had no effect on enzyme activity. To verify specificity of KillerRed phototoxic effect, horse-radish peroxidase (HRP) was added to the sample. Upon 15 min of illumination (white light, 1W/cm²) only 2% of HRP activity was lost, showing high specificity of the phototoxic effect.

	Time-course of CALI of β-galactosidase. White light exposure (1W/cm2) of β-gal fused to KillerRed (E. coli cell extract).

In the second experiment, a triple EGFP-PH-KillerRed fusion protein that allows both protein visualization and CALI was transiently expressed in mammalian cell line. Intracellular localization of EGFP signal was evaluated before and after CALI of the PH domain using confocal and fluorescence microscopy. In intact cells, the fusion is located predominantly at the plasma membrane because of specific affinity of PH domain to phosphatidylinositol 4,5-bisphosphate.

Irradiation with intense green light led to KillerRed-mediated ROS production, PH domain damage, and fusion protein dissociation from the membrane. After 10 sec of green light irradiation (63x objective, mercury lamp, 515-560 nm filter, 7W/cm²), translocation of the PH domain into cytosol was clearly visible. When irradiated for a longer period of time, considerable amount of PH domain translocated into cytosol, increasing the cytoplasm-to-membrane green fluorescent signal ratio to 0.5-0.9.

In the negative control experiments, the cellular location of a DsRed-Express (Clontech) containing construct, GFP-PH-DsRed-Express, showed no dependence on green light irradiation. Similarly, no detectable CALI of the PH domain was achieved when KillerRed was expressed in the cell separately from PH domain, in either the membrane or cytosol.

KillerRed-mediated light-induced inactivation of PLC delta-1 PH domain. (A) Schematic outline of the experimental system; (B, C) confocal images of a cell expressing EGFP-PH-KillerRed triple fusion (EGFP green fluorescent signal) before (B) and after (C) 10-s irradiation with green light (63x objective, mercury lamp, 515-560 nm filter, 7W/cm2). Note considerable increase in cytoplasmic signal.