A genetically encoded photosensitizer - Nature

© 2006 Nature Publishing Group http://www.nature.com/naturebiotechnology

LETTERS

A genetically encoded photosensitizer Maria E Bulina1,3, Dmitriy M Chudakov1,3, Olga V Britanova1, Yurii G Yanushevich1, Dmitry B Staroverov2, Tatyana V Chepurnykh2, Ekaterina M Merzlyak2, Maria A Shkrob1, Sergey Lukyanov1 & Konstantin A Lukyanov1 Photosensitizers are chromophores that generate reactive oxygen species (ROS) upon light irradiation1. They are used for inactivation of specific proteins by chromophoreassisted light inactivation (CALI) and for light-induced cell killing in photodynamic therapy. Here we report a genetically encoded photosensitizer, which we call KillerRed, developed from the hydrozoan chromoprotein anm2CP, a homolog of green fluorescent protein (GFP). KillerRed generates ROS upon irradiation with green light. Whereas known photosensitizers must be added to living systems exogenously, KillerRed is fully genetically encoded. We demonstrate the utility of KillerRed for light-induced killing of Escherichia coli and eukaryotic cells and for inactivating fusions to b-galactosidase and phospholipase Cd1 pleckstrin homology domain. In CALI, mild illumination of a photosensitizer-tagged molecule for a limited time inactivates a specific target2–4, whereas neighboring molecules remain intact. CALI has been successfully applied in functional studies of various proteins5–8. Because photosensitizermediated ROS production on a large scale damages tissues9–11, photosensitizers that accumulate in tumors have been used in photodynamic cancer therapy12. All known photosensitizers, including genetically targeted FlAsH13 and ReAsH14 (both of which involve a tetracysteine motif and a biarsenical dye15–17) require that a compound be added to living systems exogenously. This limitation has prompted the search for a fully genetically encoded photosensitizer. Although GFP from Aequorea victoria and its homologs are genetically encoded, GFP is an inefficient photosensitizer4,7,17, presumably because the protein shell screens the chromophore and prevents ROS generation. However, GFP can generate singlet oxygen18, albeit ineffectively. Prolonged visualization of GFP-expressing cells can result in physiological changes and eventually cell death19. Although GFP has been used for CALI in certain systems7, the ability of GFP homologs to produce ROS in response to irradiation has apparently not been studied. To identify an efficient photosensitizer, we screened a collection of GFP homologs for phototoxic effects on E. coli cells. Proteins tested were representatives of distinct spectral classes derived from anthozoan, hydrozoan and copepod species, and included fluorescent

variants of different colors, nonfluorescent chromoproteins and circularly permuted variants. The majority of proteins tested had little or no effect on the viability of bacterial cells (Fig. 1). Mutant W94F of the chromoprotein asulCP from Anemonia sulcata showed a weak phototoxic effect, resulting in decreased bacterial survival. Only one of the proteins, KillerRed, showed a strong phototoxic effect (Fig. 1a,b). KillerRed (GenBank accession number AY969116) is a dimeric red fluorescent protein with fluorescence excitation/emission maxima at 585/610 nm, an extinction coefficient of 45,000 M–1cm–1 at 585 nm and a fluorescence quantum yield of 0.25 (Fig. 1c). We derived it from the hydrozoan chromoprotein anm2CP20 by including the substitutions T145N and C161A (corresponding to positions 148 and 165 in Aequorea victoria GFP), which are spatially close to the chromophore and drastically affect protein fluorescent properties, as well as the folding- and brightness-improving mutations V21E, D24G, V30I, K73R, N76D, T91S, G114D, I118V, K136R, S154P, Y162F, S175G, L181M, K190E, V199I, I201T, K205R and T220A. Both Asn145 and Ala161 appeared to be indispensable for KillerRed’s phototoxic effect, although other substitutions are also important (see Supplementary Data online for mutational analysis of KillerRed phototoxic properties). We characterized KillerRed phototoxic activity as a function of irradiation time. KillerRed kills 96% of E. coli cells after 10 min and almost all cells after 20 min of irradiation with white light (1 W/cm2) (Fig. 1b). We also verified that cell killing depends on the wavelength of the incident light. KillerRed demonstrated a strong cell killing effect in green light (540–580 nm) and almost no effect upon irradiation with blue light (460–490 nm) of equal intensity (35 mW/cm2). This correlates well with the fluorescence excitation and absorption spectra of KillerRed (Fig. 1c, Supplementary Fig. 1a online) and suggests that the KillerRed red chromophore mediates the phototoxic effect. We also tested the ability of KillerRed to kill cultured eukaryotic cells in a light-mediated manner. To investigate the effects of cytosolic KillerRed, we irradiated 293T human kidney cells that transiently expressed KillerRed with intense green light for 10 min (100 objective, 535–575 nm excitation filter, 5.8 W/cm2). Death of irradiated KillerRed-expressing cells reached 40–60%, depending on the protein concentration (Fig. 2a). No phototoxic effect was observed for DsRed2 in identical control experiments.

1Institute 3These

of Bioorganic Chemistry RAS, Miklukho-Maklaya 16/10, Moscow 117997, Russia. 2Evrogen JSC, Miklukho-Maklaya 16/10, Moscow 117997, Russia. authors contributed equally to the work. Correspondence should be addressed to K.A.L. ([email protected]).

Received 6 January; accepted 8 November; published online 20 December 2005; doi:10.1038/nbt1175

NATURE BIOTECHNOLOGY

VOLUME 24

NUMBER 1

JANUARY 2006

95

LETTERS

b

c 100

80

80

21.2

585 nm 610 nm

DsRed

40 20

KillerRed

60 40 20 0

0 1

3 2

5

4

7 9 11 13 15 6 8 10 12 14 Sample number

d 100 ONP absorbance

10

zFP506-Y66A

60

Fluorescence

100

1

1.0

80 60 40 20 0

0

10 20 30 40 50 Irradiation time (min)

400 450 500 550 600 650 700 Wavelength (nm)

60

0

e

5 10 15 20 Irradiation time (min)

25

1.4

1.2 Figure 1 Identification and characterization of the phototoxic properties of KillerRed. (a) Decrease of viable 1 bacterial counts after white light illumination of E. coli expressing the following GFP-like proteins: 1. AmCyan (Clontech); 2. ZsGreen (Clontech); 3. DsRed2 (Clontech); 4. asulCP, kindling fluorescent protein from Anemonia 0.8 sulcata35; 5. EGFP (Clontech); 6. AsRed2 (Clontech); 7. asulCP-W94F, asulCP red fluorescent mutant35; 0.6 8. dendGFP, green-to-red photoconvertible fluorescent protein from Dendronephthia sp.36; 9. gtenCP, 0.4 nonfluorescent chromoprotein from Goniopora tenuidens37; 10. AcGFP1 (Clontech); 11,12. Circularly permuted 0.2 AcGFP1 variants, protein backbone gap at positions 144–148 (unpublished data); 13. anm2CP, chromoprotein 0 20 Buffer Azide, SOD + Mannitol, from an unidentified anthomedusa ; 14. KillerRed; 15. zFP506-Y66A, noncolored, nonfluorescent ZsGreen 33 mM catalase 66 mM mutant, which doesn’t form a chromophore. (b) Time-course of light-induced killing of E. coli expressing zFP506-Y66A, DsRed2 or KillerRed. (c) Excitation (black line) and emission (red line) spectra for KillerRed. 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 fold decrease in viable cells after 30-min irradiation. (d) Time course of CALI of b-gal. KillerRed-b-gal fusion was irradiated for various time periods and b-gal activity was measured by O-nitrophenyl (ONP) absorption after 15 min O-nitrophenyl-D galactoside (ONPG) incubation. (e) Quenching of CALI effect in b-gal assay. Light and dark gray columns show b-gal activity for the intact and irradiated samples, respectively. ONP absorbance


100

1,000 Viable cells (%)

CFU decrease (fold)

a

Mitochondrial localization of photosensitizers is known to be effective in photodynamic therapy, resulting primarily in apoptotic cell death9. To increase the phototoxic effect, we targeted KillerRed to mitochondria by cloning it in-frame with two tandem copies of an N-terminal mitochondrial localization signal (2MLS)21. Exposure of B16 melanoma cells expressing 2MLS-KillerRed to green light for 15 min using less intense irradiation (40 objective, 535–575 nm excitation filter, 3.3 W/cm2) killed almost all cells within 45 min after irradiation (Fig. 2b). When preincubated with the pancaspase inhibitor zVAD-fmk (10 mM), the cells survived an identical irradiation treatment and preserved their native shapes for at least 1.5 h after illumination. This indicates that 2MLS-KillerRed-mediated cell death proceeds via apoptosis, but not necrosis (necrosis is insensitive to this pancaspase inhibitor). Apart from the immediate phototoxic effect, photosensitizers can mediate postponed cellular responses, such as cell growth arrest or cell death via a long-term apoptotic mechanism9. Owing to the lower

Figure 2 Light-induced killing of eukaryotic cells expressing KillerRed. (a) Cells coexpressing KillerRed and AcGFP1 were irradiated with green light for 10 min (TRITC filter set, 100 objective). The illumination resulted in profound KillerRed photobleaching, while cell fate still could be monitored by AcGFP1 green fluorescence. In the typical experiment shown, the cell of interest characterized with high expression levels of both AcGFP1 and KillerRed, is located near the center of the field. This cell started to change shape several minutes after irradiation and disintegrated within 1.5 h, indicating KillerRed phototoxicity. At the same time, neighboring cells with lower KillerRed expression level survived irradiation, presumably because of the lower concentration of the phototoxic agent. First frame shows KillerRed red fluorescent signal (TRITC filter set) before irradiation. Further AcGFP1 green fluorescence (FITC filter set) is shown, images taken once per 15 min. Time zero is set immediately after irradiation with green light. (b) Cells expressing 2MLS-KillerRed were irradiated with green light for 15 min (TRITC filter set, 40 objective). First frame shows KillerRed red fluorescent signal (TRITC filter set) before irradiation. Further cells are shown in the white field, images taken once per 15 min. Time zero is set immediately after irradiation with green light.

96

irradiation intensity required, such applications of photosensitizers may be significant in a medical context. We used mixed populations of cells expressing mitochondrially targeted KillerRed or enhanced (E)GFP to monitor long-term cell survival. Mixtures of cells were irradiated with 30-fold lower intensity green light (3.7 objective, 535–575 nm excitation filter, 115 mW/cm2) for 45 min. Sixteen hours after irradiation, no red fluorescent cells were observed, whereas the green fluorescent cells remained viable. This experiment showed that mitochondrially localized KillerRed can mediate cell death through long-term mechanisms in response to irradiation by low-light intensities that are feasible for use in photodynamic therapy22. We compared the phototoxicities of mitochondrial KillerRed and tetramethylrhodamine (TMRM), a red fluorescent dye capable of producing ROS23 in mitochondria. Cells were preincubated with 0

a

45

b

VOLUME 24

15

60 0

NUMBER 1

30

75 15

30

JANUARY 2006

90 45

60


LETTERS (PLC d1) as a model protein. The PH domain locates to the inner leaflet of the plasma 0.8 Membrane membrane as a result of its high affinity for KillerRed Cytoplasm PH membrane phospholipids24–26. One anticiROS 0.6 domain pates that in the event of direct protein N C inactivation by dye-generated ROS, the PH 0.4 N C DsRed domain will lose its membrane affinity and 0.2 become evenly distributed throughout the KillerRed EGFP cell. We constructed a triple EGFP-PH-Killhν 0.0 300 250 200 0 50 100 150 erRed fusion protein that allows both protein Irradiation time (s) visualization and CALI (Fig. 3a). EGFP-PHKillerRed was transiently expressed in a c d e f g mammalian cell line. Intracellular localization of the EGFP signal was evaluated before and after CALI of the PH domain using confocal and fluorescence microscopy. As expected, light-induced damage of the PH domain Figure 3 KillerRed-mediated light-induced inactivation of the PLC d1 PH domain. (a) Schematic affected its membrane affinity dramatically outline of the experimental system. Normally, the EGFP-PH-KillerRed triple fusion protein is localized (Fig. 3b–e). In resting cells, most of the predominantly at the plasma membrane as a result of the specific affinity of the PH domain for fluorescent signal located to the cell memphosphatidylinositol 4,5-bisphosphate. Irradiation with intense green light leads to ROS generation by KillerRed and damage of the adjacent PH domain. As a result, the fusion protein dissociates from the brane, with the cytoplasm:membrane EGFP membrane. (b) Dependence of EGFP-PH-KillerRed (magenta circles) and EGFP-PH-DsRedExpress (red signal intensity ratio being B0.2. Translocasquares) membrane-to-cytoplasm translocation on the irradiation time. (c,d,e) A confocal image of a tion of the PH domain into the cytoplasm cell expressing EGFP-PH-KillerRed triple fusion (EGFP green fluorescent signal) before (c), after (d) and was clearly visible after 10 s of irradiation 1.5 h after (e) 10-s irradiation with green light. Note considerable increase in cytoplasmic signal. with green light (63 objective, mercury (f,g) Control experiment showing cell expressing EGFP-PH-DsRedExpress triple fusion (EGFP green lamp, 515–560 nm filter, 7 W/cm2) fluorescent signal) before (f) and after (g) 1-min irradiation with green light. No change in signal distribution within the cell was observed. (Fig. 3b,d). If irradiated for longer times, a considerable amount of the PH domain translocated into the cytoplasm, increasing 100 nM TMRM for 20 min at 37 1C. We applied 3.3 W/cm2 green the cytoplasm-to-membrane signal intensity ratio to 0.5–0.9 (Fig. 3b). light irradiation (40 objective, 535–575 nm) to the TMRM-loaded In the negative control experiments, the cellular location of a DsRedcells for 10 s, 1 min, 5 min and longer. Although irradiation for 10 s containing construct, GFP-PH-DsRedExpress, showed no dependence did not kill TMRM-loaded cells within 1.5 h, irradiation for 5 min or on green light irradiation (Fig. 3b,f,g). Even prolonged exposure of longer resulted in immediate cell death. Irradiation for 1 min led to cells to green light for times and under intensities sufficient to cause cell death within 30–40 min. The effect was similar to that obtained by nearly complete bleaching of DsRed fluorescence did not affect 15-min irradiation of the KillerRed-expressing cells under the same membrane tagging of the green signal. Similarly, no detectable CALI conditions. Based on this experiment, we estimate that the mito- of the PH domain could be achieved when KillerRed was expressed in chondrially localized KillerRed generated by expression of the cyto- the cell separately from the PH domain, in either the membrane or the megalovirus immediate-early promoter exerts about 7% of the cytosol. These controls indicate that KillerRed was acting in cis against the fused PH domain. phototoxic effect of TMRM. Experiments to assess cell killing using the PH-CALI system To determine whether KillerRed is suitable for CALI, we tested it against two model target proteins in prokaryotic in vitro and in vivo indicated that a single 10-s illumination reproducibly resulted in systems and in eukaryotic cells. First, we assayed the ability of effective CALI, whereas monitoring cell shape and morphology for KillerRed to inactivate b-galactosidase (b-gal), as this enzyme was the subsequent 1.5 h revealed no changes associated with cell death used to establish the CALI technique4. Upon green light irradiation (Fig. 3e). Cells irradiated for 1 to 5 min also survived for at least 1.5 h, (540–580 nm, 30 min, 360 mW/cm2), b-gal fused to KillerRed was although 5 min continuous irradiation occasionally affected cell inactivated in E. coli. On the contrary, no effect of green light on the morphology. Longer irradiation intervals led to considerable changes activity of the enzyme was detected in control cells containing the in cell morphology and to lethality. Generally, irradiation time unmodified b-gal gene. We then measured CALI-mediated loss of required for cell death correlated with the level of expression of the b-gal activity as a function of time (Fig. 1d). An E. coli cell extract with KillerRed fusion, whereas the efficiency of CALI did not. In our model b-gal fused to KillerRed lost 99.4% of the enzymatic activity within with highly overexpressed EGFP-PH-KillerRed, we observed approxi25 min of exposure to white light (1 W/cm2), with an inactivation mately 10- to 20-fold differences between the light doses required for half-time t1/2 of B5 min. Irradiation of E. coli extracts containing CALI and for cell death. We anticipate that this difference will be even unfused b-gal protein alone or b-gal mixed with KillerRed had no greater for KillerRed fusion proteins expressed at low levels. To reveal which ROS mediates KillerRed’s phototoxicity, we effect on the enzyme activity. To verify the specificity of the KillerRed phototoxic effect, we added horseradish peroxidase (HRP) to the first examined CALI in eukaryotic cells. Oxygen partial pressure is sample. Upon 15 min illumination (white light, 1 W/cm2) only known to affect the efficiency of light-induced, oxygen-based ROS 2% of HRP activity was lost, showing the high specificity of the generation17. In our previous CALI experiments, mammalian cells phototoxic effect. were exposed to normal oxygen pressure. When cells were vacuum We further tested KillerRed-mediated CALI in mammalian cells sealed and the oxygen pressure was allowed to decrease as a conusing the pleckstrin homology (PH) domain of phospholipase C d1 sequence of cell breathing, CALI of the PH domain was impeded. The

b

1.0


Cytoplasm/membrane fluorescence ratio

a


VOLUME 24

NUMBER 1

JANUARY 2006

97


LETTERS effect could be reversed by restoring oxygen. Therefore, KillerRed generates oxygen-based ROS upon irradiation. If the mediator was a hydroxyl radical, oxygen partial pressure would have little effect on the CALI ratio1. We then assayed production of singlet oxygen and superoxide, the two primary oxygen-derived ROS27. Direct ROS measurement using chemical probes indicated that KillerRed is capable of both singlet oxygen and superoxide generation upon irradiation with green light (see Supplementary Data and Supplementary Fig. 1 online). We further probed the mechanism of KillerRed-mediated effect by adding ROS quenchers to the in vitro b-gal CALI system described above (Fig. 1e). In combination with catalase (500 u/ml), the superoxide-specific quencher superoxide dismutase (500 u/ml) impeded CALI slightly. A similar effect was observed when mannitol (67 mM), a superoxide anion and hydroxyl radical quencher, was added to the sample. Finally, sodium azide (20 mM), a strong specific quencher of singlet oxygen, significantly reduced CALI of b-gal. The quenching data suggest that singlet oxygen is the primary damaging agent generated by KillerRed. To verify the role of singlet oxygen, we studied CALI of PH-domain experiments in the presence of quenchers of singlet oxygen (30 mM imidazole)17 or superoxide anions (100 mM superoxide dismutase mimetic Mn-TBAP28). Whereas Mn-TBAP had no visible influence on CALI effectiveness, imidazole suppressed translocation of the PH domain, confirming the dominant role of singlet oxygen in mediating CALI. Genetically encoded photosensitizers such as KillerRed will likely be used in a variety of applications. First, the possibility of control by a specific promoter provides a unique opportunity to investigate cell fate in developing and adult organisms by spatially and temporally controlled cell killing. Second, genetically encoded photosensitizers open new perspectives for photodynamic therapy. Recent studies on nude mice demonstrated accumulation of GFP-expressing bacteria or viruses within various tumors29. Bacteria or viruses expressing a genetically encoded photosensitizer might also allow light-induced killing of tumor cells. Finally, genetically encoded photosensitizers should offer new opportunities for CALI of specific molecules in living cells. Recent work demonstrated that CALI with organic dyes is suitable to identify membrane proteins involved in metastasis30. Use of a genetically encoded photosensitizer extends the use of CALI beyond surfaceexposed proteins to intracellular proteins. Potentially, a genetically encoded photosensitizer could be fused to collections of open reading frames, allowing high-throughput analysis of protein function in living cells. Nucleic acids could be targeted specifically by fusing a genetically encoded photosensitizer to an RNA- or DNA-binding domain, similar to the well-established labeling of individual RNA molecules and genomic loci with GFP31. If genetically encoded photosensitizer activation induces breaks or damage in the nucleic acid strand, it should be possible to inactivate expression of target genes transiently or even permanently by a pulse of light. KillerRed induces CALI far more effectively than GFP. Whereas a fusion of b-gal to KillerRed completely loses enzymatic activity upon 25 min of irradiation with white light (1 W/cm2), analogous experiments involving b-gal–GFP fusions resulted in incomplete (up to 40% at best) inactivation upon sevenfold more powerful irradiation4. In contrast to genetically targeted photosensitizers, such as FlAsH and ReAsH13,15,17, KillerRed does not require an exogenous supply of chemical compounds and can be used in stably transformed cell lines and transgenic animals with comparable efficiency of CALI. ReAsH irradiation at 17 W/cm2 for 25 s was required for 95% reduction in

98

protein function17 versus about 21 min at 1 W/cm2 to achieve the same effect for KillerRed. Thus, KillerRed is only threefold less efficient than ReAsH in terms of the light dose consumed. Moreover, ReAsH produces up to 21% nonspecific proteins inactivation17, whereas KillerRed generates only 2% nonspecific inactivation. We discovered the photosensitizer properties of KillerRed by testing various fluorescent proteins and chromoproteins for phototoxic effects in bacteria. Most probably, genetically encoded photosensitizers that are even more efficient than KillerRed could be discovered by additional mutagenesis and screening. METHODS Screening GFP-like proteins for phototoxic effects in E. coli. E. coli cells (XL1-Blue, Invitrogen) were transformed with the pQE30 vector encoding the corresponding fluorescent protein or chromoprotein gene. A mutant of Zoanthus sp. fluorescent protein (zFP506) containing the chromophoreeliminating mutation Y66A was used as a negative control. Protein-expressing cells were obtained from a single E. coli colony, diluted into 1 ml of PBS buffer and divided into two equal portions. One of them was irradiated with white light (1 W/cm2, light source Fiber-Light from Dolan-Jenner Industries) for 1 h, 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 (that is, colony forming units, CFU). CFU number for the irradiated E. coli portion was compared to the nonirradiated one, thus allowing estimation of the relative phototoxic effects for each of the proteins tested. Singlet oxygen measurement. Singlet oxygen measurement was conducted using trans-1-(2¢-Methoxyvinyl)pyrene (MVP) from Molecular Probes as described in http://www.probes.com/handbook/sections/1802.html and in references32,33. Protein samples (5-10 ml) were placed into 1 ml of measurement buffer (20 mM Tris-HCl, pH 7.3, 100 mM SDS, 16 mM MVP) and irradiated for 60 s (520-600 nm filter, 0.7 W/cm2, light source Fiber-Light) in a quartz cuvette. The luminescence of pyrene-1-carboxaldehyde was measured at 470– 490 nm within the first 2 min after irradiation in the Varian Cary Eclipse Fluorescence Spectrophotometer. Alternatively, 20 ml of protein sample containing 10 mM MVP was diluted in 980 ml of buffer and irradiated for 10, 20 or 30 min. Fluorescence emission spectra upon excitation at 366 nm were measured. The degree of MVP conversion to the pyrene-1-carboxaldehyde was estimated by the decrease of the emission peak at 420 nm and the increase of the emission peak at 460 nm. Superoxide measurement. Superoxide measurement was conducted using Amplex UltraRed Reagent (Molecular Probes). Superoxide dismutase (4 units/ml, Sigma) was added to each sample to convert superoxide to hydrogen peroxide (H2O2). In the presence of horseradish peroxidase (Sigma), nonfluorescent Amplex UltraRed (33.3 mM) reacted with H2O2 in a 1:1 stoichiometric ratio to produce fluorescent reaction product (excitation/ emission maxima B568/581 nm), as described in http://www.probes.com/ handbook/sections/1802.html. The increase in amplex UltraRed fluorescence upon continuous irradiation with green light was measured using a Varian Cary Eclipse Fluorescence Spectrophotometer. Intrinsic protein fluorescence was subtracted as background. In a control experiment, superoxide dismutase was not added to the mix. This resulted in a very small Amplex UltraRed fluorescence increase upon irradiation, confirming superoxide production by the bleached KillerRed. Gene construction. DNA encoding the a-fragment of b-gal (MTMITP SAQLTLTKGNKSWESYYARSLAVVLQRRDWENPGVTQLNRLAAHPPFASW RNSEEARTDRPSQQLRSLNGE), derived from a pBlueScript plasmid, was linked by PCR to the KillerRed-encoding DNA and cloned into PQE30 plasmid, using BamHI and HindIII cloning sites. pQE30 plasmid with an insert of a-fragment of b-gal alone was used as a negative control. To demonstrate CALI on the PLC d1 PH domain, the 224 N-terminal amino acids of PLC d1 were amplified from Fetus Brain cDNA (Clontech) using the PCR primers 5¢-CCCCTCGAGCTATGGACTCGGGCCGGGACTT-3¢ and

VOLUME 24

NUMBER 1

JANUARY 2006


LETTERS


5¢-CCCGGATCCTCATGGGCCCGCGGCCTCGGCGA-3¢, and the fragment was cloned into pEGFP-Tub plasmid (Clontech) in frame with EGFP using XhoI and BamHI restriction sites (underlined). Alternatively, PH domain was PCR-amplified with 5¢-CCCCTCGAGCTATGGACTCGGGCCGGGACTT-3¢ and 5¢-CCCGGTACCTGGGCCCGCGGCCTCGGCGA-3¢ primers and subcloned simultaneously with KillerRed or DsRed-Express genes, using XhoI, KpnI and BamHI restriction sites (underlined). Chromophore-assisted light inactivation of b-gal. Individual E. coli colonies (XL1-Blue) expressing the a-fragment of b-gal or a fusion of the a-fragment to KillerRed were streaked on Petri dishes and grown at 37 1C overnight. The resulting streaks were transferred onto a Hybond-C membrane and illuminated with green light (540–580 nm, 360 mW/cm2) for 10 min using a SZX-12 fluorescence stereomicroscope (Olympus). Membranes were then wetted (PBS/ MgSO4 buffer with 1 mg/ml MgSO4 and 1 mg/ml X-Gal substrate) and incubated at 37 1C for 30 min. Blue coloration, corresponding to the absorbance maximum of cleaved substrate, was evaluated for irradiated and nonirradiated cells. We quantified the effect of KillerRed on the activity of b-gal disrupting ultrasonically XL1-Blue cells expressing a fusion of the a-fragment to KillerRed in PBS buffer and analyzing 50-ml aliquots of clarified lysate. Aliquots were either incubated under white light (1 W/cm2, light source Fiber-Light) or kept in the dark. The entire volume of the tube was illuminated. b-gal activity was assayed in light-illuminated and ‘dark’ aliquot colorimetrically. The activity of samples was measured by an orthonitrophenyl galactoside assay as described previously34. Chromophore-assisted light inactivation of PH domain. In PH domain inactivation experiments, the 293T cell-derived Phoenix Eco cell line was transiently transfected with constructs described above. Fluorescence was monitored using a DM IRE2 TCS SP2 confocal microscope (Leica). Green fluorescent signal was acquired at excitation of 488 nm (laser line) and collected in the 492–538 nm wavelength range. The sample was illuminated with green light (515–560 nm, mercury lamp source, 63 magnification) to initiate CALI of PH domain. Note: Supplementary information is available on the Nature Biotechnology website.

ACKNOWLEDGMENTS We are grateful to Alexei V. Feofanov for valuable advice and help in light intensity measurements. This work was supported by Howard Hughes Medical Institute grant HHMI 55005618, the Russian Academy of Sciences for the program ‘‘Molecular and Cell Biology’’ and the EC FP-6 Integrated Project LSHG-CT-2003-503259. COMPETING INTERESTS STATEMENT The authors declare that they have no competing financial interests. Published online at http://www.nature.com/naturebiotechnology/ Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions/ 1. Liao, J.C., Roider, J. & Jay, D.G. Chromophore-assisted laser inactivation of proteins is mediated by the photogeneration of free radicals. Proc. Natl. Acad. Sci. USA 91, 2659–2663 (1994). 2. Beermann, A.E. & Jay, D.G. Chromophore-assisted laser inactivation of cellular proteins. Methods. Cell. Biol. 44, 715–732 (1994). 3. Jay, D.G. Selective destruction of protein function by chromophore-assisted laser inactivation. Proc. Natl. Acad. Sci. USA 85, 5454–5458 (1988). 4. Surrey, T. et al. Chromophore-assisted light inactivation and self-organization of microtubules and motors. Proc. Natl. Acad. Sci. USA 95, 4293–4298 (1998). 5. Wong, E.V., David, S., Jacob, M.H. & Jay, D.G. Inactivation of myelin-associated glycoprotein enhances optic nerve regeneration. J. Neurosci. 23, 3112–3117 (2003). 6. Rubenwolf, S. et al. Functional proteomics using chromophore-assisted laser inactivation. Proteomics 2, 241–246 (2002). 7. Rajfur, Z., Roy, P., Otey, C., Romer, L. & Jacobson, K. Dissecting the link between stress fibres and focal adhesions by CALI with EGFP fusion proteins. Nat. Cell Biol. 4, 286–293 (2002).


VOLUME 24

NUMBER 1

JANUARY 2006

8. Hauptschein, R.S. et al. Functional proteomic screen identifies a modulating role for CD44 in death receptor-mediated apoptosis. Cancer Res. 65, 1887–1896 (2005). 9. Moor, A.C. Signaling pathways in cell death and survival after photodynamic therapy. J. Photochem. Photobiol. B 57, 1–13 (2000). 10. Plaetzer, K., Kiesslich, T., Krammer, B. & Hammerl, P. Characterization of the cell death modes and the associated changes in cellular energy supply in response to AlPcS4-PDT. Photochem. Photobiol. Sci. 1, 172–177 (2002). 11. Cho, K.S., Lee, E.H., Choi, J.S. & Joo, C.K. Reactive oxygen species-induced apoptosis and necrosis in bovine corneal endothelial cells. Invest. Ophthalmol. Vis. Sci. 40, 911–919 (1999). 12. Calin, M.A., Gruia, M., Herascu, N. & Coman, T. The monitoring of the accumulation of protoporphyrin IX in Walker tumours by subcutaneous administration of delta-aminolevulinic acid. J. Exp. Ther. Oncol. 4, 247–251 (2004). 13. Griffin, B.A., Adams, S.R. & Tsien, R.Y. Specific covalent labeling of recombinant protein molecules inside live cells. Science 281, 269–272 (1998). 14. Adams, S.R. et al. New biarsenical ligands and tetracysteine motifs for protein labeling in vitro and in vivo: synthesis and biological applications. J. Am. Chem. Soc. 124, 6063–6076 (2002). 15. Marek, K.W. & Davis, G.W. Transgenically encoded protein photoinactivation (FlAsHFALI): acute inactivation of synaptotagmin I. Neuron 36, 805–813 (2002). 16. Poskanzer, K.E., Marek, K.W., Sweeney, S.T. & Davis, G.W. Synaptotagmin I is necessary for compensatory synaptic vesicle endocytosis in vivo. Nature 426, 559–563 (2003). 17. Tour, O., Meijer, R.M., Zacharias, D.A., Adams, S.R. & Tsien, R.Y. Genetically targeted chromophore-assisted light inactivation. Nat. Biotechnol. 21, 1505–1508 (2003). 18. Greenbaum, L., Rothmann, C., Lavie, R. & Malik, Z. Green fluorescent protein photobleaching: a model for protein damage by endogenous and exogenous singlet oxygen. Biol. Chem. 381, 1251–1258 (2000). 19. Dixit, R. & Cyr, R. Cell damage and reactive oxygen species production induced by fluorescence microscopy: effect on mitosis and guidelines for non-invasive fluorescence microscopy. Plant J. 36, 280–290 (2003). 20. Shagin, D.A. et al. GFP-like proteins as ubiquitous metazoan superfamily: evolution of functional features and structural complexity. Mol. Biol. Evol. 21, 841–850 (2004). 21. Filippin, L. et al. Improved strategies for the delivery of GFP-based Ca2+ sensors into the mitochondrial matrix. Cell Calcium 37, 129–136 (2005). 22. Loschenov, V.B., Konov, V.I. & Prokhorov, A.M. Photodynamic therapy and fluorescence diagnostics. Laser Physics 10, 1188–1207 (2000). 23. Zorov, D.B., Filburn, C.R., Klotz, L.O., Zweier, J.L. & Sollott, S.J. Reactive oxygen species (ROS)-induced ROS release: a new phenomenon accompanying induction of the mitochondrial permeability transition in cardiac myocytes. J. Exp. Med. 192, 1001–1014 (2000). 24. Ferguson, K.M., Lemmon, M.A., Schlessinger, J. & Sigler, P.B. Structure of the high affinity complex of inositol trisphosphate with a phospholipase C pleckstrin homology domain. Cell 83, 1037–1046 (1995). 25. Fujii, M. et al. Real-time visualization of PH domain-dependent translocation of phospholipase C-delta1 in renal epithelial cells (MDCK): response to hypo-osmotic stress. Biochem. Biophys. Res. Commun. 254, 284–291 (1999). 26. Shimohama, S. et al. Mutation in the pleckstrin homology domain of the human phospholipase C-delta 1 gene is associated with loss of function. Biochem. Biophys. Res. Commun. 245, 722–728 (1998). 27. Foote, C.S. Definition of type I and type II photosensitized oxidation. Photochem. Photobiol. 54, 659 (1991). 28. Milano, J. & Day, B.J. A catalytic antioxidant metalloporphyrin blocks hydrogen peroxide-induced mitochondrial DNA damage. Nucleic Acids Res. 28, 968–973 (2000). 29. Yu, Y.A., Timiryasova, T., Zhang, Q., Beltz, R. & Szalay, A.A. Optical imaging: bacteria, viruses, and mammalian cells encoding light-emitting proteins reveal the locations of primary tumors and metastases in animals. Anal. Bioanal. Chem. 377, 964–972 (2003). 30. Sloan, K.E. et al. CD155/PVR plays a key role in cell motility during tumor cell invasion and migration. BMC Cancer 4, 73 (2004). 31. Kremser, L., Petsch, M., Blaas, D. & Kenndler, E. Labeling of capsid proteins and genomic RNA of human rhinovirus with two different fluorescent dyes for selective detection by capillary electrophoresis. Anal. Chem. 76, 7360–7365 (2004). 32. Thompson, A., Nigro, J. & Seliger, H.H. Efficient singlet oxygen inactivation of firefly luciferase. Biochem. Biophys. Res. Commun. 140, 888–894 (1986). 33. Posner, G.H. et al. A chemiluminescent probe specific for singlet oxygen. Biochem. Biophys. Res. Commun. 123, 869–873 (1984). 34. Horstkotte, E. et al. Toward understanding the mechanism of chromophore-assisted laser inactivation—evidence for the primary photochemical steps. Photochem. Photobiol. 81, 358–366 (2005). 35. Lukyanov, K.A. et al. Natural animal coloration can be determined by a nonfluorescent green fluorescent protein homolog. J. Biol. Chem. 275, 25879–25882 (2000). 36. Labas, Y.A. et al. Diversity and evolution of the green fluorescent protein family. Proc. Natl. Acad. Sci. USA 99, 4256–4261 (2002). 37. Gurskaya, N.G., Savitsky, A.P., Yanushevich, Y.G., Lukyanov, S.A. & Lukyanov, K.A. Color transitions in coral’s fluorescent proteins by site-directed mutagenesis. BMC Biochem. 2, 6 (2001).

99