During Photoinduced Cell Death

Transcription

1 Supplementary information doi: /nchem.120 Imaging Intracellular Viscosity of a Single Cell During Photoinduced Cell Death Marina K. Kuimova, 1,5* Stanley W. Botchway, 2 Anthony W. Parker, 2 Milan Balaz, 3 Hazel A. Collins, 3 Harry L. Anderson, 3 Klaus Suhling 4 and Peter R. Ogilby 5* 1 Chemistry Department, Imperial College London, Exhibition Road, SW7 2AZ, UK 2 Central Laser Facility, Science and Technology Facilities Council, Rutherford Appleton Laboratory, Harwell Science and Innovation Campus, Didcot, Oxfordshire, OX11 0QX, UK 3 Oxford University, Department of Chemistry, Chemistry Research Laboratory, Oxford, OX1 3TA, UK 4 Department of Physics, King s College London, Strand, London WC2R 2LS, UK 5 Center for Oxygen Microscopy and Imaging, Department of Chemistry, University of Aarhus, Århus DK-8000, Denmark m.kuimova@imperial.ac.uk progilby@chem.au.dk Table of Contents for Supplementary Information (1) Materials and Methods...2 (2) Developments in the approach to detect singlet oxygen...2 (3) Time resolved detection of singlet oxygen phosphorescence...3 (4) Verification of the fluorescence-based ratiometric viscosity calibration plot for (5) References...10 nature chemistry 1

2 (1) Materials and Methods The butadiyne-linked lipophilic porphyrin dimer 1 was synthesised as reported previously. 1 5,10,15,20-Tetrakis(N-methyl-4-pyridyl)-21H, 23H-porphine (TMPyP) was obtained from Sigma- Aldrich and used as received. Spectroscopic grade solvents methanol and glycerol (Aldrich) were used as received for all measurements. The viscosity of the methanol/glycerol mixtures was measured using an advanced rheometric expansion system rheometer (ARES) 2 equipped with a force-rebalance transducer (torque range, Nm). A double-walled Couette-type rheometric tool was used for the measurements (cup outer radius, mm; cup inner radius, mm; bob outer radius mm; bob inner radius, mm, resulting in a 1 mm external gap and a 0.77 mm inner gap). The temperature was kept at (22.0±0.1) C with a bath circulator (Julabo F33). Each measurement used ca. 7 ml of sample. The viscosity was measured as a function of shear rate in the range from 1 to 100 s -1 and was found to be constant within this range. Absorption spectra in solutions of different viscosity were measured using a Pelkin Elmer Lamda 950 UV-Vis spectrometer. Fluorescence spectra were measured using a Jobin Yvon Fluoromax 3 fluorimeter and corrected for excitation light intensity and sensitivity of emission detection. The ratiometric images were obtained in ImageJ software. The procedure included aligning the original images collected using different interference filters with StackReg plugin 3 and using image calculator to divide one aligned image by another. The brightness level in the resulting ratiometric image was adjusted to cover the whole dynamic greyscale range of the image, and an artificial intensity two-colour scale was applied to help visualization of the viscosity gradient. (2) Developments in the approach to detect singlet oxygen Singlet oxygen production and decay in cells with TMPyP as the sensitizer has been investigated in some detail previously. 4-6 In these early studies, the time resolved O 2 (a 1 g ) signal was quantified in a photon counting experiment in which the sampling window was moved, in time, nature chemistry 2

3 across the time range studied. One disadvantage of this approach is that, when recording a time resolved O 2 (a 1 g ) trace, a sample is exposed to a significant dose of light. The difference in the present approach is the use of a multichannel scaler (Becker&Hickl, MSA 300), 7 which allows for much faster acquisition times to obtain a time resolved trace of singlet oxygen phosphorescence (3-5 min) compared to previous gated detection (ca. 30 min). Moreover, since the intensity counts in all time channels are accumulated simultaneously, the singlet oxygen traces obtained with the multichannel scaler are less susceptible to distortion which might occur during longer signal acquisition. Using the improved setup, we are now able to monitor the evolution of time resolved singlet oxygen in a cell as a function of irradiation time. (3) Time resolved detection of singlet oxygen phosphorescence For the triplet state photosensitized production of singlet oxygen, the evolution of the 1270 nm O 2 (a 1 g ) phosphorescence signal, P, in time should follow Eq. 1. 5,8 P( t) = k rem K k T ktt kremt ( e e ) ( 1 ) where K is a scaling parameter that includes the efficiency of singlet oxygen production, k T is the rate constant for all channels of sensitizer triplet state deactivation, and k rem is the rate constant that accounts for all channels of singlet oxygen removal. In our experiments, the latter can be expressed as the sum of two terms (Eq. 2), k rem = k d + k q [Q] (2) where k d is the pseudo-first-order rate constant for solvent-induced deactivation (i.e., quenching by water), and k q [Q] accounts for physical or chemical quenching by any other molecules (e.g., intracellular proteins). Similarly, k T can be expressed as the sum of two terms (Eq. 3), k T = k 0 T + k Ox [O 2 ] (3) where k 0 T is the intrinsic decay rate of the sensitizer triplet state in the absence of quenchers, and k Ox [O 2 ] accounts for quenching of the triplet state by ground state oxygen. nature chemistry 3

4 It is important to note that the diffusion dependent bimolecular quenching terms k q [Q] in Eq 2 and k Ox [O 2 ] in Eq 3 are dependent on the viscosity of the medium. Therefore, we expect that both k rem and k T will decrease as the viscosity of the environment increases. In systems such as ours, where k T ~ k rem, the observed O 2 (a 1 g ) phosphorescence signal will indeed appear as a difference of two exponential functions, Eq. 1. In the analysis of such data, one cannot, a priori, assign the rising portion of the observed signal to k T and the falling portion of the signal to k rem. Rather, independent experiments must ascertain what is the rate-limiting step. It is now well-established that (i) changes in the concentration of dissolved oxygen only influence k T, not k rem, and (ii) changes in the H/D isotopic composition of the surrounding medium only influence k rem, not k T. 5,9 Earlier experiments concerning the photosensitised generation of singlet oxygen using TMPyP in solution and in cells 6,10 have characterised the rate constants for the singlet oxygen and sensitizer triplet decay channels. These experiments indicate that in an atmosphere of 100% oxygen, the rising portion of the time resolved O 2 (a 1 g ) signal in D 2 O incubated cells is always attributed to the deactivation rate of the sensitizer triplet state, while the falling portion of the signal corresponds to the singlet oxygen decay rate. Dimer 1 has not been previously used for the photosensitised generation of singlet oxygen in cells, and the corresponding sensitizer triplet and singlet oxygen decay rates have not been characterised. It has been previously established in low temperature glass experiments for a conjugated porphyrin dimer structurally similar to 1 11 that a weak dimer phosphorescence is observed at 1100 nm. We were unable to detect phosphorescence of 1 in solution, to independently measure k T for 1. Instead, we have recorded the time resolved traces of singlet oxygen phosphorescence following excitation of 1 in air-saturated D 2 O, Fig S1. The trace obtained upon irradiation of TMPyP is also shown. For both sensitizers, the fast decay component at early times nature chemistry 4

5 arises from sensitizer fluorescence. For TMPyP, this initial fast signal of high intensity is the dominant signal between 0 1 s, while for 1 it is dominant between 0-7 s. The singlet oxygen phosphorescence trace obtained for TMPyP at time delays > 1 s is well described by Eq. 1, giving the time constants of (3±1) µs and (67±2) µs for the rising and falling components, respectively. For 1 (Fig S1, a), the fast signal is more intense, due to the considerably red-shifted fluorescence spectrum, λ max = 780 nm. At time delays > 7 s, monoexponential decay is observed at 1270 nm, with the rate constant of (64±3) µs, indicating that the rising part of the signal in this trace is obscured by the fast decay, due to the red-shifted fluorescence of 1. Since the singlet oxygen lifetime in D 2 O is known to be 67 µs, 12 we assign the longer lifetimes in both traces (1 and TMPyP) to the decay of O 2 (a 1 g ). The triplet decay rate for 1 in D 2 O is probably too short to be observed with the background of initial fluorescence signal. The estimated triplet decay lifetime, k -1 T, ca 1-5 µs, is consistent with the triplet state lifetime (k 0 T ) -1,determined for a conjugated dimer of similar structure in degassed toluene, 400 µs, 11 a concentration of ground state oxygen in water of 1 mm, and a typical bimolecular reaction rate constant for triplet state deactivation by ground state oxygen of > 10 9 s -1 M -1. nature chemistry 5

6 Phosphorescence counts at 1270 nm Phosphorescence counts at 1270 nm (a) Time / µs (b) Time / µs Fig S1. Time-resolved singlet oxygen phosphorescence traces recorded at 1270 nm in air-saturated D 2 O (a) for 1 following 790 nm excitation and (b) for TMPyP following 420 nm excitation Thus, in single-cell experiments we expect that the rising portion of the singlet oxygen phosphorescence signal at 1270 nm for both 1 and TMPyP corresponds to the sensitizer triplet decay, while the falling portion corresponds to decay of O 2 (a 1 g ). As has been noted above, the triplet state deactivation rate can decrease if the concentration of ground state oxygen in the environment surrounding the photosensitiser is low, or if the diffusion coefficient of the surrounding medium is low (e.g. caused by high viscosity), according to Eq. 3. However, a simultaneous change in both k T and k rem, as well as the increased intensity of the integrated O 2 (a 1 g ) signal during elapsed irradiation, cannot be explained by depletion in the ground state oxygen concentration. In the present study, we have detected (i) changes in rate constants corresponding to both the rising and the falling portions of the time-resolved phosphorescence traces at 1270 nm, and (ii) an increase in the integrated intensity of 1270 nm signal with elapsed nature chemistry 6

7 irradiation time (Fig 3, main text). Therefore, we can now unequivocally assign these changes to an increase in the local viscosity of the intracellular domains. (4) Verification of the fluorescence-based ratiometric viscosity calibration plot for 1. We have utilized the ratiometric fluorescence measurements for 1 in methanol-glycerol mixtures of various viscosity to create a viscosity calibration plot, Fig. 2a, main text. However, it is important to recognize that the emission spectrum of a fluorophore can also be influenced by other factors. We demonstrate here that the effects of solvent polarity and solute aggregation should not influence the use of our viscosity calibration plot to comment on intracellular viscosity. Normalised emission intensity log (fluorescence peaks ratio) 1,0 DCM 2-propanol ethanol methanol DMSO toluene 0,5 0, Wavelength / nm ,1 (a) (b) 0, log (viscosity) Fig. S2. (a) Normalised emission spectra obtained upon 473 nm excitation of 1 in solvents of different polarity: toluene (ε= 2.4), dichloromethane, DCM (ε= 9.1), 2-propanol (ε= 20), ethanol (ε= 24), methanol (ε= 33), DMSO (ε= 47); (b) double logarithmic plot of the intensity ratio (the higher energy emission peak divided by the lower energy emission peak) vs the solution viscosity. The data point obtained in toluene is shown in red and is attributed to the aggregation of 1. The data at high viscosity are taken from Fig 2a (inset) and the straight line obtained here exactly matches the one obtained in Fig 2a (inset). nature chemistry 7

8 We have verified that solvent polarity does not significantly affect the shape of the emission spectra of 1, Fig S2. The small blue shift in emission maximum is observed upon going from the least polar solvent dichloromethane (λ max = 820 nm) to the most polar solvent DMSO (λ max = 770 nm). The intensity ratio of the higher energy emission peak vs the lower energy emission peak, plotted against the solution viscosity in double logarithmic coordinates for solvents of different polarity, falls on the same straight line as that obtained in methanol/glycerol mixtures, Fig. S2b. Relative extinction coefficient (a) DCM 2-propanol ethanol methanol DMSO Wavelength / nm Relative extinction coefficient (b) methanol water cell culture medium toluene Wavelength / nm Fig. S3. Absorption spectra of 1 in the series of solvents It is important to realise that, in solutions of high viscosity, the ratio between the emission peaks is dependent upon the amount of light absorbed by the twisted vs the planar conformers of 1. The spectral shift of the absorption bands of either of the conformers with polarity might result in significantly different values of apparent intracellular viscosity. Since the fluorescence maxima of 1 in cells are very close to those observed in the methanol/glycerol mixtures, and the absorption of nature chemistry 8

9 both conformers of 1 significantly overlap at 473 nm for all solvents studied, Fig. S3, we assume that such spectral shifts are minor contributors to ratiometric data obtained from 1 in cells. 1 dissolved in toluene is an exception to a general trend of fluorescence peak ratio vs solution viscosity, Fig S2b (red point). Additionally, emission of 1 in toluene is characterised by a much lower fluorescence quantum yield. There are also characteristic differences in the absorption spectra of 1 in toluene and more polar solvents, Fig. S3. Mainly, the intensity of the Soret and Q absorption bands are significantly reduced and red shifted in toluene. Similar changes in absorption spectra are observed in water and cell culture medium, Fig S3b. We were unable to detect fluorescence of 1 in water. 13 These changes in absorption and fluorescence spectra have been observed for 1 previously 13 and are characteristic of aggregation. Thus 1 is aggregated in toluene, (aggregated species continue to be present even in the presence of 1% of pyridine, which is known to coordinate to the central Zn in a dimer and eliminate aggregation) and in the aqueous environment. The presence of aggregates in cells might influence the emission spectra of 1 and lead to inconsistencies between extracellular and intracellular viscosity calibration. We argue that intracellular emission of 1 is not determined by aggregates based on the following observations: (i) intense emission is observed from cells incubated with 1; (ii) the emission spectrum obtained for 1 in cells prior to or after irradiation (Fig 2, main text) is not the same as the emission spectrum of 1 in toluene, which is characterised by the presence of aggregates. These data suggest that the extracellular viscosity calibration plot, obtained in methanol/glycerol mixtures, is a pertinent reference for the intracellular studies and thus gives the correct values of the intracellular viscosity. nature chemistry 9

10 (5) References Balaz, M., Collins, H. A., Dahlstedt, E., and Anderson, H. L., Synthesis of biocompatible conjugated porphyrin dimers for one-photon and two-photon excited photodynamic therapy at NIR wavelengths. Org. Biomol. Chem. in press, DOI: /b814789b (2009). Kaliviotis, E. and Yianneskis, M., On the effect of dynamic flow conditions on blood microstructure investigated with optical shearing microscopy and rheometry. Proc. Inst. Mech. Eng., Part H-J.Eng. Medicine 221, 887 (2007). Thévenaz, P., Ruttimann, U. E., and Unser, M., A Pyramid Approach to Subpixel Registration Based on Intensity. IEEE Trans. Image Processing 7 (1), 27 (1998). Snyder, J. W., Skovsen, E., Lambert, J. D. C., and Ogilby, P. R., Subcellular, time-resolved studies of singlet oxygen in single cells. J. Am. Chem. Soc. 127 (42), (2005). Snyder, J. W., Skovsen, E., Lambert, J. D. C., Poulsen, L., and Ogilby, P. R., Optical Detection of Singlet Oxygen from Single Cells. Phys. Chem. Chem. Phys. 8, 4280 (2006). Hatz, S., Poulsen, L., and Ogilby, P. R., Time-Resolved singlet oxygen Phosphorescence measurements from photosensitized experiments in Single cells: the effect of oxygen diffusion and oxygen concentration. Photochem. Photobiol. 84 (5), 1284 (2008). Jiménez-Banzo, A., Ragàs, X., Kapusta, P., and Nonell, S., Time-resolved methods in biophysics. 7. Photon counting vs. analog time-resolved singlet oxygen phosphorescence detection, Photochem. Photobiol. Sci. 7, 1003 (2008). Egorov, S. Y. et al., Rise and decay kinetics of photosensitized singlet oxygen luminescence in water. Measurements with nanosecond time-correlated single photon counting technique. Chem. Phys. Lett. 163, 421 (1989). Schweitzer, C. and Schmidt, R., Physical mechanisms of generation and deactivation of singlet oxygen. Chem. Rev. 103, 1685 (2003). Kuimova, M. K., Yahioglu, G., and Ogilby, P. R., Singlet Oxygen in a Cell: Spatially- Dependent Lifetimes and Quenching Rate Constants. J. Am. Chem. Soc. DOI: /ja807484b (2009). Kuimova, M. K. et al., Determination of the triplet state energies of a series of conjugated porphyrin oligomers. Photochem. Photobiol. Sci. 6, 675 (2007). Ogilby, P. R. and Foote, C. S., The effect of solvent, solvent isotopic substitution, and temperature on the lifetime of singlet molecular oxygen. J. Am. Chem. Soc. 105, 3423 (1983). nature chemistry 10

11 13 Kuimova, M. K. et al., Photophysical Properties and Intracellular Imaging of Water-Soluble Porphyrin Dimers for Two-Photon Excited Photodynamic Therapy. Org. Biomol. Chem. in press, DOI: /b814791d (2009). nature chemistry 11