Fluorescent/phosphorescent dual-emissive conjugated polymer dots for hypoxia bioimaging† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c4sc03062a Click here for additional data file.

Fluorescent/phosphorescent dual-emissive conjugated polymer dots were designed and synthesized, ere used for tumor hypoxia sensing via ratiometric imaging and photoluminescence lifetime imaging.


Introduction
Hypoxia has been found to be closely related to various diseases, such as solid tumors, 1 brain abnormalities 2 and retinal diseases. 3 Especially, hypoxia is a feature of tumor tissues, and the median oxygen (O 2 ) concentration in some solid tumors is around 4% and may even decrease to 0% locally. 4 The real-time monitoring of O 2 concentration in living cells and tissue can not only lead to accurate diagnosis of cancer, but can also be used to evaluate therapeutic effects. Thus far, enormous efforts have been focused on the development of hypoxia sensing. Hypoxia can be selectively detected by utilizing immunostaining, 5 magnetic resonance imaging, 6 positron emission tomography imaging, 7 and optical imaging techniques. 8-10 Among them, non-invasive optical imaging offers a powerful approach to map oxygen in living cells and tissues with high sensitivity and spatial resolution. Many uorescence probes for optical imaging that employ organic dyes containing a nitro group, quinone group, or azo group as the hypoxia-sensing moiety have been developed. 10 Most of these probes, however, usually show an irreversible uorescence change upon bioreduction reaction under hypoxia, which limits their applications in real-time uorescence monitoring of oxygen concentration.
Phosphorescent probes based on transition-metal complexes can be used for fully reversible real-time monitoring of O 2 levels in vitro with high-resolution by utilizing the energy transfer between the triplet excited state of the metal complex and the triplet ground state of O 2 . 8,11 Thus, phosphorescent metal complexes with a long-lived triplet excited state, such as platinum(II), 12 palladium(II), 13 ruthenium(II), 14 and iridium(III) 15 complexes, have been regarded as a kind of fascinating hypoxia probe. The phosphorescence intensities of these complexes are reversibly dependent on the change in O 2 concentration. However, most of these reported probes are based on the variation in single phosphorescent emission ("ON-OFF" or "OFF-ON" type). The single intensity-based reporting signal is easily inuenced by the external environment, such as the probe's concentration, temperature, or pH value. The standardization and use of accurate readout of O 2 concentration are also difficult. Therefore, an advisable choice is to develop ratiometric O 2 probes that can allow for accurate measurement of O 2 concentration through the ratio change of the emission intensities at two different wavelengths.
For the design of ratiometric O 2 probes, herein, we hope to demonstrate an effective way to construct a Förster resonance energy transfer (FRET) system with an O 2 -insensitive uorophore as the energy donor and an O 2 -sensitive phosphorescent metal complex as the acceptor (Fig. 1). The occurrence of FRET can signicantly improve the sensitivity of sensing by promoting the population of the oxygen-sensitive triplet excited state of the acceptor. We introduced phosphorescent platinum(II) porphyrin (O 2 -sensitive) into a uorene-based conjugated polyelectrolyte (O 2 -insensitive). Fluorescent/ phosphorescent dual-emissive polymer dots (FP-Pdots, see Fig. 1A) were successfully prepared as excellent O 2 probes. The as-prepared FP-Pdots enabled ratiometric measurements of hypoxia in living cells with the advantages of eliminating photobleaching and excitation power uctuation by using the O 2insensitive uorescence from polyuorene moieties as a reference and O 2 -sensitive phosphorescence from Pt(II) porphyrins as the sensing signal. 16 Taking advantage of the long excitedstate lifetime of phosphorescent Pt(II) porphyrins, time-resolved luminescence imaging of O 2 levels in living cells, including photoluminescence lifetime imaging microscopy (PLIM) and time-gated luminescence imaging (TGLI), was realized. These techniques can eliminate external inuences and auto-uorescence interferences from biological samples. 17,18 Furthermore, we applied the FP-Pdots probe to nude mice for the luminescence imaging of tumor hypoxia. These experiments have shown that FP-Pdots are an excellent class of luminescent probes for practical application in determining O 2 concentrations in living samples.

Results and discussion
Design, synthesis and properties of the FP-Pdots We chose conjugated polymer as the oxygen-insensitive energy donor to construct FRET-based ratiometric O 2 probes ( Fig. 1) because of their high light absorptivity, extraordinary uorescence brightness, excellent photostability, and good water-dispersibility. 19,20 The synthetic route and chemical structures of the conjugated polyuorene electrolyte are shown in Scheme 1. The Pt(II) complex-polymer precursor P1 was prepared via a Suzuki polycondensation reaction, and its structure was characterized via 1 H NMR and 13 C NMR. The number-average molecular weight of the polymer was determined to be 25 400 with a polydispersity index of 2.49, measured via gel permeation chromatography (GPC) in THF by using the calibration curve of polystyrene standards. The target Pt(II) complex-polymer P2 was obtained by the quarternization of the precursor P1 (Scheme 1). The initial feed ratio of Pt(II) porphyrins to the total monomers was 10 mol%. The actual Pt(II) porphyrin content in the copolymer, which was estimated via 1 H NMR, was about 7 mol%. This value is lower than that of the feed ratio probably because of the reaction activity or steric hindrance.
The morphology of P2 in the phosphate buffer solution (PBS) was investigated via transmission electron microscopy (TEM) and dynamic light scattering (DLS), as shown in Fig. 2A and S1. † Each of the FP-Pdots have a diameter of approximately 5 nm and they were formed via self-assembly caused by their amphiphilic structures with hydrophobic backbones and hydrophilic side chains. Thus, the as-prepared FP-Pdots were well dispersed in PBS. In addition, the zeta-potential of P2 was measured to be 54.00 mV, which further demonstrated the good dispersibility and stability of the FP-Pdots in PBS.
A particular degree of spectral overlap between the emission spectrum of Pt(II)-free polyuorene P3 (see Scheme S1 †) and the absorption spectrum of Pt(II) porphyrin ensures the efficient energy transfer from polyuorene segments as host to Pt(II) porphyrin moieties as guest (Fig. S4 †). The photophysical properties of P2, specically of the FP-Pdots, are shown in Fig. 2B. The UV-vis absorption spectrum of the FP-Pdots is dominated by a strong featureless transition centered at 395 nm, which consists of a mixture of Soret bands of Pt(II) porphyrin 21 and the p-p* transition of uorene units. An additional two weak and broad absorptions, which peaked at 510 and 540 nm, respectively, were also found. These results correspond to Q(1,0) and Q(0,0) of the Pt(II) porphyrin moieties. 21 FP-Pdots excited at 375 nm exhibit strong blue uorescence, attributed to the p-p* emission of the uorene moieties, as well as relatively weak red phosphorescence of the Pt(II) porphyrin moieties (Fig. 2B). In addition, the FP-Pdots of P2 also show more efficient FRET than the blend of Pt(II) porphyrin (1) and Pt(II)-free polyelectrolyte P3 according to the PL spectra ( Fig. S5 †). Although the mixture of 1 and P3 showed the same molar ratio of Pt(II) porphyrin and uorene to P2, the emission from the Pt(II) porphyrin was weaker than that in P2. The Förster radius is calculated to be 4.7 nm, 15a demonstrating the effectiveness of covalent incorporation of Pt(II) porphyrin into P2. It also indicated that the dose of Pt(II) porphyrin was reduced when FRET was present. Therefore, the nanoprobe will be less toxic to cells and animals. Furthermore, due to the high molar absorption coefficient and quantum yield of polyuorene, the FRET from polyuorene to Pt(II) porphyrin will be of benet for better O 2 -sensitivity than for non-FRET systems.

Luminescence response to O 2 content
The PL spectra of the FP-Pdots at various O 2 concentrations in water at room temperature are shown in Fig. 2C. Under an atmosphere of 21% O 2 , the FP-Pdots showed strong emission centered at 425 nm with a shoulder peak at 450 nm attributed to polyuorene moieties as well as weak emission at 656 nm with a shoulder peak at 721 nm from Pt(II) porphyrins. The solution exhibited very bright blue emission under UV excitation (Fig. 2C, inset). With decreasing O 2 content, the blue uorescence intensities showed little change, but the red phosphorescence intensities grew progressively. This result is in agreement with the O 2 sensitivity of the Pt(II) porphyrin complexes. The red phosphorescence became dominant under an atmosphere of 0% O 2 . Thus, the color of the emission changed from very bright blue to red under UV excitation (Fig. 2C, inset).  The ratiometric oxygen sensing of the FP-Pdots was analyzed quantitatively based on the data of phosphorescence intensities shown in Fig. 2C, according to the Stern-Volmer equation: 22 where K sv is the Stern-Volmer constant, pO 2 is the oxygen partial pressure, R I 0 ¼ (I 656 0 /I 422 0 ) and R I ¼ (I 656 /I 422 ) represent the ratios of the phosphorescence (656 nm) intensity of Pt(II) porphyrin to the uorescence (422 nm) intensity of the poly-uorene moieties in the absence and presence of O 2 , respectively. A good linearity of R I 0 /R I as a function of O 2 concentration was observed ( Fig. 2D and S6 †), according to eqn (1). The K sv value is 1.67 Â 10 À2 mmHg À1 . At the same time, no change in the uorescence intensities of the FP-Pdots monitored at 425 nm from the polyuorene moiety was observed at various O 2 concentrations (Fig. 2C), which means that the ratio of the phosphorescence intensity at 656 nm to the uorescence intensity at 422 nm for the FP-Pdots is dependent on O 2 concentration. Therefore, the FP-Pdots can be used as a ratiometric phosphorescent probe with high reliability for measuring oxygen levels. The limit of detection of the FP-Pdots was determined to be 0.5 mmHg.

Lifetime response to O 2
To conrm the reliability of the O 2 sensing measurements of the FP-Pdots nanoprobe, we performed similar O 2 quenching experiments based on lifetime measurements of phosphorescence from Pt(II) porphyrin moieties, according to eqn (2): where K sv is the Stern-Volmer constant, pO 2 is the oxygen partial pressure, and s 0 and s are the phosphorescence lifetimes in the absence and presence of O 2 , respectively. The lifetime (s) monitored at 656 nm decreased evidently with increasing O 2 concentration, as shown in Fig. 2E. For example, the phosphorescence lifetimes of the FP-Pdots in N 2 and air were 33.7 and 9.9 ms, respectively. As illustrated in Fig. 2F, a good linear relationship between s 0 /s and pO 2 was observed. The constant K sv based on the excited state lifetime was 1.57 Â 10 À2 mmHg À1 , which is very close to that (1.54 Â 10 À2 mmHg À1 ) obtained from luminescence intensity measurements. However, the uorescence lifetime of the FP-Pdots monitored at 425 nm from the polyuorene moiety remains constant at various O 2 concentrations (Fig. S7 †). Therefore, the FP-Pdots can be used as a lifetime-based phosphorescent probe with high reliability for measuring oxygen levels.

Ratiometric imaging of intracellular O 2 levels
We then used the FP-Pdots as a ratiometric luminescence nanoprobe to investigate intracellular O 2 levels. The cells were cultured for 24 h at 37 C at 21% and 2.5% O 2 concentrations. The excitation wavelength was 405 nm. As shown in Fig. 3, the emission intensities collected at wavelengths from 420 nm to 460 nm (I blue ), which correspond to the reference uorescence of polyuorene moieties, are almost the same at 21% and 2.5% O 2 concentrations. By contrast, the emission images taken in the wavelength range of 630 nm to 680 nm (I red ), which corresponds to the phosphorescence of Pt(II) porphyrin complexes, show a much brighter image at a condition of 2.5% O 2 than at 21% O 2 . The excellent intracellular ratiometric O 2 sensing properties of the FP-Pdots nanoprobe can also be conrmed by the evident change in the I red /I blue ratio values in living cells cultured at different O 2 concentrations, as shown in ratio images in Fig. 3.

Time-resolved luminescence imaging of intracellular O 2 levels
Considering the response of long phosphorescence lifetime from Pt(II) porphyrins to oxygen, we have applied our FP-Pdots in time-resolved luminescence imaging of intracellular O 2 Fig. 3 Confocal luminescence imaging and ratiometric luminescence imaging (l ex ¼ 405 nm) of HepG2 cells incubated with the FP-Pdots (10 mg mL À1 ) at 21% or 2.5% O 2 concentrations. In luminescence imaging, the emission channels at wavelengths of 420-460 nm and 630-680 nm were collected. In ratiometric imaging, the ratio of emission intensity at 630-680 nm to that at 420-460 nm was chosen as the detected signal. Fig. 4 Photoluminescence lifetime images (l ex ¼ 405 nm) of HepG2 cells incubated with 10 mg mL À1 FP-Pdots at 37 C for 2 h at 21% and 2.5% O 2 concentrations, respectively. The magnification of the objective lens is 40Â. The luminescence signals were collected in the range of 420-680 nm. levels. First, the PLIM of intracellular O 2 level was performed using the FP-Pdots as a lifetime-based nanoprobe, as shown in Fig. 4. The average emission lifetime s was about 17 ns at 21% O 2 concentration and increased to about 95 ns when the O 2 concentration was lowered to 2.5%. The interference from short-lived autouorescence was eliminated because of the relatively long average lifetime of FP-Pdots. This result demonstrated that the FP-Pdots probe exhibits an evident difference in emission lifetime at different O 2 concentrations and it can be applied as an excellent PLIM probe for sensing intracellular O 2 levels.
To further demonstrate the ability of anti-interference when the FP-Pdots were applied for O 2 sensing in living cells, we performed TGLI measurements, and the luminescence intensity images at different time ranges were collected (Fig. 5). When the signal was collected at a time range of 0 ns to 2000 ns, the images of FP-Pdot-treated HepG2 cells exhibited a high signal intensity at 21% and 2.5% O 2 concentrations because of the presence of reference uorescence from polyuorene moieties. Once a particular time delay was exerted, such as in the image collected at a time range of 250 ns to 2000 ns, the signal intensity of the FP-Pdot-treated HepG2 cells at 21% O 2 concentration could not be observed. By contrast, the signal intensity collected at a time range of 250 ns to 2000 ns at 2.5% O 2 concentration was still high enough to be observed, although the decrease in the luminescence intensity of the FP-Pdottreated HepG2 cells is reasonable. When the time delay was increased to 500 ns, the TGLI image of the FP-Pdot-treated HepG2 cells can still be measured. This phenomenon is attributed to the long phosphorescence lifetime of the FP-Pdots. The difference in the emission intensity at O 2 concentrations of 21% and 2.5% became increasingly more evident, which indicated that O 2 sensing can be made more sensitive by collecting the signal at a long time range via the TGLI technique. As far as we know, this study is the rst to use Pdots-based TGLI for intracellular oxygen detection.

Luminescence imaging of tumor hypoxia in nude mice
To further conrm the potential of the probe for practical application, we applied the FP-Pdots probe for luminescence imaging of tumor hypoxia in nude mice. In vivo and ex vivo luminescence imaging were performed using a modied Kodak in vivo imaging system. The luminescence signals were collected at 660 AE 13 nm. The mice were intratumorally injected with the FP-Pdots solution (10 mL, 2 mg mL À1 ) without being deoxygenized. For the purpose of comparison, the mice were subcutaneously injected with the FP-Pdots probe under the same conditions. As shown in Fig. 6A, a very intense signal from the area of the tumor (ROI 1) was detected. When compared to the area of the subcutaneous injection (ROI 2), an obviously high signal-to-noise ratio in the emission intensity within the tumor area can be observed. To intuitively estimate the change in luminescence intensity, imaging dynamic curves were plotted, as shown in Fig. 6B. The luminescence intensity of the FP-Pdots located in tumor ROI 1 increased abruptly following 3 seconds of injection into the tumor and subsequently increased. The intensity was up to 3.3 times greater than that at 660 AE 13 nm of the FP-Pdots (ROI 3) compared with the phosphorescence intensity at ROI 2. Therefore, the FP-Pdots exhibited excellent oxygen-sensitivity and can be applied as tumor hypoxia probes in vivo.
Lastly, the cytotoxicity of the FP-Pdots nanoprobe toward HepG2 cells was determined via an MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. The plot of cellular viabilities (%) vs. incubation concentrations (0-50 mg mL À1 ) of the probe in PBS buffer at 37 C for 24 h is illustrated in Fig. S8. † The viabilities of the HepG2 cells still retained a value higher than 70% even aer incubation with a high concentration (50 mg mL À1 ) of the nanoprobe. Under the living cell imaging experimental conditions (10 mg mL À1 concentration, 2 h incubation time), the FP-Pdots nanoprobe showed negligible cytotoxicity toward HepG2 cells with cell viabilities of more than 95%. These data demonstrated that the use of the FP-Pdots as a hypoxia probe shows excellent biocompatibility.

Conclusions
In conclusion, we have presented a hypoxia nanoprobe based on uorescent/phosphorescent dual-emissive FP-Pdots, which was prepared from conjugated polyelectrolyte by using poly-uorenes as an O 2 -insensitive uorophore and Pt(II) porphyrins  as an O 2 -sensitive phosphor. The FP-Pdots nanoprobe was demonstrated to be a kind of excellent ratiometric O 2 nanoprobe with full reversibility and low toxicity toward living cells. Importantly, the FP-Pdots can perform excellently in PLIM and TGLI techniques, which will evidently improve the sensing sensitivity and reliability. In addition, the FP-Pdots nanoprobe was applied to tumor-bearing mice for luminescence imaging of tumor hypoxia. These initial studies showed that the uorescent/phosphorescent dual-emissive FP-Pdots are a promising kind of hypoxia nanoprobe, which will be helpful in designing a ratiometric and time-resolved luminescent probe for monitoring tumor hypoxia.