Photo-induced uncaging of a specific Re(I) organometallic complex in living cells

: In the last decades, a large number of organometallic complexes have shown promising anti-proliferative activity towards different cancer cell lines. However, these compounds generally had low cellular uptake and low selectivity towards cancer cells over healthy cells. The use of external triggers (e.g. light, ultra-sound, temperature, etc.) to modify the cytotoxic effect of a prodrug and the coupling of a targeting vector (e.g. peptides, antibodies, etc.) to a drug were found to be very successful techniques to tackle these drawbacks. Here, we envisioned combining these two methods, namely an external trigger (i.e. light activation) and a targeting vector, in an organometallic compound. More specifically, a Re(I) tricarbonyl N,N-bis(quinolinoyl) complex (Re-NH2) was derivatised with a photo-labile protecting group (PLPG) to cage Re-NH2 by formation of Re-PLPG. For organelle/cellular specificity, Re-PLPG was then further coupled to a nuclear localization sequence (NLS) or a bombesin peptide derivative to give Re-PLPG-NLS or Re-PLPG-Bombesin, respectively. Photolysis experiments in PBS buffer (pH 7.4) demonstrated that Re-NH2 was completely photo-released from Re-PLPG-NLS and Re-PLPG-Bombesin using a very low irradiation dose (1.2 J cm−2). To the best of our knowledge, these are the first two examples of the selective photo-release of an intact organometallic compound from a bioconjugate. Of high interest, both derivatives showed toxicity comparable to that of cisplatin towards cervical cancer cells (HeLa) upon light irradiation, although the phototoxic index (PTI) varied greatly with the targeting peptide. The cell death mechanism of Re-PLPG-NLS was explored using different techniques, including fluorescence microscopy, ICP-MS, gel electrophoresis, flow cytometry and transmission electron microscopy (TEM). It could be demonstrated that HeLa cells treated with Re-PLPG-NLS in the dark and upon irradiation showed severe cell stress (nucleolar segregation, pyknosis and vacuolation). The data obtained from an Annexin V/propidium iodide (PI) assay indicated that, after an early apoptotic stage, the onset induced by Re-PLPG-NLS led to cell death, with features ascribable to late apoptosis and necrosis, which were more marked for the treatment involving irradiation. In the last decades, a large number of organometallic complexes have shown promising anti-proliferative activity towards di ﬀ erent cancer cell lines. However, these compounds generally had low cellular uptake and low selectivity towards cancer cells over healthy cells. The use of external triggers ( e.g. light, ultra-sound, temperature, etc. ) to modify the cytotoxic e ﬀ ect of a prodrug and the coupling of a targeting vector ( e.g. peptides, antibodies, etc. ) to a drug were found to be very successful techniques to tackle these drawbacks. Here, we envisioned combining these two methods, namely an external trigger ( i.e. light activation) and a targeting vector, in an organometallic compound. More speci ﬁ cally, a Re( I ) tricarbonyl N , N -bis(quinolinoyl) complex ( Re-NH 2 ) was derivatised with a photo-labile protecting group (PLPG) to cage Re-NH 2 by formation of Re-PLPG . For organelle/cellular speci ﬁ city, Re-PLPG was then further coupled to a nuclear localization sequence (NLS) or a bombesin peptide derivative to give Re-PLPG-NLS or Re-PLPG-Bombesin , respectively. Photolysis experiments in PBS bu ﬀ er (pH 7.4) demonstrated that Re-NH 2 was completely photo-released from Re-PLPG-NLS and Re-PLPG-Bombesin using a very low irradiation dose (1.2 J cm (cid:1) 2 ). To the best of our knowledge, these are the ﬁ rst two examples of the selective photo-release of an intact organometallic compound from a bioconjugate. Of high interest, both derivatives showed toxicity comparable to that of cisplatin towards cervical cancer cells (HeLa) upon light irradiation, although the phototoxic index (PTI) varied greatly with the targeting peptide. The cell death mechanism of Re-PLPG-NLS was explored using di ﬀ erent techniques, including ﬂ uorescence microscopy, ICP-MS, gel electrophoresis, ﬂ ow cytometry and transmission electron microscopy (TEM). It could be demonstrated that HeLa cells treated with Re-PLPG-NLS in the dark and upon irradiation showed severe cell stress (nucleolar segregation, pyknosis and vacuolation). The data obtained from an Annexin V/propidium iodide (PI) assay indicated that, after an early apoptotic stage, the onset induced by Re-PLPG-NLS led to cell death, with features ascribable to late apoptosis and necrosis, which were more marked for the treatment involving irradiation.

Photo-induced uncaging of a specific Re(I) organometallic complex in living cells † Introduction Photodynamic therapy (PDT) is a medical technique that allows the destruction of target cells by irradiation of a photosensitiser (PS) with light. Offering important advantages, such as temporal and spatial control, PDT is currently used to treat certain cancers that are readily accessible with lamps or optic bers (e.g. some types of head and neck, skin, bladder, lung and oesophagus cancers). 2 Other applications of PDTtreatment of multi-drug resistant bacterial infections in particularare also being actively explored. [3][4][5][6][7][8] However, PDT suffers from a general drawback, namely its reliance on an oxygen-dependent mechanism. In fact, the two main PDT modes of action (type I which involves reactive oxygen species (ROS) production, and type II in which the damage is directly produced by 1 O 2 ) rely on the conversion of oxygen from its triplet ground state to its singlet excited state by a PS. This can be problematic since tumours, in their inner central part in particular, are oen hypoxic. 9,10 Hence, it would be extremely interesting to develop a light-mediated strategy that does not rely on the presence of oxygen. In this sense, singlet oxygen generation is not the only effect that light can induce. Chemical bonds can also be rearranged and/or cleaved upon light irradiation. Indeed, the activity of a compound may be modi-ed by caging some of its functional groups with a photoremovable moiety. The release of the original compound can then be controlled by light irradiation. Such light-activated molecules have proven themselves useful tools in biological/ chemical biology/medicinal chemistry studies, such as in the analysis of neurological and enzymatic processes, gene expression, cell signalling and many others. 11,12 Several drugs and drug candidates have also been photo-caged and efficiently released upon light irradiation. [13][14][15][16][17][18][19] Among them are aspirin, 13 ibuprofen, 13 ketoprofen, 13 carbonic anhydrase II inhibitors, 14 as well as a few anti-cancer agents such as uxoridine, 15 paclitaxel, 16 tegafur, 17 doxorubicin, 18 and a photoreactive DNA intercalator. 19 Recently, this photo-caging strategy has also been combined with singlet oxygen generation: an anti-cancer drug candidate, combretastatin A-4, has been coupled to a porphyrin photosensitiser via a cleavable singlet oxygen linker. 20 Surprisingly, although many studies reporting the use of photo-caged organic molecules have been published, examples of the specic release of intact (organo)metallic compounds using light as a trigger are very scarce. 21 Surely, some complexes are known to undergo light-induced reactions involving the metal centre itself. This property was used not only to enhance their cytotoxicity, [22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39] but also to design metalbased photo-cages. [40][41][42][43][44] As shown by Franz and co-workers using a photo-labile o-nitrophenyl moiety, the release of a labile metal complex 45 or the release of metal ions 46,47 could also induce cytotoxicity. Of note, the photo-release of metal ions alone to explore neurological response and cell signalling is currently an intensively investigated research area. [48][49][50][51][52] However, to the best of our knowledge, we were the rst to report the photo-release of an intact cytotoxic substitutionally inert Ru(II) complex upon light irradiation. 21 While most drugs on the market are organic molecules, the platinum(II)-based drug cisplatin and its derivatives oxaliplatin and carboplatin are playing a pivotal role in anti-cancer chemotherapy. In addition to purely metal coordination complexes, a signicant number of organometallic compounds have been reported to be strongly toxic towards cancer cells in recent years. [53][54][55][56][57] Among them, rhenium tricarbonyl complexes may not be the most numerous, but are particularly attractive, as their isostructural 99m Tc analogues permit the use of imaging to observe the compound's behaviour in vivo. In this work, we are building on the information gained during a previous study on a Re(I) tricarbonyl bis(quinolinoyl) complex that showed a strong PDT prole 1 to achieve multi-modal activity. We observed that synthetic modications of the alkyl chain of these Re(I) complexes (see Scheme 1) strongly inuenced both their dark and light-mediated cytotoxicities. 1,58 Hence, photocaging and peptide conjugation of these compounds could provide temporal and spatial control. Moreover, we envisaged boosting the uptake of our derivative by malignant tissues via conjugation of targeting peptides. Herein, we report the synthesis and characterization of photo-activated targeted Re(I) tricarbonyl N,N-bis(quinolinoyl) bioconjugates, as well as their photo-products. Their bioactivity towards cervical (HeLa) and prostate (PC-3) cancer cells, as well as towards normal lung broblasts (MRC-5), upon light activation is described. The possible modes of cytotoxic action are also explored in depth.

Results and discussion
Synthesis and characterization of the organometallic caged bioconjugates Re-NH 2 59 was caged with an o-nitrophenyl-based photo-linker, namely the Dmochowski photo-labile protecting group (PLPG). 60 This type of photo-linker has been systematically studied and applied in numerous biochemical systems. 12,[61][62][63][64] Besides, it allows coupling of additional moleculesa property we used to attach targeting peptides to our caged Re(I) derivative. More specically, the primary amino group of Re-NH 2 was rst reacted with PLPG via N-hydroxysuccinimide ester coupling to give Re-PLPG (Scheme 1). 60 For this purpose, the original Dmochowski protocol was slightly modied. The phosphate buffer was replaced by organic solvents to solubilise the organometallic compounds, and a weak base was added to catch the protons liberated during the reaction. Although the presence of a base can induce maleimide ring opening, 60 only a small amount of by-product was observed. The formation of Re-PLPG was conrmed by ESI-MS, which showed the expected mass of [M] + at m/z 943.2, as well as the isotopic pattern characteristic of rhenium (Fig. S3 in the ESI †). In the 1 H NMR spectrum, a new peak corresponding to the amide proton of the newly formed carbamate bond appeared at approximately 5.7 ppm. As expected, a shi of the signals of the protons in the immediate vicinity of the carbamate was observed. In Re-PLPG, the signals of the methyl group of the PLPG moiety and its neighbouring tertiary carbon proton shied upeld compared to PLPG. Slight (both upeld and downeld) changes in chemical shis were also observed for the protons of the organometallic part of Re-PLPG compared to Re-NH 2 ( Fig. S5-S7 †).
The photo-caged complex Re-PLPG was then conjugated to two different targeting peptides via Michael addition of the cysteine thiol of a peptide to the maleimide bond of the PLPG (Scheme 1). Re-PLPG was attached to (1) a nucleus localization signalling (NLS) peptide to target the compound to the nucleus in close proximity to DNA and (2) a bombesin derivative 65,66 to improve the uptake of the compound by cancer cells overexpressing the gastrin-releasing peptide receptor (GRPR). A cysteine residue was added to each peptide to allow conjugation to the PLPG maleimide. Aer preparative HPLC purication and lyophilisation, light yellow solids of Re-PLPG-NLS and Re-PLPG-Bombesin were obtained.  S9-S16 †). The purity of the bioconjugates was unambiguously conrmed by LC-MS, in which two peaks in very close proximity with the same mass pattern were observed, corresponding to the two possible diastereomers formed during the Michael addition. Such behaviour was previously reported for biotin conjugated compounds photo-caged with a similar PLPG. 18,67 Two small extra m/z peaks were sometimes observed in the mass spectra during LC-MS. They could be assigned to the photo-uncaged but not yet decarboxylated metal complex, as well as to the fully uncaged complex. This uncaging process clearly happens during the electrospray ionization process due to the harsh ionization conditions used for LC-MS, since the uncaged complex is eluted over a different time frame (Fig. S12 and S16-S18 in the ESI †).

Photo-release of the organometallic complex
In order to verify the photo-uncaging efficiency of our system, the photo-release of Re-NH 2 from the two prepared bioconjugates, namely Re-PLPG-NLS and Re-PLPG-Bombesin, was then investigated in phosphate buffer solution (PBS, pH 7.4). To determine the light dose required to photo-release Re-NH 2 from Re-PLPG-NLS and Re-PLPG-Bombesin in cells, the two compounds were irradiated using a Rayonet UV reactor (350 nm) that was later used in the cell culture experiments. Both Re-PLPG-NLS and Re-PLPG-Bombesin completely photo-released Re-NH 2 using a very low irradiation dose of 1.2 J cm À2four times lower than the dose used to completely photolyze the caged Ru complex previously reported by our group. 21 During the irradiation, the composition of the solutions was monitored by LC-MS. As shown in Fig. 1 for Re-PLPG-NLS, prior to irradiation, only two peaks at 12.6-13.3 min, corresponding to the two diastereomers, were present. As the caged complex was irradiated, new peaks arose in the chromatograms. The rst small peaks at 9.8-10 min were assigned to the by-products of the UV decomposition of the PLPG (Fig. S17 and S19 †) that have been reported for o-nitrobenzyl photo-linkers. 12 The released Re-NH 2 complex is detected at 14.5 min. In the case of Re-PLPG-Bombesin, the order of peak appearance changes, as bombesin is more lypophilic than NLS (Fig. S18 †). The released Re-NH 2 is eluted rst, followed by o-nitrobenzyl moiety by-products, while the remaining Re-PLPG-Bombesin peak is eluted last. Photolysis quantum yields were then measured using a 355 nm laser setup. As expected, the targeting peptide had no inuence, and both Re-PLPG-NLS and Re-PLPG-Bombesin achieved a good photolysis quantum yield of 10 AE 2% (value obtained using  The cytotoxicity of the photo-caged bioconjugates Re-PLPG-NLS and Re-PLPG-Bombesin, as well as of the Re-NH 2 complex, towards cancerous (HeLa) and non-cancerous (MRC-5) cells was evaluated in the presence and the absence of light. Cisplatin was used as a positive control (although it is not a PDT agent), in order to compare the cytotoxicity of our compounds to that of a successful metal-based drug. Re-PLPG-Bombesin was additionally tested on prostate cancer cells (PC-3), as the analogue of bombesin used in this study has been optimised on this cell line, which over-expresses GRPR. 65,66 HeLa cells also overexpress receptors belonging to the bombesin receptor family, namely the bombesin receptor subtype 3 (BSR-3), for which our bombesin derivative should also have some affinity. 69 As shown in Table 1, 48 h of incubation, photo-caging and peptide conjugation strongly affected both the dark and light toxicity of the compounds. In general, the bioconjugates with NLS peptide (Re-NLS and Re-PLPG-NLS) showed a marked increase in dark cytotoxicity compared to Re-NH 2 (up to 13-fold), while the bombesin derivatives displayed only moderate toxicity towards MRC-5 cells and remained inactive up to the highest concentration used in this study (100 mM) towards HeLa and PC-3. While the presence of the PLPG rendered Re-PLPG-NLS almost 2.5 times more toxic towards HeLa cells than Re-NLS, it did not affect the cytotoxicity of the bombesin derivative Re-PLPG-Bombesin towards HeLa and PC-3, and actually attenuated (up to 2-fold) the effect of the compounds on the non-cancerous cell line MRC-5.
The inuence of light on the activity of Re-PLPG-NLS and Re-PLPG-Bombesin was studied by incubating cells with the compounds for 4 h. This incubation time frame has been shown to allow sufficient uptake of Re-NH 2 for its phototoxic effect to be observed. 1 Aer incubation, the medium was replaced with a compound-free fresh medium, the cells were irradiated at 350 nm (UV reactor) with a low light dose of 2.58 J cm À2 , and the viability of the cells was quantied 44 h later. Light irradiation increased the toxicity of almost all tested compounds towards HeLa, PC-3 and MRC-5. The UVA irradiation by itself did not affect cell viability (Fig. S22 †). Interestingly, although light signicantly increased the cytotoxicity of the Re-NH 2 complex towards HeLa (10-fold) and MRC-5 cells (2.5-fold), irradiated PC-3 cells remained unaffected at a Re-NH 2 concentration of up to 100 mM. Derivatization of Re-NH 2 with bombesin allowed much more efficient targeting of this particular cancer cell line (5-7 phototoxic index (PTI)). In addition, the bombesin conjugates Re-Bombesin and Re-PLPG-Bombesin showed good light-induced increases in cytotoxicity towards the HeLa cell line (10-20 fold). However, no signicant difference (within the error range) in phototoxicity between Re-Bombesin and Re-PLPG-Bombesini.e. no inuence of the PLPGwas observed. These results suggest a predominant singlet oxygen-dependent mode of cytotoxic action that, together with an acceptable PTI towards cancer cell lines and moderate toxicity towards healthy MRC-5 cells, makes these two bombesin derivatives interesting for PDT applications. In fact, some commercially available PSs have comparable PTIs (at a similar light dose of 1.5 J cm À2 ), e.g. Photofrin (PTI 3), Hypercin (PTI 27) and ALA-induced PPIX. 70 Only Foscan and Fospeg achieve PTIs of 268 and 4695 respectively. 70 Compared with our previously reported caged Ru(II) species that use a similar photo-caging moiety, Re-Bombesin and Re-PLPG-Bombesin possess better phototoxicity characteristics (higher PTI and higher light toxicity). 21 However, in the area of photoactive metal compounds acting via singlet oxygen production or ligand lability, similar or higher (typically in the 100-200 range) PTIs with both UVA and visible light have already been achieved, 34,35,44,[71][72][73][74][75][76][77][78] although oen with higher light doses (up to 100 J cm À2 ) and longer irradiation times.
Upon light irradiation, the NLS derivatives behaved quite differently to the bombesin bioconjugates. Re-NLS and Re-PLPG-NLS possess lower PTIs (2 or less), and show a more pronounced difference between the PLPG-peptide conjugate (Re-PLPG-NLS) and the peptide-only derivative (Re-NLS). In this case, the presence of the PLPG confers a certain advantage. Re-PLPG-NLS had a stronger (2-fold) anti-proliferative effect on malignant HeLa cells compared with Re-NLS. In addition, Re-PLPG-NLS affected non-cancerous MRC-5 cells less, compared with Re-NLS. To explore the origin of these differences, the effect of Re-PLPG-NLS on cells and cellular components was studied in-depth.

Cellular localization
To investigate the biological behaviour of Re-PLPG-NLS in more detail, its cellular localisation was assessed. The known luminescence 1,59,79 of the Re-NH 2 complex allowed us to visualize the sub-cellular localization of its bioconjugates (Fig. 2). While the complex by itself 1 and its photo-caged version Re-PLPG were observed exclusively in the cytoplasm ( Fig. 2B and C), conjugation to an NLS peptide (Re-PLPG-NLS) brought the compound inside the nucleus, as designed (Fig. 2D). However, the compound did not show homogeneous distribution inside the nucleus, but accumulated in a particular sub-nuclear domainthe nucleolus. We previously reported this cellular distribution for Re-NH 2 directly conjugated to NLS peptide. 1 Responsible mainly for ribosomal RNA transcription, processing and assembly, nucleoli also play an important role in cell cycle regulation and stress response. 80 As up-regulated ribosomal RNA synthesis and the consequent enlargement of nucleoli are characteristic of many cancer cells, these organelles have gained attention as promising drug targets in the eld of anticancer research. 81 Quantication of Re-PLPG-NLS uptake in whole cells and nucleoli Due to possible quenching effects inside the cells, uorescence microscopy is not always reliable when it comes to quantifying sub-cellular accumulation of a compound. Such effects with this class of compounds were recently observed by our group. 58,82 The cellular distribution of Re complexes can be quantitatively determined by isolating the cellular compartments of interest and measuring the Re content by ICP-MS. According to this method, we incubated HeLa cells with 20 mM of Re-NH 2 , Re-NLS and Re-PLPG-NLS for 2 h, and collected their pellets (whole cell samples) or isolated their nucleoli through differential centrifugation (nucleoli samples), following an established method with opportune modications (Fig. S23 in the ESI †). 83 Both nucleoli and whole cell extracts were then lysed, the protein content was measured, and the obtained samples were lyophilized and analysed by ICP-MS. The Re content of the whole cell samples followed the order of accumulation Re-NH 2 ( Re-NLS < Re-PLPG-NLS. Interestingly, these data demonstrated very efficient uptake of Re-PLPG-NLS by HeLa. Indeed, in just 2 h, the cells internalised almost half of all Re available in solution. Moreover, evaluation of the accumulation of the target Re compound in the nucleoli indicated that approximately 25% of Re-PLPG-NLS taken up by the cells was concentrated in these organelles (Fig. S23 †). The results obtained from the quantitative determination are in good agreement with the microscopic experiments and emphasize the fast and signicant uptake of the Re complex.

Effect of Re-PLPG-NLS on nucleic acids
In order to achieve a better understanding of the possible mode of action of our Re compound, we investigated the possible damage to DNA or RNA induced by Re-PLPG-NLS in the dark and upon irradiation. Given the high positive charge of Re-PLPG-NLS (5+) at physiological pH and the fact that the NLS sequence is arginine-and lysine-rich, its DNA/RNA aggregating behaviour could be hypothesized. Electrostatic forces indeed play an important role in the interactions of small molecules and proteins with nucleic acids, due to the polyanionic nature of the latter. Arginine and lysine groups in peptides and proteins are heavily involved in H-bonds in DNA/RNA-protein complexes. Most of the small molecule compounds targeting nucleic acids carry multiple positive charges as well. 84,85 DNA photo-cleavage experiments were performed on a closed circular plasmid DNA (namely pcDNA3, see Fig. 3A). The DNA was incubated with various concentrations of Re-PLPG-NLS, and irradiated with UVA for 10 minutes. Re-PLPG-NLS cleaved the plasmids upon light irradiation in a concentration-dependent manner, converting the supercoiled DNA form (intact) to a circular relaxed/nicked form (damaged). Linearization of pcDNA3 was observed at a concentration as low as 10 mM, and complete conversion of the supercoiled to the nicked conguration occurred at a concentration of 50 mM (Fig. 3A). DNA incubated with the rhenium bioconjugate in the dark did not show a signicant effect up to the highest concentration used in this study (50 mM). It was also observed that the plasmid treated with Re-PLPG-NLS displayed a smear band for high concentrations of Re-PLPG-NLS in the dark. This effect is due to NLS/DNA interactions, which produced the deformed band, as conrmed by DNA photo-cleavage experiments on NLS alone (see Fig. S24 in the ESI †).
As RNA is one of the major components of nucleoli, we additionally investigated the effect of Re-PLPG-NLS on a 638 nucleotide long RNA sequence (D135-L14 ribozyme derived from the Sc.ai5g group II intron of S. cerevisiae). 86 Denaturing polyacrylamide gel electrophoresis (Fig. 3B) revealed a considerable decrease in intensity and shi of the RNA band due to Re-PLPG-NLS/RNA interactions and formation of non-migrating RNA agglomerates (Fig. S26B †) at high concentrations of Re-PLPG-NLS (100 mM). This effect occurred both in the absence and the presence of light, and no difference was observed for UVA-irradiated samples. Re-NH 2 or Re-PLPG alone did not affect the RNA (Fig. S27 and S28 †). Given the localization of Re-PLPG-NLS in RNA-rich nucleoli, it is possible that the strong dark toxicity of this compound is linked to its effect on RNA.

Morphological features induced by the targeted rhenium complex
The deep cellular onset provoked by the target Re-PLPG-NLS is not just related to an isolated biochemical pathway or biological structure, but is also mirrored by macroscopic morphological features. Major alterations can be conveniently monitored by transmission electron microscopy (TEM), which is a particularly powerful technique to distinguish sub-cellular compartments, hence giving precious insights into the overall cell stress response. [87][88][89][90] Due to its traceability through confocal microscopy, its valuable anti-proliferative effects both in the dark and upon irradiation, and its strong interaction with DNA and RNA, Re-PLPG-NLS was chosen for this further biological characterization. TEM images of irradiated and non-irradiated HeLa cells incubated for 2 h with Re-PLPG-NLS were recorded. As shown in Fig. 4, the resulting images (selection from a total pool of 40-60 images per sample, see Fig. S29 in the ESI † for other examples) suggested that Re-PLPG-NLS induced the onset of severe cellular stress, with features ascribable to late apoptosis and necrosis. The sudden occurrence of membrane blebs, cytoplasm condensation, organelle packaging, extended vacuolation and nuclear pyknosis indicated an initial triggering of apoptosis. However, the extremely rapid uptake and activity of Re-PLPG-NLS (conrmed by ICP-MS measurements, see

View Article Online
"Quantication of Re-PLPG-NLS uptake in whole cells and nucleoli" section) caused the balance to shi towards necrosisrelated effects, such as loss of membrane integrity and nuclear fragmentation (see Fig. 4). Moreover, coherent with the insights obtained from the cellular localization studies, major attention was paid to the specic effects triggered in nucleoli by Re-PLPG-NLS. Treatment with the rhenium conjugate inuenced the nucleoli morphology. In good agreement with established data, the target sub-organelle presented a stress-altered morphology, characterized by indicators such as the formation of nuclear caps around the nucleolar body, separation of the brillar centre from the granular centre, unravelling of the brillar centre and migration of the whole nucleolar body to the peripheral part of the nucleus (Fig. S30 in the ESI †). 91,92 All these features represent the main characteristics of nucleolar segregation, which could contribute to the initial apoptotic event, as well as to the overall cytotoxicity of Re-PLPG-NLS.

Mechanism of cell death induced by Re-PLPG-NLS
The mobilization of phosphatidylserine from the inner to the outer layer of the cellular membrane promotes phagocytosis and is considered a clear marker of an ongoing early apoptotic event. 93 Under severe stress, the loss of membrane integrity enables a large pool of solutes to penetrate the plasma membrane (which is normally impermeable to them), and represents a signal of the occurrence of necrosis. The combination of two staining methods allows evaluation of the cell death mechanism in large cell populations by ow cytometry, using the Annexin V/propidium iodide (PI) assay. The rapid triggering of cell death processes aer short incubation treatment of HeLa cells with Re-PLPG-NLS investigated by TEM showed a combination of apoptotic and necrotic stimuli. Hence, it was of interest to study the contributions of these two pathways to the cytotoxicity of Re-PLPG-NLS. In agreement with the TEM data, Re-PLPG-NLS induced a drastic cellular effect aer only 2 h incubation with HeLa cells, with less than 10% of cells still viable (see Fig. 5 and S31 in the ESI †). The majority of the cell population displayed membrane symmetry due to phosphatidylserine migration (Annexin V positive) and membrane permeabilisation, which allowed PI internalization (PI positive). The data support the thesis that Re-PLPG-NLS induces severe cellular stress aer a short incubation time, to which cells react by activating apoptosis. However, the very dynamic onset of these events resulted in a pronounced late apoptotic prole (see Fig. 5). These effects could be observed for both probes in the dark and upon irradiation, and were irreversible since the cells did not show any recovery over a time frame of 24 h. Camptothecin was used as a positive control.

Conclusions
In this article, we described the efficient caging of a Re(I) tricarbonyl N,N-bis(quinolinoyl) complex (Re-NH 2 ) with a photolinker (PLPG) to give Re-PLPG. The latter was further conjugated to two different peptides for potential organelle/cellular uptake (i.e. bombesin or NLS) to give Re-PLPG-NLS or Re-PLPG-Bombesin. Photolysis experiments showed that complete release of Re-NH 2 could be achieved using a very low light dose (1.2 J cm À2 ). To the best of our knowledge, these are the rst two examples of the selective release of an organometallic complex from a bioconjugate using light activation. A rst biological screening on different cancer and non-tumorigenic cell lines demonstrated the very valuable toxicity of the target complexes, with IC 50 values in the range of the established anti-cancer drug cisplatin. It is noteworthy that, while Re-PLPG-Bombesin showed a more marked 1 O 2 -dependent cytotoxicity at low irradiation energy (2.58 J cm À2 ), with a certain preference for cancer cells over MRC-5, Re-PLPG-NLS showed a stronger (2-fold) antiproliferative effect on malignant HeLa cells compared with Re-NLS. In addition, Re-PLPG-NLS affected non-cancerous MRC-5 cells less, compared with Re-NLS. Due to this interesting behaviour, Re-PLPG-NLS was chosen for in-depth biological characterization. Cellular biodistribution studies using ICP-MS indicated an outstanding cellular uptake, close to 50% of all available Re, with very rapid and strong accumulation in the nucleoli. The penetration into sub-cellular compartments (nucleus/nucleoli) allowed interaction between the Re complex and nucleic acids. Re-PLPG-NLS was able to photo-cleave DNA and created non-migrating RNA agglomerates. This molecular dysfunction contributed to the induction of late apoptotic and necrotic effects, mirrored by morphological features such as nucleolar segregation, pyknosis and extended vacuolation, which resulted in cell death.
In conclusion, in this article, we present, for the rst time, a strategy (organometallic complex -PLPGtargeting vector) to selectively release an organometallic complex upon light irradiation. Current efforts are aimed at increasing the ratio of dark vs. UV light toxicity and producing this photo-cytotoxic action in the PDT window.

Materials
Chemicals and solvents were of reagent grade or better, purchased from commercial suppliers and used without further purication unless otherwise specied. The plasmid pcDNA3 was obtained from Invitrogen.

Synthesis and characterization
Re-NH 2 . The complex was synthesised following a previously published procedure. The analytical data were in agreement with the literature report. 59 PLPG. 1-5(1-(5-(N-maleimidomethyl)-2-nitrophenyl)ethanol NHS ester). The photo-linker was synthesised following a previously published procedure. The analytical data were in agreement with the literature report. 60 Re-PLPG. PLPG (30 mg, 0.072 mmol) and Re-NH 2 (14.5 mg, 0.072 mmol) were dissolved in dry acetonitrile (1 mL). N,N-Diisopropylethylamine (DIPEA) (23.5 mg, 31 mL, 0.18 mmol) was then added to the mixture, which was stirred for 2 h at 35 C. The solvent was then removed under vacuum to give a sticky brown solid, which was then puried by preparative HPLC to afford Re-PLPG as a light yellow solid. Yield: 58%. Characterisation data: 1  Peptide synthesis. The peptide synthesis was performed using solid-phase techniques following a previously published procedure. 1 NLS (Cys-Arg-DArg-Arg-Lys(CONH 2 )). The peptide was synthesised following a previously published procedure. The analytical data were in agreement with the literature report. 1 Bombesin (Cys-bAla-bAla-Gln-Trp-Ala-Val-Gly-His-Cha-Nle-(CONH 2 )). The peptide was synthesised following a previously published procedure. The analytical data were in agreement with the literature report. 1 Re-PLPG-NLS. Re-PLPG (86.0 mg, 0.0841 mmol) and the NLS peptide (89.5 mg, 0.107 mmol) were pre-dissolved in DMSO (100 mL), and then in a 1 : 2 : 2 mixture of methanol-acetonitrilewater (45 mL). The reaction mixture was stirred overnight at room temperature. MeOH was removed using a rotary evaporator, while water and acetonitrile were removed by lyophilisation. The compound was puried by preparative HPLC to afford a yellow-white solid. Re-PLPG-Bombesin. Re-PLPG (86.0 mg, 0.0841 mmol) and bombesin peptide (89.5 mg, 0.107 mmol) were dissolved in a 1 : 1 mixture of acetonitrile-water (45 mL). The reaction mixture was stirred overnight at room temperature. The solvent was removed using a lyophiliser, and the crude compound was puried by HPLC to afford a yellow-white solid. Determination of photo-uncaging quantum yields 20 mM DMSO stock solutions of the caged compounds were diluted in phosphate buffer (pH ¼ 7.4) to achieve an optical density at 355 nm (OD(l ¼ 355 nm)) of approximately 0.2. Aliquots of 100 mL were irradiated at 355 nm using an Edinburgh LP-920 setup equipped with a Continuum Surelite laser. The laser was slightly misaligned to reach a suitable irradiation power. Aer certain irradiation intervals (number of laser shots), the solutions were transferred to amber HPLC vials with 200 mL inlets. 20 mL of the solutions were injected into an LC-MS system. The percentage of remaining compound was plotted against the number of laser shots and tted with a single exponential decay curve: where C is the remaining concentration of the compound (percentage relative to the initial concentration), C 0 is the initial concentration (percentage), k is the rate constant and x is the irradiation dose.
The initial slope (less than 20% conversion) of the curve (S compound ) was determined by taking the tangent to the rst (<20% decomposition) points of the exponential t. It was then used to calculate the photolysis quantum yield by comparing it to the quantum yield of azobenzene photoisomerization. For that purpose, trans-azobenzene was dissolved in methanol to obtain an OD(l ¼ 355 nm) of approximately 0.2. It was then irradiated under the same conditions as the compounds for the measurements. UV/vis absorbance of the azobenzene solution at 355 nm was then monitored. The remaining amount of transazobenzene was then plotted against the number of laser shots, and the initial slope of the curve (S reference ) was then obtained. The photolysis quantum yields were then calculated using: Concentration % plotted as a function of photolysis time (or irradiation dose) could be tted with a single exponential (rst order kinetics law): where C 0 is the initial concentration, k is the rate constant and x is the irradiation dose.

Cytotoxicity studies
The cytotoxicity of the rhenium complexes and their bioconjugates towards HeLa, MRC-5 and PC-3 cells in the presence and absence of UV irradiation was measured using a uorometric cell viability assay using resazurin (Promocell GmbH). Cells were seeded in triplicate in 96-well plates at a density of 4 Â 10 3 cells per well in 100 mL medium 24 h prior to treatment.
To assess cytotoxicity, cells were treated with increasing concentrations of the compounds for 48 h, whereas, to assess phototoxicity, cells were treated with increasing concentrations of the compounds for 4 h only. Aer that, the medium was removed and replaced by fresh complete medium prior to 10 min UVA irradiation (2.58 J cm À2 ). Cells were then returned to the incubator for 48 h. Aer incubation, the medium was replaced by 100 mL complete medium containing resazurin (0.2 mg mL À1 nal concentration). Aer 4 h of incubation at 37 C (overnight incubation for PC-3 due to slower metabolism of this cell line), uorescence of the highly red uorescent resorun product at 590 nm emission was quantied using a SpectraMax M5 microplate reader with a 540 nm excitation wavelength.

In vitro uorescence evaluation
Cellular localization of the luminescent rhenium complexes and bioconjugates was assessed using uorescence microscopy. HeLa cells were grown in 2 mL medium on 18 mm Menzelglaser coverslips at a density of 1 Â 10 5 cells per mL and incubated with rhenium complexes at 50 mM for 2 h. Cells were xed in 4% formaldehyde solution and mounted on slides for viewing by confocal microscopy on a CLSM Leica SP5 microscope. The rhenium complexes were excited at 405 nm, and emission above 420 nm was recorded.

DNA photo-cleavage
The DNA photo-cleavage effect provoked by complex Re-PLP-NLS was investigated by electrophoresis. Supercoiled pcDNA3 plasmid (0.10 mg) was treated with increasing concentrations of the rhenium compounds in buffer (50 mM

RNA
Preparation of RNA. The plasmid pT7D135-L14, encoding the D135-L14 ribozyme sequence derived from the Sc.ai5g intron from S. cerevisiae, was digested with HindIII and subsequently transcribed in vitro with home-made T7 RNA polymerase under standard conditions 86 to give D135-L14 RNA. D135-L14 was puried by denaturing polyacrylamide gel electrophoresis. 94 Gel electrophoresis studies. D135-L14 86 RNA (1 mM) was incubated with increasing concentrations of the rhenium compounds in buffer (50 mM Tris-HCl, 18 mM NaCl, pH 7.2) for 30 min at room temperature and irradiated at 350 nm for 5 min (Rayomet Chamber Reactor Complex, 1.29 J cm À2 ). Control experiments were carried out in the dark. An equal volume of denaturing loading buffer (11.7 M urea, 40 mM Tris buffer, pH 7.5, 0.1% xylene cyanol, 0.1% bromophenol blue, 230 mM sucrose, 0.8 mM EDTA, pH 8.0) was added to each sample, and denaturing polyacrylamide gel electrophoresis was performed using 5% acrylamide in TBE 1Â buffer for 3 h at room temperature. RNA bands were detected by UV shadowing. Additionally, RNA bands were stained with ethidium bromide and detected by UV excitation at 260 nm (Gel Visualization System from Vilber).

Determination of the rhenium complex content in biological samples
Sample preparation for ICP-MS Whole cells. To perform ICP-MS measurements, HeLa cells were seeded two days before treatment at a concentration of 4 Â 10 5 cells per mL in a 75 cm 2 cell culture ask until 80% conuence, and incubated with the target complexes, namely Re-NH 2 , Re-NLS or Re-PLPG-NLS, at 20 mM for 2 hours. The medium was removed, and the cells were washed with PBS and trypsinized. Aer re-suspension in PBS, the pellets were collected by centrifugation at 5500 rpm for 4.5 minutes. The pellets were re-dissolved in 500 mL of PBS, lysed by a freeze-thaw cycle and treated in an ultrasonic bath (Digitana AG) for 20 minutes. The lysates were lyophilized using an Alpha 2-4 LD plus (CHRIST).
Cell nucleoli. Nucleoli of HeLa cells were collected following an established procedure with appropriate modications. 81 Briey, HeLa cells were seeded two days before treatment at a concentration of 4 Â 10 5 cells per mL in a 75 cm 2 cell culture ask until 80% conuence. Cells were treated with 20 mM Re-NH 2 , Re-NLS or Re-PLPG-NLS for 2 h, and the pellets were collected and washed 3 times with ice cold PBS. The pellets were incubated with lysis buffer (10 mM HEPES, 10 mM KCl, 1.5 mM MgCl 2 , 0.5 mM dithiothreitol, pH 7.9) containing a protease inhibitors cocktail (1/500 v/v, batch number: P8340, Sigma-Aldrich) for 10 minutes, the membranes were disrupted and the cells were homogenized in a 7 mL Dounce Homogenizator with a tight pestle (type A, 20 strokes). The suspensions were centrifuged at 250g at 4 C in a Centrifuge 5810R (Eppendorf) for 5 minutes, and the supernatant was removed. The pellets were re-dissolved in 3 mL of a sucrose solution (0.25 M sucrose, 10 mM MgCl 2 ) and layered with 3 mL of a second hypertonic sucrose solution (0.35 M sucrose, 0.5 mM MgCl 2 ). The suspensions were centrifuged at 1450g at 4 C for 5 minutes. The supernatant was discarded, and the pellets were re-suspended in 3 mL of the second sucrose solution. The suspensions containing the nuclear fraction were sonicated 5-6 times in 10 s bursts (Bandelin Sonoplus GM70, 50% cycle, intervals of 10 s), layered with a third hypertonic sucrose solution (0.88 M sucrose, 0.5 mM MgCl 2 ) and centrifuged at 3000g at 4 C for 10 minutes. The pellets were re-suspended in the second sucrose solution (0.35 M sucrose, 0.5 mM MgCl 2 ), centrifuged at 1450g at 4 C for 5 minutes to obtain the pure nucleoli extracts (see Fig. S17 †), and stored at À80 C. All the steps of the isolation procedure were monitored under a phase contrast microscope on Menzelgläser coverslips (Olympus IX81). The collected nucleoli solutions were lyophilized using an Alpha 2-4 LD plus (CHRIST).
ICP-MS measurements. To measure the Re content, the biological samples were sent to Mikroanalytisches Labor Pascher (Remangen, Germany). The samples were dissolved in a PFA pressure system with nitric and hydrochloric acid at 180 C, and Re187 (Ir193 as internal standard) was detected by ICP-MS (Perkin Elmer Elan 5000).

Transmission electron microscopy
HeLa cells were seeded in 6-well plates containing carboncoated sapphire discs and glass cover slips at a concentration of 100 000 cells per mL and cultured at 37 C/5% CO 2 for one day (>80% conuence on the discs). The medium was then removed and replaced with medium containing 20 mM Re-PLG-NLS. The cells were further incubated for 2 hours and irradiated at 350 nm for 10 minutes (2.58 J cm À2 ) or le non-irradiated (control). Subsequently, the cells were processed by chemical xation and high-pressure freezing, as follows.
Chemical xation (Fig. S23 †). Cells were xed for 20 minutes at room temperature with 2.5% glutaraldehyde and, subsequently, with 1% OsO 4 for 1 hour, both in PBS (pH 7.3). The cells were then incubated with 1% uranyl acetate in water for 1 hour, dehydrated in an ethanol series and embedded into Epon/ Araldite.
High-pressure freezing (Fig. 4). Cells on sapphire discs (6 mm diameter) were frozen in an EM HPM100 high-pressure freezing machine (Leica Microsystems, Vienna, Austria) using the 6 mm cartridge system and a middle plate with ridge. Subsequently, the samples were freeze-substituted in an EM AFS2 unit (Leica) for 8 h at À90 C, 6 h at À60 C, 4 h at À30 C and 1 h at 0 C, with periodic temperature transition gradients of 30 C h À1 . Aerwards, the cells were incubated for 1 hour in 1% uranyl acetate in anhydrous acetone at 4 C prior to embedding in Epon/Araldite. Ultrathin 50 nm sections of all specimens were contrasted with uranyl acetate and lead citrate, and analysed with a Tecnai G2 Spirit or CM 100 transmission electron microscope (FEI, Eindhoven, The Netherlands) using an ORIUS 1000 CCD camera (Gatan, Munich, Germany).

Annexin V/PI assay
The induction of apoptosis and necrosis in HeLa cells treated with Re-PLPG-NLS was evaluated by means of the well-established Annexin V/PI assay, using ow cytometry (BD Pharmingen, Bioscience) according to the manufacturer's instructions with minor changes (BD Pharmingen, Mat. no. 556419). Briey, cells were seeded in a Petri dish at a concentration of 2 Â 10 5 cells mL À1 and cultured at 37 C/6% CO 2 for 24 h. The medium was replaced with fresh medium containing 20 mM Re-PLPG-NLS. The cells were incubated for 2 h, and then the medium was replaced with fresh medium and the Petri dish was irradiated at 350 nm for 10 minutes (2.58 J cm À2 ). To observe the direct effects aer irradiation, the cells were immediately collected for FACS analysis, whereas, for the recovery experiments, the medium was replaced and the cells were further incubated for 24 h at 37 C/6% CO 2 . The cells were then trypsinized and pelleted, washed twice with ice cold PBS, centrifuged at 250g for 5 minutes and re-suspended in binding buffer (10 mM HEPES, 140 mM NaCl, and 2.5 mM CaCl 2 , pH 7.4) at a concentration of 1 Â 10 6 cells per mL. 100 mL of the obtained suspension was transferred into an FACS culture tube (1 Â 10 5 cells), and 5 mL of FITC Annexin V complex solution and 5 mL of a 50 mg mL À1 propidium iodide (PI) solution were added. Samples were incubated for 15 minutes at room temperature (25 C) in the dark, 400 mL of binding buffer was added, and the probes were analyzed using a CynAn ADP9 ow cytometer with the FITC (for Annexin V-FITC, excitation ¼ 488 nm, emission ¼ 515-545 nm) and PE-Texas Red channels (for PI, excitation ¼ 488 nm, emission ¼ 564-606 nm). The data were analyzed using Summit v4.3 soware. Camptothecin, used at 5 mM for 2 or 24 h, and untreated cells in the dark or under irradiation were used as positive controls.