Flexible synthesis of cationic peptide–porphyrin derivatives for light-triggered drug delivery and photodynamic therapy

Amphiphilic cell-penetrating peptide–porphyrin conjugates have been developed for application in light-based therapeutic techniques.


Introduction
Photodynamic therapy (PDT) is a light-based therapeutic technique that shows great promise for the selective treatment of a variety of cancers and non-malignant conditions. 1 The principle of PDT is that highly localised destruction of tumours, diseased tissue or pathogenic organisms is achieved with light following the administration of a light-activated photosensitising drug or photosensitiser.This generates various cytotoxic reactive oxygen species such as singlet oxygen that can interact with and damage cellular components, leading to cell death.Many photosensitisers have been developed and have received approval or clinical trials for use in the treatment of cancer, 2 age-related macular degeneration, 3 pre-cancerous conditions such as Barret's oesphagus, 4 and oral sterilisation in dental procedures. 5Among these, porphyrins and other tetrapyrrolerelated compounds remain the broadest class of compound under study for PDT and other light-based therapies. 6The effectiveness of PDT in a particular application depends upon the wavelength at which the photosensitiser may be activated, and this has led to the development of many modified tetrapyrroles that show enhanced absorption at the red end of the spectrum where absorption by tissue is relatively weak, thus allowing a photodynamic effect at a greater depth in vivo. 6,7nother important consideration in the development of novel photosensitisers is their efficient delivery to the target site, in particular achieving selective delivery of photosensitisers to tumours compared to normal tissues.In this context, in recent years there has been considerable interest in the development of peptide and protein-targeted photosensitisers as a means of improving the pharmacokinetic properties, solubility and tissue specificity of various otherwise hydrophobic derivatives.Conjugation of photosensitisers to antibodies and a range of synthetic peptides has been explored and has been found to provide significant enhancements in both the efficiency and selectivity of cellular uptake of photosensitisers for PDT in a range of cancer models. 8A key innovation in the development of such targeted photosensitisers has been the application of biorthogonal ligation techniques 9 which facilitate the efficient attachment of a variety of photosensitisers to unprotected, multifunctional peptides in solution.
1][12] Such molecules, of either natural or synthetic origin typically consist of 8-30 amino acid residues and possess the ability to translocate across biological membranes and transport diverse molecular cargoes, either covalently or non-covalently attached, which would otherwise be poorly internalised.The conjugation of tetrapyrrole-based photosensitisers and other derivatives to CPPs has indeed been shown to provide an attractive means for enhancing photosensitiser delivery for both PDT of cancer [13][14][15] and in antimicrobial PDT applications. 16,17Importantly, CPP-conjugation also offers a means to control the sub-cellular localisation of a particular photosensitiser in eukaryotic systems to potentially maximise therapeutic effect. 10,14e recently demonstrated that the conjugation of a porphyrin derivative to a cationic CPP is an effective way of generating a novel water-soluble amphiphilic photosensitiser suitable for light-triggered drug delivery by photochemical internalisation (PCI). 18PCI is a novel technology for effecting the light-activated intracellular release of biologically active molecules at specific sites in living tissue, often where conventional methods of drug administration prove unsuccessful. 19,20The effectiveness of various macromolecular therapeutic agents can often be severely impaired by their limited ability to reach their intracellular targets due to sequestration in endosomes and lysosomes after uptake by endocytic mechanisms.PCI provides a highly effective means to selectively trigger endosomal escape and intracellular relocalisation of entrapped drugs, by applying the basic principle of the PDT approach.In PCI, a sub-lethal light dose is applied to a photosensitiser that localises in endo/lysosomal membranes, which is sufficient to cause partial rupture of these intracellular organelles (mediated by singlet oxygen).This allows the entrapped drugs to escape but does not compromise the viability of the cells themselves. 20Recent clinical studies have demonstrated both the feasibility and safety of PCI for the treatment of advanced and recurrent solid malignancies. 21 critical requirement in PCI is that the photosensitiser used must possess features that make it localise in the same intracellular vesicles (lysosomes, endosomes) as the administered drug, i.e. they must be lysosomotropic and therefore amphiphilic. 20Many well-known PDT photosensitisers are unsuitable for PCI as they partition non-selectively to other cellular organelles (e.g.mitochondria, Golgi apparatus, endoplasmic reticulum), however as shown in Fig. 1, attachment of an otherwise hydrophobic photosensitiser derivative to a cationic CPP may transform it into an amphiphilic compound that is both water-soluble and amenable to cellular uptake by endocytosis. 18The hydrophobic porphyrin macrocycle then has the potential to localise in the lipid bilayer of the endosomal membranes, for selective oxidative damage, 22,23 with the hydrophilic peptide in the fluid phase. 18These properties are ideally suited for PCI (see above).
The application of CPP-targeted photosensitisers for PCI and PDT depends on the availability of reliable synthetic protocols to derivatise typical cationic CPPs in a site-specific fashion 24 with easily accessible functionalised photosensitisers of maximum clinical potential. 25The aim of this study was to investigate the potential of a number of bioconjugation reactions for the efficient modular preparation of CPPporphyrin conjugates, as a general strategy to convert classical PDT agents into amphiphilic conjugates suitable for targeted PDT and PCI.

Photosensitiser synthesis
The preparation of the ligatable photosensitisers used in this study is outlined in Scheme 1.As in previous studies by us, meso-tetrakistetraphenylporphyrin 1 provided the template for these compounds, and was transformed to the mono-amino derivative 2 by a modification of the method of Luguya et al. 26 Treatment of 1 with sodium nitrite and trifluoroacetic acid under the reported conditions (3 min contact time) gave a mixture of mono-and dinitrated porphyrins that were reduced directly with sodium borohydride and palladium on charcoal.2 was then easily separated from disubstituted products and unchanged 1 in reproducible yields of around 30% over the two steps.The porphyrin scaffold was then elaborated with several functionalities that would permit biorthogonal ligation with a suitably modified CPP component, via acylation with a suitable carboxylic acid component using either EDC/HOBt or PyBOP activation. 27Compounds 3-8 were obtained in satisfactory yields via either activation protocol, and in the cases of 3 and 8 were used directly with their complementary ligatable CPP partners (see below).The azido and alkynyl derivatives 5-7 were further converted into the corresponding zinc complexes by treatment with Zn(OAc) 2 in DCM in excellent yields, ready for use in copper-catalysed ligation reactions.

Peptide synthesis
All peptides were synthesised by 9-fluorenylmethoxycarbonyl (Fmoc) peptide synthesis on Rink amide resin, using PyBOP activation.Peptides 10-14 were based on the decapeptide CPP sequence derived from the transcriptional activator (Tat) protein from human immunodeficiency virus 1 (HIV-1), commonly known as HIV-1 Tat(48-57), 28 which provided the template for N-terminally ligatable derivatives in this study.In each case, the final peptide was functionalised on-resin to introduce a complementary reactive group to permit one of the following ligation chemistries with the appropriate porphyrin derivative.Peptides 10-12 were obtained via straightforward acylation of the resin-bound peptide with either Fmoc-L-Cys-OH (for 10), azidoundecanoic acid (for 11), or pentynoic acid (for 12), using HATU activation. 27For peptide 13, an N-terminal azido function was introduced by direct diazo transfer upon the resin-bound Tat sequence using the Stick reagent (imidazole-1-sulfonyl azide hydrochloride), 29 with completion of the reaction being assessed by the Kaiser test, 30 as for the on-resin acylations to generate 10-12.In the case of the keto-functionalised derivative 14, acylation of the resinbound peptide could only be satisfactorily achieved using the preformed succinimido ester of pyruvic acid. 31Other activation methods (DIC, PyBOP, HATU) failed to give desired product, presumably due to the tendency of keto acids such as pyruvate to undergo Claisen-type condensations under strong carboxyl activation conditions. 32The C-terminally ligatable peptides 15-18 were obtained directly by standard solid phase peptide synthesis.Cys-containing peptide 15, was originally reported by Santra et al. 33 and is also based upon the HIV-1 Tat(48-57) sequence.It provided the template for our prototype CPPtargeted photosensitiser for PCI and applications in anti-microbial PDT. 17,19Azidopeptide 16 was derived from 15, by replacement of the terminal Cys residue with ε-azidolysine 34 and insertion of a Gly residue at the C-terminus itself, to facilitate loading of the Rink amide resin.Azidopeptides 17 and 18 were obtained in an analogous fashion by respectively the replacement of one Lys residue in the sequence of pVEC, a CPP derived from murine vascular endothelial-cadherin protein, or a fragment of penetratin, a CPP derived from the Antennapedia homeoprotein. 12The final N-terminally functionalised CPPs 11-14 were obtained in yields of 35-58%, and the C-terminally functionalised derivatives 15-18 in yields of 59-67% after cleavage from the resin and side chain deprotection with TFA/TIS/H 2 O or TFA/TIS/H 2 O/EDT, for Cys-containing peptides 10 and 15 (Scheme 2).

Ligation reactions
Thiol-maleimide and oxime ligation.There are many examples of the ligation of porphyrin-type photosensitisers to multifunctional peptides and proteins with a suitably placed Cys residue via conjugate addition to a maleoyl-type or other α,β-unsaturated carbonyl function. 9,35Recently, Bryden et al. 36 have used dibromopyridazinediones as the electrophilic unit as part of a ligation strategy to couple cationic porphyrins to a targeting antibody.In order to compare the efficiency of N-and C-terminal ligation with the cationic Tat sequence, peptides 10 and 15 were reacted with a two-fold excess of porphyrin 31 17,37 in DMSO as shown in Scheme 3.This furnished the desired CPP conjugates 19 and 20 in excellent yields, with no significant difference in the efficiency of N-terminal or C-terminal attachment.This is in contrast to a report by Kitagishi et al. 37 who observed that an analogous ligation to the N-terminus of an octaarginine derivative proceeded only in very low yield.Indeed, using an alternative polar ligation at the N-terminus of the Tat-CPP unit via keto peptide 14 also proved to be highly effective.In this case, the Boc-protected porphyrin 4 was first treated with TFA in DCM to generate aminoxyderivative 9, which isolated as the free base in 94% yield.The complementary fully deprotected peptide 14 was then treated with a four-fold excess of 9 in 0.1% TFA/DMSO to give the desired oxime conjugate in 75% yield following purification by semipreparative HPLC.DMSO proved to be the solvent of choice for these ligations as a compromise between the otherwise mutually exclusive solubilities of the highly hydrophobic porphyrins and the unprotected peptides, since preliminary studies with water/PEG systems were found to give very sluggish reactions and variable yields.Notwithstanding this, it was typically possible to greatly simplify the isolation of 19-21 and similar conjugates by precipitation of the crude products or removal of the reaction solvent by isolation on a solid-phase extraction cartridge prior to semi-preparative HPLC.
Copper-catalysed azide-alkyne cycloaddition (CuAAC).Various examples of the ligation of photosensitisers to peptide derivatives using copper-catalysed azide-alkyne cycloaddition (CuAAC) have been described, 38 however the potential of this approach appears to have been very little exploited with polycationic CPPs as the targeting agent. 39,40Although maleimide and oxime ligations proved quite satisfactory for our applications, ligation by copper-catalysed CuAAC offered a number of advantages.As well as generating a triazole-containing linkage that is completely stable under physiological conditions, an additional attractive feature was the possibility to easily introduce one or other of the azide and alkyne functions into either photosensitiser or CPP components, at the N-or C-termini, and with a range of spacer units As a direct comparison for the chemistry of N-terminal attachment, the alkynyl and azido porphyrins 5 and 6 were first combined with the corresponding functionalised Tat peptides 11 and 12 respectively to generate isomeric conjugates 22 and 24 that differ only in the location of the triazole moiety within the linker (Scheme 3).Preliminary tests aimed at devising the optimum reaction conditions confirmed the need for the Zn(II) complexes of both 5 and 6 to avoid sequestration of the copper catalyst by the macrocycle. 41When peptides 11 and 12 were thus reacted with an excess of complexes Zn-5 and Zn-6 in the presence of CuSO 4 and sodium ascorbate in DMSO/H 2 O/ tBuOH, the N-linked corresponding conjugates were successfully obtained in reproducible yields of 45% and 77% respectively.Conjugates 22 and 24 were isolated by preparative HPLC, and the zinc was easily and quantitatively removed from the porphyrin unit by diluting the reaction mixture with aqueous TFA prior to purification.When the porphyrin component was replaced with azido derivative Zn-7, applying the same reaction conditions with Tat peptide 11 also gave the desired conjugate 25 in very good yield (78%); shortening the spacer between the porphyrin unit and the azido function thus appeared to have no serious adverse effect on the cycloaddition reaction.No improvement in yields was observed when the copper source was replaced with a Cu(I) reagent in these reactions (e.g.Cu(OTf )), 42 although the conjugations also proceeded, though generally less efficiently, in the absence of tert-butanol as cosolvent.Reaction of alkynyl-porphyrin 5 and peptide 13, in the presence of CuSO 4 and ascorbic acid in DMSO in fact gave the desired conjugate again in very high yield (78%).Thus, even incorporating the azido function directly at the N-terminus of the polyfunctional CPP sequence (mutating Gly to azidoacetyl) appeared to be well tolerated for this ligation.
Strain-promoted azide-alkyne cycloaddition (SPAAC).While it was found to be possible to efficiently ligate simple azido or alkynyl porphyrins to the N-terminus of a complementary Tat peptide in either sense via CuAAC chemistry, a general drawback of this approach is the need to protect the porphyrin macrocycle via pre-formation of a Zn complex.Furthermore, preliminary studies with other porphyrin and chlorin derivatives suggested that significant reaction optimisation might be required on a case-by-case basis depending on the precise photosensitiser structure. 43ecent advances in biorthogonal ligation chemistry [44][45][46] have yielded a range of novel variants on the original triazole approach in which the use of a strained alkyne component removes the need for copper catalysis, and this has indeed recently been shown to be suitable for generating specific porphyrin-antibody conjugates. 36We chose to investigate SPAAC ligations for the effective synthesis of both N-and C-terminally linked CPP-porphyrin conjugates, and as described above, porphyrin derivative 8 bearing a strained dibenzocyclooctyne (DBCO) function could be readily prepared by acylation of 2 with the commercially available carboxylic acid derivative.Owing to the chemical reactivity of the DBCO unit, the preferred sense of combination was with the azido function in the peptide component as detailed above, either through N-terminal acylation with a suitable azidoalkanoic acid (Tat peptide 12), or by incorporation of a C-terminally located azidolysine ( peptides 16-18).Reaction of 2 eq. of DBCO porphyrin 8 with the N-terminal azido Tat peptide 12 proceeded smoothly in DMSO to give the expected conjugate 26 in 65% yield as an inseparable mixture of triazole regioisomers. 44gain there was no significant loss of efficiency upon switching from an N-to C-terminal ligation: combination of 8 with the relevant C-terminally ligatable Tat peptide 16 also proceeded effectively to give conjugate 27 in 50% yield after HPLC isolation.When 8 was combined with the penetratin and pVEC derivatives 17 and 18, once again the desired C-terminally ligated conjugates 28 and 29 were obtained with high efficiency irrespective of the length of the peptide component and without requiring a significant excess of 8 (see Scheme 3).

Uptake and subcellular localisation of CPP conjugates
The efficient uptake and endolysosomal localisation of the amphiphilic CPP-photosensitiser conjugates was determined by microscopy in MC28 rat fibrosarcoma cells.All the N-and C-terminally linked conjugates studied showed the expected sub-cellular localisation predicted for their overall amphiphilic character (see Fig. 1) and which would render them potentially suitable photosensitisers for PCI applications.Fig. 2 shows typical images demonstrating the efficient uptake of C-linked derivative 20.This parallels the behaviour of 20 previously observed in the clinically important HN5 head and neck cancer cell line, wherein greatly enhanced cellular uptake of the porphyrin moiety is provided by the function of the CPP to which it is conjugated. 18Endolysosomal localisation of the conjugate was confirmed by colocalisation of the porphyrin fluorescence with Lysotracker Green as visualised by the yellow colour in the merged image in Fig. 2C.All the conjugates were highly water soluble, unlike the parent porphyrin 1 and functionalised but non-ligated porphyrins 2-8.

Phototoxicity
The localisation of CPP-targeted porphyrin derivatives in lysosomal compartments should result in a highly efficient targeted PDT effect, if as desired, the photosensitiser is localised in the membrane of these vesicles where singlet oxygen generated can effectively photo-oxidise unsaturated lipid materials. 20The photoxicities of selected N-and C-linked conjugates were therefore examined in monolayer culture in MC28 and MCF-7 human breast cancer cells.A significant photoinduced reduction in cell viability was observed with both Nand C-terminally linked maleoyl conjugates 19 and 20 in MC28 cells, which was enhanced with increased photosensitiser concentration and dose of blue light.Comparison of 19 and 20 however revealed little difference in performance as a result of switching the photosensitiser location within the peptide carrier (Fig. 3).
Importantly, Fig. 3 shows that control experiments in the dark for both conjugates resulted in no chemical toxicity at concentrations that were significantly greater than those typically employed for in vitro PCI experiments (see below).This confirmed that the observed reduction in cell viability was light-induced rather than being associated with any membrane-disrupting effects due to the CPP moiety.The triazole conjugates 24 and 26 also displayed a concentration and light dose-dependent increase in the PDT effect (Fig. 4), again  without dark toxicity under the experimental conditions.Table 1 shows the porphyrin dose required to induce 50% toxicity (LD 50 ) after 5 min illumination for the four conjugates 19, 20, 24, and 26.These values show that while the PDT performance is indeed similar for 19 and 20 (N-vs.C-terminal porphyrin conjugation), connecting the hydrophobic porphyrin and the polycationic hydrophilic peptide components via the more extended triazole-based linker of 26 seems to be highly beneficial.Notwithstanding this observation, the spectroscopic properties of the porphyrin unit did not appear to be affected by peptide ligation for these or any of the conjugates (see ESI †).
Finally, the C-terminally linked triazole conjugate 29 exemplifying SPAAC ligation was evaluated in MCF7 cells.Once again, an effective concentration and light dose-dependent reduction in cell viability was observed (Fig. 5), with LD 50 = 449 nM for 7 min irradiation.As expected, no dark toxicity was shown when the cells were incubated with 29 at various concentrations up to 1 µM.

Light-triggered drug delivery
Two of the N-terminally linked Tat conjugates, 19 and 26 were studied further for their ability to promote light-induced relocalisation and endosomal escape of a co-administered protein toxin.Saporin is a 30 kDa ribosome inactivating protein, which has been widely used in neuronal research and as a model compound for PCI studies.Upon cellular uptake, it is ordinarily entrapped within lysosomes, thus severely limiting its toxicity.PDT treatment and PCI of saporin with 19 and 26 resulted in almost the same level of cytotoxic response for each compound (Fig. 6).This data shows that for the PCI experiments the effect of combining either conjugate with saporin is synergistic rather than additive.Indeed, with both conjugates a significant reduction in cell viability was observed in PCI-treated cells versus PDT-treated cells ( p < 0.0001).For example at 5 min illumination in Fig. 6B for 26, PCI resulted in a four-fold reduction in viability compared to PDT.PCI efficacy is defined as the ratio of the viability measured using PDT over the PCI viability, and since the reduction in viability using saporin alone is small (approximately 10% in either case) this ratio gives a measure of the enhancement in saporin cytotoxicity induced by PCI. 47In the absence of effective light-triggered endolysosomal disruption and saporin release, the PDT : PCI ratio would be near unity.The ratios observed here of 3.3 for 19 and 4.0 for 26 showing that effective PCI is induced and to similar extents in both conjugates, notwithstanding the difference in the CPP-photosensitiser linkage.The method of Luguya et al. 26 was modified as follows.
General procedure: a stirred solution of amino derivative 2 (0.08 mmol) in anhydrous DCM (2.5 mL) was treated with the carboxylic acid (0.16 mmol), followed by EDC (0.16 mmol), HOBt (0.16 mmol) and DIPEA (0.47 mmol).The solution was protected from light and stirred under nitrogen at room temperature overnight.The reaction mixture was diluted with DCM, washed with 0.1 M HCl, saturated aq.NaHCO 3 , and brine.The organic phase was dried (MgSO 4 ), filtered, and the solvent was evaporated.The crude product was purified by column chromatography (0-20% EtOAc/DCM gradient).
Following the acylation step, the resin beads were filtered off under vacuum, and washed thoroughly with DMF, DCM, MeOH, and Et 2 O, and dried in vacuo.For the cleavage and deprotection of the ligatable peptides, the acylated peptidyl resin was swollen in DCM for 20 min, then it was treated with TFA/ TIS/H 2 O (95/2.5/2.5) for 4 h.Cysteine-containing peptide (10)  was instead treated with TFA/TIS/H 2 O/EDT (95/2.5/2.5/1) for 3 h.The resin beads were filtered off, washed with TFA, and the combined filtrates were evaporated to a small volume then anhydrous Et 2 O was added.The resulting precipitate was collected by centrifugation and was washed twice more with Et 2 O.The precipitated material was dissolved in 1% aq.TFA, filtered using a 0.2 μm syringe filter and the resulting solution was directly purified by semi-preparative HPLC.The purified peptides were then freeze-dried.
( deprotection was removed as described above.The resin was washed, and then swollen again in mixture of DCM/MeOH/H 2 O (2/2/1) for 1 h.The resin was then treated with imidazole-1-sulfonyl azide hydrochloride (121 mg, 0.7 mmol) DIPEA (273 μL, 1.40 mmol), and CuSO 4 •5H 2 O (17 mg, 0.07 mmol) in DCM/MeOH/H 2 O (2/2/1, 2 mL).The resin was shaken overnight, then the solution was discarded and the beads were washed with deionised water, MeOH, and DCM.Cleavage from the resin and isolation was carried out as described above to give 12 (24 mg, 55%), t R (Method Preparation of C-terminally ligatable peptides ( 15)- (18).For peptides 16-18, the azidolysine-containing Tat, penetratin and pVec sequences were assembled as described above on Rink Amide MBHA resin (0.25 g, 0.60 mmol g −1 For each ligatable peptide, a sample of the peptide resin (50 mg) was swollen in DCM, then the N-terminal Fmoc protection was removed.Cleavage from the resin and side chain deprotection were achieved by treatment with TFA/TIS/H 2 O (95/2.5/2.5/1)(3 mL).All the solvents were evaporated under vacuum and the residue was dissolved in TFA and cold Et 2 O was added.The white solid precipitated was isolated by centrifugation.16 and 18 were obtained directly in >90% purity as judged by HPLC (see ESI †).For 17, the precipitate collected by centrifugation was dissolved in 0.1% aq.TFA, filtered using a 0.2 μm syringe filter the resulting solution was directly purified by semipreparative HPLC.The purified peptide were then freeze-dried.C-terminal Cys-containing peptide 15 was obtained in analogous fashion to the N-terminal Cys-containing peptide 10 (see above).

Cell lines and cultivation
MC28 cells, a methylcholanthrene-induced fibrosarcoma cell line, were grown in DMEM supplemented with 10% FCS.Human breast cancer cells (MCF-7) were grown in DMEM-F12 medium containing 10% FCS.Cells were incubated at 37 °C in a humidified atmosphere containing 5% CO 2 .Unless otherwise stated materials for the cell studies were purchased from Sigma-Aldrich Gillingham, UK.
Cellular uptake and confocal microscopy.MC28 cells were seeded in small Petri dishes with a glass cover slip bottom (Fluorodish, World Precision Inst.UK) and allowed to attach overnight.Cells were then incubated for 24 h with conjugate 20 (2.5 µM).Afterwards, culture medium was removed and replaced with fresh medium containing LysoTracker Green (100 nM) 30 min before microscope imaging.Cells were then washed 2 times with PBS and incubated with drug-free/phenol red-free medium for the confocal imaging using an Olympus laser scanning confocal microscope (FluoView FV1000, 60× magnification, NA 1.20, Olympus UK Ltd, Essex, UK).Fluorescence from the photosensitiser was recorded within the range of 620-720 nm using a 405 nm laser for the excitation.For the LysoTracker Green imaging, cells were illuminated at 488 nm and the fluorescence signal was recorded at 500-550 nm.Colocalisation analysis and image processing were performed with ImageJ software.
In vitro PDT phototoxicity study.MC28 and MCF-7 cells were seeded out at a density of 5000 cells in 96 well plates.The cells were allowed to attach to the bottom of the wells for 24 h.The next day, cells were incubated with various concentrations of photosensitiser for 18 h.The cells were then washed twice with PBS and incubated with drug-free culture medium for 4 h.Afterwards, the cells were exposed to increasing doses of light using a blue LumiSource® flatbed lamp with peak emission at 420 nm and 7 mW cm −2 output (PCI Biotech, Oslo, Norway).Cell viability was evaluated 48 h after light illumination using a standard MTT (3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide) assay.Each experiment was carried out in triplicate.
In vitro PDT/PCI phototoxicity study.MC28 cells were seeded out at a density of 3000 cells in 96 well plates overnight.then treated with 19 or 26 (50 nM) or saporin (Sigma-Aldrich Gillingham, UK) (20 nM) for 18 h separately.Another group of cells were co-incubated with saporin (20 nM) and 19 or 26.Cells were then washed twice with PBS and incubated for a further 4 h with fresh full medium.Irradiation was carried out for up to 5 min using a blue LumiSource® flatbed lamp with peak emission at 420 nm and 7 mW cm −2 output (PCI Biotech, Oslo, Norway).Cell viability was determined by means of the MTT assay 72 h after the light exposure.Each experiment was carried out in triplicate.

Statistical analysis
Data were analysed using two-tailed Students T-test with appropriate testing post hoc using Prism 6 software.The dose required to reduce viability by 50% (LD 50 ) was also calculated using Graphpad Prism.Error bars from the mean show ±standard deviation (SD).Values of P < 0.05 were considered to be significant.

Conclusions
Bioconjugation strategies offer a highly efficient and flexible means to generate conjugates between hydrophobic porphyrins and polyfunctional CPPs under very mild conditions.These amphiphilic photosensitisers with regiospecific attachment of a tetrapyrrole component within the peptide backbone are effective agents for both targeted PDT and PCI.As exemplified in this study, strain-promoted azide-alkyne ligations offer an ideal way to repurpose simple porphyrin derivatives for PCI by attachment to a variety of CPPs.This approach opens the way for adapting clinically relevant photosensitisers with the most attractive spectroscopic properties that are currently used in PDT for minimally invasive therapies in conjunction with selected targeted CPPs.

Fig. 1
Fig. 1 Amphiphilic CPP-targeted photosensitiser.The hydrophobic photosensitiser unit localises in the lipid bilayer of endosomal/lysosomal vesicles while the cationic peptide carrier resides in the fluid phase.

Fig. 2
Fig. 2 Cellular uptake and colocalisation of conjugate 20 with LysoTracker Green in MC28 cells using laser scanning confocal microscopy.Cells were incubated with 20 (2.5 µM) for 24 h.LysoTracker Green (100 nM) was applied to cells 30 min before imaging.A: 20 alone, B: LysoTracker Green, C: merged A and B inset figure highlighting the overlap of fluorescence from 20 and LysoTracker Green.Scale bar: 20 µm.

Fig. 3
Fig.3PDT effect of 19 (A) and 20 (B) in MC28 cells.Cells were incubated with Tat-porphyrin conjugates at various concentrations and were illuminated with a blue lamp for up to 5 min.MTT assay was carried out 48 h after light exposure.Data are presented as mean value ± standard deviation (SD) of three independent experiments.

Fig. 4
Fig.4PDT effect of 24 (A) and 26 (B) in MC28 cells.Cells were incubated with Tat-porphyrin conjugates at various concentrations and were illuminated with a blue lamp for up to 5 min.MTT assay was carried out 48 h after light exposure.Data are presented as mean value ± SD of three independent experiments.

Fig. 5
Fig.5PDT effect of 29 in MCF-7 cells.Cells were incubated with photosensitiser at various concentrations and were illuminated with a blue lamp for up to 7 min.The MTT assay was carried out 48 h after light exposure.Data are presented as mean value ± SD of three independent experiments.

( 1 .
60 g, 2.60 mmol) in TFA (80 mL) in an air-open roundbottom flask was treated at 18 °C with NaNO 2 (321 mg, 4.64 mmol).After 3 min, the reaction was quenched with H 2 O (40 mL), then transferred to a separatory funnel and further diluted with H 2 O (80 mL) and DCM (160 mL).After separation, the aqueous phase was extracted with DCM (2 × 150 mL).The organic phases were combined, washed with saturated aq.NaHCO 3 (2 × 175 mL) and brine (175 mL), then dried over MgSO 4 .The combined organic phases were filtered and evaporated to give a residue (1.89 g), which was used directly in the next step.Reduction: a solution of the preceding crude mixture of unreacted 1 and mono-and dinitrated products in DCM (640 mL) and MeOH (160 mL) in an air-open round-bottom flask was treated with 5% Pd/C (320 mg).NaBH 4 (241 mg, 63.7 mmol) was added in small portions with stirring during 10 min, then the mixture was stirred for an additional 15 min.The reaction was quenched by the addition of H 2 O (150 mL), and the resulting suspension was transferred to a separatory funnel.The organic phase was separated, washed with brine (100 mL), dried over MgSO 4 , filtered and evaporated.The crude product was purified by column chromatography, eluting first with toluene to remove unreacted 1 (589 mg, 37% recovery), then 0-10% EtOAc in DCM to elute 2 (564 mg, 34% from 1) and then diaminoporphyrins (368 mg, 22% from 1).

Fig. 6
Fig. 6 Light-induced cytotoxic response for 19 (A) and 26 (B) in MC28 cells, showing PDT (without saporin) and PCI (with saporin) effects.Cells were incubated with photosensitiser (50 nM) with or without saporin (20 nM) for 18 h.Irradiation was carried out for 3 and 5 min with a blue lamp.MTT assay was carried 72 h after irradiation.The experiments were repeated 3 times, and representative data are shown.Error bars = SD.Statistically significant difference between PCI conjugate 19-saporin versus PDT conjugate 19 ( p < 0.0001) is indicated by ***.Likewise, statistically significant difference between PCI conjugate 26-saporin versus PDT conjugate 26 ( p < 0.0001) is indicated by ***.