The FASEB Journal express article 10.1096/fj.02-0888fje. Published online April 8, 2003. Recombinant modular transporters for cell-specific nuclear delivery of locally acting drugs enhance photosensitizer activity Andrey A. Rosenkranz,*,� Vladimir G. Lunin,*,� Pavel V. Gulak,* Olga V. Sergienko,� Maria A. Shumiantseva,� Olga L. Voronina,�,§ Dinara G. Gilyazova,� Anna P. John,¶ Anna A. Kofner,# Andrey F. Mironov,# David A. Jans,¶,** and Alexander S. Sobolev*,�

*Department of Molecular Genetics of Intracellular Transport, Institute of Gene Biology, Russian Academy of Sciences, 119334 Moscow; �Department of Biophysics, Biological Faculty, Moscow State University, Vorobyevy Gory, 119899 Moscow; �Laboratory of Molecular Diagnostics and Gene-Engineering Constructs, All-Russia Institute of Agricultural Biotechnology, Russian Academy of Agricultural Sciences, 127550 Moscow; §Laboratory of the Enzyme Systems, Bach Institute of Biochemistry, Russian Academy of Sciences, 119071 Moscow, Russia; ¶Nuclear Signaling Laboratory, John Curtin School of Medical Research, Australian National University, Canberra ACT 2601, Australia; #Department of Chemistry and Technology of Fine Organic Chemicals, Moscow Academy of Fine Chemical Technology, 117571 Moscow, Russia; and **Nuclear Signaling Laboratory, Department of Biochemistry and Molecular Biology, Monash University, Clayton, Victoria 3168, Australia Corresponding author: Alexander S. Sobolev, Department of Molecular Genetics of Intracellular Transport, Institute of Gene Biology, Russian Academy of Sciences, 34/5 Vavilov St., 119334 Moscow, Russia. E-mail: asobol@1.biophys.bio.msu.ru, alexander_s_sobolev@hotbox.ru ABSTRACT

The search for new pharmaceuticals that are specific for diseased rather than normal cells in the case of cancer and viral disease has raised interest in locally acting drugs that act over short distances within the cell and for which different cell compartments have distinct sensitivities. Thus, photosensitizers (PSs) used in anti-cancer therapy should ideally be transported to the most sensitive subcellular compartments in order for their action to be most pronounced. Here we describe the design, production, and characterization of the effects of bacterially expressed modular recombinant transporters for PSs comprising 1) α-melanocyte-stimulating hormone as an internalizable, cell-specific ligand; 2) an optimized nuclear localization sequence of the SV40 large T-antigen; 3) an Escherichia coli hemoglobin-like protein as a carrier; and 4) an endosomolytic amphipathic polypeptide, the translocation domain of diphtheria toxin. These modular transporters delivered PSs into the nuclei, the most vulnerable sites for the action of PSs, of murine melanoma cells, but not non-MSH receptor-overexpressing cells, to result in cytotoxic effects several orders of magnitude greater than those of nonmodified PSs. The modular fusion proteins described here for the first time, capable of cell-specific targeting to particular subcellular compartments to increase drug efficacy, represent new pharmaceuticals with general application.

Key words: targeted drug delivery • intracellular transport • nucleo-cytoplasmic transport • cancer • melanoma

M

any macromolecular medicines acting within the cell (e.g., foreign DNA, antisense oligonucleotides, polypeptide toxins) must be transported into specific intracellular compartments in order to exert their effects. By themselves, many of these locally acting compounds are unable to reach the site of their action, needing additional

macromolecular systems for targeted intracellular delivery. The list of locally acting drugs includes not only macromolecules but also low-molecular weight drugs that act over very short distances comparable with the dimensions of the cell compartments, where additionally, the sensitivities of distinct cell compartments to the action of these drugs are different. Such drugs accordingly need to be transported efficiently to the most sensitive/vulnerable site(s) of the target cell. Photosensitizers (PSs), used for anti-cancer photodynamic therapy, are an example of this type of locally acting drugs, with cytotoxic action not exceeding 40 nm from the site of subcellular localization since it is effected via reactive oxygen species (mainly, singlet oxygen and free radicals úOH and HOú2), which have a very short mean range within the cell (1). Keeping in mind that the efficacy of cell killing induced by reactive oxygen species depends on the site of their action, and that the cell nucleus is the most vulnerable/sensitive site, in contrast to cell membranes and other cytoplasmic organelles (1), there is little doubt that PS anti-tumor efficiency will depend on the subcellular distribution of PSs. Consistent with this, many laboratories have demonstrated that PSs of medical interest localize in various cytoplasmic structures, but not in the cell nucleus, and that the ultrastructural damage caused by PSs coincides with their localization (1, 2). Earlier (3, 4), we tried to change the subcellular distribution of PSs in favor of the cell nucleus, by exploiting targeting signals recognized by the cellular nuclear transport machinery. PSs can be successfully redirected within cells by using cross-linked modular polypeptide transporters possessing 1) an internalizable ligand providing for cell-specific delivery; 2) a nuclear localization sequence (NLS) conferring interaction with importins, the cellular proteins mediating active translocation into the nucleus; and modules enabling 3) escape from endosomes and 4) attachment of the PSs. The PSs transported to the cell nucleus by the modular carriers proved to be several orders of magnitude more efficient than nonmodified, free PSs, the best results being obtained using model modular transporters composed of 1) insulin as the internalizable ligand, 2) an optimized NLS from the SV40 large tumor antigen (T-ag), 3) attenuated human adenovirus Ad5 dl312 as an endosomolytic component, and 4) bacterial β-galactosidase as a cross-linked carrier (3, 4). Analogous results have been recently obtained by others (5). Notably, internalized PSs are more efficient in cell killing than those localized at the cell surface (6-8), whereas PSs transported to the nuclei are more efficient than those internalized (3, 4) and, as just mentioned, substantially more efficient than free, nonmodified PSs. The nucleus is thus a hypersensitive site for photodynamic damage. Although insulin was used as a model ligand conferring both cell-type specificity and receptor-mediated endocytosis of a photosensitizing construct, the approach is clearly applicable to a wide variety of cancer cell types, if ligands are used that are recognized by receptors overexpressed on these cells, e.g., α-melanocyte-stimulating hormone (MSH) specific for a number of melanomas (9-12), epidermal growth factor receptor family binding ligands (13, 14), and somatostatin (15, 16). There are options to simplify the transporting constructs, e.g., by replacing adenoviruses

with a simpler component mediating release from endosomes, such as amphipathic oligopeptides that have already been shown to promote the liberation of macromolecules from the endocytotic pathway compartments (17). An important aspect is also the technological feasibility of producing the transporting constructs since multicomponent transporters (see above) produced in a laborious and expensive fashion through covalent linkage of different peptide modules via bifunctional cross-linking reagents, even if they prove highly efficient in vivo, are unlikely to find broad clinical application. It is thus expedient to develop recombinant vehicles that would include modules for addressed delivery both to specific target cells and into the most vulnerable compartments thereof. Here we describe the design, production, and characterization of bacterially expressed modular recombinant transporters (MRTs) for PSs comprising 1) MSH as the internalizable ligand, 2) the optimized T-ag NLS, 3) the Escherichia coli hemoglobin-like protein HMP as a carrier, and 4) an endosomolytic amphipathic peptide. The MRTs delivered the PSs into the nuclei of murine melanoma cells and provided for several orders of magnitude greater cytotoxicity than free PSs. MRTs of the type described here represent new pharmaceuticals with general application in drug delivery.

METHODS Cell culture The B16-F1 and M3 mouse melanoma cell lines (American Type Culture Collection, Rockville, MD), both expressing MSH receptors (18,19), were used, where the latter displayed less autofluorescence and was thus used for microscopic investigations. Normal mouse embryonic fibroblast cell lines, C3H/10T1/2 and NIH/3T3 (a kind gift of S. L. Kiselev, Institute of Gene Biology, Moscow, Russia), as well as the above melanoma cell lines were maintained in Dulbecco�s modified Eagle�s medium (DMEM) supplemented with 10% fetal calf serum at 37 oC in 5% CO2. Bacteriochlorin p6 preparation Bacteriopurpurin p6 was obtained from Rhodobacter purpuratus by alkaline hydrolysis (20). The resulting substance was dissolved in dimethyl formamide, where the anhydride ring is opened to give bacteriochlorin p6. Production of MRT-coding plasmids, expression, and purification of MRTs All molecular biological procedures including DNA isolation, endonuclease cleavage, phosphorylation, ligation, cell transformation, and polymerase chain reaction (PCR) were performed according to standard protocols. The NLS module was generated by PCR using Deep Vent polymerase (Promega, Madison, WI), the primers 5'-GTGAGATCTGGGTTCTTCTACCTTTCTCTTC-3' (forward) and

5'-GTGAGATCTGCGCGTAATGAGCTCCTTGCAAAC-3' (reverse; restriction site underlined, as below � see Fig. 1 for assignment), and plasmid pPR28 (21) as a template. The MSH module was generated synthetically from the oligonucleotides M1, 5'-GATCCTACTCCATGGAACACTT-3'; M2, 5'-CCGTTGGGGCAAGCCGGTATA-3'; M3, 5'-AGCTTATACCGGCTTGCCC-3' (stop codon complement in bold, as below); and M4, 5'-CAACGGAAGTGTTCCATGGAGTAG-3'. GALA was produced analogously using G1, 5'-CATGGGATCCTGGGAAGCTGCTC-3;' G2, 5'-TGGCTGAAGCACTGGCAGAGGCAGC-3'; G3, 5'-GGCCGCTGCCTCTGCCAGTGCT-3'; G4, 5'-TCAGCCAGAGCAGCTTCCCAGGATCC-3'; G5, 5'-GGCCGAACACCTGGCTGAGGCA-3'; G6, 5'-CTTGCAGAGGCTCTCGAGGCTCT-3'; G7, 5'-GGCTGCAGGTGGGCCCAGATCTA-3'; G8, 5'-AGCTTAGATCTGGGCCCACCT-3'; G9, 5'-GCAGCCAGAGCCTCGAGAGCCTC-3'; and G10, 5'-TGCAAGTGCCTCAGCCAGGTGTTC-3'. The HMP module was generated by PCR using Taq polymerase (SibEnzyme, Novosibirsk, Russia), the PCR primers 5'-GCAAAAAAAGGGATCCCATATGCTTGACGCTC-3' (forward) and 5'-CCGGCAACTCTAGATCTCAGCACCTTATGCG-3' (reverse), and with E. coli chromosomal DNA as the template. The translocation domain of diphtheria toxin together with the natural spacer between the toxin domains (DTox module), residues 198-384 of the whole toxin, was similarly generated by Taq PCR using the primers 5'-GTAGGTGGATCCGGGTCATCCATAAATCTTGATTGG-3' (forward) and 5'-CCCGTCATCCGGAAATGGTTAAGATCTATGCCCCGG-3' (reverse) and plasmid DNA containing the cloned diphtheria toxin gene (kindly provided by Y. V. Vertiev, N. F. Gamaleya Institute of Epidemiology and Microbiology, Moscow, Russia) as a template. Plasmid construction is detailed in Fig. 1, A-F. Expression was carried out in E. coli strain M15 (carrying plasmid pREP4) according to the QIAGEN protocol (22), except for plasmid pR522-encoded protein, which was produced in strain DH5α carrying plasmid pREP4. The cells were lysed in 10 mM HEPES-NaOH, pH 7.5, 0.1 mM EDTA, 1 mM DTT, 3 mM phenylmethylsulphonyl fluoride, and 10 µg/ml lysozyme (all from Sigma, St. Louis, MO); sonicated (20 kHz); and centrifuged (17,000 rpm, JA-20 Beckman rotor) for 25 min. The supernatant was loaded onto blue Sepharose CL-6B (Pharmacia, Uppsala, Sweden), and proteins were eluted using a 0.7�1.6 M NaCl gradient. Protein purity was assayed with 10% polyacrylamide gel electrophoresis (Fig. 1, E' and F') according to Laemmli (23). Functional activity of MRT modules MSH activity of MRTs was tested by assaying melanogenesis in B16-F1 cells (18). Briefly, cells were seeded at a density of 2500/well in 96-well plates. After 24 h, serial dilutions of MSH (ICN, Moscow, Russia) or MRTs were added to the wells for 36 h, the cells were grown for 2 days, and the melanin absorbance was measured at 405 nm. Endosomolytic modules were tested

by calcein release from liposomes at pH 3�7 (24). NLS-containing MRTs were tested for recognition by α/β importins, expressed in bacteria as glutathione-S-transferase fusion proteins, using an ELISA-based binding assay (25). Preparation of PS-MRT conjugates Chlorin e6 (Porphyrin Products, Logan, UT) and bacteriochlorin p6 were conjugated (26) with MRTs (2:1 molar ratio) using cyclohexyl-3-(2-morpholinoethyl)carbodiimide metho-4-toluene sulfonate and 1,6-diaminohexane (both from Sigma). Video-intensified and image-ratio microscopy To visualize the intracellular localization of MRTs (used as MRT-chlorin e6 or -bacteriochlorin p6 conjugates), 10 µM 2',7'-dichlorodihydrofluorescein diacetate (Serva, Heidelberg, Germany), an indicator of reactive oxygen species, was added for 10 min to washed M3 mouse melanoma cells incubated for 18 or 24 h with MRT conjugates and then the cells were washed again, illuminated, and examined as previously described (3, 4, 26). 2',7'-Dichlorofluorescein production was then visualized immediately using a cooled CCD camera (AT200, Photometrics, Tucson, AZ, or SenSys:0400) coupled to an Axioplan microscope (Carl Zeiss, Jena, Germany). Fluorescence image processing (deconvolution, nearest-neighbor deblurring) was performed according to Agard et al. (27). The pH of the intracellular microenvironment of MRT conjugates at the sites of reactive oxygen species generation was determined by image-ratio video-intensified microscopy (28). In brief, two specimen fluorescence images were obtained by 2',7'-dichlorofluorescein excitation at 496 and 461 nm; the image ratio reflecting the local pH was translated into conventional colors using the PMIS program (Photometrics). Cytotoxic assays of PS-MRT conjugates The B16-F1, C3H/10T1/2, and NIH/3T3cells were seeded into 48 (2000 cells/well)- or 96-well (1000 cells/well) plates for 24 h and then incubated for 18 h with PS or PS-MRT conjugates. The cells were washed three times with Hanks' solution, transferred into DMEM with 2 mg/ml bovine serum albumin (Serva, Heidelberg, Germany), illuminated with a slide projector, with filters (730�1000 nm), 75 kJ/m2 (100% cell survival without PSs or PS-MRTs), and grown under 5% CO2. Cell viability was determined in 3�4 days using methylene blue staining according to Finlay et al. (29). RESULTS Design of modular plasmids encoding MRTs of PSs, synthesis, and purification of the MRTs We designed the gene modules encoding the corresponding polypeptide module according to the scheme: BamHI site � module sequence � BglII site � stop codon � HindIII site. This structure allows every gene module to be placed at any position along the hybrid gene because it is flanked with BamHI and BglII restriction sites with identical sticky ends. All the constructs were

assembled through consecutive cloning and the strong T5 bacteriophage promoter used for protein expression in bacteria (see Methods). As mentioned above, MSH was chosen as the ligand module conferring recognition by specific target cells (melanoma) through binding to melanocortin receptors with subsequent internalization (30). The peculiarities of MSH dictated its localization at the C terminus of MRTs: only the N terminus of the hormone can be modified without loss of MSH activity/binding to its receptors (31, 32). The MSH gene was prepared synthetically using codons optimized for expression in E. coli. The optimized nuclear targeting module had the amino acid sequence Ser-Ser112-Asp-Asp-Glu-Ala-Thr-Ala-Asp-Ala-Gln-His-Ala-Ala124-Pro-Pro-Lys-Lys128-Lys-Arg-Lys-Val-Glu-Asp-Pro135, where the numbers refer to the T-ag amino acid sequence, with the NLS (residues 126�132) underlined; Ser112 is the site of CK2 phosphorylation that enhances T-ag nuclear import, and Thr124 (replaced with Ala) is the site of phosphorylation by Cdc2 kinase that inhibits nuclear import (21). The endosomolytic modules used were either 1) the GALA peptide shown to make pores in membranes at acidic pH (17,33), or 2) the amphipathic translocation domain of diphtheria toxin (DTox) also capable at acidic pH of permeating endocytotic membranes even in supramolecular complexes with chimeric polypeptides and DNA (34, 35). The latter has been widely used, either alone or together with the catalytic domain, as a component of membrane penetrable fusion proteins (36-38). The GALA gene module was composed of two parts: four oligos for the peptide N-proximal part and six oligos for the C-proximal part; codons were optimized for E. coli expression, introducing Gly and Pro into the N- and C-terminal regions. The diphtheria toxin translocation domain encoding sequence was taken together with the adjacent natural spacer between the toxin domains (DTox module). Thus, we generated plasmids (Fig. 1) encoding a number of MRTs, including pR522: HMP-NLS-MSH; pR523: GALA-HMP-NLS-MSH; pR676: DTox-HMP-NLS-MSH. The extent of MRT expression ranged from 5�8% of total protein for construct pR523 to 20�30% for constructs pR522 and pR676, with solubility 60�70%. Note that the yield of GALA-HMP-NLS-MSH (pR523) was considerably lower than that of the MRT with the other endosomolytic module (DTox-HMP-NLS-MSH, pR676). The MRTs were purified on blue Sepharose, the single-step purification providing 90�95% purity (Fig. 1). Functional testing of MRT modules The purified chimeric MRTs were initially tested to assess whether their individual modules retained their functional activities and were able to contribute to the overall goal of cell-specific nuclear PS delivery.

MSH and its analogs induce melanin synthesis in melanocyte-related cells through activation of melanocortin receptors (39). Stimulation of melanogenesis in B16-F1 cells by the MRTs showed that they were able to bind with the melanocortin receptors of these cells to evoke a biological response. The concentrations producing a half-maximal effect (EC50) were similar for the two MRTs: 14 nM for HMP-NLS-MSH and 17 nM for DTox-HMP-NLS-MSH, compared to the EC50 for native MSH of 0.36 nM (Fig. 2A). Recombinant peptides designed similarly but not containing the MSH module did not induce melanogenesis in B16-F1 cells (Fig. 2A). MRTs delivered into cells by receptor-mediated endocytosis are internalized into endosomes, enclosed membranous structures with weakly acidic internal pH, which they must exit in order to be targeted subsequently to their final intracellular destination, in this case the nucleus, mediated through the action of importins in the cytosol. The propensity of a polypeptide to make pores in membranes in an acidic medium can be assessed from its ability to effect leakage of dye-loaded liposomes at different pHs (33). In experiments with calcein-loaded liposomes, maximal activity of GALA-HMP-NLS-MSH was observed at about pH 3.5-4 (Fig. 2B), which is more acidic than in endosomes (40). Liposome leakage under the action of DTox-HMP-NLS-MSH (Fig. 2B) was observed in two pH intervals: 3.5-4.5, which was attributable to the HMP because it alone showed maximal activity at pH 3.5-4.5 (Fig. 2B), and 4.5-6, which is close to the endosomal pH (40) and was attributable to activity of the DTox moiety. This, together with the fact that GALA-HMP-NLS-MSH was more difficult to express in bacteria (see above), encouraged us to concentrate on DTox-HMP-NLS-MSH for the studies below. Assessment of the recognition of the MRTs by the nuclear transport-mediating importin α/β heterodimer using an ELISA-based assay (25) indicated that the NLS in the context of the MRTs is capable of mediating high affinity interaction with importins (Fig. 2C); the apparent dissociation constants (Kd) for importin binding were 1.9 and 2.5 nM for HMP-NLS-MSH and DTox-HMP-NLS-MSH, respectively. By comparison, a control T-ag NLS peptide including the NLS and optimized phosphorylation site (25) showed a Kd of 2 nM, whereas an NLS mutated (Thr128-substituted) control peptide, or non-NLS-containing MRT such as DTox-HMP (Fig. 2C), displayed only very low importin binding, precluding accurate estimation of the Kd. Importantly, covalent attachment of PSs did not affect the functional activities of any of the MRT modules (data not shown). Intracellular localization of MRTs Analysis of the localization and sites of intracellular photodynamic action of PS-containing DTox-HMP-NLS-MSH and HMP-NLS-MSH in M3 mouse melanoma cells using the permeant dye 2′,7′-dichlorodihydrofluorescein diacetate as an indicator of reactive oxygen species production (4) revealed that the MRTs accumulated in the cells, with the pattern of accumulation depending on the presence of a functional endosomolytic module in their composition. This cell line also expresses melanocortin receptors (19) but possesses less autofluorescence and therefore was used for microscopic investigations. Production of reactive oxygen species due to DTox-HMP-NLS-MSH was detected in the nuclei of 87.5% of cells, whereas that due to HMP-NLS-MSH was present only in the nuclei of 12.2% of cells (P < 0.05); see also the deconvoluted fluorescence micrographs of Fig. 3, A and B.

Results for probing the pH of the intracellular environments of the MRTs by image-ratio video-intensified microscopy were consistent with the above results. Figure 4, A and B, displays the pH-microenvironment of the PS-containing MRTs probed by the pH-sensitive dye 2′,7′-dichlorofluorescein produced intracellularily from 2′,7′-dichlorodihydrofluorescein by reactive oxygen species generated by the PS-MRTs upon illumination. Thus, a pH-specific fluorescence (shown in conventional colors in Fig. 4) develops at the subcellular sites where PS-MRT and the dye colocalize. The HMP-NLS-MSH conjugate lacking an endosomolytic module was found in acidic regions of the cells (reddish spots in Fig. 4A; see also Fig. 4C for match of conventional colors and pH values). No such acidic regions were revealed in the vicinity of DTox-HMP-NLS-MSH conjugate localization (Fig. 4B), which localized in rather less acidic (yellowish) sites, meaning that the MRT and the dye do not colocalize within the acidic (<5.4) sites in the optical section.

Cytotoxic action of PSs delivered by MRTs to the nuclei of mouse melanoma cells Evaluation of the photocytotoxic effect on mouse B16-F1 melanoma cells, which overexpress MSH receptors, a property of many melanomas (9-12), showed that the efficacy of PS is greatly enhanced by its attachment to MRTs (Fig. 5A). A half-maximal effect of (bacteriochlorin p6)-DTox-HMP-NLS-MSH was attained at a concentration (EC50 = 22 nM), which is lower than that required for free bacteriochlorin p6 (EC50 = 4990 nM) by a factor of 230. (Bacteriochlorin p6)-DTox-HMP-NLS-MSH conjugate was not photocytotoxic to normal C3H/10T1/2 (Fig. 5A) or NIH/3T3 (data not shown) mouse fibroblast lines, demonstrating cell-specific activity of the MRT through the MSH module. (Bacteriochlorin p6)-HMP-NLS-MSH conjugate, lacking the endosomolytic module, was 5.3 times less active than (bacteriochlorin p6)-DTox-HMP-NLS-MSH, possessing this module; PS-MRT conjugates lacking NLS module showed less photocytotoxic activity than the above two conjugates. Free MRTs, not carrying PSs, did not affect viability of B16-F1 melanoma cells (Fig. 5B). DISCUSSION Site-specific drug-delivery systems are one of the most promising approaches in the drug delivery field including anti-cancer therapy (41-43). This approach will permit delivery of very specifically acting compounds including proteins and nucleic acids to particular sites of target cells. The physical properties of these compounds, e.g., high molecular mass (>500 Da) or inappropriate solubility, etc., are the main obstacles precluding their application. These compounds may be called �locally acting� drugs because they must be transported into specific intracellular compartments in order to exert their effects. The list of locally acting drugs also includes low-molecular weight drugs such as PSs that act over very short distances comparable with the dimensions of the cell compartments, where the sensitivities of distinct cell compartments to the action of these drugs are different. Such drugs accordingly need to be transported to the most sensitive/vulnerable site of the target cell. To circumvent this transportation problem, several approaches have been developed including the use of 1) antibodies (44), in some cases able to be internalized (45); 2) ligands to internalizable receptors (46); and 3) protein transduction domains (47). The latter do not confer cell-specific delivery, which is desirable for anti-cancer and other therapies; in contrast, the first two afford cell-

specific delivery and uptake but do not provide for transportation to particular intracellular targets. We think that a combination/modification of the aforementioned approaches as well as inclusion of intracellular targeting signals into the future delivery constructs allows the problem of cell-specific intracellular targeted drug delivery to be overcome. The recombinant transporters of the type described in our present paper fulfil the above requirements and thus represent, we believe, novel, new types of pharmaceuticals with wide application in drug delivery. The results obtained show that it is possible to design modular chimeric genes encoding MRTs and to obtain these MRTs as 90�95% pure polypeptides after a simple purification procedure. The arrangement of the ligand, carrier, endosomolytic, and NLS modules within the MRTs is governed mainly by the possible dependence of their activity on their position within the whole polypeptide. The modules of the chimeric MRTs retain their functional activities so that they are able to interact with the corresponding internalizable receptors overexpressed on target cells (e.g., mouse melanoma), escape from acidic compartments such as endosomes after internalization, be transported into the cell nucleus, and also carry PSs. These data taken together clearly indicate that NLS-mediated nuclear localization of PSs enhances their activities, again confirming that the nucleus is a hypersensitive site for photodynamic damage (1, 3, 4, 48, 49). The cytotoxic activity of the MRT-linked PSs exceeded free PSs by several orders of magnitude. It is not clear at this stage whether the DTox-HMP-NLS-MSH conjugates detected in the nuclei of melanoma cells on the basis of PS-generated fluorescence are fully intact, but since nuclear fluorescence is observed (Figs. 3 and 4), it is clear that PSs have been delivered into the nuclei by the MRT. These results are indicative of the prospects of using recombinant chimeric multicomponent vehicles for these and, possibly, for other locally acting drugs such as alpha-particle emitting radionuclides where the dose of radioactivity necessary to kill 63% of cells (D0), of 211-astatine, delivered to human hepatoma cell nuclei by our modular transporters, is one order of magnitude less than that of free 211At- (unpublished observations). MRTs of the type described here, capable of cell-specific targeting to particular subcellular compartments to increase drug efficacy, represent new pharmaceuticals with potential general application. The different modules of the MRTs, which are highly expressed and easily purified to retain full activity of each of the modules, are, of course, interchangeable, meaning that they can be tailored for particular applications, e.g., targeting signals for lysosomal, membrane, or mitochondrial location can substitute the NLS in the case of drugs with optimal activity in these compartments. Similarly, ligands other than MSH can be used depending on the desired target cell type, e.g., an MRT containing epidermal growth factor as a ligand and conjugated with bacteriochlorin p6 showed higher photocytotoxic efficacy than free bacteriochlorin p6 by a factor of 900 (human A431 epidermoid carcinoma cells) (unpublished observations). The HMP moiety can also be substituted by domains binding other types of drugs/molecules with high affinity; an example of this is the use of DNA-binding domains for DNA delivery in gene therapy (50). Our modular/combinatorial approach to drug delivery based on MRTs, we believe, thus constitutes the first step toward developing a new generation of pharmaceuticals.

ACKNOWLEDGEMENTS The authors are indebted to G. P. Georgiev (Institute of Gene Biology, Moscow) for constructive discussions of particular subjects and this work as a whole. This study was supported at various times by grants from the International Union Against Cancer (ICRETT Grant No. 69, 1999, and No. 129, 2000); International Association for the Promotion of Cooperation with Scientists from the New Independent States of the former Soviet Union (INTAS; Grant No. 01-0461); Russian Ministry of Industry, Science and Technologies; Russian Foundation for Basic Research (Grant No. 00-04-48118); and Moscow Program for the �Development of New Methods and Remedies for Diagnostics and Treatment of Cancer and Their Introduction into Clinical Practice.� D. A. Jans is an Australian National Health and Medical Research Council senior fellow (ID# 143790).

REFERENCES 1. Sobolev, A. S., Jans, D. A., and Rosenkranz, A.A. (2000) Targeted intracellular delivery of

photosensitizers. Prog. Biophys. Mol. Biol. 73, 51-90 2. Rosenkranz, A. A., Jans, D. A., and Sobolev, A. S. (2000) Targeted intracellular delivery of

photosensitisers to enhance photodynamic efficiency. Immunol. Cell Biol. 78, 452-464 3. Akhlynina, T. V., Jans D. A., Rosenkranz, A. A., Statsyuk, N. V., Balashova, I. Y., Toth, G.,

Pavo, I., Rubin, A. B., and Sobolev, A. S. (1997) Nuclear targeting of chlorin e6 enhances its photosensitizing activity. J. Biol. Chem. 272, 20328-20331

4. Akhlynina, T. V., Jans, D. A., Statsyuk, N. V., Balashova, I. Y., Toth, G., Pavo, I.,

Rosenkranz, A. A., Naroditsky, B.S., and Sobolev, A. S. (1999) Adenovirus synergize with nuclear localization signals to enhance the nuclear delivery and photodynamic action of internalizable conjugates containing chlorin e6. Int. J. Cancer 81, 734-740

5. Bisland, S. K., Singh, D., and Gariépy, J. (1999) Potentiation of chlorin e6 photodynamic

activity in vitro with peptide-based intracellular vehicles. Bioconjugate Chem. 10, 982-992 6. Akhlynyna, T. V., Gulak, P. V., Serebryakova, N. V., Rosenkranz, A. A., and Sobolev, A. S.

(1990) Photodynamic action of concanavalin A-chlorin e6 conjugate on human fibroblasts. Bull. Exp. Biol. Med. (USSR) 109, 183-184

7. Akhlynyna, T. V., Rosenkranz, A. A., Jans, D. A., Gulak, P. V., Serebryakova, N. V., and

Sobolev, A. S. (1993) The use of internalizable derivatives of chlorin e6 for increasing its photosensitizing activity. Photochem. Photobiol. 58, 45-48

8. Hamblin, M. R., Miller, J. L., and Ortel, B. (2000) Scavenger-receptor targeted

photodynamic therapy. Photochem. Photobiol. 72, 533-540 9. Jiang, J., Sharma, S. D., Fink, J. L., Hadley, M. E., and Hruby, V. J. (1996) Melanotropic

peptide receptors: membrane markers of human melanoma cells. Exp. Dermatol. 5, 325-333

10. Funasaka, Y., Sato, H., Chakraborty, A. K., Ohashi, A., Chrousos, G. P., and Ichihashi, M.

(1999) Expression of proopiomelanocortin, corticotropin-releasing hormone (CRH), and CRH receptor in melanoma cells, nevus cells, and normal human melanocytes. J. Investig. Dermatol. Symp. Proc. 4, 105-119

11. Wikberg, J. E., Muceniece, R., Mandrika, I., Prusis, P., Lindblom, J., Post, C., and Skottner,

A. (2000) New aspects on the melanocortins and their receptors. Pharmacol. Res. 42, 393-420

12. Salazar-Onfray, F., Lopez, M., Lundqvist, A., Aguirre, A., Escobar, A., Serrano, A.,

Korenblit, C., Petersson, M., Chhajlani, V., Larsson, O., and Kiessling, R. (2002) Tissue distribution and differential expression of melanocortin 1 receptor, a malignant melanoma marker. Br. J. Cancer 87, 414-422

13. Mendelsohn, J., and Baselga, J. (2000) The EGF receptors family as targets for cancer

therapy. Oncogene 19, 6550-6565 14. Gibbs, J.B. (2000) Anticancer drug targets: growth factors and growth factor signaling. J.

Clin. Invest. 105, 9-13 15. De Jong, M., Valkema, R., Jamar, F., Kvols, L. K., Kwekkeboom, D. J., Breeman, W. A.,

Bakker, W. H., Smith, C., Pauwels, S., and Krenning, E. P. (2002) Somatostatin receptor-targeted radionuclide therapy of tumors: preclinical and clinical findings. Semin. Nucl. Med. 32, 133-140

16. Kwekkeboom, D. J., and Krenning, E. P. (2002) Somatostatin receptor imaging. Semin. Nucl.

Med. 32, 84-91 17. Plank, C., Oberhauser, B., Mechtler, K., Koch, C., and Wagner, E. (1994) The influence of

endosome-disruptive peptides on gene transfer using synthetic virus-like gene transfer systems. J. Biol. Chem. 269, 129180-12924

18. Siegrist, W., and Eberle, A. (1986) In situ melanin assay for MSH using mouse B16

melanoma cells in culture. Anal. Biochem. 159, 191-197 19. O'Hare, K.B., Duncan, R., Strohalm, J., Ulbrich, K., and Kopeckova, P. (1993) Polymeric

drug-carriers containing doxorubicin and melanocyte-stimulating hormone: in vitro and in vivo evaluation against murine melanoma J. Drug Target. 1, 217-229

20. Mironov, A. F., Kozyrev, A. N., and Brandis, A. (1992) Sensitizers of second generation for

photodynamic therapy of cancer based on chlorophyll and bacteriochlorophyll derivatives. Proc. SPIE 1922, 204-208

21. Jans, D. A., Ackermann, M., Bischoff, J. R., Beach, D. H., and Peters, R. (1991) p34cdc2-mediated phosphorylation at T124 inhibits nuclear import of SV40 T-antigen proteins. J. Cell Biol. 115, 1203-1212

22. The QIAexpressionist. (QIAGEN, ed. 5, 2001), http://www.qiagen.com/literature

/handbooks/qxp/qxp/1016852QXPHB_0301WW.pdf 23. Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of

bacteriophage T4. Nature 227, 680-685 24. Rosenkranz, A. A., Antonenko, Y. N., Smirnova, O. A., Yurov, G. K., Naroditsky, B. S., and

Sobolev, A. S. (1997) Avian adenovirus induces ion channels in model bilayer lipid membranes. Biochem. Biophys. Res. Commun. 236, 750-753

25. Hübner, S., Xiao, C.-Y., and Jans, D. A. (1997) The Protein kinase CK2 site (Ser 111/112)

enhances recognition of the simian virus 40 large T-antigen nuclear localization sequence by importin. J. Biol. Chem. 272, 17191�17195

26. Akhlynina, T. V., Rosenkranz, A. A., Jans, D. A., and Sobolev, A. S. (1995) Insulin-

mediated intracellular targeting enhances the photodynamic activity of chlorin e6. Cancer Res. 55, 1014-1019

27. Agard, D. A., Hiraoka, Y., Shaw, P., and Sedat, J. W. (1989) Fluorescence microscopy in

three dimensions. Methods Cell Biol. 30, 353-377 28. Silver, R. B. (1998) Ratio imaging: practical considerations for measuring intracellular

calcium and pH in living tissue. Methods Cell Biol. 56, 237-251 29. Finlay, G. J., Baguley, B. C., and Wilson, W. R. (1984) A semiautomated microculture

method for investigating growth inhibitory effects of cytotoxic compounds on exponentially growing carcinoma cells. Anal. Biochem. 139, 272-277

30. Wong, W., and Minchin, R. F. (1996) Binding and internalization of the melanocyte

stimulating hormone receptor ligand [Nle4, D-Phe7] alpha-MSH in B16 melanoma cells. Int. J. Biochem. Cell. Biol. 28, 1223-1232

31. Sahm, U. G., Olivier, G.W., Branch, S. K., Moss, S. H., and Pouton, C. W. (1994) Influence

of alpha-MSH terminal amino acids on binding affinity and biological activity in melanoma cells. Peptides 15, 441-446

32. Wen, Z., Tao, X., Lakkis, F., Kiyokawa, T., and Murphy, J. R. (1991) Diphtheria toxin-

related α-melanocyte-stimulating hormone fusion toxin. Internal in-frame deletion from Thr387 to His485 results in the formation of a highly potent fusion toxin which is resistant to proteolytic degradation. J. Biol. Chem. 266, 12289-12993

33. Nir, S., and Nieva, J. L. (2000) Interactions of peptides with liposomes: pore formation and fusion. Progr. Lipid Res. 39, 181-206

34. Uherek, C., Fominaya, J., and Wels, W. (1998) A modular DNA carrier protein based on the

structure of diphtheria toxin mediates target cell-specific gene delivery. J. Biol. Chem. 273, 8835-8841

35. Barati, S., Chegini, F., Hurtado, P., and Rush, R. A. (2002) Hybrid tetanus toxin C fragment-

diphtheria toxin translocation domain allows specific gene transfer into PC12 cells. Exp. Neurol. 177, 75-87

36. Frankel, A. E., Powel,l B. L., Hall, P. D., Case, L. D., and Kreitman, R. J. (2002) Phase I trial

of a novel diphtheria toxin/granulocyte macrophage colony-stimulating factor fusion protein (DT388GMCSF) for refractory or relapsed acute myeloid leukemia. Clin. Cancer Res. 8, 1004-1013

37. Chenal, A., Nizard, P., Forge, V., Pugnière, M., Roy, M.-O., Mani, J.-C., Guillain, F., and

Gillet, D. (2002) Does fusion of domains from unrelated proteins affect their folding pathways and the structural changes invovled in their function? A case study with the diphtheria toxin T domain. Protein Eng. 15, 383-391

38. Ramage, J. G., Vallera, D. A., Black, J. H., Aplan, P. D., Kees, U. R., and Frankel, A. E.

(2003) The diphtheria toxin/urokinase fusion protein (DTAT) is selectively toxic to CD87 expressing leukemic cells. Leuk. Res. 27, 79-84

39. Slominski, A., Wortsman, J., Luger, T., Paus, R., and Salomon, S. (2000) Corticotropin

releasing hormone and proopiomelanocortin involvement in the cutaneous response to stress. Physiol. Rev. 80, 979-1020

40. Ceresa, B. P., and Schmid, S. L. (2000) Regulation of signal transduction by endocytosis.

Curr. Opin. Cell Biol. 12, 204�210 41. Atkins, J. H., and Gershell, L. J. (2002) Selective anticancer drugs. Nature Rev. Drug Discov.

1, 491-492 42. Langer, R. (2001) Drugs on target. Science 293, 58-59 43. Schwarze, S. R., and Dowdy, S. F. (2000) In vivo protein transduction: intracellular delivery

of biologically active proteins, compounds and DNA. Trends Pharmacol. Sci. 21, 45-48 44. Trail, P. A., and Bianchi, A. B. (1999) Monoclonal antibody drug conjugates in the treatment

of cancer. Curr. Opin. Immunol. 11, 584-588 45. Poul, M. A., Becerril, B., Nielsen, U. B., Morisson, P., and Marks, J. D. (2000) Selection of

tumor-specific internalizing human antibodies from phage libraries. J. Mol. Biol. 301, 1149-1161

46. Rihova, B. Receptor-mediated targeted drug or toxin delivery. (1998) Adv. Drug Deliv. Rev.

29, 273-289 47. Schwarze, S. R., Ho, A., Vocero-Akbani, A., and Dowdy, S. F. (1999) In vivo protein

transduction: delivery of a biologically active protein into the mouse. Science 285, 1569-1572

48. Wiseman, H., and Halliwell, B. (1996) Damage to DNA by reactive oxygen and nitrogen

species: role in inflammatory disease and progression to cancer. Biochem. J. 313, 17-29 49. Liang, H., Shin, D. S., Lee, Y. E., Nguyen, D. C., Kasravi, S., Do, T., Aurasteh, P., and

Berns, M. W. (2000) Subcellular phototoxicity of photofrin-II and lutetium texapherin in cells in vitro. Laser Med. Sci. 15, 109-122

50. Chan, C. K., and Jans, D. A. (2001) Enhancement of MSH receptor- and GAL4-mediated

gene transfer by switching the nuclear import pathway. Gene Therapy 8, 166-171

Received November 13, 2002; accepted February 14, 2003.

Fig. 1

Figure. 1. Construction of MRT modules. MSH- (A); NLS- (B); carrier, HMP- (C); and endosomolytic, DTox- (D) modules of HMP-NLS-MSH (E) and DTox-HMP-NLS-MSH (F). The corresponding purified MRTs (10% polyacrylamide electrophoresis) are E′ and F′, respectively. Lane 1, total E. coli soluble protein; lane 2, total soluble protein of E. coli expressing the MRT; lane 3, the purified MRT.

Fig. 2

Figure 2. Functionality of MRT modules. A) Induction of melanogenesis in B16-F1 cells: free MSH (♦), HMP-NLS-MSH (▲), DTox-HMP-NLS-MSH (●), DTox-HMP (■); maximal (100%) melanin production corresponds to 26.1 ± 1.2 µg of melanin/well. B) Liposome leakage induced by MRTs: DTox-HMP-NLS-MSH (●), GALA-HMP-NLS-MSH (∆), HMP-NLS-MSH (▲), and HMP (□). Egg yolk phosphatidylcholine liposomes were loaded with fluorescent calcein up to the concentration of fluorescence quenching; the liposome leakage resulted in appearance of fluorescence; 100% liposome leakage (maximum) corresponds to the fluorescence intensity observed after addition of Triton X-100 (0.5% final concentration) to the calcein-loaded liposomes. C) Binding, B, of α/β-importin heterodimer to the MRTs as quantitated using ELISA-based binding assay: DTox-HMP-NLS-MSH (●), HMP-NLS-MSH (▲), and DTox-HMP (■). Bmax, maximal binding. Results are means ± SE (n=3-6).

Fig. 3

Figure 3. Visualization of the sites of generation of reactive oxygen species in M3 mouse melanoma cells. Ten minute incubation with 10 µM 2′,7′-dichlorodihydrofluorescein diacetate, followed by washing and photoactivation. A and B) Deconvoluted fluorescence micrographs of the cells incubated with (bacteriochlorin p6)-DTox-HMP-NLS-MSH or (bacteriochlorin p6)-HMP-NLS-MSH, respectively.

Fig. 4

Figure 4. The pH of (chlorin e6)-DTox-HMP-NLS-MSH (A) and (chlorin e6)-HMP-NLS-MSH (B) microenvironment. The pH, shown in conventional colors (refer to insert C), was determined by image-ratio video-intensified microscopy upon formation of 2′,7′-dichlorofluorescein through generation of reactive oxygen species under the same conditions as described in the legend to Fig. 3. D and E) The corresponding brightfield images. C) pH vs. fluorescence intensity, I, ratio (I461×10/I496) plot with corresponding color scale.

Fig. 5

Figure 5. Enhancement of photocytotoxic activity due to cell-specific intranuclear delivery of the PS bacteriochlorin p6. A) Cytotoxic action of bacteriochlorin p6 delivered into B16-F1 cell nuclei in the form of (bacteriochlorin p6)-DTox-HMP-NLS-MSH conjugate (●) compared to free bacteriochlorin p6 (○); cytotoxicity in C3H/10T1/2 normal mouse fibroblasts of (bacteriochlorin p6)-DTox-HMP-NLS-MSH conjugate (▼) and free bacteriochlorin p6 (�) is also shown. B) Influence of MRT- bacteriochlorin p6 conjugates lacking different modules on B16-F1 cell survival tested at the same experimental conditions as in A: (bacteriochlorin p6)-HMP-NLS-MSH (▲), (bacteriochlorin p6)-HMP-MSH (�), (bacteriochlorin p6)-DTox-MSH (▲), and DTox-HMP-NLS-MSH without bacteriochlorin p6 (●). Results are means ± SE (n=3).