Porphyrins have been widely utilized as biochemical and biological functional chromophores which can operate under visible-light irradiation. Water-soluble porphyrins have been used as the drug for photodynamic therapy (PDT) and photodynamic inactivation (PDI). Although usual water-solubilization of porphyrins has been achieved by an introduction of an ionic group such as ammonium, pyridinium, sulfonate, phosphonium, or carboxyl to porphyrin ring, we proposed the preparation of water-soluble P and Sb porphyrins by modification of axial ligands. Alkyl (type A), ethylenedioxy (type E), pyridinium (type P), and glucosyl groups (type G) were introduced to axial ligands of Sb and P porphyrins to achieve water-solubilization of Sb porphyrin and P porphyrins. Here, we review their water-soluble P and Sb porphyrins from the standpoints of preparation, bioaffinity, and photosensitized inactivation.

1. Introduction

Porphyrin complexes with high absorptivity in visible-light regions have been widely utilized as chemical and biological functional chromophores [1]. Porphyrin complexes are capable of versatile catalytic reactions through electron transfer or energy transfer under visible-light irradiation. Recently bioactive porphyrins have received much attention in connection with photoinactivation [2, 3] and photodynamic therapy (PDT) [47]. In general, the porphyrin compound tends to form face-to-face aggregates, resulting in less solubility even in organic solvents. For the biological application of porphyrins, water solubility is an important characteristic in handling the porphyrins in aqueous solution. Usual water-solubilization of metalloporphyrins is achieved by the modification of porphyrin ring by an ionic group such as ammonium [8], pyridinium [913], sulfonate [1416], phosphonium [13, 17], or aminocarboxylate [18]. However, the preparation of these complexes is not so easy. Recently we have developed water-solubilization of P and Sb porphyrins by modification of axial ligands [1926]. Tetraphenylporphyrin (H2tpp) is a typical and commercially available porphyrin and can easily be converted to the P(V) and Sb(V) complexes. The P and Sb porphyrins are able to connect covalently with axial ligands through oxygen and nitrogen atoms, resulting in the highly stabilized complexes. Therefore the axial-ligand modified P and Sb porphyrins can conveniently provide a variety of water-soluble porphyrins with bioactivity and bioaffinity as well as high electron accepting ability. Here, we review water-soluble P and Sb porphyrins from the standpoints of preparation, bioaffinity, and photosensitization.

2. Water-Solubilization of Porphyrins

2.1. By Modification of Axial Ligands

It is believed that (1a) (1b) and (tpp)Sb(OMe)OH+ (2a) are water-soluble since they are cationic compounds with hydrophilic hydroxogroup (Scheme 1). However, their water solubilities (/mM), which were defined as saturated concentration in water, were still low. Therefore, the of Sb porphyrins was enhanced by an introduction of a variety of axial ligands. Modification of axial ligands of P and Sb porphyrins has been easily performed by alkylation of 1a and 1b with alkyl halides as well as ligand exchange of and with alcohols. Matsumoto et al. and Yasuda et al. have modified axial ligands by alkyl (named as type A), ethylenedioxy (type E), N-alkyl pyridyl (type P), and glucosyl groups (type G) to provide water-soluble Sb porphyrin (2) [1921] and P porphyrins (35) [2225]. Table 1 lists the of P and Sb porphyrins as well as spectroscopic data. Also, oil-solubility (/mM) which was defined as solubility in 1,4-dioxane is listed in Table 1.

Type A of porphyrins (2a2g, Scheme 2) was prepared by the ligand exchange of (tpp)Sb(OMe)Br+Br with alcohols. Introduction of axial alkyloxo ligands enhanced the values of 2c2g [1921]. For example, (1.09 mM) of hexyloxo(methoxo)tetraphenylporphyrinatoantimony (2c) was much higher than that of the parent 2a ( = 0.10 mM). As increase of carbon number () of axial alkyl ligands, the of 2 increased until reached 2.40 mM at , as shown in Figure 1. It was thought that hydrophobic alkyl ligand induced the formation of micelle type of aggregation. Therefore, the structure of 1 in aqueous solution was examined by the dependence of half-width of bands in the absorption spectra [12] and surface tension on the concentration of 2 in range from 1 μM to 100 μM. It was estimated that 2c2g with the larger than 1 mM were present as aggregates in a concentration higher than 10 μM. In the NMR analysis in D2O, the anisotropic higher field shifts compared with the NMR spectra in CD3OD were observed at terminal methyl group on axial alkyl ligand. It was deduced that the alkyloxo ligands of 2c2g were arranged alternately in the aggregates, as illustrated in Figure 1 (inset). The long alkyl axial ligands with were requisite for the high solubility in water. The diameter of the aggregates of 2c2g in water was determined to be around 100 nm by the dynamic light scattering method.

P porphyrins with two symmetric axial ligands in upper and lower positions can be easily synthesized while it is difficult to introduce ligands unsymmetrically to axial position. The introduction of two axial hexyloxo ligands did not enhance the : that is, of 3i was still low (0.023 mM). Therefore, hydrophilic ethylenedioxy group (–CH2CH2O–) was incorporated into the axial ligands in order to enhance [22]. Bis(polyoxoalkyloxo)(tetraphenylporpyrinato)phosphorus (3a3k, Scheme 3) was prepared by the reaction of polyoxa-alcohols (H(CH2)n(OCH2CH2)mOH) with (tpp) in MeCN in the presence of a small amount of pyridine. The introduction of ethylenedioxy group improved . For example, the of 3h (1.11 mM) became relatively higher compared with 3i without the ethylenedioxy group. As the carbon number () of the alkyl group on H(CH2)n(OCH2CH2)2– ligands decreased, the values increased: = 2.07 (; 3g), 5.38 (; 3e), 13.0 (; 3d), and 13.9 mM (; 3c) at . The maximum was observed at 3b (m = 3, ), whose was 17.4 mM. Thus, the could be varied from 0.02 to 17.4 mM by the adjustment of and . The bioaffinity and photoinactivation of 2 and 3 to Saccharomyces cerevisiae (yeast) will be discussed in Sections 3.1 and 4.3.

The tricationic P porphyrins (4a4c, Scheme 4) were prepared by the reactions of 3-alkylpyridine with bis(5-bromo-3-oxapentyloxo)tetraphenylporphyrinato-phosphorus(V) chloride, which was prepared by the reaction of 1a with di(2-bromoethyl) ether [24]. Another type of tricationic P porphyrins (4d4f) was prepared by N-alkylation of bis[3-(4-pyridyl)propoxo]tetraphenylporphyrinatophosphorus(V), which was prepared by the reaction of (tpp) with 3-(4-pyridyl)-1-propanol. The 4a4c containing both N-alkyl-pyridinio and ethylenedioxy units had large (>63 mM) whereas 4d4f containing only N-alkylpyridinio group had >3 mM of . It is well known that porphyrins tend to aggregate in aqueous solution to cause broadening of the absorption spectra. Therefore, we checked whether 4a4f aggregated or not in aqueous solution by absorption spectra. We have previously reported that half-width (HW) of Soret band of P porphyrins was about 16 nm in the case of monomer form whereas HW became more than 20 nm when P porphyrins formed aggregates [22]. The absorption spectra of 4a4f showed Soret band with 16.6–18.2 nm of HW, showing no aggregation even in phosphate buffer solution.

Koenigs-Knorr synthesis is a glycosylation method of alcohols in the presence of soft Lewis acids such as silver carbonate which plays a role to activate glycosyl halides. For the glycosylation of 1a which had chloride ion, we modified Koenigs-Knorr synthesis. Tetra-O-acetyl-5-bromo-α-D-glucopyranose was firstly activated with Ag2CO3 in the presence of molecular sieves in MeCN and then subjected to glycosylation of 1a P porphyrin which was stereoselectively introduced at β-position of glucose. This modification provided 5a in 44% yield (Scheme 5). The deacetylation of 5a was quantitatively transformed to 5b by Zemplen reaction which was carried out in MeOH with a catalytic amount of NaOMe [25]. The of 5a was determined to be 131 mM, as listed in Table 1. The of 5b was too large to determine precisely (>150 mM). Moreover, solubility of 5b in physiological saline was maintained at 6.2 mM, enough to handle it in biological systems. The values are listed in Table 1.

2.2. By Modification of Porphyrin Rings

As has been reviewed by Kalyanasundaram [26], the modification of porphyrin ring by hydrophilic N-methylpyridiniumyl groups is the most popular method for water-solubilization of porphyrins. Water-soluble metalloporphyrins (metal = Zn [29, 30], Fe [31], Mn [32], Cu [33], Co [34], and Au [35]) containing N-methylpyridiniumyl group have been reported. Also, we prepared P porphyrin containing N-alkyl pyridyl group on the porphyrin (6a6d, Scheme 6). The values of 6b6d were large (>45 mM) as listed in Table 1.

3. Bioaffinity of Porphyrins

Bioaffinity of porphyrins can be conveniently evaluated using microorganisms such as S. cerevisiae and Escherichia coli and specific proteins such as human serum albumin (HSA) and concanavalin A (ConA) acting as drug delivery proteins and a glucose-binding protein, respectively. In particular S. cerevisiae is a large size microorganism whose diameter is 5 μm and can survive in not only buffer solution but also pure water. Therefore, it is easy to handle in sterilization experiments and quantitative analysis.

3.1. Bioaffinity of 2 and 3 toward S. cerevisiae

Bioaffinity of complexes 2 and 3 towards S. cerevisiae cells was examined. The amounts of 2 and 3 incorporated into yeast cells were determined by quantitative analysis using a confocal laser scanning microscope (CLSM) as follows [21, 23]. An aqueous solution of 2 and 3 (0.5 mL, 50 μM) was added to the cell suspension (1.0 mL; ca. 2.5 × 104 cell mL−1) of S. cerevisiae. However it is difficult to perform CLSM analysis since cells moved around by their Brownian motion. Therefore, an aqueous solution of agar (1 wt%; 1.0 mL) was added to the solution in order to stop the motion of cells during the CLSM analysis. A portion of the prepared aqueous solution containing 2 and 3 (10 μM), S. cerevisiae (ca. 1.0 × 104 cell mL−1), and agar (0.4 wt%) was taken on a space (1 cm × 1 cm) surrounded by silicone spacer (thickness 50 μm) put on a glass slide. The glass slide was set on the stage to be subjected to absorption spectrophotometry with the CLSM. The saturated adsorption concentrations () of porphyrins (2 and 3) on yeast were determined by absorption spectrophotometry at the Q-band using molar coefficiency (Table 1), path length (b), and absorbance (A) according to Lambert-Beer’s law:. The was determined by the CLSM fluorescence image where the fluorescence was emitted from the inside rather than the walls of cells. The observed values are summarized in Table 1.

Figure 2 shows the dependence of on carbon number (n) of the alkyl group on axial ligands of 1 and 2. In the cases of 2a2f, values depended on . The values reached maximum at (2c). When (2f), the was less than a measurable lower limit (<1.65 mM). Thus bioaffinity of 2 and 3 to yeast can be evaluated by value. Also, in the cases of 3a3f, the values depended on n, as shown in Figure 2. The values reached the maximum at (3h). The 3h was concentrated into 171 mM in S. cerevisiae at 17100-fold from 10 μM of an aqueous solution. However, when , the decreased to be less than a measurable lower limit. Thus was optimized to be 6. Photosensitized inactivation of S. cerevisiae using 2 and 3 will be described in Section 4.3.

Moreover, adsorption mechanism of 3 in cells was discussed by the concentration-dependence aggregation in aqueous solution [20]. The structure of 3 in aqueous solution of the concentration range from aqueous 1 μM to 1.0 mM was estimated by measuring surface tension, absorption spectra, and NMR spectra. As results, it was found that the 3e3h ( and 6) were solved as monomers below 10 μM and formed aggregates above a range of 10–50 μM (Scheme 7(b)). This showed that the 3e3h with a high of 123–171 mM could pass through the cell wall to accumulate inside the cell. Once inside the cell, where they reached a concentration of more than 10 μM, they formed aggregates that were unable to escape to the outside of the cell. We called this the ship-in-a-bottle effect. On the other hand, the 3b3d did not form aggregates at any concentrations in the range from 1 μM to 1 mM. This indicated that they easily passed through the cell wall in both directions (Scheme 7(a)). Therefore, values remained <25 mM. In the case of 3i3j, whose values were very low, passage through the cell wall was not possible because aggregates formed even at <1 μM (Scheme 7(c)). Although the adsorption of P porphyrins (4) into S. cerevisiae was examined in a similar manner as 2 and 3, the adsorption of 4 into S. cerevisiae was not observed.

3.2. Bioaffinity of 4 toward HSA

HSA (MW 66500) which is the most abundant plasma protein has the ability to bind many kinds of drugs and is applied to drug delivery systems [36] (Scheme 8). Therefore, interaction between P porphyrins (4) and HSA was examined by the fluorescence quenching of HSA by 4 (Method A) as well as spectral changes in absorption (Method B) and fluorescence (Method C) spectra of 4 by the addition of HSA [24].

Analysis by Methods A and B showed that the adsorption obeyed Langmuir-type adsorption with binding constants (K): log K = 6.3 (4a), 4.3 (4b), 4.1 (4c), 4.6 (4d), 4.7 (4e), and 5.1 (4f). HSA has one tryptophan residue on major drug-biding site (Site I) [37]. The ratio of the number of molecules of 4 adsorbed on Site I to the total number of molecules of 4 adsorbed on whole HSA, which was defined as values, can be analyzed by Method C using modified Stern-Volmer plots [38]. By the analysis of Method C, compounds 4 were classified into three categories according to their values: 1, 0.6, and 0.2. The 4b () and 4c () with values of 1.0 bind to Site I site-specifically (Scheme 8). However, 4a (), 4d (), and 4e () with values of 0.6 bind to HSA at two types of sites with similar affinities. In the case of 4f () with an value of 0.2, most molecules of 4f adsorbed on the HSA far from the tryptophan residue leading to inefficient quenching. Thus, the length of the alkyl chain affects the site-specific binding to HSA. When 4 had moderately long alkyl chains on the pyridinio group, 4 was adsorbed strongly on the hydrophobic pocket near a tryptophan residue of HSA.

3.3. Bioaffinity of 5 toward Concanavalin A

ConA (MW 25700 for one subunit) is a glucose-binding protein which is composed of four identical subunits under basic conditions (pH > 7) [39]. Each subunit has one sugar-binding site which includes two tyrosine residues (tyrosine-12 and tyrosine-100) [40, 41] (Scheme 9). Moreover, ConA emitted fluorescence at 335 nm under excitation at 230 nm, since the subunit contains four tryptophan and seven tyrosine residues which have absorption at UV region. There are tyrosine-12 and tyrosine-100 at the sugar-binding site and they are able to interact with sugar moieties [41, 42]. However, four tryptophan residues do not exist near the sugar-binding site.

According to Methods A, B, and C which have been reported for the interaction of 4 with HSA, the interaction of ConA with 5 was analyzed. Analysis by Method A showed that 5a and 5b strongly bound ConA near to tryptophan and/or tyrosine residues. Analysis by Method B provided 4.5 × 104 and 6.4 × 104 M−1 of the binding constants (K) for 5a and 5b, respectively.

Fluorescence spectral changes of 5 with L-tryptophan and L-tyrosine were examined in the buffer solutions. L-tryptophan quenched the fluorescence of 5, whereas L-tyrosine did not quench. Free energy change (ΔG) for an electron transfer from L-tryptophan to the excited singlet state of 5 (5*) was assumed according the Rehm-Weller equation: ΔG = , where denotes an oxidation potential of L-tryptophan (0.638 V versus Ag/Ag+), denotes a reduction potential of 5a (–0.895 V) and 5b (–0.835 V), and denotes excitation energies of 5a and 5b which were 2.02 eV (612 nm) and 2.03 eV (614 nm), respectively. Thus, the ΔG values for electron transfer from L-tryptophan to 5* were estimated as negative values for 5a (−0.547 eV) and 5b (−0.497). Therefore, the fluorescence quenching of 5 with ConA might proceed through the photoinduced electron transfer from tryptophan residues to 5* whose fluorescence lifetimes were determined to be 5.6 ns in aqueous solution. Actually, the fluorescence of 5a was quenched with ConA whereas the fluorescence of 5b was not quenched with ConA (Method C). The 5a adsorbed near these tryptophan residues without site-specificities. The 5b adsorbed at the site near tyrosine residues but far from tryptophan residues.

4. Photosensitization of Porphyrins

4.1. Mechanism of Photosensitization

The photosensitization by porphyrins mainly proceeds through two processes such as electron transfer from biomolecules to the photoexcited photosensitizer (Type I mechanism) and energy transfer from photoexcited photosensitizer to oxygen molecule to generate singlet oxygen (1O2) (Type II mechanism). The 1O2 is a very important reactive oxygen species for the photosensitized reaction of porphyrins, because 1O2 can be easily generated by a wide range of wavelengths (ultraviolet region ~ near-infrared region) including visible light. Formation of other reactive oxygen species, such as superoxide and hydroxyl radicals, in general, requires ultraviolet radiation. The electron transfer and 1O2 generation processes lead to the oxidation of guanine residues of DNA and certain amino acid residues of protein.

4.2. Photosensitized DNA Damage by 1b and 1a

DNA is an important targeting biomacromolecule of photosensitized reaction [4244]. The 1b demonstrates bactericidal activity during photoirradiation and this effect has been applied for water sterilization [45]. To elucidate the mechanism of phototoxic effect of 1b, its photodamaging activity for biomacromolecule was investigated using DNA as a model targeting biomacromolecule. Photoirradiated 1b damaged [32P]-5′-end-labeled DNA fragments [46]. 1b induced markedly severe photodamage to single-stranded DNA rather than to double-stranded DNA. Photoexcited 1b frequently caused DNA cleavage at the guanine residues after E. coli formamidopyrimidine-DNA glycosylase or piperidine treatment. These results indicated the formation of 8-oxo-7,8-dihydro-2′-deoxyguanine (8-oxo-G), a typical oxidized product of guanine. HPLC measurement also confirmed the formation of 8-oxo-G and showed that the content of 8-oxo-G in single-stranded DNA is larger than that in double-stranded DNA. Because the single-stranded DNA can be easily oxidized by 1O2, this result suggests the contribution of 1O2-mediated oxidation of guanine. The effects of scavengers of reactive oxygen species on DNA damage also supported the involvement of 1O2. These results have shown that the mechanism via 1O2 formation mainly contributes to the phototoxicity of 1b.

On the other hand, 1b induced DNA damage specifically at the underlined G of 5′′-GG, 5′′-GGG, and 5′′-GGGG in double-stranded DNA. The sequence-specificity of DNA damage is quite similar to that induced by the Type I photosensitizers [47]. The redox potential of one-electron oxidation of guanine is lowest in the four nucleobases [48]. Furthermore, the molecular orbital (MO) calculations have revealed that the consecutive guanines in double-stranded DNA significantly lower the highest occupied MO energy [49, 50]. Consequently, the consecutive guanines are selectively damaged through the electron transfer mechanism. These results showed that photoinduced electron transfer slightly participates in the photosensitized reaction of 1b.

1b induces DNA photodamage via the generation of 1O2 and electron transfer. Similar mechanisms can damage other biomacromolecules, such as protein and the phospholipid membrane. The damage to biomacromolecules via these mechanisms may participate in the phototoxic effect of 1b.

The phosphorus(V) porphyrin also induces an oxidative electron transfer reaction [51]. The 1a, a cationic water-soluble porphyrin, induced DNA photodamage via similar mechanism of 1b, 1O2 generation and electron transfer [51]. The study of near-infrared emission measurements demonstrated the 1O2 generation by photoexcited 1a. The fluorescence quenching of 1a by DNA supported the electron transfer mechanism. Under aerobic conditions, 1a-photosensitized damage was more severe for single-stranded DNA compared to its double-stranded counterpart. Photoirradiated 1a damaged every guanine residue in single-stranded DNA. HPLC measurements confirmed the formation of 8-oxo-G. The guanine-specific DNA damage and the enhancement in single-stranded DNA suggest that the 1O2 generation mainly contributes to the mechanism of DNA photodamage by 1a similarly to that by 1b. On the other hand, for double-stranded DNA, photosensitized damage at consecutive guanines was much less pronounced. Because the consecutive guanines act as a hole trap [4750], this DNA-damaging pattern suggests the partial involvement of electron transfer mechanism. However, DNA damage by electron transfer mechanism was not a main mechanism, possibly due to the reverse electron transfer mechanism.

4.3. Photosensitized Protein Damage by 7

Protein is also an important targeting biomacromolecule. The photosensitized protein damage by phosphorus(V) porphyrin has been investigated using the abovementioned HSA as a protein model [28]. A water-soluble porphyrin, dimethoxyP(V) tetraphenylporphyrin chloride (Scheme 10, 7a), photosensitized HSA damage. The quantum yield of 1O2 generation (ΦΔ) by 7a (0.64 in ethanol) was comparable with that of typical porphyrin photosensitizers. Absorption spectrum measurement demonstrated the binding interaction between 7a and HSA. HSA has one tryptophan residue. Since tryptophan, a relatively strong fluorescent amino acid, is easily oxidized by both mechanisms (1O2 generation and electron transfer), leading to fluorescence quenching (Figure 3), the fluorometry of tryptophan is a convenient method to evaluate the protein damage. Photoirradiated 7a damaged the amino acid residue of HSA, resulting in the decrease of the fluorescence intensity from the tryptophan residue of HSA. Sodium azide (NaN3), a 1O2 quencher, partially inhibited the HSA damage, supporting the 1O2-mediated protein damage. However, the effect of sodium azide is not completed, suggesting that the electron transfer mechanism contributes to protein damage as does 1O2 generation. The decrease of the fluorescence lifetime of 7a by HSA supported the electron transfer mechanism. The estimated contribution of the electron transfer mechanism is 0.64. These results suggest that the activity of 7a can be preserved under lower oxygen concentration condition such as tumor.

The specific characteristics of the phosphorus(V) porphyrin are the variety of the substituted axial ligand and the relatively low redox potential of the one-electron reduction in the photoexcited state. The fluorination effect of the axial ligand of phosphorus(V) porphyrin (Scheme 10) on the photosensitized reaction has been examined [52]. As a target protein model, HSA was used. The activity of 1O2 generation by diethoxy (tetraphenylporphyrinato)phosphorus(V) complex (7b) was slightly improved by the fluorination of the ethoxy chains (7c). The quantum yields of 1O2 generation in sodium phosphate buffer were 0.59 and 0.68 for 7a and 7b, respectively. Absorption spectrum measurements demonstrated the binding interaction between these phosphorus(V) porphyrins and HSA. Photoirradiated phosphorus(V) porphyrins damaged the amino acid residue of HSA, resulting in the decrease of the fluorescence intensity from the tryptophan residue. A 1O2 quencher, NaN3, could not completely inhibit the damage of HSA, suggesting that the electron transfer mechanism contributes to protein damage. The decrease of the fluorescence lifetime of these porphyrins by HSA supported the electron transfer mechanism. The estimated contributions of the electron transfer mechanism are 0.57 and 0.44 for the fluorinated and nonfluorinated P(V) porphyrins, respectively. The total quantum yield of the protein photooxidation was slightly enhanced by this axial fluorination.

4.4. Photosensitized Inactivation of S. cerevisiae by 2 and 3

In order to use 1a and 1b as photocatalyst in aqueous solution, these porphyrins were fixed on silica gel to produce the composite of porphyrins with silica gel. These composites were utilized for photochemical bactericidal reaction of E. coli [53], S. cerevisiae (yeast) [54], and Legionella pneumophila [45]. In particular, the composite was applied to the bactericidal reaction of Legionella species occurring in cooling tower and public fountain under irradiation of fluorescent lamp and sun light. Details have been published in review [2]. Here, we review the photosensitized inactivation of S. cerevisiae by water-soluble Sb and P porphyrins.

The photosensitized inactivation S. cerevisiae by Sb porphyrins (2a2f) and P porphyrins (3b3h) was examined. The photoinactivation under an argon-saturated atmosphere did not occur at all. Therefore, 1O2 was responsible for the active species of the photoinactivation. Under visible-light irradiation, 2 and 3 were excited to singlet state to transform to triplet state with high efficiency. The energy transfer from the triplet state of 2 and 3 to oxygen molecules in a ground state (3O2) took place to generate 1O2 through Type II mechanism. The values of ΦΔ for these porphyrins were determined in an aqueous solution to be 0.48 (1b), 0.65 (2b), 0.53 (2d), 0.62 (3c), 0.69 (3d), and 0.73 (3e). These values were comparable values with other metalloporphyrins: 0.56 (H2(tpp)), 0.65 (Zn(tpp)), and 0.62 (Mg(tpp)) [55]. The lifetimes of triplet state of 3b3d were estimated from the time profile of 1O2 emission and were determined to be 1.5–1.6 μs.

The photoinactivation of S. cerevisiae was performed for an aqueous solution (10 mL) of S. cerevisiae NBRC 2044 (1 × 104 cell mL−1) and 2 and 3 (5–500 nM) in an L-type tube under irradiation by a fluorescent lamp on a reciprocal shaker. The photoinactivation was evaluated by the activity factor (/μM−1 h−1) which was derived by using the minimum effective concentrations of 2 and 3 (/μM) and half-life (/h) which is the time required to be reduced to one-half initial concentration of S. cerevisiae. As listed in Table 1, the large values (12.6–188.7 M−1 h−1) were obtained in 2c2e and 3e and 3g-3h which have large values (22.5–171 mM) and long alkyl chain (–12). It was requisite that 2 and 3 were both water-soluble and oil-soluble. Presumably the oil-soluble feature was advantageous in passing through the cell wall of the yeast which consisted of hydrophobic peptidoglycan and the water-soluble feature was advantageous in the adsorption occurring at hydrophilic sites inside the cell, causing fatal damage to the yeast.

4.5. Modeling of PDT Using a Phospholipid Liposome

As a model for PDT, the photosensitization of 2e was investigated in a phospholipid liposome (LP) where target cell’s substrate was modeled by 9,10-dimethylanthracene (DMA) [26]. The 2e photosensitized the oxidation of DMA to 9,10-epidioxy-DMA (epidioxy-DMA) [56, 57] in LP. The LP was prepared by Banghom method [58] using 1,2-dipalmitoyl-sn-glycerol 3-phosphocholine (DPPC) (Scheme 11). A chloroform solution of DPPC (2.0 mM, 1.0 mL) was added to a 25 mL flask, and then CHCl3 was removed with a rotary evaporator under reduced pressure at 40°C. All traces of the organic solvent were then removed by drying with a vacuum pump. The resulting thin film was hydrated with pure water in an ultrasonic bath at 70°C for 30 min to give an LP solution. The LP solution was cooled to room temperature and maintained at 25°C overnight. The LPs were found to have diameters of 50–240 nm (average diameter of 83 nm) by dynamic light scattering, measured at 25°C. DMA and 2e were incorporated into LP as follows. Next, an aqueous solution of 2e (0.1 mM, 6 μL) and a MeOH solution of DMA (1 mM, 2.7–15.0 μL) were added to an aqueous LP solution (3.0 mL, DPPC = 0.6 μmol) to give an LP solution (3.0 mL) containing 2e (0.2 μM; 6 nmol) and DMA (2.7–15.0 nmol). The adsorption isotherms of 2e into the LP showed Langmuir-type adsorption with binding constants in 2.87 × 106 M−1 through the hydrophobic interactions with the dodecyloxo ligand and the core of the LP membrane. The solution was irradiated at 550 nm with gentle stirring under an aerated atmosphere at 25°C. The photoreaction was monitored by the fluorescence coming from DMA at 430 nm under an excitation at 375 nm, since the epidioxy-DMA was nonfluorescent.

From the kinetic analysis, the limiting quantum yield was determined to be 0.73. The reaction was not suppressed by the addition of NaN3 which is known to be a quencher for 1O2. Moreover, the electron transfer from DMA to the 12e* was a favorable exothermic process, since free energy change was calculated to be −0.51 eV by the Rehm-Weller equation. Therefore, it was postulated that the photosensitized oxygenation reaction of DMA with 2e in the LP occurred through Type I mechanism where photoinduced electron transfer from DMA to the excited singlet state of 1e (11e*) took place to generate and . The reacted with which was generated from the reduction of O2 by to produce epidioxy-DMA.

5. Conclusions

A variety of porphyrins have been applied as the drug for PDT [59]. In general, PDT can damage cancer cells through generation of 1O2 by photosensitized energy transfer (Type II mechanism) [60]. Free base type of porphyrin compounds (e.g., photofrin) has been used as Type II photosensitizer because their protonated species have water solubility and strong absorption in visible region with sufficient triplet energy to produce the excited singlet state of oxygen (1; excited energy = 0.98 eV, 1270 nm) [61]. Also the photosensitized damage through generation of ion radicals by photoinduced electron transfer (Type I mechanism) has received much attention because Type I mechanism can operate even under low O2 concentration in tumor cell. Metalloporphyrins are an attractive candidate because of relatively high electron-accepting power compared with free base porphyrins. Metalloporphyrin complexes have redox properties depending on the interaction between central metal and tetraphenylporphyrinato (tpp) chromophore. Reduction potentials of five-valent P and Sb complexes such as 2b (−0.50 V versus SCE) [28] and 1b (−0.51 V) [2] were relatively shifted to positive compared with those of two-valent metal complex such as (tpp)Zn (−1.31 V), (tpp)Ni (−1.18 V), and (tpp)Pb (−1.10 V) [2]. Therefore it is expected that water-soluble P(V) and Sb(V) porphyrins can inactivate cancer cells through both Types I and II mechanisms.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.