Introduction

Photodynamic inactivation (PDI) of bacteria has received considerable attention as a methodology leading to the medical application such as photodynamic therapy for tumor cells. PDI refers to the use of a visible-light source, oxidizing agents (e.g., O2), and photosensitizers. Photosensitizers absorb light energy that causes an energy transfer to O2, which leads to the formation of reactive oxygen such as singlet oxygen (1O2), thereby inactivating cells and bacteria. Preliminary studies on the photodynamic action for biological systems started in 1930s by the PDI of phages using methylene blue (Clifton 1931; Perdrau and Todd 1933). Among the large variety of photosensitizers developed for PDI over the last 60 years, porphyrins and metalloporphyrins become attractive sensitizers owing to their strong absorption band in the visible-light region (Pandey and Zheng 2000; Nyman and Hynninen 2004; Shiragami et al. 2005; Ethirajan et al. 2011).

For biological applications of porphyrins, their water solubility is an important characteristic for handling the aqueous solution. However, in general, porphyrin derivatives have poor water-solubility. The introduction of cationic groups into porphyrins is the most popular method to enhance the water solubility of porphyrins. In particular, the introduction of an alkylpyridinium (Apy) group into porphyrins is a convenient method to make porphyrins water-soluble (Kalyanasundaram 1984; Girek and Sliwa 2013). A typical Apy-bonded porphyrin is represented by meso-tetra[4-(1-methylpyridinium)]porphyrin (TMP). The first application of TMP to PDI was reported by Ben Amor et al. in 1998 (Ben Amor et al. 1998). For the last two decades, a variety of Apy-bonded porphyrins have been prepared and studied for PDI (Kano et al. 2000; Kubát et al. 2000; Trommel and Marzilli 2001; Lang et al. 2004; Banfi et al. 2006; Haeubl et al. 2009; Batinic-Haberle et al. 2012). We have interested in axially Apy-bonded tricationic P-porphyrins (1, Scheme 1) (Matsumoto et al. 2013; Matsumoto et al. 2016; Matsumoto et al. 2017a). It is advantageous that the water solubilization is easily achieved through the modification of the axial ligands. Recently we have reported the PDI of E. coli using several kinds of axially Apy-bonded P-porphyrins (1e1j) (Matsumoto et al. 2017a). Here, to develop more efficient 1O2 sensitizers than 1e1j, we assessed the potential of another type of Apy-bonded P-porphyrins (1a1d) in the PDI of E. coli.

Scheme 1
scheme 1

Tricationic P-porphyrins bonded to alkylpyridinium (Apy) (1a1j)

Materials and methods

Instruments

1H nuclear magnetic resonance (NMR) (400 MHz) and 13C NMR (100 MHz) spectra were obtained with a Bruker AV 400 M spectrometer in CD3OD solutions using SiMe4 as an internal standard. High-resolution mass spectra (HRMS) were measured on a Thermo Scientific Q Exactive mass spectrometer equipped with an electrospray ionization source. The molar absorption coefficients (ε) of 1 at the Soret and Q bands were determined from the visible spectra measured in MeOH using a JASCO V-550 spectrophotometer. The fluorescence spectra of solutions were measured on a Shimadzu RF-5300PC spectrometer. Time-resolved fluorescence lifetime were measured via time-correlated single photon counting using a lifetime fluorescence spectrometer (Horiba, DeltaFlex) equipped with a PPD detector and a pulsed laser diode (Horiba, DeltaDiode DD-415L, peak wavelength: 419 nm, pulse width: 70 ps). Fluorescence at 595 nm emitted from MeOH solution of 1 under aerated conditions was collected with a monochrometer on the spectrometer.

General procedure for the preparation of Apy-bonded tricationic P-porphyrins (1a–1d)

As we have described in previous reports (Matsumoto et al. 2013), [(HO)2P(tpp)]Cl (2b, tpp = tetraphenylporphyrino), which was prepared by hydrolysis of [Cl2P(tpp)]Cl (2a; 300 mg) by refluxing in a mixed solvent of MeCN (160 mL) with pyridine (60 mL) and H2O (60 mL), was used as starting material (Scheme 2). Bis(5-bromo-3-oxa-pentyloxo)tetraphenylporphyrinatophosphorus(V) chloride (2c) was synthesized by the alkylation of 2b (80 mg) with di(2-bromoethyl) ether (1 mL) in the presence of K2CO3 (19 mg) and 18-crown-6 ether (4.2 mg) in MeCN (5 mL) at 50 °C. The compound bis[5-(3-methyl-1-pyridinio)-3-oxa-pentyloxo]tetraphenylporphyrinatophosphorus(V) dibromide, chloride (1a) was prepared via the reaction of 2c (51 mg) with 3-methylpyridine (1.0 mL) in dry MeCN (10 mL) at 100 °C for 20 h. Similarly bis[5-(4-ethyl-1-pyridinio)-3-oxapentyloxo]tetraphenylporphyrinatophosphorus(V) dibromide chloride (1k) was prepared via the reaction of 2c (63 mg) with 4-ethylpyridine (1.0 mL) in dry MeCN (10 mL) at 100 °C for 20 h. Spectroscopic characterizations of 1b1d have been reported in previous studies (Matsumoto et al. 2013; Matsumoto et al. 2017b). The spectral data of newly prepared 3-methyl analog (1a) and 4-ethyl analog (1k) are provided as follows.

Scheme 2
scheme 2

Preparation of tricationic P-porphyrins (1a and 1k)

Bis[5-(3-methyl-1-pyridinio)-3-oxapentyloxo]tetra-phenylporphyrinatophosphorus(V) dibromide chloride (1a)

Yield 73% from [(HO)2P(tpp)]Cl. 1H NMR (400 MHz, CD3OD): δ −2.20 (dt, JP–H = 11.1 Hz, J = 4.6 Hz, 4H, P–OCH2CH2O–), 0.70 (brs, 4H, P–OCH2CH2O–), 2.25 (s, 6H, CH3), 2.43 (t, J = 4.7 Hz, 4H, –OCH2CH2N–), 3.84 (t, J = 4.7 Hz, 4H, –OCH2CH2N–), 7.46 (dd, J = 8.0, 6.0 Hz, 2H, H-5 of C5H4N), 7.74 (d, J = 6.0 Hz, 2H, H-4 of C5H4N), 7.80–7.86 (m, 12H, H-3, H-4, and H-5 of C6H5), 7.99–8.01 (m, 8H, H-2, and H-6 of C6H5), 8.04 (s, 2H, H-2 of C5H4N), 8.20 (d, J = 8.0 Hz, 2H, H-6 of C5H4N), 9.18 (d, JP–H = 2.9 Hz, 8H, pyrrole β); 13C NMR (100 MHz, CD3OD): δ 18.31 (CH3), 60.85 (–OCH2CH2N), 61.68 (d, JP–C = 13.9 Hz, P–OCH2CH2O–), 68.25 (d, JP–C = 18.9 Hz, –OCH2CH2N), 68.74 (–OCH2CH2N), 117.63 (meso), 127.73 (C-5 of C5H4N), 129.6 (C-3 of C6H5), 131.0 (C-4 of C6H5), 134.6 (d, JP–C = 5.1 Hz, pyrrole β), 134.7 (C-2 of C6H5), 136.7 (C-1 of C6H5), 140.4 (C-4 of C5H4N), 140.5 (pyrrole α), 142.6 (C-2 of C5H4N), 145.1 (C-3 of C5H4N), 147.2 (C-6 of C5H4N); HRMS calcd for C64H58N6O4P3+ [M3+]: 1005.4241, m/z 335.1414. Found: 335.1415.

Bis[5-(4-ethyl-1-pyridinio)-3-oxapentyloxo]tetra-phenylporphyrinatophosphorus(V) dibromide chloride (1k)

Yield 72% from [(HO)2P(tpp)]Cl. 1H NMR (400 MHz, CD3OD): δ −2.18 (dt, JP–H = 10.3 Hz, J = 5.1 Hz, 4H, P–OCH2CH2O–), 0.72 (brs, 4H, P–OCH2CH2O–), 1.26 (t, J = 7.5 Hz, 6H, CH3), 2.41 (t, J = 4.5 Hz, 4H, –OCH2CH2N), 2.84 (q, J = 7.5 Hz, 4H, –CH2CH3), 3.80 (t, J = 4.5 Hz, 4H, –OCH2CH2N), 7.50 (d, J = 6.2 Hz, 4H, H-3 of C5H4N), 7.82–7.86 (m, 12H, H-3, H-4, and H-5 of C6H5), 7.87 (d, J = 6.2 Hz, 4H, H-2 of C5H4N), 8.00–8.02 (m, 8H, H-2 and H-6 of C6H5), 9.17 (d, JP–H = 2.7 Hz, 8H, pyrrole β); 13C NMR (100 MHz, CD3OD): δ 13.83 (–CH2CH3), 29.71 (–CH2CH3), 60.27 (–OCH2CH2N), 62.00 (brs, P–OCH2CH2O–), 68.44 (brs, P–OCH2CH2O–), 68.94 (–OCH2CH2–N), 117.78 (brs, meso), 127.98 (C-3 of C5H4N), 129.70 (C-3 of C6H5), 131.04 (C-4 of C6H5), 134.62 (d, JP–C = 3.5 Hz, pyrrole β), 134.83 (C-2 of C6H5), 136.80 (C-1 of C6H5), 140.60 (pyrrole α), 144.82 (C-1 of C5H4N), 166.19 (C-4 of C5H4N); HRMS Calcd for C66H62N6O4P3+ [M3+]: 1033.4554, m/z 344.4851. Found: 344.4845. Molar absorption coefficients of 1k in MeOH were much smaller than that of 1b, because of broadening of Soret and Q bands in UV–Vis absorption spectra due to aggregation of porphyrin chromophores (Fig. 1). As a results, 1k did not show clear 13C NMR spectrum and the doublet peak for the carbons at P–OCH2CH2O–, meso position, and pyrrole-β.

Fig. 1
figure 1

UV–Vis absorption spectra of 1b (botted line) and 1k (solid line) in MeOH

Photoinactivation of E.coli

E. coli K-12 (IFO 3335) was cultured aerobically at 30°C for 8 h in a basal medium (pH 6.5) consisting of bactotryptone (10 g L−1), yeast extract (5 g L−1), and NaCl (10 g L−1). After centrifugation of the cultured broth at 12,000 rpm for 10 min, the harvested cells were washed with physiological saline (NaCl, 7 g L−1) and then suspended in physiological saline, resulting in a cell suspension of E. coli (6.4 × 104 cells mL−1). The cell concentrations were determined using a calibration curve and turbidity quantified by the absorbance measured at 600 nm on a UV−Vis spectrometer.

A phosphate buffer (0.1 M, pH 7.6) was prepared by dissolving Na2HPO4 (2.469 g) and NaH2PO4 (0.312 g) in 100 mL of water. The suspension of E. coli cells (1 × 105 cells mL−1, 1.0 mL), an aqueous solution of 1a1d and 1k (40–200 μM, 0.1 mL), and the phosphate buffer (0.1 M, pH 7.6, 8.9 mL) were introduced into L-type glass tubes, resulting in buffer solution (10 mL) containing E.coli (1 × 104 cells mL−1) and 1a1d and 1k (0.25−1.0 μM). The L-type glass tubes were set on a reciprocal shaker and shaken at 160 rpm at room temperature for 2 h under dark conditions (Matsumoto et al. 2017a). Irradiation was performed using a fluorescent lamp (Panasonic FL-15ECW, Japan; λ = 400–723 nm; the maximum intensity: 545 nm; 10.5 W cm–2) on a reciprocal shaker for 2 h at room temperature. A portion of the reaction mixture (0.1 mL) was taken at 20-min intervals and plated on an agar medium.

The amount of the living cells (B) was defined as the average number of E. coli colonies that appeared after an incubation period of 30 h at 30 °C in three replicate plates. The B values for 1a1d and 1k were recorded at each irradiation times, as summarized in Table 1.

Table 1 The amount of living cells (B) in PDI of Escherichia coli using 1

Results and discussion

Axially Apy-bonded P-porphyrins

The values of water solubility (CW) of 1a1k are listed in Table 2. In addition, Table 2 lists the coefficient (ε) of Soret band around 431 nm and Q-band at 562 nm. Effect of Apy group on the physicochemical properties of porphyrin was examined using [(MeO)2P(tpp)]Cl (3a, Scheme 3) as a reference porphyrins without Apy. The absorption coefficient (ε) of 3a was determined to be 3.12 × 105 M−1 cm−1 for Soret band (λmax = 424 nm) and 1.82 × 104 M−1 cm−1 for Q band (554 nm). The reduction potential (E1/2red vs. Ag/AgNO3) of 3a was measured to be −0.82 V. The fluorescence of 3a was observed at 610 nm under excitation at 550 nm. The fluorescence quantum yield (ΦF) and fluorescence lifetime (τF) of 3a was 0.0388 and 4.64 ns, respectively. In contrast, 1d with Apy group has E1/2red (−0.85 V), ΦF (0.0350), and τF (4.91 ns) (Matsumoto et al. 2015). These physicochemical values were similar to those of 3a. Thus, the introduction of the Apy on axial ligands does not affect the physicochemical parameters of the P(tpp) moiety. Moreover, the 1 could sensitize 1O2 formation in high efficiency. The quantum yields for the formation of 1O2 were measured to be 0.87 for 1d (Matsumoto et al. 2017a). The reduction potentials of the P-porphyrins shifted to positive compared with the free-base porphyrins due to more positive pentavalent phosphorus. Therefore, they are expected to have abilities of electron transfer sensitizer in addition to energy transfer sensitizer.

Table 2 The PDI of E. coli by porphyrins (1a1j)
Scheme 3
scheme 3

Structure of [(MeO)2P(tpp)]Cl (3a)

PDI activity of 1a−1d and 1k

The PDI of E. coli was performed by irradiation of buffer solution (10 mL) containing E.coli (1 × 104 cells mL−1) and 1a1d and 1k (0.25−1.0 μM) for 2 h using a fluorescent lamp at 545 nm. Based on Table 1, the survival ratios were calculated as 100B/B0 where B0 is the initial amount of E. coli. From the time-course plots of survival ratios, the half-life (T1/2 in min), i.e., the time required to reduce B from B0 to 0.5B0, was measured. The minimum concentrations of the sensitizer ([P]) were adjusted such that T1/2 attained values between 20 and 120 min. Thus, the bactericidal activity (AF in μM−1 h−1) was evaluated using the following equation: AF = 60/([P] × T1/2). Table 2 summarizes [P], T1/2, and AF values of 1a1d and 1k along with those of 1e1j.

Optimization of carbon atoms (n) in the alkyl group on the APy group

As shown in Table 2, the AF values varied by the number of carbon atoms (n) in the alkyl group on the APy group in a series of 1a1d and 1k. Figure 2 shows the dependence of the AF values on n. The maximum value of AF appeared at n = 2 (1b) whose [P] value for E. coli was 0.25 μM. Therefore, the optimized AF and [P] values of 1b, which was 3-ethyl analog, were compared with that of 4-ethyl analog (1k). It was found that the AF value of 1k was lower than that of 1b. As shown in Fig. 1, broadening of Soret and Q bands occurred due to aggregation of porphyrin chromophores. It is suggested that aggregation caused to lower the AF value of 1k. As has reported previously in a series of 1e1j (Matsumoto et al. 2017a), the optimized [P] value in PDI of E. coli was 0.40 μM in 1i whose n was 7. Thus, the optimum [P] value for E. coli was 0.25 μM of 1b (n = 2) among 1a1k totally.

Fig. 2
figure 2

Relationship between the AF values and number of carbon atoms (n) in alkyl group on the alkylpyridinium (Apy) in 1a1d (), 1e1j (◇), and 1k ()

The [P] value (0.25 μM) of 1b for E.coli was larger than the reported [P] value (50 nM) of 1b for S. cerevisiae (Matsumoto et al. 2016). In general, gram-negative bacteria, such as E. coli, have complex cell wall structures comprising phospholipids, lipopolysaccharides, lipoteichoic acids and lipoproteins, which pose an impermeable barrier to antimicrobial agents (Alves et al. 2014). Therefore, the present results are accord with these features already known.

Conclusion

Apy-bonded tricationic P-porphyrins (1) could photoinactivate E. coli. The [P] value for E. coli was optimized at bis[5-(3-ethyl-1-pyridinio)-3-oxapentyloxo]tetraphenylporphyrinatophosphorus dibromide, chloride (1b). Polycationic porphyrins have strong binding affinities to DNA (Pasternack et al. 2001; Sirish et al. 2002; Marczak et al. 2007; Haeubl et al. 2009; Tada-Oikawa et al. 2009; Kim et al. 2013) and proteins (Gyulkhandanyan et al. 2013). Alkyl chains might result in moderate hydrophobicity to take advantage of passing through cell wall. Therefore, it is important to provide the porphyrins with both polycationic character and hydrophobicity for an efficient PDI of E. coli.