1 Introduction

Multi-drug resistant (MDR) bacterial infections are already responsible for approximately 1.3 million deaths each year. In particular, MDR Escherichia coli is mentioned in World Health Organization reports as one of the most troublesome microorganisms [1], causing up to 100,000 deaths annually [2]. Photodynamic inactivation (PDI), also named antimicrobial photodynamic therapy (aPDT), emerged as a viable alternative therapy to treat topical MDR bacterial infections due to its broad antimicrobial activity spectrum and low potential for developing resistance. PDI involves the use of a designed photosensitizing molecule, along with an appropriate light source and molecular oxygen, to generate reactive oxygen species (ROS) that can eliminate bacteria [3,4,5,6]. The clinical translation of PDI to photodisinfection of external wounds, such as diabetic foot ulcers, or of upper respiratory tract, ocular and buccal infections, is still limited by scientific challenges, namely, in the design and synthesis of photosensitizers at multigram scale, structural optimization to facilitate penetration of the double barrier present in Gram-negative bacteria, and effective infiltration in biofilms [7,8,9]. Studies with positively charged photosensitizers, such as cationic meso-(1-methylpyridinium-4-yl) [10,11,12,13,14,15,16,17,18,19,20,21] and meso-(4-N,N,N-trimethylammoniumphenyl) porphyrins [22,23,24], showed promising photoinactivation of E. coli and encouraged further work with cationic porphyrin derivatives.

Cationic imidazolyl-substituted photosensitizers, namely phthalocyanines and porphyrins, were successfully employed to photoinactivate Gram-negative bacteria such as E. coli [8, 25, 26]. Pereira, Dabrowski and co-workers showed that amphiphilic imidazole-phthalocyanines with 4 or 8 positive charges and 2 or 8 imidazolyl side chains effectively photo-inactivated E. coli. A phthalocyanine with 4 positive charges and a short chain (two carbon atoms) gave the best results: 7-log inactivation at a concentration of 100 nM and 10 J/cm2 (white light irradiation) [25]. However, these imidazolyl phthalocyanines have problems of aggregation and solubility in aqueous solution, and our attention shifted to meso di- or tetra-imidazolyl porphyrins, which were also very effective in the photoinactivation of E. coli [8]. Furthermore, we also demonstrated that the size and number of charges are particularly relevant for the inactivation of S. aureus biofilms. A small size dicationic meso-imidazolyl porphyrin resulted in higher permeation through the polysaccharide polymer surface of biofilms, with consequent bacterial photoinactivation (7 log CFU reduction) with just 5.2 nM and 5 J/cm2 [8]. Despite this progress, it is difficult to completely eradicate bacteria present in infected tissues and avoid recurrence with PDI alone [10, 27,28,29,30,31,32]. Recent studies emphasized that the synergic combination of PDI with existing antibiotics is a more promising approach to tackle multi-resistant bacteria and delay antibiotic resistance [26, 33, 34]. This approach enhances the efficacy of antibacterial treatments by leveraging the strengths of both PDI and antibiotics [27, 35,36,37]. We showed [26], in agreement with literature [34, 36], that the order and number of each individual treatment sessions impact the inactivation of MRSA or E. coli, and found that PDI followed by antibiotic administration yielded the best results.

Although the combination of PDI with antibiotics is probably the best approach to treat infections that extend beyond the surface accessible to illumination, the combination of PDI with other types of antimicrobials may be sufficient to treat a superficial localized infection. Natural antimicrobials with a low propensity to generate drug resistance are desirable for such combinations, especially if they are approved for human use as GRAS (Generally Recognized as Safe). This is the case with trans-cinnamaldehyde, a major component of the cinnamon plant, known for a broad range of antimicrobial activity against various strains of bacteria [38]. In human dermatological studies, the No-Observed-Adverse-Effect-Level (NOAEL) for cinnamaldehyde sensitization was set at 591 µg/cm2 [39], which roughly corresponds to 0.5% content in a topical formulation. Several mechanisms may be involved in the antimicrobial effect of cinnamaldehyde, and their relevance may depend on the dosage. Although more research may be necessary to obtain an integrated view of its mechanism of action, there is a consensus that cinnamaldehyde alters the integrity of bacterial cell wall and their permeability [38, 40,41,42].

In this paper, we describe a pioneer study to evaluate the synergy of the combination of cinnamaldehyde with PDI in the photoinactivation of planktonic E. coli. The family of cationic di-imidazolyl-substituted porphyrin photosensitizers was expanded with a new member (IP-H-OH2+) and structure–activity relations in PDI of planktonic E. coli were investigated. Preliminary studies in vivo show the feasibility of PDI with this class of photosensitizers.

2 Results and discussion

2.1 Synthesis

The dicationic meso-imidazolyl porphyrins were synthesized through alkylation of the commercially available 5,15-bis(1-methylimidazol-2-yl)porphyrin with different agents, which are expected to impart different degrees of hydrophilicity to the final products (Scheme 1). The 5,15-bis(1,3-dimethylimidazol-2-yl)porphyrinate diiodide (IP-H-Me2+) and 5,15-bis(1-methyl-3-(3,3,3-trifluoropropyl)imidazol-2-yl)porphyrinate diiodide (IP-H-CF32+) were synthesized using our previously reported methodologies, being yields and characterization data in good agreement [8, 26, 43]. The synthesis of the new 5,15-bis(1-methyl-3-(3-hydroxypropyl)imidazol-2-yl)porphyrinate diiodide (IP-H-OH2+) was carried out in DMF, under microwave irradiation (Pmax = 125 W), at 100 ºC for 10 min, using iodopropanol as alkylating agent. The desired hydroxypropyl imidazolyl cationic porphyrin IP-H-OH2+ was precipitated from the crude with 2-methyltetrahydrofurane and was obtained in 60% isolated yield. It should be noted that the use of microwave irradiation to promote these alkylation reactions proved to be a better approach than conventional heating, giving the desired cationic imidazolyl porphyrin in just 10 min, instead of 24 h [8].

Scheme 1
scheme 1

Synthesis and structures of dicationic meso-imidazolyl porphyrins

Full size image

2.2 Photophysical and photochemical studies

The most relevant photophysical and photochemical properties of IP-H-OH2+ are summarized in Table 1 and compared with those of IP-H-Me2+ and IP-H-CF32+, previously published [43]. Electronic absorption spectrum of IP-H-OH2+ in ethanol is typical of this class of compounds, exhibiting a Soret band at 400 nm, a Qy(0,1) band (λ = 499 nm); a Qy(0,0) band (536 nm); a Qx(0,1) band (576 nm) and a Qx(0,0) band (627 nm). This is consistent with the spectra of IP-H-Me2+ and IP-H-CF32+ [43]. The molar absorption coefficient of the Soret band of IP-H-OH2+ (ε = 3.0 × 105 M−1 cm−1) and its fluorescence quantum yield (ϕF = 0.11) in ethanol are also comparable to the values of other cationic di-imidazolyl porphyrins. [43] The relatively low fluorescence quantum yields are consistent with those of other neutral and cationic 5,15-di-arylporphyrins [44].

Table 1 Summary of photosensitizers’ physicochemical properties
Full size table

The singlet oxygen quantum yield (ΦΔ) is often used to evaluate the quality of a PDI photosensitizer. We determined ΦΔ in ethanol and DMSO using time-dependent singlet oxygen phosphorescence emission at 1270 nm. ΦΔ values of dicationic meso-imidazolyl porphyrins range between 0.27 and 0.63 in ethanol and 0.39 and 0.57 in DMSO, suggesting these are promising photosensitizers. These ranges agree with those of previously reported di-cationic porphyrins photosensitizers, such as 5,15-di[4-(3-N,N,N-trimethylammoniumpropoxy)phenyl]-10,20-di(4-trifluoromethylphenyl) porphyrin iodide (ΦΔ = 0.53 ± 0.05, in DMF) [45] and 5,15-di(1-methylpyridinium-4-yl)-10,20-diphenylporphyrin iodide (ΦΔ = 0.39 in PBS) [46].

A parameter equally important in the characterization of a photosensitizer is its photostability [47, 48]. We determined the photodecomposition quantum yields (Φpd) of our dicationic meso-imidazolyl porphyrins (electronic absorption spectra in Figures S4-S6) and found that all these photosensitizers are very photostable, being IP-H-OH2+ the most promising (Table 1).

The n-octanol:water partition coefficient POW, or its logarithmic expression (logP), allows for a clear distinction between these photosensitizers. The propyl chain with a terminal CF3 group appreciably increases the lipophilicity of IP-H-CF32+ (logP = 0.99) with respect to that of IP-H-Me2+ (logP = –1.35). This value is comparable to that of dicationic 5,15-diphenyl-10,20-di(1-methylpyridinium-4-yl)porphyrin diiodide (logP = 0.87), which possesses two N-methylpyridinium groups instead of the N-methylimidazolium groups, and two additional phenyl groups in the other meso positions [49]. Although IP-H-OH2+ has a hydrophilic functional group (OH) at the end of the alkyl chain, the longer size of this chain compensates for the functional group and IP-H-OH2+ has an intermediate value of logP (logP = –0.50). This is in line with that of dicationic 5,15-di(N-(4-methoxybenzyl)-4-pyridyl)-10,20-di(4-pyridyl)porphyrin dichloride (logP = –0.52) [18]. Overall, the synthesized di-imidazolyl porphyrins allow us to investigate a series of relatively small-size dicationic photosensitizers that differ mostly in their n-octanol:water partition coefficients, to fine-tune the role of this factor in PDI efficacy.

2.3 Planktonic photodynamic inactivation

PDI of E. coli ATCC 25922 with dicationic meso-imidazolyl porphyrins was evaluated with a light dose of 1.36 J/cm2 at 415 nm (Fig. 1). Remarkably, all these porphyrins enabled full inactivation of this Gram-negative bacterium at a concentration of 1 µM. Lowering the concentrations to 100 nM revealed differences between the photosensitizers. IP-H-OH2+ allowed for a 3 logs CFU reduction, IP-H-Me2+ for 2 logs and IP-H-CF32+ for 1 log reduction. This is consistent with our earlier work with IP-H-Me2+ [8]. The longer alkyl side chain ending in a haloalkyl functional group (CF3) did not improve the photoinactivation of E. coli but the small alkyl chain with a more hydrophilic group (OH) did increase the photoinactivation of E. coli. The potency of IP-H-Me2+ at 100 nM and 1 µM with 1.36 J/cm2 is comparable to that of IP-Zn-Me2+ at the same concentrations and 2 J/cm2 [8]. Interestingly, Orlandi et al. demonstrated that 5,15-di(N-methyl-4-pyridyl)porphyrin has low E. coli inactivation in the µM range, which highlights the relevance of the type of cationic N-heterocyclic moiety in the overall photodynamic inactivation efficiency [20]. The activity of these pyridinium-derived porphyrins was improved with the introduction of more lipophilic groups in the porphyrin structure, such as benzyl [20] and phenoxyalkyl groups [13]. These results are in line with our findings, since the modulation of the lipophilicity of the dicationic meso-imidazolyl porphyrins described here also impacted their efficiency towards E. coli inactivation.

Fig. 1
figure 1

Comparative E. coli photoinactivation studies using the meso-imidazolyl porphyrins and a 1.36 J/cm2 light dose and 1 h incubation time. The labels * (p < 0.05), ** (p < 0.01) and *** (p < 0.001) represent statistical difference

Full size image

The positive charges of these imidazolyl porphyrins favor electrostatic interactions with the membranes of Gram-negative bacteria. When the interactions at the surface of the bacteria determine the efficacy of the treatment, triple washing with water is expected to reduce treatment efficacy. This washing was done to evaluate photoinactivation by photosensitizer molecules that diffused into the bacteria. Figure 2 shows that incubation with 100 nM concentrations followed by washing reduced photoactivation with IP-H-Me2+ and IP-H-CF32+ to non-significant levels, but ca. 3 log CFU reductions were achieved in incubations with 500 nM concentrations, and nearly 7 logs of inactivation were achieved at 1 µM. The observed decrease in photoinactivation efficiency after consecutive washing steps is in agreement with previous reports using other cationic porphyrins [45, 46]. Overall, this protocol confirms that IP-H-OH2+ is more potent at lower concentrations than the other two imidazolyl porphyrins, and suggests that it more rapidly partitions into inner regions of the bacteria. Moreover, the stronger photoinactivation before washings suggests that ROS generated in PDI with the photosensitizers on the surface of the bacteria can produce sufficient damage to inactivate bacteria.

Fig. 2
figure 2

Comparative E. coli photoinactivation studies using the synthesized meso-imidazolyl porphyrins and a 1.36 J/cm2 light dose, after removing unbound PS with 3 × washings. The labels * (p < 0.05), ** (p < 0.01) and *** (p < 0.001) represent statistical difference relative to the control group

Full size image

2.4 Photosensitizer uptake

The uptake of photosensitizers by bacteria was studied with stock solutions of IP-H-OH2+ and IP-H-Me2+ in PBS:DMSO (DMSO 0.5%) and IP-H-CF32+ in PBS:THF (THF 0.5%). Porphyrin accumulation in planktonic bacteria was followed as a function of time, from 10 to 120 min of incubation, for 10 μM solutions, based on the fluorescence of porphyrins after their lyse with SDS 10% (Fig. 3). IP-H-OH2+ accumulation increases rapidly in the first 30 min and then stabilizes at ca. 2.4 × 105 molecules per cell. The uptake of IP-H-Me2+ is slower, reaching 1.9 × 105 molecules per cell after 2 h of incubation and may still increase further at longer times. Interestingly, IP-H-CF32+ uptake was much smaller. The difference in the stock solutions solvents may interfere with the uptake, but the most likely reason for the low uptake of IP-H-CF32+ is the unfavorable interaction between this lipophilic photosensitizer (logP = 0.99) and the negatively charged lipid A residues of the outer membrane of E. coli [3].

Fig. 3
figure 3

Time-dependent cellular uptake of investigated porphyrins IP-H-OH2+, IP-H-Me2+ and IP-H-CF32+ (with the initial 10 μM concentration) in E. coli

Full size image

The differences in cell uptake were corroborated by flow cytometry and light scattering confocal microscopy (LSCM) studies, Figs. 4 and 5, respectively. The analysis of the histograms of the fluorescence intensities of the control bacteria (blue population), and bacteria with accumulated photosensitizer (red population) clearly demonstrates that IP-H-Me2+ and, particularly, IP-H-OH2+ have the highest cellular uptake. In addition, the LSCM representative pictures obtained for control cells (untreated bacteria, stained only with Hoechst33342) and each porphyrin also confirms the higher accumulation of IP-H-Me2+ and IP-H-OH2+ than IP-H-CF32+.

Fig. 4
figure 4

Cellular uptake was determined in E. coli based on the red fluorescence of the selected porphyrin (IP-H-CF32+, IP-H-Me2+, IP-H-OH2+) using flow cytometry equipped in red laser. Control bacteria are shown in blue

Full size image
Fig. 5
figure 5

Laser scanning confocal microscopy imaging of cellular uptake of IP-H-OH2+, IP-H-Me2+ and IP-H-CF32+ in E. coli after 2 h of incubation: (A) with magnification × 100 (scale bar: 5 µm) and (B) confocal zoom (scale bar: 2 µm). The live cells were stained with Hoechst33342 (blue fluorescence); the accumulated porphyrin is indicated in red fluorescence

Full size image

The higher uptake of IP-H-OH2+ with 1 h of incubation is consistent with its enhanced PDI efficacy at low concentrations presented in Figs. 1 and 2. IP-H-Me2+ closely follows IP-H-OH2+ in cell uptake and PDI efficacy, and IP-H-CF32+ is the least successful photosensitizer in this family. It is possible that the performance of IP-H-CF32+ is limited by its lipophilicity. A moderate, but not excessive, hydrophilicity seems to be required for higher efficacy of photoinactivation of E. coli bacteria. IP-H-OH2+ combines both hydrogen-bond donors (OH) and positive charges in the photosensitizer structure.

2.5 PDI combined with cinnamaldehyde

Cinnamaldehyde is a natural antimicrobial agent whose mechanism of action is attributed to the disruption of bacterial membrane integrity [40]. Taking advantage of this membrane targeting mechanism and of the extensive uptake of membranes by IP-H-OH2+ and IP-H-Me2+, we hypothesized that the administration of cinnamaldehyde after PDI may lead to a significant synergic effect. Thus, E. coli was incubated for 1 h with 50 nM of each photosensitizer concentration, in aqueous solution, and then irradiated with a 1.36 J/cm2 light dose. Afterwards, cinnamaldehyde from a DMSO stock solution was added to the planktonic bacteria (7.5 mM final concentration, with 5% DMSO), followed by incubation in the dark for one additional hour, at room temperature. This concentration of cinnamaldehyde corresponds to ~ 0.1% of a topical formulation and is below the NOAEL for cinnamaldehyde sensitization. The results of the combination of IP-H-Me2+ and IP-H-OH2+ with cinnamaldehyde are presented in Fig. 6.

Fig. 6
figure 6

Combination of 7.5 mM cinnamaldehyde (Cin) with PDI of E. coli using IP-H-Me2+ and IP-H-OH2+ as photosensitizers (50 nM; 1.36 J/cm2 light dose). The labels ** (p < 0.01) and *** (p < 0.001) represent statistical difference relative to the control group

Full size image

PDI with IP-H-OH2+ at a concentration of 50 nM gives a modest 1.5 log CFU reduction, and with IP-H-Me2+ the change in CFUs is not statistically significant. Interestingly, cinnamaldehyde alone at 7.5 mM also gives a 1.5 log CFU reduction. The antimicrobial activity of cinnamaldehyde is clearly evident with just 1 h of incubation at this non-toxic concentration. The combination of PDI with IP-H-Me2+ followed by cinnamaldehyde administration gives a robust 3 log CFU reduction, while using IP-H-OH2+ gives a remarkable 7 log CFU reduction. This is clear evidence of a synergy between the loss of integrity of the bacterial cell membrane due to cinnamaldehyde and the facile uptake of IP-H-OH2+ by the cell membrane.

The sequence of the treatments, first PDI and then antimicrobial treatment with cinnamaldehyde, is in agreement with our previous experience of combining ciprofloxacin with PDI [26], which may be explained by the easy oxidation of cinnamaldehyde by the burst of oxidative damage produced by PDI [50, 51].

Cinnamaldehyde has antimicrobial activity but it is not an antibiotic, and it is not associated with drug resistance. The remarkable synergy between PDI with cinnamaldehyde treatments in the inactivation of E. coli, herein observed for the first time, is a notable example of how two treatments that do not contribute to bacterial resistance can be combined to inactivate bacteria at lower doses. The success of this combination may be attributed to the membrane damage caused by ROS produced in PDI [52], which potentiate the permeabilization of the bacteria membrane by cinnamaldehyde [53].

2.6 PDI in vivo

A preliminary in vivo study to assess the feasibility of treating local infections with PDI using dicationic 5,15-di-imidazolyl porphyrins was conducted in Balb C mice with infected wounds. The backs of the mice were infected with a GFP-fluorescent E. coli to allow for quantitative and non-invasive follow-up of the infection (Fig. 7). After a single PDI treatment with IP-H-Me2+ at 25 µM and 120 J/cm2 (420 nm light) light dose, a significant reduction of infection was observed over the 7 days follow-up after PDI. This is in stark contrast with the untreated control group, where a progressive increase of the infection area was observed over the same period. In vivo treatments require significantly higher photosensitizer concentration and light dose (~ 100-fold) than in vitro studies. This is expected because: (i) the interaction of photosensitizer molecules with the host cells reduces their accumulation in bacteria; (ii) the quenching of ROS by the host biomolecules reduces oxidative stress in the bacteria; (iii) absorption and scattering light by tissues lowers the light dose [10, 29,30,31,32]. In a PDI study with a comparable model of burn infection, treatment with 50 µM methylene blue and a light dose of 150 J/cm2 (660 nm light) did not significantly reduce infection with respect to control [54]. That study showed that PDI with methylene blue was remarkably potentiated by the addition of 10 mM of KI. We obtained a reduction of the infection with IP-H-Me2+–PDI alone. We expect that further studies will show that cinnamaldehyde can offer a potentiation comparable to that of KI, although using different mechanisms, and contribute to make PDI a more valuable alternative to treat superficial infections without aggravating drug resistance.

Fig. 7
figure 7

Photographs of mice with excision wounds infected with E. coli. Panel of representative fluorescent images showing response to PDI on different days post-treatment. PDI was performed using 25 µM of IP-H-Me2+ and 120 J/cm2 blue light (420 nm)

Full size image

3 Conclusion

We synthesized a new dicationic 5,15-di-imidazolyl porphyrin (IP-H-OH2+), in just 10 min reaction time, using microwave irradiation, with 60% yield, to expand the family of cationic imidazolyl photosensitizers and investigate structure–activity relations in PDI of E. coli. IP-H-OH2+ (logP = – 0.50) has half the singlet oxygen quantum yield and twice the photostability of IP-H-Me2+ (logP = –1.35). IP-H-CF32+ (logP = 0.99) has an intermediate ΦΔ and a stability similar to IP-H-Me2+. Considering that lower ΦΔ and high photostability tend to compensate each other, these photosensitizer differ mostly in their octanol:water partition coefficients. All of them reduced planktonic E. coli CFUs by ~ 7 log units at 1 µM concentrations with 1.36 J/cm2 (415 nm light), which reveals that this class of photosensitizers is remarkably efficient in PDI of planktonic bacteria. At 100 nM, IP-H-OH2+ became the most potent photosensitizer, followed by IP-H-Me2+. IP-H-OH2+ has a faster uptake in E. coli than IP-H-Me2+, and IP-H-CF32+ showed the lowest uptake. The high affinity of IP-H-OH2+ for the bacterial membrane is likely the reason for its enhanced photodynamic efficacy.

The strong interaction of IP-H-OH2+ with bacterial membranes is consistent with the success found in its combination with cinnamaldehyde to inactivate E. coli. This natural antimicrobial agent is known to disrupt the bacterial membrane, and the addition of 7.5 mM of cinnamaldehyde to E. coli shortly after PDI with 100 nM IP-H-OH2+ afforded a dramatic photoinactivation of 7 log CFUs. Cinnamaldehyde is widely used in the food industry and as a fragrance ingredient. Its combination with photosensitizers to treat superficial infections has a very favorable safety profile and can enhance PDI efficacy.

Preliminary in vivo studies of PDI with 25 µM IP-H-Me2+ showed the feasibility of using this class of molecules to treat infected wounds when a light dose of 120 J/cm2 is delivered 30 min after the addition of the photosensitizer to the wound. Light doses of this magnitude can be delivered in 12 min without thermal effects. These treatments can be repeated without generating drug resistance. Translation to clinical applications will benefit from more efficient treatments performed in shorter times. Combination therapies, namely with cinnamaldehyde, may contribute to shorten the treatment time and increase the efficacy of PDI. We hope that topical formulations with potent photosensitizers like IP-H-OH2+ combined with antimicrobial agents will be developed to change the management of localized, superficial infections and minimize the propagation of multidrug-resistant bacteria.

4 Experimental procedures

4.1 Synthesis

All reagents and solvents were purchased from commercial sources and used without further purification. 5,15-bis(1-methylimidazol-2-yl)porphyrin was acquired from PorphyChem. 5,15-bis(1,3-dimethylimidazol-2-yl)porphyrinate diiodide (IP-H-Me2+) and 5,15-bis(1-methyl-3-(3,3,3-trifluoropropyl)imidazol-2-yl)porphyrinate diiodide (IP-H-CF32+) were synthesized as previously reported [43]. Electronic absorption spectra were recorded on a Shimadzu 2100 spectrophotometer. Nuclear magnetic resonance 1H spectra were recorded on a 400 Bruker Avance spectrometer (400 MHz), using tetramethylsilane (δ = 0.00 ppm) as internal standard. The high-resolution mass spectrometry (HRMS) analysis was carried out using a Bruker Microtof.

4.1.1 5,15-bis(1-methyl-3-(3-hydroxypropyl)imidazol-2-yl)porphyrinate diiodide (IP-H-OH2+)

In a 5 mL microwave vial, the 5,15-bis(1-methylimidazol-2-yl)porphyrin (30 mg, 0.064 mmol) was dissolved in DMF (0.2 mL) and 3-iodopropanol (0.2 mL, 0.32 mmol) were added. The reaction was carried out under microwave irradiation with Pmax = 125 W, at 100 ºC for 10 min. After reaction completion, the solvent was evaporated and the crude was redissolved in a minimum amount of methanol. The product was precipitated following the addition of 2-methyltetrahydrofurane to the methanolic solution, affording a dark purple solid after filtration (23 mg, 60% yield). 1H NMR (400 MHz, CD3OD) δ, ppm: mixture of atropisomers 10.89 (s, 2H), 9.86 (br s, 4H), 9.18 (br s, 4H), 8.48 (s, 2H), 8.41 (s, 2H), 4.25–4.23 (m, 4H), 3.86–3.84 (m, 6H), 3.69–3.66 (m, 4H), 1.82–1.81 (m, 4H); ESI–MS m/z: obtained 587.2858 [M-2I-H]+; calculated for [C34H35N8O2].+ 587.2877. See characterization spectra in Figs. S1 and S2.

4.2 Photophysical and photochemical studies

4.2.1 Fluorescence quantum yield

Fluorescence spectra were obtained in obtained in a Horiba Jobin Yvon—Spex Fluorolog 3.22. The determination of the fluorescence quantum yields (ΦF) was conducted in ethanol using a comparative method, according to Eq. (1):

$${\Phi }_{{\text{F}}}= {{\Phi }_{{\text{F}}}}^{{\text{Std}}}\frac{\mathrm{F }{{\text{A}}}_{{\text{Std}}} {\upeta }^{2}}{{{\text{F}}}_{{\text{Std}}}\mathrm{ A }{{\upeta }_{{\text{Std}}}}^{2}}$$
(1)

where F and FStd are the integrals of the fluorescence emission curves of the samples and the standard, respectively; A and AStd are the respective absorbances of the samples and standard at the excitation wavelengths, respectively; η2 and η2Std are the refractive indices of solvents used for the sample and standard, respectively. 5,10,15,20-tetraphenylporphyrin (TPP) in toluene (ΦF = 0.11) was used as reference [55].

4.2.2 Singlet oxygen quantum yield

Singlet oxygen formation quantum yields (ΦΔ) were determined in ethanol and DMSO by direct measurement of singlet oxygen phosphorescence at 1270 nm. The photosensitizer excitation was carried out using the third harmonic of a Spectra-Physics Quanta-Ray GCR-130 Nd–YAG laser (355 nm). Then, the singlet oxygen phosphorescence was collected at 1270 nm using a Hamamatsu R5509-42 photomultiplier cooled to 193 K in a liquid nitrogen chamber. First-degree exponential decay curves of the singlet molecular oxygen emission were extrapolated to time-zero for the reference (phenalenone, ΦΔStd = 0.97) [56] and for the samples to obtain a linear relationship between emission intensities and a given laser power. The actual singlet oxygen quantum yields were obtained by comparing this linear dependence between singlet oxygen phosphorescence emission and the energy of the laser pulse for the sample (S) and the reference (SStd) using (2):

$${\Phi }_{\Delta }={\Phi }_{\Delta }^{{\text{Std}}} \frac{\mathrm{S }}{1-{10}^{-{\text{A}}}}*\frac{1-{10}^{-{{\text{A}}}_{{\text{Std}}}} }{{{\text{S}}}_{{\text{Std}}}}$$
(2)

where A and AStd are the absorbance of the sample and the reference, respectively.

4.2.3 Photodegradation studies

The photodegradation studies were conducted in a photosensitizer PBS solution starting with a 0.8–1.0 absorbance in the Soret band. Then, the cuvette was irradiated with a 415 nm light, with light spot with 1 cm diameter and a total power of 0.20 W. A magnetic stirring bar was placed in the cuvette for solution homogenization during irradiation. After each irradiation period, the cuvette was weighted to determine solvent loss due to evaporation. The UV–Visible absorption spectra were recorded before and after each irradiation. The photodegradation quantum yield was calculated following the equation:

$${\Phi }_{pd}= \frac{{V}_{d}}{{V}_{p}}=0.11978\times \frac{V{A}_{0}k}{\varepsilon \lambda P\left(1-{10}^{-{A}_{i}}\right)}$$

were V is the volume of initial solution, A0 and Ai refer to the initial absorption values ​​at the peak corresponding to the molar absorption coefficient (ε) and the absorption value corresponding to the wavelength of the irradiation light, respectively. P refers to the power of the light [47].

4.2.4 Partition coefficient

The PBS/n-octanol partition coefficients (logP) were measured following standard methodologies [57]. Approximately, 1 mg of each photosensitizer were first solubilized in 1–2 drops of DMSO and then 5 mL of n-octanol, previously saturated with a solution of PBS, and 5 mL of PBS, previously saturaterd with n-octanol, were added. Then, the solutions were vigorously shaked and mixed on a vortex and the phases were separated by centrifugation (5 min, 3700 rpm, RT). Next, aliquots from each of the PBS/n-octanol phase were taken and diluted to obtain a value of absorption corresponding to Soret band less than 1. The calculation of the partition coefficient was obtained using the following equation:

$$logP =log\left(\frac{{Abs}_{n-octanol}}{{\varepsilon }_{n-octanol}}\right)/\left(\frac{{Abs}_{PBS}}{{\varepsilon }_{PBS}}\right)$$

4.3 Microbiological studies

4.3.1 Bacterial strains and culture conditions

In this study, we selected a standard E. coli strain from American Type Culture Collection (ATCC 25922), that was cultured in Mueller–Hinton broth/agar (Sigma Aldrich).

4.3.2 Light source and illumination conditions

For the photoinactivation studies, a blue LED light (λ = 415 nm) was used, with a fluence of 6 mW/cm2. The light dose reported in each experiment represents the actual light dose absorbed by each compound, corrected by LED light emission overlap with compound absorption in water (see Figure S3) [58], using the following multiplicative factors: IP-H-Me2+  = 0.29; IP-H-CF32+  = 0.43; IP-H-OH2+  = 0.24.

4.3.3 Photoinactivation studies

Stock solutions of each photosensitizer were prepared in DMSO at 1–2 mM concentrations and further diluted with doubly distilled water (ddH2O) to the desired concentration. The planktonic E. coli was cultured in Mueller–Hinton agar (MH) at 37 °C overnight. Cell density was adjusted to the turbidity of the 0.5 McFarland standard in sterile ddH2O, which corresponds to approximately 1–2 × 108 CFU/mL. In 96-well plates, 10 μL of the inoculum were added to a solution of 100 μL of photosensitizer (IP-H-Me2+, IP-H-CF32+ or IP-H-OH2+) in ddH2O at the desired concentrations (100 nM and 1 µM), for a final inoculum size of 1–2 × 107 CFU/mL. The plates were incubated in the dark, at room temperature, for 1 h. Following incubation, the wells were irradiated for a total light dose of 1.36 J/cm2 light dose. Dark controls were covered in aluminum foil during the time of irradiation. After irradiation, 10 μL aliquots were taken from each well and dilutions in water were done as needed and plated in Petri dishes with MH agar. After 37 °C incubation during 18–24 h, the colony-forming units (CFU) were counted. Experiments were performed in triplicate. In photoinactivation studies where unbound PS was washed, after bacterial dark incubation with the PS, the bacteria suspensions were centrifuged (5000 rpm, 5 min). The supernatant was removed, the bacteria were re-suspended in ddH2O and the process was repeated two additional times. Then, irradiation was performed under the aforementioned conditions. Data analysis was carried out in GraphPad Prism 8.0 (GraphPad Software, San Diego, California USA).

4.3.4 PS uptake in bacterial cells

Stock solutions of IP-H-OH2+ and IP-H-Me2+ were prepared in PBS:DMSO (DMSO 0.5%) and IP-H-CF32+ in PBS:THF (THF 0.5%). The microorganisms were incubated in suspension with a photosensitizer (10 µM) for selected time intervals (0–120 min) in the dark at room temperature. The unbound photosensitizer was removed by washing twice in PBS without Ca2+ and Mg2+. After the second wash, bacteria were lysed in 10% SDS for 24 h. The cellular uptake of the photosensitizer was evaluated by determination of fluorescence using excitation at the Soret band and emission between 630–730 nm (Tecan Infinite M200 Reader). Calibration curves were prepared in 10% SDS and used for the determination of PS concentration. Uptake values were obtained by dividing the PS amount (nmol) by the number of CFU.

4.3.5 Flow cytometry studies

Cellular uptake of the investigated photosensitizers was also determined using flow cytometry and quantified based on the red fluorescence of porphyrin. For this analysis, E. coli bacteria were incubated with each porphyrin at 10 µM for 2 h. After this incubation, cells were washed two times with PBS without calcium and magnesium ions and collected for analysis. Then, the cells were collected by centrifugation and then resuspended in 200 µL of PBS. Bacteria were then examined using BD Accuri c6 flow cytometer equipped with a red laser (640 nm).

4.3.6 Confocal microscopy studies

Accumulation of the selected porphyrin derivatives in the microorganisms was followed with confocal imaging using a Zeiss LSM880 laser-scanning microscope equipped with an argon-ion laser and the 100 × objective with oil immersion (Carl Zeiss Ltd., Jena, Germany) with a working distance of 1.46 mm. Accordingly, for the uptake studies, bacteria were incubated with the photosensitizer solution (10 μM) for an appropriate time interval determined in the uptake studies (2 h). After washing, the bacteria samples were counterstained with Hoechst33342 (10 µg/mL) for 10 min, placed on the microscopic glass slides, and imaged. Registered images were analyzed with the Zeiss ZEN software.

4.3.7 Combination with cinnamaldehyde

Following the same protocol described in the photoinactivation studies, the bacteria were incubated in the dark for 1 h with 50 nM photosensitizer concentrations an aqueous solution, and then irradiated with 1.36 J/cm2 at 415 nm. Next, 5 μL of cinnamaldehyde from a DMSO stock solution (150 mM) were added to the bacteria, yielding a 7.5 mM cinnamaldehyde concentration (half of the determined MIC for this strain) in a 5% DMSO aqueous solution. In the control experiment, 5 μL of DMSO (without cinnamaldehyde) were added. The bacteria were incubated for one additional hour in the dark at room temperature, and then 10 μL aliquots were taken from each well and dilutions in water were done as needed and plated in Petri dishes with MH agar. After 37 ºC incubation for 24 h, CFUs were counted. For comparison purposes, monotherapy control experiments were conducted: (i) in PDI monotherapy, 5 μL DMSO (without cinnamaldehyde) was added after treatment and then 10 μL aliquots were plated after 1 h dark incubation; (ii) in cinnamaldehyde monotherapy, bacteria were first irradiated without photosensitizer, then cinnamaldehyde was added and finally 10 μL aliquots were plated after 1 h dark incubation. Experiments were performed in triplicate. Data analysis was carried out in GraphPad Prism 8.0 (GraphPad Software, San Diego, California USA).

4.3.8 In vivo studies

The bacteria were grown in liquid LB medium with shaking at 120 rpm at 37 °C overnight to reach the stationary phase. One mL of this suspension was then refreshed in fresh LB to mid-log phase. Cell numbers were estimated by measuring the OD at 600 nm. The bacterial suspension was spun down, washed, and resuspended in PBS for the experiments. BALB/c mice, male 8–10-week-old were used for in vivo infection studies. Prior to the experiment, mice were anesthetized by i.p. injection of ketamine/xylazine cocktail. They were shaved using a razor blade on the dorsal surface. A surgical scalpel was used to gently scrape the epidermis off an area of skin to create abrasion wounds. The depth of the wound was no more than the shallow dermis. After creating the wounds, a 50 μL aliquot of bacterial suspension containing 107 CFU of fluorescent E. coli in PBS was topically inoculated onto each defined area of the abrasion with a pipette tip. Fluorescence images were taken immediately after the inoculation of bacteria to ensure that the number of bacteria applied to each abrasion was consistent. The infection imaging was performed using the NEWTON 7.0 Animal Optical Imaging System (Vilber).

Mice with abrasion wounds infected with E. coli were randomly divided into two groups: (1) infection control group and (2) PDT group. The infected abrasion wounds were incubated for 60 min. In the treatment group, a 50 μL aliquot of porphyrin solution (25 μM in a PBS/5% DMSO mixture) was added to the wound and 30 min later the wounds were imaged to quantify any dark toxicity of the porphyrin to the bacteria. Next, the wounds were irradiated with a 420 ± 20 nm light and a total light dose of 120 J/cm2 (12 min of illumination). Sterile saline (0.5 mL intraperitoneally) was administered to support fluid balance during recovery. For both groups, fluorescence imaging was performed at different time points. To record the time course of the extent of bacterial infection, the bacterial bioluminescence from mouse wounds was measured for 7 days after the wounds until the infections were cured (characterized by the decrease of bacterial fluorescence).