Over the past few decades, photodynamic therapy (PDT) has become widespread as an effective treatment modality of cancer [1]. PDT is now clinically approved and used to treat neoplastic and non-malignant diseases [2]. PDT employs two separately non-toxic components, which in combination and in the presence of oxygen, induce cell death [3]. The first component is a photosensitizer (PS), the second is light. When exposed to light of a specific wavelength, PS transfers light energy to molecular oxygen to generate reactive oxygen species (ROS). It is well documented that, among other ROS, singlet oxygen (1O2) is the most reactive and cytotoxic species generated by PSs [4]. Since singlet oxygen has a very short lifetime (~3 μs) and a limited diffusion distance (∼150 nm in water) [5], it appears preferentially in certain areas that have been exposed to light.

To date, many PSs based on low-molecular-weight dyes have been designed [6]. In clinical practice, tetrapyrrole PSs based on porphyrin and chlorine e6 are widely used. After administration, most PSs migrate to the diseased tissue by passive diffusion and build up in the tumor nonspecifically, accumulating in healthy tissues as well. As a result, and due to slow clearance, the use of synthetic PSs has undesirable side effects such as nonspecific photodynamic effect and long-term hypersensitivity of the skin to sunlight. To overcome these adverse effects, engineering of new PSs with the targeted delivery to cancer cells is currently underway [[7], [8], [9]].

One possible approach to creating targeted PSs is their conjugation with antibodies specific to antigens overexpressed in cancer cells [10]. However, an unstable expression of these antigens often causes significant difficulties. On the other hand, targeting cancer cell biomarkers often results in resistance due to clonal selection, which makes therapy less effective [11]. An alternative and more versatile approach of targeted PDT could be based on the differences in basic properties of normal and cancer cells. It has been previously shown that cancer cells are characterized by a reduced pH in the extracellular space (pH 6.0–6.7) [12]. Later on, a 36-residue polypeptide from bacteriorhodopsin (pHLIP) that spontaneously binds to the lipid bilayer at acidic pHs was identified [13]. In recent years, attempts have been made to use pHLIP-mediated targeting for cancer imaging, diagnosis and therapy [14]. In particular, the pH-dependent cell membrane targeting PS has been designed by solid phase pHLIP synthesis and its subsequent conjugation with the protoporphyrin IX [15].

In this study, we took advantage of miniSOG (singlet oxygen generating protein) as an effective light-driven 1O2 generator [16,17]. In our previous work we found that genetically engineered fluorescent protein -pHLIP construct demonstrates selective pH-dependent binding to cancer cells in in vitro experiments [18]. It has also been shown that the wild type pHLIP fused to the C-terminus of EGFP exhibit the highest binding efficiency with a pKa ~ 6.4. Taking these results into account, here we investigate the feasibility of utilizing fully genetically encoded miniSOG-pHLIP fusion construct for targeted PS delivery to cancer cells. An improved version of miniSOG, a W81L/H85N/M89I/Y98A/L103V mutant (SOPP3) was used, which shows the highest quantum yield of 1O2 production (ΦΔ ≈ 0.60) [19].