Cytotoxicity of Cyclodipeptides from Pseudomonas aeruginosa PAO1 Leads to Apoptosis in Human Cancer Cell Lines

Pseudomonas aeruginosa is an opportunistic pathogen of plants and animals, which produces virulence factors in order to infect or colonize its eukaryotic hosts. Cyclodipeptides (CDPs) produced by P. aeruginosa exhibit cytotoxic properties toward human tumor cells. In this study, we evaluated the effect of a CDP mix, comprised of cyclo(L-Pro-L-Tyr), cyclo(L-Pro-L-Val), and cyclo(L-Pro-L-Phe) that were isolated from P. aeruginosa, on two human cancer cell lines. Our results demonstrated that the CDP mix promoted cell death in cultures of the HeLa cervical adenocarcinoma and Caco-2 colorectal adenocarcinoma cell lines in a dose-dependent manner, with a 50% inhibitory concentration (IC50) of 0.53 and 0.66 mg/mL, for HeLa and Caco-2 cells, respectively. Flow cytometric analysis, using annexin V and propidium iodide as apoptosis and necrosis indicators, respectively, clearly showed that HeLa and Caco-2 cells exhibited apoptotic characteristics when treated with the CDP mix at a concentration <0.001 mg/mL. IC50 values for apoptotic cells in HeLa and Caco-2 cells were 6.5 × 10−5 and 1.8 × 10−4 mg/mL, respectively. Our results indicate that an apoptotic pathway is involved in the inhibition of cell proliferation caused by the P. aeruginosa CDP mix.

signals. However, the mechanism of action and physiological relevance of CDPs are poorly understood [7,8].
DKPs are cyclized molecules comprising two amino acids bound by two peptide bonds; they are produced by a wide range of organisms, from bacteria to fungi and animals (Figure 1(a)) [9,10]. DKPs belong to the nonribosomal peptides that are synthesized in microorganisms by a multifunctional assembly of enzymes known as nonribosomal peptide synthases [10] and by CDP synthases, another kind of enzymes that utilizes aminoacylated transfer RNAs as substrates instead of free amino acids [11].
CDPs are structurally diverse, and they have been implicated in multiple functions; the CDPs cyclo(D-Ala-L-Val) and cyclo(L-Pro-L-Tyr) have been identified as a new class of QS autoinducers in Pseudomonas strains, based on their ability to activate AHL-dependent biosensors [12][13][14]. The CDP cyclo(L-Phe-L-Pro) isolated from Lactobacillus plantarum exhibited an antifungal effect against Fusarium sporotrichioides and Aspergillus fumigatus [15], while the CDPs cyclo(L-Leu-L-Pro), cyclo(L-Phe-L-Pro), cyclo(L-Val-L-Pro), cyclo(L-Trp-L-Pro), and cyclo(L-Leu-L-Val) isolated from the deep-sea bacterium Streptomyces fungicidicus showed antifouling effects [16]. Moreover, synthetic CDPs such as cyclo(Phe-Pro) induced apoptosis in the HT-29 colon cancer cell line [17], and cyclo(L-Cys-L-Leu) exhibited potential for scavenging free radicals [18]. Recently, it was reported that P. aeruginosa is capable of interacting with the plant Arabidopsis thaliana via the secretion of CDPs such as cyclo(L-Pro-L-Tyr), cyclo(L-Pro-L-Val), and cyclo(L-Pro-L-Phe), appearing to mimic the biological role of auxin, a natural plant hormone [12] (Figure 1(b)). In Staphylococcus aureus, the aureusimines A/B, comprised of the CDP cyclo(L-Val-L-Tyr) and cyclo(L-Val-L-Phe), respectively, are involved in the regulation of bacterial virulence factors in a murine host [19]; similarly, the CDP cyclo(L-Phe-L-Pro) in Vibrio cholerae, V. parahaemolyticus, and V. harveyi is involved in controlling the expression of genes that are important in pathogenicity [20]. Moreover, it was reported that CDPs and DKPs may induce cell death in several cancer cell lines [21], by affecting biological processes such as microtubule polymerization; for example, cyclo(D-Tyr-D-Phe), isolated from Bacillus species, induced apoptosis via caspase-3 activation in the A549 pulmonary adenocarcinoma cell line [22]. In addition, it was reported that the CDPs cyclo(L-Leu-L-Pro) and cis-cyclo(L-Phe-L-Pro) isolated from Lactobacillus exhibited antiviral activity against the influenza A (H3N2) virus [23].
Although, in the context of bacteria-mammalian interaction, it has been suggested that CDPs could play an important role in bacterial pathogenesis, bacteria-host signaling, or mammalian cell growth, the mechanisms involved are unknown. Therefore, in this study, we focused on investigating the cellular effect of CDPs produced from P. aeruginosa strain PAO1, a pathogenic bacterium in humans that is capable of secreting the CDPs, cyclo(L-Pro-L-Tyr), cyclo(L-Pro-L-Val), and cyclo(L-Pro-L-Phe) into the culture medium (Figure 1(b)). The biological effects of these CDPs on the growth and/or pathogenesis of mammalian cells remain unknown; the P. aeruginosa CDPs could be involved in bacterial host colonization phenomena during disease episodes, where antiproliferative or anti-immune properties of these compounds could affect the host organism. In this regard, we employed the HeLa cervical adenocarcinoma and Caco-2 colorectal adenocarcinoma cell lines as host models in this study.
The P. aeruginosa CDP mix was characterized as described previously [12]. Briefly, the P. aeruginosa WT strain was placed in 100 mL of Luria Bertani (LB) medium and incubated for 24 h at 30 ∘ C for bacterial growth. Cell-free supernatants were prepared by centrifugation (10,000 ×g, 25 ∘ C for 10 min). The resulting supernatant was extracted twice with ethyl acetate supplied with acetic acid (0.1 mL/L). The extracts were evaporated to dryness using a rotavapor at 60 ∘ C (Buchi Co., Lawil, Switzerland). The residue was solubilized in methanol : acetonitrile (1 : 1) and analyzed by GC-MS as described [12]. The CDP mix is constituted by the cyclo(L-Pro-L-Tyr), cyclo(L-Pro-L-Val), and cyclo(L-Pro-L-Phe) in a 1 : 1 : 1 molar ratio. For dose-response assays, the CDP mix was evaporated to dryness, weighed out, and dissolved with DMSO to prepare a 100 mg/mL concentration as stock solution.

Cell Line Growth.
The human cancer cell lines HeLa and Caco-2 were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cell procedures were performed under class II biological safety cabinets. Cells were cultured in DMEM supplemented with 10% (v/v) FBS (complete medium) and 1% antibiotic (10,000 units of penicillin, 10 mg streptomycin, and 25 g of amphotericin B per mL) solution. The cultures were fed twice a week and maintained at 37 ∘ C under 80% humidity and incubated in an atmosphere of 5% CO 2 . HeLa and Caco-2 cells were collected by trypsinization using trypsin/EDTA buffered solution for 5 min at room temperature, followed by the addition of 5 mL of serum-enriched medium (CM) to stop trypsin action. After trypsinization the cells were collected and washed with CM. Finally, cells were counted in a hemocytometer chamber and incubated in fresh CM media.

Cell Viability Assay.
Cell viability was determined by the MTT colorimetric method using thiazolyl blue tetrazolium bromide (Sigma-Aldrich Co). Briefly, HeLa and Caco-2 cells were seeded in 96-well flat-bottomed plates at a density of

CDPs from P. aeruginosa Cultures Affect the Viability of Human Cancer Cell Lines.
In order to test the effect of CDPs from P. aeruginosa on mammalian cell growth, we used the HeLa and the Caco-2 cell lines as models in this study. The HeLa cell line has been extensively employed to test anticancer drugs [24], while the Caco-2 cell line has been used to evaluate the ability of chemicals to cross the intestinal barrier and to study their transport mechanisms [25]. A mixture of CDPs, mainly comprised of cyclo(L-Pro-L-Tyr), cyclo(L-Pro-L-Val), and cyclo(L-Pro-L-Phe) in a 1 : 1 : 1 molar ratio, was isolated from the P. aeruginosa PAO1 strain, grown on Luria Bertani broth. The CDP mix was applied in a dose-dependent manner to human cells grown in CM medium. The results obtained showed that CDPs caused a decrease in the viability of HeLa and Caco-2 cells in a dose-dependent manner, cell cultures exhibiting 75% dead cells following treatment with the CDP mix at 100 mg/mL (Figure 2(a)). The 50% inhibitory concentration (IC 50 ) for the CDP mix from the PAO1 strain was 0.53 and 0.66 mg/mL, for HeLa and Caco-2 cells, respectively (Figure 2(b)). Although CDPs incubated in serum-free (SS) medium showed slight differences in activity compared to those incubated in serumcontaining (CM) medium, these differences were not significant (see Figure S1 in Supplementary Material available online at http://dx.doi.org/10.1155/2015/197608). These findings indicate that the CDP mix from P. aeruginosa inhibited the viability of HeLa and Caco-2 cells and that this effect was independent of the presence or absence of serum. Microscopic observation of cells following treatment with the CDP mix and staining with DAPI showed that while HeLa cells treated with DMSO solvent alone did not exhibit nuclear DNA fragmentation (Figure 3(a)), HeLa cells did exhibit nuclear DNA fragmentation after treatment with the CDP mix at a concentration of 10 mg/mL for 24 h (Figure 3(b)); in addition, apoptotic bodies were clearly visible in cells treated with CDPs (Figure 3(c)). These results indicate that the CDP mix from P. aeruginosa produced a decrease in cell viability by means of a mechanism of DNA fragmentation.

Inhibition of HeLa and Caco-2 Cell Viability by CDPs from P. aeruginosa Involves an Apoptotic Pathway.
In order to identify the mechanism underlying the decrease in viability of HeLa and Caco-2 cells due to CDP treatment, flow cytometric analysis was carried out using markers for apoptosis (annexin V) or necrosis (propidium iodide; PI), with cultures of both cell lines, after their treatment with the CDP mix. The value for the percentage of fluorescent cells (PFC), corresponding to cells that were positive for the annexin V marker, was ≤14% for negative controls of both HeLa and Caco-2 cells (Figures 4(a) and 4(e)), but ≥97% for both cell lines when actinomycin D was used as an apoptosis inducer (Figures 4(b) and 4(f)). Importantly, HeLa and Caco-2 cells treated with the P. aeruginosa CDP mix at a concentration of 0.01 mg/mL showed PFC values ≥50% for the annexin V marker (Figures 4(c) and 4(g); lower-right quadrants), increasing to ≥90% when treated with 1 mg/mL and 0.1 mg/mL CDP mix (Figures 4(d) and 4(h); resp.); however, at higher concentrations of the CDP mix, the PFC values did not show a further significant increase (Figure 4(i)). With respect to PI staining, to identify necrosis in HeLa and Caco-2 cell cultures, positive cells were not detected under the same CDP-treatment conditions (Figures 4(c), 4(d), 4(g), and 4(h); upper-right quadrants). The IC 50 CDP-mix doses for apoptosis of HeLa and Caco-2 cells, after treatment for 24 h, were calculated as 6.5 × 10 −5 mg/mL and 1.8 × 10 −4 mg/mL, respectively (Figure 4(j)). These findings indicate that the CDP mix from P. aeruginosa caused inhibition of the viability of these two cancer cell lines via an apoptotic mechanism, in a dose-dependent manner. Interestingly, the CDP concentration for apoptosis induction in HeLa cells was 30-fold lower than for Caco-2 cells (Figure 4(j)). In addition, the CDP concentration for apoptosis induction was two log units lower than that for nonapoptotic cell death in both the cancer cell lines tested. These findings were confirmed by microscopic observation of cells; HeLa and Caco-2 cells exhibited apoptotic cell morphology following treatment with the CDP mix, at a concentration of 0.1 mg/mL ( Figure 5), and similar morphology was observed following treatment with actinomycin D (apoptosis control; Figures 5(b) and 5(h)). These results indicate that CDPs from P. aeruginosa caused a decrease in the cell viability of HeLa and Caco-2 cell lines by means of a mechanism that involves apoptotic pathways.
These two cell lines are the most commonly used for testing drugs, for which details of molecular mechanisms of transport and cell signaling, among other biological processes, have been described. It is noteworthy that similar compounds from other organisms induce apoptosis in different cell lines, for example, cyclo(L-Pro-L-Tyr) and cyclo(L-Pro-L-Phe) isolated from Bacillus species. Although these CDPs were not tested for their effects on mammalian cell viability, they provoked a slight decrement in the phosphorylation of the AKT1 serine/threonine kinase at a concentration of 0.01 mg/mL in U-87 MG cells from human glioblastoma, where AKT1 inactivation is an important event that leads to apoptosis [26,27]. The quest for novel molecules with properties involved in the control of cancer cell growth is a scientific field in growing demand [28]. Natural molecules with antiproliferative activity are considered more specific for their target than synthetic molecules; one of the probable reasons for this is that molecules produced from biological organisms such as P. aeruginosa tend to have more chiral