Introduction

The emergence of multidrug resistant bacterial infections is a major health threat [1]. Agencies like the World Health Organization and the Public Health Agency of Canada [2] have highlighted the need for investment in research and development of new antibiotics. A remarkable example is the Gram-positive bacterium Staphylococcus aureus, a ubiquitous commensal of the human skin. Naturally susceptible to most antibiotics, S. aureus is capable of causing infections in predisposed individuals and can acquire formidable antibiotic resistance [3]. One of the most notorious antibiotic-resistant members of the species is methicilin-resistant S. aureus (MRSA) [4], which are grouped into two categories referred as hospital-associated (HA-MRSA) and comunity-associated (CA-MRSA) [5]. MRSA is also a concerning pathogen for people with the genetic disease cystic fibrosis (CF) [6]. Studies show that lower airway inflamation in children with CF is associated with early colonization with S. aureus [7] and that the prevalence of MRSA in CF patients is high [8, 9].

Microorganisms such as bacteria and fungi produce chemical compounds that have been sucessfuly used as therapeutic antibiotics [10]. Lichenized fungi are a rich source of mostly under-explored natural products, many of which have chemical structures that suggest an origin from acetate via the polyketide pathway. One of the most widely occuring lichen natural products is the dibenzofuran usnic acid (UA), and the biological activity of this polyketide has been extensivly studied. The biosynthesis of UA has been demonstrated to proceed via the intermediate methylphloracetophenone (MPA [11], Fig. 1) which is produced from acetyl-CoA, malonyl-CoA, and S-adenosylmethionine (SAM) by a polyketide synthase (PKS). The intermediate MPA is in turn dimerized by an oxidative P450-type enzyme to produce UA. The biological activity of MPA vs. UA has recently been examined [12] and it was demonstrated that UA is more bioactive than MPA in a series of antibacterial and biofilm disruption assays. However, a systematic examination of the biological activity of other related phloracetophenone analogs has not yet been reported. We chose to explore this chemical space as a potential source of novel bioactivity. Althougth MPA is a known natural product none of the compounds reported here appear to be of natural origin.

Fig. 1
figure 1

The biosynthesis of usnic acid (UA) from acetyl-CoA, malonyl-CoA, and S-adenosylmethionine (SAM) has been demonstrated to proceed via methylphloracetophenone (MPA)

Full size image

In this study, we synthesized 12 novel phloracetophenone analogs (Compounds 1–12, Schemes 1–3, Fig. 2) and explored their biological activity against S. aureus ATCC 29213 and MRSA strains isolated from CF patients. Our results show that 4 of the 12 novel compounds displayed promising activity as a monotherapy and one strongly synergized with doxycycline. These results suggest that phloracetophenone analogs may serve as promising lead compounds for the design of novel antibiotics.

Scheme 1
scheme 1

The synthesis of compounds 1–4 is shown

Full size image
Scheme 2
scheme 2

The synthesis of compounds 5–8 is shown

Full size image
Scheme 3
scheme 3

The synthesis of compounds 9–12 is shown

Full size image
Fig. 2
figure 2

Chemical structure of the synthetized compounds. Compounds were grouped according to the synthesis scheme. Group 3 contains the most bioactive compounds

Full size image

Results and discussion

Synthesis of phloroglucinol compounds

We first set out to synthesize a series of analogs of MPA where the methyl ketone had been replaced with a sequentially longer hydrocarbon chain of 3–6 carbon atoms in length. The synthesis of compounds 1–4 is summarized in Scheme 1. Commercially available trihydroxytoluene was reacted with 0.5 equivalent of the appropriate anhydride in the presence of 1 equivalent of Lewis acid (BF3•OEt2) using dioxane as the solvent. Using the anhydride as the limiting reagent resulted in a correspondingly low yield (18–40%), but prevented over acylation of the substrate. We also undertook the synthesis of a series of acylated phloroglucinols, compounds 5–8, as described in Scheme 2. Phloroglucinol was dissolved in carbon disulfide and nitrobenzene and treated with 10 equivalents of the appropriate acid chloride using AlCl3 (10 equivalents) as catalyst. Heating for one hour followed by work-up afforded compounds 5–8 in a range of 41–53% yield. A series of diacylated phloroglucinol derivatives were synthesized as described in Scheme 3. Phloroglucinol was dissolved in dioxane and reacted with an excess of the appropriate anhydride (10 equivalents) and an excess (10 equivalents) of Lewis acid. Heating for 30 min followed by work up produced diacylphloroglucinols 9–12 in a range of 17 to 55% yield. The complete characterization data for each compound is provided in the supplementary material

Evaluation of antibiotic properties

The minimum inhibitory concentration (MIC) of UA against S. aureus ranges between 6 and 32 μg ml−1 depending on the strain [12]. To evaluate the antibiotic activity of the synthesized compounds, we performed MIC assays using a laboratory strain of S. aureus (ATCC 29213) and MRSA strains isolated from the sputum of CF patients [13]. Table 1 shows MICs of the synthesized compounds. Group 3 had the lowest MICs, ranging between 0.125 and 8 μg ml−1. Among them, compound 11 had the lowest MIC against all the strains of S. aureus tested ranging from 0.125 to 0.5 μg ml−1, while compounds 9 and 10 had MIC values ranging from 2 to 8 μg ml−1. The common structural feature for the four most active compounds (9–12) is a diketo moiety with a mutual ortho phenolic group. The compounds differ only in the chain length of the alkyl group on the anhydride that was used in the acylation step. We then tested the ability of compounds 9–12 to effectively kill S. aureus by performing minimum bactericidal concentration (MBC) assays. All compounds displayed a MBC to MIC ratio lower than 4 for all the strains tested with exception of compound 11 and strain CF188 (Table 2), indicating a bactericidal action [14]. We also evaluated the ability of the synthesized phloroglucinol compounds to eradicate biofilms by performing the minimum biofilm eradication concentration (MBEC) assay (Suppl. Table 1). None of the four compounds from group 3 that had antibacterial activity possessed biofilm eradication activity (Suppl. Table 2).

Table 1 Minimum inhibitory concentration (MIC) of the 12 novel phloroglucinol derivatives against S. aureus strains
Full size table
Table 2 MBC and ratio of MBC to MIC for the four novel phloroglucinol derivatives against S. aureus strains
Full size table

To further investigate the therapeutic potential of the four active compounds, we evaluated their hemolytic properties against red blood cells (RBC) (Table 3). Only 11 displayed minor hemolytic activity of 8% at the highest concentration tested (32 μg ml−1), suggesting it would be unsuitable for therapeutic purposes. Furthermore, the lack of hemolytic activity for compounds 9, 10, and 12 suggests that this class of compounds may have promising selectivity against bacterial cells. Hence, more research on understanding the mechanism of action and the apparent selectivity is merited.

Table 3 In vitro hemolysis of ovine red blood cells by novel phloroglucinol derivatives
Full size table

We also investigated the interaction of the compounds with clinically relevant antibiotics that have various mechanisms of action. We selected compound 9 as representative for checkerboard assays as it displayed strong activity, but without hemolysis. As shown in Table 4, Compound 9 interacted additively with oxacillin, cephalexin, rifampicin, and trimethoprim, and acted synergistically with doxycycline. This trend was seen in ATCC 29213 and the MRSA strains. Studies have shown that synergies often occur between drugs that target the same cellular process [15] and also can happen due to modulation of intracellular drug concentrations. Doxycycline targets the small ribosome subunit (30S subunit) interrupting protein synthesis. In a study conducted by Sahuquillo-Arce et al. it was shown that linezolid interacted synergistically with doxycycline [16]. Linezolid binds to the 50S ribosomal subunit inhibiting protein synthesis. Therefore, the synergistic interaction between doxycycline and linezolid occurs because the two compounds act on interacting targets [16]. It is tantalizing to speculate that the mechanism of action of compound 9 is related to the inhibition of protein synthesis, a hypothesis worth to further be tested.

Table 4 Interaction of Compound 9 with several clinically relevant antibiotics against S. aureus
Full size table

Conclusion

In the present study, we have synthesized 12 phloroglucinol derivatives and characterized their antibiotic properties against S. aureus, an important CF pathogen. Diacylated derivatives displayed the strongest bactericidal activity against MRSA clinical isolates. Furthermore, these compounds displayed apparent selectivity in their action, as they were not hemolytic. Of interest, compound 9 interacted synergistically with doxycycline. Hence, it could potentially be used as a drug combination to treat highly resistant MRSA strains.

Experimental procedures

Bacterial strains and growth conditions

Strains (Suppl. Table 1) were grown in Tryptic soy broth (TSB) (Difco) at 37 °C and 230 rpm shaking for routine overnight culturing unless otherwise stated.

MIC and MBC assays

The CLSI protocol for determining the MIC and MBC of antibiotic activities in S.aureus was followed [17]. Briefly, an overnight culture was diluted equivalent to a 0.5 McFarland standard and then further diluted to inoculate wells of a 96-well plate, containing an antibiotic dilution gradient, with 5 × 105 cells each. The plate was grown statically at 37 °C for 24 h. To determine the MIC the plates were visually inspected for growth and by reading the optical density (OD600) with a BioTek plate reader. To determine the MBC, 10−1–10−3 dilutions of cell suspensions exposed to the antibiotic dilution gradients were plated on LB plates and incubated at 37 °C for 24 h. The number of colony forming units (CFU) per dilution was recorded. As a positive control, cells suspensions treated similarly but without exposure to the compounds were used. The MBC was calculated using the formula:

$${\mathrm {MBC}} = \frac{{{\mathrm {No.}}\,{\mathrm {of}}\,{\mathrm {CFU}}\,{\mathrm {in}}\,{\mathrm {a}}\,{\mathrm {pspecific}}\,{\mathrm {concentration}}}}{{{\mathrm {No.}}\,{\mathrm {of}}\,{\mathrm {CFU}}\,{\mathrm {in}}\,{\mathrm {the}}\,{\mathrm {postive}}\,{\mathrm {control}}}} \times 100$$

where number of CFU in a specific concentration is the average CFU of each dilution at each concentration tested. A concentration was considered bactericidal if it killed 99.9% of the bacterial population [14].

Minimum biofilm inhibitory concentration (MBIC) assay

Assays were performed as previously described [18]. Briefly, an overnight culture was diluted 100-fold and incubated at 37 °C for 3 h to reach mid-log phase. The subculture of S. aureus was prepared by diluting the overnight culture in 1:1000 (∼105 CFU ml−1) into TSB. Wells of a 96-well plate were used to serially diluted the desired compound in TSB and then inoculated with 5 × 106 cells each and grown statically at 37 °C for 24 h. The wells were then rinsed twice with PBS (pH 7.2, 0.8% NaCl, 0.02% KCl, 0.17% Na2HPO4, 0.8% KH2PO4) and then followed with the resazurin cell variability assay below.

MBEC assay

Assays were performed as previously described [18]. Briefly, an overnight culture was diluted 100-fold and incubated at 37 °C for 3 h to reach mid-log phase. The subculture of S. aureus was prepared by diluting the overnight culture in 1:1000 (∼105 CFU ml−1) into TSB. Wells of a 96-well plate were inoculated with 5 × 106 cells each and grown statically at 37 °C for 24 h. The media was then removed from each well. Then serial dilutions of the desired compound in TSB were added to the wells and grown statically at 37 °C for 24 h. The wells were then rinsed twice with PBS (pH 7.2, 0.8% NaCl, 0.02% KCl, 0.17% Na2HPO4, 0.8% KH2PO4) and then followed with the resazurin cell variability assay.

Resazurin cell viability assay

Viable cells with active metabolism reduce resazurin into fluorescent resorufin [19]. To the washed wells, 20 μl of CellTitre-Blue (Promega) was added with 100 μl of PBS and incubated at 37 °C for 1 h. Fluorescence was measured (530/25 nm excitation and 590/35 nm emission) using a BioTek Synergy plate reader. Wells that had the same fluorescence as a resazurin-only control were considered to have no cell viability.

Checkerboard assay

Checkerboard assays in 96-well format were performed as previously described [20]. Briefly, the inoculum was prepared as for the MIC testing above and added to a two-dimensional gradient of the compound of interest. Plates were grown statically at 37 °C for 24 h then visually inspected for growth. The fractional inhibition concentration (FIC) index is calculated using the formula

$${\mathrm {FIC}}\,{\mathrm {Index}} = \frac{{C}_{1}}{{{\mathrm {MIC}_{1}}}} + \frac{{C}_{2}}{{{\mathrm {MIC}_{2}}}}$$

where MIC1 and MIC2 are the MICs of antibiotics 1 and 2 alone, respectively, and C1 and C2 are the concentrations of antibiotics 1 and 2, respectively, when at the combined MIC. When the FIC index was ≤0.5 the antibiotic combination was considered to be synergistic; when the FIC index was >0.5 and ≤4.0 the effect of the two antibiotics was considered to be additive. Finally, when the FIC index was ≥4.0, the antibiotic combination was defined as antagonistic [21].

Hemolysis assay

Hemolytic activity on ovine erythrocytes was assessed as previously described [22, 23]. Briefly, ovine erythrocytes were gently pelleted by centrifugation and washed thrice with PBS then resuspended in PBS at a 1:5 dilution. In 96-well format, 100 µl erythrocyte suspension was added to a dilution gradient of the desired compound and incubated statically at 37 °C for 1 h. Triton X-100 at 0.1% was included as a positive control and a dimethyl sulfoxide (DMSO) concentration gradient as a negative control. Erythrocytes were pelleted and the absorbance of the supernatant was measured at 540 nm. The percentage of lysis was calculated as follows:

$${\mathrm {\% Lysis}} = X - BT - B\,x\,100\%$$

where B is the A540 nm of the negative DMSO control, T is the A540 nm of the positive Triton X-100 control, and X is the A540 nm of the analyzed sample. A540 nm > 40% hemolysis was considered as high percent hemolysis and 5–10% hemolysis was considered as low percent hemolysis for a given compound [24].

Synthesis and characterization of compounds 1–12

Reagents and solvents were used as purchased without further purification. Thin layer chromatography (TLC) was conducted with 0.25 mm F254 silica gel on aluminum plates and visualization was conducted with 254 nm UV light. The purification of all compounds was carried out over flash grade-silica gel and fractions combined according to TLC. Solvent was removed under reduced pressure. The 1H (300 MHz) and 13C NMR (125 MHz) spectra were collected on a Bruker Avance-300 console and were referenced to solvent (CDCl3 = 7.24  ppm). The complete synthetic procedures and characterization data for compounds 1–12 are included in the electronic supplementary material.