In vitro and in vivo antibacterial activities of cranberry press cake extracts alone or in combination with β-lactams against Staphylococcus aureus

The antibiotic susceptibility testing reference strain Staphylococcus aureus ATCC 29213, the sequenced MRSA strains COL and N315, as well as the bovine mastitis isolates: Newbould (ATCC 29740), SHY97-4320 and SHY97-3906, were used in this study. Strains SHY97-4320 and SHY97-3906 were isolated from clinical mastitis and were previously described [21]. Bacteria stocks were kept frozen at −80°C and grown on Mueller Hinton agar or Cation adjusted Muller Hinton broth (CAMHB, Becton Dickinson, Mississauga, ON, Canada). Viable bacterial counts (CFU) were determined on Tryptic Soy agar (TSA, Becton Dickinson).

A commercial cranberry product, Nutricran®90 (NC90), was obtained from Decas Botanical Synergies (Wareham, MA, USA). The NC90 contained at least 88% carbohydrates, 30% organic acids, 2 to 3.8% total phenolic compounds, 0.3 to 1% anthocyanins, 0.8 to 1.5% proanthocyanidins, 300 to 435 μg/g of quercetin, and was 100% soluble in water as described by the manufacturer. Laboratory and pilot-scale preparations of extracts were performed with fresh Ben Lear cranberries obtained from a local farm (Agassiz, British Columbia, Canada) or from Ocean Spray (Langley, BC, Canada) as described previously [17]. Chloramphenicol, oxacillin, amoxicillin, norfloxacin, rifampicin and vancomycin were purchased from Sigma Aldrich (Oakville, Ontario, Canada).

Laboratory-scale fraction were obtained from 2 Kg of fresh fruits which were hammer milled through 1 cm holes then pressed in a miniature rack and cloth press at 3000 psi for 10 min to obtain juice and press cake (PC). Triplicate allotments of PC (90 ± 5 g) were placed into screw cap polyethylene centrifuge containers and covered with 95% ethanol and kept at room temperature for 48 h. Triplicate extracts were pooled, dried by rotary evaporation to remove the ethanol (Büchi rotavapor R, Brinkmann Instruments Canada Corp.; 30°C, speed 5–7, full vacuum (−25kPa; Büchi Vac V-500) with condenser temperature of 0°C), and freeze dried (Virtis 50-SRC, Gardiner, NY, USA; condenser at −56°C; vacuum at −100 milli-Torr; shelf at 25°C) to obtain fraction FC111. Replicate PC mashes were used for one and two other extractions with 95% ethanol, and the soluble fractions were freeze dried (fractions FC121 and FC131, respectively). Pilot-scale extraction procedures were previously described [17]. In brief, the fresh cranberries were pressed in a hydraulic rack and cloth press and the juice was separated from the PC. Following this step, the PC was covered with 95% ethanol for 24 h. Then, the crushed cranberries were decanted and the soluble fraction was refrigerated at 1°C (fraction FC111).

Phenolic acids, anthocyanins, flavonols and tartaric esters were determined by the method of Oomah et al. [22]. Antioxidant activity was measured through oxygen radical absorbance capacity (ORAC), as described previously [23]. The polyphenolics and antioxidant activities of the fresh cranberry fruit fractions obtained by pilot-scale ethanol extraction of the PC was previously reported [17]. As described, PC extracts including FC111 showed three to four times the phenolic acid, tartaric ester contents and antioxidant activities, as well as five to 10 times the flavonols and anthocyanins of their respective juice powders [17]. There was good similarity between the amounts of measured chemical constituents of laboratory- and pilot-scale preparations, as well as their antimicrobial activity (see result section).

Minimal inhibitory concentrations (MICs) of antibiotics and cranberry extracts were determined by a broth microdilution technique, following the recommendations of the Clinical and Laboratory Standards Institute [24]. A checkerboard protocol [25] was conducted by a broth microdilution method similar to that used for standard MIC determinations to evaluate the effect of varying concentrations of cranberry extracts on the activity of known antibiotics against S. aureus strains. The fractional inhibitory concentration (FIC) indices were calculated as follow: FIC index = FIC_A + FIC_B = A/MIC_A+ B/MIC_B, where A and B are MICs of compounds A and B in combination, MIC_A and MIC_B are the MICs of compound A and compound B alone and FIC_A and FIC_B are the FICs of compound A and of compound B. Indifference for drug interactions or an additive effect is demonstrated if the FIC index is >0.5-4, a synergy if the FIC index is ≤ 0.5, whereas an antagonistic effect is represented by a FIC index of >4 [26].

Time-kill experiments were performed to observe the bactericidal effect (i.e., a 3 log10 reduction in CFU/mL) of cranberry extracts alone or in combination with sub-inhibitory concentrations (sub-MICs) of traditional antibiotics. The CAMHB medium was inoculated with S. aureus at ~10⁵-10⁶ CFU/mL and grown at 35°C, with shaking (225 rpm) in the absence or presence of tested agents at the specified concentrations (see the figure legends). At several time points, bacteria were sampled, serially diluted and plated on TSA for CFU determinations.

The modulation of transcriptional profiles of S. aureus strains ATCC 29213, MRSA N315, SHY97-4320 and SHY97-3406 by the cranberry product NC90 was first evaluated by exploratory transcriptional profiling experiments using a sub-genomic array as described previously (in detail) [27, 28]. Bacteria (~10⁷-10⁸ CFU/mL) were exposed to NC90 at the MIC (2 mg/mL) for 30 min in CAMHB. Culture samples (~10⁸ cells) were treated with RNAprotect (Qiagen, Mississauga, ON, Canada) for at least 10 min. Bacterial pellets were then suspended in 100 μL of TE buffer containing 200 μg/mL lysostaphin and incubated at room temperature for 1 h. RNA was extracted using the RNeasy Mini Kit and treated with the RNAse-free DNAse set (Qiagen). Total RNA (2.5 μg) and 1 μL of the appropriate RNA spike from Lucidea Universal Scorecard (GE Healthcare Life Sciences, Baie D’Urfé, QC, Canada) was reverse transcribed using 5 μg of random hexamers, a dNTP mix, 5-[3-aminoallyl]-2-dUTP and the Superscript II RT (Invitrogen, Life Technologies Inc., Burlington, ON, Canada). The amino-modified cDNA was purified and labeled with NHS-Cy3 (untreated control) and -Cy5 (bioactive agent test) prior to hybridization on GAPS II slides (Corning, NY, USA). The probes (200 pmol, ≥8% incorporation) were suspended in 18.5 μL of hybridization buffer (5X SSC, 0.1% SDS, 25% formamide, 100 μg/ml mouse COT1 DNA (Invitrogen)). The prehybridization, hybridization and washing steps were done as prescribed for Corning Gaps II Slides. Hybridization signals for each spot were quantified with the ScanArrayExpress Microarray Scanner and the ScanArrayExpress software V 2.2.0.0022 (Perkin Elmer, Wellesley, MA, USA). Intensity of each dye was adjusted using the signal of the control spots from Lucidea Universal Scorecard and data were submitted to the Lowess normalization. Only signals showing intensity three times above the background were analyzed as described previously [18, 27]. Expression ratios (Cy5 [treated]/Cy3 [untreated control]) were represented in log2. Only log2 ratios with a minimum mean of ≥2.0 (or ≤ −2.0) obtained for at least two of the four S. aureus strains tested were considered. Data represent the average of triplicate spots from duplicate independent experiments for each of the four tested S. aureus strains. The modulation of expression of marker genes providing transcriptional signatures was thereafter confirmed by qPCR analyses.

Transcriptional modulations induced by cranberry fractions and observed by using exploratory microarrays were confirmed by qPCR for a selection of marker genes and strains. Additional RNAs, from three independent bacterial cultures and prepared as described above, were obtained for qPCR analyses. The qPCR assays using SYBR green (Invitrogen) and the JumpStart Taq DNA Polymerase (Sigma) were performed as described previously [27] for the marker vraS, sgtB, murZ, mrsA, lytM and mntB genes indicative of a cell wall stress [29, 30]. The relative expression ratios were calculated using the cycle threshold (C _t) of the untreated culture (no-bioactive agent control) and that of the housekeeping gyrA gene [n-fold expression = 2^{-ΔΔC t}, where ΔΔC _t = ΔC _t (treated culture)/ΔC _t (untreated culture) and ΔC _t represents the difference between the C _t of the gene studied and the C _t of gyrA]. The sequences of primers for qPCR were: gyrA-forward 5^′-CATTGCCAGATGTTCGTGAC -3^′ and gyrA-reverse 5^′-CACCAACGATACGTGCTGAT -3^′; vraR- forward AAAAGATATCGCCGATGCAG and vraR- reverse ATAACTCTGCCGCGCTTTTTC; sgtB- forward AAACCGCCGAAAAAGAAAAA and sgtB-reverse TCATCCACATTATCGCGTGT; msrA- forward GCAAATGGTGTAGGTAAGACAACT and msrA-reverse ATCATGTGATGTAAACAAAAT; murZ- forward ACGCACACTAAATGGGGAAG and murZ-reverse CCTCAGCAATTAAACCAGCA; lytM- forward ATTAACAGCAGCAGCGATTG and lytM- reverse TGTGCTTGTTGGGTGTTTGT; mntB- forward CCTGGTGTTGCCCTATCATT and mntB-reverse GGCGTCAGGTTTCGTTTTAC.

Macromolecular biosynthesis assays were performed exactly as described previously [31]. Briefly, S. aureus ATCC 292123 was grown in a complete defined medium, and transferred to a round-bottom microwell plate containing radiolabeled precursors [³H]-Leucine; [³H]-Thymine; [³H]-Uridine (Perkin Elmer, Montreal, QC, Canada) or D-[³H]-Alanine (American radiolabeled chemicals, St-Louis, MO, USA), as well as various dilutions of antibiotic agents. Incorporation was allowed to proceed for 45 min (for Leu and D-Ala), or 35 min (for Thy and Uri) at 35°C. The reaction was terminated by transferring the mixtures into pre-chilled trichloroacetic acid (10% final), followed by collection of the precipitated macromolecules on GF/C filter using a dot-blot filtration system. The dried filter was cut and the level of radiolabeled precursor incorporation was determined by liquid scintillation counting. Data represent the average of three independent experiments.

The mastitis model was previously optimized for experimental therapy [32]. CD-1 lactating mice (Charles River, St-Constant, Canada) were used 12 to 14 days after offspring birth. Pups were removed 1 h before intramammary inoculation, under anesthesia, of both L4 (on the left) and R4 (on the right) abdominal mammary glands with 100 μL of S. aureus Newbould (1 CFU/μL). Glands were treated (intramammary) 4 h after infections with amoxicillin, a cranberry extract or a combination of both. Mice were then sacrificed 14 h post-treatment for mammary gland sampling, homogenization, plating, and bacterial counts. Data represent two independent experiments using several mice and glands in each.

One milligram aliquots of each dry powdered fraction were suspended in 20% aqueous methanol, 0.1% TFA, at 1 mg/mL, prior to HPLC-MS analysis. This analysis was performed on an Agilent 1100 series LC/MSD SL system (Agilent Technologies, Mississauga, ON), equipped with a quaternary pump, in-line degasser, robotic auto sampler and diode array (DAD) and single quadropole mass spectrometry (SQD) detectors. The separation and characterizations were achieved using a Zorbax Eclipse XDB-C18 (Agilent Technologies) analytical column (4.6 × 150 mm, 5 μm particle size) fitted with a Zorbax Eclipse XDB-C18 analytical guard column (4.6 × 12.5 mm, 5 μm particle size). A tertiary solvent system was used, (A = 0.1% aqueous trifluoroacetic acid; B = water:methanol:trifluoroacetic acid 1:1:0.1; C = acetonitrile), at a flow rate of 0.5 mL/min, with a column temperature of 45°C, and 15 μL sample injection volume , according to the following gradient: Time = 0 min. A:B = 100:0; Time = 2 min. A:B = 85:15; Time = 20 min. A:B = 0:100; Time = 26 min. B = 100% isocratic; Time = 30 min. B:C = 0:100; Time = 33 min. C = 100% isocratic; post time equilibration (3 min.) A = 100%. The entire UV–vis absorption spectrum (190–900 nm) was recorded, while general phenolics (280 nm), anthocyanins (520 nm), and flavonols (360 nm) channels were monitored. Atmospheric pressure electrospray ionization mass spectra in both positive (PIM) and negative (NIM) ion modes (PIM = 150–1400 amu; NIM = 150–2000 amu) were recorded on duplicate injections of the same sample (drying gas flow 10 L/min; nebulizer pressure 30 psig; dry gas temperature 350°C; PIM capillary voltage 3000; NIM capillary voltage 3500). Peak identities were assigned using a combination of UV–vis absorption spectra, MS fragmentation patterns, congruent retention times with commercially available standards, and comparison to relevant cranberry literature [33‐35].

Statistical analyses were carried out with the GraphPad Prism Software (v.5.00). Tests used for the analysis of each experiment are specified in the figure legends.

The MIC of the cranberry product NC90 against S. aureus ATCC 29213, Newbould and MRSA COL was 2 mg/mL. In an attempt to understand the basis of the growth inhibitory effect of NC90 observed on S. aureus, we studied the modulation of bacterial gene expression for a variety of strains by microarray analyses (Table 1). Exposure of any of the four tested S. aureus strains (ATCC 29213, MRSA N315, SHY97-4320, SHY97-3906) to NC90 up-regulated strongly several genes involved in cell wall biosynthesis, similar to the up-regulation of genes caused by antibiotics known to disrupt peptidoglycan biosynthesis. Indeed, the strong expression of fmt, pbp2, murZ, sgtB, tcaA and vraS/R genes is evidence of a cell wall stress induced by the tested cranberry product NC90 (Table 1). Our data are consistent with the results from other laboratories on the transcriptional profiles obtained by exposure of S. aureus to oxacillin, bacitracin and D-cycloserine [30] or to daptomycin and vancomycin [29] and these similarities are also reported in Table 1. NC90 also up-regulated lytM, encoding a peptidoglycan hydrolase distinctively like daptomycin (Table 1), and down-regulated genes such as spa, mntABC, isdA, hla, and femC.

To follow up, extracts from press cakes (PC) were investigated. The MIC of FC111 (lab- and pilot scale extracts) was lower than that of NC90 and represented 1 mg/mL against S. aureus ATCC 29213, Newbould and MRSA COL. Besides, the MIC value for both FC121 and FC131 was higher (3 mg/mL). Since FC111 showed the lowest MIC (best efficiency), its bactericidal effect against S. aureus ATCC 29213 and MRSA COL was evaluated in time-kill experiments. FC111 at 1 mg/mL (MIC) induced more than a 3-log10 reduction of CFU/mL against both strains compared to their respective untreated control cultures after 8 h of incubation and a concentration of 2 X MIC (2 mg/mL) yielded more than a 6-log10 reduction of CFU/mL at 8 h and no detectable CFU after 24 h (data not shown).

Subsequently, we used qPCR to first confirm the transcriptional effects induced in S. aureus ATCC 29213 following exposure to NC90 in DNA array experiments (above) and secondly, to detect similar transcriptional modulations induced by the most active ethanol-extracted PC fraction FC111. Modulation of the same cell wall stress markers (lytM, vraR, sgtB, msrA, murZ and mntB) was indeed detected for both FC111 and NC90, indicating a similar mode of action against S. aureus (Figure 1).

Whole-cell macromolecular biosynthesis assays were performed to elucidate the mechanism of action of FC111 and to confirm its inhibitory effect on cell-wall biosynthesis. S. aureus ATCC 29213 was treated with several concentrations of the cranberry fraction FC111 or known antibiotics and the percentage of incorporation of radiolabeled precursors into macromolecules was evaluated (Figure 2). All known drugs targeted the expected macromolecular biosynthesis; chloramphenicol, norfloxacin, rifampicin and vancomycin inhibited protein, DNA, RNA and cell wall biosynthesis, respectively. In the presence of 2 (2 mg/ml) and 4 times (4 mg/ml) the MIC of the cranberry fraction FC111, the incorporation of D-[³H]Ala, a precursor of cell wall peptidoglycan, was significantly reduced compared to that observed for precursors of DNA, RNA and proteins (Figure 2E).

Since exposure of S. aureus to FC111 induced cell wall stress, we investigated the potential synergy between this cranberry PC fraction with amoxicillin (used for treatment of S. aureus-induced bovine mastitis in veterinary medicine) and oxacillin (used for many clinical indications in humans). Checkerboard assays were performed and a FIC index was determined for each tested combination against S. aureus ATCC 29213 (β-lactam susceptible strain) and MRSA COL (with the mecA/PBP2A-mediated β-lactam resistance). The cranberry fraction FC111 increased the inhibitory activity of amoxicillin and oxacillin against both tested S. aureus strains by triggering up to 512-fold reductions in β-lactam MIC values (Table 2). Combination of FC111, 128 or 256 μg/mL (i.e., 1/8 or 1/4 FC111 MIC, respectively), with either of the β-lactams yielded FIC indices of ≤0.25, for each of the strains tested. These FIC indices clearly showed synergy between FC111 and β-lactams against S. aureus. Data also demonstrated that this synergy can efficiently bypass the mecA-mediated mechanism of β-lactam resistance found in MRSA strain COL.

The synergistic effect arising from the combination of β-lactams and FC111 in checkerboard assays was explored in greater details by using kill kinetic studies. Combination of the cranberry fraction FC111 at its MIC (1 mg/mL) and 1/8 × MIC of amoxicillin induced a strong bactericidal effect (>3 Log10 CFU/mL) against a bovine mastitis strain (S. aureus Newbould, also used in our mastitis model of infection in the mouse [see below]) compared to the bactericidal effect observed with FC111 or amoxicillin alone (Figure 3A). Moreover, the combination of FC111 at its MIC and oxacillin (1/512 × MIC) also showed strong bactericidal effect against S. aureus MRSA COL (Figure 3B).

The in vivo relevance of the synergy between the cranberry extract FC111 with β-lactams was determined using a mouse mastitis model. The mice were infected with the bovine strain S. aureus Newbould and treated with FC111, amoxicillin or a combination of both (Figure 4). Compared to the untreated mice, there was no significant change in bacterial loads (CFU/g of gland) when mice were treated with either one of the compounds. However, a significant reduction in bacterial counts (1.8 Log10 CFU/g of gland) was observed using the FC111-amoxicillin combination.

HPLC-MS was performed on laboratory-scale cranberry fraction FC111, which had the following spectrophotometrically measured [22] classes of chemical compounds per gram of freeze-dried powder: 132.4 (total phenolics), 8.3 (tartaric esters), 9.6 (flavonols) and 23.8 (anthocyanins) mg equivalents of catechin, caffeic acid, quercetin and cyanidin-3-glucoside, respectively. The total phenolics, tartaric esters and flavonols compared well with the composition of the pilot scale sample FC111 [17]. Laboratory- and pilot-scale prepared FC111 provided identical antibacterial activity against S. aureus strains as reported above. HPLC separation of cranberry fraction FC111 constituents was achieved in less than 30 min, using the program described above. A typical chromatographic trace, recorded at 254 nm from laboratory scale FC111 (Figure 5), demonstrates the separation of iridoids, hydroxybenzoic acid, hydroxycinnamic acid, anthocyanin and flavonol components (Table 3). Compounds lacking UV-absorbing chromophores (sugars, small organic acids etc.) were omitted from analysis. In fraction FC111, qualitatively representative of all PC extracts, the principal components were anthocyanins, peaks 21, 24–29, 31, 33 (40%), flavonols, peaks 30–32, 34–44 (25%), iridoids, peaks 2a,b, 3, 16b (22%), and the conjugated glycosides of simple phenolic acids, peaks 4–15, 16a, 17–19 (9%), based on integrated peak areas recorded at 254 nm measured against total absorbance. Absolute quantification of peaks (mass balance) could not be measured, due to lack of analytical standards. Relative abundances between classes of molecules were calculated however, using integral values measured at 254 nm, while specific wavelengths were used to compare individual compounds within a class; 520, 360, and 280 nm for anthocyanins, flavonols and phenolics, respectively (Table 3, peak identities). The primary anthocyanins present were peonidin-3-O-galactoside (peak 28, 43% abundance), cyanidin-3-O-galactoside (peak 24, 18% abundance) and peonidin-3-O-arabinoside (peak 31, 15% abundance) (Table 3, abundances).