In the present work, C, ECg and EGC were isolated from hawthorn. A checkerboard method was used to determine which monomer or combination of the three monomers could produce the best ILSMR effect. The combination of C and EGC showed no ILSMR effect, with FICIs higher than 0.5 (Table S1). EGC in combination with ECg also had no ILSMR effect, with most FICIs above 0.5 and only one at 0.5 (Table S2). Significantly, the combination of C and ECg (Table 1) or of C, ECg and EGC showed strong ILSMR effects, with FICIs lower than 0.5 (Table 2), and there was no significant difference between two combinations. Because their pharmacological effects were equivalent, the more simple combination was preferentially chosen. Therefore, the combination of C and ECg was chosen and then its ILSMR effect was further investigated.

The MIC of C against WHO-2 was >1024 mg/L (no antibacterial effect), while that of ECg was 128 mg/L (very low antibacterial effect). C or ECg (32 mg/L) in combination with oxacillin produced FICIs of more than 0.5 or 0.5, demonstrating no or very weak ILSMR effect. Meanwhile, the FICI of C and ECg together with oxacillin was lower than 0.5, suggesting this combination had a strong ILSMR effect.

Two combinations of C and ECg at different concentrations, C (64 mg/L) + ECg (32 mg/L) or C (128 mg/L) + ECg (16 mg/L), combined with oxacillin produced FICIs lower than 0.375 (Table 1). Considering the good solubility of C and poor solubility of ECg in H₂O, the latter combination with the lower ECg concentration, C (128 mg/L) + ECg (16 mg/L), was selected and investigated in the subsequent experiments.

C in combination with ECg could potentiate the antibacterial effects of oxacillin against WHO-2. Therefore, it was of interest to determine whether C in combination with ECg could potentiate other β-lactam antibiotics or even non-β-lactam antibiotics and to verify the ILSMR effects, not only in WHO-2, but also in clinical MRSA strains. Herein, The 45 clinical MRSA strains were identified by drug susceptibility assay and amplification of the mecA gene. They all are resistant to oxacillin (MIC > 4 mg/L) and other five β-lactams with high levels of resistance. In the present experiment, the effects of C in combination with ECg to potentiate β-lactam antibiotics (oxacillin, ampicillin or ampicillin/sulbactam, cefazolin, cefepime and imipenem/cilastatin) and non-β-lactam antibiotics (vancomycin, linezolid and teicoplanin) against 45 clinical MRSA strains were observed.

The results showed that C or ECg each alone had no ILSMR effect, but C in combination with ECg could potentiate antibacterial effects of six β-lactam antibiotics (Table 3). Three combinations [C (32 mg/L) + ECg (4 mg/L), C (64 mg/L) + ECg (8 mg/L), and C (128 mg/L) + ECg (16 mg/L)] potentiated the effects of all six β-lactam antibiotics against almost all strains. However, none of these combinations could potentiate the effects of non-β-lactam antibiotics (Table S3).

In order to determine whether C in combination with ECg could potentiate β-lactam antibiotics in vivo, the MRSA infection model was established in mice challenged with a sublethal dose of WHO-2. Considering the t_1/2 of catechins was very short (approximately 1–4 h), blood bacterial loads were tested at 1, 2 and 4 h time points after the first treatment of C in combination with ECg [22]. The results showed there was no significant difference of bacterial blood load among mice treated with normal saline only, oxacillin only and C in combination with ECg. However, the bacterial load in mice treated with oxacillin (200 mg/kg/day) in combination with C and ECg was significantly lower than that in mice treated with oxacillin alone or C in combination with ECg (p < 0.05). The bacterial loads in mice treated with oxacillin in combination with C and ECg (100 mg/kg/day) was lowest (Figure 1). These results demonstrated C in combination with ECg potentiated oxacillin’s antibacterial effect in vivo, too.

The efflux system could confer multidrug resistance of MRSA. In order to investigate whether C and ECg demonstrated its ILSMR effect via increasing antibiotics accumulation, the influence of C and ECg on drug accumulation was investigated using fluorospectrophotometry assay and laser confocal scanning microscope observation.

Daunorubicin is not a β-lactam antibiotic but with red autofluorescence. It is used as a tracer agent because it can be observed more directly. The influence of daunorubicin was observed using drug susceptibility test and the result was shown by the time-kill curve (Figure S1). After it was confirmed that daunorubicin (20 and 40 mg/L) had no effect on the growth of WHO-2, the effect of C and ECg on the accumulation of daunorubicin within WHO-2 was detected. The results from confocal scanning microscopy showed C (256 mg/L) did not increase the daunorubicin accumulation within WHO-2 but ECg (32 mg/L) increased the drug accumulation. Significantly, C (256 mg/L) in combination with ECg (32 mg/L), increased more daunorubicin accumulation than ECg alone. Interestingly, the higher the concentration of C combined with ECg, the higher the daunorubicin accumulation (Figure 2A). The similar result was also observed from fluorospectrophotometry (Figure 2B). The above two results demonstrated that the ILSMR effect of C in combination with ECg was strongly related to the increased antibiotics’ accumulation.

In order to investigate whether C in combination with ECg manifested its ILSMR effect via inhibiting mRNA expression of efflux pump genes, the influence of C in combination with ECg on mRNA expression of efflux pump genes were investigated using reverse transcription Polymerase Chain Reaction (RT-PCR) method.

The results showed that reserpine as a positive control could downregulate mRNA expressions of norA and norC not abcA, which was in accordance with previous reports [23,24]. C or ECg alone did not influence efflux pump gene expression, but C in combination with ECg could downregulate mRNA expressions of norA, norC and abcA among eight efflux pumps of MRSA (Figure 3), thus suggesting that increased daunorubicin accumulation was related to the inhibition of three efflux pumps’ gene expressions.

In traditional Chinese medicine, hawthorn is used as a peptic agent for stimulating digestion and promoting the function of the stomach, improving blood circulation and removing blood stasis. It is used as the agent for treatment of indigestion with epigastria distension, diarrhea, abdominal pain, amenorrhea, hypertension and hyperlipidemia, as well [25]. To the best of our knowledge, this is the first report demonstrating that hawthorn possesses ILSMR effects.

On the basis of the literature data, there are flavonoids, oligomeric proanthocyanidins, phenolicacids, triterpeneacids, organicacids and sterols within hawthorn fruit. Flavonoids are considered to be the main groups of active constituents in hawthorn extracts [21]. Catechins belong chemically to flavonoids.

Catechins are present in many plants and have many pharmacological effects. Although some of them may have the capacity to sensitize MRSA strains to oxacillin and other β-lactam antibiotics, they only possess weak antibacterial properties [26]. For example, galloy catechins, such as ECg and Cg, reduced the high MIC level of β-lactams to the antibiotic breakpoint or even lower than the breakpoint [14,15,26,27], but non-galloylated catechins, such as C and EC, had no such effect [15]. In the present study, we found ECg only had weak ILSMR effects, which did not accord with the previous reports [14,15], while C had no ILSMR effect as reported. The main reason could be that the ILSMR effect of ECg was influenced by its optical activity, which was affected by the extraction process and purity. Therefore, the discrepancies are likely due to different specific optical activities of ECg from the different sources of ECg and different MRSA strains.

We also found that when C combined with ECg, the ILSMR effect was markedly increased and the ILSMR effect of the combination (C and ECg) was enhanced with the increase in the concentration of C, with a higher concentration of C resulting in a lower FICI. This phenomenon did not occur when the concentration of ECg was increased, suggesting that the ILSMR effect of ECg could be potentiated by C. Previously, the cis form of non-galloylated catechins such as (−)-EC and (−)-EGC was reported to enhance the ILSMR effect of ECg [28], but excluding non-galloylated catechins of the trans form such as C. Herein, we firstly reported the ILSMR effects of C in combination with ECg on multiple classic β-lactam antibiotics against not only a standard MRSA strain but also clinical MRSA strains in vitro and in vivo.

In addition, C was found to be more effective than EC and EGC in potentiating the ILSMR effect of ECg against WHO-2 (data not shown). C and EC are known optical isomers. As the pharmacological effect of drugs requires strict chiral recognition by biological macromolecules [29], the bioactivities of optical isomers are generally different. The difference of chemical structure between C and EGC is that EGC has one more hydroxy group in the B ring than in that of C. Overall, the different abilities of C and EGC to potentiate the ILSMR effect of ECg was presumed to be due to the difference in steric hindrance provided by the hydroxy groups in the B ring of these compounds.

Previously, the ILSMR effect of ECg in combination with β-lactam antibiotics against the standard WHO-2 strain has been described [14,15]. Here, C together with ECg also showed an ILSMR effect on β-lactam antibiotics, not only against the MRSA WHO-2 standard strain, but also against 45 clinical MRSA strains, thereby suggesting the potential clinical value of this combination.

The ILSMR effect of ECg in combination with some β-lactam antibiotics against MRSA has also been described previously [14]. In this study, C in combination with ECg showed ILSMR effects on six β-lactam antibiotics: ampicillin, ampicillin/sulbactam, cefazolin, cefepime and imipenem/cilastatin. These β-lactam antibiotics are both classic and representative β-lactam antibiotics in the clinic. However, they cannot be used in the treatment of MRSA infections due to resistance. Our results showed that C in combination with ECg could significantly reduce the MIC of these six antibiotics against almost all of the clinical MRSA strains. These results are significant since they demonstrate the possibility that these antibiotics can be used in the clinic when combined with C and ECg. Interestingly, C in combination with ECg demonstrated ILSMR effects on β-lactam antibiotics, while they had no such effect on non-β-lactam antibiotics, which has not been reported previously.

Previously, ECg was found to prevent biofilm formation and promoted cell wall thickening, leading to cell aggregation without affecting the rate or extent of growth in culture [30]. Furthermore, ECg could insert into bacterial cytoplasmic membrane to disrupt PBP2a-mediated resistance by delocalizing PBP2 [31]. In the present study, we found that the ILSMR effect of C and ECg was also related to antibiotics accumulation. Efflux pumps were related to drug accumulation.

Efflux pumps were capable of extruding antibiotics, leading to multidrug resistance. There were nine efflux pumps within MRSA; they were norA, norB, norC, mdeA, qacA/B, mepA, smr, sepA[32], and abcA[33]. In a recent study, 50% of the 232 isolates of S. aureus were discovered to pump out at least two structurally unrelated substrates [34]. Frequencies of overexpressed efflux genes were norA (23%), norB (25%), norC (17%), mdeA (11%), and mepA (4%) [34], which meant norA and norB were the most popular pumps in S. aureus. However, abcA was reported to be responsible for the β-lactam resistance in S. aureus[33]. Previously, reserpine (a phytoalkaloid) was reported as efflux pump inhibitor (EPI) against various microbes [23,24]. There was one report that ECg inhibited norA efflux pump expression [35]. Herein, our results showed that reserpine down-regulated mRNA expressions of norA and norC, but not abcA, and C, in combination with ECg, down-regulated mRNA expressions of norA, norC and abcA among eight efflux pumps, while ECg or C alone could not. The above results suggested the ILSMR effect of C and ECg was probably related to the inhibition of three pumps’ mRNA expressions. Which pump(s) played a more important role should be further investigated in the future experiments.

In the present experiments, the synergy was assayed by FICI according to the criterion suitable for combination of only two compounds. Synergism was defined as an FICI ≤ 0.5 [36]. As is known, when more than two compounds were combined against bacteria, FICI would be easily more than 0.5, suggesting there was no synergism. In fact, each agent’s concentration had markedly decreased after combination. Therefore, a criterion suitable for more than two compounds should be investigated in the future.

Chemicals used in extraction and isolation, like methanol, ethanol, ethyl acetate, normal butanol, acetone, were from Aladdin Regent Company (Shanghai, China). Macroporous D-101, silica gel and LH₂₀ were from Ouyll Biological Technology Company (Tianjin, China). Oxacillin was purchased from Southwestern Pharmaceutical Corp. LTD. (Guangzhou, China). Ampicillin, cefazolin and cefepime were purchased from North China Pharmaceutical Group Corp. (Shijiazhuang, China). A 2:1 mixture of ampicillin sodium and sulbactam sodium was purchased from Seashore Pharmaceutical Company (Shenzhen, China). A 1:1 mixture of imipenem and cilastatin sodium was purchased from Zhuhai Federal Pharmaceutical Company (Zhuhai, China). Müller-Hinton (MH) agar plates and MH broth were purchased from OXOID Company (Basingstoke, UK).

The fruits of hawthorn (Crataegus pinnatifida) were collected from Yuzhou Market of Traditional Chinese Herbs, He Nan province in China, and identified by Yimin Zheng (Chongqing University of technology, Chongqing, China). The voucher specimen (2010-0900SZ^#) of the hawthorn was deposited in our laboratory for future reference.

The standard MRSA strain WHO-2 (WHO-2) and 45 clinical MRSA strains were kindly provided by Xiaoxing Luo [37] (The Fourth Military Medical University, Xi’an, China) and Peiyuan Xia (Southwestern Hospital, Chongqing, China), respectively. The properties of WHO-2 and 45 clinical MRSA strains were confirmed by both a drug susceptibility assay according to the Clinical Laboratory Standards Institute (CLSI) guidelines of 2010 (strains were resistant to oxacillin (MIC > 4 mg/L) and amplification of the mecA gene identified by Polymerase Chain Reaction (PCR) in our lab. WHO-2 possessed a high level of resistance to oxacillin (MIC was 512 mg/L) and harbored the mecA gene. Forty-five clinical strains were all resistant to oxacillin (MIC > 4 mg/L) and harbored the mecA gene. The primers of mecA gene and 16S were listed (Table S4).

Specific-Pathogen-Free (SPF) Kunming (KM) mice (4–6 weeks old, weighing 18–22 g) were supplied by the Experimental Animal Center of the Third Military Medical University (Chongqing, China). An equal number of male and female mice were used. All animals received humane care in compliance with the Principles of Laboratory Animal Care formulated by the National Society for Medical and Research and the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences (Washington DC, USA). The experiments were approved by a medical ethics committee of The Third Military Medical University.

The fruits of hawthorn were crushed and refluxing extracted with 75% EtOH three times (each extraction period lasted 2 h). Merging the extraction together, the 75% EtOH extract was portioned between EtOAc, n-BuOH. Each extract (EtOAc extract, n-BuOH extract, H₂O extract) was measured for antibacterial activity and ILSMR effect. The active part was chromatographed on macroporous D-101 column (washed by EtOH-H₂O), and 50% EtOH-H₂O fraction was collected by the freeze-drying technique. The sample was chromatographed on a silica gel column with chloroform and acetone afforded the fraction1–8 (frs.1–8), and then these fractions were measured for antibacterial effect and ILSMR effect. Fraction 4, which showed a significant ILSMR effect, was then separated by Sephadex LH-20 column chromatography and preparative HPLC (Agilent 1100 Series equipped with G1361A binary pump, G1365B MWD UV detector, G1364B fraction collector), yielding the three monomeric compounds. The three compounds were identified as (+)-catechin, (−)-epicatechin gallate and (−)-epigallatecatechin by the methods of ¹H-NMR and spiking with reference compounds which were purchased or available in our laboratory (Figure 4).

Each single colony from the MH agar plate was inoculated into a 10 mL volume of liquid MH broth (containing 2% NaCl), according to the CLSI 2010 guidelines, and cultivated aerobically at 37 °C in a heated and shaking chamber for 12 h. These cultures were diluted 1:100 (v/v) in another 10 mL of fresh broth and cultivated aerobically until the log growth phase.

Bacteria in exponential phase (1 × 10⁵ cfu/mL) were inoculated into 96-well plates. MIC of β-lactam antibiotics, C alone and ECg alone were determined by serial two-fold dilutions in MH broth according to CLSI 2010 guidelines. MIC₅₀ and MIC₉₀ were the half and 90% MICs of β-lactam antibiotics against 45 clinical MRSA strains, respectively.

Synergistic effects were assayed by determining the FICI calculated for each combination using the following formula: FICI = (MIC of combined antibiotic/MIC of antibiotic alone) + (concentration of combined “C”/MIC of “C” alone) + (concentration of combined “ECg”/MIC of “ECg” alone). Synergism was defined as a FICI ≤ 0.5, antagonism was defined as an FICI > 4.0, and no interaction was defined with a FICI 0.5–4.0, according to previously described criteria [36].

One hundred and five mice, weighing from 18 g to 22 g, were randomly divided into seven groups (15 mice/group) and intravenously injected with a sublethal dose of live WHO-2 (4.0 × 10⁹ cfu/mL). Four hours later, the mice were given the following treatments (C and ECg mixture by intragastric injection and oxacillin by intramuscular injection): Normal saline only, oxacillin (200 mg/kg/day) only; mixture of C and ECg (total dose 100 mg/kg/day); combination of C and ECg mixture (100 mg/kg/day) with oxacillin (200 mg/kg/day); combination of C and ECg mixture (60 mg/kg/day) and oxacillin (200 mg/kg/day); combination of C and ECg mixture (30 mg/kg/day) and oxacillin (200 mg/kg/day). The total injection volume was 0.4 mL/20 g body weight. One milliliter of blood from each of three mice of each group was collected at 1-, 2- and 4-h time points after the first treatment of the agents, and samples were diluted with sterile normal saline immediately. The blood samples were then inoculated onto MH agar plates containing 50 mg/L of oxacillin. After culturing for 24 h at 37 °C, the bacterial colonies were counted and presented as lg cfu/mL of blood.

The dose of oxacillin used in this study was based on a conversion of clinical dosage regimens. Doses of oxacillin in the range from 30–100 mg/kg/day have been proposed to treat bacterial infection in adult patients, and higher doses are given for serious infections. According to the drug conversion principle applied in pharmacological studies, a dose ranging from 30–100 mg/kg/day of oxacillin in adults is equivalent to that of 300–900 mg/kg/day in mice. Therefore, the dose of 200 mg/kg/day of oxacillin chosen for mice in this study was lower than the equivalent doses used for humans in the clinic.

C and ECg were combined at the mass ratio of 8:1 and then mixed with sodium carboxymethylcellulose (0.5%) by milling. The dose of the C and ECg mixture used in mice was based on results of in vitro experiments, which showed that the doses of C at 128 mg/L and ECg at 16 mg/L (144 mg/L total) resulted in the best ILSMR effect. Because the quantity of body fluid in a mouse is generally 70%–80% of its body weight (average 20 g per mouse), the dose of the C and ECg mixture should be about 2 mg per mouse. Therefore, the highest dose of the C and ECg mixture used was calculated at 100 mg/kg/day in mice.

WHO-2 was inoculated into 10 mL of MH broth with seven different treatments: Broth only, 1% (v/v) dimethyl sulfoxide (DMSO), C (256 mg/L), ECg (64 mg/L), the combination of C (64 mg/L) and ECg (8 mg/L), the combination of C (128 mg/L) and ECg (16 mg/L), and the combination of C (256 mg/L) and ECg (64 mg/L). Then the mixtures were cultured for 6 h at 37 °C in a heated and shaking chamber. Then, the bacteria were centrifuged at 3500× g for 5 min to harvest the bacterial pellet. After washing with PBS (0.2 mM, pH 7.2) three times, the bacterial pellet was resuspended and the bacterial suspension was adjusted to an OD₆₀₀ of 1.0. After co-culture with daunorubicin (40 mg/L) at 37 °C for 2, 10, 20 and 30 min in the dark, daunorubicin accumulation within WHO-2 was observed by the two methods described below.

The first method was fluorospectrophotometry; daunorubicin accumulation within the bacteria was assayed in the absence or presence of C and ECg. The emission wavelength was 467 nm and the excitation wavelength was 588 nm. Additionally, as the second method, at 30 min, 0.5 mL of the bacteria was collected. After washing with PBS three times, the bacteria were fixed on glass cover slips and observed under a 510 Meta confocal microscope (Zeiss, Göttingen, Germany).

WHO-2 was inoculated into 100 mL of MH broth with six different treatments: broth, 1% DMSO (v/v), C (128 mg/L), ECg (16 mg/L), C (128 mg/L) and ECg (16 mg/L) and reserpine (20 mg/L). Bacteria were cultivated to an OD₆₀₀ of 0.6 and washed with normal saline. Then bacteria were harvested and milled with liquid nitrogen. 1 mL of RNAiso was added to completely lyse bacteria. Chloroform was added to the mixture with shaking for 15 s and stewing for 10 min. Bacteria were harvested by centrifugation at 10,000× g for 15 min in 4 °C. Isopropanol was added with stewing for 15 min. Bacteria were harvested again by centrifugation at 12,000× g for 10 min and washed by 75% (v/v) ethanol, and then 20 μL of RNase-free ddH₂O was added. The reverse transcription was performed by PrimeScript RT reagent Kit. The primers for mecA, norA, norB, norC, mepA, mdeA, sepA, qacA/B, abcA, smr and 16S RNA were listed (Table S1). The primers were added to the PCR tubes and subjected to the normal reverse transcription PCR (RT-PCR). RT-PCR products were determined using 1.0% agarose gel electrophoresis and Golden View staining. Images of the gels were analyzed using Quantity One software (Bio-Rad, Hercules, CA, USA).

Each experiment was repeated at least three times, and each data point was represented as the mean, with error bars denoting standard deviations. Differences in bacterial load in the whole blood of mice were examined by repeated measures ANOVA using Statistical Program for Social Sciences (SPSS) 16.0. A p value of 0.05 was considered significant, and a p value of 0.01 was considered highly significant.