Antibacterial activities and structure–activity relationships of a panel of 48 compounds from Kenyan plants against multidrug resistant phenotypes

In the current study forty eight compounds belonging to anthraquinones, naphthoquinones, benzoquinones, flavonoids (chalcones and polymethoxylated flavones) and diterpenoids (clerodanes and kauranes) were explored for their antimicrobial potential against a panel of sensitive and multi-drug resistant Gram-negative and Gram-positive bacteria. The minimal inhibitory concentration (MIC) determinations on the tested bacteria were conducted using modified rapid INT colorimetric assay. To evaluate the role of efflux pumps in the susceptibility of Gram-negative bacteria to the most active compounds, they were tested in the presence of phenylalanine arginine β-naphthylamide (PAβN) (at 30 µg/mL) against selected multidrug resistance (MDR) bacteria. The anthraquinone, emodin, naphthaquinone, plumbagin and the benzoquinone, rapanone were active against methicillin resistant Staphylococcus aureus (MRSA) strains of bacteria with MIC values ranging from 2 to 128 μg/mL. The structure activity relationships of benzoquinones against the MDR Gram-negative phenotype showed antibacterial activities increasing with increase in side chain length. In the chalcone series the presence of a hydroxyl group at C3′ together with a methoxy group and a second hydroxyl group in meta orientation in ring B of the chalcone skeleton appeared to be necessary for minimal activities against MRSA. In most cases, the optimal potential of the active compounds were not attained as they were extruded by bacterial efflux pumps. However, the presence of the PAβN significantly increased the antibacterial activities of emodin against Gram-negative MDR E. coli AG102, 100ATet; K. pneumoniae KP55 and KP63 by >4–64 g/mL. The antibacterial activities were substantially enhanced and were higher than those of the standard drug, chloramphenicol. These data clearly demonstrate that the active compounds, having the necessary pharmacophores for antibacterial activities, including some quinones and chalcones are substrates of bacterial efflux pumps and therefore should be combined to efflux pump inhibitors in the fight against MDR bacterial infections.


Background
Multidrug resistance (MDR) in bacteria is usually mediated by the expression of efflux pumps or porins involved in transport, by the expression of mutated genes coding for specific drug targets or specific enzymatic barriers. As a matter of fact, MDR remains a major obstacle hindering successful antibacerial chemotherapy (Alekshun and Levy 2007;Davin-Regli et al. 2008;Nikaido 2009).
One way of tackling the emergence of MDR is to diversify the chemical structures of anti-microbial drugs to which resistance has developed in order to extend their lifespan (De Clercq 2001;Poole 2001;Jeu et al. 2003). Alternatively, the compounds exhibiting modest to significant antibacterial activities against MDR phenotypes, could be used in combination with the efflux inhibitors, in order to improve on the accumulation of the drug in the cells for optimal activities (Kuete et al. 2011). It is not surprising that in response to anti-microbial resistance, major pharmaceutical companies have concentrated their efforts on improving chemotherapeutic agents in established drug classes (Taylor et al. 2002). However, with the portfolio of antimicrobials currently available, most of the lead structures of current drugs have already been explored. Today, large-scale empirical screening of synthetic, semi-synthetic and natural chemical entities in chemical libraries for anti-microbial activities is investigated (Kimberlin and Whitley 1996). The compounds of interest for the development of novel drugs include those that target different proteins or biochemical processes compared to those in current use and those that inhibit bacterial efflux pumps (Kuete et al. 2011). Natural products have been a particularly rich source of antiinfective agents, yielding, for example, the penicillins in 1940, the tetracyclines in 1948 and the glycopeptides in 1955 (Silver and Bostian 1990). In the current study the antibacterial activities of compounds from several families (anthraquinones, naphthoquinones, benzoquinones, chalcones, flavones, flavanones, clerodane, and kaurane diterpenoids) were determined against different bacterial strains expressing MDR phenotypes. Furthermore, the chemical moieties relevant for pharmacophore binding were analyzed by their structure activity relationships (SAR) studies of related compounds.
The compounds were tested for their ability to prevent the growth of MDR and reference strains of Gramnegative bacteria, alone and for some of the active compounds, in the presence of the efflux pump inhibitor (EPI), PAβN ( Table 2).
In a separate study the in vitro antimicrobial activities of plumbagin against selected microorganisms were reported to be significantly higher than the standard drug, streptomycin (Jeyachandran et al. 2009). The antibacterial potencies of different related benzoquinones were established against both Gram-negative and Grampositive bacteria strains. Compounds 6-8 with a 2-4 carbon alkyl side chain revealed similar activities against various microbes with MIC ranging from 16 to 256 µg/ mL. There was a marked improvement of antibacterial activities with increasing length of the lipophilic chain to 7 as shown by 9 exhibiting low MIC values ranging from 4 to 32 µg/mL. There was reduced antimicrobial activities with compounds 10 (C9) against most bacteria strains with the lowest activity recorded having MIC ≥ 256 µg/mL against E. coli ATCC 8739, AG102, AG100A Tet ; S. aureus A4, A11; K. pneumoniae ATCC11296, KP63; P. stuartii ATCC29916, NAE16; E. aerogenes EA27 and against all strains of MRSA. The activities of compound 11 (C13) against specific bacteria was comparable to that of 9 especially against the Gram-negative E. coli AG100A Tet , K. pneumoniae KP55, KP63 with MIC values ranging from 16 to 32 µg/mL. This compound exhibited selective activities against specific Gram-negative bacteria such as E. coli AG100A-Tet , K. pneumoniae KP55, KP63 and the Gram-positive MRSA 4, 6, 8 some of which were more potent than chloramphenicol (Table 2).
Increasing chain lengths of the alkyl group from C15 in compound 12 to C19 in 15-17 had no effect on further improving the activities of compound 11 (C13) against most bacteria. However, 17 showed selective activities against three strains of MRSA including MRSA4, 6 and 8. Reduction of the two keto groups of 15 and subsequent acetylation at C1, 2, 4, 5 positions as seen in 17 improved the antibacterial activities against specific bacteria including MRSA 4, 6, 8. The MIC values of 17 against MRSA4 and 8 strains were observed to be almost as low as those of the standard drug with MIC values of 32 (vs 8) and 64 (vs 32), respectively (Table 2) while that against MRSA6 was more active than the standard drug (16 (vs 64) µg/mL. This observation could imply that increasing the chain length up to C7 probably results in more profound antibacterial activities. Further increases in chain length had little effect on the improvement of the antibacterial activities but led to selectivity of the drugs towards specific bacteria, as seen in 11-13. The alkyldibenzoquinone, ardisiaquinone B (18) with two benzoquinone rings at the end of the alkyl chain (C16) showed improved activities as compared to the monomeric benzoquinone with a similar number of carbon atoms in the lipophilic chain. This compound showed interesting activities against all strains of E. coli, K. pneumoniae, P. staurtii and MRSA. This compound was also potent against P. aeruginosa PA124 and E. aerogenes EA27 but seemed to be inactive against Gram-positive S. aureus.
Compound 19 with two hydroxyl groups, one at C3′and the other in meta orientation at C5′and one methoxy group also in meta orientation to the hydroxyl group at C3′, without a substituent at C2′ was active against two strains of MRSA 4 and 6. However, interchanging the hydroxyl group at C3′ for a methoxy group and vice versa completely reduced the antimicrobial activities against these two MRSA strains. This may imply that the substitution pattern of the substituents on ring B where at least one hydroxyl group is placed at C3′, an additional hydroxyl groups in meta orientation (either at C1′ and 5′), a methoxy group in meta orientation to the C3′ hydroxyl group (either at C1′ and 5′) as observed in 19 (hydroxyl at C3′, one meta position to have methoxy and the other hydroxyl) is important for minimal activities against the two strains of MRSA. Compound 21 with two methoxy groups and two hydroxyl groups in meta orientation was the most active compound of the evaluated chalcones; against the four MRSA strains. The increased antibacterial activities of this compound could be attributed to the additional oxygenation at C2′ position. The basic requirement for activities against MRSA strains (hydroxyl at C3′, one meta position to have methoxy and the other hydroxyl) was observed here too. However, more studies need to be undertaken to understand the active scaffold in the chalcone skeleton. Compound 22 with one hydroxyl group at C5′and two methoxy groups at C1′ and C3′ did not exhibit any antibacterial activities against all the strains of bacteria including MRSA most probably due to the lack of the required scaffold. Compound 23 with two hydroxyl group in meta position in ring A but was lacking a methoxy group in second meta position was inactive against all bacteria tested including the MRSA strains, except for the Gram-negative E. aerogenes, ATCC13048, with minimal activity.
Most of the flavones tested were polymethoxylated making them lipophilic and therefore showing minimal or no activities against the bacteria tested. Previous studies have shown that antifungal compounds tend to be more lipophilic compared to antibacterial compounds and hence these compounds should be tested for their antifungal potential (McClure 1975;Omosa et al. 2014). Compound 34 showed minimal activities against specific bacteria strains with MIC values ranging from 128 to 256 µg/mL most probably due to increase in hydrophilicity (Omosa et al. 2014). The flavanones tested were also inactive due to their lipophilicity as shown in previous studies. The clerodanediterpenoids were inactive against all bacteria tested except for hautriwaic acid (Omosa et al. 2014) which showed minimal activity against some bacterial strains. However, its lactone did not exhibit antimicrobial activities ≤256 µg/mL, most probably because of increased lipophilicity. The presence of a free hydroxyl group at C19 in the clerodane skeleton appears to be necessary for minimal activities against S. aureus SA 11, 12; K. pneumoniae KP55, 63; P. aeroginosa PA01 and E. aerogenes ATCC 13048 strains. The kaurane type diterpenoids exhibited similar antimicrobial profile as the clerodanes which could be attributed to their lipophilicity.

Role of efflux pumps in the susceptibility of tested bacteria
In this study the compounds that attained certain threshold of activities including the anthraquinone, 2; naphthaquinone, 5; and some benzoquinones, 6-10, 13 and 18 were combined with the efflux pump inhibitor, EPI, PAβN, in order to assess the involvement of efflux in the activities of these compounds. When tested alone compound 2 showed minimal antibacterial activities against some Gram-negative MDR strains including E. coli AG100 and 102, as well as K. pneumoniae KP 55 and K63 (Table 2). However, there was substantial enhancement of activities of 2 with PAβN. The observed MICs without PAβN against E. coli AG102, 100A Tet ; K. pneumoniae KP 55 and K63 were 128, 16, 32, 128 μg/mL. With PAβN the activities improved by >4 fold, exhibiting MICs higher than the standard, chloramphenicol of 2, 4, <2 and 2 μg/ mL, respectively (Table 2). This compound which exhibited good antibacterial activities against MDR Grampositive bacteria alone also showed improved activities in combination with PAβN. However, the improvement was not as substantial as that observed against the Gramnegative bacteria. The naphthaquinone, plumbagin (5) which showed good activities even when tested alone exhibited minimal improvement with PAβN as compared to emodin (2) except against AG100A Tet with >8 fold increament. These results consistent with those obtained by Kuete et al. (2011). The antimicrobial activities against the Gram-positive MDR strains increased by >4 fold against MRSA 6, 8 and >2 fold for MRSA 4.
The benzoquinones, 6-10 and 18 showed enhancement in activity with PAβN most of which were >2 and 4 folds, with 13 having MICs < 10 μg/mL. These data clearly demonstrate that the quinones and chalcones tested are substrates of bacterial efflux pumps and should be combined to EPI in the fight against MDR bacterial infections. The obtained results are therefore consistent to data previously reported by Kuete et al. (2011), who highlighted the role of efflux pumps in the bacterial resistance to natural products.

Conclusion
In this study, various anthraquinones, naphthoquinones, benzoquinones, flavonoids (chalcones and polymethoxylated flavones) and diterpenoids (clerodanes and kauranes) were explored for their antimicrobial potential against different drug sensitive Gram-negative and Gram-positive bacteria.
The results show that the anthraquinone (2), naphthaquinone, plumbagin (5) benzoquinones (11, 12, 17, 18), chalcones (19, 21) were active against MRSA bacteria strains with MIC value ranging from 2 to 128 µg/ mL. Structure activity relationships of benzoquinones; which has not been carried out in previous studies, showed that antibacterial activities gradually increased with increasing side chain length from 2, 3, 4 for 6, 7 and 8, respectively, with optimal activities being realized with C7 (9). This study also showed that there is a minimum chain length that is required for maximum activity in the 2,5 dihydroxy-1,4-benzoquinone moieties. This studies showed that the minimum chain length required for optimal antimicrobial activities was C7 (compound 9) beyond which no marked activity improvement has been observed. Some compounds selectively inhibited the growth of specific but not all bacteria. These studies revealed that the presence of a hydroxyl group at C3′ together with a methoxy group and a second hydroxyl group in meta orientation in ring B of the chalcone skeleton appears to be necessary for minimal activities against MRSA 4 and 6 as elaborated in 19-24.
These data clearly demonstrate that the tested compounds are substrates of bacterial efflux pumps and should be combined to EPI in the fight against MDR bacterial infections. The obtained results are therefore consistent to data previously reported by Kuete et al. (2011), who highlighted the role of efflux pumps in bacterial resistance to natural products.

Bacterial strains and culture media
MDR isolates and reference strains of Escherichia coli, Enterobacter aerogenes, Pseudomonas aeruginosa, Klebsiella pneumoniae and Staphylococcus aureus are summarized in Table 1. Isolates were conserved at 4 °C and were grown on Mueller-Hinton agar for 24 h before minimal inhibitory concentration (MIC) testing. Mueller-Hinton broth (MHB) was used for the susceptibility tests (Kuete et al. 2008).

Determination of bacterial susceptibility
The MIC determinations on the tested bacteria were conducted using rapid INT colorimetric assay according to described methods (Eloff 1998) with some modifications (Kuete et al. 2007(Kuete et al. , 2009. First of all, the test samples and reference antibiotic (RA) were dissolved in dimethyl sulphoxide (DMSO)/Mueller-Hinton Broth (MHB) or DMSO/MHB broth. The final concentration of DMSO was lower than 2.5 % and does not affect the microbial growth. The solution obtained was then added to MHB, and serially diluted two fold (in a 96-wells microplate). One hundred microlitrer (100 µL) of inoculum 1.5 × 10 6 CFU/mL prepared in appropriate broth was then added. The plates were covered with a sterile plate sealer, then agitated to mix the contents of the wells using a plate shaker and incubated at 37 °C for 18 h. Wells containing adequate broth, 100 µL of inoculum and DMSO to at a final concentration of 2.5 % served as negative control. Choramphenicol was used as a RA. The minimum inhibition concentration (MIC) of samples was detected after 18 h incubation at 37 °C, following addition (40 µL) of 0.2 mg/mL of INT and incubation at 37 °C for 30 min. Viable bacteria reduced the yellow dye (TNT) to pink. MIC was defined as the sample concentration that prevented the color change of the medium and exhibited complete inhibition of microbial growth (Eloff 1998). All assays were carried out in triplicate and were repeated three times. To evaluate the role of efflux pumps in the susceptibility of Gram-negative bacteria to the most active compounds, they were tested in the presence of PAβN (at 30 µg/mL) against selected MDR phenotypes (Table 2) and MICs were determined as mentioned above.