The vegetal materials used in this study were the same used in our previous study [27]. They were barks and seeds of G. lucida Vesque, leaves of C. citratus (DC.) Stapf, L. vulgare, M. oleifera Lam and V. amygdalina Delile; barks of C. officinalis and E. chlorantha Oliv; and Leaves and seeds of A. indica (Neem).

Three reference cultures (purchased from American Type Culture Collection) were used to screen the antimicrobial activity of the different vegetal materials. They were S. aureus ATCC 6538 as Gram positive model, E. coliATCC 25922 as Gram negative model and C. albicans ATCC 10231 as fungi model. To assess the synergy between common antibiotics and the extracts, 11 strains of uropathogenic bacteria provided by the Department of Microbiology and Virology of the Peoples’ Friendship University of Russi were used. These strains were A. xylosoxidans 4892, C. freundii 426, E. avium 1669, E. coli 1449, K. oxytoca 3003, K. rizophilia. 1542, M. catarrhalis 4222, M. morganii 1543, P. aeruginosa 3057, S. aureus 1449 and S. agalactiae 3984. From frozen stock, the different strains were subcultured twice in BHI broth for bacteria and SAB broth for C. albicans ATCC 10231.

Brain Heart Infusion Broth (BHIB), Muller Hinton Agar (MHA), and Sabouraud Dextrose Broth (SDB) were purchased from HiMedia™ Laboratories Pvt. Ltd., India while Dimethyl sulfoxide (DMSO) was purchased from BDH Laboratories, VWR International Ltd., USA.

Plant extracts were prepared following the protocol described in previous studies by Mbarga et al. [27]. Briefly, fifty grams of plant material was weighed and introduced into separate conical flasks containing 450 mL of the solvents which were distilled water and ethanol/water (80:20, v/v), The mixture were shaken at 200 rpm for 24 h and 25 °C (Heidolph Inkubator 1000 coupled with Heidolph Unimax 1010, Germany), filtered (Whatman n°1), and concentrated at 40 °C using a rotary evaporator (IKA RV8, Germany) The semi-solid extracts obtained were kept at 4 °C for analyses.

An EDX-7000 Shimadzu energy dispersive X-ray fluorescence (XRF) spectrometer was used to identify the minerals present in the E. chlorantha extract . The equipment settings were as followed: range of measured elements — 11Na - 92U; X-ray generator — a tube with a Rh-anode, air cooled; voltage 4–50 kV, current 1–1000 $\mu$ A; irradiated area — a circle of 10 mm in diameter; silicon drift detector (SDD), counting method — a digital counting filter; the content of elements according to the value of intensity; automatic change of filters emitting the wavelengths of the corresponding elements; chamber size 300 mm $\times$ 275 mm $\times$ 100 mm. The X-ray fluorescence spectrum for each measurement was recorded at the same device settings: mylar film, collimator width - 10 mm, exposure time - 100 sec, atmosphere - air; the number of repeated measurements for one sample n = 3. To process the obtained results, we used the OriginPro 2017 software (OriginLab, Northampton, MA, USA). The results obtained using the XRF method were presented in values of irradiation intensity expressed in cps/ $\mu$ A.

All the experiments were carried out at least in triplicate. The statistical significance was set at p $\le$ 0.05. T-test was carried out using the statistical software XLSTAT 2020 (Addinsof Inc., New York, USA) and the graphs were plotted by SigmaPlot 12.5 (Systat Software, San Jose, CA, USA).

Fig. 2 presents the inhibition diameters of the different extracts on the tested pathogens. All extracts were not active against the pathogens. It is the case of the extracts (both ethanolic and aqueous extracts) from C. officinalis and G. lucida leaves. The rest of extracts were actives with inhibition diameters ranging from 5 to 36 mm.

Fig. 2.

Inhibition diameter (mm) resulting from the screening of antimicrobial activity by well diffusion method with 100 mg/mL of each extract.

Taking into consideration the extraction solvents, the highest inhibition diameters were mainly recorded with ethanol as solvent. Ethanol therefore appears as the solvent which extracted more antimicrobial compounds compared to water although the extraction yields were globally more important with water as solvent [27]. This observation could be ascribed to the insoluble nature of metabolites extracted with ethanol as solvent opposite to water. Indeed, most of bioactive compounds endowed with antimicrobial activity such as flavonoids, polyphenols, tannins and alkaloids are generally insoluble in water [35, 36, 37]. In a study conducted by Mouafo et al. [35], it was highlighted that ethanol extracted more antimicrobial compounds from plant materials opposite to water. A similar conclusion was also stated by Evbuomwan et al. [38].

With regards to the part of the plant material, extracts from bark were more actives independently of the pathogens and the extraction solvents. The highest activities on both bacterial and yeast strains were noticed with bark from E. chloranta. This could be attributed to the presence of high amount of antimicrobial alkaloids such as protoberberines (berberine, canadine, palmatine, jatrorrhizine, columbamine and pseudocolumbamine), phenanthrene alkaloids (atherosperminine and argentinine) and aporphines (7-hydroxydehydronuciferine and 7-hydroxydehydronornuciferine) in that plant as reported in the literature [26]. Several studies also highlighted the interesting antibacterial and antifungal properties of dried and fresh barks from E. chlorantha [10, 39, 40]. The extracts from E. chloranta therefore appears as a source of antimicrobials with a broad-spectrum activity. Similarly, ethanolic extracts of L. vulgare and V. amygdalina showed a broad-spectrum antimicrobial activity against all the tested pathogens. This could be ascribed to the richness of these extracts in phytochemicals such as saponins, sesquiterpenes, flavonoids, steroid glycosides, and lactones [20] for which recent studies reported its antibacterial activity [21].

Amongst leaves extracts, those derived from M. oleifera were more actives against the bacterial strains E. coli ATCC 25922 and S. aureus ATCC 6538 while extract from L. vulgare leaves was more active against C. albicans ATCC 10231. This result could be explained by the great variability of the phytochemical composition of the two vegetal materials and thus, the possible different action mechanisms against microorganisms.

When the tested pathogens are considered, it appears from Fig. 2 that S. aureus ATCC 6538 was more sensitive than E. coli ATCC 25922. This observation can be explained by the nature and composition of both cell wall and membranes which differ between Gram positive and Gram-negative bacteria. In fact, Gram-negative bacteria possess a lipopolysaccharides layer in their external membrane. This layer act as a barrier against the permeability antimicrobials [41].

A surprising observation was noticed in this study as the yeast C. albicans ATCC 10231 was more sensitive to the ethanolic and aqueous extracts of both bark and leaves extracts from E. chlorantha compared to the bacterial strains S. aureus ATCC 6538 and E. coli ATCC 25922. This result might arise from the antimicrobial action mechanisms of the bioactive compounds found in that plant. In fact, eukaryotic cells are known for their ability to resist to several antimicrobials as opposite to prokaryotic cells, they possess less phospholipids in their membranes. Phospholipids due to their anionic nature are mostly involved in the preliminary interaction with antimicrobial which will ease their penetration into cells [42, 43].

All the extracts inhibited the pathogens with MIC values which varied significantly from one plant material to another. As previously observed with qualitative tests, the extract of E. chloranta was the most active independently of the solvents and the tested pathogens. As shown in Table 2, the ethanolic extract of E. chloranta with MIC value lower than 2 mg/mL against the three pathogens was the most active extract. This important activity could result from the presence of alkaloids such as palmatine, coloumbamine and jatrorrhizine in its composition [44]. Indeed, these compounds have the ability to penetrate cells and intercalate DNA of microorganism leading to their death [45]. MIC values higher than 256 mg/mL were observed with some aqueous and ethanolic extracts against the different pathogens. This observation suggests that further analysis at concentrations higher than 256 should be performed in order to quantify the antimicrobial activity of these extracts. Moreover, extracts which have showed no activity in well diffusion qualitative test were active in liquid medium against the pathogens. Thus, suggesting that these extracts might contain antimicrobial compounds which cannot diffuse in the Muller Hinton agar.

With water as solvent, the most important activity was recorded with extract from E. chloranta for which MIC of 8 mg/mL was noticed against S. aureus ATCC 6538 and E. coli ATCC 25922, and 4 mg/mL against C. albicans ATCC 10231. These activities were higher compared to those reported by Adesokan et al. [46]. They obtained with aqueous extract of E. chloranta, MIC values of 25 and 100 mg/mL against S. aureus and E. coli, respectively. This can be ascribed to several factors which influence the plant composition such as climate and soil composition, as well as to the tested strains.

The majority of extracts showed MIC values higher or equal to 256 mg/mL against the yeast strain C. albicans ATCC 10231 independently of the extraction solvent. Only ethanolic (MIC = 8 mg/mL) and aqueous (MIC = 32 mg/mL) extracts of L. vulgare leaves, and the ethanolic (MIC 2 mg/mL) and aqueous (MIC = 4 mg/mL) extracts of E. chloranta bark. This observation suggests that C. albicans ATCC 10231 was the most resistant strain to the different extracts independently of the solvent. This resistance could be ascribed to their membrane composition which is different to those of bacteria. In fact, the higher amount of anionic phospholipids in the membrane of bacteria ease their interaction with antimicrobial compounds and thus increase their sensitivity [42, 43].

An observation of Table 2 revealed that S. aureus ATCC 6538 was the most sensitive strain as lower MIC values of most extracts were generally recorded against that strain. Thus, it clearly appears that the bacterial cell walls and membranes are involved in the antimicrobial activity mechanism of these extracts. This conclusion is different to that stated by Etame et al. [10] who found no significant difference in the MIC values of plant extracts against Gram positive and Gram-negative bacteria.

According to the classification established by Kuete [47] and Kuete and Efferth [48], the different extracts could be considered as deserving a weak antimicrobial activity independently of the extraction solvent and the tested strain as they scored MIC value higher than 0.625 mg/mL.

Minimum bactericidal concentration (MBC) and minimum fungicidal concentration (MIC) of the different extracts against three pathogenic strains were assessed and results are presented in Table 2. The MBC values ranged from 4 to more than 256 mg/mL. The strongest MBC value (4 mg/mL) against C. albicans ATCC 10231 and S. aureus ATCC 6538 was obtained with the ethanolic extract of E. chloranta bark. Against E. coli ATCC 25922, the strongest MBC value of 8 mg/mL was recorded with the same ethanolic extract of E. chloranta bark. Globally, lowest MBC values were recorded against S. aureus ATCC 6538, thus confirming its higher sensitivity to the different plant extracts.

The MBC values against E. coli ATCC 25922 and S. aureus ATCC 6538 obtained in this study with the aqueous extracts of E. chloranta were lower than that reported by Adesokan et al. [46]. The authors found with aqueous extract of E. chloranta, MBC values of 90 and 130 mg/mL against S. aureus and E. coli, respectively. In the same way, the MBC values of the ethanolic extract of V. amygdalina leaves against E. coli ATCC 25922 and S. aureus ATCC 6538 were lower than the 100 and 200 mg/mL obtained respectively against E. coli and S. aureus by Evbuomwan et al. [38] using the ethanolic extract from the same plant. This difference could be explained by the variability of phytochemical profile of the plant according to the geographical origin. Besides, the fact that microbial strain developed different resistance mechanism to antimicrobials as highlighted by several authors in the literature [49, 50] could also explained the variability in the MBC values.

It is established in the literature that an antimicrobial compound is considered as bactericidal/fungicidal against a microbial strain when the ratio MBC/MIC or MIC/MIC is $\le$ 4 [51, 52]. Based on this classification, the ethanolic extract of C. officinalis bark and M. oleifera leaves, the aqueous extract of G. lucida seeds, E. chloranta leaves and bark, the aqueous and ethanolic extracts of V. amygdalina leaves, G. lucida bark, A. indica seeds, and L. vulgare leaves can be considered as bactericidal against S. aureus ATCC 6538. The ethanolic extract of V. amygdalina leaves, the aqueous extract of G. lucida seeds, the ethanolic and aqueous extracts of G. lucida bark and L. vulgare leaves can be considered as bactericidal against E. coli ATCC 25922. The ethanolic extract of G. lucida seeds and L. vulgare leaves, and the aqueous extracts of E. chloranta bark can be considered as fungicidal against C. albicans ATCC 10231.

The use of combination therapy has been suggested as a new approach to improve the efficacy of antimicrobial agents by screening crude extracts from medicinal plants with good indications for use in combination with antibiotics [53]. The checkboard method was applied to assess the synergy between some conventional antibiotics and plant extracts which showed a valid MIC. After determining the MIC of the plant materials (Table 2), we determined the MICs and MBCs of the antibiotics (ampicillin, benzylpenicillin, cefazolin, ciprofloxacin, nitrofurantoin, and kanamycin) against S. aureus ATCC 6538 and E. coli ATCC 25922. As shown in Table 3, the MICs of the different antibiotics varied from 4–64 $\mu$ g/mL while the MBCs varied from 4–256 $\mu$ g/mL.

Antibiotics with high and quantified MICs values against the both pathogens (ampicillin, nitrofurantoin, and kanamycin) were selected for modulation assays. Similarly, regarding the plant extracts, we decided to work with extracts with low and determined MICs. Thus, it was performed a modulation of ampicillin, nitrofurantoin, and kanamycin with aqueous extract of E. chlorantha bark and ethanolic extract of G. lucida seed, A. indica seed and L. vulgare leaves. As shown in Table 4, the fractional inhibitory concentration (FIC) ranged from 0.125 to 0.750. No antagonism (FIC $>$ 4) or indifference (1 $\le$ FIC $\le$ 4) was noted between the extracts and the antibiotics. However, it was found an additional effect (0.5 $\le$ FIC $\le$ 1) in some plant + antibiotic combinations such as Kanamycin + (G. lucida seed or A. Indica seed) which had a FIC index of 0.625 against S. aureus. Regarding E. coli, an additional effect (FIC index = 0.750) in the combinations Kanamycin + G. lucida seed and Kanamycin + L. vulgare leaves, was also found. Except for the 4 cases of combinations above mentioned, all the other plant + antibiotic combinations exhibited synergistic effect (FIC $\le$ 0.5) against the two test microorganisms. The lower is the FIC index, better is the synergy [53]. The best synergies were therefore found with E. chloranta bark which well-modulated Kanamycin (FIC = 0.125 against S. aureus and 0.250 against E. coli), nitrofurantoin (FIC = 0.250 against S. aureus and 0.188 against E. coli) and ampicillin (FIC = 0.125 against E. coli). A good synergy between A. indica seed and nitrofurantoin (FIC = 0.125) was also noticed. The results gathered from this study prove that common antibiotics such as ampicillin, nitrofurantoin and kanamycin could be successfully combined with plant extracts and demonstrate better antimicrobial activity materialized here by good FIC and reduction of MIC (more than 50% in all combinations). It has been reported that some plant-derived compounds can enhance the in vitro activity of some antibiotics by directly attacking the same site as the antibiotic or multiple sites at once [54]. For the 3 antibiotics used for modulation in liquid medium, it is well known that kanamycin inhibits protein synthesis by tightly binding to the conserved A site of 16S rRNA in the 30S ribosomal subunit [55]; Ampicillin acts as an irreversible inhibitor of transpeptidase, an enzyme essential the cell wall synthesis [56], and nitrofurantoin is a broad-spectrum antibacterial agent, active against the majority pathogens [57]. Thus, the protoberberins and phenanthrene alkaloids which have been reported as being the major constituents of E. chloranta bark [26] and which inhibit protein synthesis [58] may have had a cumulative effect when combined with kanamycin or an inhibitory effect on two targets (Protein synthesis mechanism and cell wall synthesis) at the same time when combined with ampicillin, which therefore explains the good modulation between E. chlorantha bark and kanamycin or ampicillin. Similarly, the broad-spectrum antibacterial properties of Azadirachtin which is one of the major constituents of A. indica [59, 60] may also explain the 16-fold reduction of the MIC of nitrofurantoin (from 32 to 2 $\mu$ g/mL against E. coli) when associated with A. indica seed. Finally, although the combinatory assays gave positive results against both Gram + and Gram - models, further studies are needed to assess the bonds formed between the extracts and the antibiotics tested and their implication in the mechanism of action. Similarly, further preclinical and clinical trials are required to evaluate the cytotoxicity and safety issues of these combinations before they can be recommended as antimicrobials drug in the fight against antibiotic resistance issue.

The solid-medium modulation test using commercial antibiotic discs is a less complex means for synergy testing. In this study, after having observed in liquid medium (checkboard method) that most of the extracts had a synergistic effect with antibiotics on non-resistant bacteria, we undertook to perform modulation tests in solid medium to assess the extent to which extracts from the tested plants could potentially enhance the performance of conventional antibiotics against a wide range of resistant bacteria. So, we started by determining the effectiveness of antibiotic discs alone. The sensitivity of the eleven uropathogenic bacteria used in this study to eleven (11) antibiotics was determined (Table 5) and the multidrug resistance (MDR) index of each bacterium was calculated. No bacteria were resistant to imipenem and amoxiclav. Regarding the other antibiotics, we found resistance in 10/11 bacteria to ampicillin, 8/11 to trimethoprim and tetracyclin, 6/11 to cefazolin + clavulanic acid, 5/11 to ceftazidime, 4/11 to nitrofurantoin and 1/11 to ceftriaxone and ciprofloxacin. The highest MDR index (0.54) was found in E. coli 1449 which was resistant to ampicillin, ceftazidime, cefazoline + clavulanic acid, ceftriaxone, tetracycline and trimethoprim. St. agalactiae 3984 and K. rizophilia 1542 scored the same MDR index of 0.45. K. rizophilia 3984 was resistant to Ampicillin, Ceftazidime, Cefazoline + clavulanic, Nitrofurantoin and tetracycline while St. agalactiae 3984 was resistant to trimethoprim, ampicillin, ceftazidime, cefazoline + clavulanic and tetracycline The lowest MDR index (0.27) was recorded with C. freundii 426 and S. aureus 1449. This result is consistent with those obtained by other authors on the resistance of clinical strains to antibiotics [61, 62, 63].

Various means have been implemented in recent years to provide effective solutions to the antibioresistance issue, and the studies carried out target bacteria that are resistant or not. Here we focused on evaluating the modulatory effect of ethanolic extracts of plant materials on antibiotics which presented an inhibition diameter lower than 20 mm (Table 5) against uropathogenic bacteria. Tables 6,7,8,9,10 present, respectively, in the form of increase in fold area (IFA), the modulating effect of plant extracts with ampicillin (AMP) (Table 6), ceftazidime (CAZ) (Table 7), tetracycline (TE) (Table 8), nitrofurantoin (NIT) (Table 9) and trimethoprim (TR) (Table 10). As with synergy tests using the checkboard method on non-resistant bacteria (Table 5), we found that E. chlorantha bark (ECB) (indenpendently of the solvent used) had the best modulating properties on most of the antibiotics tested. Its IFA was significantly different (p 0.05) to that obtained with the other extracts. The ECB-AMP combination induced an increase in inhibition diameters of more than 35% in all bacteria tested and made P. aeruginosa 3057, K. rizophilia 1542 and M. catarrhalis 4222 more susceptible to AMP with respective AFIs of 8.00, 3.84 and 3.00. The ECB-CAZ combination has also demonstrated an interesting increase of the susceptibility to CAZ and we found IFAs of 3.00; 3.00, 4.90 and 6.11 respectively against K. rizophilia 1542, St. agalactiae 3984, E. coli 1449, and S. aureus 1449 (Fig. 3A). Similarly, E. avium 1669, K. oxytoca 3003 (Fig. 3B), M. morganii 1543 and St. agalactiae 3984 which were resistant to tetracycline became susceptible to that antibiotic after combination with ECB.

Fig. 3.

Modulation of ceftazidime (CAZ) and tetracycline (TE) with E. Chlorantha bark (ECB) respectively against S. aureus 1449 (A) and K. oxytoca 3003 (B).

The ECB-TR and ECB-NIT combinations also showed a strong increase in activity especially against Ac. Xylosoxidans 4892 (both) K. rizophilia 1542 (only ECB-NIT) P. aeruginosa 3057 (both), and K. oxytoca 3003 (only ECB-TR). Otherwise, L. vulgare leaves (LVL) and G. lucida seed (GLS) have also demonstrated some promising results when combined with certain antibiotics. For example, the combinations LVL-AMP, LVL-CAZ, LVL-CAZ, LVL-TE, LVL-TE, LVL-TE, LVL-TE, LVL-TR, LVL-TR and LVL-TR were respectively very active against P. aeruginosa 3057, E. coli 1449, P. aeruginosa 3057, E. avium 1669, K. oxytoca 3003, M. morganii 1543, St. agalactiae 3984, E. avium 1669, P. aeruginosa 3057 and St. agalactiae 3984. Interestingly, C. citratus well-modulated trimethoprim while with other antibiotics the activity was lower compared to ECB, LVL and GLS.

Generally, the activity of antibiotics to which bacteria were resistant was interestingly increased with E. chlorantha bark, L. vulgare leaves and G. lucida seed while the action of ATB-extract combinations was moderate with G. lucida bark and weak with the other extracts. Besides mechanisms previously mentioned for checkboard method, another explanation of the modulation effect observed in this study in solid media could be the fact that some plant-derived compounds enhance the activity of antimicrobial compounds by inhibiting MDR efflux systems in bacteria [54]. Bacterial efflux pumps are responsible for a significant level of resistance to antibiotics in pathogenic bacteria [5]. Indeed, efflux pumps allow bacteria to flush antibiotics out of bacterial cells and therefore reduce their sensitivity to conventional antibiotics [5]. It is likely that ethanolic extracts from our plants may contain potential efflux pump inhibitors which are likely to be broad considering that the synergistic effect of the extract was observed on both Gram-positive and Gram-negative organisms. Several studies reported the isolation of some broad-spectrum efflux pump inhibitors from plant materials [5, 54, 62, 64]. A good example of this is the work of Smith et al. [64] who reported one efflux inhibitor (ferruginol) from the cones of Chamaecyparis lawso-niana, which inhibited the activity of the quinolone resistance pump (NorA), the tetracycline resistance pump (TetK) and the erythromycin resistance pump (MsrA ) in Staphylococcus aureus.

Finally, the management of bacterial infections should be done by aggressive empiric therapy with at least two antimicrobial agents [54]. Empiric combination antimicrobial therapy is usually applied to expand antibacterial spectrum and reduce the selection of resistant mutants during treatment [54]. In addition, combinations of agents that exhibit synergy or partial synergy could potentially improve the outcome for patients with difficult to treat infections [54]. However, this approach while viable, has limitations because it might not be effective for a long period of time due the possible alteration in the susceptibility of bacteria. Therefore, the development of new classes of antimicrobial compounds is of significant importance.