The chemical structures of the compounds isolated from G. nobilis, O. suaveolens and B. camerunensis (Figure 1) were established by spectroscopic methods. The compounds were isolated from the stem bark of G. nobilis (1–5), the woods of O. suaveolens (6–11), the roots (12–17) and the stem bark of B. camerunensis (7, 18–22). The twenty two isolated compounds were identified as caroxanthone (1), 4-prenyl-2-(3,7-dimethyl-2,6-octadienyl)-1,3,5,8-tetrahydroxyxanthone (2), smeathxanthone A (3), 8-hydroxycudraxanthone G (4), morusignin I (5), stigma-5-en-3-ol (6), lupeol (7), evoxanthine (8), norevoxanthine (9), 1-hydoxy-2,3-dimethoxy-10-methylacridone(10), 1,3-dimethoxy-10-methylacridone (11), skimmianine (12), kokusaginine (13), montrifoline (14), 1-hydroxy-4-methoxy-10-methylacridone (15), syringaresinol (16), limonin (17), 1-hydroxy-3,6-dimethoxy-8-methylxanthone (18), xanthoxyletin (19), imperatorin (20), scoparone (21) and umbelliferone (22) [11–25]. Amongst the twenty-two compounds were three terpenoids (6, 7 and 17), eleven phenolic compounds (1–5, 16, 18–22), and eight alkaloids (10–17). The isolated terpenoids were steroid (6), triterpenoid (7), and limonoid (17) whilst the alkaloids included five acridones (8–11, 15) and three furanoquinolines (12–14). The phenolics obtained herein were six xanthones (1–5, 18), one lignan (16) and four coumarins (19–22). The isolation and identification of compounds 1–5 from G. nobilis was previously reported . The occurrence of alkaloids and terpernoids from O. suaveolens was reported [5, 6], and their isolation in the present study is in consistence with previous reports. The occurrence of coumarins, quinoline alkaloids, and free aliphatic acids was also reported in B. camerunensis. However, in this study, only coumarins were isolated. The crude extracts as well as the isolated compounds [excluding compounds 6, 7, 12 (known to have low or no antimicrobial activity), and 14 (isolated in very low quantities)] were tested for the antibacterial activities against Gram-negative bacteria and M. tuberculosis and the results are summarized in Tables 1 and 2.
The results of the MIC determinations (Table 1) showed that the crude extract from G. nobilis (GNB) was the most active amongst the studied extracts, its inhibitory effects being noted on 12 of the 14 tested bacteria. GNB also exhibited the best activity against M. tuberculosis ATCC 27294 (MIC of 128 μg/ml) and E. coli ATCC10536 (MIC of 64 μg/ml) than OSR, OSW and BCB. The inhibitory effects of the extracts OSR and OSW from O. suaveolens were noted on 9/14 and 7/14 studied bacteria respectively, meanwhile that of the extract BCB of B. camerunensis was observed on 4/14 pathogens tested. Interestingly, compound 2 isolated from the most active extract GNB, also exhibited the best activity, preventing the growth of all the fourteen tested microorganisms. The lowest MIC obtained with compound 2 was 8 μg/ml against M. tuberculosis ATCC 27294 and the clinical MTCS2 strains. It is noteworthy that compound 2 was more active than chloramphenicol on two Gram-negative MDR bacteria, namely E. coli AG102 and E. aerogenes CM64 (Table 1). Other compounds showed selective activities, their effects being noted on 1/14 tested bacteria for compounds 16, 17, 19 and 20; 2/14 for 1; 3/14 for 18; 4/14 for 9 and 21; 5/14 for 22; 7/14 for 3, 4, 10 and 11; 9/14 for 8; and 11/14 for 5 and 13. The results of the MBC determinations (Table 2) also showed the activities of the studied samples on some of the tested microorganisms. As observed with MIC data (Table 1), the lowest MBC value (Table 2) was recorded with compound 2 (16 μg/ml) against M. tuberculosis ATCC 27294.
Phytochemicals are routinely classified as antimicrobials on the basis of susceptibility tests that produce MIC in the range of 100 to 1000 μg/ml . The activity is considered to be significant if MIC values are below 100 μg/ml for crude extract and moderate when the MICs vary from 100 to 625 μg/ml [34, 35]. Also, the activity of compounds is considered to be significant when the MIC is below 10 μg/ml, and moderate when such values vary between 10 and 100 μg/ml [34, 35]. On the basis of such criteria, the activity of the studied crude extracts can mostly be considered as moderate, though a significant effect was observed with GNB on E. coli ATCC strains. Compound 2 was significantly active against M. tuberculosis ATCC 27294 and MTCS2 strains. However its activities can also be considered as moderate against the majority of the bacteria tested. All the tested compounds were active on at least one of the studied microorganisms, and their presence can explain the activity of the crude extracts. Nonetheless, it can be observed that the activity of GNB was not detected on both P. stuartii ATCC NAE16 and P. aeruginosa PA124, while the extract yielded at least one active compound on these bacteria (Table 1). This can be explained by the fact that the activity of the crude extract does not only depend on the presence of the active compounds, but is also influenced by the quantity and/or possible interaction with other constituents of the plant. This observation can also be applied when carefully analysing the activities of the crude extracts from O. suaveolens and B. camerunensis and their constituents (Table 1). The activity of the crude extracts and compounds studied herein (and mostly compound 2) can be considered interesting when regarding the medical importance of the studied bacteria. In fact, the clinical MDR Gram-negative bacteria tested express active efflux pumps and are involved in many therapeutic failures . To the best of our knowledge, the antibacterial activities of the extracts of G. nobilis, O. suaveolens and B. camerunensis as well as those of most of the studied compounds are being reported for the first time. Nevertheless, some of the isolated compounds were reported for their antibacterial properties. In effect, lupeol (7) is known to have low antibacterial activities, the lowest MIC value obtained against Enterococcus faecalis ATCC 29212 being 63 μg/ml . The activity of compound 7 was also not detected against the sensitive Mycobacterium smegmatis when it was tested at a concentration up to 312 μg/ml  and consequently, this compound was not tested again in this work. Previously we also reported the moderate activity of the alkaloid 13 on some sensitive Gram-negative bacteria as well as it low effect against M. smegmatis. This compound was not more active against MDR bacteria as observed in the present work, confirming its low antibacterial potential. It has also been demonstrated that compounds 3 (active only on two of 21 tested microorganisms)  and 16 (inhibition zone diameters for 10 μl discs at 104 ppm reported as 0.0; 0.2; 0.4 and 0.5 mm against K. pneumoniae, S. aureus, Pseudomonas syringae and Bacillus subtilis respectively)  have low antibacterial activities. Such reports are in accordance with the results obtained in the present study.
Though compounds 17 (limonin) and 20 (imperatonin) showed poor activities as reported herein, their antilisterial inhibitory effects were found moderate, the MIC values obtained against five Listeria monocytogenes ATCC strains varying between 15.62-31.25 μg/ml .
Compounds 19 (xanthoxyletin) and 22 (umbelliferone) were also found inactive against M. tuberculosis and poorly active against Gram-negative bacteria, consolidating the previously reported data [41, 42]. It is noteworthy that compound 12 was reported not to have any bacterial activity  and was not tested in the present work.
When regarding the structure-activity relationship, it appears that coumarins had the lowest activities, none of them been active against M. tuberculosis. The tested tetranortriterpenoid, compound 17 also showed very weak antibacterial activities. These results are in consistence with previous studies, showing the low antibacterial activities of terpenoids and coumarins against bacteria expressing MDR phenotype . Alkaloids also showed low antibacterial activities. However moderate inhibitory effects were noted with 10 (1-hydoxy-2,3-dimethoxy-10-methylacridone) and 13 (kokusaginine) respectively against E. coli ATTC and AG100 strains (Table 1). Amongst alkaloids, compound 13, one of the three isolated furanoquinolines, showed the best spectrum of activity (contrary to the five acridone alkaloids) and was found active on both M. tuberculosis and Gram-negative bacteria. Within the acridone alkaloids, and when comparing the effects of compounds 10 and 15 [antibacterial spectra (7/14 of the tested bacteria for 10 and 4/14 for 15); lowest MIC value (64 μg/ml for 10 and 128 μg/ml for 15)], it appears that the presence of methoxyl- (−OCH3) group in C2 (compound 10) increases the bacterial susceptibility (Figure 1, Table 1). A comparison of the activities of compounds 8 and 9 [the presence of –OCH3 group in C1 (compound 8) instead of –OH group (compound 9)] on one hand, and those of 11 and 15 [the presence of –OCH3 group in C1 (compound 11) instead of the hydroxyl (−OH) group (compound 15)] on another hand seems to confirm the fact that the natural substitution of –OCH3 by –OH in the studied alkaloids increases their activities (Figure 1, Table 1). The best antibacterial activities were recorded with xanthones. Within the studied xanthones and when regarding the activities of compounds 2 and 3, it appears that the presence of additional prenyl- group in C4 significantly increases the antibacterial activity (Figure 1, Table 1). Between compounds 4 and 5, it can also be deduced that the cyclisation increase the activity (Figure 1, Table 1).