*3.1. Chemistry*

Compounds **1**–**5**, **10**, **12**–**17**, **20**, **22,** and **25** (Figure 2) were isolated, purified, and characterized in our laboratory from two *Maytenus* species, *M. blepharodes* and *M. canariensis*, as previously described [20, 21,27]. Derivatives **11**, **18**, **19**, **21**, **23**, **24,** and **26**, achieved by acetylation or methylation from natural compounds, are described by González et al. [20] and Rodríguez et al. [21]. Moreover, derivatives **6**–**9** were prepared from pristimerin (Scheme 1). The semisynthesis of **6** was achieved by acetylation of pristimerin, and derivatives **7**–**9** were obtained by catalytic reduction and further acetylation of pristimerin, as described in the experimental section. Their structures were greatly aided by comparison of their spectroscopic data with those previously reported for related compounds **24**–**26** [21]. Even so, a complete set of two-dimensional (2D) NMR spectra (COSY, ROESY, HSQC and HMBC) was acquired for the new derivatives to gain the complete assignment of the 1H and 13C NMR resonances (see Experimental Section, Figures S1–S4). Derivative **7** was identified as the previously described 2,3-diacetylpristimerol [31], however, its 1H and 13C NMR assignments have not been previously reported. Moreover, derivatives **6**, **8,** and **9** are described herein for the first time, and their structures were elucidated as described below.

**Figure 2.** Natural and derivatives phenolic nor-triterpenes included in the SAR studies. SAR: structure– activity relationship.

**Scheme 1.** Synthesis of analogues **6**–**9** from pristimerin.

Derivative **6** was obtained as a pale yellow amorphous solid with [ α]<sup>20</sup> D − 39.7 (c 0.15, MeOH). The molecular formula, C34 H46 O7, was established by EI-HRMS ( *m*/*z* 556.3048 [M+], calcd. 556.3032). The IR absorptions revealed the presence of hydroxyl (3517 cm<sup>−</sup>1) and ester carbonyl (1770 and 1730 cm<sup>−</sup>1) groups and an aromatic ring (1459 and 758 cm<sup>−</sup>1). Two acetoxy and a carboxylic methyl ester group in **6** were evident from the 1H NMR signals at δ 2.29 (3H, s), 2.33 (3H, s), and 3.54 (3H, s), in combination with those in the 13C NMR (δ 20.6, 21.0, 51.9, 168.4, and 168.7). Besides the carbons of acetoxy and methoxy groups, the 13C NMR and DEPT data revealed 29 carbon resonances, including one carboxylic carbon, one aromatic ring, one trisubstituted double bond, one oxymethine carbon, five sp<sup>3</sup> quaternary carbons, one methine, seven methylenes, and six methyls. Analysis of 2D NMR data, including HSQC, 1H-1H COSY, ROESY, and HMBC experiments, and comparison with data reported for pristimerol (**24**) [30] indicated that **6** is a phenolic triterpene with a pristimerin framework. In particular, HMBC correlations of Me-23 with C-3, C-4, and C-5; H-1 with C-2, C-3, C-5, C-9, and C-10; H-6 with C-4, C-5, C-8, and C-10, and correlations of Me-25 with C-8, C-9, and C-10 established the structure of A and B rings. The regiosubstitution of the acetate groups and the relative configuration of H-6 were deduced by a ROESY experiment, showing ROEs (Rotating-frame Overhauser E ffects) between H-6/OMe-25, H-1/OAc-2, and Me-23/OAc-3 (Figure 3). Thus, the structure of **6** was established as 6 α-hydroxy-2,3-diacetoxy-pristimerol.

**Figure 3.** Selected HMBC (single arrow) and ROE (doubled arrow) correlations for compound **6**.

Derivative **7** was assigned the molecular formula C34 H46 O6, determined by EI-HRMS. The mass spectrum exhibited peaks characteristic to acetate groups (M+-15-60, *m*/*z* 475, CH3, CH3COOH). This was confirmed by the 1H and 13C NMR spectra, which included signals for two acetyl groups (δH 2.28 (s), 2.31 (s), δC 20.4 (q), 20.7 (q), 168.4 (q), 168.7 (s)). Its 1H NMR spectrum showed signals for six methyl groups (δH 0.58, 1.08, 1.17, 1.22, 1.35, and 2.06), a methoxy group at δH 3.55, a vinylic proton at δH 5.73 (d, *J* = 5.1 Hz), and an aromatic proton at δH 7.00 (s). The 13C NMR spectrum displayed 34 signals, which were assigned to one methoxy, eight methyls, eight methylenes, three methines, and fourteen quaternary carbons, including six *sp*<sup>2</sup> and three carboxylic carbons. Extensive study of the 2,3*J*C-H correlations (HMBC) allowed us to establish the structure of A and B rings and build the nor-triterpene skeleton for **7**. The regiosubstitution of the acetate groups was deduced by a ROESY experiment, showing ROE e ffects between H-1/OAc-2 and Me-23/OAc-3. This data established the structure of **7** as 2,3-diacetoxy-pristimerol.

Compounds **8** and **9**, both with the molecular formula C34 H48 O6 (EI-HRMS), di ffer from that of **7** by the presence of two additional hydrogen atoms. The most remarkable di fference in their NMR spectra, when compared to that of compound **7**, was the absence of the double bond C-7–C-8. Thus, in the 13C NMR spectra of **8** and **9**, the signals assigned to C-7 and C-8 resonated in the region of aliphatic carbons (δC 18.2 and 43.5 in **8,** and δC 30.5 and 56.0 in **9**). Moreover, di fferences between **8** and **9** were also observed for the chemical shifts of Me-25 (δH 1.60, δC 27.3 in **8**, and δH 1.46, δC 36.8 in **9**) and

Me-26 (δH 0.88, δC 15.9 in **8**, and δH 1.21, δC 25.8 in **9**). Analysis of 2D COSY, HSQC, and HMBC spectra allowed us to define their core structures. The ROE effects of H-8/Me-25 and Me-26 observed in a ROESY experiment for compound **8**, and those of H-8/Me-27 observed for compound **9**, further supported the stereochemical assignment for epimers **8** and **9**. Their structures were also confirmed by chemical correlations; hydrolysis of **8** and **9** yielded compounds whose spectroscopic data were identical to those previously reported for 6-deoxoblepharodol (**25**) and 8-*epi*-6-deoxoblepharodol (**26**) [21], respectively. Consequently, the structure of compounds **8** and **9** were deduced to be 2,3-diacetyl-6-deoxoblepharodol and 2,3-diacetyl-8-*epi*-6-deoxoblepharodol, respectively.

## *3.2. Antimicrobial Evaluation*

Compounds **1**–**9** were tested on Gram-positive and Gram-negative bacteria, and the yeas<sup>t</sup> *Candida albicans* by the broth microdilution method. Moreover, compounds **24**–**26**, closely related to **6**–**9**, were also assayed against *B. alvei*, *B. megaterium,* and *B. pumilus* for SAR studies. The results (see Table 1) showed that spore-forming bacteria, such as the genus *Bacillus* are more sensitive to these compounds, among which compounds **6** and **24** exhibiting higher activity (MIC values of 1 and 2.5 μg/mL against *B. subtilis* and *B. cereus*, respectively) than cephotaxime, used as a positive control. Furthermore, the effectiveness against the Gram-positive cocci was restricted to compound **6** (*S. epidermidis* and *E. faecalis*, MIC 5–2.5 and 20–10 μg/mL, respectively). On the other hand, compounds **1**–**9** were inactive against the Gram-negative bacteria and the yeas<sup>t</sup> *C. albicans* (MIC > 40 μg/mL), which is in line with previous studies on this type of compound [20–23,25].

**Table 1.** Minimal inhibitory concentration (MIC, μg/mL) 1 against Gram-positive bacteria of phenolic nor-triterpenes 2 and lipophilicity (clog *P*).


*S. a.*: *Staphylococcus aureus*, *S. e.*: *S. epidermidis*, *S. s.*: *S. saprophyticus*, *E. f.*: *Enterococcus faecalis*, *B. s.*: *Bacillus subtilis*, *B. a.*: *B. alvei*, *B. c.*: *B. cereus*, *B. m.*: *B. megaterium*, *B. p.*: *B. pumilus*.—not assayed. 1 The experiments were carried out in triplicate, and values represent average of MIC values. 2 All compounds were inactive against Gram-negative bacteria (MIC > 40 μg/mL). Compounds (*clog P*) **1** (6.15), **4** (6.07), **7**–**9** (7.18, 6.90 and 6.90), **11** (6.38), **18** (6.47), **19** (5.44), **21** (7.17), and **23** (5.72) were inactive against all microorganisms assayed. 3 Cephotaxime was used as a positive control.

#### *3.3. Structure–Activity Relationship Analysis*

The influence of substitution patterns of the phenolic nor-triterpenes on their antimicrobial activity was analyzed, taking into account three regions of the triterpene skeleton, the phenolic moiety, the extended B-ring conjugation, and the triterpene scaffold (pristimerin, tingenone, or iguesterin). This analysis revealed the following trends in the structure–activity relationship (SAR) (Figure 4). Regarding the results obtained against *Bacillus* spp., the SAR studies sugges<sup>t</sup> that: (a) A substituent at C-4 seems essential for the activity. Thus, comparison of the activities (MICs against *B. subtilis*) of **14**, **10**, **12,** and **1**, whose only structural difference is the substituent at C4, showed that the most effective

group at this position was a carboxylic acid (**14**), followed by hydroxyl (**10**) and aldehyde (**12**) groups, while a methyl group, as in **1**, led to a loss of activity. *B. cereus*, *B. alvei*, *B. megaterium,* and *B. pumilus* showed similar behavior to *B. subtilis* (**14** vs. **12**); (b) Acetylation and methylation reduce activity, as revealed by comparison of potency of the natural compounds with their corresponding analogues (**7** vs. **24**, **8** vs. **25**, **9** vs. **26**, **10** vs. **11**, **16** vs. **17**, **18**, **19**, and **20** vs. **21**), except for compound **6** with an hydroxyl group at C6 (**6** vs. **7**). This indicates that the hydroxyl group, which could interact with the receptor as hydrogen-bond-donor (HBD), is the best functional group, suggesting that the hydrophilicity is relevant for the activity; (c) Compounds with an unconjugated double bond at C-7 showed considerable activity (e.g., **6** and **24**), and reduction at C-7/C-8 results in a partial loss of antimicrobial activity (**24** vs. **25** and **26**); (d) The β-stereochemistry at C-8 increases two-fold the activity with respect to the H-8<sup>α</sup>, as shown by comparing the MIC values of **26** and **25**; (e) A substituent at C-7, in addition to the carbonyl at C-6, is also relevant for the activity, being a ketone preferred to a hydroxyl group (**22** > **5** vs. **4**).


**Figure 4.** Most relevant structural features for phenolic nor-triterpenes against the most sensitive microorganism, *Bacillus subtilis*.

Furthermore, the results obtained against *Staphylococcus* spp. sugges<sup>t</sup> that: (a) the presence of a carboxylic acid moiety at C-4 could strongly affect the antimicrobial activity against *S. aureus* since its substitution by an aldehyde or a methyl group reduced or fully eliminated the activity (**14** vs. **12** or **1**); (b) moreover, replacement of the carboxyl group at C-29 by its methyl ester led to a partial loss of activity (**13** vs. **12**); (c) compounds containing two carboxylic acid groups lacked activity, whereas those with one carboxylic acid showed high activity (**14** vs. **15**); (d) the reduction of the double bond at C-7/C-8 led to less activity (**24** vs. **25** and **26** against *S. epidermidis*); (e) the β-stereochemistry at the C-8 position favored the activity with respect to the H-8α (**26** vs. **25**); (f) a hydroxyl group at C-6 affected the activity against *S. epidermidis* (**6** vs. **7**). Taking into account these structural features, we propose a hypothetical compound, 4-carboxy-6α-hydroxy-pristimerol, that would have a zeylasterone A-ring (as in **14**), an unconjugated B-ring (as in **6** and **24**), and a pristimerin E-ring as being the most active against *Bacillus* and *Staphylococcus* spp. (Figure 5).

**Figure 5.** Structural requirements of phenolic nor-triterpenes based on SAR studies.

Some phenolic compounds primarily target the cytoplasmic membrane due to their hydrophobic nature, and preferentially partition into the lipid bilayer [33], as was observed in our previous works [22–25]. Therefore, for the purpose of correlating the antimicrobial activity with lipophilicity, clog *P* values [34] of this series of compounds were calculated, including that of the hypothetical compound, 4-carboxy-6 α-hydroxy-pristimerol (clog *P* 4.88), and are shown in Table 1. Slight increases in lipophilicity of **24** (clog *P* 7.09) and **25** and **26** (clog *P* 6.41) by acetylation (**7**, **8** and **9**, respectively) (clog *P* 7.18 and 6.90) led to a suppression of the antibacterial activity. Moreover, other factors must be taken into account for the expression of the activity, such as the presence of an hydrogen-bond donor (HBD) group, strategically positioned at C-4 (e.g., **14**), C-6 (e.g., **6**), or at C-2 and C-3 (e.g., **12** and **14**) with clog *P* of 5.04, 5.95, and 5.67, respectively, which seems relevant for the activity. These observations indicated that both lipophilicity and HBD factors are involved in the antimicrobial activity of this type of compound.
