*Review* **Enantiomeric Tartaric Acid Production Using** *cis***-Epoxysuccinate Hydrolase: History and Perspectives**

#### **Jinsong Xuan 1,\* and Yingang Feng 2,\***


Received: 31 January 2019; Accepted: 1 March 2019; Published: 5 March 2019

**Abstract:** Tartaric acid is an important chiral chemical building block with broad industrial and scientific applications. The enantioselective synthesis of L(+)- and D(−)-tartaric acids has been successfully achieved using bacteria presenting *cis*-epoxysuccinate hydrolase (CESH) activity, while the catalytic mechanisms of CESHs were not elucidated clearly until very recently. As biocatalysts, CESHs are unique epoxide hydrolases because their substrate is a small, mirror-symmetric, highly hydrophilic molecule, and their products show very high enantiomeric purity with nearly 100% enantiomeric excess. In this paper, we review over forty years of the history, process and mechanism studies of CESHs as well as our perspective on the future research and applications of CESH in enantiomeric tartaric acid production.

**Keywords:** *cis*-epoxysuccinate hydrolase; tartaric acid; enantioselectivity; stereoselectivity; regioselectivity; epoxide hydrolase; immobilization; whole cell catalyst; enzyme stability; biocatalyst

#### **1. Introduction**

Tartaric acid (TA) is a well-known organic acid that naturally occurs in many kinds of fruit, most notably in grapes. The chemical chirality of TA was first discovered by Jean-Baptiste Biot in 1832 [1]. The naturally-occurring form of the acid is L(+)-TA, while D(−)-TA rarely exists in natural sources [2,3]. L(+)-TA is widely used in the food, wine, pharmaceutical, chemical, and polyester industries. D(−)-TA is also important in pharmaceutical manufacturing [4–6]. Both are well-known chiral chemical building blocks with broad industrial and scientific applications [7,8]. In enantioselective chemical synthesis, TA serves not only as a resolving agent or chiral auxiliary in the synthesis of bioactive molecules, but also a source of new asymmetric organocatalysts [7–9]. Traditionally, L(+)-TA is obtained as a solid by-product during wine fermentation, and this kind of production method is strongly influenced by the growth of grapes and the climatic conditions. Chemical synthesis of L(+)-TA with maleic acid is also possible but this gives a much less soluble racemic product (DL-form) which is not suitable for inclusion in foods because D(−)-TA in the product is considered to be harmful to human health. Commercial application of the chemical method is limited by both the product form and the high production cost [10]. Currently, microbial methods are considered to be much simpler and more economical for the production of L(+)-TA and D(−)-TA.

Epoxide hydrolases (EHs, EC 3.3.2.3) are biocatalysts that are ubiquitous in Nature. They can hydrolyze racemic epoxides to their corresponding optically active epoxides and pure vicinal diols, which are versatile intermediates for chiral pharmaceutical synthesis. In general, this enzymatic

process occurs under mild conditions without the need for any cofactors, prosthetic groups, or metal ions [11]. The high transformation rate and enantioselectivity of epoxide hydrolases have gained them increasing attention in recent years, and they have found more and more applications in the organic chemical industry [11]. Epoxide hydrolases are found in a variety of sources, such as plants, insects, mammals, and microbes [12–14]. Mammalian epoxide hydrolases have been the subject of many studies because of their key role in xenobiotic detoxification in the liver, but their use as biocatalysts has been hindered by their limited availability [15]. Lately, bacterial epoxide hydrolases have been increasingly recognized as highly versatile biocatalysts owing to their abundance, high efficiency, and environmental friendliness [16].

*cis*-Epoxysuccinic acid hydrolases (CESHs) are epoxide hydrolase members that catalyze the asymmetric hydrolysis of *cis*-epoxysuccinate (CES) to form an enantiomeric tartrate [17–19]. Bacteria presenting CESH activity were discovered in the 1970s, and the synthesis of L(+)-TA was the first application of an epoxide hydrolase [20]. Since then, a large number of bacteria with CESH activity have been discovered and successfully applied for industrial TA production. The sequences and mechanism of CESHs have been partly elucidated since 2000, when it was revealed that CESH[L] and CESH[D], which produce L(+)-TA and D(−)-TA, respectively, are completely different proteins in terms of both sequence and structure. Therefore, CESHs are interesting EHs not only for TA production, but also for enantiomer biosynthesis in general. In this review, we summarize the body of literature on CESHs including both process optimization for industrial application and mechanism studies to understand how their regio- and stereoselectivity makes them efficient biocatalysts. We also provide our perspective on the use of CESHs in future research and applications.

#### **2. The History of CESH Studies**

Several patents involving bacteria with CESH activity were filed by Japanese companies in the 1970s [21–26]. However, these bacteria and their associated CEHSs did not receive much attention in the following ten years. In fact, unlike mammalian EHs, which have been subject to extensive enzymatic and biochemical study, microbial EHs were not well studied until the 1990s. Several studies reported in the 1990s showed that bacterial CESHs are promising biocatalysts for industrial synthesis, and EHs were proposed as new tools for the synthesis of fine organic chemicals [16,20]. In the 1990s, several groups from China, Belgium, Japan, and Slovakia continued to report on new bacteria containing CESH activity, and some process optimization methods, such as immobilization, were performed for industrial application [3,10,27–31]. The microbial production of L(+)-tartaric acid was successfully commercialized in the late 1990s [32]. Subsequently, new bacterial species were isolated for TA production in the 21st century [18,33–39], which demonstrates that CESHs exist broadly in bacteria.

In 2000, the first CESH encoding gene was reported from *Alcaligenes* sp. MCI3611 for the production of D(−)-TA [40]. Genes from several other CESHs, including CESH[D] from *Bordetella* sp. BK-52 and CESH[L]s from *Rhodococcus opacus*, *Nocardia tartaricans*, and *Klebsiella* sp. BK-58, have also been sequenced and cloned [17,19,41–43]. The genes and the derived amino acid sequences of CESHs provided the basis for later studies of the recombinant expression, structure, and mechanisms as well as the protein engineering of CESHs. Analyses of the amino acid sequences of CESHs indicated that CESH[D]s and CESH[L]s are completely different proteins [19]. Subsequent structural analysis and mechanism studies revealed that they have different structures and catalysis mechanisms [41,43–45]. In a recent study [46], we determined a high-resolution structure and elucidated detailed catalytic mechanisms for CESH[D], but these characteristics have not yet been reported for CESH[L].

With the knowledge of CESH sequences, structures, and mechanisms, scientists began to conduct extensive protein engineering research to improve enzyme stability and activity. As a result, many mutants with good properties have been obtained for potential industrial use [42,47,48]. However, these studies were accomplished without knowledge of the high-resolution structures

of CESHs; therefore, there is potential to further improve the use of CESHs as biocatalysts in industrial applications.

#### **3. Bacteria that Produce CESHs**

CESHs catalyze the enzymatic hydrolysis of *cis*-epoxysuccinate to form L(+)-TA or D(−)-TA with high product enantioselectivity. TA products obtained by the hydrolysis of CES using purified CESHs generally have enantiomeric excess (EE) values of near 100% [19,36,38,39,44,45]. Therefore, bacteria that produce CESH[D] or CESH[L] have been separately reported. More than twenty species have been isolated by researchers and are distributed among more than ten genera (Table 1). Isolated species with CESH[L] activity include both Gram-positive and Gram-negative bacteria, while to date, all species with CESH[D] activity are Gram-negative. Only the genus *Pseudomonas* has both type of species.


**Table 1.** Strains producing *cis*-epoxysuccinic acid hydrolases (CESHs).

<sup>1</sup> The gene and protein sequence were reported in a patent [40].

More than ten species with CESH[L] activity have been reported since the 1970s. Miura et al. discovered that microorganisms of the genus *Nocardia* can produce CESH[L]; specifically *Nocardia tartaricans nov. sp.* was identified as a preferred natural species [24]. Kamatani et al. also isolated microorganisms capable of hydrolyzing *cis*-epoxysuccinate to L(+)-TA belonging to the genera *Pseudomonas*, *Agrobacterium*, and *Rhizobium* [22]. Two strains of *Rhizopus validum* and *Corynebacterium* JZ-1 that were discovered in soil can produce L(+)-TA, and the latter has a molar conversion rate of *cis*-epoxysuccinate as high as 96% [27,28]. Screening of 65 *Nocardia* strains identified *Nocardia* sp. SW 13-57 as a high-yield strain with the ability to produce CESH[L]. Its molar conversion rate is over 90% and the CESH[L] formation is effectively induced by sodium *cis*-epoxysuccinate during fermentation [29]. In addition, *Rhodococcus ruber* M1 isolated from soil was the first strain in *Rhodococcus* reported to produce CESH[L] [34]. Further, a strain of *Klebsiella* sp. BK-58 can produce a novel CESH[L] with good thermal and pH stability, which is a promising biocatalyst for the industrial production of L(+)-TA [38].

Ten species in four genera have been reported to have CESH CESH[D] activity. Sato et al. first isolated four novel species of *Achromobacter* and two novel species of *Alcaligenes* with CESH[D] activity [21]. A strain belonging to the genus *Pseudomonas* and the microbial cells of *Alcaligenes* sp. MCI3611 also has the capability to hydrolyze *cis*-epoxysuccinate to D(−)-TA [31,33]. The DNA fragment encoding the enzyme to produce D(−)-TA was successfully obtained from the chromosomal DNA library of *Alcaligenes* sp. MCI3611 [40]. Two strains from the genus *Bordetella* (*Bordetella* sp. strain 1-3 and *Bordetella* sp. BK-52), isolated from vegetable fields in Hangzhou, were able to transform *cis*-epoxysuccinate to D(−)-TA [35,36]. Unlike traditional *Bordetella* species that are exclusively associated with humans and warm-blooded animals, both of these strains are from the natural environment. Unlike previously reported CESH[D] producing bacteria, *Bordetella* sp. strain 1-3 has also been reported to have the ability to degrade D(−)-TA as its carbon source, so some measures should be adopted to stop this degradation process to accumulate D(−)-TA. Furthermore, the molecular weight of CESH[D] from *Bordetella* sp. strain 1-3 is the same as the beta subunit of the previously reported CESH[D] from *Alcaligenes* sp. The eight amino acid sequence of the N-terminal region of CESH[D] from *Bordetella* sp. strain 1-3 has also been shown to have the same sequence as the beta subunit from *Alcaligenes* sp. [19].

The gene sequences of several CESH-producing bacteria have been reported, including CESH[L] genes from *Rhodococcus opacus*, *Nocardia tartaricans* CAS-52, and *Klebsiella* sp. BK-58, and CESH[D] genes from *Alcaligenes sp.* MCI3611 and *Bordetella* sp. BK-52 [17,19,37,40,43]. Some of these genes have been successfully expressed in *Escherichia coli* [17,19,37,43]. The amino acid sequences of CESH[L] from *Rhodococcus opacus* and *Nocardia tartaricans* CAS-52 are identical, but they only share 36% sequence identity with the CESH[L] from *Klebsiella* sp. BK-58. The two CESH[D]s from *Alcaligenes* sp. MCI3611 and *Bordetella* sp. BK-52 have identical amino acid sequences. The amino acid sequences of CESHs from most bacteria in Table 1 are still unknown.

#### **4. Stability of CESHs**

Although CESHs have excellent product enantioselectivity and high activity, pure CESHs are unstable and heat-sensitive and are thus unsuitable for industrial applications [3,17,49]. A continuous bioconversion study using *Rhodococcus rhodochrous* showed that the effect of the large increase in stability at a lower temperature was much more important than the decrease in activity [3]. To improve the stability of CESHs at the optimal pH and temperature, whole-cell immobilization was adopted for the industrial bioconversion process. Carriers including gelatin beads, pectate gel beads, and κ-carrageenan were screened, and the process of immobilization was optimized for different species [30,49–54]. Whole-cell immobilization was shown to not only increase the stability of the biocatalysts, but also improve the activity and conversion ratio.

CESH[L] and CESH[D] show different stabilities. CESH[D] from *Bordetella* sp. BK-52 has high stability and activity in a broad range of temperatures (37–45 ◦C) and pH values (4.6–9.0) with optimal conditions being 40 ◦C and pH 6.5 [19]. A comparison study indicated that CESH[D] has greater thermal and pH stability than CESH[L] [41]. However, the recently discovered novel CESH[L]s from *Klebsiella* sp. and *Labrys* sp. BK-8 have good thermal and pH stability. The former is stable up to 50 ◦C and at pH 5 to 11, while the latter is stable over a broad range of temperatures and pH values with the greatest activity occurring at 50 ◦C and 8.5 [38,39]. Therefore, they could be used as alternative biocatalysts for the production of L(+)-TA.

In addition, the stability and activity of CESHs can also be improved by protein engineering methods such as fusing a binding module to CESH or changing the protein primary structure. For example, the wild-type CESH[L] gene from *Rhodococcus opacus* has been fused with a carbohydrate binding module (CBM30), and the resulting fusion enzyme (CBM30-CESH) exhibited improved temperature and pH adaptability than free native CESH[L] [55]. The CESH[L] mutant 5X-1 from *Rhodococcus opacus* was successfully constructed by combining directed evolution with various semi-rational redesign methods. The optimal reaction temperature using mutant 5X-1 occurred at 55 ◦C, which is much higher than the optimal temperature at 35 ◦C using the wild-type enzyme. The pH range for the effective working of mutant 5X-1 extended from 8.0–9.0 to 5.0–10.0 [47]. Random

mutation by error-prone PCR and high throughput screening revealed that single point mutations on the Phe10 residue of CESH[L] from *Klebsiella* sp. BK-58 resulted in different levels of enzyme thermostability and catalytic activity. The mutant F10Q had 230% higher activity but lower stability than the wild-type enzyme [48].

#### **5. Process Optimization for TA Production Using CESHs**

Since the discovery of CESH activity in the 1970s, CESH-producing bacteria and CESHs have been utilized to produce TA with high enantiopurity. The cell lysate or crude enzyme solution is not suitable for TA production because the cellular protease degrades CESH rapidly. Therefore, significant effort has been made to optimize TA production using CESHs or CESH-producing bacteria. TA production was established in the 1970s including the surfactant addition and recovery of TA from the media [23,25]. Subsequent studies have found that these enzymes have low stability [3], so studies on process optimization have mainly focused on the methods of immobilization and recombination.

Bacterial cells can be immobilized by different carriers and show different levels of efficiency and stability. Different microorganisms have different optimal immobilization methods, for example, the best cell immobilization carriers for *Nocardia tartaricans*, *Corynebacterium*, and *Rhizobium* are gelatin, κ-carrageenan, and sodium alginate, respectively [30,52,56]. The immobilization of *Labrys* and recombinant *E. coli* cells with carrageenan is also an excellent process for TA production with high efficiency and stability [39,54]. Aside from the carriers, the subsequent processes of immobilization also have important effects on both the activity and stability. Rosenberg et al. found that although κ-carrageenan is an excellent carrier for the immobilization of *Nocardia tartaricans* cells, the use of cross-linked calcium pectate gel (CPG) is advantageous for the preparation of spherical particles with high activity and stability [50,51]. Sodium alginate–cellulose sulfate-poly(methylene-*co*-guanidine) (SA-CS/PMCG) capsules have been shown immobilize *Nocardia tartaricans* with a better performance than CPG [49]. Additionally, various surfactants can greatly enhance the activity of the immobilized cells, mainly through a change in the permeability of the cell membrane [50,52,57]. Dong et al. reported that ultrasound treatment could be used to change the cell permeability and improve the bioconversion efficiency of immobilized *E. coli* cells containing expressed recombinant CESH[D] [58].

The genes of CESH[L] and CESH[D] have been cloned and expressed in *E. coli* successfully [17,19,40,41,43]. Therefore, recombinant CESHs also have good potential to be used in industrial TA production. Some studies have reported that the stability of enzymes can be improved by immobilization. For example, we significantly improved the stability of CESH[L] by fusing it with CBM [55], and then purified and immobilized it on cellulose in one step. Wang et al. immobilized CESH[L] on agarose Ni-IDA to enhance its stability [59]. The preparation of recombinant CESH[L] was also improved by the utilization of the heat-induced promoter to avoid the chemical induction of protein expression [60]. Still, more effort is needed to optimize the stability of recombinant CESH and the process of TA production using the recombinant enzyme.

#### **6. Structure and Catalytic Mechanism of CESH[L]**

Until now, there have been no reports on the structure of CESH[L]. Only two CESH[L] sequences have been reported and they share a 36% sequence identity [17,37,43]. Both have about a 30% sequence identity with L-2-haloacid dehalogenase, of which the structure is known. Therefore, homology modeling has been performed to elucidate the catalytic mechanism of CESH[L] [41,42,44]. L-2-haloacid dehalogenase has an α/β hydrolase fold that is adopted by most EHs [61]. Therefore, it is likely that CESH[L] also adopts the α/β hydrolase fold, and the catalytic mechanism of CESH[L] has been proposed to be similar to most EHs, i.e., a two-step mechanism including an ester intermediate [61]. The two-step mechanism was confirmed by 18O experiments for both CESH[L]s [43,44].

As CESH[L] and L-2-haloacid dehalogenase only have about a 30% sequence identity and different substrates, the catalytic residues cannot be deduced from homology modeling. Two mutagenesis analyses revealed that D18, H190, and D193 are essential for the activity; therefore, they were proposed to be a catalytic triad, with D18 activating the attacking water molecule [42,44]. However, as the CES substrate is a polar hydrophilic molecule, the active site of CESH[L] contains many charged and hydrophilic residues, making it difficult to elucidate the role of each residue without a high-resolution crystal structure of the CESH[L]-substrate complex. The proton donor that may facilitate the ring opening is still not clear, and how the CES is fixed in the active site to ensure the stereoselectivity also remains unknown.

#### **7. Structure and Catalytic Mechanism of CESH[D]**

Although CESH[D] and its gene sequence were reported earlier than CESH[L], knowledge of CESH[D] catalysis was not obtained until very recently. The protein sequences of CESH[D]s from *Bordetella* sp. BK-52 and *Alcaligenes* sp. MCI3611 have about 30% sequence identity to the Kce enzyme, whose function is totally different [19,40,41]. Homology modeling using Kce as a template indicated that CESH[D] has a TIM barrel fold with a metal ion which is crucial for its activity [41,45]. The metal ion was identified to be a divalent ion, either zinc, calcium, or magnesium [41,45], which is coordinated by three residues in Kce; however, only two of these are conserved in CESH[D] [41,45]. The third coordinative residue could not be identified before the CESH[D] structure was determined. An 18O labeling experiment indicated that CESH[D] hydrolyzes CES through a one-step mechanism [45] instead of the two-step mechanism of CESH[L]. As CESH[D] and Kce have different substrates, their active sites are very different and their catalytic residues cannot be deduced by homology modeling. Mutagenesis studies have identified a large number of essential residues, so it is difficult to identify the key catalytic residue and stereoselective catalytic mechanism from these studies without a high-resolution CESH[D] structure.

The catalytic mechanism of CESH[D] was not elucidated in detail until our recent report on high-resolution CESH[D] structures [46]. The structure of substrate-free CESH[D] revealed not only the third metal coordinative residue (Glu14), but also three coordinative water molecules that formed an equilateral triangle. Trials using an inactive mutant and CES co-crystallization obtained an unexpected CESH[D] structure in complex with its reaction product, D(−)-TA. In the complex structure, three oxygen atoms of TA occupy the positions of the three coordinative water molecules in the substrate-free CESH[D] structure. The identification of the structure of the product–enzyme complex provided the details of substrate binding and positioning, from which the key catalytic residues and substrate recognition residues were elucidated. Instead of the previously proposed catalytic residue D251 [45], the catalytic residues were identified as D115 and E190, while R11 provided the proton and facilitated the ring opening. D251 played a crucial role in fixing the position of R11, which supports the importance of this residue, as identified in the previous mutagenesis analysis. This structure and catalytic mechanism explained the stereoselectivity, regioselectivity, and substrate selectivity of CESH[D] [46].

The structure of CESH[D] has some distinct features in comparison with known EHs. In contrast with the α/β hydrolase fold or LEH fold adopted by most EHs, except for LA4H, ChEH, and FosX, CESH[D] adopts a TIM-barrel fold [61]. Also, CESH[D] has a one-step mechanism, while α/β hydrolase fold EHs have a two-step mechanism. LEH adopts a one-step mechanism, but the substrate specificity is determined by the hydrophobic interaction (molecular shapes), and no metal ion is needed for LEH [62]. The EH with the most similar mechanism is FosX, which also contains a metal ion for substrate binding and adopts a one-step mechanism (Figure 1) [63]. However, FosX has a dimeric VOC family fold with paired βαβββ where the active sites are located at the dimer interface, and the substrate of FosX only occupies two coordination sites of the metal ion with square pyramidal coordination geometry [63]. Although CESH[D] is also a dimeric protein, its active sites are located at the center of each dimer's subunit, and its metal elements have octahedral coordination geometry where three of the coordination sites are occupied by the substrates [46]. Therefore, CESH[D] is a unique EH in terms of both its protein fold and catalytic mechanism.

**Figure 1.** Comparison of CESH[D] and FosX: (**a**) structure of CESH[D]; (**b**) structure of FosX; (**c**) the active site of CESH[D]; (**d**) the active site of FosX. Both proteins are dimeric and colored in green and cyan for each monomer. The metal ions are shown as yellow balls. The products are shown as magenta and red ball and sticks for CESH[D] and FosX, respectively. In (**c**,**d**), the three coordinative residues of the metal ion are shown as blue sticks.

#### **8. Perspective for CESH Research and Application**

CESHs are unique among the known EHs because the CES substrate is highly hydrophilic and mirror-symmetric. The structural features of CES suggest that CESHs have specific substrates; in other words, the CES molecule is fixed in CESH with an exact position, which leads to the high stereoselectivity and regioselectivity of CESHs. This feature is of great interest for enantioselective synthesis. Although CESHs and their host bacteria have been successfully applied in industry for TA production, there are still many questions to be addressed, and there is a lot of room for growth in the production of TA using CESHs.

Currently, the structure and the catalytic mechanism of CESH[L] are still not fully understood. Determination of the high-resolution structure of CESH[L], particularly of complexes with a substrate or product, is necessary to elucidate its catalytic mechanism. With this structure, rational engineering to enhance the stability will be possible. Furthermore, CESHs could also potentially be engineered as biocatalysts to catalyze different substrates, but this potential has not been explored in past studies.

Although many microorganisms have been reported to have CESH activity, only very few of them have been sequenced. Therefore, much work is still needed to isolate the CESHs and analyze their protein/gene sequences. The new sequences of CESHs may provide other new features that will help us to understand these enzymes and promote their applications.

Currently, the production of TA using whole cell catalysts is performed by wild strains. No metabolic engineering of these microorganisms has previously been reported. Therefore, understanding the features of these microorganisms and the development of genetic engineering are important topics for future studies. CESHs are intracellular enzymes that are present in certain

microorganisms, which cause the activity of whole cell catalysts to depend on the cell permeability. If the CESHs could be engineered as a secretive protein, or immobilized on the cell surface, their activity would be greatly enhanced. This engineering could be done in either the original species or the recombinant *E. coli* cells, and it will improve the production of enantiomerically pure TA using CESHs.

**Funding:** This research was funded by the National Natural Science Foundation of China (grant numbers 31670735 and 31661143023 to YF); and the Undergraduate Education and Teaching Reform Research Project, USTB (grant number JG2018M39 to JX).

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


63. Fillgrove, K.L.; Pakhomova, S.; Schaab, M.R.; Newcomer, M.E.; Armstrong, R.N. Structure and mechanism of the genomically encoded fosfomycin resistance protein, FosX, from *Listeria monocytogenes*. *Biochemistry* **2007**, *46*, 8110–8120. [CrossRef] [PubMed]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Synthetic Chiral Derivatives of Xanthones: Biological Activities and Enantioselectivity Studies**

**Carla Fernandes 1,2,\* , Maria Letícia Carraro 1,† , João Ribeiro 1,† , Joana Araújo 1,†, Maria Elizabeth Tiritan 1,2,3,\* and Madalena M. M. Pinto 1,2**


Academic Editor: Maurizio Benaglia Received: 1 February 2019; Accepted: 19 February 2019; Published: 22 February 2019

**Abstract:** Many naturally occurring xanthones are chiral and present a wide range of biological and pharmacological activities. Some of them have been exhaustively studied and subsequently, obtained by synthesis. In order to obtain libraries of compounds for structure activity relationship (SAR) studies as well as to improve the biological activity, new bioactive analogues and derivatives inspired in natural prototypes were synthetized. Bioactive natural xanthones compromise a large structural multiplicity of compounds, including a diversity of chiral derivatives. Thus, recently an exponential interest in synthetic chiral derivatives of xanthones (CDXs) has been witnessed. The synthetic methodologies can afford structures that otherwise could not be reached within the natural products for biological activity and SAR studies. Another reason that justifies this trend is that both enantiomers can be obtained by using appropriate synthetic pathways, allowing the possibility to perform enantioselectivity studies. In this work, a literature review of synthetic CDXs is presented. The structures, the approaches used for their synthesis and the biological activities are described, emphasizing the enantioselectivity studies.

**Keywords:** synthetic xanthones; chiral derivatives of xanthones; bioactivities; enantioselectivity; enantiomeric purity

#### **1. Introduction**

In the last few years, the relationship between chirality and biological activity has been of increasing importance in Medicinal Chemistry [1]. Chirality can now be considered as one of the majors' topics in the design, discovery, development and marketing of new drugs [2–4]. Enantiomers of drugs often present different behaviours within pharmacodynamics and/or pharmacokinetics events [5–7] as well as different levels of toxicity [8–10]. These differences makes in the majority of cases therapeutics with single enantiomers of unquestionable advantages [11].

The importance of enantioselective studies and the increase of chiral drugs in the pharmaceutical market upsurges each year due to the advantages in potency, efficacy, selectivity and safety associated with the use of single enantiomers. The advances in enantioselective synthesis [12–15] as well as enantioresolution methodologies [16–19] aligned to the stricter requirements from regulatory authorities to patent new chiral drug [20–22] boosted the research in this field.

The development of methods to obtain and analyse both pure enantiomers has acquired crucial importance in the early stage of drug development, as biological and pharmacological activity evaluation of both enantiomers are required. Thus, despite the diversity of structures with different groups and positions at the base scaffolds, the library of compounds for structure-activity relationship (SAR) studies should also include stereoisomers for enantioselectivity evaluation. To perform that type of studies, it is necessary to obtain both enantiomers with very high enantiomeric purity. The improvement of chromatographic instrumentation and the development of efficient chiral stationary phases (CSPs) [23–26] have made liquid chromatography (LC) the first choice for determination of enantiomeric purity.

Chemically, xanthones (9*H*-xanthen-9-ones) are compounds with an oxygen-containing dibenzo-γ-pyrone heterocyclic scaffold [27], being considered as a privileged structure [28,29]. Within this class of compounds a broad range of biological and pharmacological activities has been reported [30–34]. Additionally, other applications have been described for xanthone derivatives, such as preparation of fluorescent probes [35,36] or stationary phases for LC [23,24,37].

Many naturally occurring xanthones, isolated from terrestrial as well as marine sources, are chiral, and present interesting biological activities [38–40]. The biosynthetic pathway of xanthones only allows the presence of specific groups in particular positions of the xanthone scaffold, which is a limiting factor for structural diversity. For this reason, in order to enlarge chemical space in this field, total synthesis needs to be considered [41,42] allowing the access to structures that otherwise could not be reached only with natural product as a basis for molecular modification. Moreover, higher number of compounds can be obtained for SAR studies.

#### **2. Synthetic Chiral Derivatives of Xanthones**

Recently, there has been an increase interest in new bioactive xanthones obtained by synthesis, particularly in chiral derivatives of xanthones (CDXs). Some reasons that can justify this trend were the importance of this class of compounds in Medicinal Chemistry align with the fact that nature usually gives only one enantiomer and the synthetic procedures allow the preparation of both enantiomers to explore the enantioselectivity in biological screening assays. The promising biological and pharmacological activities of some chiral members of this family, the clinical advantages of a single enantiomer than a racemate, the scarce examples of synthetic CDXs described, and the possibility to perform enantioselectivity studies strengthens the obtaining of new synthetic CDXs. This review aims to gather the research findings on synthetic CDXs, reporting their biological and pharmacological activities. The structures and the bioactivity differences associated to the stereochemistry of the CDXs (enantioselectivity) as well as their enantiomeric purity are highlighted.

#### *2.1. Synthetic CDXs Inspired in Naturally Occurring Xanthones*

Natural compounds have always been a source of inspiration for the discovery of new therapeutic agents [43]. Historically, the first proposed synthesis of a xanthone was achieved by Michael, in 1883, and later by Kostanecki and Nessler, in 1891, through the distillation of *O*-hydroxy-benzoic acid, acetic anhydride and a phenol [44,45], while the first total synthesis of a naturally occurring xanthone was a euxanthone, described by Ullmann and Panchaud, in 1906 [46]. Several synthetic CDXs analogues of natural xanthones are described. The biological and pharmacological activities evaluated are summarized in Table 1.

#### 2.1.1. Synthetic Xanthonolignoids

Xanthonolignoids are a natural class of compounds isolated from plants of the *Clusiaceae* family (*Guttiferae*) [47]. They possess a phenylpropane core linked to a xanthone scaffold by a dioxane ring, formed by radical oxidative coupling [48,49]. Natural xanthonolignoids include kielcorins, cadensins, subalatin, calophyllumins and gemixanthone [49]. The first xanthonolignoid described was based in 2,3,4-trioxygenated xanthone being isolated, in 1969, from *Kielmeyera species* [50].

Considering that xanthonolignoids are bioactive molecules and very interesting templates for molecular modifications, several xanthonolignoids were isolated and synthesized [49]. Initially the main goal of their synthesis was to help in the structure elucidation of this class of compounds but subsequently, also to improve their biological and physicochemical properties. Both classic synthesis and biomimetic approaches have been used to obtain xanthonolignoids, mainly kielcorin derivatives [49]. The total synthesis of kielcorin derivatives requires several steps and drastic reaction conditions while the biomimetic way is based on natural building blocks and is achieved by an oxidative coupling of a suitable dihydroxyxanthone and a cinnamyl alcohol derivative, in the presence of an oxidizing agent at room temperature [49].

Pinto et al. [51], in 1987, reported the first biomimetic synthesis of xanthonolignoids of the kielcorin group, specifically kielcorin (**1**) and its stereoisomer, *cis*-kielcorin (**2**) (Figure 1). Both kielcorins **1** and **2** were also obtained by a classical method [52,53]. In 1999, by using a biomimetic approach, the synthesis of *trans*-kielcorin B (**3**) and *trans*-isokielcorin B (**4**) (Figure 1) was described [54].

**Figure 1.** Structures of kielcorin (**1**) and synthetic derivatives **2**–**9**.

To obtain related bioactive compounds with a kielcorin framework, other constitutional isomers were synthesized by our group, namely *trans*-kielcorin C (**5**), *trans*-kielcorin D (**6**), *trans*-isokielcorin D (**7**), *cis*-kielcorin C (**8**), and *trans*-kielcorin E (**9**) [55] (Figure 1). Once again, the synthetic approach, involving an oxidative coupling of coniferyl alcohol with an appropriate xanthone, was modelled on a biomimetic pathway. Different oxidizing agents were used (e.g., Ag2O, Ag2CO3, and K3[Fe(CN)6]) to investigate the oxidative coupling reactions. The synthesis of *trans*-kielcorin C (**5**) (also designated as demethoxykielcorin) using a classic approach was previously reported by Vishwakarma et al. [53].

Kielcorins **5**–**9** were evaluated for their *in vitro* effect on the growth of three human tumor cell lines, MCF-7 (breast), TK-10 (renal), UACC-62 (melanoma), and on the proliferation of human lymphocytes [56]. The growth inhibitory effect was moderate but dose-dependent and influenced by the isomerism of the tested compounds. The *trans*-kielcorins C (**5**) and D (**6**) were the most active. The inhibition of human lymphocyte proliferation induced by phytohemagglutinin (PHA) was detected [56]. The high potency and selectivity observed for these compounds suggested that kielcorins may be an important model for developing potent and isoform-selective protein kinase C (PKC) inhibitors [31]. Accordingly, kielcorins **5-9** revealed an effect compatible with PKC inhibition similar to that exhibited by the well-established PKC inhibitor chelerythrine [57]. The *trans*-kielcorins C (**5**) and

E (**9**) were evaluated and demonstrated protectives effects against *tert*-butylhydroperoxide-induced toxicity in freshly isolated rat hepatocytes [58].

In order to study if the growth inhibitory effects of the kielcorins **5**–**9** depended on the stereochemistry, analytical LC methods using four carbamates of polysaccharide derivative CSPs and multimodal elution conditions were developed for their enantioresolution [59]. An amylose *tris*-3,5-dimethylphenylcarbamate CSP was chosen for a preparative resolution scale-up considering not only the highest enantioselectivity obtained for these chiral compounds, but also due to its low retention factors [59]. Consequently, the enantiomers of the kielcolins **5**–**7** and **9** were efficiently separated by chiral LC on a multimilligram scale. A solid-phase injection system was developed and combined with a closed loop recycling system to increase the productivity and recovery of the preparative process [60]. The enantiomeric purity was also measured being higher than 99% for each enantiomer, except for compound **5** [60].

The inhibitory activity of the racemates **5**–**7** and **9** as well as of the corresponding enantiomers on *in vitro* growth of the human breast adenocarcinoma cell line MCF-7 was evaluated and compared. The most evident enantioselectivity was noticed between the racemate of *trans*-kielcorin D (**6**) (inactive) and the active enantiomers (+)-**6** and (−)-**6** [60].

#### 2.1.2. Derivatives of Psorospermin

Psorospermin (**10**) was isolated from the bark and roots of the African plant *Psorospermum febrifugum*, in 1980 [61,62]. It is a natural fused tetracyclic xanthone containing two stereogenic centers with (2*R*,3'*R*)-stereochemistry and a reactive epoxide (Figure 2). The importance of the configuration and the functionality of the epoxydihydrofuran group for the *in vivo* activity have been evaluated [31,50]. The total synthesis of psorospermin (**10**) was reported for the first time in 2005, by obtaining the xanthone skeleton by the method of Grover et al. [62], including thirteen steps and with an overall yield of 1.7%. Psorospermin (**10**) revealed interesting biological activities showing antileukaemic, and antitumor activity in several human cell lines [31,62].

**Figure 2.** Structures of psorospermin (**10**) and synthetic derivatives **11**–**15**.

Additionally, the (*R*,*R*)-stereochemistry of psorospermin (**10**) gave optimum DNA alkylation and antitumor activity, although all four possible stereoisomers show topoisomerase II-dependent alkylation [63].

Two ring-constrained derivatives of psorospermin were also synthesized, namely, stereoisomer **12,** a ring-constrained (2*R*,3*R*)-form, and **13,** a ring-constrained (2*R*,3*S*)-compound (Figure 2) [63]. The chlorohydrin **14** retains psorospermin-like DNA alkylation characteristics despite its rigid structure and high affinity for DNA. The chlorohydrin **14** and epoxide **13** showed increased cytotoxicity against a range number of human tumor cell lines, compared to isohydroxypsorofebrin (**11**) [63].

Another study described the synthesis of two diastereisomeric pairs of *O*-5-methyl psorospermin and evaluation of *in vitro* activity against a range of solid and hematopoietic tumors. The diastereisomeric pair having the naturally occurring enantiomer (2*R*,3*R*) (**15**) (Figure 2) was the most active across all the cell lines tested. In subsequent studies using the four isomers of *O*-5-methyl psorospermin, the order of biological potency was (2*R*,3*R*) > (2*R*,3*S*) = (2*S*,3*R*) > (2*S*,3*S*) [64]. The compound (2*R*,3*R*) psorospermin (**15**) showed to be as effective as gemcitabine (chemotherapeutic drug) in slowing tumor growth *in vivo* in pancreatic cancer model [64].

#### 2.1.3. Derivatives of Muchimangins

In many tribes and folk medicine use, plants and other organisms are commonly used to treat several conditions. For example, in Africa, the roots of *Securidaca longepedunculata* are used to treat sneezing, syphilis, gonorrhea, rheumatic pain, headache, feverish pain, malaria, sleeping sickness, among other conditions [65]. Muchimangins are a minor constituent of this specie and their biological activities have not been fully explored [66]. Dibwe et al. [67] reported the promising antiausteric activity of one natural occurring muchimangin against human pancreatic cancer PANC-1 cell line. Besides the anticancer promising activity, Kodama et al. [66] explored the antimicrobial activity of these structures and performed SAR studies. Accordingly, they synthesized several muchimangins derivatives **16**–**20** (Figure 3), and analyzed their antimicrobial activity.

To synthesize the muchimangins derivatives **16**–**20**, they etherified commercially available 1,2,4-trihydroxybenzene with dimethyl sulfate, producing 1,2,4-trimethoxybenzene. Then, by acylation 2,4,5-trimethoxybenzophenone was obtained. This compound was further reduced to afford 2,4,5-trimethoxydiphenylmethanol, part of the muchimangin skeleton. Afterwards, the corresponding xanthone moiety was obtained using Eaton's reagent. To finalize, both structural moieties were coupled by a Bronsted acid-catalyzed nucleophilic substitution, to produce the corresponding racemates [66]. In order to clarify the effect of chirality, Kodama et al. [66] separated the most promising derivatives using a CSP and to identify their optical rotation via polarimetry.

The preliminary SAR studies suggested that the presence of a hydroxyl group at C-6 was important for the antibacterial activity. Moreover, enantioselectivity occurred for compound **18**, with the dextro (+) enantiomer being more active against *S. aureus* than the levo (-) enantiomer and the racemate [66].

#### 2.1.4. Derivatives of Mangiferin

Mangiferin (**21**, Figure 4) is a natural occurring chiral xanthone with a large spectrum of biological activities, which have been explored for many years [68–71]. Several authors have compiled information about the biological properties of mangiferin and derivatives [72,73].

As previously reported by Araújo et al. [74], mangiferin derivatives present a large spectrum of antimicrobial activities. Singh et al. [75,76] developed new mangiferin derivatives **22**–**27** (Figure 4) and screened their antipyretic and antimicrobial activities. The synthetic strategy used equivalent molar proportions of mangiferin and an appropriate base (*R*-aromatic amine) at reflux to give the corresponding derivative.

**Figure 4.** Structures of mangiferin (**21**) and some synthetic derivatives **22**–**27** with antipyretic and antimicrobial activities.

According to the results, it was suggested that the antipyretic activity may be attributed to the anti-inflammatory and antioxidant potential of mangiferin and its derivatives [76]. However, further investigations are required to understand the mechanism of action to prove its antipyretic activity. Regarding to the antimicrobial activity, the same compounds showed powerful inhibition of the growth of *S. virchow* and significant antibacterial activity against *B. pumilus* and *B. cereus*. On the other hand, all tested compounds revealed poor growth inhibition of *P. aeruginosa* and low antifungal activity [75].

In other studies, the analgesic, antioxidant and anti-inflammatory activities of other mangiferin derivatives **28**–**34** (Figure 5) were explored [77,78]. Dar et al. [77] analyzed the analgesic and antioxidant activities of acetylated (**28**), methylated (**29**), and cinnamoylated (**30**) mangiferin, where compound **12** was acetylated to afford **14.** Mahendran et al. [78] observed the analgesic and anti-inflammatory activities of mangiferin with benzoyl (**31**), benzyl (**32**), and acetyl (**33**) groups.

**Figure 5.** Structures of mangiferin derivatives **28**–**34** with antioxidant, anti-inflammatory and analgesic activities.

The results demonstrated that mangiferin derivatives substituted with benzoyl (compound **32**) and acetyl (compound **34**) groups displayed better antioxidant activity than mangiferin (**21**) in lipid

peroxidation, p-NDA, deoxyribose and alkaline DMSO assays, while neither compound had analgesic nor anti-inflammatory activities. In all of these methods, standard drugs showed better activity than mangiferin and its derivatives **28**–**34** [78].

Mangiferin (**21**) is also known to possess antidiabetic activity [79,80]. This biological activity was further investigated for other mangiferin derivatives **35**–**42** and **46**–**53**, by Hu et al. [81,82] and **43**–**45** by Li et al. [83] (Figure 6). These works evaluated the properties of mangiferin derivatives as protein tyrosine phosphatase 1B (PTP1B) inhibitors in order to demonstrate their hypoglycaemic activity. The PTP1B has an important role in type 2 diabetes and obesity [84], which is primarily responsible for the dephosphorylation of the activated insulin receptor and thus downregulates insulin signalling [85]. For this reason, PTP1B inhibitors are a good strategy for diabetes mellitus treatment.

**Figure 6.** Structures of mangiferin derivatives **35**–**53** with antidiabetic activity.

According to Hu et al. [81,82], mangiferin (**21**) is a weak PTP1B inhibitor, whereas some derivatives such as **36**, **38** and **39** showed good inhibition of this protein. The SAR studies suggested that the substitution of free hydroxyl at C-3, C-6, C-7 of mangiferin (**21**) remarkably enhanced the inhibition, and the mono- or dichloro benzylated derivatives displayed better inhibitory activity than other groups [81,82]. However, further modification and biological studies are still in progress [81].

Li et al. [83] also demonstrated that the esterified-derivatives of mangiferin **43**–**45** (Figure 6) could repair damaged islet cells, and had higher lipid-solubility, and more potent hypoglycaemic activity than the mangiferin (**21**). The SAR studies indicated that the larger the esterification moieties or the higher lipid-solubility, the more potent hypoglycaemic activity was displayed by the derivative. Thus, esterification proved to be an effective way to improve the activity of mangiferin as a potential antidiabetic drug [83].

Correia-da-Silva et al. [86] developed new sulfated xanthones **54**–**57**, inspired by the mangiferin scaffold, to study their anticoagulant and antiplatelet properties (Figure 7). The synthetic approach included the sulfation of commercially available mangiferin affording mangiferin-2 ,3 ,4 ,6 tetrasulfate. It was found that an increase of the quantities of sulfating agent furnished the 2 ,3, 3 ,4 ,6,6 ,7-heptasulfated derivative. The precursor of the other derivatives could be a suitable xanthone scaffold, where 3,6-(*O*-β-glucopyranosyl)xanthone was obtained after deprotection of the glycosylated xanthone 3,6-(2,3,4,6-tetra-*O*-acetyl-β-D-glucopyranosyl)xanthone. The sulfation of 3,6-dihydroxyxanthone allowed preparation of persulfated 3,6-(O-β-glucopyranosyl)xanthone. Two polysulfated xanthonosides proved to be inhibitors of thrombosis, combining anticoagulant and antiplatelet effects in a single molecule [86].

**Figure 7.** Structures of mangiferin derivatives **54**–**57** with anticoagulant and antiplatelet activities.

#### 2.1.5. Derivatives of α-Mangostin

One of the most studied xanthones found in Nature is α-mangostin, isolated from tropical fruits of *Garcinia mangostana* which have been used for centuries in folk medicine to treat many conditions [87,88]. Several studies have reported its anticancer and antimicrobial activities, among others, which have prompted researchers all over the world to synthesize diverse derivatives [74,87–94].

One strategy concerned cationic antimicrobial peptides (CAMPs),amphipathic structures whose hydrophobic moiety penetrates the membrane core, while the cationic residues disrupt bacterial membranes [95–98]. Due to the manufacturing costs and poor stability of peptides, Koh et al. [95,99] developed small-molecules **58**–**70** with CAMP characteristics by combining the α-mangostin core with basic amino acids moieties (Figure 8). The purpose of the work was to confirm if lipophilic chains enhance the membrane permeability and to examine the role of the cationic moieties conjugating the xanthone scaffold with basic amino acids [99]. This strategy was also used to develop new anti-mycobacterial derivatives **65**–**70** [95], and by Lin et al. [97] for studies seeking new antimicrobial and hemolytic compounds.

**Figure 8.** Structures of derivatives of α-mangostin **58**–**70**.

According to the results, the amphiphilic xanthone derivatives **65**–**70** (Figure 8) possessed promising mycobacterial activity without resistance mechanisms, which may contribute to the development of an entirely new class of therapeutics for tuberculosis [95]. Besides the interesting activity of these compounds, the cationic and hydrophobic moieties enhanced the water-solubility, and also lead to high membrane selectivity and excellent antibacterial activity against Gram-positive bacterial strains, including methicillin-resistant Staphylococcus aureaus and vancomycin-resistant Enterococcus. It is important to point out that the membrane selectivity of these compounds was higher than several membrane-active antimicrobial agents in clinical trials [97,99].

#### 2.1.6. Derivatives of Caged Xanthones

The *Garcinia* genus contains caged xanthones which mainly occur in a few species like *G. morella*, *G. hanburyi*, *G. bracteata*, *G. gaudichaudii*, and *G. scortechinii*, widely distributed in Southeast Asia [100–103]. The caged core is responsible for the vast range of bioactivities of this class of compounds, such as anti-viral, and antibacterial effects, among others [104–107]. Many reports have described the potential antitumor activity of gambogic and morrelic acids [100,101,104,105,107,108]. According to this, many caged xanthones **71**–**155** have been synthesized and studied through the last years, with a diversity of purposes (Figures 9–14).

**Figure 9.** Structures of caged xanthones **71**–**82** with antimicrobial activity.

**Figure 10.** Structures of caged xanthone **83**–**96**, with antimalarial and antitumor activities.

**Figure 11.** Structures of caged xanthones **97**–**123**, with antitumor activity.

**Figure 12.** Structures of caged xanthones **124**–**138**, with antitumor activity.

**Figure 13.** Structures of caged xanthones **139**–**148**, with antitumor activity.

**Figure 14.** Structures of gambogic acid derivatives **149**–**155**, with antitumoral activity.

Chaiyakunvat et al. [105] inspired by the biological properties of caged xanthones, synthesized a few morrelic acid derivatives and evaluated their antimicrobial activity. They started with the synthesis of methylated morrellic acid and, afterwards, they synthesized derivatives **71**–**82** with amino acid moieties via solid-phase synthesis (Figure 9).

The morellic acid derivatives showing more inhibition of bacterial growth were the ones with an amino acid-containing hydrophobic side chain like **71**, **72**, **76**, **78** and **79** (Figure 9) [105]. This is in agreement with previous reports where the antimicrobial activity was higher in the structures with hydrophobic and/or aromatic amino acids [99,105].

Theodorakis et al. [106,109–111] synthesized new caged xanthones and studied their properties. They developed a Claisen/Diels–Alder reaction cascade that, in combination with a Pd(0)-catalyzed reverse prenylation, provided a rapid and efficient access to the caged xanthone pharmacophore. Afterwards, various A-ring oxygenated derivatives of cluvenone (**83**) were further synthesized and analyzed (Figure 10) [106,109–111].

The SAR studies showed that their activity could be substantially improved by attaching a triphenylphosphonium group at the A ring of the caged xanthone. Derivatives **93** and **94** (Figure 10) were found to be highly effective as antimalarials against *Plasmodium falciparum* [106]. The conjugation of these compounds with a phosphonium salt improved their efficacy, resulting in lead compounds with a promising therapeutic window [106]. It was suggested that, further modification of the caged xanthone could increase the selective cytotoxicity and lead to a promising lead candidate [106].

Cluvenone (**83**) was also reported to induce cell death via apoptosis, presenting similar cytotoxicity in multidrug-resistant and sensitive leukemia cells [109,110]. The caged xanthone derivatives proved to be active with cytotoxicity at low to sub-micromolar concentrations in solid and non-solid tumor cell lines, respectively. Additionally, they induced apoptosis in HUVE cells. Remarkably, similar IC50 values were obtained for the compounds tested in the HL-60 and HL-60/ADR cell lines, suggesting that these compounds were not subject to a drug resistance mechanism. Therefore, it was suggested that members of this family of compounds may have therapeutic potential in relapsed cancers typically resistant to standard chemotherapeutic agents. In addition, the cytotoxicity observed in HUVE cells suggested that these compounds may be interesting leads for the development of new angiogenesis inhibitors [111]. Elbel et al. [112] synthesized selected A-ring hydroxylated analogues and evaluated their effect on cell growth, mitochondrial fragmentation, mitochondriotoxicity and Hsp90 client protein degradation. They found out that both the C6 and C18 hydroxylated cluvenones inhibited the growth of CEM cells at low micromolar concentrations and induced cell death via the mitochondrial pathway. In addition, cluvenone (**83**) and the hydroxylated cluvenones induced Hsp90-dependent protein client degradation at low micromolar concentrations [112].


**Table 1.** Summary of the biological activities of synthetic CDXs inspired in natural xanthones.

Zhang et al. [113–116] synthesized a series of caged xanthone derivatives to improve the physicochemical properties and *in vivo* cytotoxic potency. For that, they relied on MAD28 synthesis and characterization of the derivatives. The structural modifications revealed that the presence of a carbamate moiety was useful for obtaining comparable cytotoxicity and improved aqueous solubility and permeability (Figure 11).

It is important to highlight that compound **137** (named DDO-6306, Figure 12) displays growth inhibition in Heps transplanted mice, and is now undergoing further evaluation as a candidate for cancer chemotherapy [115]. In a more recent study, compound **105** (Figure 11), considered as the lead compound and called MAD28, successfully led to the discovery of a novel series of natural-product-like triazole-bearing caged xanthones with improved drug-like properties as orally-active antitumor agents *in vivo* [115].

Regarding the caged xanthone derivatives containing carbamate scaffolds **109**–**123** (Figure 11), the results showed a potent antiproliferative activity and good physicochemical properties, which contributed to improving their *in vivo* activities. The compound **122** (DDO-6337) showed moderate inhibitory activity toward Hsp90 ATPase and resulted in the degradation of Hsp90 client proteins, such as HIF-1, which ultimately contributed to its antitumor and anti-angiogenesis activities [116].

Compounds **140**–**143** (Figure 13) exhibited micromolar inhibition against several cancer cell lines. Some interesting SAR considerations have been highlighted, such as the importance of the periphery gem-dimethyl groups in maintaining the anti-tumor activity, the effect of the substituent at C-1 position of B-ring on activity, since hydroxyl group at C-1 position enhanced the potency while prenyl group reduces it, and, that the change of hydroxyl or prenyl groups in carbons C-2, C-3 and C-4 had no significant effect on the anti-tumor activity. These events indicated that referred sites can be used to improve drug-like properties [114].

In another study, Miao et al. [117] developed small molecule entities inspired by the structure of gambogic acid. They focused on modifications of the prenyl moiety of the caged xanthones which led to synthesize seven derivatives **149**–**155** (Figure 14), which were tested for anti-tumor activity [117].

The SAR studies suggested that compounds **151**, **153**, and **154** showed similar cytotoxicity to gambogic acid against A549 cells, whereas compounds **149**–**151** and **152** were less active than gambogic acid. Although these experiments were preliminary, the results suggested that promising agents with anti-tumor activities could be obtained by modification at C-2 position of the B ring and at C-21/22 or C-23 position of the prenyl group in the caged scaffold. The formation of dihydroxy and epoxy groups of the double bond at C-21/22 and the introduction of an electron-withdrawing group at C-23 evidently affected the anti-proliferation activity [117]. In Table 1 a summary of the synthetic CDXs inspired in natural xanthones and their described biological activities are presented, as well as the associated references.

#### *2.2. Synthetic CDXs Obtained by Binding/Coupling Chiral Moieties to the Xanthone Scaffold*

Another approach to acquire synthetic CDXs is by binding chiral moieties to the xanthone scaffold using different strategies. The biological and pharmacological activities evaluated of the various CDXs obtained by this strategy are included in Table 2.

#### 2.2.1. XAA and DMXAA Analogues

DMXAA (5,6-dimethylxanthone-4-acetic acid, vadimezan, ASA404, **157**, Figure 15) is a simple carboxylated xanthone discovered by SAR studies involving a series of xanthone-4-acetic acids (XAA, **156**, Figure 15) related to the parent compound flavone acetic acid [121]. DMXAA is one of the most studied xanthones considering not only its remarkable pharmacological profile [122–131], but also its physicochemical and pharmacokinetic properties [130,132,133]. It is a tumor vascular-disrupting agent leading to a fast, vascular collapse and tumor necrosis by immunomodulation and cytokines induction. DMXAA (**157**) also demonstrated antiviral [134], antiplatelet and antithrombotic [135] activities. This synthetic xanthone is not chiral but, it is evident that structurally and from a biological

activity perspective, it may be an attractive scaffold for the development of other bioactive analogues and derivatives.

**Figure 15.** Structures of XAA (**156**), DMXAA (**157**) and chiral analogues **158**–**164**.

Rewcastle et al. [136], in 1991, synthetized the DMXAA chiral analogues 2-(5-methyl-9-oxo-9*H*xanthen-4-yl)propanoic acid (**158**) and 2-(9-oxo-9*H*-xanthen-4-yl)propanoic acid (**159**) as racemates (Figure 15). CDX **159** was synthetized via bromination of 4-ethylxanthenone followed by conversion of the resulting 4-(1-bromoethyl)xanthenone to compound **159** via the nitrile. CDX **158** was obtained by reaction of 2 -hydroxyacetophenone with benzyl chloride under phase-transfer conditions, giving 2 -benzyloxyacetophenone, which was converted successively to an alcohol, using NaBH4, and a chloride, with anhydrous CaCl2 and HCl. After three reaction steps, 2-(2-hydroxyphenyl)propanoic acid was obtained and then reacted with 2-iodo-3-methylbenzoic acid via a copper/TDA-catalysed condensation. Finally, the resulting diacid was ring-closed using H2SO4 to give the racemic compound **158**. For this CDX both enantiomers were separated by indirect method employing (*R*)-(−)-pantolactone as chiral resolving agent. The obtained mixture of diastereomers was separated by chromatography on silica gel. Further, hydrolysis of the esters under non-enolizing conditions afforded both enantiomers, (*S*)-(+)-**158** and (*R*)-(−)-**158** (Figure 15) [136].

The racemic compounds **158** and **159**, as well as enantiomers (*S*)-(+)-**158** and (*R*)-(-)-**158** were tested in *in vitro* and *in vivo* tumor assays [136]. It was found that all the compounds were active. Moreover, enantioselectivity was observed, being the *S*-(+)-enantiomer much more dose-dependent than the *R*-(-)-enantiomer. It was suggested that the enantiomers have different intrinsic activities, rather than differing in their metabolism [136]. CDX **159** had been tested previously as anti-inflammatory agent [137].

Marona et al. reported the synthesis [138] of three new chiral analogues of XAA **160**–**162** (Figure 15) and the evaluation of their cytotoxicity against J7774A.1 cells [139]. Compounds **160** and **161** were obtained by condensation of 2-hydroxyxanthone and 2-methyl-6-hydroxyxanthone, respectively, with α-bromopropionic acid and compound **162** by esterification of compound **161** [138]. Regarding the biological activity tested, it was found that all CDXs showed weak cytotoxicity [139].

Recently, Zelaszczyk et al. [140] synthesized two new chiral XAA derivatives **163** and **164** (Figure 15) by the reaction of 3-hydroxyxanthone with ethyl 2-bromopropanoate, followed by ester hydrolysis [140]. These compounds were found to have anti-inflammatory and analgesic activities [34,140].

#### 2.2.2. Synthetic Aminoalkanolic CDXs

Our group has a vast experience in synthesis and biological/pharmacological activity evaluation of xanthone derivatives [86,141–147] and, recently, reported the synthesis of a library of CDXs **165-179** in an enantiomerically pure form (Figures 16 and 17) [148,149]. Among the synthesized CDXs, the compounds **166-171** and **174**–**179** are aminoalkanolic, while CDXs **165, 172** and **173** comprise of simple amines with a *p*-tolyl moiety (compounds **165** and **173**) or an aminoester (compound **172**) [148,149].

**Figure 16.** Structures of aminoalkanolic CDXs **166**–**171** and analogues **165** and **172**.

**Figure 17.** Structures of aminoalkanolic CDXs **174**–**179** and analogue **173**.

Considering that carboxyxanthone derivatives are suitable molecular entities to perform molecular modifications to obtain new bioactive derivatives [34], the synthesis of all CDXs **165-179** were achieved by using two carboxyxanthone derivatives as substrates, namely 6-methoxy-9-oxo-9*H*-xanthene-2 carboxylic acid and 2-((9-oxo-9*H*-xanthen-3-yl)oxy)acetic acid. The synthetic strategy used was the coupling of the carboxyxanthone derivatives with both enantiomers of eight commercially available chiral building blocks, using *O*-(benzotriazol-1-yl)-*N*,*N*,*N* ,*N* -tetramethyluronium tetrafluoroborate (TBTU) as coupling reagent [148,149]. TBTU has been widely used as efficient reagent for the synthesis of diverse classes of compounds, including peptides [150], esters [151,152], phenylhydrazides [153], acid azides [154], among others. However, this was the first report of the use of TBTU to synthesize CDXs [149]. All used commercial chiral blocks included both enantiomers of enantiomerically pure

building blocks with no tendency towards racemization or enantiomeric interconversion and having a primary amine as reactive group for amide formation. Amino alcohols, amines, and amino esters were selected. The coupling reactions were performed at room temperature, showing short reactions times and excellent yields (ranging from 94% to 99%) [148,149]. The synthetic methodology used to obtain the referred CDXs provided to be very efficient, broad-scope applicability, and operationally simplest. Moreover, it was found that the synthesis of the CDXs was easily scaled-up for both enantiomers in order to obtain available quantities for biological and pharmacological assays as well as other applications.

LC using different types of CSPs, namely polysaccharide-based [149], macrocyclic antibiotics [155,156], and Pirkle-type [157,158] was used for enantioresolution studies and determination of the enantiomeric purity of the synthesized CDXs. The enantioselective LC method using polysaccharide-based CSPs under multimodal elution conditions afforded very high resolutions with short chromatographic runs. The best performances were achieved on amylose *tris*-3,5 dimethylphenylcarbamate stationary phase under polar organic elution conditions. The resolution achieved allowed the determination of the enantiomeric purity for all CDXs, affording values higher than 99% [149].

Considering the macrocyclic antibiotic-based CSPs, four commercially available columns were used, namely Chirobiotic TTM, Chirobiotic RTM, Chirobiotic VTM and Chirobiotic TAGTM, under multimodal elution conditions. The optimized chiral LC conditions were successfully employed for the accurate determination of the enantiomeric purity, always higher than 99%. The studies also explored the influence of different mobile phase compositions and pH on the chromatographic parameters as well as of the structural features of the CDXs on their chiral discrimination by the macrocyclic antibiotic-based CSPs [155,156].

Regarding the Pirkle-type CSPs, the (*S*,*S*)-Whelk-O1® CSP showed the best performance for the resolution of the CDXs evaluated, presenting very high enantioselectivity for CDXs with aromatic group linked to the chiral moiety. Polar organic elution mode presented the best chromatographic parameters allowing good resolutions and lower run time [157,158].

The overall results proved that, for the same enantiomeric pair of CDXs, the polysaccharide-based CSPs were the most efficient to separate the enantiomers of this group of compounds, since all the CDXs were enantioseparated with excellent enantioselectivity and resolution [159].

For each enantiomeric pair of the synthesized CDXs **165**–**179** the *in vitro* growth inhibitory activity in three human tumor cell lines, A375-C5 (melanoma), MCF-7 (breast adenocarcinoma), and NCI-H460 (non-small cell lung cancer), were evaluated [149]. The results obtained demonstrate that some CDXs exhibited interesting growth inhibitory effects on the tumor cell lines. The most active CDX in all human tumor cell lines was compound (**1***R*,**2***S*)-**179**. Furthermore, it is important to highlight that the effects on the growth of the human tumor cell lines were attributed not only to the nature and positions of substituents on the xanthonic scaffold, but also, in some cases, were associated with the stereochemistry of the CDXs concerning enantioselectivity results. Interesting examples of enantioselectivity were observed between the enantiomeric pairs of CDXs **165**, **167**, and **171** [149].

Recently, other enantioselectivity studies associated with biological activity were conducted, specifically the *in vitro* and *in silico* inhibition of cyclooxygenases (COX-1 and COX-2) for the enantiomeric pairs of CDXs **166**, **168** and **169** [160]. All CDXs exhibited COX-1 and COX-2 inhibition being, in general, the inhibitory effects similar for both COXs. The only exception was the enantiomeric pair of compound **166**, being the (*R*)-(+)-enantiomer more active at inhibiting COX-2 than COX-1. Interestingly, all pairs demonstrated enantioselectivity for COX-1. Concerning COX-2, the % of inhibition was also dependent of the stereochemistry being (*S*)-(-)-**166** and (*S*)-(+)-**169** more active [160].

Additionally, for the same enantiomeric pairs of CDXs **166**, **168** and **169**, human serum albumin (HSA) binding affinity was evaluated by spectrofluorimetry and *in silico* studies, by a docking approach [160]. All CDXs demonstrated to bind with high affinity to HSA and enantioselectivity was observed for compound **168**.

Taking into account that these CDXs have molecular moieties structurally very similar to local anaesthetics, the ability to block compound action potentials (CAP) at the isolated rat sciatic nerve was also investigated [161]. The CDXs (*S*)-**165,** (*S*)-**166** and (*S*)-**167** were chosen for biological evaluation and the results suggested that the nerve conduction blockade might result predominantly from an action on Na<sup>+</sup> ionic currents. It was also investigated if the CDXs could prevent hypotonic haemolysis on rat erythrocytes. However, data suggested that all tested CDXs caused no significant protection against hypotonic when applied in concentrations high enough to block the sciatic nerve conduction in the rat [161].

Besides the potential as new drugs, CDXs present structural features with interest as chiral selectors for LC [23]. In this context, some of these small molecules ((*S*)-**167**, (*R*)-**168**, (*S*)-**168**, (*R*)-**176**, (1*R*,2*S*)-**179** and (1*S*,2*R*)-**179**) were selected and bound to a chromatographic support for a new application as CSPs in LC [24]. The new xanthonic CSPs afforded promising enantioresolution results, high stability and reproducibility. Accordingly, CDXs have important applications in the field of Medicinal Chemistry, not only as candidates for potential new drugs but also as analytical tools for enantioseparation in LC [37].

Recently, our group also performed enantioselectivity studies with chiral thioxanthones, *S*-analogues of xanthones, as modulators of P-glycoprotein (P-gp) [162]. It was found that one of the enantiomers modulated P-gp expression differently from its pair.

Marona *et col.* [163–169] also described the synthesis and biological activity evaluation of a series of aminoalkanolic CDXs **180**–**205**, **217**–**220**, **238** (Figures 18–20). More recently, they synthesized the aminoalkanolic CDXs **206**–**243** (Figures 18–20), being tested for anticonvulsant, antimicrobial and cardiovascular activities [170–174]. CDXs **180**–**194**, **198**–**202** were synthesized by condensation of an appropriate 2-bromomethylxanthone or 2-bromomethyl-7-chloroxanthone with the adequate aminoalkanol in toluene, in the presence of anhydrous potassium carbonate [163,164,175]. The exchange of secondary amino group of compound **189** for a tertiary amino group (compound **193**) was generated by reductive *N*-methylation [163]. Compound **210**, however, was synthesized through the chlorination of compound **186** with thionyl chloride in toluene [170].

Compounds **195**–**197**, **221**–**223**, **228**–**232** and **241**–**243** were synthesized by the aminolysis of 3- or 4-((oxiran-2-yl)methoxy)xanthone in *n*-propanol, or by the amination of 2-methyl-6-hydroxy-xanthone using propylene epichlorohydrin, in the presence of sodium hydroxide and water [167,172–174].

The synthesis of compounds **203**–**204**, **206**–**209**, **213**–**216**, **224**–**227** and **240** involved a multi-step process. At first, a substituted benzoic acid reacted with 2- or 4-methylphenol in two steps involving an Ullmann condensation and electrophilic addition. The intermediate methyl derivatives of substituted xanthone were used in the reaction with *N*-bromosuccinimide giving appropriate bromide derivatives. The last step comprised an aminolysis by means of appropriate aminoalkanol carried out in toluene in the presence of anhydrous K2CO3 [171,176].

Most of the synthesized CDXs **180**–**209**, **211**–**220**, **224**–**227**, **233**–**238**, **240** were evaluated for anticonvulsant activity [163–167,169,171,175,177–179]. The studies involved three kind of tests: maximal electroshock-induced seizures (MES), subcutaneous pentetrazole seizure threshold (scMet), and neurological toxicity (TOX).

In one of the first MES assays in mice, 2-amino-1-propanol-, 1-amino-2-propanol and 1-amino-2-butanol derivatives of 6-methoxy- or 6-chloroxanthone were the most interesting compounds. In fact, the results indicated that compound **184** was the most active [163]. Further study compared the anticonvulsant activity of CDX **184** (racemate) with the single enantiomers ((*R*)-**184**, (*S*)-**184**). All the compounds showed excellent results, and no significant differences were observed in the anticonvulsant activity of the single enantiomers compared with the racemate [166].

Additionally, the enantiomeric purity of (*R*)-**184** and (*S*)-**184** was determined by a liquid chromatography–mass spectrometry method with an electrospray ionization interface (ESI-LC/MS). The separation of the two enantiomers ((*R*)-**184** and (*S*)-**184**) was carried out on the commercially

available cellulose *tris*-(3,5-dimethylphenyl carbamate) CSP, Chiralcel® OD-RH, giving enantiomeric purity values higher than 99.9% [166].

Interesting anticonvulsant results were also observed with alkanolic chiral derivatives **189** and **193** of 7-chloroxanthone, which displayed anti-MES activity corresponding with that for phenytoin, carbamazepine and valproate [164]. Moreover, it is important to highlight that in this study some cases of enantioselectivity were observed. For example, although enantiomers (*R*)-**189** and (*S*)-**189** showed anticonvulsant activity, the (*S*)-enantiomer was more neurotoxic. Furthermore, considering the compound **193** (racemate), the (*R*)-enantiomer ((*R*)-**193**) in comparison to (*S*)-enantiomer ((S)-**193**) showed higher anticonvulsant activity [164,177]. Recently, several aminoalkanolic chiral derivatives substituted in positions 2, 3, or 4 evaluated for their anticonvulsant activities with 2-xanthonoxy derivatives **206, 207** and **209**, being the most active and exhibiting neurotoxicity after 30 min after administration at the dose of 100 mg/kg [171]. A further study including chiral aminoalkanol derivatives substituted in position 2 of the xanthonic scaffold (structures not shown) also emphasized the importance to examine biologically enantiomers other than racemates [178].

**Figure 18.** Structures of 2-aminoalkanolic CDXs **180**–**210**.

**Figure 19.** Structures of 4-aminoalkanolic CDXs **211**–**232**.

**Figure 20.** Structures of 3-aminoalkanolic CDXs **233**–**243**.

Additionally, a structure-anticonvulsant activity relationship study was described including series of aminoalkanol derivatives **204, 213**–**216** of 6-methoxy- or 7-chloro-2-methylxanthone as well as 6-methoxy-4-methylxanthone [176]. All the compounds showed activity in the MES screening which is recognized as one of the two most widely used seizures models for early identification of candidates as anticonvulsants. The tested compounds were evaluated in the form of racemic mixture and some additionally in the form of single enantiomers to determine stereochemistry-activity relationship. In fact, as demonstrated before [176], stereochemistry is one of the factors that can potentially influenced anticonvulsant activity of the CDXs. However, considering anti-MES activity it was not possible to establish relationship between stereochemistry and anticonvulsant properties because all sets of compounds gave different results. Racemate and enantiomers showed either similar results or diverged in duration of activity or lower effective doses. However, the anticonvulsant activity was associated with both aminoalkanol type and respective configuration as well as the location of substitution in the xanthone scaffold [176]. The overall results from several studies of Marona *et col.* [163–167,171,175–179] are quite encouraging and suggested that in the group of xanthone derivatives new potential anticonvulsants might be found.

Some of alkanolic CDXs were also evaluated for cardiovascular activity [167,173–175,179,180], including antiarrhythmic, hypotensive, α1- and β1-adrenergic blocking activities, effect on the normal electrocardiogram and influence on the central nervous system (CNS) [169]. Among the investigated compounds, some of them exhibited significant antiarrhythmic and/or hypotensive activity. For example, compounds **218** and **219** revealed the strongest anti-arrhythmic activity in the adrenaline-induced model of arrhythmia. Additionally, compound **219** was also the most potent concerning hypotensive activity [169]. Recently, compounds **231** and **232** were also evaluated for their cardiovascular activity, through both α1- and β1-adrenergic blocking. These compounds, classified as beta-blockers with vasodilating properties, exhibited also hypotensive vasorelaxant activities comparable to those of carvedilol [173].

The effects on platelet aggregation of racemic CDXs **180**–**182**, **195** and **196**, and the enantiomeric pure CDX (*R*)-**193** were evaluated and showed motivating results. The most active and promising compound was (*R*)-**193** which nearly completely inhibited the thrombin aggregation concentration (TAC) [181]. The results indicated that the presence of the 2-*N*-methylamino-1-butanol at position 2 and the chloride atom at the 7-position of the xanthone scaffold promoted antiplatelets activity.

Alkanolic CDXs **180**, **183**, **184** and **194** were used to assess mutagenic and antimutagenic activities in assays using the *Vibrio harveyi* test [182]. According to the obtained results, the most beneficial mutagenic and antimutagenic profiles were observed for compound **194**. This compound was shown to have strong antimutagenic activity towards the BB7 *V. harveyi* strain while failing to induce mutagenic responses in the tested strains. The modification of the chemical structure of compound **194** through chlorination of the hydroxyl group, improved considerably the antimutagenic activity maintaining the inability to induce mutagenic responses in the strains. Thus, antimutagenic potency reached a maximum with the presence of tertiary amine and one chloride atom in the side chain. Minimal activity was showed to compound **184** and no antimutagenic activity was observed for compound **180** [182].

In recent years, several aminoalkanolic CDXs **210, 228**–**230, 241**–**242** have been evaluated for their antibacterial and antifungal activities [170,172]. Compound **210** was evaluated against several dermatophytes, moulds and yeasts, exhibiting good activity results against ten strains of the *Aspergillus fumigatus, niger* and *flavus* moulds, being among the tested compounds the most active [170]. This CDX also exhibited moderate to good activity against some strains of the *Trichophyton mentagrophytes* and *rubum* dermatophytes, while being inactive against the *Candida albicans* yeast [170].

The antimicrobial activity of CDXs **228**–**230, 241**–**242** against 12 strains of the bacteria *Helicobacter pylori* was evaluated through the Kirby-Bauer method, by measuring the diameters of inhibition zones, showing that compounds **228**, **230**, **241**–**242** exhibited strong activity against the strains ATCC 43504, 700684 and 43504. Actually, those CDXs were considered the best of the tested compounds, while compound **229**, demonstrated weak antibacterial activity [172].

In a recent study on the influence of reactive oxygen species (ROS) in the anticancer activity of aminoalkanolic derivatives, it was reported that in the case of CDXs **219** and **223**, ROS were of great importance to their proapoptotic activity [183]. These encouraging results suggested that aminoalkanolic CDXs might be interesting structures for potential use in anticancer therapy [183].

#### 2.2.3. CDXs Conjuged with Amines, Amino Esters and Amino Acids

Inspired by natural xanthones properties, Rakesh et al. [184] synthesize xanthone derivatives with conjugated L-amino acids (**244**–**263**, Figure 21), to determine the corresponding antimicrobial and anti-inflammatory activities. The same research group recently reported the evaluation of *in vitro* anticancer activity of those compounds, against three different cancer cell lines, MCF-7, MDA-MB-435 and A549, validated by DNA binding and molecular docking approaches [185].

**Figure 21.** Structures of CDXs **244**–**263**, with antimicrobial and anti-inflammatory activities.

The synthetic strategy used to obtain the compounds was accomplished using 2-chlorobenzoic acid and resorcinol in anhydrous zinc chloride to give 2-chlorophenyl-(2,4-dihydroxyphenyl) methanone and cyclized with DMSO and NaOH to give 3-hydroxyxanthone. This chemical substrate was, afterwards, conjugated with different protected amino acids [184,185].

The compounds with the best antimicrobial and anti-inflammatory activities were those conjugated with L-phenylalanine, L-tyrosine and L-tryptophan, followed by compounds conjugated with L-cysteine, L-methionine and L-proline [184]. Additionally, the compounds with amino acids with high aromaticity and hydrophobicity, presented more stable amphiphilic structures.

The antimicrobial effect comes from the penetration of the amino acid hydrophobic chains in the bacterial membranes where the cationic moiety of the amino acids interacts with the membrane phospholipids disturbing bacterial membrane. This strategy proved to be effective to develop new antimicrobial agents [184], as the microorganism die without developing mutations or resulting in loss of recognition by the antibiotics [96]. Other studies accomplished the same conclusions [168,172,186]. Regarding the antitumor activity, the compounds (*S*)-**248**, (*S*)-**249** and (*S*)-**250** exhibited potent inhibition against the tested tumor cell lines as well as DNA binding. The SAR studies showed that the aromatic and hydrophobic amino acids, such as phenylalanine, tyrosine, and tryptophan, favored the DNA binding studies and antitumor activities; whereas, aliphatic amino acids showed

lower activity. The derivatives with glycine, alanine, valine, leucine, and isoleucine showed less or moderate anticancer properties [185].

#### 2.2.4. CDXs Containing Piperazine Moieties and Analogues

Several CDXs containing piperazine moieties (**267**–**270**, **272**–**294**) and analogues (**264**, **265**, **271**) were synthesized (Figures 22–24) and their biological activity evaluated by Marona's group [165,167, 168,172,173,181,187–192].

2-Hydroxyxanthone was the building block used to synthesize compounds **267**–**273** and **277** using epichorhydrin in the presence of pyridine [165]. Compounds **274**–**276, 278, 280**–**282** were synthesized by amination of 2-(2-bromoethoxy)-9*H*-xanthen-9-one and derivatives in *n*-propanol or toluene in the presence of K2CO3 [187]. The compounds **283**–**284**, **292**–**293** were obtained by aminolysis of 4-[(2,3-epoxy)propoxy]xanthone with appropriate 1-piperazine derivatives in *n*-propanol, while **279** was synthesized using the same methodology through the aminolysis of 3-[(2,3-epoxy)propoxy]xanthone [172,173,188].

Chiral compounds **287**–**291** were obtained by amination of respective parent compounds [189] with appropriate amines in *n*-propanol. In addition, compound **287** was obtained from compound **286** by acetylation. In order to optimize synthetic methodologies, CDX **286** was obtained using an alternative method including (*R*,*S*)-4-(3-chloro-2-hydroxypropoxy)-9*H*-xanthen-9-one as intermediate [189].

**Figure 22.** Structures of 2-piperazine derivatives **264**–**276**.

**Figure 23.** Structures of 3-piperazine derivatives **277**–**279**.

**Figure 24.** Structures of 4-piperazine derivatives **280**–**294**.

CDXs **267**–**273** and **277** were evaluated for anticonvulsant activity in the MES- and subcutaneous pentylenetrazole-induzed seizures in mice and rats [165]. Among them, the most promising compound was CDX **268** which was active in both the anticonvulsant tests.

Moreover, the influence on the platelet aggregation of CDXs **264**, **265, 269** and **271** was evaluated by using adenosine-5 -diphosphate (ADP), sodium arachidonate (AA) or thrombin as the aggregating agents [181]. CDXs **265** and **271** were active, inducing 80-90% inhibition of thrombin-stimulated platelet aggregation.

Considering that the xanthone itself proved to possess vasorelaxing properties in thoracic aorta isolated from rats [193] and the strongest hypotensive effects were observed for compounds containing piperazine moiety [189], several compounds that combine the xanthone nucleus and piperazine rings (**274**–**276, 278**–**291**) were evaluated for anti-arrhythmic and/or antihypertensive activities. It is important to emphasize that CDXs **274** and **280** demonstrated to possess significant anti-arrhythmic activity in the adrenaline-induced model of arrhythmia [187]. The strongest hypotensive activity which persisted for 60 min was also associated to compound **89**.

Additionally, in another study related to the same biological activities, compound **96** was the most promising considering its effect on circulatory system. Moreover, this CDX diminished arterial blood pressure by about 40% during one hour [188].

A recent cardiovascular activity study of several CDXs **286**–**291** was described, including the following pharmacological experiments: the binding affinity for adrenoceptors, the influence on the normal electrocardiogram, the effect on the arterial blood pressure and prophylactic antiarrhythmic activity in adrenaline induced model of arrhythmia (rats, iv) [189]. The CDXs **286** and **287** revealed to act as potential antiarrhythmics in adrenaline induced model of arrhythmia in rats after intravenous injection. In another study, CDX **279** was reported as a promising hypotensive with its activity attributed to the blockage of α1-adrenoreceptors [173]. The results obtained were quite encouraging and suggested that in the group of xanthone derivatives new potential antiarrhythmics and hypotensives might be found.


**Table 2.** Summary of the biological activities of synthetic CDXs obtained by binding/coupling chiral moieties to the xanthone scaffold.

A series of some chiral derivatives of 2-xanthones **267**–**273, 277** with a piperazine moiety was evaluated for their activity against *M. tuberculosis*. The highest level of activity against *M. tuberculosis* was observed for compound **270**, which exhibited 94% growth inhibition. This compound was also examined for its anti-*M. avium* activity as well as cytotoxicity, showing insignificant anti *M. avium* activity and cytotoxic effects [188].

Recently, CDXs **284**–**285, 290, 292, 293** were evaluated for their antibacterial activity against 12 strains of the bacteria *Helicobacter pylori* [172]. CDX **285** showed strong activity against *H. pylori* strain ATCC 43504 and 700684, being the only compound to show higher activity against clarithromycin-resistant *H. Pylori* strains, than to the one resistant to metronidazole [172].

#### 2.2.5. CDXs Containing Other Moieties

Recently, Cherkadu et al. [194] reported the synthesis of some CDXs **295**–**310** (Figure 25) containing moieties other than piperazine and aminoalcohols [194]. The synthesis of these CDXs, a series of 2-(aminobenzothiazol)methylxanthones, was performed through the reaction of 3-hydroxy-xanthone, aromatic aldehydes and 2-aminobenzothiazoles in DMF at 120◦C with FeCl3 as catalyst [194]. To the best of our knowledge, no biological activities were reported for those CDXs.

A summary of the synthetic CDXs obtained by binding/coupling chiral moieties to the xanthone scaffold, their biological activities as well as the associated references are presented in Table 2.

**Figure 25.** Structures of CDXs **295**–**310**.

#### **3. Conclusions**

The synthetic CDXs, inspired in natural sources, and obtained by coupling chiral moieties to the xanthone scaffold, demonstrated potential to perform a large variety of biological activities, including antitumor, antimicrobial, anticonvulsant, antimalarial, anti-inflammatory, antiplatelet, anti- thrombotic, antipyretic, analgesic, antioxidant, antidiabetic, anticoagulant, among others. Nevertheless, for this family of compounds the main biological activities reported were antitumor and antimicrobial.

The more studied chiral moieties were amines, amino alcohols and amino acids. The cationic moieties of the amino acids have been indicated as a good approach to develop new antimicrobial agents both for CDXs inspired in natural xanthones and obtained by coupling chiral moieties to the xanthone scaffold. Regarding enantioselectivity, some studies reported the importance in SAR studies, but the majority neglected the influence of stereochemistry in the biological activity.

**Author Contributions:** C.F, M.L.C, J.R. and J.A collected the primary data and contributed in writing of the manuscript. M.E.T., C.F. and M.M.M.P. supervised the development of the manuscript, and assisted in data interpretation, manuscript evaluation, and editing.

**Funding:** This work was supported by the Strategic Funding UID/Multi/04423/2013 through national funds provided by FCT – Foundation for Science and Technology and European Regional Development Fund (ERDF), in the framework of the programme PT2020, the project PTDC/MAR-BIO/4694/2014 (reference POCI-01-0145-FEDER-016790; Project 3599 – Promover a Produção Científica e Desenvolvimento Tecnológico e a Constituição de Redes Temáticas (3599-PPCDT)) as well as by Project No. POCI-01-0145-FEDER-028736, co-financed by COMPETE 2020, under the PORTUGAL 2020 Partnership Agreement, through the European Regional Development Fund (ERDF), and CHIRALXANT-CESPU-2018.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
