**1. Introduction**

Natural heterocyclic products consisting of fused benzene and α-pyrone rings are designed as coumarins [1]. The great structural diversity of coumarins can be found both in simple coumarins, whose chemical structures contain only two rings, and in coumarins that contain an additional ring, such as furano- and pyranocoumarins. In all coumarin scaffolds, hydroxy or methoxy groups at position 7 are common structural features [2].

Natural coumarins are widespread in edible and medicinal plants as secondary metabolites [3]. These chemicals display broad biological activities [4], including antioxidant [5], antibacterial [6], antiviral [7], anti-inflammatory [8], antidepressant [9], and antitumoral activities [10], among others. In addition to other secondary metabolites, the biosynthesis of coumarins is controlled by several enzymes, and their accumulation is directly influenced by biotic and abiotic factors [11]. Given the wide pharmaceutical, chemical, and alimentary importance of coumarins, several alternative ways to achieve them have been developed in an attempt to replace the natural compound obtained by bioprospecting [12,13]. As a result, limitations associated with obtaining such substances in low yields from natural sources may be circumvented.

Since the first synthetic coumarin [14], coumarin-like compounds have been mainly used in the pharmaceutical industry as precursors for the synthesis of anticoagulant drugs or in the production of fragrances. Several chemical compounds bearing a coumarin moiety continue to be produced by using synthetic methodologies [15,16], but some specific structural modifications and substituents insertions are barely performed. Sometimes, the synthetic methods require corrosive catalysts and long reaction times, and they generally generate byproducts along with the desired product [17].

Chemical derivatization using microorganisms (defined as biotransformation) represents a powerful tool for obtaining derivatives for fine chemical, pharmaceutical, and agrochemical industries according to green chemistry principles [18]. Biotransformation using microorganisms may operate in mild conditions (neutral pH, room temperature, and atmospheric pressure), and generally, it is conducive to high regio-, stereo-, and chemoselectivity at low cost [19]. Biotransformation processes have an interesting diversity, which allows for diverse products to be obtained from a single substrate [20]. Moreover, some chemical transformations that cannot be performed through the traditional synthetic methods are readily obtained using the biotransformation approach in ecofriendly reactions [21]. Within this context, the microbiological transformation of diterpenes has been studied to achieve the specific hydroxylation of unactivated C–H bonds, which is difficult to prepare using chemical methods [22].

Microbial biotransformation can be described as a reaction or a set of simultaneous reactions in which a precursor molecule is converted, rather than a fermentation process where molecules are produced from a carbon and energy source. Biotransformation could involve the use of enzymes or whole cells, or combinations thereof. In general, whole cells are more popular than isolated enzymes in industrial biotransformation [23] because the former allows for a great quantity of catalysts in small volumes and high turnover rates of enzymes and cofactors. Besides, filamentous fungi are especially useful as industrial enzyme sources because of their broad diversity in the production of proteins, which are quickly secreted [24].

Recently, our research group published a review on the biotransformation of simple, furano-, and pyranocoumarins using different genera of filamentous fungi [25]. The survey showed that *Cunninghamella* sp. and *Aspergillus* sp. are the most common genera used as catalysts in the biotransformation of several types of coumarins. These findings prompted us to investigate the chemical specificities of *Cunninghamella* and *Aspergillus* genera in transforming coumarin cores.

Therefore, the present study reports on the biotransformation of a variety of coumarin compounds using *Cunninghamella elegans* ATCC 10028b and *Aspergillus brasiliensis* ATCC 16404. Our results provide an interesting way of mapping these microbial systems onto the derivatization of coumarin scaffolds. The approach used by us highlights some chemical features in the coumarin structures that allow for their biotransformation. Additionally, the biotransformation of coumarins using the selected fungi strains occurred under chemo- and stereoselectivity. One of the assays led to a selective hydroxylation at an unexpected position of the coumarin core.

#### **2. Results and Discussion**

Coumarins represent an important class of natural products and also synthetic oxygen-containing heterocycles. Several reports have claimed that the most common structural pattern in natural coumarins is hydroxylation at C-7. Based on this, the present study focused on an evaluation of the biotransformation of a panel of coumarins that contained one or two hydroxyl groups in the aromatic ring.

Initial screening of the biotransformation of nine coumarin compounds (**1** to **9**, Figure 1) was carried out using the filamentous fungi *Cunninghamella elegans* ATCC 10028b and *Aspergillus brasiliensis* ATCC 16404. The biotransformations were carried out for 72 h, and samples were analyzed using HPLC every 24 h. Both fungi were able to transform only two of the coumarins (**4** and **7**) after 72 h of incubation.

**Figure 1.** Chemical structures of coumarins used as substrates in biotransformation through *Cunninghamella elegans* ATCC 10028b and *Aspergillus brasiliensis* ATCC 16404.

HPLC analysis (Figure 2) showed that *C. elegans* and *A. brasiliensis* biotransformed (exclusively) the coumarins **4** and **7** into more polar derivatives. The chemical structures of the coumarin substrates **4** and **7** contained two common features: aromatic rings from both contained only one hydroxyl group at C-7, and both coumarins contained a bulky group at C-4. None of the coumarins whose aromatic ring contained two hydroxyl groups (**2**, **3**, **5**, **6**, **8**, and **9**) were biotransformed by *C. elegans* or *A. brasiliensis*. Moreover, none of the coumarins with a methyl group at C-4 (**1**, **2**, and **3**) were biotransformed by the two evaluated filamentous fungi.

**Figure 2.** Reverse-phase HPLC elution profiles (λ = 211 nm) of the ethyl acetate extracts of *C. elegans* ATCC 10028b cultures incubated for 72 h with coumarin **4** (chromatogram A) and *A. brasiliensis* ATCC 16404 cultures incubated for 72 h with coumarin **7** (chromatogram B). The derivatives of coumarins **4** and **7** can be visualized in peaks **C1** and **C2**, respectively. AU: absorbance unit.

A more detailed analysis of the HPLC chemical profiles of the crude extracts of the biotransformations catalyzed by *C. elegans* and *A. brasiliensis* showed that both fungi biotransformed coumarins **4** and **7** into the same derivatives. However, the yields of derivatives **C1** and **C2** were different. Concerning the biotransformation of coumarin **4**, the yield of the assay catalyzed by *C. elegans* was five-fold higher than that catalyzed by *A. brasiliensis*. On the other hand, *A. brasiliensis* was more efficient in the biotransformation of coumarin **7**, generating almost double the yield. Similarly to our study, Lee and coworkers reported the biotransformation of isoflavone by *C. elegans* and *A. niger* [26]. The authors highlighted that *A. niger* gave a more complex metabolite profile than did *C. elegans*, but some derivatives were found in the crude extracts from both fungi.

Next, the biotransformation experiments were repeated to facilitate the isolation and characterization of **C1** and **C2**. Scale-up biotransformations of coumarin-related compounds **4** and **7** by *C. elegans* and *A. brasiliensis*, respectively, led to the isolation of their main derivatives.

The derivative isolated from the biotransformation of coumarin **4** was identified as 7,9-dihydroxy-2,3-dihydrocyclopenta[*c*]chromen-4(1*H*)-one (**C1**), which is reported for the first time in the present study. Its molecular formula was determined as C12H10O4 by HRESIMS (*m*/*z* 217.0507 [M − H]−, Figure S6 in the Supplementary Materials), indicating that one new hydroxyl group was introduced into **4** as a result of the biotransformation. A comparison between the **4** and **C1** 1H NMR spectra (Figure S1 in the Supplementary Materials) showed that their chemical structures were very similar, except for C-9 (δ 75.1) in the **C1** chemical structure. The chemical shift of C-9 (δ 75.1) and the presence of only one hydrogen attached at C-9 (δ 5.20–5.22, ddd, H-9) indicated hydroxylation at this position. Heteronuclear HSQC and HMBC analysis, along with 13C NMR data analysis (see spectra in Figures S2–S4 in the Supplementary Materials), allowed for the unequivocal structural identification of **C1** (Figure 3). The substitution of one hydrogen for the hydroxyl group at C-9 explained the greater polarity of **C1** compared to its precursor, **4** (see chromatogram A, Figure 2).

The hydroxylation at C-9 generated a chiral center in the **C1** chemical structure, which explained the multiplicity of the diastereotopic hydrogens attached at C-10 (δ 2.42–2.48, dddd, H-10a; and 2.00–2.05, dddd, H-10b) and C-11 (δ 3.17–3.23, dddd, H-11a; and 2.95–3.00 ddd, H-11b). The ddd δ 5.20–5.22 was attributed to the enantiotopic hydrogen attached at C-9. The optical rotation measurement of the **C1** sample ([α] 20D = <sup>−</sup>16◦) showed that the biotransformation approach used herein was stereoselective with a preference for levogyre stereoisomer production.

The magnitude of the *J* between two adjacent C–H bonds (<sup>3</sup>*JHH*) is directly dependent on the dihedral angle α between these two bonds [27]. Therefore, the *J* analysis of the 1H NMR signals of the hydrogens belonging to the **C1** cyclopentanol ring led to some conclusions about the relative stereochemistry of its substituents. The largest *J* value of ddd δ 5.20–5.22 (<sup>3</sup>*JHH* = 6.9 Hz) was attributed to the vicinal coupling between H-9 and H-10a (δ 2.42–2.48) that should occupy the opposite face of the cyclopentanol ring. By the same rationale, the middle *J* value (<sup>3</sup>*JHH* = 2.2 Hz) was attributed to the vicinal coupling between H-9 and H-10b (δ 2.00–2.05) that should occupy the same face of the cyclopentanol ring. Finally, the smallest *J* value of the ddd (4*JHH* = 1.4 Hz) was attributed to the W-coupling between H-9 and H-11a (δ 3.17–3.23). The W-coupling was herein confirmed through COSY homonuclear analysis (Figure S5 in the Supplementary Materials).

**Figure 3.** Chemical structures of **C1** and **C2**, which were isolated from the biotransformation of coumarins **4** and **7***,* respectively.

Several coumarins and their derivatives have displayed interesting biological activities, and they are useful as a starting point for drug development [25]. The biotransformation approach employed in the present study provided chemo- and stereoselective hydroxylation of an unactivated C–H bond

(position 9 of coumarin **4**). The ability of microorganisms to hydroxylate chemically inaccessible centers is a powerful synthetic tool because the functionalization of unactivated C–H bonds is a true challenge to organic synthesis [28]. Traditional chemical methods generally require highly reactive oxidizing agents, which causes difficulties in the regio- and stereocontrols.

Considering the poor conversion rate of coumarin **4** and the consequent low yield of **C1** (9.0% yield), we investigated the biotransformation of **4** by *C. elegans*, intending to improve its conversion rate. The biotransformation of coumarin **4** by *C. elegans* was then carried out by using twice the amount of fungus and also by changing the incubation time to 120 and 168 h. We concluded that the biotransformation assay that used twice the amount of fungus for 72 h was the best one for the biotransformation of **4** into **C1**. Next, we repeated the biotransformation of **4** by *C. elegans* by using the optimized conditions, and we isolated **C1** with a yield of 22.0%.

The next step of our study was the identification of the chemical structure of the derivative **C2**, which was achieved through biotransformation of the coumarin **7** by *A. brasiliensis* with a yield of 35.0%. The **C2** 1H NMR spectrum analysis (Figure S7 in the Supplementary Materials) showed that the biotransformation of **7** by *A. brasiliensis* led to hydrolysis at C-10. The ester group at C-10 of the chemical structure of **7** was converted into a carboxyl acid group in **C2**. The **C2** and **7** 1H NMR spectra were almost identical. Meanwhile, no methyl hydrogen was present in the **C2** 1H NMR spectrum. The 1H NMR of coumarin **7** contained a singlet at δ 3.83 with an integral value of 3. This signal was not found in the **C2** 1H NMR spectrum. All **C2** NMR data were in accord with those previously reported for 2-(7-hydroxy-2-oxo-2*H*-chromen-4-yl) acetic acid or 7-hydroxycoumarinyl-4-acetic acid [29].

It was previously established that C-7 substitution in the coumarin nucleus increases its antioxidant and antifungal activities [30]. Within this context, Molnar and coworkers have synthesized some coumarinyl thiosemicarbazides from **C2**. The synthesized compounds displayed interesting antibacterial activity against *Bacillus subtilis* [31].

As part of our studies on the biotransformation of coumarins, we recently reported that transformations at C-7, reductions at C3–C4, and lactone-ring opening were the most frequent reactions in coumarin cores submitted to biotransformation using filamentous fungi [25]. Unlike what was expected, the biotransformations described in the present study led to hydroxylation at an unexpected position and hydrolysis reaction.

In summary, our investigations into the biotransformation of coumarin compounds by *C. elegans* and *A. brasiliensis* provided useful information about structural features that are important for the microbial transformation of this kind of compound. Coumarins that contained aromatic monohydroxylation and a bulky group at C-4 in their chemical structures were efficiently biotransformed by the fungi strains. Although both fungi strains converted the coumarins into the same derivatives, the reaction yields were distinct. We demonstrated the occurrence of a stereoselective hydroxylation at an unactivated carbon of one of the coumarins with a preference for levogyre stereoisomer production. The spectroscopic analysis allowed for the identification of a new chiral coumarin, whose relative stereochemistry was identified.

#### **3. Materials and Methods**
