Evaluation of Microbial Transformation of 10-deoxoartemisinin by UPLC-ESI-Q-TOF-MSE

10-deoxoartemisinin is a semisynthetic derivative of artemisinin that lacks a lactone carbonyl group at the 10-position, and has stronger antimalarial properties than artemisinin. However, 10-deoxoartemisinin has limited utility as a therapeutic agent because of its low solubility and bioavailability. Hydroxylated 10-deoxoartemisinins are a series of properties-improved derivatives. Via microbial transformation, which can hydroxylate 10-deoxoartemisinin at multiple sites, the biotransformation products of 10-deoxoartemisinin have been investigated in this paper. Using ultra-performance liquid chromatography-electrospray ionization-quadrupole time-of-flight mass spectrometry (UPLC-ESI-Q-TOF-MSE) combined with UNIFI software, products of microbial transformation of 10-deoxoartemisinin were rapidly and directly analyzed. The hydroxylation abilities of nine microorganisms were compared using this method. All of the microorganisms evaluated were able to hydroxylate 10-deoxoartemisinin, and a total of 35 hydroxylated products were identified. These can be grouped into dihydroxylated 10-deoxoartemisinins, monohydroxylated 10-deoxoartemisinins, hydroxylated dehydrogenated 10-deoxoartemisinins, and hydroxylated hydrogenated 10-deoxoartemisinins. Cunninghamella echinulata and Cunninghamella blakesleeana are able to hydroxylate 10-deoxoartemisinin, and their biotransformation products are investigated here for the first time. Cunninghamella elegans CICC 40250 was shown to most efficiently hydroxylate 10-deoxoartemisinin, and could serve as a model organism for microbial transformation. This method could be used to generate additional hydroxylated 10-deoxoartemisinins for further research.


Introduction
Artemisinin is a first-line antimalarial agent, which is first derived from Chinese herb by Professor Tu [1]. 10-deoxoartemisinin is a bioactive derivative of artemisinin, first prepared by the simple transformation in 1990 [2]. It lacks the C-10 lactone carbonyl group of artemisinin, but retains its antimalarial endoperoxide activity [3]. Compared with artemisinin, the derivate was eight times more active against multidrug-resistant malaria in vitro [2]. However, poor solubility and bioavailability are the barriers for further application in the clinic. Moreover, 10-deoxoartemisinin also has impressive antitumor and antiangiogenesis activities [4][5][6]. As a part of our ongoing projects [7][8][9][10], novel 10-deoxoartemisinin derivatives with improved the properties are attractive to develop for further research. Hydroxylation, as the main modification route, could not only improve chemical properties, but provide candidates of novel drug development.
In recent years, microbial transformation has become an efficient tool to generate novel active products. Microorganisms can achieve predetermined selectivity to add functional groups in an environment-friendly manner because of their abundant cytochrome P450 enzymes [11][12][13]. Several fungi have been reported to exhibit particularly strong hydroxylation capabilities, including Cunninghamella, Aspergillus, Mucor, etc. [14][15][16]. Up to now, five microbial transformation products of 10-deoxoartemisinin from four strains have been documented. These include 5β-hydroxy-10-deoxoartemisinin from Mucor ramannianus [17], 4α-hydroxy-1, 10-deoxoartemisinin and 7β-hydroxy-10-deoxoartemisinin from both Mucor ramannianus and Cunninghamella elegans [18], 15-hydroxy-10-deoxoartemisinin from Aspergillus Niger [19], and a 13-carbon rearranged product from Aspergillus ochraceus [20], none of which were reported to be bioactive. Even though, 10-deoxoartemisinin provides an excellent biotransformation substrate for the production of new potentially antimalarial active, and there remains a need for more exploited. The scheme synthesis of 10-deoxoartemisinin is shown in Figure 1. develop for further research. Hydroxylation, as the main modification route, could not only improve chemical properties, but provide candidates of novel drug development.

Identification of 10-deoxoartemisinin
The substrate 10-deoxoartemisinin was used to establish a reference fragmentation pattern.

Identification of Microbial Transformation Products
Microbial transformation products were identified using the characteristic mass spectrometric behavior of 10-deoxoartemisinin, including parent ions, internal cleavage of the ion, and retention times. Altogether, 35 transformation products were identified; primarily dihydroxylated 10-deoxoartemisinin, monohydroxylated 10-deoxoartemisinin, hydroxylated dehydrogenated 10-deoxoartemisinin, and hydroxylated hydrogenated 10-deoxoartemisinin. While all strains tested were able to produce hydroxylated 10-deoxoartemisinins, MT1 produced the most derivatives, with 18 distinct products. The 18 hydroxylated products from MT1 all had a peroxide bridge. These results demonstrate that MT1 is an ideal microorganism model to obtain hydroxylated 10-deoxoartemisinin for further research. Meanwhile, the biotransformation products from Cunninghamella echinulata and Cunninghamella blakesleeana are first to be investigated. The identified microbial transformation products are shown in Table 1, and the structures of predicted transformation products are shown in Figure 4.

Identification of Microbial Transformation Products
Microbial transformation products were identified using the characteristic mass spectrometric behavior of 10-deoxoartemisinin, including parent ions, internal cleavage of the ion, and retention times. Altogether, 35 transformation products were identified; primarily dihydroxylated 10-deoxoartemisinin, monohydroxylated 10-deoxoartemisinin, hydroxylated dehydrogenated 10-deoxoartemisinin, and hydroxylated hydrogenated 10-deoxoartemisinin. While all strains tested were able to produce hydroxylated 10-deoxoartemisinins, MT1 produced the most derivatives, with 18 distinct products. The 18 hydroxylated products from MT1 all had a peroxide bridge. These results demonstrate that MT1 is an ideal microorganism model to obtain hydroxylated 10-deoxoartemisinin for further research. Meanwhile, the biotransformation products from Cunninghamella echinulata and Cunninghamella blakesleeana are first to be investigated. The identified microbial transformation products are shown in Table 1, and the structures of predicted transformation products are shown in Figure 4.

Identification of Microbial Transformation Products
Microbial transformation products were identified using the characteristic mass spectrometric behavior of 10-deoxoartemisinin, including parent ions, internal cleavage of the ion, and retention times. Altogether, 35 transformation products were identified; primarily dihydroxylated 10-deoxoartemisinin, monohydroxylated 10-deoxoartemisinin, hydroxylated dehydrogenated 10-deoxoartemisinin, and hydroxylated hydrogenated 10-deoxoartemisinin. While all strains tested were able to produce hydroxylated 10-deoxoartemisinins, MT1 produced the most derivatives, with 18 distinct products. The 18 hydroxylated products from MT1 all had a peroxide bridge. These results demonstrate that MT1 is an ideal microorganism model to obtain hydroxylated 10-deoxoartemisinin for further research. Meanwhile, the biotransformation products from Cunninghamella echinulata and Cunninghamella blakesleeana are first to be investigated. The identified microbial transformation products are shown in Table 1, and the structures of predicted transformation products are shown in Figure 4.  . The products were identified by comparing their parent ions to 10-deoxoartemisinin, which showed a series of fragment ions resulting from H2O and CO loss. The dihydroxylated products were more likely to lose two water molecules compared with 10-deoxoartemisinin, and the fragment ions at m/z 265 and 247 were generated by successive water loss from m/z 283. The fragment ion at m/z 237 was generated by CO loss from m/z 265 or HCOOH loss from m/z 283. The fragment ion at m/z 219 was generated by water loss from m/z 237. Mass spectra and the fragmentation diagram for dihydroxylated 10-deoxoartemisinin products are shown in Figure 5.  . The products were identified by comparing their parent ions to 10-deoxoartemisinin, which showed a series of fragment ions resulting from H 2 O and CO loss. The dihydroxylated products were more likely to lose two water molecules compared with 10-deoxoartemisinin, and the fragment ions at m/z 265 and 247 were generated by successive water loss from m/z 283. The fragment ion at m/z 237 was generated by CO loss from m/z 265 or HCOOH loss from m/z 283. The fragment ion at m/z 219 was generated by water loss from m/z 237. Mass spectra and the fragmentation diagram for dihydroxylated 10-deoxoartemisinin products are shown in Figure 5.  . The products were identified by comparing their parent ions to 10-deoxoartemisinin, which showed a series of fragment ions resulting from H2O and CO loss. The dihydroxylated products were more likely to lose two water molecules compared with 10-deoxoartemisinin, and the fragment ions at m/z 265 and 247 were generated by successive water loss from m/z 283. The fragment ion at m/z 237 was generated by CO loss from m/z 265 or HCOOH loss from m/z 283. The fragment ion at m/z 219 was generated by water loss from m/z 237. Mass spectra and the fragmentation diagram for dihydroxylated 10-deoxoartemisinin products are shown in Figure 5. Monohydroxylated 10-deoxoartemisinin transformation products were detected between 3.14 and 6.91 min. The products were identified by comparing their parent ions to 10-deoxoartemisinin, with a 16 Da mass shift attributed to an oxygen atom. The molecular ion at m/z 307 ([M + Na] + ) was observed. The fragment ion at m/z 267 resulted from water loss from m/z 285. The fragment ions at m/z 249 and m/z 231 were generated by water loss from m/z 267 and m/z 249, respectively. The fragment ion at m/z 221 resulted from CO loss from m/z 249. Mass spectra and the fragmentation diagram for monohydroxylated 10-deoxoartemisinin products are shown in Figure 6. Monohydroxylated 10-deoxoartemisinin transformation products were detected between 3.14 and 6.91 min. The products were identified by comparing their parent ions to 10-deoxoartemisinin, with a 16 Da mass shift attributed to an oxygen atom. The molecular ion at m/z 307 ([M + Na] + ) was observed. The fragment ion at m/z 267 resulted from water loss from m/z 285. The fragment ions at m/z 249 and m/z 231 were generated by water loss from m/z 267 and m/z 249, respectively. The fragment ion at m/z 221 resulted from CO loss from m/z 249. Mass spectra and the fragmentation diagram for monohydroxylated 10-deoxoartemisinin products are shown in Figure 6.     Hydroxylated hydrogenated 10-deoxoartemisinin products were detected at about 4.50 min. The fragment ions were similar to monohydroxylation products of 10-deoxoartemisinin. The parent ion was detected at m/z 309 ([M + Na] + ). Fragment ions at m/z 249 and 231 were generated by successive water loss from m/z 267. The fragment ion at m/z 221 was generated by CO loss from m/z 249 or HCOOH loss from m/z 267. The fragment ion at m/z 203 resulted from water loss from m/z 221. The mass spectra and the fragmentation diagram for hydroxylated hydrogenated 10-deoxoartemisinin products are shown in Figure 8.

Discussion
Hydroxylation is the main route for molecular modification and also a primary metabolic reaction in vivo involving artemisinin and its derivatives [20,26]. Hydroxylated 10-deoxoartemisinins are a series of products that have a peroxide bridge and may retain antimalarial activity, with better physiological solubility compared with unmodified 10-deoxoartemisinin. Meanwhile, these hydroxylated compounds provide the new reactive site to link another functional group for developing a novel active drug.
Microbial transformation is an increasingly popular method used to modify small molecules, particularly adept at generating hydroxylated derivatives, because many enzymes catalyze hydroxylation reactions, and a variety of microorganisms can be used in this method. The microbial transformation was used in this study to find out novel hydroxylated 10-deoxoartemisinins. The hydroxylation abilities of nine microorganisms from the genera Cunninghamella and Mucor was evaluated, because previous studies have shown that species in these two genera are rich in hydroxylases [27]. Additionally, C. elegans, C. echinulata, C.blakesleeana, and M. circinelloides have been successfully employed to hydroxylate other substrates [28][29][30]. The hydroxylation ability of these microorganisms was estimated by identifying and comparing their hydroxylation products. A total of 35 hydroxylated products were identified from the nine microorganisms.
The rapid and direct analysis method developed in this paper is significant for efficiently screening numerous of Microorganisms transformation. The results could be the guidance for further isolation research to obtain more derivates. This work is foundational for producing novel, potentially bioactive, hydroxylated 10-deoxoartemisinins.

Discussion
Hydroxylation is the main route for molecular modification and also a primary metabolic reaction in vivo involving artemisinin and its derivatives [20,26]. Hydroxylated 10-deoxoartemisinins are a series of products that have a peroxide bridge and may retain antimalarial activity, with better physiological solubility compared with unmodified 10-deoxoartemisinin. Meanwhile, these hydroxylated compounds provide the new reactive site to link another functional group for developing a novel active drug.
Microbial transformation is an increasingly popular method used to modify small molecules, particularly adept at generating hydroxylated derivatives, because many enzymes catalyze hydroxylation reactions, and a variety of microorganisms can be used in this method. The microbial transformation was used in this study to find out novel hydroxylated 10-deoxoartemisinins. The hydroxylation abilities of nine microorganisms from the genera Cunninghamella and Mucor was evaluated, because previous studies have shown that species in these two genera are rich in hydroxylases [27]. Additionally, C. elegans, C. echinulata, C.blakesleeana, and M. circinelloides have been successfully employed to hydroxylate other substrates [28][29][30]. The hydroxylation ability of these microorganisms was estimated by identifying and comparing their hydroxylation products. A total of 35 hydroxylated products were identified from the nine microorganisms.
The rapid and direct analysis method developed in this paper is significant for efficiently screening numerous of Microorganisms transformation. The results could be the guidance for further isolation research to obtain more derivates. This work is foundational for producing novel, potentially bioactive, hydroxylated 10-deoxoartemisinins.

Materials and Reagents
10-deoxoartemisinin was synthesized from artemisinin, as previously described [2]. M. circinelloides CGMCC 3.3421, M. circinelloides CGMCC 3.49, C. blakesleeana CGMCC 3.910 and 3.5802, C. echinulata CGMCC 3.5771, C. echinulata 3.4879, and C. elegans CGMCC 3.4832 were obtained from the China General Microbiological Culture Collection Center (Beijing, China). C. elegans CICC 40250 was obtained from the China Center of Industrial Culture Collection (Beijing, China). C. elegans ATCC 9245 was obtained from the American Type Culture Collection (Virginia, United States). Acetonitrile and methanol were purchased from Fisher (Geel, Belgium), and formic acid, EtOAc (ethyl acetate), and other chemicals used in this research were purchased from Beijing Chemical Works (Beijing, China).
Two-stage fermentation was employed in this study [33,34]. The substrate was dissolved in MeOH at a concentration of 25 mg/mL, and 2 mL of the solution was added into each flask after the second fermentation stage, producing a 10-deoxoartemisinin concentration of 0.5 mg/mL. Cultures were incubated at 28 • C and shaken at 180 rpm for 14 days. Mycelia and broth were then separated by filtration. The filtrate was extracted with EtOAc (1:1, v/v) three times, and the extract was evaporated under vacuum to produce a brown residue, which was dissolved in MeOH for LC-MS analysis.

LC-MS E Conditions and Data Processing
Mass spectrometry conditions were established in previous studies [32]. The UPLC-ESI-Q-TOF-MS E system consisted of a Waters ACQUITY I-class UPLC and Xevo G2-XS Q-TOF Mass Spectrometer (Waters, Manchester, UK). Chromatography was conducted using an Acquity UPLC BEH C 18 (2.1 mm × 100 mm, 1.7 µm (Waters, Manchester, UK)). The mobile phase consisted of solvent A (H 2 O containing 0.1% formic acid, v/v) and solvent B (acetonitrile containing 0.1% formic acid, v/v). The gradient program included three segments: 5%-100% B from 0 to 15 min; 100%-5% B from 15 to 17 min; followed by a 3 min post-run for column equilibration. The flow rate was 0.4 mL/min, and the temperature was 25 • C throughout the analysis.
The MS was operated in positive ionization mode across a scan range of 50 to 1200 m/z. The low collision energy was set at 6 eV, and the high collision energy was ramped from 12 to 25 eV. MS E analysis was conducted using multiple reactions monitoring positive-ion electrospray ionization.
All data processing was performed using UNIFI 1.9 (Waters, Manchester, UK). Components were identified using the following 3D peak detection features: Low-energy limits of 150 and high-energy limits of 20, isotope clustering, and high-to-low energy association within a 0.5 fraction of the chromatographic and drift peak width, with a mass accuracy of ± 2 mDa. The maximum number of allowed fragment ions per match was set to 10.

Conclusions
This research establishes microbial transformation as a method for obtaining various hydroxylated products. Transformation products from nine microorganisms were identified using UPLC-ESI-Q-TOF-MS E , and the hydroxylation ability of the nine microorganisms was evaluated. Transformation products of 10-deoxoartemisinin were categorized as hydroxylated 10-deoxoartemisinins, hydroxylated dehydrogenated 10-deoxoartemisinins, and hydroxylated hydrogenated 10-deoxoartemisinins. MT1 (C. elegans CICC 40250) exhibited the best hydroxylation ability among the selected microorganisms, and could be used as a transformation model to prepare hydroxylated 10-deoxoartemisinins for further research. Additionally, this study was the first to investigate the transformation products of 10-deoxoartemisinin from C. echinulata and C. blakesleeana.
Author Contributions: L.Y. and Y.M. conceived the idea for this project, designed, and conducted experiments. Y.B. collected and analyzed the data and wrote the manuscript with contributions from other authors. D.Z. participated in research design and assisted in the data analysis. Y.Z. and X.C. participated in performing the microbial transformation experiments. P.S. contributed to the writing.