Polyoxygenated Terpenoids and Polyketides from the Roots of Flueggea virosa and Their Inhibitory Effect against SARS-CoV-2-Induced Inflammation

Six new polyoxygenated terpenoids, podovirosanes A–F (1–6), and two known polyketides (7 and 8) were isolated from the roots of F. virosa. Their structures, along with absolute configurations, were deduced using spectroscopic analysis as well as computational calculations, including TDDFT calculation of ECD spectra and GIAO NMR calculations combined with DP4+ probability analysis. Compounds 2, 3, 5, and 8 were found to reduce the phosphorylation levels of NF-κB p65 in SARS-CoV-2 pseudovirus-stimulated PMA-differentiated THP-1 cells.


Introduction
Coronavirus-induced disease COVID-19, a highly infectious disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has resulted in a major public health crisis since 2019 [1]. Although most of the infectious patients are mildly symptomatic or asymptomatic, the acute infectious case might result in an aggressive lung inflammation initiated by increased production of inflammatory cytokines from innate immune cells [2,3].
Flueggea plants, formerly classified asthe genus Securinega, were known to have 16 different species that were widely spread in Asia, Africa, and South America [4,5]. Flueggea virosa, belonging to the Euphorbiaceae family, were traditionally used as folk medicines in Asia and China and have been used for treating diverse ailments [4]. Although the anti-virus constituents have been reported from the title plants [6][7][8], the constituents for anti-inflammatory properties remain largely uninvestigated. Our group has developed an assay for screening the anti-inflammatory activity of compounds against SARS-CoV-2-induced inflammation, which was performed in this study. As part of our ongoing search for bioactive constituents from natural sources [9][10][11][12], six new ent-podocarpane-related terpenoids and two known polyketides ( Figure 1) were isolated from the roots of F. virosa. Their structures were characterized using spectroscopic methods as well as quantum chemical calculations, including ECD and gaugeindependent atomic orbital (GIAO) NMR calculations, of which the latter was combined with DP4+ analysis. Herein, the isolation, the structural elucidation, and the ability of isolates against SARS-CoV-2-induced inflammatory response were described.
A 2-methyl-4,5-disubstituted anisole fragment was evidenced by HMBC correlations, which were observed from H [3][4][5][6][7][8][9][10][11][12][13][14][15]. In combination with the two spin systems from H 2 -1 to H-3 and H-5 to H-7 established by analysis of COSY correlations, the HMBC correlations from H-14 to C-7, from H-11 to C-10, from H 3 -18 to C-3, C-4, C-5, and C-19, and from H 2 -1 to C-20, C-5, and C-10 allowed the establishment of a 13-methylpodocarpane framework in 1 (Figures 2 and S7). Moreover, the HMBC correlations from oxymethylene protons (H 2 -19) to the carbonyl carbon (C-20) allowed the connectivity of the ester linkage between C-19 and C-20 ( Figure 2). The relative configuration of the ester bridge was assigned as α-oriented being opposite to H-5 according to the NOE correlation of H-5/H 3 -18 (Figures 3 and S8). The small 3 J H,H value of H-3 and its NOE correlations with both H 3 -18 and one proton of H 2 -19 (δ H 4.04), of which the latter displayed a 4 J W-coupling (1.7 Hz) with H-5, suggested that H-3 is located on equatorial position and assigned as αproton. The relative configuration of C-3 was further confirmed by GIAO NMR calculations of the two possible candidates with 3β-OH and 3α-OH substituents ( Figure S48), followed by DP4+ probability analysis [13]. The result revealed that the all-data probability of the 3β-OH candidate was found to be 100% (Tables 3 and S1) and thus suggested it to be the correct structure. Finally, a comparison of experimental and calculated ECD data established the absolute configuration of 1 as 3S, 4R, 5R, 10S ( Figure 4).  Podovirosane B (2) has a molecular of C 19 H 26 O 4 as determined from the (-)-HRAPCIMS and NMR data (Figures S9-S12). The 1 H and 13 C NMR data of 2 (Tables 1 and 2) showed remarkable similarity to a known analog, 7α,20-epoxy-3α-hydroxy-12-methoxy-13-methyl-entpodocarp- 8,11,13-triene (2a), isolated from the same plant [6], and the obvious differences between these two compounds were the chemical shifts of C-6, of which the aliphatic methylene (δ C 29.9; δ H 2.13, 1.58) in 2a [6] was replaced by a hydroxy-containing methine (δ C 68.9; δ H 4.21) in 2. Detailed analysis of HMBC, HSQC, and COSY spectra (Figures S13-S15) gave evidence to support the structure of 2, as shown in Figure 2. The NOE correlations of H-5/H-3 and H-5/H-1a suggested that ring A adopted a chair conformation and implied that these protons were cofacial (β-orientation) ( Figure S16). A four-bond W-coupling observed between H-20b (δ H 2.72) and H-5 (δ H 0.91) suggested the α orientation of this ether functionality between C-7 and C-20. Although H-5 was found to have an NOE correlation with H-6, a comparison of 6α-OH and 6β-OH models disclosed that in both candidates, the proton distances between H-5 and H-6 were close enough to give NOE enhancements ( Figure S49). Thus, the NOE correlations were unable to correlate the relative configuration of C-6. Using the same GIAO NMR calculation method as described above and applying the DP4+ probability on possible candidates of 2 ( Figure S48), the (6β-OH)-2 was suggested with a high probability of 100% (Tables 3 and S4). 5 according to the NOE correlation of H-5/H3-18 (Figures 3 and S8). The small 3 JH,H value of H-3 and its NOE correlations with both H3-18 and one proton of H2-19 (δH 4.04), of which the latter displayed a 4 J W-coupling (1.7 Hz) with H-5, suggested that H-3 is located on equatorial position and assigned as α-proton. The relative configuration of C-3 was further confirmed by GIAO NMR calculations of the two possible candidates with 3β-OH and 3α-OH substituents ( Figure S48), followed by DP4+ probability analysis [13]. The result revealed that the all-data probability of the 3β-OH candidate was found to be 100% (Tables 3 and S1) and thus suggested it to be the correct structure. Finally, a comparison of experimental and calculated ECD data established the absolute configuration of 1 as 3S,4R,5R,10S ( Figure 4).  which the latter displayed a 4 J W-coupling (1.7 Hz) with H-5, suggested that H-3 is located on equatorial position and assigned as α-proton. The relative configuration of C-3 was further confirmed by GIAO NMR calculations of the two possible candidates with 3β-OH and 3α-OH substituents ( Figure S48), followed by DP4+ probability analysis [13]. The result revealed that the all-data probability of the 3β-OH candidate was found to be 100% (Tables 3 and S1) and thus suggested it to be the correct structure. Finally, a comparison of experimental and calculated ECD data established the absolute configuration of 1 as 3S,4R,5R,10S ( Figure 4).      Podovirosane B (2) has a molecular of C19H26O4 as determined from the (-)-HRAP-CIMS and NMR data ( Figures S9-S12). The 1 H and 13 C NMR data of 2 (Tables 1 and 2) showed remarkable similarity to a known analog, 7α,20-epoxy-3α-hydroxy-12-methoxy- 13-methyl-ent-podocarp-8,11,13-triene (2a), isolated from the same plant [6], and the obvious differences between these two compounds were the chemical shifts of C-6, of which the aliphatic methylene (δC 29.9; δH 2.13, 1.58) in 2a [6] was replaced by a hydroxy-containing methine (δC 68.9; δH 4.21) in 2. Detailed analysis of HMBC, HSQC, and COSY spectra (Figures S13-S15) gave evidence to support the structure of 2, as shown in Figure 2. The NOE correlations of H-5/H-3 and H-5/H-1a suggested that ring A adopted a chair conformation and implied that these protons were cofacial (β-orientation) ( Figure S16). A four-bond W-coupling observed between H-20b (δH 2.72) and H-5 (δH 0.91) suggested the α orientation of this ether functionality between C-7 and C-20. Although H-5 was found to have an NOE correlation with H-6, a comparison of 6α-OH and 6β-OH models disclosed that in both candidates, the proton distances between H-5 and H-6 were close enough to give NOE enhancements ( Figure S49). Thus, the NOE correlations were unable to correlate the relative configuration of C-6. Using the same GIAO NMR calculation method as described above and applying the DP4+ probability on possible candidates of 2 ( Figure S48), the (6β-OH)-2 was suggested with a high probability of 100% (Tables 3 and  S4).
Podovirosane C (3) was found to have a negative [M -H] ion peak in HRESIMS, which, in combination with the NMR data (Tables 1 and 2), indicated the same molecular formula as that of 2. A comparison of the NMR data of 3 with those of 2 disclosed that they are epimers, with the difference being the NMR chemical shifts of C-3. Similarly, the chair-form ring A was evidenced by the NOE correlations of H-5/H-1a and H3-19/H-2a. Moreover, the proton signal of H-3 displayed as a broad singlet revealed the equatorial nature of this proton, suggesting the 3-OH to be β-oriented. The absolute configurations of 2 (3R,5S,6S,7S,10S) and 3 (3S,5S,6S,7S,10S) were also determined by the TDDFT-ECD calculations ( Figure 4).
The ( (Tables 1 and 2) were closely related to those of 9(10→20)-abeo-ent-podocarpanes [7], and differences should be substitution patterns in ring B. The HMBC correlations from H-7 to C-9, C-14, and C-10 as well as correlations from H 2 -20 to C-8, C-9, and C-10, indicated that an ether linkage should be assigned between C-7 and C-10 (Figures 2 and S37). The COSY correlations of H-6 with H-6 and H-7 and of H 2 -2 with H 2 -1 and H-3 suggested the C-5/C-6/C-7 and the C-1/C-2/C-3 spin systems ( Figure S38). Further HMBC correlations from H 3 -18 (or H 3 -19) to C-3, C-4, and C-5 connected the two spin systems and thus established the planar structure of 5. The NOE correlation of H 3 -18/H-1 suggested ring A should present as boat conformation [14]. The NOE correlation of H-6/H-3 arbitrarily assigned these two protons as αoriented, whereas the correlations of H-5/H-20b and H-5/H 3 -18 suggested the β-orientation of H-5 and the α-oriented C-7-C-10 ether linkage (Figures 3 and S39). The absolute configuration of 5, as shown in Figure 1, was further corroborated by comparing the calculated and experimental ECD data (Figure 4).
The (+)-HRESIMS of podovirosane F (6) displayed a sodiated molecular ion [M + Na] + , revealing a molecular formula of C 20 H 30 O 4 ( Figure S40). The 1 H and 13 C NMR data (Tables 1 and 2) of 6 suggested it to be an analog of the known compound, 3β,10αdihydroxy-12-methoxy-13-methyl-9(10 → 20)-abeo-ent-podocarpa- 6,8,11,13-tetraene (6a), which is reported from the leaves of the same plant [15]. Analysis of 13 C NMR data of 6 disclosed that it has an additional hydroxyl group at ring A as compared to 6a. A 2,3dihydroxy substituent could be readily assigned by the COSY correlations between H-2 and H-3 and between H 2 -1 and H-2 ( Figure 2). Further HMBC and COSY correlations (Figures 2, S45 and S46) confirmed such a planar structure of 6. The NOE correlations of H-2/H-3, H-3/H 3 -19, H-3/H 3 -18, and H-5/H 3 -18 implied that H-5 and H-3 should be located on axial and equatorial positions, respectively; however, the orientation of H-2 remains unclear as either equatorial or axial H-2 is likely to show NOE correlation with equatorial H-3 (Figures 3 and S47). In turn, analysis of the coupling constant between H-2 and H 2 -1 (Table 1) revealed that H-2 should be located in an equatorial position, suggesting the α-orientation of the attached OH group at C-2. On the other hand, the relative configuration of C-10 was also unable to be determined due to the lack of crucial NOE correlations. In order to clarify the configuration of C-10 and also to confirm the relative configuration of C-2, the GIAO NMR calculation and DP4+ probability analysis were performed for the four possible candidates, (2α,10α-OH)-6, (2β,10α-OH)-6, (2α,10β-OH)-6, and (2β,10β-OH)-6 ( Figure S48). The result suggested that (2α,10α-OH)-6 with 2α,10α-dihydroxy substituents has a high probability of 100% (Tables 3 and S7). With a similar procedure of ECD analysis upon 5, the absolute configuration of 6 was determined as 2R,3R,5R,10S ( Figure 4).
Since SARS-CoV-2 Spike protein can stimulate inflammation [19][20][21][22], we established a platform to screen the activity of SARS-CoV-2 induced-inflammation for the isolated compounds. A preliminary evaluation of the cytotoxicity of the isolates toward the human monocytic cell line (THP-1) was performed using an MTS assay. The results revealed that compounds 1-8 were not found to exhibit toxicity toward THP-1 cells at a concentration of 25 µM (Figure 5A,C). The anti-inflammatory activity was performed by evaluating the expression of Phospho-NF-kappaB p65 (p-NF-κB p65) in SARS-CoV-2 pseudovirusstimulated phorbol 12-myristate 13-acetate (PMA)-differentiated THP-1 cells. The result of western blot analysis revealed that compounds 2, 3, 5, and 8, at a concentration of 25 µM, reduced the phosphorylated status of NF-κB in SARS-CoV-2 pseudovirus-stimulated THP-1 macrophages ( Figure 5A,B).

Plant Material
The roots of F. virosa were collected from Pingtung County, Taiwan, in September 2011 and authenticated by Prof. C. S. Kuoh of the Department of Life Sciences, National Cheng Kung University, Tainan, Taiwan. A voucher herbarium specimen, FV-Chao001, has been deposited in the School of Pharmacy, China Medical University, Taichung, Taiwan.

Extraction and Isolation
The dried roots of F. virosa (13.0 kg) were minced and soaked in MeOH (4 × 20 L) at room temperature. The combined extract was concentrated to give an oily brown residue and subsequently partitioned between CHCl 3 and H 2 O. The CHCl 3 layer was washed with 3% tartaric acid to afford an alkaloid-free extract (93 g) which was fractionated using 311+G(d,p) level of theory in MeOH (compounds 1-3, and 5) and ACN (compound 4), while compound 6 was performed at SMD/M062x/6-311+G(d,p) level of theory in MeOH. The spectra were averaged according to the Boltzmann distribution using energies calculated at the same level of theory.

NMR Calculations and DP4+ Analysis
The conformers for possible candidates of 1, 2, and 6 were obtained from calculations at B3LYP/6-31G(d) level of theory using Gaussian 16 software. The GIAO NMR calculation, using PCM/mPW1PW91/6-31G+(d,p) level in CHCl 3 , was applied on the resulting conformers with relative energy less than 2 kcal/mol from the global minimum. The Boltzmann-weighted NMR data of possible candidates and the experimental data were subjected to DP4+ probability analysis with the aid of the Excel sheet provided by Grimblat et al. (Tables S1-S11) [13].

Generation of SARS-CoV-2 Spike Pseudotyped Viruses
SARS-CoV-2 Spike pseudotyped viral particles have recently been reported using HIVbased lentiviral particles [24]. The pseudotyped lentiviral particles with SARS-CoV-2 Spike were then generated with some modifications. In brief, 293T cells were co-transfected with three plasmids, including pcDNA3.1-2019-nCoV-S∆18 (from RNAi Core, Academia Sinica, Taipei, Taiwan), pCMV d8.91 and pAS3W Fluc, to produce pseudotyped viral particles. The supernatant containing pseudotyped viral particles was collected at 24 h, 36 h, and 48 h post-transfection, following centrifugation at 3000 rpm for 10 min to remove cellular debris. The supernatant was then passed through a 0.45 µm filter and concentrated. The viral particle number was determined by real-time RT-PCR to quantify the RNA copies of FLuc reporter gene.

MTS Assay
MTS assay was performed as previously described [6]. In brief, the THP-1 cells were seeded on 96-well plate at a density of 1 × 10 4 cells/well and treated with 5 ng/mL of PMA for 48 h. The differentiated THP-1 cells were treated with the tested compounds at the concentration of 25 µM for additional 18 h. The cells were then incubated with MTS solution for 4 h, and the absorbance at 490 nm was measured using a microplate spectrophotometer (Molecular Devices, San Jose, CA, USA).

Compound Screening to Inhibit SARS-CoV-2 Induced-Inflammation
In brief, the human monocytic cell line THP-1 was treated with 5 ng/mL of phorbol 12-myristate 13-acetate (PMA; sigma) to differentiate into macrophages in RPMI-1640 medium with 10% FBS for 24 h. After changing fresh medium, the cells were incubated for further 48 h [25]. The differentiated THP-1 cells were re-seeded into 12 well at a density of 5 × 10 5 cells per mL, and the differentiated macrophages were infected with SARS-CoV-2 pseudotyped viruses (purchased from RNAi Core, Academia Sinica, Taipei, Taiwan) for 2 h following treated with each tested compound (25 µM) for 18 h. The cell lysates were collected, and the expression level and phosphorylated status of NF-κB were detected by Western analysis.

Conclusions
Six new polyoxygenated terpenoids, podovirosanes A-F (1-6), and two known polyketides (7 and 8) were isolated from the roots of F. virosa. Among them, the polyketides 7 and 8 likely originated from root symbiotic fungi as these two compounds and related analogs were reported from fungi [16][17][18]26]. The present work represents the first report for the non-alkaloid constituents of the title plant that exhibited potent anti-inflammatory activity against SARS-CoV-2-induced inflammation.