Oxy-Polybrominated Diphenyl Ethers from the Indonesian Marine Sponge, Lamellodysidea herbacea: X-ray, SAR, and Computational Studies

Polybrominated diphenyl ether (PBDE) compounds, derived from marine organisms, originate from symbiosis between marine sponges and cyanobacteria or bacteria. PBDEs have broad biological spectra; therefore, we analyzed structure and activity relationships of PBDEs to determine their potential as anticancer or antibacterial lead structures, through reactions and computational studies. Six known PBDEs (1–6) were isolated from the sponge, Lamellodysdiea herbacea; 13C NMR data for compound 6 are reported for the first time and their assignments are confirmed by their theoretical 13C NMR chemical shifts (RMSE < 4.0 ppm). Methylation and acetylation of 1 (2, 3, 4, 5-tetrabromo-6-(3′, 5′-dibromo-2′-hydroxyphenoxy) phenol) at the phenol functional group gave seven molecules (7–13), of which 10, 12, and 13 were new. New crystal structures for 8 and 9 are also reported. Debromination carried out on 1 produced nine compounds (1, 2, 14, 16–18, 20, 23, and 26) of which 18 was new. Debromination product 16 showed a significant IC50 8.65 ± 1.11; 8.11 ± 1.43 µM against human embryonic kidney (HEK293T) cells. Compounds 1 and 16 exhibited antibacterial activity against Gram-positive Staphylococcus aureus and Gram-negative Klebsiella pneumoniae with MID 0.078 µg/disk. The number of four bromine atoms and two phenol functional groups are important for antibacterial activity (S. aureus and K. pneumoniae) and cytotoxicity (HEK293T). The result was supported by analysis of frontier molecular orbitals (FMOs). We also propose possible products of acetylation and debromination using analysis of FMOs and electrostatic charges and we confirm the experimental result.


Introduction
Natural marine polybrominated diphenyl ethers (PBDEs) are promising medicinal lead structures for anticancer [1,2] and antibacterial [3] drugs. The latter can be rationalized by the structural resemblance to the man-made antibacterial, triclosan (2-(2 ,4 -molecules (1, 2, 14, 16-18, 20, 23, 26) ( Figure 1). We propose the calculated 13 C NMR data of 30 PBDEs and evaluate their RMSEs for which we have experimental data. In regard to the exploration reaction, we also found that acetylation occurred in the presence of Ac 2 O and sonication without a catalyst and without solvent at room temperature for 2-3 h. Two compounds (8 and 9) have new crystal structures, whereas compounds 10, 12, and 13 are entirely new structures. Compound 18 is a new derivative obtained via debromination. One unpublished 13 C NMR of 6 is also reported and secured its assignment by its calculated 13 C NMR chemical shifts. This article also describes structure-activity relationships of marine polybrominated natural products against human embryonic kidney (HEK293T) cells and Gram-positive Staphylococcus aureus, as well as the Gram-negative bacterium Klebsiella pneumoniae. Computational studies further support their activities and account for acetylation and debromination.
Molecules 2021, 26, x FOR PEER REVIEW 3 of 22 interesting debromination in the presence of a bromine scavenger to give nine additional molecules (1, 2, 14, 16-18, 20, 23, 26) ( Figure 1). We propose the calculated 13 C NMR data of 30 PBDEs and evaluate their RMSEs for which we have experimental data. In regard to the exploration reaction, we also found that acetylation occurred in the presence of Ac2O and sonication without a catalyst and without solvent at room temperature for 2-3 h. Two compounds (8 and 9) have new crystal structures, whereas compounds 10, 12, and 13 are entirely new structures. Compound 18 is a new derivative obtained via debromination.

Results and Discussion
Six known O-PBDEs (1-6) have been isolated from the Indonesian marine sponge, L. herbacea, collected at Ujung Kulon. MS and 1 H NMR spectra of 1−6 were in good agreement with previously reported data [14,15,[18][19][20]. Compound 6 was first isolated by Fu et al., in 1995 [20]. To the best of our knowledge, 13 C NMR data for compound 6 have not been reported previously. This work also gave us an opportunity to assign the 13 C NMR for compound 6 ( Table 1). Complete 13 C NMR data of 6 (HREIMS m/z 605.6311 [M] + , C 13 H 7 O 3 79 Br 5 , ∆ −0.0001 mmu) were derived with the aid of DEPT, HMQC, and HMBC experiments and analogies from the literature. The HMBC experiment showed correlations with H2/C1,3,4,6; H4 /C2 , 3 , 5 , 6 ; H6 /C1 ,2 , 3 , 5 and -OMe/C2 leading to the order of the 13 C data of 6 (Table 1). Moreover, the assignment was secured by the protocol of Hehre et al. [6] to give RMSE < 4.0 ppm (RMSE obtained was 3.2 ppm; RMSE for aromatic compounds were around 3.2 [6]). Intriguing with the RMSE of the isolated natural products, we conducted a set of experiments in order to obtain theoretical 13 C NMR chemical shifts of 1-5 with six steps: (1) conformational search using the MMFF molecular mechanics model; (2) calculation of equilibrium geometries using the HF/3-21G model; (3) calculation of energies using the ωB97X-D/6-31G* density functional model; (4) calculation of equilibrium geometries using the ωB97X-D/6-31G* density functional model; (5) calculation of energies of using the ωB97X-V/6-311+G(2df,2p)(6-311G*) density functional model; and (6) calculation of 13 C NMR chemical shifts using the ωB97X-D/6-31G* density functional model, correction of 13 C NMR chemical shifts based on the empirical parameters, and correction 13 C NMR chemical shifts based on the Boltzmann weight obtained in step 5. We found that compounds 2-6 gave RMSEs <4.0 ppm, except for 1 (RMSE >4.0 ppm, for 1a 4.3 ppm [18], for 1b 4.4 ppm [21] in DMSO-d 6 ). These values suggested a possible error in the 13 C chemical assignments. The error could be in an assignment of 13 C chemical shifts in ring A (lack of H atoms). Based on the calculated 13 C chemical shifts, we propose an RMSE for 1 as 3.5 ppm as in 1c (Me 2 CO-d 6 ). The assignment of 13 C chemical shifts for 1 and 6 can be seen in Figure 2. In order to see the effect of phenol in our assays, seven-modified phenolic compounds (7)(8)(9)(10)(11)(12)(13) were prepared by methylation (7-9) and acetylation (10)(11)(12)(13). Two new crystal structures (8 and 9) were obtained from a methylation reaction using TMSCHN 2 with three new synthetic PBDE derivatives (10, 12, 13) from an acetylation reaction using green chemistry.
Compound 9 (C14H9Br5O3) was identified as 2,3,5-tribromo-6-(3′,5′-dibromo-2′-methoxyphenoxy) anisole after X-ray analysis. The torsion angle of 9 is φ1 = −17.1° (C6′-C1′-O-C6); φ2 = −69.0° (C1′-O-C6-C1), whereas the rest are 117.3 (6)° and 77.5°. Compound 9 was crystallized as orthorhombic in the space group Pca21 using CHCl3−Me2CO−MeCN (1:1:1). Crystal structures of 1, 8 and 9 are depicted in Figures 3 and 4, while theoretical 13 C NMR chemical shifts of 7-9 are shown in Figure 5. The RMSE of 7 showed 3.8 ppm after comparing with the literature data [19].  Figure 5. The RMSE of 7 showed 3.8 ppm after comparing with the literature data [19].  with displacement ellipsoids drawn at the 50% probability level. The crystal structu of 1 is presented to compare the material for structure modification with two phenolic groups and debromination.    We initially screened conditions for acetylation and found that 1 reacts with Ac2O under sonication for 1 h without DMAP and Et3N at room temperature. Conversion of 1 to 11 was estimated at >95% using TLC. A larger amount of 1 was used under the above conditions to give acetylation products (10-13) with slightly longer time (2-3 h). To date, acetylation with acyl halides or acid anhydrides has been reported using solvents and catalysts at room temperature or higher (>40 °C). Acetylation methods have been reported without catalysts at high solvent temperatures (60-70 °C) or with a variety of catalysts from room temperature to 110 °C [22]. With regard to PBDE molecules, only 11 has been obtained from acetylation using Ac2O with pyridine at room temperature for 24 h [23]. Therefore, our method is new and is consistent with green chemistry concepts.
Compound 10 is a partial acetylation product (Rf 0.23 Hex/  We initially screened conditions for acetylation and found that 1 reacts with Ac 2 O under sonication for 1 h without DMAP and Et 3 N at room temperature. Conversion of 1 to 11 was estimated at >95% using TLC. A larger amount of 1 was used under the above conditions to give acetylation products (10-13) with slightly longer time (2-3 h). To date, acetylation with acyl halides or acid anhydrides has been reported using solvents and catalysts at room temperature or higher (>40 • C). Acetylation methods have been reported without catalysts at high solvent temperatures (60-70 • C) or with a variety of catalysts from room temperature to 110 • C [22]. With regard to PBDE molecules, only 11 has been obtained from acetylation using Ac 2 O with pyridine at room temperature for 24 h [23]. Therefore, our method is new and is consistent with green chemistry concepts.
Compound 10 is a partial acetylation product (R f 0.23 Hex/  6 . Because ring A is fully substituted, no HMBC correlation can be observed. The -OAc group can be placed on ring A by assuming that ring A is more reactive than ring B, on the basis of molecular orbital analysis of 1 with DFT dispersion correction level of theory and a triple-ζ-basis set ( Figure 6a). Moreover, analysis of the electrostatic charge of 1 also justified that the -OH group in ring A is more electronegative (−0.560) than that of ring B (−0.478) ( Figure 6a). Therefore, the first acetylation is more likely to proceed in ring A. In addition, comparison of 10 with 3 and 7 containing an anisole group on ring A showed that the 1 H chemical shift is more upfield than ring B. Compound 10 was revealed as 2,3,4,5-tetrabromo-6-(3 ,5 -dibromo-2 -hydroxyphenoxy) phenyl acetate. The same situation as 10 was also observed for 12. The -OAc group could be attached to ring A. Highest occupied molecular orbital (HOMO) analysis 2 showed that ring A is more reactive than ring B which is justified by electrostatic charge analysis of the -OH group in ring A (−0.565) and in ring B (−0.472) (Figure 6a). Compound 12 was confirmed as 2,3,5-tribromo-6-(3 , 5 -dibromo-2 -hydroxyphenoxy) phenyl acetate.
Next, we examined the effect of lower bromine atoms on two phenol rings by employing debromination. Compound 1, isolated in gram quantities, and possessing six bromine atoms on two phenol rings, was subjected to refluxing using HBr and Na 2 SO 3 as scavengers of bromine in the presence of acetic acid. Compounds (1, 2, 14, 16-18, 20, 23, 26) were isolated and confirmed from the reaction. MS and 1 H NMR spectra of compounds (1, 2, 16, 17, 20, 23, 26) were in a good agreement with previous reports [14,15].
Debromination of 1, including its product, can be supported by DFT ωB97X-D/6-311 + G(2d,p) calculations. Combination analysis of frontier molecular orbital (HOMO) for 1, 2, 14, 16-18, 20, 23, 26 and their electrostatic charge are shown in Figure 7 allowing us to propose the three putative debromination pathways. OAc group on ring A could be δ 2.24, while the other -OAc could assigned to δ 2.36 (in Me2CO-d6). Compound 13 is determined as 2,3,5-tribromo-6-(3′,5′-dibromo-2′-acetoxyphenoxy) phenyl acetate. To complete the characterization of acetylation products 10-13, theoretical 13 C NMR chemical shifts of the compounds are proposed as in Figure 6b.  scavengers of bromine in the presence of acetic acid. Compounds (1, 2, 14, 16-18, 20, 23, 26) were isolated and confirmed from the reaction. MS and 1 H NMR spectra of compounds (1, 2, 16, 17, 20, 23, 26) were in a good agreement with previous reports [14,15]. Debromination of 1, including its product, can be supported by DFT ωB97X-D/6-311 + G(2d,p) calculations. Combination analysis of frontier molecular orbital (HOMO) for 1, 2, 14, 16-18, 20, 23, 26 and their electrostatic charge are shown in Figure 7 allowing us to propose the three putative debromination pathways.   Only o, p-bromines were selectively reduced; hence, we could deduce the isolated products with the three putative debromination pathways A-C shown in Figure 7. Debromination in this study gave nine compounds, (1, 2, 14, 16-18, 20, 23, 26). In addition, there were 8 hypothetical compounds (21, 22, 24, 25, 27-30) that we have predicted on the basis of computational studies. This includes 15, which was reported previously [21,24,25]. Based on analysis of FMO and electrostatic charge for 1 on ring A, three reactive sites were readily identified in HOMO region as C2 (electrostatic charge: −0.125), C5 (−0.035), and C6 (−0.032) corresponding to Br atoms in the ortho and para positions toward the -OH group and the Br atom in the ortho position toward the −OR group, respectively which led to the three putative pathways A-C. Moreover, combination analysis of FMOs and electrostatic charges for the three pathways support the presence of debromination products as in 1, 2, 14, 16-18, 20, 23, and 26. The high electron density and negative charge gave the possibility of attack through the ortho or para positions of the ring toward the lowest unoccupied molecular orbital (LUMO) HBr orbital leading to the selectivity of the reaction. Reaction products were predicted to occur via the ortho and para pathways toward hydroxy or phenoxy groups. This was because of competition between two activating groups (regioselective). Reaction products predicted to occur via the ortho toward hydroxy groups (pathway A) were 1 → 14 → 16 (the scheme continues on the pathway B); 1 14 → 19. Reaction products predicted to occur via the para toward hydroxy groups (pathway B) Reaction products predicted to occur via the para toward phenoxy groups (pathway Based on Figure 7, we also observed the possibility of isomerization, which also supports the debromination mechanism proposed by Effenberger [17]. This may explain other products that are not directly obtained by removing bromine atoms in the ortho or para positions, to the electron-donating groups (EDG) -OH or-OR. Among the debromination products, compound 26 was the most stable (∆E LUMO-HOMO 9.38 eV) followed by 16 (9.14 eV), while less stable compound was 1 (8.62 eV). All hypothetical compounds (21, 22, 24, 25, 27-30) were relatively stable (9.03-9.29 eV) except for 24 (8.86 eV). Apparently, the molecules may exist during the reaction, offering a challenging task to isolate and to assay them with our interest assay panel because of nature of the molecules. In this computational study, we also proposed theoretical 13 Table 2). The result was supported by the shape of HOMO and LUMO for 1, 16, and 30 ( Figure 9). Compounds 1 and 16 showed the HOMO-LUMO orbitals located in different molecular regions and this characteristic may explain their potent antibacterial activity, as shown in many studies [9,26,27]. In contrast, 30 had HOMO-LUMO orbitals dispersed across almost the entire molecular region. While the MID of 1 and 16 were the same, the clear inhibition zone of 1 is slightly stronger than that of 16. This was supported by E LUMO of 1 (−0.05 eV) < E LUMO of 16 (0.38 eV) while E LUMO of 30 was 1.13 eV. It is clear that the higher number of bromine atoms, as in 1 stabilizes LUMO energy and gives slightly stronger antibacterial activity against the Gram-positive bacterium, S. aureus [27]. In contrast, the antibacterial activity of 16 against Gram-negative bacterium, K. pneumoniae, was slightly more potent than that of 1.    (−0.05 eV) < ELUMO of 16 (0.38 eV) while ELUMO of 30 was 1.13 eV. It is clear that the higher number of bromine atoms, as in 1 stabilizes LUMO energy and gives slightly stronger antibacterial activity against the Gram-positive bacterium, S. aureus [27]. In contrast, the antibacterial activity of 16 against Gram-negative bacterium, K. pneumoniae, was slightly more potent than that of 1.    Compound 16 was discovered previously in the marine sponge, Dysidea herbacea collected from Australia [29] and later in Lamellodysidea sp. [24] collected from Papua New Guinea. It showed a wide range of antibiotic activities against Staphylococcus aureus (MIC 1.25-1.6 µg/mL), Enterococcus faecium (MIC 1.6-3.1 µg/mL), Escherichia coli (MIC 50 µg/mL), Pseudomonas aeruginosa and Candida albicans (MIC >50 µg/mL), while it also had an IC 50 >50 µg/mL against Bsc-1 cells [24]. Methylation and acetylation of 1 reduced cytotoxicity, with an IC 50 >10 µg/mL against HEK293T, as in 7-13. Meanwhile, antibacterial activity of methylation and acetylation products was weaker, with MID >0.078 µg/disk against Gram-positive Staphylococcus aureus and Gram-negative Klebsiella pneumoniae.
In summary, we found new derivatives 10, 12, 13 from acetylation by comparing data published between 1969 and 2020. We performed the acetylation with a new method using green chemistry at room temperature and determined the structures using analysis of FMOs and electrostatic charges. In addition, we also discovered a new debromination product 18 and new crystal structures of methylated 8 and 9 with twist conformations (ϕ 1 , ϕ 2 > 0 • ) [30]. The presence of debromination products (1, 2, 14, 16-18, 20, 23, 26) and hypothetical molecules (21, 22, 24, 25, 27-30) were predicted using analysis of FMOs and electrostatic charges showing regioselectivity ortho and para positions toward the EDG -OH or -OR groups in PBDEs. The analysis also showed the possibility of isomerization among the PBDEs. All products methylations, acetylations, and debrominations were characterized for their theoretical 13 C NMR chemical shifts. Novel 13 C NMR data 6 are also reported (RMSE 3.2 ppm). PBDE compounds that lose hydroxyl groups, due to methylation and acetylation, have weaker biological activity. Fewer bromines, as in 16, resulted in a significant IC 50 8.65 ± 1.11; 8.11 ± 1.43 µM against HEK293T cells. Compound 16 also showed antibacterial activity against Gram-positive Staphylococcus aureus, as well as Gramnegative Klebsiella pneumoniae, with an MID = 0.078 µg/disk. Cytotoxicity and antibacterial assays of derived compounds show that two phenolic hydroxyl groups and four bromine atoms are important for these activities. The result of active compounds was supported by analysis of FMOs. Additional information can be found in the Supplementary Materials.

General Methods
NMR spectra were measured on a 500 MHz Bruker Avance III spectrometer (MA, USA) or a 500 MHz JEOL (Tokyo, Japan) or a 500 MHz Varian (CA, USA). Chemical shifts were referenced to tetramethylsilane (TMS) or acetone (Me 2 CO) signals. MS spectra were recorded on a Waters Acquity Xevo G2-S ESIQTOF in positive mode or an HRESITOFMS JEOL T100LP, or EIMS were measured on a Hitachi M-2500 instrument. UV and IR spectra were obtained on a Perkin Elmer Spectrum One FTIR and on a Shimadzu Pharmaspec 1700 spectrophotometer. X-ray analysis was performed on a Rigaku AFC10 goniometer equipped with a Saturn 724+ detector. High-performance liquid chromatography (HPLC) separations were carried out on a Hitachi L-6000 pump fitted with Shodex RI-101 refractive index and SPD-20A Shimadzu UV detectors, or a Shimadzu HPLC with Prominence LC-20AD, DGU-20A5, SPD-20A. A Cosmosil 5SL-II-MS (10 × 250 mm) column was used for HPLC. Analytical thin-layer chromatography (TLC) was performed on Merck silica gel 60 F 254 plates and visualized with sulfuric acid with cerium sulfate. All solvents used were reagent grade.

Animal Material
Marine sponges were collected by hand while scuba diving in Banten Province, Indonesia at a depth of 5-10 m. Samples were then stored in EtOH. Sponges were identified as Lamellodysidea herbacea by NJdeV.

Extraction and Isolation
A fresh marine sponge specimen (wet weight 300 g) stored in EtOH was extracted with MeOH. The combined extract was concentrated under vacuum, and the resulting residue was partitioned between hexane and aqueous MeOH (90%). The latter layer was further partitioned between CH 2 Cl 2 and aqueous MeOH (50%). Finally, the aqueous MeOH (50%) was removed and adjusted with water, followed by extraction with n-BuOH. The three layers: Hexane, CH 2 Cl 2 and BuOH layer were evaluated for activity against Gram-positive and Gram-negative bacteria. Recrystallization of the CH 2 Cl 2 layer using CHCl 3 -Me 2 CO-MeCN gave 1 (1.36 g). A non-crystalline fraction was separated using either a silica gel column or silica HPLC eluted with hexane/EtOAc/MeOH, followed by recrystallization to give compound 2 (5.4 mg), a mixture of compounds 2 and 3 (5.6 mg), compound 4 (7.3 mg), and compound 5 (0.8 mg). Additional 2 (8.3 mg) and 3 (11.9 mg) were isolated from the hexane fraction after open column chromatography. Compounds 4 (163.5 mg) and 6 (50.6 mg) were isolated from the hexane layer collected from another L. herbacea.

Methylation
To a solution of 1 (9.1 mg) in MeOH (1 mL), 2 M TMSCHN 2 in hexane was added dropwise. The reaction was monitored by TLC. The solution was allowed to stand at room temperature and concentrated to dryness under a stream of nitrogen followed by purification using HPLC (RP 18, MeOH, MeCN + 0.1% TFA) to give 7 (3 mg) and 8 ( To a solution of 2 (5.2 mg) in MeOH (1 mL) excess 2 M TMSCHN 2 in hexane was added and monitored by TLC. The solution was allowed to stand at room temperature and concentrated to dryness under a stream of nitrogen to give the total methyl derivative 9 (5.2 mg). Compound 9: R f 0.63 (Hex/EtOAc 8:1, n-silica, UV λ 254 nm), 1

Acetylation
Screening of acetylation was performed under seven conditions with a variety of catalysts and solvents, without catalyst and solvent, with or without sonication. Acetylation of 1 with Ac 2 O and sonication for 1 h proceeded to give 10 and 11. Acetylation of 2 with Ac 2 O and sonication for 1 h proceeded to give 12 and 13. The reaction was performed without DMAP, Et 3 N, and solvent.

Debromination
HBr (10 mL) and Na 2 SO 3 (20 equivalents) in AcOH (20 mL) were added to compound 1 (60 mg). The mixture was refluxed for 6 h and then neutralized with dilute KOH to pH 7 and partitioned using EtOAc and H 2 O. The organic layer formed was separated and purified by HPLC (RP18, MeOH) to give 1 (1. HBr (2.5 mL) and Na 2 SO 3 (15 equivalents) in AcOH (20 mL) were added to compound 1 (55.5 mg). The mixture was refluxed for 6 h and then dilute KOH was added to raise the pH to 10-11, after which it was partitioned using EtOAc and H 2 O. The organic layer formed . To 1 (60.6 mg) in AcOH HBr (10 mL) and Na 2 SO 3 (20 equivalents) in AcOH (20 mL) were added. The mixture was refluxed for 24 h and then neutralized with dilute KOH to pH 7 and partitioned using EtOAc and water. The organic layer formed was separated and purified by HPLC (RP 18, MeCN + 0.1% TFA) to give 16 (4.1 mg) and 20 (0.5 mg).

X-ray Study
Single crystals of C 12 H 6 Br 6 O 4 (1), C 14 H 8 Br 6 O 3 (8), and C 14 H 9 Br 5 O 3 (9) were supplied. A suitable crystal was selected and mounted on Rigaku Saturn 724 Plus with AFC10. The crystal was kept at 113 K or 123.15 K during data collection. Using Olex2 [31], the structure was solved with the SHELXT [32] structure solution program using direct methods and refined with the SHELXL [33] refinement package using least squares minimization.

Agar-Plate Diffusion Assay
Staphylococcus aureus ATCC 6538 and Klebsiella pneumoniae were used for biological evaluation. Concentrations assayed ranged from 0.08 to 1.25 µg/disks [14,15]. DMSO was used to dissolve the compounds, while vancomycin and oxacillin were used as positive controls for Staphylococcus aureus and gentamycin was used as a positive control for Klebsiella pneumoniae. The minimum inhibitory dose (MID, µg/disk) was defined as the minimum dose that induced an obvious inhibition zone (1-1.5 mm) [28]. The disk diameter was 6 mm.

In Vitro Cytotoxicity Assay
In vitro cytotoxicity was determined against human embryonic kidney (HEK293T) cells. The assay was performed in 96-well treated tissue culture plates. Cells were seeded in the wells (5000 cells in 100 µL media containing RPMI1640, FBS, penicillin and streptomycin) and incubated for 24 h. Samples were then added and plates were again incubated for 48 h. MTT was then added and the plates were incubated for 4 h at 37 • C. Formazan crystals formed were dissolved in EtOH and absorbances were read at λ = 595 nm. The result was analyzed by using Prism 9 software (Graphpad, San Diego, CA, USA) to obtain IC 50 values.