Svalbamides A and B, Pyrrolidinone-Bearing Lipodipeptides from Arctic Paenibacillus sp.

Two new secondary metabolites, svalbamides A (1) and B (2), were isolated from a culture extract of Paenibacillus sp. SVB7 that was isolated from surface sediment from a core (HH17-1085) taken in the Svalbard archipelago in the Arctic Ocean. The combinational analysis of HR-MS and NMR spectroscopic data revealed the structures of 1 and 2 as being lipopeptides bearing 3-amino-2-pyrrolidinone, d-valine, and 3-hydroxy-8-methyldecanoic acid. The absolute configurations of the amino acid residues in svalbamides A and B were determined using the advanced Marfey’s method, in which the hydrolysates of 1 and 2 were derivatized with l- and d- forms of 1-fluoro-2,4-dinitrophenyl-5-alanine amide (FDAA). The absolute configurations of 1 and 2 were completely assigned by deducing the stereochemistry of 3-hydroxy-8-methyldecanoic acid based on DP4 calculations. Svalbamides A and B induced quinone reductase activity in Hepa1c1c7 murine hepatoma cells, indicating that they represent chemotypes with a potential for functioning as chemopreventive agents.


Introduction
Marine habitats were generally recognized as extreme environments exposing organisms to conditions of high salt, high pressure, and hypoxia, forcing them to develop unique physiologies in comparison to their terrestrial counterparts. Among the marine organisms, bacteria have always contributed significantly as a source for the discovery of new marine natural products, with 232 new compounds found in 2019 alone [1]. However, most (62.5%) of the new marine-bacterial molecules were derived from the single genus Streptomyces. We also reported the dimeric benz[a]anthracene thioethers donghaesulfins A and B, and rearranged angucyclinones donghaecyclinones A-C from marine-derived Streptomyces sp. SUD119 in 2019 and 2020, respectively [2,3]. Even though Streptomyces is chemically prolific and still provides numerous new bioactive compounds, the desire for compounds with greater structural diversity has brought about chemical investigations of a wider diversity of bacteria that extend beyond conventional phylogenetically biased chemical studies. The chemical examination of bacteria inhabiting the Arctic Ocean-a more extreme habitat than tropical or subtropical oceans that remains poorly investigated-represents a promising strategy for the discovery of new bioactive molecules.
In our continuing efforts to search for new bioactive microbial compounds from extreme marine environments, we explored the chemistry of bacterial strains from the Arctic Ocean. Our initial chemical profiling of Arctic strains led to the discovery of articoside and In our continuing efforts to search for new bioactive microbial compounds from extreme marine environments, we explored the chemistry of bacterial strains from the Arctic Ocean. Our initial chemical profiling of Arctic strains led to the discovery of articoside and C-1027 chromophore-V-two new benzoxazine-bearing compounds that inhibit Candida albicans isocitrate lyase-from Streptomyces sp. ART5 collected from the East Siberian continental margin [4]. In this study, we diversified the phylogeny of bacteria for chemical analysis and focused on non-Streptomyces bacterial strains inhabiting the Arctic Ocean. The Paenibacillus sp. SVB7 strain was isolated from sediment collected at the continental shelf (depth = 322 m) off Wijdefjorden, Svalbard, during a marine-geoscientific cruise to North Spitsbergen in 2017. Cultivation in liquid media and the LC/MS-based chemical examination of the strain Paenibacillus sp. SVB7 identified the production of previously unreported molecules with the molecular ions at m/z 384. Scaling-up of the culture enabled us to purify two new compounds, svalbamides A and B, and subsequently elucidate their structures by spectroscopic analysis, chemical derivatization, and quantum mechanics-based calculation. Here, we report the structural determination of svalbamides A and B (1, 2; Figure 1) along with their biological activity.

Phylogenetic Analysis
Sequence comparison using the almost complete 16S rRNA gene sequence of strain SVB7 (1440 bp) in BLASTn and EzBioCloud searches revealed that strain SVB7 belongs to the genus Paenibacillus of the Paenibacillaceae family. According to 16S rRNA gene sequence similarities, strain SVB7 was most closely related to P. maysiensis SX-49 T (99.30% similarity), followed by P. terrae AM141 T (98.26%), and P. peoriae DSM 8320 T (97.42%). In all phylogenetic trees inferred by maximum likelihood, neighbor-joining, and minimumevolution methods, strain SVB7 was located within the Paenibacillus clade and formed a robust clade with P. maysiensis and P. terrae, providing clear support for its genus being classified as Paenibacillus ( Figure 2). Based on the formation of the robust clade with P. maysiensis SX-49 T and > 98.7% 16S rRNA gene sequence similarity, it is likely that strain SVB7 is a member of Paenibacillus maysiensis. However, inclusion of strain SVB7 must be confirmed using whole-genome sequencing analysis.

Phylogenetic Analysis
Sequence comparison using the almost complete 16S rRNA gene sequence of strain SVB7 (1440 bp) in BLASTn and EzBioCloud searches revealed that strain SVB7 belongs to the genus Paenibacillus of the Paenibacillaceae family. According to 16S rRNA gene sequence similarities, strain SVB7 was most closely related to P. maysiensis SX-49 T (99.30% similarity), followed by P. terrae AM141 T (98.26%), and P. peoriae DSM 8320 T (97.42%). In all phylogenetic trees inferred by maximum likelihood, neighbor-joining, and minimumevolution methods, strain SVB7 was located within the Paenibacillus clade and formed a robust clade with P. maysiensis and P. terrae, providing clear support for its genus being classified as Paenibacillus ( Figure 2). Based on the formation of the robust clade with P. maysiensis SX-49 T and > 98.7% 16S rRNA gene sequence similarity, it is likely that strain SVB7 is a member of Paenibacillus maysiensis. However, inclusion of strain SVB7 must be confirmed using whole-genome sequencing analysis.

Structural Elucidation
Svalbamide A (1) was isolated as a white powder. The molecular formula of 1 was assigned as C 20 H 37 N 3 O 4 , which has an unsaturation number of 4 based on high-resolution electrospray ionization (HR-ESI) mass spectrometry ([M + H] + at m/z 384.2851, calculated as 384.2857) along with 1 H and 13 C NMR spectra. The 13 C NMR spectra of 1 showed three carbonyl carbon (δ C 174.2, 171.0, and 170.8), one oxygenated methine carbon (δ C 67.5), and two α-amino methine carbon (δ C 57. 2
The interpretation of COSY, TOCSY, and HMBC NMR spectra enabled us to determine the partial structures of 1.  The interpretation of COSY, TOCSY, and HMBC NMR spectra enabled us to determine the partial structures of 1. First, the 2-NH (δ H 8.10)/H-2 (δ H 4.30) COSY correlation connected 2-NH to the C-2 α-carbon (δ C 49.4). The TOCSY and COSY correlations among H-2, H-3a and H-3b (δ H 1.82 and 2.26), H 2 -4 (δ H 3.16), and 4-NH (δ H 7.78) secured the spin system from 2-NH to 4-NH. The HMBC correlations from 4-NH to C-1 (δ C 174.2), C-2 (δ C 49.4), C-3 (δ C 28.0), and C-4 (δ C 38.0), and from 2-NH to C-1 (δ C 174.2), led to elucidation of the substructure as a 3-amino-2-pyrrolidinone. The structure of valine was assigned based on the 1 H-1 H couplings of 6-NH (δ H 7.83), H-6 (δ H 4.18), H-7 (δ H 1.96), H 3 -8 (δ H 0.84), and H 3 -9 (δ H 0.88) in the COSY and TOCSY NMR spectra along with the H-6/C-5 (δ C 171.0) HMBC correlation. The remaining part of the molecule was composed of a lipophilic acyl chain. The protons in the substructure from C-11 to C-20 belong to a single spin system as revealed by COSY/TOCSY correlations of the protons. These protons showed correlation peaks with an exchangeable proton at δ H 4.65 in the TOCSY spectrum, indicating that the exchangeable proton is also included in the spin system ( Figure 3). , and C-18 (δC 28.9), were finally assigned to 3-hydroxy-8-methyldecanoic acid ( Figure 3). Consequently, the four unsaturation equivalents were fully explained by one pyrrolidinone ring containing one carbonyl group and two more carbonyl functional groups. Thus, svalbamide A (1) must not possess an additional ring and comprises a combination of the three substructures as a linear molecule.
Svalbamide B (2) was isolated as a white powder, and its molecular formula was determined to be C 20 H 37 N 3 O 4 , which contains four double bond equivalents, using highresolution electrospray ionization (HR-ESI) mass spectrometry ([M + H] + at m/z 384.2845, calculated as 384.2857). This molecular formula was identical to that of svalbamide A (1). The 1 H and 13 C NMR data of 2 in DMSO-d 6 were extremely similar to those of 1 (Table 1), but distinct differences in chemical shifts were found mainly in the 3-amino-2pyrrolidinone unit. Specifically, H-2 in 1 (δ H 4.30) was shifted upfield in 2 (δ H 4.27), while signals for H-3a and H-3b in 1, at δ H 1.82 and 2.26, were detected at δ H 1.76 and 2.29 in 2. C-3 (δ C 28.0) also shifted slightly to the deshielded region by 0.3 ppm in svalbamide B (2). Comprehensive analysis of 1D and 2D NMR data indicated the planar structure of 2 to be the same as 1 (Figure 3). Based on the observation that the distinct chemical shift differences were found in 3-amino-2-pyrrolidone unit, the structure of svalbamide B (2) was expected to have stereochemical modification in this residue.

Biological Evaluation
The biological activities of svalbamides A (1) and B (2) were evaluated in several ways. First, we measured cytotoxicity against various cancer cell lines [5], including HCT116 (human colorectal cancer cells), MDA-MB-231 (human breast cancer cells), A549 (human lung cancer cells), SK-HEP-1 (human liver cancer cells), and SNU-638 (human gastric cancer cells), but 1 and 2 showed no significant cytotoxicity against the tested cell lines even at 50 μM. Therefore, we evaluated the detoxification ability by measuring quinone reductase (QR) activity. QR is a major phase II detoxification enzyme, and the in-

Biological Evaluation
The biological activities of svalbamides A (1) and B (2) were evaluated in several ways. First, we measured cytotoxicity against various cancer cell lines [5], including HCT116 (human colorectal cancer cells), MDA-MB-231 (human breast cancer cells), A549 (human Mar. Drugs 2021, 19, 229 6 of 10 lung cancer cells), SK-HEP-1 (human liver cancer cells), and SNU-638 (human gastric cancer cells), but 1 and 2 showed no significant cytotoxicity against the tested cell lines even at 50 µM. Therefore, we evaluated the detoxification ability by measuring quinone reductase (QR) activity. QR is a major phase II detoxification enzyme, and the induction of QR activity is considered as a strategy to increase the chemoprevention effect. Svalbamide A (1) enhanced QR activity by 1.45-, 1.98-, and 2.54-fold at 10, 20, and 40 µM, respectively, in a concentration-dependent manner. In addition, svalbamide B (2) effectively induced QR activity by 1.93-, 2.64-, and 2.98-fold at 10, 20, and 40 µM, respectively (Figure 5). At a concentration of 40 µM, it exhibited a comparable level of QR activity induction in the positive control of 1 µM β-naphthoflavone (β-NF). These results suggest that 1 and 2 are potential chemotypes with chemopreventive activity.

Biological Evaluation
The biological activities of svalbamides A (1) and B (2) were evaluated in several ways. First, we measured cytotoxicity against various cancer cell lines [5], including HCT116 (human colorectal cancer cells), MDA-MB-231 (human breast cancer cells), A549 (human lung cancer cells), SK-HEP-1 (human liver cancer cells), and SNU-638 (human gastric cancer cells), but 1 and 2 showed no significant cytotoxicity against the tested cell lines even at 50 μM. Therefore, we evaluated the detoxification ability by measuring quinone reductase (QR) activity. QR is a major phase II detoxification enzyme, and the induction of QR activity is considered as a strategy to increase the chemoprevention effect. Svalbamide A (1) enhanced QR activity by 1.45-, 1.98-, and 2.54-fold at 10, 20, and 40 μM, respectively, in a concentration-dependent manner. In addition, svalbamide B (2) effectively induced QR activity by 1.93-, 2.64-, and 2.98-fold at 10, 20, and 40 μM, respectively (Figure 5). At a concentration of 40 μM, it exhibited a comparable level of QR activity induction in the positive control of 1 μM β-naphthoflavone (β-NF). These results suggest that 1 and 2 are potential chemotypes with chemopreventive activity.   2) showed QR induction activity, with 2.54-and 2.98-fold increases, respectively, at 40 µM. All data represent the mean ± SD (n = 3). *** p < 0.001 compared to the control.

General Experimental Procedures
Optical rotations were measured using a JASCO P-2000 polarimeter (sodium light source, JASCO, Easton, PA, USA) with a 1 cm cell. IR spectra were obtained using a Thermo NICOLET iS10 spectrometer (Thermo, Madison, CT, USA). 1 H, 13 C, and 2D NMR spectra were recorded on a Bruker Avance 800 MHz spectrometer (Bruker, Billerica, MA, USA) at the Research Institute of Pharmaceutical Sciences, Seoul National University. ESI low-resolution LC/MS data were recorded using an Agilent Technologies 6130 Quadrupole mass spectrometer (Agilent Technologies, Santa Clara, CA, USA) coupled with an Agilent Technologies 1200 series high-performance liquid chromatography (HPLC) instrument using a reversed-phase C 18 (2) column (Phenomenex Luna, 100 × 4.6 mm). HR-ESI mass spectra were acquired on a high-resolution LC/MS-MS spectrometer (Q-TOF 5600) at the National Instrumentation Center for Environmental Management (NICEM) in the College of Agriculture and Life Sciences at Seoul National University.

Isolation, Cultivation, Phylogenetic Analysis, and Extraction of Bacteria
During the Korea-Norway Joint marine-geoscientific cruise with RV Helmer Hanssen to North Spitsbergen in 2017, a sediment core (HH17-1085) was taken at a water depth of 322 m with a giant box corer from the continental shelf off Wijdefjorden (80 • 16.469' N, 016 • 12.625' E) in Svalbard ( Figure 6). Surface sediment corresponding to 1 cm depth was collected from core HH17-1085. A portion of the sample (2 g) was diluted in 20 mL sterilized water and vortexed. The mixture was spread on YEME isolation solid medium (500 mL of sterilized water, 9 g agar, 100 mg cycloheximide, 2 g yeast, 5 g malt, and 2 g glucose) for two weeks. Strain SVB7 was isolated in YEME medium after one week of incubation for further study.
The 16S rRNA gene sequence of strain SVB7 was obtained by Sanger sequencing using PCR products amplified with the universal primers 27F and 1492R. The resultant 16S rRNA gene sequence (1440 bp) was queried in a BLASTn search implemented at GenBank and was also identified by the "16S-based ID service" in the EzbioCloud database [6]. For phylogenetic analysis, sequences of strain SVB7 and its close relatives retrieved from the EzBioCloud database were aligned with the SINA online aligner [7]. Using the aligned sequences, phylogenetic trees were inferred by maximum likelihood, neighbor-joining, and minimum-evolution algorithms implemented in MEGA software version 7.0 [8].

Conformational Search and DP4 Analysis.
A conformational search was carried out using a mixed sampling method of torsional/low-mode using MacroModel (version 9.9, Schrodinger LLC) in the Maestro suite (version 9.9, Schrodinger LLC). A total of 158 conformers of the diastereomers were identified with relative potential energies below 10 kJ/mol using the MMFF force field. The shielding tensor values of the optimized conformers were calculated based on the equation below, where is the calculated NMR chemical shift for nucleus x, and σ o is the The SVB7 strain was cultured in 50 mL modified K medium (4 g yeast extract, 5 g malt extract, 5 g soytone, 5 g soluble starch, 5 g mannitol, 2 g glucose, and 6 g glycerol in 1 L deionized water) in a 125 mL Erlenmeyer flask. After cultivation for 2 days on a rotary shaker at 200 rpm and 30 • C, 5 mL of the culture medium was inoculated in 200 mL of modified K medium in a 500 mL Erlenmeyer flask. After cultivation for 2 days under the same incubation conditions, 10 mL of the culture medium was inoculated in 1 L of modified K medium in 2.8 L Fernbach flasks (200 ea × 1 L) at 170 rpm and 30 • C for 6 days. The whole culture of SVB7 was extracted with 300 L of EtOAc. The EtOAc and water layers were separated, and the remaining water in the EtOAc layer was removed by adding anhydrous sodium sulfate. The extract was concentrated using a rotary evaporator, yielding 50 g of dry material.

Isolation of Svalbamides A and B
The crude extract was divided into ten equal parts and fractioned over a C 18 reversedphase open column (φ 6.5 × 10 cm) with 500 mL of 20%, 40%, 60%, 80%, and 100% MeOH-H 2 O. The 80% MeOH fraction was subjected to a reversed-phase HPLC (Kromasil C 18 , 5 µm, 250 × 10 mm, flow rate = 2 mL/min) using a gradient solvent system from 35% to 75% CH 3 CN-H 2 O over 40 min with 210 nm UV detection. Svalbamides A and B were collected as one broad peak at 25 min. The recorded 1 H NMR spectra of the peak indicated a diastereomeric mixture, prompting further purification on a chiral HPLC column (CHIRALPAK IB, 5 µm, 250 × 4.6 mm, flow rate = 0.6 mL/min) using a step gradient 45% CH 3 CN isocratic solvent system over 10 min, followed by a 60% CH 3 CN-H 2 O isocratic solvent system from 10 to 50 min. Svalbamides A (10.0 mg) and B (7.8 mg) were isolated at 30 and 31.5 min, respectively.

Conformational Search and DP4 Analysis
A conformational search was carried out using a mixed sampling method of torsional/ low-mode using MacroModel (version 9.9, Schrödinger LLC) in the Maestro suite (version 9.9, Schrödinger LLC). A total of 158 conformers of the diastereomers were identified with relative potential energies below 10 kJ/mol using the MMFF force field. The shielding tensor values of the optimized conformers were calculated based on the equation below, where δ x calc is the calculated NMR chemical shift for nucleus x, and σ o is the shielding tensor for the proton and carbon nuclei calculated at the B3LYP/6-31++ level. These values were averaged via the Boltzmann population with the associated Gibbs free energy and utilized for the DP4 analysis, which was facilitated using an Excel spreadsheet provided by the authors of [9] and as described in their publication.

Quinone Reductase Assay
Hepa1c1c7 murine hepatoma cells (American Type Culture Collection, Manassas, VA, USA) were used to investigate QR induction activity. The test cells were seeded (3 × 10 4 cells/mL) and incubated at 37 • C for 24 h with 5% CO 2 containing humidified atmosphere. The plates were then exposed to svalbamides A and B (1, 2), including a positive control compound, β-naphthoflavone (β-NF). After 24 h, the media were decanted from the wells, and the cells in each well were lysed by incubation at 37 • C with 250 µL of a mixed solution consisting of 10 mM Tris-HCl pH 8.0, 140 mM NaCl, 15 mM MgCl 2 , and 0.5% NP-40 (IGEPAL CA-630, Sigma, St. Louis, MO, USA) for 10 min. A 1 mL aliquot of the complete reaction mixture (12.5 mM Tris-HCl pH 7.4, 0.67 mg/mL bovine serum albumin (BSA), 0.01% Tween-20, 50 µM flavin adenine dinucleotide (FAD), 1 mM glucose-6-phosphate, 2 U/mL glucose-6-phosphate dehydrogenase, 30 µM NADP, 50 µg/mL 3-(4,5-dimethylthiazo-2-yl)-2,5-diphenyltetrazolium bromide (MTT), and 50 µM menadione) was added to each of the wells, and the plates were incubated at 25 • C for the colorimetric reaction. The rate of NADPH-dependent menadiol-mediated reduction of MTT in this reaction was measured at 610 nm, and cytotoxicity was determined by crystal violet staining of an identical set of the test plates. The quinone reductase activity was calculated from the following equation: absorbance change for MTT per min/absorbance of crystal violet × 3345 nmol/mg. The value of 3345 nmol/mg represents the ratio of the extinction coefficient of MTT and the proportionality constant of crystal violet. The relative QR activity was normalized using controls [10].

Conclusions
Our chemical study of the Arctic sediment-derived Paenibacillus sp. SVB7 led to the discovery and structural elucidation of two new pyrrolidinone-bearing lipodipeptides, svalbamides A (1) and B (2), in which QR activity could be induced. Based on spectroscopic analysis, advanced Marfey's analysis, and DP4 calculation, these two compounds were identified to have a diastereomeric relationship with alternative absolute configurations at the 3-amino-2-pyrrolidione unit. Svalbamides A and B are structurally unique as they contain 3-amino-2-pyrrolidinone amino acid. This amino acid unit was rarely reported in natural products, with the only example being of actinoramide E, an antimalarial peptide from a marine-derived Streptomyces strain [11]. 3-Hydroxy-8-methyldecanoic acid is another interesting component. This saturated fatty acid was occasionally found in natural products from Paenibacillus and related bacteria. For example, tridecaptins A-C were first isolated from Bacillus polymyxa in 1978 and were studied in various fields, with reports of new derivatives and biosynthesis undertaken to date [12]. A series of new tridecaptin compounds containing 3-hydroxy-8-methyldecanoic acid were discovered in a Paenibacillus strain collected in the deep oligotrophic Krubera-Voronja cave. However, the absolute configuration of this fatty acid was not determined [13][14][15]. Octapeptin and cerexin from the Bacillus sp. bear the same fatty acid, but no experiments have yet been conducted to reveal the absolute stereochemistry [16,17]. Therefore, svalbamides A and B are the first metabolites for which the stereochemistry of 3-hydroxy-8-methyldecanoic acid was addressed. Discovering these new bioactive secondary metabolites from Paenibacillus from the polar region indicates that chemical studies of underinvestigated bacterial taxa in marine extreme habitats, such as the Arctic Ocean, could lead to the discovery of significant natural chemical diversity with pharmaceutical potential in terms of drug discovery.