Securamine Derivatives from the Arctic Bryozoan ... - ACS Publications

Aug 21, 2017 - Department of Chemistry, UiT The Arctic University of Norway, Breivika, N-9037 Tromsø, Norway. •S Supporting Information. ABSTRACT: ...
1 downloads 0 Views 3MB Size
Article Cite This: J. Nat. Prod. 2017, 80, 3276−3283

pubs.acs.org/jnp

Securamine Derivatives from the Arctic Bryozoan Securif lustra securif rons Kine Ø. Hansen,*,† Johan Isaksson,‡ Annette Bayer,‡ Jostein A. Johansen,‡ Jeanette H. Andersen,† and Espen Hansen† †

Marbio, UiT The Arctic University of Norway, Breivika, N-9037 Tromsø, Norway Department of Chemistry, UiT The Arctic University of Norway, Breivika, N-9037 Tromsø, Norway



S Supporting Information *

ABSTRACT: Bryozoans belonging to the Flustridae family have proven to be a rich source of structurally unique secondary metabolites. As part of our continuing search for bioactive secondary metabolites from Arctic marine invertebrates, the organic extract of Securif lustra securif rons was examined. This resulted in the isolation of three new halogenated, hexacyclic indole-imidazole alkaloids, securamines H−J (1−3), together with the previously reported compounds securamines C (4) and E (5). The structures of the new compounds were elucidated by spectroscopic methods including 1D and 2D NMR and analysis of HRMS data. Through NMR and HRMS analysis, we were also able to prove that 1, 2, 4, and 5, when dissolved in MeOH, were converted into their corresponding artifacts, the securamine MeOH adducts m1, m2, m4, and m5. When redissolved in a nonnucleophilic solvent, the native variants were re-formed. We also found that 3 was a MeOH addition product of a native variant. Even though the structures of several securamines have been reported, their bioactivities were not examined. The securamines displayed various degrees of cytotoxicity against the human cancer cell lines A2058 (skin), HT-29 (colon), and MCF-7 (breast), as well as against nonmalignant human MRC-5 lung fibroblasts. Compounds 1, 2, and 5 were found to be active, with IC50 values against the cancer cell lines ranging from 1.4 ± 0.1 to 10 ± 1 μM. The cytotoxicity of 1 was further evaluated and found to be time-dependent. nimals belonging to the phylum bryozoa are benthic filter feeders living in intricate colonies where zooids are interconnected and cooperate for feeding, breeding, and defense.1−3 For the latter they largely rely on a well-developed arsenal of secondary metabolites, which is reflected in the persistent yield of compounds with complex structures isolated from bryozoans. 4 Their value as lead structures for pharmaceutical development is indicated by their notable biological activities.5 Examples of such compounds are the caulibugulones, cytotoxic isoquinoline quinone alkaloids from Caulibugula intermis,6 and the wilsoniamines, hexahydropyrrole imidazoles with antimalaria and antitrypanosomal activity, isolated from Amathia wilsoni.7 Another example is the wellknown bryostatins, a group of macrolide lactones isolated from Bugula neritina, known for their anticancer,8 antiviral,9 and memory-enhancing10 properties. The bryostatins are, as many of the secondary metabolites isolated from collected bryozoans, produced by a species-specific endosymbiotic microorganism.5,11−13 In this paper, the organic extract of Securif lustra securif rons was investigated. Previously, we have reported the compound securidine A isolated from the aqueous extract of the same organism.14 S. securif rons is part of the Flustridae family, from which the compounds flustramines,15 chartel-

A

© 2017 American Chemical Society and American Society of Pharmacognosy

lines,16 and terminoflustrindoles,17,18 in addition to securidine A,14 have been reported. Despite being isolated from the same collected sample of S. securif rons, securidine A and the herein reported securamines have unrelated chemical structures. In this continuing examination of the chemistry of S. securif rons and as part of our ongoing search for bioactive secondary metabolites from Arctic marine invertebrates,19 fractions of the organic extract from S. securif rons were found to be cytotoxic to the human melanoma cell line A2058. Through detailed chemical investigation and an exhaustive isolation protocol, we were able to isolate three new compounds, securamines H−J (1−3), along with two previously reported variants, securamines C (4)20 and E (5).21 The structures of the previously reported securamines were confirmed by comparing obtained NMR spectra to published data.20−22 The compounds all vary in their decoration of a common modified isoprene−histamine− tryptamine scaffold. Through NMR and HRMS analysis, it became apparent that 1, 2, 4, and 5 were converted into their corresponding artifacts (m1, m2, m4, and m5) when dissolved Received: August 21, 2017 Published: December 8, 2017 3276

DOI: 10.1021/acs.jnatprod.7b00703 J. Nat. Prod. 2017, 80, 3276−3283

Journal of Natural Products

Article

columns with C18, fluoro-phenyl, and phenylhexyl stationary phases. This led to the isolation of five related compounds, the new securamines H−J (1−3) and the two previously reported compounds securamines C (4) and E (5). The structures of the latter were confirmed by comparison of obtained HRMS data, as well as hydrogen and carbon chemical shifts, with published values (4: δ[13C] mean error: 0.24 ppm, 5: δ[13C] mean error: 0.26 ppm).20−22 The observed ROESY patterns were consistent with the energy-minimized structures of the reported relative configurations of 4 and 5. Securamine H (1) was isolated as a light yellow powder. The molecular formula was calculated to be C20H16Br3ClN4O2 by HRESIMS, suggesting the presence of 13 degrees of unsaturation. The 1H and 13C NMR data of 1 (Table 1) resembled those reported for 4 and 5, and 2D NMR analysis (COSY, HMBC, and HSQC, key correlations in Figure 1) indicated that 1 possessed the same hexacyclic moieties as 4 and 5. Based on the calculated elemental composition, the assumption was made that the hexacyclic moiety of 1 was substituted with three bromines. ROESY correlations between H-20 (δH 3.72) and H-11a/H-11b (δH 2.69−2.53) revealed the relative orientations of the C, D, E, and F rings. A ROESY correlation between H3-23 (δH 1.36) and H-15 (δH 7.13) determined the spatial orientation of C-23 (δC 17.2). Furthermore, through a ROESY correlation between H-2 (δH 7.52) and H-10 (δH 4.69), the spatial orientation of the latter was confirmed. Structural differences between 1 and 4 were detected by the lack of HSQC correlations at C-17 (δC 120.1) and C-18 (δC 123.8). This, and the slightly shielded C-17 (δC 120.1) and C-18 (δC 123.8) shift values compared to 4, revealed the location of the two additional bromine atoms found in 1 compared to 4. C-17 (δC 120.1) could readily be assigned by a characteristic strong 3JCH to H-15 (δH 7.13), C-18 (δC 123.8) was assigned by a weak HMBC correlation with H20, while C-16 was assigned by a weak HMBC correlation with H-15. The possible alternative placement of the aromatic proton at C-18 was ruled out by a weak ROESY correlation from H-15 (δH 7.13) to H3-23 (δH 1.36). Consequently, the structure of 1 was assigned. Securamine I (2) was isolated as a white powder. The molecular formula of 2 was the same as that of 5, C20H17Br2ClN4O2, as deduced from HRESIMS data. Analysis of 1D and 2D NMR data recorded for 2 (Table 1) revealed that compound 2 had the same general planar structure and the same relative configuration of the C, D, E, and F rings as 1. The bromination pattern of ring A was determined by the two aromatic protons (H-15 (δH 6.94) and H-18 (δH 7.29)) being devoid of vicinal 3JHH couplings. Individual assignments were primarily based on ROESY correlations between H-15 (δH 6.94) and H3-23 (δH 1.44) and between H-18 (δH 7.29) and H21b (δH 2.91). The C-16 (δC 125.3) and C-17 (δC 116.7) assignments were distinguished by a weak four-bond HMBC correlation from H-20 (δH 3.69) to C-17 (δC 116.7) as well as consistency with chemical shift predictions. Thus, the structure of 2 was elucidated. Securamine J (3) was obtained as a light yellow wax, and its molecular formula was calculated to be C21H20Br3ClN4O4 based on HRESIMS (12 degrees of unsaturation). Compared to 1 this corresponds to one additional carbon, four additional hydrogens, and two additional oxygens, indicating that the securamine skeleton of 3 was decorated with a methoxy and a hydroxy group. The 1H and 13C NMR data of 3 (Table 1) showed that the securamine skeleton of 3 has the same

in MeOH. When dissolved in a non-nucleophilic solvent, m1, m2, m4, and m5 reverted back to their parent compounds. We also found that 3 is a product of chemical modification of a native variant. Although the structures of several securamines have been published, no bioactivity data are available for these compounds. Compounds 1−5 were evaluated for in vitro cytotoxic activity against the human malignant cell lines A2058 (melanoma), HT-29 (colon adenocarcinoma), and MCF-7 (breast adenocarcinoma). Nonmalignant lung fibroblasts (MRC-5) were used to assess the general toxicity of the compounds. Furthermore, the effect of varying exposure time of 1 to the above-mentioned cell lines was evaluated. Herein, we describe details of the isolation, structure elucidation, and bioactivity assessment of 1−5, in addition to their chemical instability when exposed to MeOH.



RESULTS AND DISCUSSION Specimens of S. securif rons (336.91 g) were collected off the coast of Hjelmsøya, Norway, in 2014. After freeze-drying and extraction, an aliquot of the organic extract was fractioned by RP flash chromatography, and the eight fractions were assayed for their ability to inhibit cell viability using the human melanoma cell line A2058. Fractions 4 and 5 (eluting at 75% and 100% MeOH, respectively) gave 2% and 5% remaining cell survival, respectively. The remaining six fractions did not affect cell survival. Based on the observed activity and the knowledge regarding bryozoan secondary metabolite production, the active fractions were prioritized for further chemical investigation. UHPLC-HRMS analysis of these fractions revealed a rich presence of halogenated secondary metabolites, and the calculated elemental compositions of the compounds indicated that they were related. After a first round of collection of the compounds from the organic extract by mass-guided preparative HPLC using a C18 column, each compound was carefully isolated by an extensive series of isolation steps, using 3277

DOI: 10.1021/acs.jnatprod.7b00703 J. Nat. Prod. 2017, 80, 3276−3283

Journal of Natural Products

Article

Table 1. 1H and 13C NMR Assignments for Securamines H (1), I (2), and J (3) (1H 600 MHz, 13C 150 MHz, CDCl3) securamine H (1) position 2 3 4 6 7 8 9 10 11a 11b 12 14 15 16 17 18 19 20 21a 21b 22 23 24 3-OH 2-OMe a

δC, type 136.0, 101.8, 187.1, 166.8,

CH CH C C

δH (J in Hz) 7.52, d (10.4) 5.97, dd (10.4, 0.9)

securamine I (2) δC, type 136.0, 101.8, 187.4, 167.1,

δH (J in Hz)

CH CH C C

7.46, d (10.5) 5.95−5.86, m

7.04, s 85.6, 44.3, 59.0, 42.2,

C C CH CH2

88.0, C 146.7, C 115.1, CH 126.2, C 120.1, C 123.8, C 129.4, C 48.3, CH 32.9, CH2 170.0, C 17.2, CH3 21.2, CH3

4.69, dd (12.7, 4.5) 2.69−2.53, ma 2.69−2.53, ma

7.13, s

3.72, d (7.9) 3.64, d (19.0) 3.04, dd (19.0, 7.9) 1.36, s 1.06, s

securamine J (3) δC, type

δH (J in Hz)

84.0, CH 69.8, CH 189.8, C 164.0, C

5.56, d (4.8) 5.26, dd (5.2, 5.2)

7.25, s 85.8, 44.1, 59.3, 41.9,

C C CH CH2

4.69, dd (12.4, 4.8) 2.68−2.58, ma 2.68−2.58, ma

89.2, C 146.0, C 116.1, CH 125.3, C 116.7, C 129.4, CH 130.6, C 45.1, CH 34.3, CH2

6.94, s

7.29, d (1.2) 3.69, d (7.0) 3.06, dd (18.2, 7.2) 2.91, d (18.2)

170.0, C 17.3, CH3 21.3, CH3

1.44, s 1.08, s

6.43, brd 87.5, 44.5, 59.0, 42.4,

C C CH CH2

86.4, C 148.3, C 115.1, CH 126.1, C 118.1, C 123.3, C 129.6, C 53.7, CH 33.3, CH2 178.4, C 17.2, CH3 20.4, CH3 57.7, CH3

5.89, ddd (12.9, 5.0, 1.6) 2.61, dd (13.4, 5.0) 2.42, t (13.1)

6.84, s

3.69, d (7.4) 3.01, dd (18.4, 7.5) 3.59, d (18.4) 1.28, 1.05, 3.99, 2.65,

s s brd s

Signals are overlapping.

spectrum of 3, correlations between methoxy protons (H-2OMe, δH 2.65) and C-2 (δC 84.0) and between a broad hydroxy proton peak (H-3-OH, δH 3.99) and C-2 (δC 84.0) and C-4 (δC 189.8) showed that the C-2 to C-3 double bond found in 1 and 2 had been saturated with a methoxy group on C-2 and a hydroxy group on C-3. The relative orientation of the methoxy and the hydroxy groups is trans, as evidenced by the presence of simultaneous H-3 (δH 5.26) to H3-2-OMe (δH 2.65) and H-2 (δH 5.56) ROEs, as well as H-3-OH (δH 3.99) to H-10 (δH 5.89) and H3-24 (δH 1.05). Antiproliferative Activities of 1−5. Even though the structures of several securamines are known,20,21 their bioactivities have not been previously reported. As fractions from the organic extract of S. securif rons were found to be active against the human melanoma cell line A2058, compounds 1−5 were evaluated for their in vitro cytotoxicity against three cancer cell lines (A2058 melanoma, HT-29 colon adenocarcinoma, and MCF-7 breast adenocarcinoma) and one nonmalignant cell line (MRC-5 lung fibroblasts). The results of the cytotoxicity assessment of the tested compounds after 72 h of incubation are given in Table 2. Among the evaluated compounds, 1, 2, 4, and 5 were found to affect cell viability, and compounds 1, 2, and 5 were the most potent, with IC50 values ranging from 1.4 ± 0.1 to 10 ± 1 μM against the malignant cell lines. Compound 3 had no significant activity at the highest test concentration (50 μM). The compounds affecting cell viability (1, 2, 4, and 5) possessed a C-2 to C-3 double bond, whereas in the inactive compound (3) the C-2 to C-3 bond was saturated. Comparing the observed levels of activity, taking into account the structural similarities and differences of compounds 1−5, led us to conclude that a methoxylation of C-2, and the associated

Figure 1. Selected 2D NMR correlations obtained for securamine H (1) and a 3D model showing important NOE contacts and stipulated distances (Å).

halogenation pattern, and 2D NMR analysis confirmed the relative configuration of the C, D, E, and F rings to be the same as 1. In the 13C NMR spectrum of 3, the carbon chemical shifts for C-2 (δC 84.0) and C-3 (δC 69.8) vary from that of 1 (C-2 δC 136.0 and C-3 δC 101.8), indicating that 3 differs from 1 with respect to the substituents attached in these positions. The structure was assigned by HMBC analysis. In the HMBC 3278

DOI: 10.1021/acs.jnatprod.7b00703 J. Nat. Prod. 2017, 80, 3276−3283

Journal of Natural Products

Article

lowest observed IC50 value of 1.4 ± 0.1 μM after 74 h of exposure. A similar pattern was observed for the HT-29, MCF7, and MRC-5 cell lines. (HT-29 and MCF-7, Figure S31). The nonmalignant MRC-5 cell line was, however, significantly less affected after 24 h of exposure (IC50 value >10 μM). Methanol Adducts of 1, 2, 4, and 5. During isolation, the compounds were exposed to a mixture of CH3CN and H2O, both with 0.1% formic acid. The isolated samples were dried, dissolved in MeOH, transferred to a small glass tube, dried, weighed, and dissolved in CDCl3 for NMR analysis. A minor component, m1, m2, m4, and m5, was observed in the initial NMR spectra of 1, 2, 4, and 5, respectively. The amount of the minor components decreased over a few hours, while the resonances belonging to the major components as well as a MeOH impurity increased correspondingly in intensity. It thus appeared that the minor components were unstable in weak acid (acid isolated with the compound from the isolation solvents). For 1, the observed reduction of the minor component, m1, was significantly slower than for the minor components of 2, 4, and 5. The initial NMR spectrum (CDCl3 used as solvent) indicated the presence of ∼20% of m1, and after 15 h ∼10% remained (Figure 3), allowing a full NMR analysis of m1 before it disappeared. The most significant difference was the strong shielding of C-2 from δC 136.0 to δC 77.6 and of C-4 from δC 187.1 to δC 140.8, together with the appearance of a new methoxy resonance at δH 3.25, δC 51.1 and a new NH at δH 7.9. The methoxy could be attached to the C-2 via an HMBC correlation, which in turn was firmly anchored in the carbon frame with HMBC correlations from C-3, C-4, C12, and C-22. Thus, m1 had a methoxy group at C-2, was protonated at N-5, and had a double bond between C-3 and C4 instead of between C-2 and C-3 (Figure 4). The relative orientation of the methoxy group was determined to be pointing up based on the comparison between the observed ROE pattern (most notably H-2 to H-10) and energy minimized 3D models of the two possible configurations. This observation was consistent also with the proton spectra at t = 0 for 2, 4, and 5. On the basis of these observations, we hypothesized that native securamines 1, 2, 4, and 5 reacted with MeOH under acidic conditions during sample handling between isolation and NMR analysis to provide the methoxylated securamine artifacts m1, m2, m4, and m5 (Figure 4). However, the structural changes were reversible,

Table 2. Cytotoxicity Data of Compounds 1−5 toward Cancer Cell Lines A2058, HT-29, and MCF-7 and the Normal Cell Line MRC-5 after a 72 h Exposure Time IC50a (μM) ± SD compound

A2058

HT-29

MCF-7

MRC-5

1 2 3 4 5

1.4 ± 0.1 2.7 ± 0.3 >50 20 ± 1 6.7 ± 0.3

1.9 ± 0.1 2.5 ± 0.2 >50 21 ± 2 10 ± 1

2.1 ± 0.1 2.4 ± 0.2 >50 23 ± 1 8.5 ± 0.4

2.7 ± 0.1 5.3 ± 1.1 >50 30 ± 2 9.6 ± 0.3

a

IC50 values refer to the concentration (μM) causing 50% inhibition of cell survival and were calculated from regression using eight different concentration points. Values are means ± SD based on three independent determinations.

saturation of the C-2 to C-3 double bond, was detrimental for the bioactivity. One can speculate that the activities of compounds 1, 2, 4, and 5 are related to their Michael acceptor reactivity, as seen in many biologically active compounds,23 while compound 3 cannot act as a Michael acceptor. In addition, it appeared that two or more bromines on the aromatic ring were beneficial. As the securamines are bulky compounds, it is probable that their cytotoxic activity is influenced by a combination of factors, including the spatial orientation of their functional groups and their overall threedimensional structures, in addition to being correlated with the presence or absence of specific substituents. It is also worth noting that none of the active securamines showed significant differences in their activities against the malignant cell lines and the nonmalignant MRC-5. To evaluate the cell death kinetics related to the response of the four cell lines (A2058, HT-29, MCF-7, and MRC-5), each was treated with different concentrations of 1 for 4, 24, 48, and 72 h. This revealed that the cytotoxic activity of 1 was timedependent, yielding decreasing IC50 values with increasing exposure times (Figure 2). Compound 1 had no apparent activity against A2058 after 4 h of exposure, indicating that the observed activity is not due to an unspecific and rapid interaction with the cell membrane causing necrotic cell death by lysis. Activity was seen after 24 h of exposure, when the IC50 value was determined to be 2.8 ± 0.6 μM. A smaller decrease in IC50 value was observed after 48 h of exposure, 1.8 ± 0.2 μM, before dropping down to the

Figure 2. Cytotoxicity of securamine H (1) against malignant cell line A2058 and nonmalignant cell line MRC-5 was time-dependent. Cells were incubated with different concentrations of 1 for 4, 24, 48, and 72 h. Data were obtained from three independent experiments and represented as mean ± standard deviation. 3279

DOI: 10.1021/acs.jnatprod.7b00703 J. Nat. Prod. 2017, 80, 3276−3283

Journal of Natural Products

Article

Figure 3. 1H NMR spectra of 1 acquired immediately (lower panel) and 15 h (upper panel) after dissolving the sample in CDCl3. In the t = 0 spectrum, a minor component, m1, is clearly visible (signals assigned in red). At t = 15 h the m1 signals are clearly reduced in intensity compared to the signals of 1 (assigned in blue), while the MeOH signal is increased. For 2, 4, and 5 there were little or no remains of the minor compound after 15 h.

Figure 4. Structures of the native securamines 1, 2, 4, and 5 produced by S. securif rons and their corresponding methoxylated artifacts m1, m2, m4, and m5.

leading to the regeneration of the native securamine under MeOH-free conditions during the NMR experiments. To further explore the mechanism, m4, dissolved in methanol-d3, was diluted 1:100 into ordinary MeOH and analyzed using HRMS. The mass spectra for m4 directly after injection and 10 min later are shown in Figure 5. The H:D ratio changed from 1:1 at t = 0 min to 2:1 at t = 10 min, showing that a CH3O group is replacing the CD3O group slowly over time. As the steric hindrance at C-2 rules out an SN2 substitution by MeOH on m4, this strongly indicates that the methoxylated compound m4 exists in a slow equilibrium with the unsaturated form plus bulk MeOH from the solvent. These results supported our hypothesis that a huge excess of MeOH (when the compound is dissolved in MeOH) will result in the transformation of 1, 2, 4, and 5 into m1, m2, m4, and m5. In CDCl3, without added MeOH, the adducts m1, m2, m4, and m5 reversed back to the parent compound. A mechanism for the formation and degradation of MeOH adducts of the

Figure 5. HRMS analysis of methoxylated securamine C (m4) dissolved in methanol-d3 after being diluted 1:100 in 1H-MeOH. Over 10 min the incorporation of 2H-methoxy drops from 1:1 to 1:2 in favor of 1H-methoxy from the bulk solvent.

securamines is proposed in Figure 6, exemplified with the structures of m4 and 4. For the degradation of the MeOH adduct m4, a proton is abstracted from N-5, leading to the formation of a conjugated double bond system between N-5, C4, C-3, and C-2 upon loss of the methoxy group at C-2 after protonation. Formation of MeOH adducts is the opposite 3280

DOI: 10.1021/acs.jnatprod.7b00703 J. Nat. Prod. 2017, 80, 3276−3283

Journal of Natural Products

Article

their corresponding macrocyclic indoles securines A and B when exposed to DMSO. The rest of the previously reported securamines, C−J, all have hexacyclic structures and have not been reported to degrade into securines in DMSO. However, as the herein reported compounds were dissolved in DMSO for bioassay purposes, the samples were analyzed using UHPLCHRMS before and after the bioactivity testing. We could not detect formation of securines in our reported samples. In addition, the analysis confirmed that the bioassayed samples were in fact the native variants of 1, 2, 4, and 5, as the MeOH adducts not were detected during these analyses.

Figure 6. Proposed mechanism for the degradation of MeOH adducts (in red from left to right) and formation of MeOH adducts (in blue from right to left) of the securamines. The mechanism is proposed on the basis of NMR analysis of the securamines and HRMS analysis of methoxylated securamine C (m4) dissolved in deuterated, then normal MeOH and is exemplified with the structures of m4 and securamine C (4).



CONCLUSIONS In conclusion, three new securamine analogues, securamines H−J (1−3), were isolated from the organic extract of S. securif rons along with the two previously reported securamines, C (4) and E (5). Their structures were elucidated by 1D and 2D NMR analyses. Through the structure elucidation process, we revealed that 1, 2, 4, and 5 were transformed into their corresponding MeOH adducts m1, m2, m4, and m5, respectively, when dissolved in MeOH. Dissolving the compounds in non-nucleophilic solvents reverted m1, m2, m4, and m5 back to their parent compound. In addition, our results indicated that 3 was a MeOH adduct of a native precursor. Compounds 1, 2, 4, and 5 inhibited the viability of four cell lines, with IC50 values as low as 1.4 ± 0.1 μM. Furthermore, the effect of 1 was found to be time-dependent. The mode of action behind the observed activity remains to be investigated. These results further highlight the capacity of Arctic marine bryozoans as producers of bioactive secondary metabolites with potential as lead structures for the development of new pharmaceuticals.

process, involving protonation of N-5, rearrangement of the double bond system, and addition of MeOH to C-2. Formation of Securamine J (3). For 3, no structural changes were observed during NMR analysis. It can however be hypothesized that 3 is the product of a related chemical instability in the presence of MeOH as described for compounds 1, 2, 4, and 5. Compound 3 was also, as the other securamines, exposed to a mixture of CH3CN and H2O (both with 0.1% formic acid) during isolation. MS data of pre-3 recorded during this isolation process indicated a mass of 631.85 and a chemical formula of C20H16Br3ClN4O3, which would correspond to a securamine with three attached bromines, one oxygen, and one hydrogen. The isolated sample was dried, dissolved in MeOH, transferred to a small glass tube, dried, weighed, and dissolved in CDCl3 for NMR analysis. New MS analysis after the acquisition of NMR data instead indicated a mass of 663.87 and a chemical formula of C21H20Br3ClN4O4, which corresponds to the addition of MeOH compared to the isolated compound. On the basis of the determined structure of 3 and analogy with the other securamines’ sensitivity to MeOH, we propose structures n3a and n3b (Figure 7) as possible



EXPERIMENTAL SECTION

General Experimental Procedures. Optical rotations were measured on an AA-10R automatic polarimeter (Optical Activity LTD) in dichloromethane. NMR spectra were acquired in chloroformd1 on a Bruker Avance III HD spectrometer operating at 600 MHz for protons, equipped with an inverse TCI cryo-probe enhanced for 1H, 13 C, and 2H. All NMR spectra were acquired at 298 K, in 3 mm solvent matched Shigemi tubes using standard pulse programs for proton, carbon, HSQC, HMBC, DQCOSY, and ROESY experiments with gradient selection and adiabatic versions where applicable. 1H/13C chemical shifts were referenced to the residual solvent peak (chloroform-d1: δH = 7.26, 13C δ = 77.16). HRESIMS were performed using a Waters LCT Premier time-of-flight MS with an Acquity UPLC (Waters), using MS grade solvents. Compound isolation was performed using a preparative-HPLC-MS system consisting of a 600 HPLC pump, a 2996 photodiode array UV detector, a 3100 mass detector, and a 2767 sample manager (Waters). The following columns were used: SunFire Prep C18 (5 μM, 10 × 250 mm), Atlantis Prep C18 (10 μM, 10 × 250 mm), Xselect CSH Prep Fluoro-Phenyl (5 μM, 10 × 250 mm), and Xselect CSH Phenyl-Hexyl Prep (5 μM, 10 × 250 mm) (all from Waters). Flash chromatography was carried out using a Biotage HPFC SP4 system equipped with a Biotage SNAP column filled with 8 g of Diaion HP-20SS (Supelco Analytical). All solvents used for extraction and fractionation were of HPLC grade, and Milli-Q H2O was used. Biological Material. Specimens of the bryozoan Securif lustra securif rons (class Gymnolaemata, order Cheilostomatida, family Flustridae) were collected as previously described.14 The organism was identified by Robert A. Johansen, at the Norwegian national biobank (Marbank), and a voucher specimen (reference number: M14064) was deposited in Marbank, Tromsø, Norway. Extraction and Prefractionation. The 1:1 CH2Cl2/MeOH extract of S. securif rons was prepared as previously described.14 An

Figure 7. Proposed structures of pre-3 (n3a and n3b) and 3D view of n3b illustrating the MeOH attack at C-2 in the formation of securamine J (3).

structures of pre-3. Structure n3b contains an epoxide, which could be obtained from enzymatic epoxidation24 of the C-2 to C-3 double bond of securamine H (1). A 3D model of n3b shows that the epoxide is correctly oriented in space to allow for MeOH attack at C-2 from the least hindered face of the D ring to form 3. UHPLC-MS analysis of the organic extract shows the presence of 3 in small amounts. It thus appears that the conditions, to which the compound was exposed during the extraction process, are less favorable for the formation of 3 from pre-3, compared to the condition pre-3 was exposed to during sample handling between the isolation step and the NMR analysis. As previously described,20 the pentacyclic compounds securamines A and B are known to exist in equilibrium with 3281

DOI: 10.1021/acs.jnatprod.7b00703 J. Nat. Prod. 2017, 80, 3276−3283

Journal of Natural Products

Article

aliquot of the extract (2.001 g) was fractionated using RP flash chromatography. The following stepwise elution method with a flow rate of 12 mL/min was used: MeOH/H2O (5:95, 25:75, 50:50, 75:25; 6 min per step), MeOH (100% over 12 min), MeOH/acetone (50:50 over 6 min), and acetone (100% over 6 min). Each fraction was collected for 6 min, resulting in eight fractions that were dried under reduced pressure. The dry fractions were dissolved in DMSO to a concentration of 40 mg/mL and stored in cryo-tubes at −23 °C until used for bioactivity testing. Cytotoxicity Screening (MTS Assay). The flash fractions from the organic extract of S. securif rons were screened for bioactivity at a concentration of 50 μg/mL against the human melanoma cell line A2058. A2058 cells were cultured and assayed in Dulbecco’s modified Eagle’s medium (D-MEM, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS, Merck) and 5 mL of L-alanyl-Lglutamine (200 mM, Merck) and incubated in 5% CO2 at 37 °C. Briefly, the cells were seeded in 96-well microtiter plates (Thermo Fisher Scientific) at 2000 cells/well and incubated for 24 h to allow the cells to adhere before the cell media was replaced and the fractions were added. The plates were left to incubate for 72 h. Subsequently, 10 μL of MTS solution (Cell Titer 96 Aqueous One Solution Reagent, Promega) was added to each well, and the cells were incubated for 1 h at 37 °C. The absorbance was measured at 485 nm using a DTX multimode detector (Beckman Coulter). The negative control was defined as cells assayed with their respective cell media with 1% DMSO (1% was the highest DMSO concentration used in wells with fraction/compound), and positive control as cells assayed with 0.5% Triton X-100 (Sigma-Aldrich). Cell viability was calculated as follows: Cell survival (%) = (absorbance treated wells − absorbance positive control)/(absorbance negative control − absorbance positive control) × 100. Isolation. Compounds 1−5 were isolated from an aliquot of the organic extract of S. securif rons (2.01 g). The extract aliquot was partitioned between 150 mL of hexane and 150 mL of 90% MeOH three times to remove lipophilic sample components that might harm the HPLC column. The pooled MeOH fractions were dried under reduced pressure and redissolved in 7 mL of CH3CN. Aliquots of this sample were repeatedly injected onto a SunFire Prep C18 column and eluted with a mobile phase consisting of solvent A (H2O with 0.1% formic acid (FA)) and B (CH3CN with 0.1% FA), delivered in a gradient mode at 6 mL/min, starting from B at 10% to 70% over 20 min and then B at 100% for 3 min. The securamines were collected with retention times of 1: 21.1, 2: 18.3, 3: 20.5, 4: 16.3, and 5: 19.2 min, respectively. Final separation of the compounds from sample impurities was achieved by utilizing columns with C18, fluoro-phenyl, and phenyl-hexyl stationary phases employing isocratic elution with H2O (0.1% FA)/CH3CN (0.1% FA) at 6 mL/min, 60:40 over 10 min. In the end, five securamines were isolated in amounts that enabled structure determination by NMR analysis: 1 (9.8 mg), 2 (8.5 mg), 3 (1.1 mg), 4 (2.8 mg), and 5 (1.21 mg). HRMS Degradation Analysis. Compound m4 was dissolved in methanol-d3 (1 mg), diluted 1:100 into ordinary MeOH, and analyzed using HRMS. The sample was injected immediately after dissolution and again after 10 min. Securamine H (1): light yellow powder; [α]20D −270 (c 0.3 CH2Cl2); UV (CH3CN/H2O) λmax 227; 1H and 13C NMR data, Table 1; HRESIMS m/z 616.8572 [M + H]+ (calcd for C20H17Br3ClN4O2, 616.8584). Securamine I (2): white powder; [α]20D −443 (c 0.3 CH2Cl2); UV (CH3CN/H2O) λmax 252; 1H and 13C NMR data, Table 1; HRESIMS m/z 538.9457 [M + H]+ (calcd for C20H18Br2ClN4O2, 538.9480). Securamine J (3): light yellow wax, UV (CH3CN/H2O) λmax 227; 1 H and 13C NMR data, Table 1; HRESIMS m/z 664.8774 [M + H]+ (calcd for C21H21Br3ClN4O4, 664.8796). Cytotoxicity of Compounds 1−5. The cytotoxicities of compounds 1−5 were determined against A2058, MCF-7 (breast adenocarcinoma), and HT-29 (colon adenocarcinoma) human cancer cell lines and the nonmalignant human lung fibroblasts MRC-5 using the MTS assay described above. The HT-29 cell line was cultured and assayed in Roswell Park Memorial Institute (RPMI, Merck KGaA).

MCF-7 and MRC-5 cell lines were cultured and assayed in Earle’s minimal essential medium (E-MEM, Merck). D-MEM and RPMI were supplied with 10% FBS (Merck) and 5 mL of L-alanyl-L-glutamine (200 mM, Merck). E-MEM was supplementd with 10% FBS, 5 mL of L-alanyl-L-glutamine, 5 mL of nonessential amino acids (100×, Merck), 5 mL of sodium pyruvate (100 mM, Merck), and 5 mL of sodium bicarbonate (7.5%, Merck). All cells were incubated in 5% CO2 at 37 °C. Briefly, the cells were seeded in 96-well microtiter plates at 2000 cells/well for the malignant cell lines and 4000 cells/well for the nonmalignant cell line. After 24 h, each cancer cell line was treated with various concentrations of compounds 1−5 for 72 h. The results were measured and calculated as described above. The half-maximal inhibitory concertation (IC50) was calculated using Graph Pad Prism software. The time dependency of the activity of 1 was further evaluated against the four cell lines. Adhered cells were added to ranging concentrations of 1, 80, 60, and 36 h after seeding, to evaluate the effect of 1 after 4, 24, and 48 h exposure times. Results were measured and evaluated as described above. All viability assays were performed in triplicate in three independent experiments. Statistical Analysis. Data analyses were performed with GraphPad Prism 5.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jnatprod.7b00703. Copies of 1D and 2D NMR data for securamines H−J (1−3), C, and E (4 and 5) in CDCl3; experimental 13C NMR shift values compared to reference shift values of securamines C (4) and E (5); dose- and time-dependent activity of securamine H (1) against cancer cell lines HT29 and MCF-7 (PDF)



AUTHOR INFORMATION

Corresponding Author

*Tel: +47 77 64 92 72. Fax: +47 77 64 60 20. E-mail: kine.o. [email protected]. ORCID

Kine Ø. Hansen: 0000-0002-9023-1958 Annette Bayer: 0000-0003-3481-200X Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We are grateful to Marbank for collecting the specimens of S. securif rons and to R. Johansen for taxonomic identification of the bryozoan.



REFERENCES

(1) Fortunato, H. Reg. Stud. Mar. Sci. 2015, 2, 32−44. (2) McKinney, F. K.; Jackson, J. B. C. Bryozoan Evolution; University of Chicago Press, 1991. (3) Kutyumov, V. A.; Maltseva, A. L.; Kotenko, O. N.; Ostrovsky, A. N. Cell Tissue Biol. 2016, 10, 152−159. (4) Blunt, J. W.; Copp, B. R.; Keyzers, R. A.; Munro, M. H. G.; Prinsep, M. R. Nat. Prod. Rep. 2016, 33, 382−431. (5) Sharp, J. H.; Winson, M. K.; Porter, J. S. Nat. Prod. Rep. 2007, 24, 659−73. (6) Milanowski, D. J.; Gustafson, K. R.; Kelley, J. A.; McMahon, J. B. J. Nat. Prod. 2004, 67, 70−73. (7) Carroll, A. R.; Duffy, S.; Sykes, M.; Avery, V. M. Org. Biomol. Chem. 2011, 9, 604−609. (8) Farokhzad, O. C.; Mun, E. C.; Sicklick, J. K.; Smith, J. A.; Matthews, J. B. Surgery 1998, 124, 380−387.

3282

DOI: 10.1021/acs.jnatprod.7b00703 J. Nat. Prod. 2017, 80, 3276−3283

Journal of Natural Products

Article

(9) Mehla, R.; Bivalkar-Mehla, S.; Zhang, R.; Handy, I.; Albrecht, H.; Giri, S.; Nagarkatti, P.; Nagarkatti, M.; Chauhan, A. PLoS One 2010, 5, e1116010.1371/journal.pone.0011160. (10) Kuzirian, A. M.; Epstein, H. T.; Gagliardi, C. J.; Nelson, T. J.; Sakakibara, M.; Taylor, C.; Scioletti, A. B.; Alkon, D. L. Biol. Bull. 2006, 210, 201−214. (11) Figuerola, B.; Sala-Comorera, L.; Angulo-Preckler, C.; Vázquez, J.; Jesús Montes, M.; García-Aljaro, C.; Mercadé, E.; Blanch, A. R.; Avila, C. Mar. Environ. Res. 2014, 101, 52−59. (12) Lopanik, B. L.; Targett, M. T.; Lindquist, N. Mar. Ecol.: Prog. Ser. 2006, 327, 183−191. (13) Piel, J. Nat. Prod. Rep. 2009, 26, 338−362. (14) Michael, P.; Hansen, K.; Isaksson, J.; Andersen, J.; Hansen, E. Molecules 2017, 22, 1236. (15) Peters, L.; Konig, G. M.; Terlau, H.; Wright, A. D. J. Nat. Prod. 2002, 65, 1633−7. (16) Chevolot, L.; Chevolot, A. M.; Gajhede, M.; Larsen, C.; Anthoni, U.; Christophersen, C. J. Am. Chem. Soc. 1985, 107, 4542− 4543. (17) Maltseva, A. L.; Kotenko, O. N.; Shabalin, K. A.; Shavarda, A. L.; Winson, M. K.; Ostrovsky, A. N. Studi Trent. Sci. Nat. 2014, 94, 163− 168. (18) Maltseva, A. L.; Kotenko, O. N.; Kutyumov, V. A.; Matvienko, D. A.; Shavarda, A. L.; Winson, M. K.; Ostrovsky, A. N. Nat. Prod. Res. 2017, 31, 1840−1848. (19) Svenson, J. Phytochem. Rev. 2013, 12, 567−578. (20) Rahbaek, L.; Anthoni, U.; Christophersen, C.; Nielsen, P. H.; Petersen, B. O. J. Org. Chem. 1996, 61, 887−889. (21) Rahbæk, L.; Christophersen, C. J. Nat. Prod. 1997, 60, 175−177. (22) In an article from 1997 (ref 21), carbon signals at 120.6 and 123.8 ppm were listed for C-16 and C-18, respectively, but the shifts were noted as being interchangeable. HMBC analysis of these two carbon signals clearly shows that the two shifts should be interchanged: C-16, 124.0 ppm, and C-18, 120.8 ppm. (23) Martin, H. J. Mini-Rev. Med. Chem. 2012, 12, 1330−1344. (24) May, S. W. Enzyme Microb. Technol. 1979, 1, 15−22.

3283

DOI: 10.1021/acs.jnatprod.7b00703 J. Nat. Prod. 2017, 80, 3276−3283