Pinnatifidenyne-Derived Ethynyl Oxirane Acetogenins from Laurencia viridis

Red algae of Laurencia continue to provide wide structural diversity and complexity of halogenated C15 acetogenin medium-ring ethers. Here, we described the isolation of three new C15 acetogenins (3–5), and one truncated derivative (6) from Laurencia viridis collected on the Canary Islands. These compounds are interesting variations on the pinnatifidenyne structure that included the first examples of ethynyl oxirane derivatives (3–4). The structures were elucidated by extensive study of NMR (Nuclear Magnetic Resonance) data, J-based configuration analysis and DFT (Density Functional Theory) calculations. Their antiproliferative activity against six human solid tumor cell lines was evaluated.


Introduction
Medium-ring haloethers of Laurencia, a significant subset of biologically active marine natural products [1,2], continue to challenge innovative efforts as attractive targets for the synthesis of stereochemically rich medium-sized oxacyclic compounds [3,4] or to explore biogenetic hypothesis [5,6].
Z-Pinnatifidenyne and E-pinnatifidenyne (1-2) are representative and were reported in 1982 after isolation from the red algae Laurencia pinnatifida collected at Canary Islands [7]. Their structures were established by espectroscopic methods, X-ray diffraction analysis and chemical correlations, although, in 1991, the absolute configurations were reassigned based on a later X-ray analysis [8]. Recent synthetic approaches and biosynthetic studies of these eight-membered compounds have been reported [9,10].
As part of our continuing interest on the chemistry of the genus Laurencia [11][12][13][14] and during the course of our anticancer drug discovery program, we report the isolation of three new C 15 acetogenins and one truncated derivative from Laurencia viridis (Figure 1). The structures were elucidated based on (9R,10S)-Epoxy-Z-pinnatifidenyne (5) Pinnatifidehyde (6)

Results and Discussion
(3R,4S)-Epoxy-pinnatifidenyne (3)  According to the molecular formula, the 13 C NMR data (Table 2) along with the analysis of the edited HSQC (Heteronuclear Single Quantum Correlation) spectrum confirmed the presence of two olefinic carbons (δ C 131.0 and 128.9), six heteroatom-bearing methynes (δ C 83.1, 77.5, 65.8, 61.1, 55.3 and 45.6), four methylenes (δ C 34.5, 34.3, 30.0 and 27.1) and one methyl (δ C 12.8). Careful examination of the homonuclear and heteronuclear NMR correlations exhibited in the COSY (Correlation SpectroscopY) and HSQC spectra allowed to observe a single 1 H-1 H spin system, C-3→C-15, containing a double bond between C-9 and C-10 together with heteroatoms located on carbons C-3, C-4, C-6, C-7, C-12 and C-13. Moreover, HMBC (Heteronuclear Multiple Bond Correlation) cross-peaks from H-6 (δ H 4.12) to C-12 (δ C 83.1) established an ether linkage between C-6 and C-12, thus indicating the presence of a 3,4,7,8-tetrahydro-2H-oxocin heterocycle. The presence of a terminal epoxy alkyne was evident by the characteristic 1 H and 13 C signals (δ H /δ C , 3.25, ddd, J = 3.8, 4.0, 8.4 Hz/55.3 (CH-4), 3.52, dd, J = 1.7, 4.0 Hz/45.6 (CH-3), 78.6 (C-2), and 2.39, d, J = 1.7 Hz/74.5 (CH-1)) and by IR (Infrared) absorption at 3023 (it is the strongest signal) and 2134 cm −1 . The relative configuration of 3 was determined by a combination of NOESY (Nuclear Overhauser Spectroscopy) data and J-based configuration approach. 1D-NOE (Nuclear Overhauser Effect) correlations observed between H-6 and H-7/H-12 located all these protons on the same face of the heterocycle (Figure 2), consistent with the orientation previously observed in the pinnatifidenynes. The relationship between the configurations of C-12 and C-13 was established comparing chemical shifts and coupling constants of 3 with those of pinnatifidenynes sharing the same C-15-C-6 fragment [7]. To complete the structural determination, the relative configuration of the epoxide as well as its stereochemical relationship with C-6, the homo-and heteronuclear J couplings were measured [15,16].  (Figure 2c). Finally, the relative configurations at C-6 and C-4 can be conveniently started from the large coupling constant displayed by the protons H-6 and H-5a ( 3 J H6-H5a = 9.7 Hz) and the observed value of 2 J C6-H5b = −1.5 Hz that suggested an anti relationship for H-6 and H-5a (δ H 2.35) and a gauche orientation between H-6 and H-5b (δ H 1.64), confirmed by the dipolar correlation between H-5b and H-7. Similarly, it was established a threo-configuration between H-4 and H-5a explained by the observed values of 3 J H4-H5b = 8.4 Hz, 3 J C3-H5a = 1.2 Hz and 2 J C4-H5a = −2.2 Hz. Further support was obtained from additional long-range heteronuclear coupling constants shown in Figure 3.   Compound 4 has the same molecular formula as 3 together with a close structural relationship. Comparison of their 1 H and 13 C NMR chemical shifts and the analysis of their 2D NMR data allowed us to establish an identical planar structure for both compounds (see Tables 1 and 2 and Experimental Section).
The stereochemical relationships between the different stereogenic centres, including those of epoxide, were performed using the above-described methods. The results of the NMR configurational analysis are shown in Figure 4. The conclusion was that compounds 3 and 4 share the relative configuration at centers C-6, C-7, C-12 and C-13, as well as the cis configuration of the epoxide, while the oxirane rings present an opposite relative configuration. Although a J-based configurational analysis on 3 and 4 was done to determine its relative configuration, a complementary computational study was undertaken. Chemical shift calculations of carbon atoms attached to halogens are not accurate or difficult to calculate due to spin-orbit contributions. However, Kutateladze et al. [17] have recently reported the use of parametric corrections to calculate 13 C chemical shifts by DFT using inexpensive computations at the B3LYP/6-31G(d) level for structure optimization and using the ωB97xD/6-31G(d) level for the chemical shift calculations. Using this methodology, we built all possible stereoisomers for compound 3 followed by the corresponding conformational searches, structure optimizations and chemical shift calculations for each calculated conformer. As a result, it turned out that the 3R,4S,6S,7S,12S,13S isomer showed the lowest RMSD (Root Mean Square Deviation) for compounds 3 and 4 (see Supplementary Materials, Table S1). This result was coincident with our NMR-based Compound 4 has the same molecular formula as 3 together with a close structural relationship. Comparison of their 1 H and 13 C NMR chemical shifts and the analysis of their 2D NMR data allowed us to establish an identical planar structure for both compounds (see Tables 1 and 2 and Experimental Section).
The stereochemical relationships between the different stereogenic centres, including those of epoxide, were performed using the above-described methods. The results of the NMR configurational analysis are shown in Figure 4. The conclusion was that compounds 3 and 4 share the relative configuration at centers C-6, C-7, C-12 and C-13, as well as the cis configuration of the epoxide, while the oxirane rings present an opposite relative configuration.  Compound 4 has the same molecular formula as 3 together with a close structural relationship. Comparison of their 1 H and 13 C NMR chemical shifts and the analysis of their 2D NMR data allowed us to establish an identical planar structure for both compounds (see Tables 1 and 2 and Experimental Section).
The stereochemical relationships between the different stereogenic centres, including those of epoxide, were performed using the above-described methods. The results of the NMR configurational analysis are shown in Figure 4. The conclusion was that compounds 3 and 4 share the relative configuration at centers C-6, C-7, C-12 and C-13, as well as the cis configuration of the epoxide, while the oxirane rings present an opposite relative configuration. Although a J-based configurational analysis on 3 and 4 was done to determine its relative configuration, a complementary computational study was undertaken. Chemical shift calculations of carbon atoms attached to halogens are not accurate or difficult to calculate due to spin-orbit contributions. However, Kutateladze et al. [17] have recently reported the use of parametric corrections to calculate 13 C chemical shifts by DFT using inexpensive computations at the B3LYP/6-31G(d) level for structure optimization and using the ωB97xD/6-31G(d) level for the chemical shift calculations. Using this methodology, we built all possible stereoisomers for compound 3 followed by the corresponding conformational searches, structure optimizations and chemical shift calculations for each calculated conformer. As a result, it turned out that the 3R,4S,6S,7S,12S,13S isomer showed the lowest RMSD (Root Mean Square Deviation) for compounds 3 and 4 (see Supplementary Materials, Table S1). This result was coincident with our NMR-based Although a J-based configurational analysis on 3 and 4 was done to determine its relative configuration, a complementary computational study was undertaken. Chemical shift calculations of carbon atoms attached to halogens are not accurate or difficult to calculate due to spin-orbit contributions. However, Kutateladze et al. [17] have recently reported the use of parametric corrections to calculate 13 C chemical shifts by DFT using inexpensive computations at the B3LYP/6-31G(d) level for structure optimization and using the ωB97xD/6-31G(d) level for the chemical shift calculations. Using this methodology, we built all possible stereoisomers for compound 3 followed by the corresponding conformational searches, structure optimizations and chemical shift calculations for each calculated conformer. As a result, it turned out that the 3R,4S,6S,7S,12S,13S isomer showed the lowest RMSD (Root Mean Square Deviation) for compounds 3 and 4 (see Supplementary Materials, Table S1). This result was coincident with our NMR-based proposal for the relative configuration of all stereogenic centers within the oxirane ring. However, it was clear that it was unable to solve the uncertainty within the oxirane ring. Taking into account that the previously mentioned methodology uses only 13 C data, we decided to use chemical shift calculations of 1 H for a second comparison as it has been shown that the combination of both chemical shifts yields better results [12,13,[18][19][20]. Thus, models of the two possible diastereoisomers, 3S,4R,6S,7S,12S,13S (3a) and 3R,4S,6S,7S,12S,13S (3b), were built and conformational searches on each one, consisting on 5000 steps of a hybrid MCMM (Monte Carlo Multiple Minimum), Low-Mode sampling using the MMFF94 (Merck Molecular Force Field) force field, were completed. Redundant conformers within a 12 kJ/mol energy window of the global minimum found were eliminated using an RMSD cutoff of 1.0 Å. Next, all the resulting structures (seven conformers for 3a and five conformers for 3b) were geometrically optimized using DFT calculations [21] at the B3LYP/6-31G** level of theory with the LACVP basis set in gas phase [22]. NMR shielding constants (σ) were calculated according to the calculated relative Boltzmann populations for each conformer. Finally, NMR chemical shifts were obtained scaling the calculated values by linear regression analysis of experimental and computed data. In this analysis, the carbon nucleus with the attached bromine atom and the acetylenic proton were not included due to their high deviations. Correlation coefficients were almost identical for both stereoisomers (0.9885 for 3a vs. 0.9881 for 3b) using the 13 C NMR data but clearly better for isomer 3b using 1 H-NMR data (0.9867 vs. 0.9696) ( Figure 5 and Supplementary Materials, Table S2). Calculation of the DP4 parameter [23] showed that the 3R,4S,6S,7S,12S,13S diastereoisomer (3b) is the most likely solution, with a probability 99.9% using both 13 C and 1 H data, therefore supporting our NMR-based proposal. Moreover, the same procedure was applied to fit the data of compound 4 to the calculated values for diasteroisomers 3a and 3b ( Figure 6 and Supplementary Materials, Table S2). In this case, it turned out that the experimental data of 4 fitted better with calculated data for the 3S,4R,6S,7S,12S,13S diasteroisomer (3a), confirming that compounds 3 and 4 are the two-possible cis oxirane stereoisomers.  (Tables 2 and 3) compared to reported data for 1 [7] indicated that compound 5 differs to 1 containing an extra disubstituted epoxide situated at C-9-C-10. The relative configuration of 5 was determined by observation of dipolar correlations. Accordingly, the observed NOE enhancements from H-10 to H-9 and H-12, and from H-6 to H-7 and H-12 located all these protons on the same face of the molecule (Supplementary Materials, Figure S23). The above data and the coupling constant for 3 JH9-H10 = 4.0 Hz provide evidence of the cis-orientation. This structure corresponds to a previously reported synthetic epoxide obtained by epoxidation of 1 and published by our group in 1982 [7]. However, the data published at that time and those obtained for our natural product did not correlate well. Therefore, to confirm the structure of 5 as (9R,10S)-epoxy-Z-pinnatifidenyne, we decided to repeat the selective epoxidation of 1. A product identical to 5 in all respects was obtained (see Experimental part).
Pinnatifidehyde 6 is a yellow amorphous solid that shows the molecular formula C12H18BrClO2, evidenced by the presence of three pseudomolecular [M + Na] + ions in the HR-ESI-MS spectrum at m/z 331.0086, 333.0056 and 335.0033 (ratio 74:100:36, calcd. 331.0076, 333.0047, 335.0026). 1 H and 13 C NMR spectra (Tables 2 and 3) were reminiscent of the corresponding partial spectral signals of pinnatifidenynes 1-2 or compounds 3-4. All those compounds share identical oxacyclic and bromopropyl terminal chain (C-6C-15 moiety), whereas the only notable differences were fixed going towards C-4 were the structure appears truncated. Thus, the deshielded signals of a methylene at δH 3.   (Tables 2 and 3) compared to reported data for 1 [7] indicated that compound 5 differs to 1 containing an extra disubstituted epoxide situated at C-9-C-10. The relative configuration of 5 was determined by observation of dipolar correlations. Accordingly, the observed NOE enhancements from H-10 to H-9 and H-12, and from H-6 to H-7 and H-12 located all these protons on the same face of the molecule (Supplementary Materials, Figure S23). The above data and the coupling constant for 3 J H9-H10 = 4.0 Hz provide evidence of the cis-orientation. This structure corresponds to a previously reported synthetic epoxide obtained by epoxidation of 1 and published by our group in 1982 [7]. However, the data published at that time and those obtained for our natural product did not correlate well. Therefore, to confirm the structure of 5 as (9R,10S)-epoxy-Z-pinnatifidenyne, we decided to repeat the selective epoxidation of 1. A product identical to 5 in all respects was obtained (see Experimental part).
Pinnatifidehyde 6 is a yellow amorphous solid that shows the molecular formula C 12 (Tables 2 and 3) were reminiscent of the corresponding partial spectral signals of pinnatifidenynes 1-2 or compounds 3-4. All those compounds share identical oxacyclic and bromopropyl terminal chain (C-6→C-15 moiety), whereas the only notable differences were fixed going towards C-4 were the structure appears truncated. Thus, the deshielded signals of a methylene at δ H 3.10 (br.dd, J = 8.0, 18.7 Hz) and 2.71 (dd, J = 4.0, 18.7 Hz), as well as the methine at δ H 4.46 (ddd, J = 2.9, 4.0, 8.0 Hz), equivalent respectively to H 2 -5 and H-6 in pinnatifidenynes, showed correlations in the HMBC with a carbonyl signal at δ C 200.2 corresponding to one aldehyde group. The structure of compound 6 possesses a truncated C 12 carbon skeleton that may be derived from compounds 1 or 2 based on a probable oxidative process. Finally, analysis of the ROESY experiment confirmed the relative configuration of all stereogenic centers presented in the molecule as equivalent to those observed in the pinnatifidenynes (1, 2). Pinnatifidehyde 6 has been a synthetic intermediate target in the total synthesis of compounds 1 and 2 [10,24]. The spectroscopic data of synthesized intermediate by the Snyder group [10] was in agreement with the spectroscopic data of the natural product 6. As far as we know, pinnatifidehyde (6) is the third example of C 12 acetogenins; the other two, okamuragenin (7) and desepilaurallene (8) (Figure 7), have been isolated from Laurencia okamurai [25,26]. It has to be noted that the new compounds with the ethynyl oxirane unit (3 and 4) could be considered biogenetic precursors of pinnatifidehyde.
Mar. Drugs 2017, 16, 5 8 of 12 The structure of compound 6 possesses a truncated C12 carbon skeleton that may be derived from compounds 1 or 2 based on a probable oxidative process. Finally, analysis of the ROESY experiment confirmed the relative configuration of all stereogenic centers presented in the molecule as equivalent to those observed in the pinnatifidenynes (1, 2). Pinnatifidehyde 6 has been a synthetic intermediate target in the total synthesis of compounds 1 and 2 [10,24]. The spectroscopic data of synthesized intermediate by the Snyder group [10] was in agreement with the spectroscopic data of the natural product 6. As far as we know, pinnatifidehyde (6) is the third example of C12 acetogenins; the other two, okamuragenin (7) and desepilaurallene (8) (Figure 7), have been isolated from Laurencia okamurai [25,26]. It has to be noted that the new compounds with the ethynyl oxirane unit (3 and 4) could be considered biogenetic precursors of pinnatifidehyde. Okamuragenin (7) Desepilaurallene (8) Figure 7. Structures of C12 metabolites isolated from Laurencia okamurai, okamuragenin (7) and desepilaurallene (8).
The antiproliferative activity against six representative human solid tumor cell lines was evaluated for compounds 3-6 [27]. The results showed that compound 5 was the most potent compound of the series, with a modest activity against four of the cell lines tested (GI50 13-48 μM). Compound 4 was active against two of the cell lines (GI50 33-45 μM), whilst the remaining compounds were inactive (GI50 > 50 μM) (see Supplementary Materials, Table S15). The antiproliferative activity against six representative human solid tumor cell lines was evaluated for compounds 3-6 [27]. The results showed that compound 5 was the most potent compound of the series, with a modest activity against four of the cell lines tested (GI 50 13-48 µM). Compound 4 was active against two of the cell lines (GI 50 33-45 µM), whilst the remaining compounds were inactive (GI 50 > 50 µM) (see Supplementary Materials, Table S15).

General Experimental Procedures
Optical rotations were measured at room temperature in CHCl 3 on a PelkinElmer-241 polarimeter (Waltham, MA, USA) by using a sodium lamp. IR spectra were recorded on a Bruker IFS55 spectrophotometer (Ettlingen, Germany) using methanolic solutions over NaCl disk. NMR spectra were recorded on a Bruker Avance 600 instrument (Karlsruhe, Germany) equipped with a 5-mm TCI (Triple Resonance CryoProbe) inverse detection cryo-probe. 1 H and 13 C NMR chemical shifts were referenced either to the CDCl 3 or C 6 D 6 solvent peaks at 300 K (CDCl 3 : δ H 7.26, δ C 77.0). COSY, HSQC, HMBC and ROESY experiments were performed using standard pulse sequences. 3 J H,H values were measured from 1D 1 H NMR. The HSQC-HECADE pulse sequence was used to measure long-range heteronuclear coupling constants. All experiments were performed in the phase-sensitive mode (States-TPPI (Time-Proportional Phase-Incrementation frequency discrimination) or echo-antiecho for quadrature detection in F1) and used gradient coherence selection. The HSQC-HECADE experiment was recorded using DIPSI (Decoupling in the Presence of Scalar Interactions) during the 40 ms of the isotropic mixing period using a bandwidth of 10 kHz, and a J-scale factor of 1 was used. Prior to Fourier transformation, zero filling was performed to expand the data to at least double the number of acquired data points. HR-ESI-MS data were obtained on a LCT Premier XE Micromass spectrometer (Waters, Milford, CT, USA). HPLC (High performance liquid chromatography) separations were carried out with a LKB 2248 system (Bromma, Sweden) equipped with a photodiode array detector. TLC (Thin layer chromatography) (Merck, Darmstadt, Germany) was visualized by spraying with phosphomolybdic acid reagent (10% in EtOH) and heating.

Computational Methods
Conformational searches were undertaken using the Macromodel software (version 8.5, Schrödinger Inc., San Diego, CA, USA) and the MMFF94 force field. Solvation effects of CHCl 3 were simulated using the generalized Born/surface area (GBSA) solvation model. Extended non-bonded cutoff distances (a van der Waals cutoff of 8.0 Å and an electrostatic cutoff of 20.0 Å) were used. Local minima within 10 kJ of the global minimum were saved and analysis of the results was undertaken using Maestro software. Quantum mechanical calculations were carried out with the Jaguar package (Jaguar; Schrödinger LLC, New York, NY, USA). Single-point energy calculations were performed at the DFT theoretical level in the gas phase. The B3LYP hybrid functional with the LACVP ** basis set was used. Chemical shifts were calculated using the gauge-including atomic orbital (GIAO) method. Chemical shifts were calculated from their shielding constants that were first averaged according to their relative Boltzmann populations using a Schrödinger Inc. python script. Proton chemical shifts for each methyl group were averaged due to their conformational freedom.

Antiproliferative Activity
Human solid tumor cell lines A549, HBL-100, HeLa, SW1573, T-47D and WiDr were a kind gift from Prof. G. J. Peters (VU Medical Center, Amsterdam, The Netherlands). The cell lines were cultured in RPMI 1640 medium (Flow Laboratories, Irvine, UK), supplemented with 5% fetal calf serum (FCS, Gibco, Grand Island, NY, USA), 2 mM L-glutamine (Merck, Darmstadt, Germany), 100 U/mL of penicillin G and 0.1 mg/mL of streptomycin (Sigma, St. Louis, MO, USA) at 37 • C in a 95% humidified with 5% CO 2 atmosphere. The in vitro antiproliferative activity was evaluated using the sulforhodamine B (SRB, Sigma, St. Louis, MO, USA) assay with slight modifications [27]. Briefly, pure compounds were initially dissolved in DMSO (Sigma, St. Louis, MO, USA) at 400 times the desired final maximum test concentration. Cells were inoculated onto 96-well plates in a volume of 100 µL per well at densities of 2500 (A549, HBL-100, HeLa and SW1573) or 5 000 (T-47D and WiDr) cells per well, based on their doubling times. Control cells were exposed to an equivalent concentration of DMSO