Optical rotations were measured on a Jasco P-2000 polarimeter. UV and CD spectra were measured using a JASCO J-810 spectropolarimeter. NMR spectra were recorded at 25 °C on a Bruker Avance 500 spectrometer at 500/125 MHz (¹H/¹³C). Chemical shifts were reported in ppm using residual CDCl₃ (δ 7.26 for ¹H and 77.2 for ¹³C) as internal reference. Coupling constants (J) are given in Hertz. HPLC purification was carried out on a Waters 600 system equipped with a Waters 717 plus autosampler, a Waters 996 photodiode array detector, and a Sedex 55 evaporative light-scattering detector (Sedere, France). GC-MS was performed on an Agilent technologies series 6890 GC coupled with the Agilent Technologies 5973 MSD. High resolution mass spectra (HRESIMS) were conducted on a LTQ Orbitrap mass spectrometer (Thermo Finnigan).

One specimen of Haliclona fulva was collected by scuba diving along the coast of Saint-Jean-Cap-Ferrat (43° 41’ 29.9646”, 7° 19’ 11.8668”), at a depth of ca. 25 m.

After extraction of the freeze-dried sponge (50g) by a mixture of methanol and dichloromethane at a ratio of 1:1, the solvent was evaporated under reduced pressure yielding to 13.8g of crude extract. This extract was then fractionated by RP-C₁₈ Vacuum Liquid Chromatography (elution with a decreasing polarity gradient of H₂O/MeOH from 1:0 to 0:1, then MeOH/CH₂Cl₂ from 1:0 to 0:1). After an HPLC analysis of the five fractions previously obtained, we realized a further fractionation by Flash chromatography on a diol phase using a gradient of cyclohexane/EtOAc/MeOH (from 100% cyclohexane to 100% MeOH). Asubsequent HPLC purification of the cyclohexane/EtOAc/MeOH 3:1:0 fraction was realized usinga Symmetry C₁₈ (7.8 × 300 mm;7 µm) Waters™ column with a gradient of H₂O/MeOH (from H₂O/MeOH 50:50 to 0:100 in 16 min, flow rate = 3.0 mL.min⁻¹) and afforded compounds 1 and 2 (1 mg). Separation of both compounds 1 and 2 was performed after methyl esterification using the same HPLC conditions.

3-epi-cladocroic acid (1): ¹H NMR (500 MHz) and ¹³C (125 MHz) NMR (CDCl₃), see Table 1;

cladocroic acid (2): ¹H NMR (500 MHz) and ¹³C (125 MHz) NMR (CDCl₃), see Table 1;

The mixture of 1 and 2 (3 µmol) was taken in a micro reaction vessel. Freshly distilled chlorotrimethylsilane (6 µmol) was added slowly and stirred with a magnetic stirrer. Methanol (2 mL) was added to the reaction, and the resulting solution was stirred at room temperature until completion (reaction was monitored by TLC). The mixture was then evaporated under reduced pressure to obtain the products methyl ester 1a and 2a.

3-epi-cladocroic methyl ester (1a): colorless solid; = + 21.5 (MeOH, c 0.1); UV (MeOH) λ_max (log ε) 210 (0.34); CD (MeOH, c 3.15× 10⁻⁴ M) ∆ε (λ_max nm) 0.81 (212); ¹H NMR (500 MHz) NMR (CDCl₃), see Table 2; EIMS m/z 318 M^+•; HRESIMS m/z 341.24362 [M+Na]⁺; calcd for C₂₁H₃₄O₂Na, 341.24510, ∆ -4.33618 ppm).

cladocroic methyl ester (2a): colorless solid; = + 12.1 (MeOH, c 0.1); UV (MeOH) λ_max (log ε) 210 (0.54); CD (MeOH, c 3.15× 10⁻⁴ M) ∆ε (λ_max nm) 0.32 (212); ¹H NMR (500 MHz) NMR (CDCl₃), see Table 2; EIMS m/z 318 M^+•; HRESIMS m/z 341.24353 [M+Na]⁺; calcd for C₂₁H₃₄O₂Na, 341.24510, ∆ -4.60447 ppm).

GC-MS analysis was conducted using an Agilent technologies series 6890 GC coupled with the Agilent Technologies 5973 MSD by injecting 0.5 µL with a 1/80 split on a Varian VF1-MS column (60 m × 0.25 mm × 0.25 µm). The temperature of the injector was set to 270 °C. A temperature gradient of 5 °C.min⁻¹ was applied to the column, starting from 60 °C to 270 °C. Helium was used as carrier gas (flow rate = 1 mL.min⁻¹). The mass spectral detection was achieved using the quadrupole detector with an electronic ionization at 70 eV.

Conformational analysis was performed using the conformer research algorithm implemented in the Conflex-Barista software [19,20,21,22]. All the Density Functional Theory (DFT) calculations were performed using the Gaussian 03 program package [23] at the B3LYP/6-311+G* level [24] using the integral equation formalism variant (IEFPCM) [25]. TD-DFT was employed to calculate excitation energy (in eV) and rotatory strength R in dipole velocity (R_vel) and dipole length (R_len) forms on the most stable conformers (∆E < 2 kcal.mol⁻¹). The calculated rotatory strengths were simulated in ECD curve by using the corrected Gaussian function [26]: where ∆ is half the bandwidth at peak height and ∆E and R are the excitation energies and the rotatory strengths for transition i , respectively. For the Figure 2, R_vel was used.

Compounds 1 and 2 were obtained as a colorless mixture, and determination of the planar structures was realized without any further purification step. ¹H NMR signal at δ 1.26 ppm indicated that compounds 1 and 2 share a long alkyl chain (see Table 1). The signal at δ 81.3 (C, C-18) ppm on the ¹³C NMR spectrum, together with a thin doublet at δ 3.07 (d, J = 2.3 Hz, 1H, H-19) ppm on the ¹H NMR spectrum, were attributed to an acetylenic proton. COSY correlations between this acetylenic proton and ethylenic protons at δ 5.44 (ddd, J = 10.9, 3.5, 1.4 Hz, 1H, H-17) and δ 6.00 (dt, J = 10.4, 7.3 Hz, 1H, H-16) ppm evidenced a Z enyne terminus substituted by three spin-coupled methylenes at δ 2.32 (q, J = 7.6 Hz, 2H, H-15), 1.41 (m, 2H, H-14) and 1.26 (m, 2H, H-13) ppm. The Z configuration of the C=C double bond was confirmed by the shielding of the the adjacent vinylic methylene at δ 30.4 (CH₂, C-15) ppm. Downfield shifted signals (> 35 ppm) are usually observed for an E configuration [27].

Inspection of the ¹H NMR spectrum also evidenced characteristic cyclopropane signals at δ 0.78 (ddd, J = 8.0, 6.5, 4.1 Hz, 1H, H-20a), 0.96 (dt, J = 10.2, 7.1, 5.2Hz, 1H, H-20a) and 1.08 (ddd, J = 8.5, 8.0, 4.6 Hz, 1H, H-20b) and 1.22 (m, 1H, H-20b) ppm. Indeed an important delocalisation of the σ electrons in a cyclopropane can exert a considerable influence on the chemical shifts of neighbouring protons. The resulting ¹H chemical shifts are lower than those observed for usual methylenes [28].

Analysis of the HMBC spectrum gave supplementary information on the structure. HMBC correlations H-16/C-18, H-16/C-17 and H-16/C-15 confirmed the position of the enyne terminus. Another HMBC correlation was observed between the non equivalent protons of the cyclopropane ring H-20a and H-20b with the C-4 connecting the ring with the rest of the molecules. Although two carboxylic acid were identified on the HMBC spectrum, the absence at 2.2 ppm in the ¹H NMR spectrum of the characteristic triplet signal of a methylene adjacent to a carbonyl, suggested that the cyclopropane ring was alpha connected to the carboxylic acid function. This conclusion was supported by the HMBC correlations H-20a/C-1 and H-20b/C-1. After anlysis of the HMBC and COSY spectra, two sets of correlations were identified for two distinct cyclopropane rings, but no correlations where observed to link them together. We came to the conclusion that two diastereoisomers of the cyclopropane ring (cis and trans) were present in the mixture. A comparison with ¹H and ¹³C NMR data obtained by dAuria et al. for cladocroic acid (1) was consistent with this conclusion [15]. Indeed, in the previously published paper, only two signals for the cyclopropane ring were reported at δ 0.96 and 1.08 ppm while four signals were observed on the ¹H NMR spectrum at δ 0.78, 0.96, 1.08 and 1.22 ppm in our case.

To separate both compounds 1 and 2, a methyl esterification was realized in order to increase the hydrophobic interactions of the compounds with the reverse phase column [29]. A mixture of the two methyl ester derivatives was obtained and subsequently purified using the same conditions than those previously described. ¹H NMR spectra of collected peaks were recorded to confirm that both compounds were successfully separated (Table 2).

Finally, the length of the methyl ester derivatives alkyl chain was reliably determined by GC-MS. A fairly similar fragmentation pattern and the same molecular ion at m/z 318 were observed for compounds 1a and 2b. Identification of the fragments led to the conclusion that the alkyl chain contained 12 methylenes (Figure S11).

Comparing NMR data and the signs of the optical rotation, we concluded that compound 2a was cladocroic acid previously isolated from C. invurvata. Since the authors were unable to assign the absolute configuration [15], compound 2 can be first described as (2S^*, 3R^*)-cladocroic acid while compound 1 can be identified as (2S^*,3S^*)-cladocroic acid. To determine the absolute configuration of both compounds, we then decided to use ECD, a nondestructive and very sensitive method that has been increasingly employed in the last few years in natural product chemistry [16,17,18].

To take advantage of the presence of the ester function chromophore close to the chiral centers of the two compounds 1a and 2a, a study of the absolute configuration was performed by ECD. The double bond of the ester implied two possible electronic transitions, the allowed π → π^* transition that gives a high ε value, and the forbidden n → π^* transition which has a far less important ε value [30]. Since the π → π^* transition has a λ_max = 185 nm that falls below the measurement window (> 200 nm), the study was realized on the n → π^* forbidden transition whereby the energy level is lower (λ_max = 205 nm). The linear methylenes induced a high degree of free rotation during the conformational analysis; the theoretical calculations were performed on a simplified analog (i.e., the cyclopropane substituted by the ester and an ethyl chain). This approximation was considered to be realistic as the effects of the enyne terminus were considered negligible.

Only one Cotton effect (CE) was observed on the ECD spectrum at λ = 210 nm, and the difference of 5 nm between the value and the experimental observation was explained by a slight bathochrome effect of the aliphatic chain (Woodward-Fieser rules). TD-DFT calculations using the integral equation formalism variant (IEFPCM) have been performed on the most stable conformer to simulate the ECD spectrum of both compounds 1a and 2a (Figure 2).

A simple comparison of the sign of the CE at ca. 210 nm indicated that both compounds shared the same absolute configuration on C-2 and could be describe as (2S,3S)-epi-cladocroic methyl ester (1a) and (2S,3R)-cladocroic methyl ester (2a). Since the methyl esterification did not change the absolute configuration of 1 and 2, these results also provided information on the absolute configuration of the previously isolated cladocroic acid (2), which also had the same S configuration at C-2.