2.1. Chemical Structural Analysis
Kallopterolide A (
1) was obtained as a pure yellowish oil,
+20.0 (c 1.0, MeOH). Nine degrees of unsaturation were deduced from its molecular formula C
20H
24O
5, established by HR-FAB-MS of the pseudomolecular ion [M + H]
+ at
m/
z 345.1702 (calcd 345.1702). The IR spectrum of
1 indicated the presence of olefin (3093, 1660, 1626 cm
−1), ester (1775, 1757 cm
−1), and aldehyde (2858, 2772 cm
−1) functionalities. The UV spectrum (MeOH) showed absorption maxima at λ
max 209 nm (ε 26,500) and λ
max 245 nm (ε 15,800) in accord with the presence of α,β-unsaturated-γ-lactone and α,β-unsaturated aldehyde arrays in
1. The
13C NMR spectrum recorded in CDCl
3 exhibited all twenty signals (
Table 1), and a DEPT NMR experiment indicated the presence of four methyl, three methylene, and six methines in addition to seven quaternary carbon atoms. Furthermore, the
13C and DEPT NMR spectra of
1 indicated the presence of eleven sp
2-hybridized carbon atoms in the molecule corresponding to four C–C double bonds between δ
C 156.8 and 117.1, two α,β-unsaturated-γ-lactone carbonyls at δ
C 173.8 (C, C-3) and 172.9 (C, C-20), and one aldehyde carbonyl at δ
C 190.6 (CH, C-2). Two oxygenated sp
3 methine carbons at δ
C 79.8 (CH, C-6) and 80.1 (CH, C-8) were also identified. These combined
13C NMR data for kallopterolide A revealed four carbon–carbon and three carbon–oxygen double bonds accounting for seven sites of unsaturation. Thus, the remaining two degrees of unsaturation suggested that compound
1 must be bicyclic.
The 1H-NMR spectrum of 1 indicated the presence of two trisubstituted olefins [δH 7.06 (dd, 1H, J = 1.2, 1.0 Hz, H-5); 7.09 (dd, 1H, J = 1.6, 1.2 Hz, H-9)]; an isopropenyl group [δH 4.89 (br s, 1H, H-18α); 5.07 (br s, 1H, H-18β); 1.72 (br s, 3H, Me-19)]; and a vinyl methyl group [δH 1.92, (d, 1H, J = 1.7 Hz, Me-16)]. Furthermore, combined 1H-NMR and COSY spectroscopic data revealed signals consistent with the presence of an isolated pair of mutually coupled methylene groups [δH 2.52 (m, 2H, H2-12); 2.29 (m, 2H, H2-11)], three adjacent sp3 methines [δH 5.34 (ddd, 1H, J = 3.9, 2.0, 1.9 Hz, H-6); 5.16 (ddd, 1H, J = 10.3, 1.4, 1.0 Hz, H-8); 2.21 (dd, 1H, J = 10.0, 4.3, H-7)] and an α-substituted-β,β-dimethyl-α,β-unsaturated aldehyde [δH 10.10 (s, 1H, H-2); 2.21 (s, 3H, Me-15); 2.03 (s, 3H, Me-14)].
The first partial structure deduced from the 2D NMR spectral data was the α-substituted-β,β-dimethyl-α,β-unsaturated aldehyde functionality (
Figure 3a). HMBC correlations of H-2 (δ
H 10.10, s, 1H) with C-1 (δ
C 135.5, C) and C-13 (δ
C 156.8, C) along with HMBC correlations of the latter carbon atoms with the vinylic methyls at δ
H 2.03 [s, 3H, Me-14] and 2.21 [s, 3H, Me-15] confirmed the presence in
1 of the trisubstituted α,β-unsaturated aldehyde moiety. The
1H-
1H COSY spectrum showed a cross-peak between H-11 and H-12, which, together with key HMBC correlations between H-2 with C-12 and between H
2-12 with C-2 connected the spin system –CH
2-CH
2– (C-11 to C-12) to C-1. These HMBC correlations validated the presence of a (CH
3)
2C=C(CHO)-CH
2-CH
2- fragment in
1. Moreover, H-9 [δ
H 7.09, dd, 1H,
J = 1.6, 1.2 Hz] showed a
1H-
1H COSY correlation with H-8 [δ
H 5.16, ddd, 1H,
J = 10.3, 1.4, 1.0 Hz] which in turn gave strong HMBC correlations with C-20 (δ
C 172.9, C) and C-9 (δ
C 147.4, CH). Further HMBC correlations of C-10 [δ
c 134.6, C] with H-9 and H
2-11, combined with the intense IR absorption band at 1757 cm
−1, vouched for the presence in
1 of an α,β-unsaturated-γ-lactone functionality (C-8 to C-10 and C-20) connected to the prior –CH
2-CH
2-(CHO)C=C(CH
3)
2 fragment through C-10.
The proton resonances at δ
H 5.07 [br s, 1H, H-18β]; 4.89 [br s, 1H, H-18α]; and 1.72 [br s, 3H, Me-19] and the
13C NMR signals at δ
C 137.8 (C, C-17), 117.1 (CH
2, C-18), and 23.7 (CH
3, C-19) were assigned to an isopropenyl group on the basis of HMBC correlations between H
3-19 with C-17 and C-18. HMBC correlations from H
2-18αβ and H
3-19 to C-7 placed the isopropenyl group at the C-7 position. The presence of an additional α-substituted-α,β-unsaturated-γ-lactone unit was deduced from
1H NMR signals at δ
H 5.34 [ddd, 1H,
J = 3.9, 2.0, 1.9 Hz, H-6]; 7.06 [dd, 1H,
J = 1.2, 1.0 Hz, H-5] and
13C NMR resonances at δ
C 173.8 (C, C-3), 130.9 (C, C-4), 146.8 (CH, C-5), and 79.8 (CH, C-6). HMBC correlations from H
3-16 with C-3, C-4, C-5, H-6 with C-4, and H-5 with C-3 corroborated the presence of the latter functionality. Interestingly,
1H-
1H COSY cross-peaks between H-7 with both oxymethines H-6 and H-8 established the remaining spin system =CH-CH-CH-CH-CH= (H-5 through H-9). These combined
1H-
1H COSY and HMBC correlations established the planar structure of
1 (
Figure 3a).
The relative stereochemistry of
1 was established through a combination of 2D-NOESY (
Figure 4a) and coupling constant (
J) data analyses (
Table 1,
Figure 3b,c) in tandem with molecular modeling studies. This task was facilitated by the fact that the stereogenic centers within
1 are contiguous. First off, the
1H NMR spectrum of
1 was re-recorded in CD
3OD (see
Table S1 in the Supplementary Materials), which enhanced signal splitting and facilitated the
J analysis of the pivotal proton H-7. The improved resolution revealed that the conformational flexibility of
1, particularly alongside the C-6–C-7–C-8 bonds, was somewhat restricted, allowing some helpful conclusions to be drawn from these data. In point, the proton resonance ascribed to H-7, which occurs as a doublet of doublet, is coupled with both oxymethines H-6 and H-8. The magnitude of the coupling constant between H-7 and H-8 (
JH7–H8 = 10.0 Hz) indicated that rotation along the C-7–C-8 bond is restricted and that these protons adopt an anti-periplanar arrangement (
Figure 4a). If we place H-7 above the plane (β-configuration), then H-8 must have the α-configuration (below the plane). The absence of a NOESY cross-peak between these vicinal protons supported this contention. On the other hand, H-7 showed a NOESY correlation with H-6, which established their spatial proximity on the β face of the molecule.
This conclusion was validated by the small axial-equatorial coupling constant between these protons (
JH6–H7 = 4.1 Hz). Furthermore, the fixed conformation depicted in
Figure 4a was supported by the strong NOESY correlations observed between H-5/Me-16, H-5/H-7, H-7/Me-19, Me-19/H-18β, and H-8/H-18α. Farther along the eastern quadrant, the peak assignment for Me-14 and Me-15 in the
1H-NMR spectrum of
1 was based on the NOESY cross-peak between H-2 and Me-15. While these methods allowed us to establish the assignments described, we should point out that the relative configuration drawn for
1, namely, 6S*, 7S*, and 8R*, correlates well with the known absolute configuration of other diterpenes co-isolated during this investigation (namely, kallolide A, kallolide A acetate, kallolide C, bipinnapterolide A, and gersemolide). This observation aligns with our contention that, most likely,
1 originates following oxidation/cleavage at C-2/C-3 of a suitable pseudopterane-based precursor (see
Figure 1 and
Scheme S1 in the Supplementary Materials).
Kallopterolide B (
2), an optically active yellowish oil,
+5.0 (c 1.0, MeOH), showed a pseudomolecular [M + 1]
+ ion peak at
m/
z 345.1696 in the HR-FAB-MS corresponding to a molecular formula of C
20H
25O
5 (calcd 345.1702). The IR and UV spectroscopic data for
2 were very similar to those recorded for kallopterolide A (
1). Further examination revealed that their
1H and
13C NMR data in CDCl
3 were also almost identical, indicating that both compounds possess identical functionality, namely, two α,β-unsaturated-γ-lactones, one α-substituted-β,β-dimethyl-α,β-unsaturated aldehyde, and one isopropenyl group. Therefore, we concluded that these compounds must be diastereomers. A detailed side-by-side comparison of the
1H NMR spectra of
1 and
2 (
Table 1) revealed that the minor differences observed could be explained by inverting the configuration in
2 at C-6. The reversal at C-6 from S* in
1 to R* in
2 was rendered by subtle differences in the
1H NMR chemical shifts and coupling constants for H-6 [δ
H 5.34 (ddd, 1H, 3.9, 2.0, 1.9 Hz) in
1 vs. δ
H 5.05 (dd, 1H, 1.7, 1.6 Hz) in
2] and H-7 [δ
H 2.21 (dd, 1H, 10.0, 4.3 Hz) in
1 vs. δ
H 2.60 (dd, 1H, 7.0, 7.0 Hz) in
2] (
Table 2).
Moreover, the coupling constant values between H-6/H-7 (
JH6-H7 = 7.0 Hz) and H-7/H-8 (
JH7–H8 = 7.0 Hz) [
9] (
Figure 5a,b), combined with molecular modeling studies and key NOESY correlations between H-6 and H-9 and between H-6 and H-18α (
Figure 4b) suggested that these proton pairs lie within spatial proximity toward the α face. Interestingly, the chemical shift of H-6 (δ
H 5.05) in
2 appears upfield when compared to that of H-6 (δ
H 5.34) in
1. This shielding albeit small can be explained by the proximity of H-6 to Δ
17 (anisotropic effect). Conversely, when H-6 has the opposite equatorial-like orientation (as in
1) the olefin functionality and the latter proton lie too far away from each other (
Figure 4b).
Given that kallopterolides A (
1) and B (
2) possess very similar NMR spectra, we sought additional computational support for our stereochemical assignment. For this, we used a machine learning-augmented DFT method, DU8ML, which in the past proved both fast and accurate for natural products of this size [
10,
11]. As shown in
Table 3, calculated chemical shifts alone are not sufficient to differentiate between the potential diastereomers. This was not unexpected, given that the actual experimental-experimental RMSD value for
13C NMR shifts of the two compounds is a mere 0.38 ppm. We, therefore, have included all three parameters, i.e., RMSDs for the
1H-
1H spin–spin coupling constants,
1H chemical shifts, and
13C chemical shifts, presented as triads in
Table 3, e.g., {1.73/0.19/1.49}. Analysis of the four diastereomers reveals that the most important differentiating factor is the calculated proton spin–spin coupling constants, with RMSD values for the wrong diastereomers exceeding 1.5 Hz. The two correct diastereomers have shown good matches across all three calculated RMSDs. For details, see
Scheme S2 in the Supplementary Materials.
Kallopterolide D (
4),
–10.0 (
c 0.9, CHCl
3), was isolated as an optically active yellowish oil. The molecular formula C
20H
24O
6, deduced from HR-FAB-MS analysis of its pseudomolecular ion (
m/
z [M + H]
+ 361.1652, calcd for C
20H
25O
6 361.1651), required nine sites of unsaturation. The IR spectrum of
4 indicated the presence of hydroxyl (3461 cm
−1), aldehyde (2870 cm
−1), ester (1775, 1752 cm
−1), and olefin (3090, 1663, 1625 cm
−1) functionalities. The UV spectrum (MeOH) showed maxima at λ
max 210 nm (ε 16,800) and λ
max 263 nm (ε 19,800). The
13C NMR spectrum displayed twenty signals (
Table 4), of which eight were olefinic and three were carbonyl carbon resonances, suggesting that compound
4 was also bicyclic. Interpretation of the 1D and 2D NMR spectra revealed the presence in
4 of the following fragments: –CH
2–CH
2–C(CHO)=C(CH
3)
2 (C-1, C-2 and C-13 through C-17) (see
Figure 6c) as well as a tertiary methyl carbinol linked to a α,β-unsaturated-γ-lactone through a methylene carbon (C-8 through C-12 and C-19 through C-20) (see
Figure 6b). The connectivity between the C-9 methylene bridge and the internal γ-butenolide was accomplished from
1H-
1H COSY correlations between oxymethine H-10 [δ
H 5.15, ddd, 1H,
J = 8.2, 4.2, 2.6 Hz)] with H
2-9αβ [δ
H 2.11 (dd, 1H,
J = 14.6, 4.3 Hz) and 1.92 (dd, 1H,
J = 14.6, 8.3 Hz)]. Key HMBC correlations between H
2-9αβ with C-8 (δ
C 71.8, C), C-10 (δ
C 80.9, CH), and C-19 (δ
C 30.1, CH
3) allowed us to complete the assembly of unit
b. Further analyses of the 2D NMR data revealed the presence of a relatively unusual 5-ethylidenyl-3-methyl-2(5H)-furanone functionality (C-3 through C-7 and C-18) (see
Figure 6a). The presence in
4 of subunit
a was deduced from the carbon resonances at δ
C 170.0 (C, C-3), 129.3 (C, C-4), 138.6 (CH, C-5), 146.9 (C, C-6), 118.3 (CH, C-7), and 10.5 (CH
3, C-18), and the proton signals at δ
H 7.05 (dd, 1H,
J = 1.5, 1.0 Hz, H-5), 5.43 (s, 1H, H-7), and 2.01 (s, 3H, H
3-18). The following key HMBC correlations supported this contention: H-5/C-4, C-6; H-7/C-5, C-6; and H
3-18/C-3, C-4, C-5. Additionally, the presence of strong absorptions in the IR (1775 cm
−1) and UV spectra [λ
max 263 nm (ε 19,800)] of
4 supported our conclusions. Finally, strong HMBC correlations between H-7 with C-8, C-9, and C-19 secured the connectivity between partial units
a and
b. In turn, units
c and
b were connected through HMBC correlations between H-13 with C-11, C-12, and C-20, and those between H-14 with C-12. These combined data established the structure of kallopterolide D (
4) devoid of all stereochemical elements.
In all, three stereogenic centers are present in kallopterolide D: the two asymmetric carbon atoms at C-8 and C-10 and the Δ
6 trisubstituted double bond. From the outset, we realized that the relative stereochemistry of kallopterolide D was going to be difficult to ascertain given the acyclic nature of its structure as well as the non-adjacency of its two chiral carbons [
12]. It should, therefore, be noticed that, except for the Z-configuration assigned to Δ
6 (vide infra), the 8R* and 10S* configurations depicted in
4 should be taken as tentative. First off, since the 10S absolute configuration for bipinnatin J (see
Scheme S1 in the Supplementary Materials), a likely biogenetic precursor to
4, has been established by asymmetric synthesis, we assigned the S* configuration at C-10 in
4 [
13]. Interestingly, after conducting a series of molecular modeling studies and 2D-NOESY experiments we envisioned that kallopterolide D has the propensity to adopt the S-shaped conformation shown in
Figure 7.
Strong NOESY correlations between H-7 with both H-5 and H-10 quickly established the Z geometry about the Δ
6 olefin and vouched for our assignment for the 10S* relative stereochemistry in
4. Furthermore, the conspicuous absence of NOEs between H-10 and Me-19 and between H-7 and H
2-9αβ, combined with strong NOEs between Me-19 and H
2-9αβ, all argued for the 8R* configuration. Additional validation for the proposed S-shape conformation of
4 stems from the multiplicity and coupling constant values for H-10 (ddd,
JH-9α/H-10 = 4.2 Hz,
JH-9β/H-10 = 8.2 Hz, and
JH-10/H-11 = 2.6 Hz) as well as the supplementary NOESY correlations depicted in
Figure 7. These arguments notwithstanding, the shown stereochemical assignments for
4 (8R*,10S*) are subject to confirmation.
Kallopterolide E (
5) was isolated as an optically active yellowish oil,
–11.4 (c 0.7, CHCl
3). HR-EI-MS of
5 showed a molecular ion [M]
+· at
m/
z 360.1573 appropriate for a molecular formula of C
20H
24O
6. The IR and UV spectra were quite like those recorded for kallopterolide D (
4). The 1D NMR data (
Table 4) and subsequent analysis of 2D NMR data suggested that
5 possessed the same partial structures and identical interconnectivity as those of
4. Thus, we concluded that kallopterolides D (
4) and E (
5) are diastereomers. Following side-by-side comparisons of their
1H and
13C NMR spectra, we quickly realized that the minor spectral differences observed were ascribable to a change in relative stereochemistry in
5 at the C-8 position. Specifically, we argue that the change at C-8 from R* in kallopterolide D (
4) to S* in kallopterolide E (
5) could be inferred from subtle differences in the
13C chemical shifts of C-7 [δ
C 118.3 (CH) in
4 vs. 119.2 (CH) in
5] and C-19 [δ
C 30.1 (CH
3) in
4 vs. 28.9 (CH
3) in
5]. Furthermore, the small differences observed in
1H NMR chemical shifts and coupling constant data for H-7 [δ
H 5.43 (s, 1H) in
4 vs. 5.37 (s, 1H) in
5], H-9α [2.11 (dd, 1H,
J = 14.6, 4.3 Hz) in
4 vs. 2.40 (dd, 1H,
J = 14.7, 3.5 Hz in
5], and H-9β [1.92 (dd, 1H,
J = 14.6, 8.3 Hz) in
4 vs. 1.85 (dd, 1H,
J = 14.6, 5.3 Hz) in
5] also supported this conclusion.
As in
4, molecular modeling analyses in combination with 2D-NOESY experiments indicated that at 20 °C, a solution of kallopterolide E (
5) in CDCl
3 does not adopt a linear conformation either. Instead, compound
5 attains a more stable S-shape conformation (
Figure 8). As we saw before in
4, strong NOESY correlations were observed in
5 between H-5/H-7 and H-5/H-18, which strongly argued for the Z-geometry of Δ
6. This time, however, and contrary to what was observed for kallopterolide D (
4), a strong NOESY correlation between H-7 and H-9α combined with the conspicuous absence of NOESY cross-peak between H-10 and H-7 clearly upholds the 8S*, 10S* configuration shown in kallopterolide E (
5) (
Figure 8). The proposed change in relative configuration at C-8 is consistent with the variations observed in coupling constant values for H-10 (
JH-9α/H-10 = 3.4 Hz, J
H-9β/H-10 = 5.1 Hz,
JH-10/H-11 = 1.7 Hz). As with
4, the 8S* and 10S* stereochemical assignments for kallopterolide E (
5) are subject to confirmation.
Kallopterolide C (
3) was isolated as an optically active yellowish oil,
+12.5 (c 0.4, MeOH). The LR-EI-MS of
3 exhibited its molecular ion [M]
+· at
m/
z 360, appropriate for a molecular formula of C
20H
24O
6. However, attempts to measure the exact mass of
3 using HR-MS techniques (HR-EI-MS, HR-FAB-MS and HR-ESI-MS) failed to secure this information. Interestingly, the IR spectrum of
3 was quite similar to those of kallopterolide D (
4) and kallopterolide E (
5). Side-by-side comparisons of the 1D-NMR (
Table 4) and 2D-NMR spectroscopic data of
3 with those for stereoisomers
4 and
5 quickly revealed the presence in
3 of the already familiar partial structures
a–
c (devoid of relative stereochemistry) previously remarked in
Figure 6. On the other hand, when the UV spectra in MeOH of these stereoisomers were compared,
3 revealed a subtle hypsochromic effect (λ
max 257 nm for
3 vs. 263 nm for
4), suggesting a change in the geometry of kallopterolide C (
3) about the 5-ethylidenyl-3-methyl-2(5H) furanone functionality [
7]. Careful comparisons of the
1H- and
13C-NMR spectra of these compounds supported this contention [
14,
15,
16,
17]. For instance, the change in geometry at Δ
6 from Z in kallopterolide E (
5) to E in kallopterolide C (
3) was clearly implied by the differences in the
13C chemical shifts of C-4 [δ
C 129.5 (C) in
5 vs. 131.3 (C) in
3], C-5 [δ
C 138.5 (CH) in
5 vs. 136.5 (CH) in
3], C-6 [δ
C 146.0 (C) in
5 vs. 149.5 (C) in
3], and C-7 [δ
C 119.2 (CH) in
5 vs. 117.4 (CH) in
3]. Additionally, differences observed in the
1H NMR chemical shift and coupling constant data for H-5 [δ
H 7.02 (dd, 1H,
J = 1.5, 1.0 Hz) in
5 vs. 7.88 (q, 1H,
J = 1.5 Hz) in
3] and H-7 [5.37 (s, 1H) in
5 vs. 5.62 (s, 1H) in
3] plus the conspicuous absence of NOESY correlations between H-5 and H-7 firmly established the E geometry for Δ
6 in
3. As in
5, extensive molecular modeling studies of the lowest-energy conformation shown in
Figure 9 combined with 2D-NOESY experiments allowed us to assign the remaining relative stereochemistry for
3 as 8S*, 10S*. Only with the shown configuration could we explain the coupling constant values for H-10 (
JH-9α/H-10 = 2.8 Hz and
JH-9β/H-10 = 10.3 Hz) and the most salient NOESY correlations observed for kallopterolide C (
3) (
Figure 9).
The results of DU8ML calculations for the potential candidate structures of kallopterolides C (
3), D (
4), and E (
5) are presented in
Table 5. The fact that C-8 is quaternary presents an additional challenge of stereochemical assignment, as the
1H-
1H spin-spin coupling constants are not informative. In this case, the assignment was solely based on
13C data. Two unambiguous matches were identified: kallopterolide C (
3) as the SS-E isomer, RMSD (δ
13C) = 1.23 ppm, and kallopterolide E (
5) as the SS-Z isomer, RMSD (δ
13C) = 1.12 ppm. The stereoconfiguration of kallopterolide D (
4) was then confirmed as SR-Z, i.e., as the remaining choice between the SR-Z and SR-E stereoconfigurations.
Kallopterolide F (
6) and kallopterolide G (
7) were isolated as optically active yellowish oils with similar optical rotations and spectroscopic data. The HR-MS analysis of each compound suggested the same molecular formula of C
20H
24O
5, which indicated nine degrees of unsaturation. The IR,
1H, and
13C NMR spectra indicated the presence of olefin, aldehyde, and ester functionalities. Careful analysis of the
1H,
13C (
Table 6), DEPT-135, HMQC,
1H-
1H COSY, and HMBC (
Figure 10) revealed the presence of partial structures –CH
2–CH
2–C(CHO)=C(CH
3)
2 (C-1 to C-2 and C-13 to C-17), α-methyl-α,β-unsaturated-γ-lactone (C-3 to C-6 and C-18), and α,β-unsaturated-γ-lactone (C-10 to C-12 and C-20) as in kallopterolides C–E (
3–
5).
The difference of sixteen mass units in the molecular formula of
6 and
7 suggested that the tertiary hydroxy group in kallopterolides C–E (
3–
5) must have been replaced by a trisubstituted alkene across C-7/C-8. This observation was corroborated by the absence of a broad IR absorption band ascribable to hydroxy functionality in the IR spectra of kallopterolide F (
6) and kallopterolide G (
7). The presence of a distinct trisubstituted olefin in kallopterolide F (
6) was inferred by the proton signals at δ
H 5.08 (dd, 1H,
J = 8.6, 1.0 Hz, H-7) and 1.87 (d, 3H,
J = 1.0 Hz, H
3-19), combined with carbon resonances at δ
C 122.3 (CH, C-7), 138.1 (C, C-8), and 18.1 (CH
3, C-19). The –CH=C(CH
3)CH
2– (C-7 to C-9) connectivity was established by key HMBC correlations: H
3-19/C-7, C-8, C-9; H-7/C-8, C-19; H-7/C-9 (
Figure 10).
Key HMBC correlations between H-6 with C-8, H-7 with C-5, C-9, in addition to those of H-10 with C-8, C-9 connected the α-methyl-α,β-unsaturated-γ-lactone (C-3 to C-6) at C-7, and the α,β-unsaturated-γ-lactone at C-8 (C-10 to C-12 and C-20). Concurrent 1H-1H COSY experiments distinctively indicated the coupled proton spin systems across all these substructures. Careful evaluation of the overall 2D NMR data recorded for kallopterolide F (6) and kallopterolide G (7) demonstrated that these compounds shared the same planar structure.
The “open-chain” nature of structures
6 and
7 severely hampered our ability to assign their relative stereochemistry. In addition, DU8ML failed to differentiate compounds
6 and
7’s relative configurations confidently, thus our assignments should be taken as tentative. As a convenient starting point, we adopted for
6 and
7 the same 10S configuration as that usually found in cembranolides from other sea plume species belonging to the
Pseudopterogorgia genus. Moreover, careful analysis of the NOESY spectra of compounds
6 and
7, as well as side–side comparisons of their coupling constant data together with molecular modeling studies, established that both molecules share a similar 3D conformation (
Figure 11). In particular, the pivotal H-10 proton, in
7, showed a NOESY cross-peak with the H
3-19 protons. The latter methyl protons, in turn, showed a NOESY correlation with H-6. If we assume that H-10 lies in the β face, H-6, too, must be assigned to the same face. The anti-periplanar relationship between H-6 and H-7 was deduced from the coupling constant value of 8.6 Hz.
Interestingly, the NOESY spectrum of kallopterolide G (
7) displayed similar NOESY correlations except for the key correlation between H-10 and H
3-19 that was not present between these protons in the NOESY spectrum of kallopterolide F (
6). These dissimilarities connote that
bis-butenolides
6 and
7 are epimers at C-6 (
Figure 11). Molecular modeling experiments corroborated these observations and established that the most likely relative stereochemistry for kallopterolides F (
6) and G (
7) is 6R*,10S* and 6S*,10S*, respectively. In both
6 and
7, the E geometry was assigned to the Δ
7-trisubstituted olefin based on the shielded methyl carbon resonance at δ
C 18.1 in kallopterollide F (
6) and δ
C 17.2 in kallopterolide G (
7), respectively. The respective absence of a NOESY cross-peak between H-7 and H
3-19 in the spectrum of each compound confirmed the proposed geometry.
Kallopterolide H (
8) was isolated as a yellowish oil,
+56.9 (c 1.0, acetone). The HR-ESI-MS exhibited a pseudomolecular ion [M + H]
+ at
m/
z 361.1654 (calcd 361.1651, C
20H
25O
6), appropriate for a molecular formula of C
20H
24O
6. The latter required nine degrees of unsaturation, which was supported by
13C NMR and DEPT NMR data (
Table 6). A difference of sixteen mass units in the molecular formula of kallopterolide G (
7) in relation to kallopterolide H (
8) revealed the presence of an extra oxygen atom in compound
8. The
1H and
13C NMR data for kallopterolide H (
8) were very similar to those for kallopterolide G (
7), but they did not show the characteristic aldehyde resonances at δ
H 10.10 (s, 1H, H-2) or δ
C 190.7 (C, C-2). On the other hand, the appearance of a shielded carbonyl resonance at δ
C 172.3 (C, C-2), in addition to a broad IR absorption band at 3446 cm
−1, corroborated the presence of a carboxylic acid functionality. This information suggested that kallopterolide H (
8) is the C-2 carboxylic acid derivative of kallopterolide G (
7). Because similar NOEs and 1D NMR data (
1H and
13C NMR) were observed for each compound, it was concluded that most likely they possess identical stereochemistry. Unfortunately, the DU8ML method failed to assign the relative configuration for
8 confidently due to intra- or inter-molecular H-bonds in the conformational equilibrium.
Lastly, kallopterolide I (
9) was isolated as a homogeneous yellowish oil. Unfortunately, ensuing decomposition of this compound after purification made it impossible for us to obtain IR, [α]
D, UV, or HR-MS data. The planar structure of this compound was however elucidated using 2D NMR data collected prior to its decomposition. The overall 1D NMR data (
Table 1) for compound
9 showed fourteen carbon resonances corresponding to two carbon-carbon and three carbon-oxygen double bonds indicating six sites of unsaturation. Careful analysis of the
1H,
13C DEPT-135, HMQC,
1H-
1H COSY, and HMBC quickly revealed the presence of the –CH
2–CH
2–C(CHO)=C(CH
3)
2 and α,β-unsaturated-γ-lactone substructures. The appearance of two carbon resonances at δ
C 204.5 (C, C-8) and 30.5 (CH
3, C-19), combined with the proton signal at δ
H 2.21 (s, 3H, H
3-19) swiftly led us to identify a methyl ketone functionality. The presence of the latter functionality was confirmed by HMBC correlations of C-8 with H
2-9 and H
3-19. The lactone moiety was connected with the C-9 methylene from HMBC correlations of H
2-9 to C-10 and C-11. This compound must likely stem from the C-7/C-8 oxidative cleavage of kallopterolides F (
6), G (
7) or H (
8). The 10S stereochemistry depicted in
9, although as likely as not to be correct, is implied and thus subject to confirmation.