1. Introduction
Protium subserratum Engl. (Engl.) (Burseraceae) is a Neotropical tree species belonging to the frankincense and myrrh family. It is widely distributed in lowland rainforests in South America, where it grows on a variety of soil types, ranging from nutrient-poor white sand to relatively nutrient-rich clay [
1]. Because of this,
P. subserratum is considered a habitat generalist. Nevertheless, ecotypes of this species that have distinct morphologies and phylogenetic histories [
2] are associated with the different soil types and can be found growing just meters apart. It is not known what factors promote and maintain habitat specialization in this taxon. A recent field experiment, however, showed that insect herbivore communities interact with soil types to reinforce habitat boundaries [
3]. As part of a larger investigation of plant-herbivore interactions and the diversification of plant lineages, we are examining leaf secondary metabolite chemistry in white-sand and clay-soil ecotypes of
P. subserratum that grow in Amazonian Peru. We focus on young, expanding leaves because they are the preferred food source of most herbivorous insects [
4] and because their defense chemical profiles are often markedly different than mature leaves of the same plant [
5,
6]. In this paper, we report on the structure of a previously unknown squalene derivative (
Figure 1) that accumulates primarily in the leaves of the
P. subserratum ecotype growing on white-sand soils.
Figure 1.
25,30-dicarboxy-26,27,28,29-tetraacetoxy-10,11,14,15-tetrahydrosqualene (1) from Protium subserratum young leaf tissue. Scalar coupling networks (bold bonds) and pertinent HMBC and NOESY correlations (red and blue arrows respectively) are indicated for one mirror half of the molecule.
Figure 1.
25,30-dicarboxy-26,27,28,29-tetraacetoxy-10,11,14,15-tetrahydrosqualene (1) from Protium subserratum young leaf tissue. Scalar coupling networks (bold bonds) and pertinent HMBC and NOESY correlations (red and blue arrows respectively) are indicated for one mirror half of the molecule.
2. Results and Discussion
Compound
1 (
Figure 1) was isolated from dried, ground
P. subserratum young-leaf tissue following flash chromatography fractionation of the 80% ethanol leaf extract and HPLC purification. The high resolution FTICR mass spectrum of
1 gave an [M + H]
+ ion of
m/z 707.40728, indicating a molecular formula of C
38H
58O
12. A proton-decoupled
13C-NMR spectrum acquired at 26 °C, however, showed only 16 resonances. When the acquisition temperature was increased, first to 75 and then to 100 °C, 19 resonances were resolved (
Table 1). A qDEPT experiment showed five unprotonated carbons, three with shifts indicative of carboxy carbonyl carbons (
δ 167.7, 169.0 and 169.2) and two with olefinic shifts (
δ 127.0 and 130.0). DEPT also showed the presence of three methyl, eight methylene and three methine peaks. Six of the methylene carbons had aliphatic shifts while two were oxygen bound (
δ 60.4 and 65.5). In addition, two of the methine resonances arose from olefinic carbons (
δ 132.3 and 139.4) and one from an aliphatic carbon (
δ 35.8). Examination of the
1H-
1H COSY spectrum, with supporting evidence from an HSQC experiment, gave two scalar coupling networks involving three and seven carbons respectively (
Figure 1). An HMBC experiment provided the long-range
1H-
13C correlations that permitted assembly of a structure uniquely consistent with the NMR spectroscopic data. This structure, however, possessed a
sp3-hybridized terminal methylene carbon (
Figure 1, carbon 12). Moreover, despite the fact that every resonance was accounted for, the compound had exactly half the atoms predicted by the high-resolution mass spectrum. The sum of these data led to the conclusion that
1 was a symmetrical molecule consisting of halves that produced identical NMR spectra, each with a nominal mass of 353 amu.
Table 1.
13C- and 1H-NMR Chemical Shifts a and 1H Multiplicities b for 1 in DMSO-d6, 100 °C.
Table 1.
13C- and 1H-NMR Chemical Shifts a and 1H Multiplicities b for 1 in DMSO-d6, 100 °C.
C number c | Assignment | 13C (ppm) | 1H (ppm) | Multiplicity (J, Hz) |
---|
1, 24 | CH3 | 11.1 | 1.75 | br s |
2, 23 | C | 127.0 | | |
3, 22 | CH | 139.4 | 6.63 | dd (7.2, 7.2) |
4, 21 | CH2 | 26.1 | 2,28 | m |
5, 20 | CH2 | 32.6 | 3.17 | dd (7.2, 7.2) |
6, 19 | C | 130.0 | | |
7, 18 | CH | 132.3 | 5.44 | dd (7.3, 7.3) |
8, 17 | CH2 | 23.7 | 2.11 | m |
9, 16 | CH2 | 30.3 | 1.36 | m |
10, 15 | CH | 35.8 | 1.65 | m |
11, 14 | CH2 | 29.8 | 1.30 | o |
12, 13 | CH2 | 25.5 | 1.30 | 0 |
25, 30 | COOH | 167.7 | | |
26, 29 | CH2 | 60.4 | 4.59 | br s |
27, 28 | CH2 | 65.5 | 3.95 | m |
31, 37 | COOH | 169.2 | | |
32, 38 | CH3 | 19.5 | 2.00 | s |
33, 35 | COOH | 169.0 | | |
34, 36 | CH3 | 19.5 | 2.01 | s |
The stereochemistry of the double bonds (
Figure 1) in
1 was determined using a 2D-NOESY experiment. In order to reduce the risk of sample degradation, the NOESY experiment, as well as a second
1H-
1H COSY experiment, was acquired at 26 °C. Data from the NOESY experiment were analyzed in the context of proton chemical shifts that were slightly altered as a result of the change in experimental temperature. NOE correlations between the carbon 1 and 4 protons, as well as the absence of any correlation between carbon 1 and 3 protons, established the
E-configuration for the carbon 2-3 double bond. The presence of NOE correlations between both the carbon 5 and 7 protons and the carbon 8 and 26 protons, as well as the absence of any correlation between the carbon 7 and 26 protons, indicated the
Z-configuration for the carbon 6-7 double bond. The relative stereochemistry at carbon 10 was not determined and there is no record of this squalene derivative ever having been synthesized. Comparisons to standards were therefore not possible.
Based on chromatographic and UV absorbance properties (
Figure 2),
P. subserratum populations of eastern Peru accumulate several related forms of oxidized terpenes, with
1 being the most abundant. At present, the significance of these metabolites is unknown. But given the fact that
1 comprises at least 0.1% of leaf dry weight, its relatively high concentration points to an allelochemical function. Squalene itself has been shown to be an effective synomone: It is synthesized in apple leaf tissue in response to leaf miner attack and attracts the parasitic wasp
Pholester bicolor to probe leaves, even when applied to the leaf surface in the absence of any host herbivore [
7]. Whether the modifications to terpenoid metabolites that we observe in
P. subserratum are related to their signalling properties remains to be investigated. For the purposes of the chemical analyses carried out in this study, only undamaged leaves were collected. Thus, the leaf chemistry we characterized more likely represents a constitutive metabolic profile than one generated in response to wounding.
Figure 2.
Diode array detector (upper panel) and evaporative light-scattering detector (lower panel) chromatograms from HPLC analysis of the fraction containing 1. The extract is from a white-sand ecotype of P. subserratum growing near Jenaro Herrera, Peru. Based on retention times and UV absorbance, several forms of oxidized terpenes are likely present in the leaves, with 1 being the most abundant. Related forms were not isolated in sufficient quantity to permit structure solution.
Figure 2.
Diode array detector (upper panel) and evaporative light-scattering detector (lower panel) chromatograms from HPLC analysis of the fraction containing 1. The extract is from a white-sand ecotype of P. subserratum growing near Jenaro Herrera, Peru. Based on retention times and UV absorbance, several forms of oxidized terpenes are likely present in the leaves, with 1 being the most abundant. Related forms were not isolated in sufficient quantity to permit structure solution.