All the Euplotes strains collected were established starting from single naturally occurring cells to obtain cellular line clones, which were fed on both microalgae and bacteria. Mass cultures were grown in salt water (32‰ salinity), sterilized, and inoculated with microalgae, with the latter first incubated at 23 ± 1 °C in a 12 h light/dark cycle for at least 10 days using a daylight (Osram Daylight lamp, 36 W/10) and fluorescent (Osram Fluora lamp, 40 W/77) illuminated incubator system. Ciliates thrive in a continuously percolating culture of the autonomously reproducing autotrophic microalgae, representing a renewable source.

Starved cells from mass cultures of the strains were pelletted by successive centrifugation rounds to provide from 0.01–0.1 mL of closely packed ciliates for small-scale cell cultures to 0.1–20 mL for large-scale samples. Small-scale cell cultures were used for the morphological and phylogenetic analyses. Morphological information was obtained from living and fixed specimens, employing differential interference contrast microscopy (DIC) and scanning electron microscopy (SEM). The silver nitrate impregnation protocol devised by Foissner et al. [16] and the standard Feulgen staining procedure were applied after cell fixation in Bouin’s fluid to reveal the nuclear apparatus. In assigning strains to the appropriate morphospecies, the intra- (within and between strains) and inter-population variability was considered. Eleven characters commonly used for morphological distinctness among Euplotes taxa were chosen. The determination of the phenotypic traits was carried out on 10 selected specimens of each preparation at 1000X magnification on a Leica DMR microscope, using the dedicated Leica IM1000 version 1.0 software STATISTICA (StatSoft, Inc., USA). Data produced by this morphometric procedure were analyzed using the multivariate technique of discriminating functions.

To gain insight into the taxonomic and phylogenetic relationships among congeneric species of Euplotes, the nuclear small subunit rRNA (SSU-rRNA) gene was selected to determine the phylogenetic trees. The nuclear SSU-rRNA gene sequences were amplified by polymerase chain reaction (PCR) of DNA prepared from small-scale cell cultures concentrated by centrifugation and incubated in lysis buffer (0.5 M EDTA, 1% SDS, 10 mM Tris-HCl, pH 9.5) containing proteinase K (200 μg/mL), at 55 °C for 12–15 h. The universal eukaryotic forward primer 18S F9 (5′-CTGGTTGATCCTGCCAG-3′) [17] and the 18S R1513 Hypo reverse primer (5′-TGATCCTTCYGCAGGTTC-3′) were used [18]. The PCR products were directly sequenced in both directions with one forward [F783 (5′-GACGAAATCAAAGAAATACCGTC-3′)] and two reverse [R536 (5′-CTGGAATTACCGCCGGCTG-3′) and R1052 (5′-AACCTTAAGGAACCCCCG GCCATGGCAA-3′)] internal primers [18]. The SSU-rRNA gene sequences were aligned using the program ClustalX [19] with default parameter settings. The alignment was adjusted interactively with the program BioEdit Sequence Alignment Editor [20] to optimize the base-pairing scheme of the rRNA gene molecule secondary structure [21]. The phylogenetic analysis was performed in PAUP version 4.0b10 [22], applying a neighbor joining method and using Modeltest 3.06 [23] to select the appropriate model of substitution. The Akaike Information Criterion (AIC) indicated that the Tamura-Nei model [24], which considers unequal base frequencies, was the most appropriate. The reliability of the internal branches of the phylogenetic tree was assessed using the bootstrap method [25].

Biological/ecological tests were carried out on various strains of the different morphospecies. Responses of strains to metabolite treatments were assessed by examining the impact on cell fission rates and cell tolerance. Cleaving cells were isolated from well-fed cultures, thus producing groups of physiologically synchronized single progeny individuals. These fission progeny cells normally underwent at least one fission in ciliates of larger size, or two fissions in ciliates of smaller size. Using a micropipette, they were individually transferred to wells of three-spot hemispheric depression slides for 16 h incubation in salt water, sterilized, and then assessed in the absence (control) or in the presence of metabolite. The cultures were kept under observation by light microscopy. The effects of each experimental treatment, as well as the controls, were scored on a series of nine cells, individually distributed into the depressions. The fission rate of a cell’s progeny was calculated using the formula: n = log₂ n_x; where n is the fission number performed by a single cell since its isolation into a depression, and n_x is the number of cells scored after 16 h. Cells that were unable to swim or creep on the bottom of the depression, or with motionless cirri (or both), were considered dead.

All the terpenoids reported here, including those isolated from E. focardii, have been tested using an unbiased sample of marine interstitial ciliates in order to verify their biological activities.

Since secondary metabolites in ciliates are not-dispersed in the culture-medium, cells obtained by centrifugation were extracted in ethanol and stored in the same solvent (at -20 °C) prior to analysis. Rapid identification and quantitative estimates of known compounds were obtained from the ethanol extracts using de-replication techniques based on liquid chromatography-tandem mass spectrometry (LC-MS/MS) methods. Photo diode array was used for UV-Vis detection, while sample MS ionization used an ESI or APCI ion-source (in either positive- or negative-ion mode) coupled to an ion trap mass analyzer. An aliquot of 5 μL of the ethanol solution of each sample (the supernatant of the cell suspension containing about 0.2 mL of cell-pellet in 1 mL of alcohol) was injected into a C18 column (4.6 × 150 mm, 3.5 μm) with a 7:3 mixture of CH₃CN/H₂O as the eluent and a flow rate of 0.9 mL/min.

In order to isolate pure metabolites, the ethanol solution obtained through cell pellet filtration was evaporated and the residue partitioned between hexane-ethyl acetate 9:1 (organic part) and methanol-water 9:1 (aqueous part). The organic extract was then subjected to reversed-phase flash chromatography (RP-FC) on a Lichrolut RP18 column with a CH₃CN-MeOH gradient elution, collecting 5 fractions of about 2 mL. The first fractions that eluted usually contained the terpenoid metabolites, which could be further purified by RP semi-preparative HPLC (RP18, CH₃CN-H₂O 7:3, 5 mL/min). The latter FC fractions contained different classes of lipids such as sterols, di- and triacylglycerols, free fatty acids and phospholipids. Structural elucidation of the compounds under investigation was performed using nuclear magnetic resonance (NMR) spectroscopy, mass spectrometry (MS) and infrared spectroscopy (IR), combined with the results from molecular mechanics (MMX and MM3 force fields) and density functional theory (DFT).

When investigation into marine ciliated protozoa in the subclass Hypotrichida first began, morphological data proved ambiguous in deciding whether Euplotes vannus and Euplotes crassus were separate species or part of a species complex. Indeed, the Euplotes vannus-crassus-minuta group is a complex of closely related, cosmopolitan, marine ciliates, for which the ability to distinguish between them continues to be contentious. Three morphospecies have been identified with E. vannus being the largest form, E. minuta the smallest, and E. crassus presenting intermediate morphological characteristics. In particular, the morpho-variability of E. vannus and E. crassus is low and their microscopic identifying features are not clearly separated (Figure 2, left).

Starting with the most frequently sampled hypotrichid E. crassus, strains collected in different geographical sites have all been found to give the same secondary metabolites, referred to as the euplotins (Scheme 2). These are sometimes also accompanied by their putative acyclic precursor, preuplotin [26]. Thin-layer chromatography (TLC) can be used to distinguish the protozoan E. crassus as it is the only euplotin producer within the E. vannus-crassus-minuta complex of sibling species. As shown in Figure 2, for E. crassus extracts, three UV active spots ascribed to euplotin A, B and preuplotin are detected, along with a dark spot (Retention factor, R_f 0.62) for euplotin C, which appears upon staining. For E. minuta, only preuplotin is detected, while strains of the E. vannus morphospecies contain neither euplotin nor preuplotin.

A typical chromatogram obtained from analysis of an E. crassus strain (Pdr1) is reported in Figure 3, showing the chromatographic separation of the euplotins, characterized by peak-selective UV and ESI-MS ion spectra. Depending on the environment of the electrospray ionization chamber, concentration and composition of the eluate, the sodium adduct [M + Na]⁺ of the molecular ion is usually detected in the positive-ion mode ESI-MS spectrum, followed by potassium adducts [M + K]⁺ and/or dimeric ion-species [2M + 2Na]²⁺. This analysis provides both qualitative and quantitative data and is currently employed in the initial de-replication analysis of new strains. With a single HPLC-UV-MS run, three important pieces of molecular information are obtained simultaneously: polarity (retention time, t_R), chromophoric groups (UV-Vis spectra), and molecular weight (MS spectra) of every metabolite eluted by the LC system. These data can also be used to track metabolic changes during the life-cycles of the microorganisms.

In all investigations on different strains of E. crassus, the acetylated sesquiterpenoid euplotin C (7) was the most abundant metabolite, and can thus be used as a chemotaxonomic marker of this morphospecies. Euplotin C is often followed by at least one of its structural analogs, euplotin A (5) or euplotin B (6), and occasionally by the putative acyclic precursor, preuplotin (8) (Scheme 2) [27].

The euplotane skeleton is a tricyclic ring system, built on a trialdehyde-masked system wherein an unusual 2,6-trans-stereochemistry imposes a high degree of strain on the overall structure. According to ab initio DFT calculations, the 2,6-trans ring junction leads to an increase in the euplotin strain energy of about 12 kcal/mol compared to that of the diastereoisomeric 2,6-cis analog (Scheme 2, bottom). Interestingly, the structural features of the latter are present in udoteatrial hydrate [28], a diterpenoid isolated from the marine alga Udotea flabellum.

Another feature of the euplotins is their high chemical instability [16]. In mild basic conditions, euplotin C readily undergoes a cascade of irreversible transformations, leading to four major products, as regio- and diastereoisomeric mixtures (9–12 in Scheme 3). Loss of the acetate group at C(15), releases the strain imposed by the trans 2–6 ring junction in the euplotane skeleton, since the incipient C(14) aldehyde in the hypothetical trialdehyde intermediate is far from the C(1) aldehyde group, regardless of the relative stereochemistry. Curiously, while C(1) and C(15) are strongly involved in mutual acetal condensations induced by nucleophilic attack of MeOH, the C(14) aldehyde group does not cause any acetalization. The analysis of the end-products of this reaction confirmed the euplotins structure, and also allowed the establishment of their absolute configuration through the application of the modified Mosher method [30,31] involving MTPA esterification of the hemiacetal groups of the regioisomers 9 and 10 [16].

Different results were obtained when euplotin C (7) was subjected to mildly acidic conditions (e.g., p-toluensulfonic acid in methanol). NMR analysis of the reaction products indicated complete acetalization at C(1) and C(15) (no hemiacetals present) and, more importantly, that the incipient C(14) aldehyde group was trapped by the sp² C(12) in a carbonyl-ene type reaction, affording the expected hydroxy-cyclopentane ring [as an epimeric mixture at C(14)] as in the derivatives 13 and 14. The latter derivative, 14, is derived from acid-catalyzed electrophilic attack of MeOH on the iso-propylidene moiety of 13. As it will become clearer further on, the possibility of acid-catalyzed carbonyl-ene processes is of paramount importance since this kind of intramolecular cyclization seems to play a dominant role in the biosynthesis of the majority of terpenoids isolated from marine ciliates.

Euplotins are also amphiphilic molecules, with a “water-like” polar part defined by an unusual sequence of electron-withdrawing oxygen atoms, as well as a hydrophobic tail. It is therefore not surprising that they give rise to micellar aggregates in water, at least above their critical concentration (CMC). Although the CMC value of euplotins in aqueous solution or the size of their aggregates has not been measured, evidence of micellar phases can be obtained from line-width analysis of ¹H-NMR spectra taken on aqueous samples containing euplotins at different concentrations. In particular, NMR experiments carried out on euplotin C (7) supported the formation of large molecular aggregates by the appearance of strong signal broadening in aqueous ¹H-NMR spectra due to significant decreasing of their transversal relaxation times T₂. These aggregates are only partially destroyed by extensive sonication treatment, as evidenced by NMR spectra signal re-sharpening.

Euplotins A–C (5–7) are even unstable in water, but the end-products are different from those discussed above in methanol solutions. In general terms, the increase of the reactant concentration in the micellar phase is expected to lead to a considerable acceleration of the hydrolysis reaction due to the well known micellar catalysis effect [31]. With euplotins as reactants, however, it could be demonstrated through LC-UV-MS techniques and deuterium incorporation experiments that three consecutive chemical processes occur. First, there is an initial fast process leading to hydrolytic cleavage of the C(15) acetate. Second, there is an electrophilic addition of water to the activated C(7)=C(14) vinyl ether double bond. Third, there is an electrophilic addition of water, catalyzed by acetic acid (deriving from process 1), to the C(10)=C(11) trisubstituted double bond of the prenyl side-chain. While the first two reactions proceed with similar rate constants, the last is much slower and becomes significant only after complete disappearance of the starting reactant, euplotin C (7). Free aldehydes are expected to be labile intermediates playing an equilibration role among the hemiacetal intermediates, but in these experiments they were undetectable.

Only one report has so far been published on the total synthesis of euplotins. Euplotin A (5, Scheme 2) has been synthesized [33] by exploiting the thermal reactivity of suitable heterodienes (2-acylalkenals) that undergo cycloaddition reactions giving the key 5-acyldihydropyran intermediate with the correct euplotane trans 2–6 stereochemistry. Racemic euplotin A has been obtained by functional group transformations.

Since euplotins are rather hydrophobic compounds, their low water solubility represents a serious limitation to the reliability of any bioassay method. In order to overcome these difficulties, inclusion complexes of euplotin C with cyclodextrins (CD) were prepared [34]. After various trials with different CD derivatives, heptakis(2,6-di-O-methyl)-β-cyclodextrin (DIMEB) was finally chosen because of its binding capability and enhanced water solubility. The complex stoichiometry, binding constants and three dimensional structure of the 1:1 inclusion complex of euplotin C(7)-DIMEB were determined using 1D/2D-NMR techniques and LC-ESI tandem MS measurements. The data suggests that, in aqueous solution, 7-DIMEB has a 1:1 stoichiometry, although a 1:2 ratio cannot be ruled out at least at high DIMEB/7 molar ratio. By using LC(PDA)-MS/MS as the probe technique, kinetic investigations of the hydrolysis of the 7-DIMEB complex demonstrated that complexed-7 gives rise to the same end-products as free-7 alone, although every step of the overall hydrolytic process occurs at a lower rate. Thus, the increased solubility and the lower hydrolysis rate of complexed-7 with respect to free-7, suggested that these inclusion complexes can be used in biological tests as reliable sources of euplotin C (7), avoiding the use of organic solvents such as DMSO as the carrier agent.

A putative biosynthetic route to the euplotins from farnesyl pyrophosphate (FPP) via preuplotin (8) as a key intermediate is shown in Scheme 4. After C1 pyrophosphate hydrolysis, it is proposed that intermediate 15 could undergo C6=C7 double bond isomerization, along with acetylation of the OH groups at C(1) and C(15), to give the 1,4-dienol acetate moiety of preuplotin (8). Hydrolysis of 8 followed by nucleophilic addition of the electron rich C(2) to the electron poor C(6) would then afford the trialdehyde terpenoid 16. Thereafter nucleophilic attack by water at C(14), followed by a cascade of internal acetalizations, would lead to euplotin C (7) via ß-elimination of water and O-C(15) acetylation of the hemiacetal diastereoisomeric mixture 17. Biosynthetic experiments with labeled isotopic precursors are planned by the authors of this review, in order to support the proposed biosynthetic route described above and to establish whether or not terpenoid biosynthesis in marine ciliates follows the classical mevalonate pathway [35]. It should be noted, however, that no biogenetic data on the euplotins is currently available.

Three main aspects of the biosynthetic proposal must be highlighted because they are atypical of classical terpenoid biosynthesis. The first is that intramolecular cyclizations are not usually driven by nucleophilic additions but by carbocation-mediated, electrophilic reactions [36,37]. In fact, these cyclizations are typically initiated by a regiochemical addition of an electrophile to a sp² hybridized carbon generating a carbocation, which in turn undergoes subsequent ring forming reactions. In marine organisms this process is often initiated by bromine induced electrophilic attack of double bonds leading to a plethora of haloterpenoids [38]. The second point is that in terpenoid biosynthesis, C(1) usually plays a pivotal role as the electrophilic site of the first intermediate after allylic diphosphate hydrolysis, whilst the C(14) and C(15) methyl groups are rarely used in ring forming processes. The opposite is proposed to occur in the E. crassus metabolism, where the key biogenetic step would be the stereoselective C(2)–C(6) bond formation, possibly deriving from conjugate addition of electron rich C(2) to the electrophilic site at C(6). In fact, the oxidation of C(14) to -CHO inverts the polarity of the C(6)=C(7) double bond making the β-carbon at C(6) particularly electron-poor. At the same time, the hydrolysis/oxidation of C(1) to -CHO makes the α-carbon C(2) an electron-rich, nucleophilic site. Finally, in classical terpenoid biogenesis, oxidations are usually thought of as final processes (sometimes even occurring as isolation artefacts), whereby the end-products (or late carbocationic intermediates) are ultimately quenched by the capture of an external nucleophile such as water.

Enzymes such as terpene cyclases that would be responsible for converting the sesquiterpene precursor FPP into species-specific secondary metabolites in ciliates are expected to be structurally and functionally different from classical cyclases involved in terpenoid biosynthesis of other phyla. This is because they are expected to recognize oxidized FPPs as substrates. Marine ciliates could, however, share the enzymatic pool leading from mevalonic acid (MVA), or from 1-deoxyxylulose-5-phosphate (DXP) [37] to isopentenyl diphosphate (IPP) and FPP.

Euplotes crassus is a cosmopolitan and highly diffused marine morphospecies, with an ability to cope with changing environmental conditions [4,39,40]. This may play an important role in ecological significance of this morphospecies, while the production of the euplotins could be another element in their survival strategy [27]. The branching of the phylogenetic tree denotes, within the morphospecies analyzed, the presence of four distinct groups, all including representatives of cosmopolitan origin (Figure 4). Inter-population variation in genetic determinants following geographical separation would be expected to occur, and would result in the production of distinct secondary metabolites.

Although most strains of E. crassus produce the same secondary metabolites, there are some exceptions. For example, several strains do not produce preuplotin (8), while other strains do not produce euplotin A (5). Some strains produce euplotin C (7) only in very low quantities accompanied by trace amounts of preuplotin (8). Furthermore, the population SL represents an exception within E. crassus being the only strain which does not produce any euplotins at all. From an evolutionary point of view, euplotins are the result of a metabolic pathway, which appears useful in natural selection, regardless of the environmental conditions. This may explain its ubiquity and larger distribution with respect to other morphospecies that share the same habitats, like E. vannus and E. minuta.

At the onset of the investigations, there were two hypotheses regarding the way in which E. crassus exchanges chemical information: (i) by the release of substances into the extracellular medium, or (ii) by direct cell-to-cell contacts. Reported studies, carried out in vivo, suggested that euplotins are utilized to inhibit populations of competitors via cell-to-cell contact, a good strategy to colonize new habitats or defend settlements [16,27]. Evidence that the toxic effect is a result of euplotins action comes from separate experiments carried out in vitro where a non-producing strain of E. vannus (TB6) was treated with varying concentrations of pure 7 and a 7-DIMEB complex at 10 μg/mL. In both cases a 100% reduction on the viable protozoa was observed within one hour. These results confirm that the complexation process does not influence the activity of 7, while control experiments verified that the effect of dimethylsulfoxide (DMSO) or DIMEB was negligible. At the functional level, euplotin C kills non-producing Euplotes cells, whereas at sublethal levels it alters cell cycle, ciliary motility and cell shape [27]. These findings suggest that euplotin C may play an ecological role, representing an adjuvant factor of the multicomponent strategy pursued by E. crassus in broadening its niche size.

These results, in conjunction with the fact that many natural products from marine resources show pharmaceutical and medical relevance, prompted the investigation into the potential in vitro bioactivity of euplotin C (7) against the non-marine pathogenic protozoa Leishmania major and Leishmania infantum, the opportunistic yeast Candida albicans and several prokaryotic microorganisms [40]. The toxicity of 7 against the macrophage-like cell line J774 was used as a mammalian host cell control. DMSO at a non-toxic concentration and DIMEB as a solubilizing agent in water for 7 were also used in the assays. Both free-7 and complexed-7 (i.e., DIMEB-7 1:1 inclusion complex) demonstrated good antileishmanicidal activity at 20 μg/mL, causing irreversible fatal damage. After 24 h, LD₅₀ values of 4.6 ± 0.5 and 8.1 ± 1.8 μg/mL for free-7 and complexed-7 respectively, were recorded. By contrast, the LD₅₀ value was >200 μg/mL for the control J774 cell line. Subsequent studies on the activity of E. crassus (a euplotin-producer) and E. vannus (a non-producer) towards both L. major and L. infantum were in agreement with the data from the isolated compound, and found that cellular contact with E. crassus SSt22 strain causes irreversibly damage to parasitic cells, while the E. vannus TB6 strain does not induce any effect. Although the required amount of euplotin C is too high to be interesting for drug applications, the absence of any cytotoxicity towards the macrophage cells is quite promising.

Further bioactivity results revealed that at high concentrations of >100 μg/mL, 7 showed an inhibitory effect on the yeast Candida albicans and on several strains of the bacteria Streptococcus and Burkholderia spp., whilst at lower concentrations only marginal antimicrobial activity was observed [40]. Interestingly, the complexed-7 form was noted to be more active in longer test times than free-7, possibly due to the higher chemical stability of the former. An inhibitory effect has also been observed for euplotin C on food vacuole formation and fluid phase endocytosis in Paramecium primaurelia, a process that might be mediated by an altered function of cell membranes as well as by modification of the microtubule network [42,43].

The underlying cellular and molecular basis of the effects of euplotin C were recently elucidated using mouse and rat tumor-derived cell lines, followed by the analysis of the nuclear and cytosolic changes usually associated with apoptosis in specific cell lines [29,45]. The apoptotic process is an auto-programmed mode of cell death that is dependent upon energy, as it requires ATP to transmit the death signal from the cytoplasm to the nucleus. It also causes well-defined morphological and biochemical changes in the cell such as cell shrinkage, chromatin condensation, nucleosomal degradation and changes in intracellular cation concentration such as Ca²⁺. The cytotoxicity of 7 was tested on the murine AtT-20 line derived from anterior pituitary tumor cells and the PC12 line originating from a rat chromaffin cell tumor, which expresses properties common to both neurons and neurosecretory cells. It was found that treatment of mouse and rat tumor cells with euplotin C resulted in a dose-dependent cytotoxic effect and in the induction of apoptotic cell death by cellular stress that triggered the activation of the caspase cascade.

Finally, cellular responses to euplotin C in E. vannus have recently been investigated by analyzing the cytosolic changes likely to be associated with cytotoxicity and apoptosis, i.e., intracellular Ca²⁺ concentration, organelle components and related caspases [46]. The most important and general outcome of this study is that the chemical signal carried by euplotin is transduced in a rapid impairing of the membrane electrical properties. This causes motility disorders which in turn hinders the ciliate’s ability to escape the external threat. After some minutes the fate of non producers is sealed because the ion channel pumps are completely impaired [43]. These results suggest that euplotin C exerts a marked disruption of homeostatic mechanisms, which play a determinant role in cell-environment interactions.

While euplotins are built on an unsaturated dioxa-tricyclic system, the structural feature of E. raikovi metabolites is the bicyclo[3.2.0]heptane ring system [47]. However, different populations of E. raikovi produce C(10)-epimeric metabolites, providing evidence for intra-species variability. In particular, the strain Mor1 collected on the Casablanca coast contains raikovenal (14) and its putative biogenetic precursor preraikovenal (15), while all other strains so far examined [48] (GA8, SB8, 39W, LPSA5) produce the C(10) epimer of raikovenal (epiraikovenal, 16) and its seco-analog (secoepiraikovenal, 17) (Scheme 5).

The chemical and spectroscopic properties of epimeric sesquiterpenoids 18 and 20 are, as expected, quite similar. Although this close structural similarity complicates the de-replication process, the two epimers can easily be distinguished by NMR spectroscopy since the ¹³C resonance of Me-C(10) in 20 is almost 6 ppm downfield with respect to the same carbon resonance in 18. The opposite trend is observed for the proton resonance of Me-C(10), which results in an upfield shift of 0.05 ppm in 20 with respect to 18. This turns out to be very useful in the de-replication process, since this signal can easily be detected, even in ¹H-NMR spectra performed on crude cell extracts.

While the structure elucidation of these metabolites was in progress, interesting investigations began into the photochemical conversions among xenicane diterpenes isolated from a brown alga of the genus Dictyota [49,50]. Motivated by the success of these studies, it was hypothesized that the bicyclo ring system embedded in raikovenal could be obtained by direct photo-irradiation of its putative precursor, preraikovenal (19) (Scheme 6). Since [2 + 2]-cycloadditions are photochemically allowed processes, it was decided to sacrifice the small amount of the natural isolated compound 19, in an attempt to obtain an intramolecular photochemical cyclization leading to 18 and/or to 20. Irradiation at α 350 nm of a chloroform solution of preraikovenal produced raikovenal as a minor product, while the major product had a ¹H-NMR spectrum that was superimposable to that of epiraikovenal (Scheme 6). According to the proposal [48], the photochemically produced raikovenal 18 could derive from a [2 + 2] concerted reaction (passing through the diradical intermediate 23) from the boat-like conformation of preraikovenal (19b). On the other hand, the photochemically produced epiraikovenal 20 can be explained only by assuming 19 adopts a chair-like conformation (19a), and passes through intermediate 22. However, if one looks at Scheme 6, the expected stereoisomer from the intramolecular addition of the reactive chair-like conformation is not 20, but its enantiomer ent-20. Indeed, it was demonstrated that the major product (epiraikovenal 20) obtained after purification of the photoaddition product mixture showed a circular dichroic (CD) spectrum opposite to that of natural epiraikovenal 20, thus indicating the compound to be ent-20. Although the absolute configuration at the chiral center at C(7) of natural 19 has not yet established, the outcome of this experiment clearly indicates not only that the photoaddition is a strongly stereocontrolled process but also that the inversion of chirality at C(7) is required in order to obtain natural epiraikovenal 20.

Two independent and different strategies for the total synthesis of racemic 18 and 20 have been reported. The first synthesis described exploits the photochemical methodology in the key step of the bicyclo[3.2.0]heptane ring construction [51]. In the same year (1997), Rosini et al. [52], relying on a completely different approach, reported another interesting total synthesis of racemic raikovenal.

According to the results discussed above, if it is assumed that preraikovenal (19) is the putative biogenetic precursor of natural raikovenal (18), this would suggest that the enzymes involved in the bicyclic ring formation preferentially accommodate and recognize substrate 19 in a boat-like conformation (19b). This is supported by MM calculations, which show this conformation to be relatively stable due to the release of the repulsive 1,3-diaxial interaction of the methyl groups in the relevant transition state between preraikovenal and the intermediate 25 (Scheme 7). In order to explain the production of 20 and 21, an enantiomeric chair-like preraikovenal form, ent-19a, would be required. Since the photocyclization of preraikovenal (19) gave mainly ent-20, the conformer ent-19a is proposed herein to be the biogenetic precursor of epiraikovenal (20).

Euplotes raikovi appears as a group with intra-morphospecific heterogeneity, characterized by the synthesis of one or two related terpenes. In particular, the strain Mor1 contains raikovenal and its putative biogenetic precursor preraikovenal, while all other strains so far examined produce the C(10) epimer of raikovenal (epiraikovenal) and its seco-analog (secoepiraikovenal). Considering that one metabolite and its epimer require different biogenetic pathways, it was hypothesized that the production of functional alternatives embedded in the same bicyclo[3.2.0]heptane skeleton is an ecological strategy of adaptation that could also be seen by phylogenetic studies. The phylogenetic tree obtained for E. raikovi shows the evolutionary split between the strain Mor1 and the fairly homogeneous group comprised of representatives of other strains with cosmopolitan distribution (Figure 5). This is in fair agreement with the natural product divergence discussed above. As with the euplotins for E. crassus, the production of raikovenal/epiraikovenal structures by all strains enables their use as taxonomic markers at the morphospecies level.

The terpenoids from E. raikovi are less active than euplotins against other ciliates belonging to the genus Euplotes. Greater activity was observed [48] against Litonotus strains, ciliates with a characteristic predacious life-style which are killed by 18 at concentration as low as 10 μg/mL, and by 19–21 at higher concentrations. It is worth to note that euplotin C (7) was ineffective against the same predator suggesting the possibility of fine-tuned recognition phenomena in ciliated cells. Biological assays on raikovenal and related molecules demonstrated no killing activity against other ciliates even at concentrations above 20 μg/mL. Two particular features were noted in these experiments. First, epiraikovenal was able to depress the fission rate in E. vannus and E. rariseta strains. However, this effect does not seem to be specific, as these strains are sensitive to various aldehyde-terpenoids. Second, it appears that certain E. raikovi strains use a combination of preraikovenal-raikovenal and epiraikovenal-secoepiraikovenal as a defensive strategy. These compounds appear to be effective only in synergy against predators such as Litonotus, accounting for the non-palatability of the strains [48].

As summarized in Scheme 8, from tropical strains (Sil21, BUN3) of E. vannus morphospecies, two unusual compounds with a novel C30 backbone, named vannusal A (26) and vannusal B (27) were isolated [53]. On the other hand, investigation of two different strains (TB6 and CM1) led to the isolation of new regular sesquiterpenoids such as hemivannusal (28) and prevannusadial A (29) and B (30) [54].

The structure elucidation of vannusals was extremely challenging [53]. Although the molecular formula suggested mono- and di-acetylated triterpenoid structures, it soon became evident that these were not regular triterpenes deriving from squalene or squalene-oxide internal cyclizations, and in fact cannot be considered triterpenoids at all. There is high structural complexity in the central part of these structures, wherein five- and six-membered rings are interlaced in fused, bridged and spiro forms. A further difficulty arose from the violation of the “isoprenic rule” in these metabolites in the central and more intricate part of the molecules. Natural products chemists’ penchant for this rule is well known since it poses strong constrains on the overall skeleton, thereby minimizing the number of candidate structures compatible with the spectral data. Much assistance in the structure determination was obtained by functional group transformation of 26 into a series of derivatives [53], which ultimately revealed the intricate nature of the vannusal A structure.

The relative stereochemistry of several chiral centers in the vannusals has recently been revised (Scheme 8 and Figure 6) [54–61]. The stereoselective total synthesis of vannusal A and B (26 and 27) carried out by Nicolaou et al. confirms the overall structure and main stereochemical architecture of vannusals. In fact, if the relative stereochemistry at C(25) is not considered, the only difference in terms of the relative configurations between the proposed [53] and the revised structures [54,56,60,61] of vannusals derive from the mis-assigned relative stereochemistry at C(6)–C(7). This led to the assignment of the wrong relative stereochemistry of the chiral centers in the “eastern” domain of the molecule with respect to those in “western” one. The brilliant syntheses designed and executed in recent years in the Nicolaou laboratory has finally solved this intricate matter through the total synthesis of eight diastereomeric vannusal B structures [60,61].

Even the structural elucidation of hemivannusal 28 [54] was not a simple task. In particular, the assignment of the relative stereochemistry of its 3 chiral centers, which requires correlation between the stereochemistry of the chiral centers on the rings linked together via a carbon-carbon single bond (C6–C7), around which free rotation is expected to occur. Thus, both the relevant vicinal coupling (³J_6,7) and the dipolar mutual effects (NOE) between the same protons H(6) and H(7) represent averaged values among several C6–C7 rotamers. Assistance in resolving this problem was obtained through an extended conformational search performed within a molecular mechanics approach and refined by ab initio quantum chemical calculations.

The minor metabolites of the CM1 strain, prevannusadial A (29) and prevannusadial B (30), were isolated in very low amounts. Their structures were established [54] by direct comparison of their ¹H-NMR spectra with those reported in the literature for the “diterpenoid 9” [62] isolated by Paul and Fenical from a Caribbean Udotea species (U. flabellum) and for petiodial isolated by Fattorusso et al. [63] from the Mediterranean U. petiolata. Prevannusadial B represents the deprenyl analog of these Udotea diterpenoids. This situation has a strong resemblance with that already discussed of the linear sesquiterpenoid preuplotin (8), with udoteal isolated from the tropical green alga U. flabellum [28].

The possible biosynthetic origin of E. vannus metabolites is proposed in Scheme 9 along similar lines to that of the E. crassus terpenoids. The biosynthesis of hemivannusal 28 may derive from FPP and prevannusadial A (29) precursors through three enzymatic processes. After C1 pyrophosphate hydrolysis, C(14) and C(15) methyl group oxidation would afford the putative intermediate 15 (see also Scheme 4) bearing two α,β-unsaturated aldehyde functions with Z stereochemistry. Following Z→E isomerization of the C6,7 double bond, the intermediate 31 could then lead i) to 29, after Z→E isomerization of the C2,3 double bond and ii) to the intermediate 32, after oxidation of the allylic hydroxyl group of 29 and enolization of the C(15) aldehyde group. The intermediate 32 can be converted via intramolecular addition of the activated nucleophilic center C(2) to C(6), an electrophilic center due to its α position in the conjugated unsaturated system, leading to prevannusadial B (30). The latter could readily react through an intramolecular C(14)–C(10) carbonyl-ene process affording the bicyclic intermediate 33. The main metabolite of the strain TB6, hemivannusal (28), can be obtained after i) acetylation of the free primary –OH group and ii) internal redox reaction taking place in the saturated cyclopentane ring, a process having precedent in literature reports on carbonyl-ene synthetic procedures [63].

As already outlined in the proposed euplotin biosynthesis (Scheme 4), the unusual route involving methyl oxidation and nucleophilic additions seems to be a feature shared by strains belonging to the morphospecies E. vannus. Divergence among secondary metabolites of the two species occurs at the level of the putative common intermediate 15. Furthermore, it relies on the strong preference for acetalization processes in E. crassus leading to the highly strained tricyclic skeleton of euplotins, compared to carbonyl-ene reactions in E. vannus strains leading to the hemivannusane skeleton. Further evidence of this is given by the C30 terpenoids vannusal A and B (26 and 27). Although first thought to arise from a deviation of the squalene pathway to triterpenes, they are now considered as end-products of the nucleophilic condensation of two molecules possessing the regular hemivannusane sesquiterpenoid skeleton (28), via the hypothetical aldol-derived intermediates 34 and 35 (Scheme 10). The C30 terpenoids are possibly derived from a concerted condensation of the “peripheral” C1 and C15 electron-poor carbonyl groups of the first hemivannusane intermediate with the “internal” C2′ and C3′ electron-rich centers of the second intermediate. This biosynthetic proposal was exploited by Nicolaou et al. in their recent stereoselective total synthesis of the vannusals 26 and 27 [60,61].

In the morphospecies E. vannus the chemodiversity of its secondary metabolites mirrors its genetic diversity, as revealed by the phylogenetic analysis. This significant intra-morphospecific variability in the secondary metabolite production is also reflected in the phylogenetic tree. E. vannus branches in at least two distinct genetic groups, with each group comprising representatives producing only one of the different classes of secondary metabolites isolated (Figure 7).

Phylogenetic analysis suggests that all the strains examined to-date constitute a monophyletic clade. From a natural products viewpoint this is advantageous. It means that the evolution of E. vannus utilizes different secondary molecular characters, not only useful as chemical biomarkers to distinguish different populations of the same morphospecies, but also as important representative of new chemical structures that could be exploited as new bioactive molecules.

Analysis of the sesquiterpenoids isolated from E. rariseta (Scheme 11) suggests a scenario similar to that already discussed for E. raikovi. For example, different strains of the same morphospecies collected in different geographical areas produce two epimeric sesquiterpenoids, rarisetenolide (36) (and its epoxide 37) and the corresponding C(10) epimer, epirarisetenolide (38) [65], built on a novel octahydroazulene ring system. So far, only the relative configurations of these terpenoids have been established.

In a series of experiments aimed at elucidating the structures of these terpenoids, the expected mixture of hemiacetal epimers at C(14) was obtained in good yields by DIBAL reduction of the lactone group of 36. At the same time, it was also demonstrated that the α-ß conjugated lactone could be removed by nucleophilic addition at the unsaturated C(3)=C(4) double bond. Indeed, UV irradiation (α 254 nm) of a methanol solution of rarisetenolide (36) in a quartz cuvette at room temperature gave good yields of the corresponding 4-methoxy derivative. Addition of MeOH rather than loss of conjugation under UV irradiation is in accordance with the proposed lactone structures.

As for the C(10) epimeric raikovenals 18/20, the structural similarity between the C(10) epimeric metabolites 36/38 complicates the dereplication process. However, the problem is readily solved using NMR spectroscopy. The ¹H resonance of Me-C(10) (δ_H 2.90, dt) in rarisetenolide is 0.5 ppm downfield with respect to the same proton resonance in epirarisetenolide (δ_H 2.49, dt), and their vicinal coupling patterns are also different. These signals can easily be detected in ¹H-NMR spectra, even in crude cell extracts, and thus it is possible to distinguish which one of the two epimers, 36 or 38, a given strain produces. Until now, a strain of E. rariseta that produces both epimers has not yet been found. This is indicative of a selective and/or different enzymatic pool in populations of this ciliate or the use of enantiomeric biogenetic precursors.

The proposed biosynthesis of rarisetenolide from FPP via the intermediate 44 is shown in Scheme 12. After C1 pyrophosphate hydrolysis of FPP, oxidation of the C(14) and C(15) methyl groups would afford the putative intermediate 39 bearing a saturated aldehyde at what was the C(6)=C(7) double bond. Aldol condensation of this aldehyde with the activated nucleophilic center at C(2), followed by proton tautomerization leads to the intermediate 40. This could then react through an intramolecular C(14)–C(10) carbonyl-ene process affording the bicyclic intermediate 42, which upon dehydration produces 43. This intermediate is proposed to play a pivotal role undergoing acetalization/oxidation processes leading to rarisetenolide-like metabolites through intermediate 44. Herein, C(14) is thus considered as the electrophilic center for the nucleophilic attack of both C(2) and C(10). If it is considered that FPP methyl groups are rarely involved in intramolecular terpenoid cylizations, the enzymatic processes able to build the octahydroazulene ring system of 36 or 38 appears to be a unique feature of ciliate secondary metabolism. It should be noted however, that the putative linear precursor (i.e., “prerarisetenolide”) has not yet been isolated. The C(6)/(7) hydrogenated analog 39, the C(14) oxidized form of preraikovenal (19), should contain all the structural and stereochemical features required to produce the octahydroazulene ring system of 36 and 38. Furthermore, it could also give rise to divergent routes towards 36 or 38, depending on the absolute configuration at its C(7) chiral center. Thus, it is proposed that evolutionary differentiation among E. rariseta strains could rely on C(7) enantiomeric FPP oxidized precursors, such as 39, in a similar way as differentiation among E. raikovi could rely on the enantiomeric FPP oxidized precursors 19 and ent-19.

The strains of the morphospecies E. rariseta analyzed for the secondary metabolite production have been collected from various coastal sites, including those of northern and southern Australia (strains PD16 and PBH1, respectively), southern Brazil (strain BR1), Canary Islands (strain GRH5), and New Zealand (NZ2) [65]. All of them are found to be producers of non-aldehydic terpenoids, which is quite unusual in the Euplotes genus. The three sesquiterpenoids identified, rarisetenolide, epoxyrarisetenolide, and epirarisetenolide, share the same octahydroazulene moiety, which seems to be a convincing chemotaxonomic marker of the morphospecies.

As with E. crassus and E. vannus, this morphospecies also shows intra-morphospecific variability in its secondary metabolite production that is reflected in the phylogenetic partition. The four strains of E. rariseta (PD16, PBH1, BR1 and GRH5) produce both rarisetenolide and epoxyrarisetenolide. These strains, although collected from geographically distinct locations, are all phylogenetically grouped together in the same clade (Figure 8). It is difficult to believe that this is a coincidence. The production of secondary metabolites is related to enzymes coded by the organism’s genome, their presence reflecting the expression of functional genes, and the inference of a close phylogenetic relationship among the above strains is warranted. In contrast to the other strains, the strain NZ2 gives epirarisetenolide, revealing a subtle variability in secondary metabolism for this ciliated morphospecies. The epimeric relationship of this metabolite with rarisetenolide implies the presence of different metabolic pathways in strain NZ2 with respect to the other strains. This is indicative of genetic differentiation of strain NZ2 as revealed by the phylogenetic tree (Figure 8).

Cytotoxicity of rarisetenolide (36) and its congeners (37 and 38) against representatives of the marine interstitial ciliate community, comprising E. rariseta itself, was detected only at relatively high doses, i.e., >20 μg/mL. The only exception observed was towards the ciliate Litonotus lamella, which proved sensitive to 36 at lower doses. It is worth noting that L. lamella differs from all other ciliates involved in the bioassays due to its behavior as a predator. It can be proposed that the effectiveness of 36 against L. lamella is part of a defensive strategy of E. rariseta to avoid predation. Since lower cytotoxicity of the chemically more reactive epoxide 37 with respect to 36 or 38 was found in the ecological tests, interactions with membrane receptors of the Euplotes morphospecies might be non-covalent in nature.

Strains assigned to the species E. focardii have been found to produce the focardins (Scheme 13) [66], which show much of the structural characters of the rarisetenolide/epirarisetenolide family. From the morphospecies collected from Antarctica (strain TN1) and inter-tropical coastal waters (strains GBS1 and GBS5), a new skeletal class of diterpenoid, epoxyfocardin (46), was isolated. Focardin (42), which is a possible biogenetic precursor of 46, was also isolated as a minor component. Focardins constitute the first example of ecologically relevant metabolites from ciliate species that inhabit polar ecosystems.

Focardin (45) was isolated as a 70:30 mixture of hemiacetals 45a and 45b. These equilibrated through the unstable ring-opened, aldehyde intermediate 47 (Scheme 14), to ultimately give epoxyfocardin (46) as an 85:15 mixture of its corresponding hemiacetals 46a and 46b. Evidence of the equilibrium of the diasteroisomeric hemiacetals came from the analysis ¹H-NMR spectra of pure 46a obtained by reversed-phase chromatography of raw extract containing 46 which was superimposable to that obtained before the separation. An unaltered ratio of products clearly suggests equilibrium conditions among the epimers. The absolute configuration of 46a/46b was determined from Mosher’s method [30] through esterification of the hemiacetal hydroxyl group.

The proposed biosynthesis of the E. focardii metabolites described here commences from geranylgeraniol pyrophosphate (GGPP) and involves (in the first steps) similar intermediates to those involved in rarisetenolide biosynthesis, albeit with one prenyl unit more (Schemes 12 and 15). Herein, C(19) is playing the same role as C(14) in the raristenolide pathway, but a striking difference between rarisetenolide and focardin biosynthesis is represented by the different oxidation states of C(1) and C(15)/C(20). Whereas in the formation of 36 (or 38), C(1) is in a lower oxidation state than C(15), in 45 (or 46) it is in a higher oxidation state than C(20).

By morphological and morphometrical analysis, all Antarctic and inter-tropical populations of E. focardii constituted a single cluster of indistinguishable units, thus allowing their attribution to the same morphotype. Nevertheless, within the morphospecies two distinct groups, which corresponded to the two geographical areas of sampling (Antarctic and inter-tropical regions), were detected by phylogenetic analysis (Figure 9). Interestingly, this genetic polymorphism is also reflected by secondary metabolite production. In fact 45 and 46 can be isolated in higher yields from the inter-tropical strains GBS1 and GBS5, whereas the Antarctic TN1 strain produces them only in low amounts (Figure 9). This observation suggests that competition for exploitable niches and food varies with latitudes, thus the survivor strategies are also expected to vary. In particular in the inter-tropical area the metabolite expression would be expected to be higher, since E. focardii inter-tropical populations must cope with higher species competition than the Antarctic strains of the same morphospecies.

The paucity of suitable test organisms prevented the investigation of the ecological role of focardin and epoxyfocardin, but the high bioactivity of the latter against predacious ciliates such as L. lamella and L. cygnus suggests a defensive role for these substances. This would indicate that ciliates are in strong competition even in Antarctica. Of the two diterpenoids, focardin (45) showed the least cytotoxic activity. E. focardii showed a significant cytotoxic activity vs. various ciliate morphospecies, including congeneric species of the ciliate Litonotus, a notorious predator of Euplotes specimens. Due to difficulties in accomplishing long lasting experiments using psychrophilic ciliate strains (growing at temperatures close to water-freezing), after 24 h of Litonotus predaction activity, the disappearance of Euplotes representatives of different morphospecies in the sample community was significantly larger than that observed when the sample ciliate community included specimens of E. focardi. Toxic activity counteracting predation by Litonotus may be guessed for the substances produced by E. focardii. Focardin’s low bioactivity and low abundance in E. focardii infers a minor role for this metabolite in nature.