**1. Introduction**

Phytoplankton, unicellular photosynthetic microorganisms, are ubiquitous in all aquatic environments. As primary producers, they are responsible for nearly half of the global primary production of organic carbon [1]. Photosynthesis, the process whereby energy is absorbed by pigments in algae and transformed into chemical energy, relies on the presence of energy trapping pigments. The main pigments, chlorophylls, carotenoids and phycobilins, absorb Photosynthetically Available Radiation (PAR) from 400–700 nm wavelengths [2]. However, pigments may also serve several functions including metabolic regulation, light harvesting (antenna pigments), electron donation or acceptance (in reaction centers), and photoprotection. The combination of different pigments and functions result in maximum efficiency and economy [3–6]. The kind of pigments produced and their relative proportions characterize the different phytoplankton groups. 

In recent years, high performance liquid chromatography (HPLC) has been used to estimate phytoplankton composition by identifying photosynthetic pigments. Some pigments found exclusively in particular algal classes or genera may serve as useful taxonomic markers [7–13]. Such indicator pigments are termed 'finger prints'. Pigment analyses offer a valuable technique in oceanography for mapping phytoplankton populations and monitoring their abundance and composition [14– 17]. 

Phytoplankton blooms occur naturally in coastal waters particularly during spring and summer seasons. However, a small number of microalgae are harmful, and although each individual is small, they may occur in huge numbers known as blooms [18–21]. Among the estimated phytoplankton species, about 7% (300 species) are known to produce red tides and of those, only 2% are actually harmful or toxic [22]. In marine and brackish water environments, most toxic species belong in the Dinophyceae, but also the Diatomophyceae, Haptophyceae, Raphidophyceae, and Cyanophyceae comprise toxic species [23–28]. The algal toxins may cause damage to other flora and fauna directly or they may accumulate through the food web in e.g., shellfish or finfish, thereby causing harm to their predators including humans [29–34]. Harmful algal blooms (HABs) are an ever more frequent phenomenon expanding in coastal regions on a world scale [35–38]. These have received much attention from researchers and local regulatory authorities due to their impact on the ecosystem and human health, influencing local economic issues [39]. 

Monitoring of coastal waters for harmful species is costly and labor-intensive and the possibility to recognize a potentially harmful algal species by means of chemical or biochemical analyses significantly reduces the time and costs of such monitoring. The one caveat is that the analysis, pigment or biochemical, involves a species specific marker for the HAB species in question [40]. Pigment signatures in the study of HABs have been very limited, particularly in monitoring programs [33,41,42]. 

During the summer period of 2000, ten fish mortality events occurred from unidentified causes in the Inland Bays of Delaware, USA. During the final fishkilling event of 28 August, 2000, over two million menhaden (*Brevoortia tyrranus*) perished when a bloom of an unidentified microalgal flagellate was observed [36]. This flagellate, accompanied by the presence of a potent neurotoxin, was tentatively called *Chattonella* cf. *verruculosa* since it resembled a fish killing species found in Japan thought to be of the class Raphidophyceae. Since none of the previously described Raphidophyceae completely agreed with the molecular features (18S rDNA; 16S rDNA) [43,44], further studies are underway to define its taxonomic position. 

This work describes the isolation and structural elucidation of a new pigment (**1**) found in *C. cf. verruculosa* cultures and in natural samples where this species was dominant, which has been called moraxanthin after the berry like algal cell (Figure 1). Moraxanthin, which is a new acylated form of vaucheriaxanthin (**2**), is unique to *C. cf. verruculosa*, indicating that it can be used as a marker of its presence when HPLC analyses of natural blooms are performed. 

**Figure 1.** Structures of moraxanthin (**1**) and vaucheriaxanthin (**2**).

## **2. Results and Discussion**

The chromatogram of the pigments of the *C. cf. verruculosa* culture showed a major peak (Figure 2, peak 4), whose retention time (Table 1) and UV spectrum (Table 1 and Figure 3) did not fit those of any known pigments, although the UV spectrum clearly showed the pigment to be a carotenoid. 

**Table 1.** Total pigments found in *C. cf. verruculosa* with relative retention times and specific absorption maxima. 


*Mar. Drugs* **2011**, *9*, 242–255.


The new pigment was then isolated to determine its structure by spectroscopic (MS and NMR) means. A large-scale (10 L) culture of *C. cf. verruculosa* was grown, and harvested by continuous flow centrifugation into an algal pellet and supernatant. The algal pellet (4 g) was extracted exhaustively with MeOH, and the extract was subjected to repeated HPLC separation, yielding 1.1 mg of the pure carotenoid moraxanthin (**1**). When re-injected in the same HPLC conditions as for the chromatogram in Figure 2, the isolated compound **1** showed retention time and UV spectrum identical to that of peak 4. Compound **1** showed a [M + Na]+ pseudomolecolar ion peak in the ESI high-resolution mass spectrum at *m*/*z* 779.4879, in accordance with the formula C48 H68NaO7. Compared to the C40 carotenoid skeleton, this formula contains eight additional carbon atoms. In addition, the ESI mass spectrum also contained a peak at *m*/*z* 663 (C42 H56NaO5, [M + Na ƺ C6H12 O2]+), which can be accounted for by the in-source loss of hexanoic acid. 

Most of the information used for structure elucidation came from one- and twodimensional NMR spectroscopy. The general features of the proton NMR spectrum (C6D6) resembled those of carotenoids, with several olefinic protons between Έ 6 and Έ 7, and 10 methyl singlets between Έ 1.87 and 1.08. However, one of the ten methyl signals (Έ 1.74) was part of an acetyl group, as shown by its correlation peak with the carbonyl carbon atom at Έ 169.2 in the HMBC spectrum. Other notable features of the proton NMR spectrum were (i) the AB system at Έ 5.13 and 5.07 (H2-19') of an oxymethylene group, (ii) two oxymethine protons at Έ 5.69 (H-3) and Έ 3.78 (H-3'), a methyl triplet at Έ 0.78 (H3-6'') and a methylene triplet at Έ 2.14 (H3-2''), indicative of an acyl chain and (iv) an olefinic proton singlet at Έ 6.03 (H-8), which showed correlation peaks in the HMBC spectrum with two non-protonated carbon atoms at Έ 202.1 (C-7) and 117.8 (C-6), and was therefore part of an allene system. The observed structural features were suggestive of a structure similar to vaucheriaxanthin (**2**), but a direct comparison of spectral data was hindered by the presence of the two additional acyl groups. A detailed analysis of the correlation peaks observed in the COSY, HSQC, and HMBC (Figure 4) spectra demonstrated that the planar structure of moraxanthin is indeed the same as that of vaucheriaxanthin, except for the presence of an acetyl group at position 3 and a hexanoyl group at position 19'. 

In addition to all the expected geminal and vicinal couplings, the COSY spectrum revealed several proton-proton long-range couplings. Among them, the quite large W couplings of H-2 ΅ with H-4 ΅

(2.2 Hz) and of H-2' ΅ with H-4' ΅ (1.7 Hz) indicated the 1,3-diequatorial relationship of these two pairs of protons. Furthermore, the methyl protons on the polyene system showed weak correlation peaks with the olefinic protons, arising from the usual allylic 4*J*HH couplings, but also from 6*J*HH couplings (H3-19/H-12, H3-20/H-15', H3-20'/H-15) and even one remarkable 8*J*HH coupling (H3-19/H-14). To the best of our knowledge, this is the first report of a 8*J*HH coupling in a carotenoid. 

The *E* configuration of double bonds at positions 11, 15, 7', and 11' was evident from the large *trans* coupling constant values of the relevant protons (see Table 2). The *E* configuration of double bonds at positions 9, 13, and 13' and the *Z* configuration of the double bond at position 9' were determined from the ROESY spectrum, displaying correlation peaks of H3-19 with H-11, H3-20 with H-15, H3-20' with H-15', and H2-9' with H-11'. 


The ROESY spectrum also provided information on the relative configuration of the two terminal six-membered rings (Figure 5). The allene terminus is in the chair conformation, with the two W-coupled H-2 ΅ and H-4 ΅ protons in the equatorial orientation. The large coupling constants of H-3 with the axial H-2 Ά and H-4 Ά (Table 2) showed the former proton to be axial, and therefore on the ΅ face of the ring; as a consequence, the OAc group at C-3 must be Ά. The ROESY correlation of the methyl protons H3-19 with H-2 Ά and H-4 Ά determined the axial chirality of the allene functionality as *R*. Finally, the ROESY correlation of H3-19 with H3-18 located C-18 on the Ά face of the ring, and therefore the OH group at C-5 on the ΅ face. 

As for the other terminal ring, the W coupling (1.7 Hz) of the pseudoequatorial H-2' ΅ and H-4' ΅ suggests a half-chair conformation of this ring. The *trans* relationship between the epoxide ring and the hydroxyl group was established from the ROESY correlation peaks of the two geminal methyl groups H3-16' and H3-17' with, respectively, H-3' and H-7' (Figure 5), showing that H-3' and H-7' are on opposite faces of the six-membered ring. This was confirmed by the prominent peak between the psudoaxial H-4 Ά and H3-18 in the same spectrum. The relative configuration determined for moraxanthin matches that of vaucheriaxanthin (**2**), and it may be assumed that also the absolute configuration of moraxanthin is the same as in vaucheriaxanthin. 

**Figure 6.** HPLC absorbance chromatogram of natural water sample collected at the 

fish-kill site of Torque Canal, Delaware on 28 August 2000 during a *C. cf. verruculosa* bloom. The arrow indicates the moraxanthin peak. 

To investigate the utility of using moraxanthin as a marker for the toxic alga *C. cf. verruculosa*, natural bloom samples from the fish-kill site at Torque Canal, Delaware, collected on 28 August 2000 on Whatman GFF glass fiber filters and stored at ƺ80 °C, were extracted in methanol and subjected to HPLC analysis. The HPLC chromatogram (Figure 6) definitely showed a peak for moraxanthin in the natural sample. The moraxanthin peak clearly separated from all the other pigment peaks having no overlap with other pigments. In addition, a peak with the same retention time and absorbance characteristics was present in the HPLC chromatogram from water samples collected in 2003–2007 at various sites in Delaware's Inland Bays where *C. cf. verruculosa* blooms occurred (data not shown). This shows that the HPLC analysis may provide a simple and rapid tool for detecting harmful blooms of *C. cf. verruculosa*. 

## **3. Experimental Section**
