In order to determine whether E. gracilis produces additional euglenatides or euglenatide biosynthetic intermediates, we cultured E. gracilis using similar methods to those reported by Aldholmi et al. [6]; that is, we used minimal medium in which the carbon source was sodium acetate and the nitrogen source was glutamate, and we cultured the algae under photosynthetic conditions. In our hands, euglenatide B was more than tenfold more abundant than euglenatides A and C, euglenatide D was only minorly present, and euglenatide E was not detectable. Using molecular networking and LC–MS analysis of E. gracilis crude extracts, we noted that E. gracilis produces not only the euglenatides, but also a corresponding set of euglenatide analogs that we named euglenatides A2-C2 (Fig. 2) [10, 11]. These analogs have the same exact masses and MS fragmentation patterns as euglenatides A-C (Fig. 3, Fig. S1), but are slightly more polar based on their retention times (Fig. 2b). Unlike euglenatides A–C, which have a UV–Vis chromophore characteristic of a conjugated triene, euglenatides A2–C2 do not absorb at 280 nm and do not possess a conjugated triene (Fig. 2b, c). Each euglenatide is produced at roughly similar levels as its corresponding euglenatide analog. We have not been able to enhance the production of the euglenatide analogs relative to the euglenatides by harvesting the E. gracilis culture at a particular time point or utilizing a particular nitrogen source (glutamate versus asparagine).

E. gracilis produces the euglenatide analogs, euglenatides A2, B2, and C2. a Molecular networking analysis of crude E. gracilis extracts analyzed in positive (ESI+) mode. The orange node represents the [M-CH4O + H]+ ion for A2, the yellow nodes represent the [M-CH4O + H]+, [M-CH6O2 + H]+, and fragment ions of B, the green nodes represent the [M + H]+, [M-CH4O + H]+ and fragment ions of B2, the pink node represents the [M-CH4O + H]+ ion of C, and the purple node represents the [M-CH4O + H]+ ion of D. b Extracted ion chromatograms (EICs) of euglenatides A–C and their analogs A2–C2 analyzed in negative (ESI-) mode, along with the absorbance at 280 nm. The fraction analyzed here has roughly equal amounts of the different euglenatides and their analogs. However, in the crude extracts, euglenatide B and B2 are much more abundant than euglenatides A, A2, C, and C2. c UV–Vis spectrum of euglenatide B (top) and euglenatide B2 (bottom)

MS–MS spectra of the euglenatides and their analogs. Mirrored MS–MS spectra for euglenatides A and A2 (a), euglenatides B and B2 (b), euglenatides C and C2 (c)

To further characterize the chemical structures of the euglenatide analogs, we purified the most abundant one, euglenatide B2, as well as euglenatide B for comparison purposes. Chromatographic fractionation of defatted methanolic extracts of E. gracilis by reversed phase C18 flash chromatography, LH-20 Sephadex chromatography, and semipreparative HPLC allowed the purification of the two euglenatides (Fig. S2). Analysis of the 1D and 2D NMR (DMSO-d6) data for euglenatide B (Table 1; Table S1 and Figs. S3, S4) in comparison to literature [6] confirmed its identification (Fig. 1).

Comparison of the 1D NMR (DMSO-d6) data for euglenatide B2 with that of euglenatide B disclosed common peptide and polyketide substructures, with the significant differences being the presence of two extra allylic methylene resonances in euglenatide B2 at C-29 (δ2H 2.10, m; δC 32.2) and C-30 (δ2H 2.04, m; δC 31.7), accompanied by a change in the pattern of olefinic methine resonances (Fig. 4, Table 1; Table S2, and Figs. S5–S13). These NMR changes in euglenatide B2 suggested that one double bond (Δ31) is separated from the diene (Δ25 and Δ27) by two methylenes at C-29 and C-30, which explains the absence of the conjugated triene chromophore in the UV–Vis spectrum. Diagnostic 2D NMR correlations (shown in Fig. 4) confirmed the double bond shift and the molecular connectivity of euglenatide B2, while NMR similarities with euglenatide B, including comparable J values, and biosynthetic considerations suggested a common absolute configuration. Finally, Double Quantum Filtered (DQF)-COSY NMR analysis of euglenatide B2 allowed the assignment of the Δ31 double bond geometry as E (trans) based on a coupling constant (J31,32) of 16.2 Hz (Figs. S10, S11). This assignment is further supported with the upfield chemical shifts of the allylic methylenes at C-30 (δC 31.7) and C-33 (δC 31.9) [12,13,14,15]. Thus, we conclude that the double bond at C-29 in euglenatide B has been shifted by two carbons to C-31 in euglenatide B2 (Fig. 5a).

Key 2D NMR correlations for the euglenatide B analog, euglenatide B2

Chemical structures of euglenatides A2, B2, and C2 (a) and the antiproliferative activity of euglenatides B and B2 against A549 cells (b)

Similarly, our data suggest that euglenatides A2 and C2 are exactly the same as euglenatides A and C, but one of the double bonds in the triene has been shifted by two carbons (Fig. 5a). Specifically, our data show that (1) euglenatides A-C and A2-C2 have the same exact masses and MS/MS fragmentation patterns (Fig. 3, Fig. S1), (2) the conjugated triene chromophore in euglenatides A-C is absent in A2-C2 (Fig. 2b, c), and (3) the euglenatides and their analogs have regularly spaced elution times by LC–MS (Fig. 2b). Several trials were made to purify euglenatides A, A2, C and C2 for NMR analysis, but unfortunately their low abundance and co-elution with other metabolites hindered their purification.

Previously, it was shown that euglenatide B had antiproliferative activity in the mid-nanomolar range against mammalian cancer cell lines, including A549 lung adenocarcinoma cells. To test the antiproliferative activity of the euglenatide analogs and thus the importance of the conjugated triene motif, we tested the activity of both euglenatides B and B2 against A549 cells (Fig. 5b). Our data confirm the potency of euglenatide B in the mid-nanomolar range, with an IC50 of 358 nM, in comparison to the previously reported IC50 of 773 nM. The data also show that euglenatide B2 is about tenfold less potent than euglenatide B, with an IC50 of 2987 nM. Thus, the conjugated triene is quite important for the activity of the euglenatides, and this information should inform future efforts to characterize the structure–activity relationships of the euglenatides.