Synthesis of Marine α-Methoxylated Fatty Acid Analogs that Effectively Inhibit the Topoisomerase IB from Leishmania donovani with a Mechanism Different from that of Camptothecin

Sponges biosynthesize α-methoxylated fatty acids with unusual biophysical and biological properties and in some cases they display enhanced anticancer activities. However, the antiprotozoal properties of the α-methoxylated fatty acids have been less studied. In this work, we describe the total synthesis of (5Z,9Z)-(±)-2-methoxy-5,9-eicosadienoic acid (1) and its acetylenic analog (±)-2-methoxy-5,9-eicosadiynoic acid (2), and report that they inhibit (EC50 values between 31 and 22 µM) the Leishmania donovani DNA topoisomerase IB enzyme (LdTopIB). The inhibition of LdTopIB (EC50 = 53 µM) by the acid (±)-2-methoxy-6-icosynoic acid (12) was studied as well. The potency of LdTopIB inhibition followed the trend 2 > 1 > 12, indicating that the effectiveness of inhibition depends on the degree of unsaturation. All of the studied α-methoxylated fatty acids failed to inhibit the human topoisomerase IB enzyme (hTopIB) at 100 µM. However, the α-methoxylated fatty acids were capable of inhibiting an active but truncated LdTopIB with which camptothecin (CPT) cannot interact suggesting that the methoxylated fatty acids inhibit LdTopIB with a mechanism different from that of CPT. The diunsaturated fatty acids displayed low cytotoxicity towards Leishmania infantum promastigotes (EC50 values between 260 and 240 µM), but 12 displayed a better cytotoxicity towards Leishmania donovani promastigotes (EC50 = 100 µM) and a better therapeutic index.


Introduction
A selected group of marine sponges, such as Calyx podatypa, Tropsentia roquensis or Higginsia tethyoides, biosynthesize unusual α-methoxylated fatty acids with saturated, monounsaturated, and diunsaturated alkyl chains [1]. The most ubiquitous monounsaturated α-methoxylated fatty acids in the phospholipids of sponges are those with a Δ 6 double bond, while among the diunsaturated α-methoxylated fatty acids the Δ 5,9 double bonds predominate akin with the propensity of sponges to biosynthesize very long-chain Δ 5,9 fatty acids [2,3]. More recently, the anticancer activity of this interesting group of fatty acids has received renewed attention, and there is substantial evidence indicating that α-methoxylation increases the anticancer properties of fatty acids by probably decreasing the fatty acid critical micelle concentration (CMC) [4].
The antiprotozoal activity of the marine α-methoxylated fatty acids has been less explored, but some interesting findings are now beginning to emerge, in particular promising inhibitory results against the DNA topoisomerase IB (LdTopIB) from Leishmania donovani, the key parasite responsible for visceral leishmaniasis. LdTopIB has a number of distinctive features that makes it a perfect therapeutic target. Unlike its human homologue of monomeric nature, the enzyme is a heterodimer and LdTopIB is comprised of two subunits encoded by different genes, which are independently regulated [5]. When LdTopIB is compared to the corresponding human enzyme (hTopIB), a low level of conservation of residues that comprise both proteins is observed. If sequence alignments between the two subunits of LdTopIB and hTopIB are performed, it is found that the degree of similarity between the largest subunit of LdTopIB and hTopIB is very low (45%) and even much lower (29%) when the small subunits are compared. These differences are important because it is just in these non-conserved regions where we find the residues responsible for drug sensitivity. A recent drug development effort targets these non-conserved regions by finding new drugs that can interfere with LdTopIB without harming human cells. Camptothecin (CPT) and some of its derivatives such as gimatecan, topotecan, and irinotecan, to name just a few, have been investigated as LdTopIB poisons [6].
While the saturated α-methoxylated fatty acids do not inhibit LdTopIB, the monounsaturated α-methoxylated Δ 6 fatty acids do display interesting inhibition of the enzyme. For example, our group recently synthesized the (±)-2-methoxy-6Z-heptadecenoic acid, a naturally occurring fatty acid from the sponge Calyx podatypa, and showed that it inhibits LdTopIB with an EC 50 of 41 ± 6 µM [7]. Moreover, we found that the synthetic alkynoic analog (±)-2-OMe-6-heptadecynoic acid displayed a better inhibition of the enzyme with an EC 50 of 17 ± 1 µM [7]. These findings led us to conclude that an alkynoic α-methoxylated fatty acid could be a more effective inhibitor of LdTopIB than an alkenoic α-methoxylated fatty acid provided that the carbon atoms in the acyl chain remain constant [7].
The Caribbean sponge Erylus goffrilleri is unusual in the sense that it biosynthesizes α-methoxylated Δ 5,9 fatty acids [2]. Among these compounds, the (5Z,9Z)-(±)-2-methoxy-5, 9-eicosadienoic acid (1) was identified in the phospholipids of E. goffrilleri together with other shorter-chain α-methoxylated Δ 5,9 fatty acids [2]. Due to the low natural abundance of these α-methoxylated fatty acids in the sponge (ca. 0.3%), it was not possible to study their antiprotozoal or other related biological properties in the absence of a synthetic methodology that could provide enough quantities of fatty acids for biological screenings. In addition, the complete characterization of 1 was not possible. Therefore, in the present work we present the first total synthesis of 1 utilizing our previously developed methodology for this type of fatty acids [8], and show that 1 inhibits LdTopIB with a mechanism different from that of CPT. In addition, taking advantage of the same synthetic route developed for 1, and based on previous findings that alkynoic fatty acids are better inhibitors of LdTopIB than alkenoic fatty acids [7], the unnatural analog (±)-2-methoxy-5,9-eicosadiynoic acid (2) was also synthesized expecting 2 to display a better inhibition of LdTopIB than 1.

Synthesis of the α-Methoxylated Fatty Acids 1 and 2
The synthesis of 1 followed the synthetic strategy previously developed in our group for this type of fatty acids and it is shown in Scheme 1 [8]. The synthesis started with the versatile starting material 1,5-hexadiyne (50% in pentane), which was coupled with 2-(2-bromoethyl)-1,3-dioxolane using n-BuLi in THF-HMPA at −78 °C, resulting in a 61% yield of the desired dioxolane 3 (Scheme 1). The next step called for coupling of dioxolane 3 with 1-bromodecane using similar reaction conditions as the first step resulting in the dioxolane 4 in a 69% yield. In order to install the correct stererochemistry for the two cis double bonds required for the final product, dioxolane 4 was hydrogenated under Lindlar conditions in hexane affording 5 (82% yield). The 100% cis stereochemistry for the two double bonds in 5 was confirmed by 13 C NMR spectroscopy and capillary gas chromatography-mass spectrometry (GC-MS). In 13 C NMR spectroscopy, the allylic carbon resonances are strongly dependent upon the stereochemistry of the adjacent double bonds and there is a significant difference (around 5 ppm) between the allylic carbon resonances (around 32 ppm) adjacent to a trans double bond and the allylic carbon resonances (around 27 ppm) adjacent to a cis double bond [9]. With the GC-MS analysis we were able to confirm that we only obtained the cis, cis product and no other combination of double bond stereochemical alternatives such as, for example, cis, trans. Removal of the dioxolane in 5 was most conveniently achieved under acidic conditions (HCl) using acetone/water (1:1) as solvent and heating at 60 °C for 24 h, resulting in a 64% yield of dienal 6. Aldehyde 6 was then reacted with trimethylsilyl cyanide (TMSCN) in dichloromethane and catalytic amounts of triethylamine at 0 °C, yielding the nitrile 7 in an 82% yield. Nitrile 7 was then transformed into the desired methyl ester 8 in two subsequent steps. First, acid hydrolysis of 7 in 2-methyltetrahydrofuran (2-MeTHF) at 60 °C for 24 h afforded the (5Z,9Z)-(±)-2-hydroxy-5,9-eicosadienoic acid, which was esterified in HCl/methanol, without isolation, resulting in the desired methyl ester 8 in a combined 49% yield for the last two steps. The final acid 1 was obtained in two more steps. First, methylation of 8 with sodium hydride and methyl iodide in THF, afforded the methyl (5Z,9Z)-(±)-2-methoxy-5, 9-eicosadienoate and after saponification in 1M KOH/ethanol and final purification the expected acid 1 was obtained in a combined 67% yield for the last two steps. The overall yield for this nine-step synthetic sequence was 5.9%. A GC-MS co-injection of methyl (5Z,9Z)-(±)-2-methoxy-5, 9-eicosadienoate with the fatty acid methyl ester (FAME) mixture from E. goffrilleri confirmed that we synthesized the same α-methoxylated dienoic fatty acid as the naturally occurring fatty acid, thus corroborating the structure of the natural fatty acid as well as the stereochemistry of the two cis double bonds [2]. We are also reporting, for the first time, the complete spectral data for 1. Taking advantage of the previous synthetic sequence, acid 2 was also prepared by utilizing dioxolane 4 as the key intermediate. Removal of the dioxolane in 4 was also possible under acidic conditions (HCl) using acetone/water (1:1) as solvent and heating at 60 °C for 24 h, resulting in an 85% yield of diynal 9. Similar to the synthetic procedure described above, diynal 9 was reacted with TMSCN in dichloromethane and catalytic amounts of triethylamine at 0 °C, resulting in a 94% yield of nitrile 10. Diynonitrile 10 was then transformed into the desired methyl ester 11 in two steps. First, acid hydrolysis of 10 in 2-MeTHF at 60 °C for 24 h afforded the (±)-2-hydroxy-5,9-eicosadiynoic acid, which was esterified in HCl/methanol, without isolation, resulting in the desired methyl ester 11 in a combined 61% yield for the last two steps. Methylation of 11 with sodium hydride and methyl iodide in THF afforded the methyl (±)-2-methoxy-5,9-eicosadiynoate, and after saponification in 1M KOH/ethanol and final purification, the desired acid 2 was obtained in a combined 55% yield for the last two steps. The overall yield for this eight-step synthetic sequence was 11.2%.

Inhibition of LdTopIB by Acids 1 and 2
Based on our previous results with the α-methoxylated Δ 6 fatty acids [7], we concentrated on the inhibition of LdTopIB by acids 1 and 2 so as to be able to compare the effect of a Δ 6 unsaturation versus a Δ 5,9 diunsaturation on the degree of inhibition of LdTopIB ( Figure 1, Table 1). As a mode of comparison, the inhibition of LdTopIB by the (±)-2-methoxy-6-icosynoic acid (12) (Figure 2) was also investigated since it has the same carbon chain length as acids 1 and 2 and its synthesis was recently described [4]. As explained in the Introduction, LdTopIB is a worthwhile target to study since recent strategies against leishmania take advantage of the structural differences between LdTopIB and hTopIB, and because the unorthodox heterodimeric TopIB of kinetoplastid parasites, such as LdTopIB, can be used for the development of novel drugs aimed at LdTopIB without interfering with the host [10]. For this reason, the inhibition of both LdTopIB and hTopIB by acids 1 and 2 was examined. As predicted, 2 was the most efficient inhibitor of LdTopIB with an EC 50 = 22 ± 1 μM followed by 1 with an EC 50 = 31 ± 2 μM and finally 12 with an EC 50 = 53 ± 3 μM (Table 1). Therefore, the effectiveness of LdTopIB inhibition followed the order 2 > 1 > 12 and the data seems to support our hypothesis that as the degree of unsaturation increases, the inhibition of LdTopIB by the fatty acid increases as well. Reaction products were resolved in agarose gel and subsequently visualized by ethidium bromide staining. The relative position of the negatively supercoiled DNA substrate is indicated by Sc, R is the relaxed DNA, whereas the ladder of relaxed DNA topoisomer bands is shown in between. Reactions were stopped with a mixture of 1% SDS and 6.1 µg of proteinase K. Lane 1 contains 0.5 µg of pSK plasmid DNA and lane 2, indicated by a 0, is 10% DMSO.  Figure 2. Structure of the (±)-2-methoxy-6-icosynoic acid (12).
In the latter experiment, the inhibition of hTopIB by acids 1 and 2 was also compared to the inhibition observed for LdTopIB and the results are shown in Table 1 and Figure 3. While all of the α-methoxylated fatty acids studied herein were able to inhibit LdTopIB at concentrations between 53 and 22 μM, they were not effective against hTopIB (EC 50 > 100 μM). These results, once again, demonstrate that it can be possible to preferentially inhibit LdTopIB without inhibiting hTopIB, a finding that could have medicinal applications. It is evident that LdTopIB is more sensitive to inhibition by the α-methoxylated fatty acids than hTopIB. To further study the mechanism by which the α-methoxylated fatty acids inhibit LdTopIB we explored if these acids inhibit the enzyme with a mechanism similar or different to that of CPT, a well-known topoisomerase I inhibitor [11]. Towards this purpose we carried an assay whereby instead of using the wild type Leishmania enzyme we used a doubly truncated enzyme [LdTopIB CPT R (LdTopIL 1−561/S175end )] from which the sections responsible for interacting with CPT were removed and yet the enzyme still retained TopIB activity [12,13]. In Figure 4, we can see that CPT (at levels of 100 µM) is not able to inhibit the activity of LdTopIB CPT R nor has it the ability to increase the intensity of the CL band. However, we can also observe in Figure 4 (lanes 4 and 5) that both acids 1 and 2 are able to completely inhibit LdTopIB CPT R as judged by the band corresponding to the supercoiled (Sc) DNA. Therefore, this experiment clearly demonstrates that acids 1 and 2 inhibit the LdTopIB-mediated DNA relaxation by a complete different mechanism as CPT. Despite the fact that the exact mechanism by which fatty acids inhibit LdTopIB still remains elusive, the best explanation to date, as we have hypothesized in previous studies, is for the α-methoxylated fatty acids to be interacting with LdTopIB by binding in a region close to the topoisomerase active site and either inhibiting the enzyme binding to DNA or blocking the cleavage reaction step [14]. 1, 2, and 12 towards L. infantum and L. donovani As a logical next step after the LdTopIB inhibitory studies, the toxicity of acids 1 and 2 towards L. infantum, which is closely related to L. donovani, was studied to determine if the inhibitory effects towards LdTopIB translates into toxicity towards L. infantum promastigotes. Both acids 1 and 2 were weakly toxic to L. infantum promastigotes with EC 50 values between 260 and 240 µM (Table 2). However, acid 2 was slightly more toxic to L. infantum than 1, a finding that seems to correlate with the LdTopIB studies. In a separate experiment, it was also found that acid 12 was more effective towards L. donovani promastigotes with an EC 50 of 100 µM. From the in vitro studies we can conclude that the α-methoxylated fatty acids 1 and 2 are weakly toxic to Leishmania promastigotes, but surprisingly, 12 displayed the best therapeutic index, as it was more toxic to L. donovani promastigotes and less toxic to murine macrophages (Table 2). Murine macrophages were chosen as reference for this study since Leishmania spp. normally infects macrophages and in macrophages is where Leishmania promastigotes are transformed into Leishmania amastigotes. From these studies we can conclude that 12 displays the best correlation between the enzyme inhibitory studies and the parasite growth inhibition data with only a two-fold difference between the enzyme inhibitory EC 50 (Table 1) and the toxicity IC 50 (Table 2).

Instrumentation
1 H NMR (300 or 500 MHz) and 13 C NMR (75 or 125 MHz) were either recorded on a Bruker DPX-300 or a Bruker DRX-500 spectrometer. 1 H NMR chemical shifts are reported with respect to internal (CH 3 ) 4 Si, 13 C NMR shifts are reported in part per million relative to CDCl 3 (77.0 ppm). GC/MS analysis were recorded at 70 eV using either a Hewlett Packard 5972A MS Chem Station or an Agilent 5975C MS Chem Station coupled to an Agilent 7890A. Both GC were equipped with a 30 m × 0.25 mm special performance capillary column (HP-5MS) of polymethyl siloxane crosslinked with 5% phenyl methylpolysiloxane. IR spectra were recorded on a Spectrum One FT-IR Spectrometer (PerkinElmer). High-resolution mass spectra data was performed at the Emory University Mass Spectrometry Center on a Thermo LTQ-FTMS using APCI as the probe.

LdTopIB and hTopIB Inhibitory Assays
The LdTopIB and hTopIB inhibitory bioassays were performed as previously described [7]. A brief description of the experimental procedure can be found as the header of Figures 1 and 3.

Comparative Inhibition of Recombinant LdTopIB by CPT, and Acids 1 and 2
The experimental procedure employed can be found as the header of Figure 4 and the details of the preparation of LdTopIB CPT R (LdTopIL 1−561/S175end ) were previously described in the literature [12,13].
3.6. Cytotoxicity of Acids 1, 2, and 12 towards L. infantum and L. donovani The experimental procedures employed to assess the toxicity of acids 1 and 2 towards L. infantum and L. donovani are the same as the ones previously described [5,15]. CPT was used as the control drug.

Conclusions
The first total synthesis of 1 was achieved in nine steps and in a 5.9% overall yield. It was possible to corroborate both the structure and the cis double bond stereochemistry for the two double bonds in 1 with that of the natural fatty acid from the sponge E. goffrilleri. In addition, full spectral data is reported, for the first time, for the natural acid 1. Both the natural acid 1 and its diynoic analogue 2 were good inhibitors of LdTopIB with a mechanism different from that of CPT. Acids 1 and 2 did not display significant toxicity towards L. infantum promastigotes, but the synthetic acid 12 displayed a better toxicity towards L. donovani promastigotes, less toxicity towards murine macrophages, and therefore, a better therapeutic index. Therefore, it can be concluded that although a Δ 5,9 diynoic α-methoxylated fatty acid displays a better inhibition of LdTopIB than a Δ 6 monoynoic α-methoxylated fatty acid with the same carbon chain length, the latter is more effective and selective towards L. donovani promastigotes.