Synthesis of an Electrophilic Keto-Tetraene 15-oxo-Lipoxin A₄ Methyl Ester via a MIDA Boronate

Abstract

15-oxo-Lipoxin A₄ (15-oxo- LXA₄) has been identified as a natural metabolite of the fatty acid signaling mediator Lipoxin A₄. Herein, we report a total synthesis of the methyl ester of 15-oxo-LXA₄ to be used in investigations of potential electrophilic bioactivity of this metabolite. The methyl ester of 15-oxo-LXA₄ was synthesized in a convergent 15 step (9 steps longest linear) sequence starting from 1-octyn-3-ol and 2-deoxy-D-ribose with Sonogashira and Suzuki cross-couplings of a MIDA boronate as key steps.

Graphic Abstract

Arachidonic acid (AA) plays a key role in cell signaling, as it has been shown to be the biosynthetic source of eicosanoids that mediate the function of leukocytes and other inflammatory responses via G-protein-coupled receptor (GPCR) ligand activity. These AA metabolites include prostaglandins, prostacyclins, thromboxanes and leukotrienes.¹ One class of AA-derived mediators termed lipoxins A₄ and B₄ ((5S,6R,15S)-trihydroxy-eicosa-7E,9E,11Z,13E-tetraenoic acid and (5S,14R,15S)-trihydroxy-6E,8Z,10E,12E-eicosatetraenoic acid, or LXA₄ and LXB₄) are trihydroxytetraenes produced by the successive lipoxygenase oxidations of AA² (5-LOX and 12/15-LOX) by stimulated leukocytes. When administered as pure synthetic compounds, these species display pleiotropic actions described³ as anti-inflammatory and pro-resolving and, because of low concentrations and rapid degradation, are difficult to detect⁴, ⁵ in vivo. They, along with the 15(R)-stereoisomers⁶ (15-epi-LXA₄), have been investigated as leads⁷ for the treatment⁸ of several chronic and acute inflammatory conditions.

In vivo lipoxin A₄ is rapidly metabolized⁹ via 15-hydroxyprostaglandin dehydrogenase¹⁰ (15-PGDH) oxidation of the 15(S)-hydroxy to 15-keto (Fig. 1), followed by reduction of the neighboring C13–14 double bond and finally reduction of the 15-keto to a hydroxyl. While the initial oxidation itself results in ablation of GPCR ligand activity,¹¹ the formation of an α ,β-unsaturated keto group in the 15-oxo intermediate suggests the potential for alternative electrophilic reactivity. Lipids containing α,β-unsaturated carbonyls have been shown to form covalent Michael adducts with thiol-containing enzymes¹² in addition to glutathione (via glutathione S-transferase¹³). Electrophilic signaling mediators typically exert effects at higher concentrations than GPCR-specific activity, but the effects can be pleiotropic and have extended kinetics. Recent work^14–16 has shown structurally similar oxo-eicosatetraeneoates (e.g. 15-oxoETEs) to modulate inflammatory responses via electrophilic signaling, and given our long-standing interest in electrophilic redox signaling^{17, 18} the de novo synthesis of a representative member of this structural family to explore this potential was encouraged.

Figure 1.

Structures of lipoxin A₄ (LXA₄) and 15-oxo-lipoxin A₄ (15-oxo-LXA₄) produced in vitro by enzymatic oxidation with 15-PGDH/NAD⁺. 15-oxo-LXA₄ is expected to contain electrophilic site(s) in conjugation to the keto group.

Multiple synthetic approaches to the parent LXA₄^19–22 and isomers^{23, 24} (as well as di- and tri-hydroxy docosapentaenoates^{25, 26} and docosahexaenoates²⁷) have been described in the literature, providing useful starting points. As we intended to utilize palladium-catalyzed coupling reactions to provide the desired (E,E,Z,E)-tetraene stereochemistry, an additional development of interest was the use of N -methylimidodiacetyl (MIDA)-protected boronates^{28, 29} as part of an iterative, modular approach to polyene natural products, which we considered as a novel approach to the tetraene core. Finally, concerns about lipophilicity and cell permeability in administering the product to biological targets prioritized the methyl ester over the free acid. With these considerations in mind a synthetic approach to 15-oxo-(5S,6R)-dihydroxy-eicosa-7E,9E,11Z,13E-tetraenoate methyl ester (15-oxo-LXA₄Me, 1 ) was designed.

In our retrosynthetic design (Fig. 2) the target was broken into two portions: an eight-carbon haloalkene ester (2) and a twelve-carbon MIDA-protected boronic ester (3) which could serve as a Suzuki coupling partner. Haloalkene ester 2 was developed from a structurally similar aldehyde³⁰ that we proposed to transform directly to an iodovinyl group through a Takai reaction.^{31, 32} The (5S,6R) diol unit could be obtained in turn from chiral pool 2-deoxy-D-ribose (4) and protected to ensure specific oxidation and retention of stereochemistry.

Figure 2.

Retrosynthetic analysis of 15-oxo-LXA₄Me (1).

Both of the interstitial conjugated E,Z double bonds of the tetraene were potentially accessible through halovinyl MIDA boronates, but limited commercial availability of the required cis-halovinyl MIDA boronate discouraged a direct double-MIDA approach. After considering similar work,²³ we proposed dien-yne 3 which could be derived from inexpensive and commercially available racemic 1-octyn-3-ol (5). While several literature precedents also took the precaution of protecting the alcohol of 5 (e.g. as a TBDMS ether) a palladium-catalyzed coupling approach did not seem to require such protection;³³ and instead the alcohol would essentially act as a protecting group for the desired carbonyl. The final steps of alkyne reduction, specific oxidation, and deprotection could then proceed after assembling and coupling the framework.

The synthesis of iodovinyl 2 (Scheme 1) commenced with protection of the diol of 2-deoxy-D-ribose (4) as an acetonide³⁴ by etherification with isopropenyl methyl ether in the presence of catalytic PPTS (pyridinium p -toluenesulfonate, 44%). The hemiacetal portion of the protected ribose 6 was then homologated³⁴ via Wittig-type conditions with methyl (triphenylphosphoranylidene) acetate and catalytic benzoic acid (86%). The resulting α ,β-unsaturated ester 7 was reduced³⁰ quantitatively with hydrogen in the presence of rhodium on alumina catalyst and subsequently oxidized to aldehyde 8 by either Swern³⁰ (oxalyl chloride, DMSO) or Parikh-Doering^{35, 36} (sulfur trioxide-pyridine complex, DMSO) conditions in similar yields (76%). Finally under Takai conditions^{31, 37} aldehyde 8 was transformed to iodovinyl ester 2 in reasonable yield (63%, Z :E 1:5). The Z :E product ratio was continued through the next step although it could be separated chromatographically. Note that the analogous bromovinyl Takai conditions were not investigated.

Scheme 1: Reagents and conditions:

(a) isopropenyl methyl ether, cat. PPTS, EtOAc, rt, 44%; (b) Ph₃P=CHCO₂Me, cat. BzOH, MeCN, reflux, 86%; (c) cat. Rhodium (5w/w% Al₂O₃), 1 atm H₂, MeOAc, rt, quant.; (d) (COCl)₂, TEA, DMSO/ CH₂Cl₂, −78°C to rt or SO₃-Pyr, TEA, DMSO/CH₂Cl₂, 0°C to rt, 76%; (e) CHI₃, CrCl₂, rt, 1,4-dioxane-THF, 63% (E:Z ratio 5:1)

Racemic 1-octyn-3-ol (5) was brominated³⁸ (Scheme 2) using N -bromosuccinimide in acetone with catalytic silver nitrate. Several precedents suggested that lithium aluminum anhydride in combination with aluminum chloride would reduce the bromo-alkyne; however, this combination was unreactive in our hands. Instead the less-common method³⁹ of DIBAL with aluminum chloride gave virtually quantitative yields (80% over two steps). The resulting bromo-vinyl allylic alcohol 10 was then coupled²⁹ by standard Sonogashira methods to trimethysilylacetylene. Afterwards, the TMS group was deprotected⁴⁰ with potassium carbonate (88% over two steps) before repeating the Sonogashira protocol with bromo (or iodo-)vinyl MIDA boronate 12 to afford the desired dien-yne 3 in reasonable yield (67%).

Scheme 2: Reagents and conditions:

(a) NBS, 10% AgNO₃, acetone, 0°C to rt, 3h; (b) DIBAL (4 eq.), AlCl₃ (2 eq.), ether, 0°C to rt, 80% for 2 steps; (c) TMSA, 5% Pd(PPh₃)₄, 10% CuI, piperidine, THF, rt; (d) K₂CO₃, MeOH, rt, 88% for 2 steps; (e) 12 , 5% Pd(PPh₃)₄, 10% CuI, piperidine, THF, rt, 67%. MIDA = (O₂CCH₂)₂NMe

The coupling of substrates 2 and 3 (Scheme 3) proceeded after first hydrolyzing the MIDA protecting group of 3 with sodium hydroxide to afford the boronic acid, which was coupled to iodovinyl 2 by Suzuki methods²⁹ in 64% yield. The resultant trien-yne 13 retained the Z :E isomer ratio of 2 (e.g. 1:5; 3 was apparently stereospecific) but could be isomerized²² to an all-trans configuration (clean by 1H) through exposure to catalytic iodine in benzene. Trien-yne 13 was subsequently reduced to tetraene 14 via a solvolytic zinc-based reduction (75%). The zinc used to produce the Zn(Cu/Ag) alloy was activated^{23, 41} immediately prior to reaction by cleaning of the zinc surface with acid followed by reaction with cupric acetate and silver nitrate. While stereoselective, this method was not consistently scalable in our hands, and attempts at larger scale were met with noticeably variable yields. Note that intermediate 14, as an acetonide-protected methyl ester of 15-rac-Lipoxin A₄, provides a convenient route to the corresponding diastereomeric LXA₄.

Scheme 3: Reagents and conditions:

(a) (i) 3, 1 M NaOH; (ii) 2, 5% Pd(OAc)₂, 10% XPhos, NaOH, THF, rt,; (b) cat. I₂, PhH, rt, 64% for 2 steps (Z:E 1:5 to clean E ); (c) Zn(Cu/Ag) alloy, MeOH-water, rt, 75%; (d) Dess-Martin Periodinane, CH₂Cl₂, rt, 82%; (e) 2M aq. HCl, MeOH, rt, 63%

Oxidation of the unprotected alcohol proceeded cleanly using Dess-Martin periodinane (82%) and the acetonide protecting group was removed (HCl in methanol, 63%) to yield 1 with only minimal formation of a cyclized δ -lactone side product (removed by chromatography). The spectral data of the keto- tetraene compounds 1 and 15 confirmed our expectations: the UV-Vis spectra was red-shifted from 312 nm to 345 nm, and the ¹H NMR spectrum of the vinyl region showed protons shifted downfield (δ6.69 to δ7.64 ppm), indicating significant deshielding characteristic of electron withdrawal. However, a general instability of the tetraenone frequently led to impurities in the post-oxidation steps, and products from 14 on were thus treated with minimal heating and stored as methanol solutions, cold, in reduced-light conditions and under argon or nitrogen to minimize intermolecular reaction. Indeed, a prior synthesis of 5-oxoETE noted^{42, 43} the facile cis -trans isomerization of the keto-diene, and a recent report⁴⁴ on a keto-triene described the inherent instability of an analogous Z-isomer in an electron deficient polyene as being an electronic rather than steric effect. This relative configurational stability may be further investigated as additional electrophilic polyenes are synthesized.

In conclusion we have developed a successful synthesis of 15-oxo-lipoxin A₄ methyl ester in preparation to evaluate the potential electrophilic reactivity and biological signaling associated with this and related α,β-unsaturated ketones. Our methods obtained a 3.7% overall final yield from a convergent 15 step (9 steps longest linear) sequence, and our use of a MIDA boronate strategy simplified our palladium-catalyzed coupling approach to the tetraene motif.

Highlights.

First total synthesis of keto-tetraene 15-oxo-Lipoxin A4 has been achieved.

Synthetic approach: iterative palladium-catalyzed coupling using a MIDA boronate

Dess-Martin oxidation of 15-position hydroxyl to electrophilic keto

Key steps: Sonogashira, Suzuki couplings; Zinc(Copper/Silver) reduction of trien-yne