First Generation Route to the Isatis indigotica-Derived Alkaloids 

Get Complete Project Material File(s) Now! »

CHAPTER FOUR BIOMIMETIC TOTAL SYNTHESIS OF THE ISATIS INDIGOTICA-DERIVED ALKALOIDS

Second Generation Approach to  24a and 24b using a Bioinspired

Thio-Diels-Alder Reaction
The Thio-Diels-Alder Reaction
The Thio-Diels-Alder Reaction in the Biosynthesis of 24a and 24b

Although two biosyntheses were proposed for the Isatis indigotica derived alkaloids 24a and 24b (section 1.1.4.2),⁶ the recent isolation of insatindigothiadiazoles A-D (33a–d) in a 2:4:1:2 ratio from the same source,⁷ supports the proposed biosynthesis by a late-stage thio-Diels-Alder reaction which is discussed in detail henceforth (Scheme 78).
Epiprogoitrin (25a) and progoitrin (25b) are present in I. indigotica in a 2:1 ratio and are the putative precursors for the biosynthesis.¹⁹⁸ The myrosinase-catalysed hydrolysis of 25a and 25b is known to liberate thiohydroximate-O-sulfonate 27,¹⁹⁸ which subsequently breaks down into nitrile 29 and imidothioate 30.¹⁹⁹ Nitrile 29a/b then undergoes heterocyclisation with imidothioate 30a/b to form insatindigothiadiazoles 33a-d in a 2:4:1:2 ratio of stereoisomers, which is in line with the isolation report. A selective monodehydration of insatindigothiadiazoles 33a-d then generates diene 31, which undergoes a thio-Diels-Alder reaction with 3-thioisatin 32, [proposed to be derived from glucobrassicin (26)],⁶ followed by double bond isomerisation to give 24a and 24b. Although the 2:1 ratio of enantiomers 24a and 24b suggests that the 2 »’ stereocentre is derived from the 2:1 mixture of glucosinolates 25a and 25b, there is some uncertainty surrounding the diastereoselectivity of the thio-Diels-Alder cycloaddition. The original report suggested that the C-2′ stereocentre in diene 31 is too distant from the forming spiro-centre to exert any diastereocontrol in this cycloaddition. As such, diastereomer 144 is also proposed to exist in Isatis indigotica (presumably as a 2:1 ratio of 144a:144b).

The Thio-Diels-Alder Cycloaddition Reaction

Introduction

The [4+2] cycloaddition of a diene and dienophile to form a cyclohexene product is known as the Diels-Alder reaction, however if a sulfur atom is present in the diene or dienophile the reaction is known as a thio-Diels-Alder reaction, and a dihydrothiopyran 305 is formed (Scheme 79, A). Thia-1,3-butadienes 306 predominantly react with electron-deficient dienophiles in normal electron demand Diels-Alder reactions.¹⁹³ In contrast, thiocarbonyl compounds 307 have been referred to as having ‘superdieneophilic’ character due to their ability to undergo normal and inverse electron demand interactions with dienes at temperatures as low as -78 °C.^193,200 To the best of our knowledge, examples of the thioDiels-Alder reaction occurring in Nature are limited to the assembly of the vinyldithiins (308 and 309) from thioacrolein (310) in Allium sativum (Scheme 79, B),^201–203 making it an extremely rare reaction in Nature.

Mechanism

Thio-Diels-Alder reactions are [4+2] cycloadditions that proceed via concerted transition state 311 (Scheme 80, A).⁹¹ The frontier molecular orbital (FMO) theory of Diels-Alder reactions explains that normal electron-demand Diels-Alder reactions usually occur between the HOMO of an electron rich diene and the LUMO of an electron poor dienophile, while inverse demand occurs between the LUMO of an electron poor diene and the HOMO of an electron rich dienophile (Scheme 80, B).^200,204 The ability of thiocarbonyl compounds 312 to undergo both normal and inverse electron demand interactions is attributed to the characteristically small gap in frontier orbital energies of the thiocarbonyl dienophile, allowing orbital interaction with either the HOMO or the LUMO of the diene.²⁰⁵

Applications of Thio-Diels-Alder Reactions in the Synthesis of Biologically Active Molecules

Vedejs and co-workers employed a thio-Diels-Alder reaction in the total synthesis of (R)-otonecine (313), which is thought to exist as a tautomeric mixture of pyrrolizidine 313a and azocane 313b (Scheme 81, A).²⁰⁶ Thioketone 314, generated in situ by photolytic fragmentation of sulfide (±)-315, underwent thio-Diels-Alder reaction with Danishefsky’s diene 316 to form dihydrothiopyran (±)-317. Treatment of (±)-317 with TFA then generated bicyclic thioaminal (±)-318, which was converted to (±)-otonecine (313) in an additional 14 steps. A thio-Diels-Alder reaction was also employed in the synthesis of potassium channel activator, trans-aprikalim (319) (Scheme 81, B).²⁰⁷ A Lewis-acid mediated thio-Diels-Alder reaction of dithioester 320 with butadiene (304) formed the desired dihydrothiopyran 321 along with thiopyran 322 (3:1) resulting from elimination of the methylsulfanyl moiety. An additional 5 steps were then used to convert dihydrothiopyran 321 to trans-aprikalim (319)

Revised Synthesis of Natural 24a/b and Putative 144a/b using a Bioinspired Thio-Diels-Alder Reaction

It was thought that model diene 323 could be used to examine the viability of the biomimetic thio-Diels-Alder reaction with 3-thioisatin 32 (Scheme 82). If successful, the isomerised cycloadduct (±)-324 could be subjected to a Suzuki-Miyaura sp³-sp² coupling with Molander salt 325 to form (±)-326. The same deprotection, oxidation and Grignard addition steps previously outlined in Scheme 67, Section 3.5.1.1, would transform (±)-326 to the natural product (±)-24 and putative natural product (±)-144.

Synthesis of Diene 323

Retrosynthetic Analysis of Diene 323

It was thought that 5-butadienyl-1,2,4-thiadiazole 323 could be prepared using a palladium(0)-catalysed Suzuki-Miyaura sp²-sp² coupling between dienylboronate 327 and commercially available 3-bromo-5-chloro-1,2,4-thiadiazole (328) (Scheme 83). Dienylboronate 327 could in turn be accessed from acrolein (329) and diborylmethane 330 using a known Boron-Wittig procedure.²⁰⁸
Preparation of Dienylboronate 327
Dienylboronate 327 was prepared from commercially available diborylmethane 330 and acrolein (329) using a Boron-Wittig reaction²⁰⁸ (Scheme 84, A). Lithium tetramethylpiperidide (LiTMP) (prepared in situ from n-butyllithium and 2,2,6,6-tetramethylpiperidine) was used to lithiate diborylmethane 330, which, upon reaction with acrolein (329) forms the sterically favoured trans-oxaboretane intermediate 331 (Scheme 84, B). Rearrangement of trans-oxaboretane 331 then forms E-dienylboronate 327 along with 1,3,2-dioxaborolan-2-ol 332 following aqueous work-up.

The Suzuki-Miyaura Reaction

Introduction

The Suzuki-Miyaura cross-coupling reaction is one of the most widely used reactions in organic chemistry and is broadly defined as the palladium(0)-catalysed cross-coupling of alkenyl or aryl halides 234 with boronic acids or esters 333 (Scheme 85, A). This reaction was discovered by Akira Suzuki and Norio Miyaura in 1979 (Scheme 85, B),²¹⁰ and as previously mentioned, earned Suzuki the 2010 Nobel Prize in Chemistry.¹⁷⁵ The broad application of the Suzuki-Miyaura cross-coupling reaction is attributed to the stability and accessibility of the organoboron reagents, the mild and scalable reaction conditions, and the high E-selectivity for the diene products.^175,211

Mechanism

The mechanism of the Suzuki-Miyaura cross-coupling reaction begins with oxidative addition of the palladium(0) catalyst 334 to the aryl or vinylhalide 234 to form palladium(II) complex 335 (Scheme 86). Base mediated transmetallation of 335 with boronic acid/ester 333 is the rate-determining step, and requires ligand dissociation for the formation of bisaryl/alkenylpalladium(II) species 336. Reductive elimination of 336 and ligand reassociation then regenerates the palladium(0) catalyst 334 and forms the product 337.

Applications of the Suzuki-Miyaura Cross-Coupling of Vinyl Boronates in Natural Product Total Synthesis

A Suzuki-Miyaura cross-coupling of Z-vinyl boronate 338 with alkenyl iodide 339 was employed as the final step in the asymmetric synthesis of (‒)-exiguolide (340) (Scheme 87, A).²¹² The mild and rapid Suzuki-Miyaura reaction conditions (room temperature, 50 minutes) were likely enabled by the triphenylarsine ligand, which reportedly facilitates rapid reaction times due to the ease of ligand dissociation during the rate-limiting transmetallation step.²¹³ The total synthesis of (‒)-exiguolide (340) enabled the biological testing of (‒)-340 against a panel of 39 human cancer cell lines, which found (‒)-exiguolide (340) to exhibit potent antiproliferative in vitro activity.²¹² E-Vinyl boronates (S)-341a and (R)-341b were used in a Suzuki-Miyaura coupling with aryl bromide 342 to yield chiral alkenes (S)-343a and (R)-343b respectively. These key intermediates were then converted to the tetrahydroisoquinoline alkaloids (S)-trolline (344a), (R)-crispin A (345), and (R)-oleracein E (344b) (Scheme 87, B).²¹⁴ An extremely mild Suzuki-Miyaura coupling of alkenyl iodide 346 with E-vinyl boronate 347 was employed by Shimizu and co-workers to form protected (‒)-spirofungin A (348) in excellent yield (Scheme 87, C).²¹⁵ Thallium ethanoate was likely employed as a base in this key reaction since formation of insoluble thallium(I) iodide assists the transmetallation step of the Suzuki-Miyaura coupling and thus improves the reaction rate.²¹⁶ Suzuki-Miyaura product 348 then underwent simple deprotection steps to complete the asymmetric total synthesis of the polyketide-type antifungal antibiotic (‒)-349.

Suzuki-Miyaura  Cross-Coupling  of  Dienylboronate  327 with  3-Bromo-5-chloro-1,2,4-thiadiazole (328)

With dienylboronate 327 in hand, we turned our attention to the preparation of 5-dienyl-1,2,4-thiadiazole 323. A Suzuki-Miyaura coupling of 327 with commercially available 3-bromo-5-chloro-1,2,4-thiadiazole (328) was envisaged for this transformation since 3,5-dihalo-1,2,4-thiadiazoles are known to undergo selective Suzuki-Miyaura couplings at C-_5.125,145,217 Dienylboronate 327 and thiadiazole 328 were subjected to several catalytic systems known to effect the Suzuki-Miyaura coupling of 5-halo-1,2,4-thiadiazoles,^{145,146,150,218} with use of carbonate bases and non-nucleophilic solvents to prevent SNAr side-reactions (Table 11, entries 1-3). Neither palladium(II) acetate and tri-tert-butylphosphonium tetrafluoroborate or PdCl2(dppf) led to the formation of the desired product 323 with only starting materials 327 and 328 visible by TLC after 44 or 24 hours at 100 °C respectively (entries 1 and 2). Trace 323 was thought to be obtained using freshly prepared Pd(PPh3)4 (entry 3), however mainly starting material was observable by TLC after 44.5 hours, possibly due to degradation of the air sensitive catalyst. Use of air stable palladium(II) acetate and triphenylphosphine with K2CO3 in acetonitrile gratifyingly yielded the desired diene 323 in 25% yield (entry 4), providing a good platform for further optimisation. K2CO3 was identified as the best base for the desired Suzuki-Miyaura coupling of 327 with 328 since changing the base to Cs2CO3, Na2CO3 or K3PO4 failed in increase the yield of 323 (entries 5-7). Screening a number of solvents (toluene, dioxane, aqueous dioxane, DMF and THF) entries (8-12) found that the yield could be improved to 33% using 9:1 dioxane/H2O (entry 10). The bidentate phosphine ligands SPhos and XPhos (entries 13 and 14) were then trialled and SPhos was found to improve the yield of 323 to 44%. However the yield of 323 was increased further to 47% with use of bis[di-tert-butyl(4-dimethylaminophenyl)phosphine]dichloropalladium(II) [Pd(amphos)Cl2], a catalyst known to be particularly affective for the Suzuki couplings of 3-bromo-5-chloro-1,2,4-thiadiazole (328) (entry 15).¹²⁵ When the scale was increased from 0.0555 mmol to 0.280 mmol, and the reaction was conducted in a sealed tube, the optimised conditions [Pd(amphos)Cl2, K2CO3, aqueous dioxane, 80 °C] gratifyingly afforded diene 323 in 51% yield (entry 16). Diene 323 was found to form an insoluble polymer when stored neat in the fridge for several days, and as such had to be freshly prepared for subsequent reactions.
Diene 323 was too unstable under electrospray ionisation-HRMS (ESI-HRMS) conditions for the desired molecular ion to be identified. However low resolution gas chromatography-MS (GC-MS) of 323 identified molecular ion peaks at 215.9 and 217.9 (in approximately 1:1 ratio) corresponding to the ⁷⁹Br and ⁸¹Br isotopes of 323 (C6H5BrN2S + H⁺). This indicated that the Suzuki coupling had occurred at the desired C-5 position of 3-bromo-5-chloro-1,2,4-thiadiazole (328).
The chemoselectivity of cross-coupling reactions typically favours palladium(0) insertion into C-Br bonds over C-Cl bonds due to the markedly lower bond dissociation energies (BDEs) of C-Br bonds.^219,220 However, the Suzuki-Miyaura coupling of 3-bromo-5-chloro-1,2,4-thiadiazole (328) selectively occurs at the C-5 position due to the similarity in BDEs of the C-Cl and C-Br bonds of 328 (88.2 and 80.4 kcal/mol respectively).¹²⁵ In such a case, FMO theory governs the chemoselectivity of palladium(0) insertion.^219,220 In FMO theory the magnitude of stabilisation of a key π* LUMO – Pd dxy HOMO interaction determines the chemoselectivity in palladium catalysed cross-coupling reactions (Figure 3).^125,219,220 A π* LUMO was identified at the 5-position of 3-bromo-5-chloro-1,2,4-thiadiazole (328).¹²⁵ As such, oxidative addition of palladium(0) occurs selectively at the C-5 position of 328 due to secondary orbital interactions stabilising the oxidative addition transition state at this _position.125,219,220

Thio-Diels-Alder Cycloaddition of Diene 323 with Dienophile 32

With ample quantities of 5-butadienyl-1,2,4-thiadiazole 323 in hand, we then turned our attention to the biomimetic thio-Diels-Alder reaction of diene 323 with 3-thioisatin 32.

Attempted Synthesis of Dienophile 32 from Isatin (350) and Subsequent Thio-Diels-Alder with Diene 323

Attempted Synthesis of 3-Thioisatin 32 from Isatin (350)
3-Thioisatin (32) was thought to be attainable by thionation of isatin (350) since the 3-carbonyl of 350 is known to be much more electrophilic than the 2-carbonyl.²²¹ Treatment of 350 with 0.6 equivalents of Lawesson’s reagent at 85 °C formed a single product by TLC (Scheme 88, conditions A). However upon isolation, this bright red solid was identified as indirubin (351), which was assumed to be forming through condensation of isatin (350). When the reaction was repeated using 1.2 equivalents of Lawesson’s reagent at 40 °C, indirubin (351) was again observed to form by TLC (Scheme 88, conditions B). Changing the solvent to THF allowed the reaction to be conducted at room temperature (since both 350 and LR are soluble in THF at room temperature), and a single new product formed by TLC after 30 minutes (Scheme 88, conditions C). However, upon removal of the solvent in vacuo, this product was rapidly converted into a multitude of products including isatin (350), indirubin (351) and indigo (352) by TLC. Subsequent efforts to repeat the reaction and identify the new product by NMR were met with failure as degradation occurred immediately following removal of solvent. HRMS of the reaction mixture allowed identification of a molecular ion corresponding to that of the desired 3-thioisatin 32 {[ESI, (M + Na)⁺] found 185.9986, [C8H5NOS + Na⁺] requires 185.9984}. It was thought that the unstable product was more likely to be 3-thioisatin 32, and not the 2-thio-regiosomer 353 due to the enhanced reactivity of the 3-carbonyl of isatin (350) towards nucleophiles.²²¹ However, N-protection of isatin (350) might confer added stability to the thionation product and thus assist with structural elucidation.

Attempted One-pot Thionation-Thio-Diels-Alder Reaction

As the dienophile was found to be unstable, it was thought that the envisaged thio-Diels-Alder reaction could be conducted in a one-pot procedure following in situ generation of suspected dienophile 32.
Thionation of isatin (350) was carried out using Lawesson’s reagent as described previously, and following observed formation of the suspected dienophile by TLC, diene 323 was added to the reaction (Scheme 91). After 16.5 hours at room temperature, a single new product was formed by TLC. HRMS of this product indicated the presence of one oxygen and two sulfur atoms, while a strong signal at 1611 cm^-1 in the IR spectrum indicated the presence of a carbonyl (not a thiocarbonyl which has a characteristic IR signal at ~1100 cm^-1). Thus it appeared a spiro-dihydrothiopyran had been formed, and not a dihydropyran. However we were unable to elucidate which regioisomer had formed by 2D NMR analysis and efforts to grow crystals of sufficient quality for X-ray crystallographic analysis were unsuccessful. It was thought that if double bond isomerisation of the thio-Diels-Alder product could be achieved, the structure of the cycloaddition product could be elucidated through 2D NMR analysis of the isomerised product and/or X-ray crystallographic analysis. Treating the suspected cycloaddition product (±)-355 with palladium(II) chloride in methanol was thought to form the isomerised product (±)-356 by TLC after 22 hours at room temperature. However, the product was found to be inseparable from the starting material, precluding characterisation of the suspected isomerised product (±)-356 (Scheme 91, conditions A). Addition of triethylamine to a crude sample of suspected cycloadduct (±)-355 in THF was gratifyingly found to effect complete isomerisation of the double bond into conjugation with the thiadiazole (Scheme 91, conditions B). Moreover, it was found that triethylamine could be added to the thio-Diels-Alder reaction following formation of the cycloadduct by TLC, to effect isomerisation in situ. X-ray analysis of the product allowed for unequivocal structural confirmation that the isomerised cycloadduct was 2-spiro-3-oxindole (±)-356, despite a poor quality crystal.

Table of Contents
Preface 
Abstract 
Acknowledgements 
Abbreviations 
Chapter One: Introduction 
1.1 Natural Products Containing 1,2,4-Thiadiazoles
1.2 Synthetic Routes to 1,2,4-Thiadiazoles
Chapter Two: Total Synthesis of Polycarpathiamines A and B 
2.1 Retrosynthetic Analysis
2.2 Oxidative Cyclisation of Thioacylguanidines
2.3 Total Synthesis of Polycarpathiamines A and B
2.4 Summary
Chapter Three: First Generation Route to the Isatis indigotica-Derived Alkaloids 
3.1 Proposed Synthesis of 24a and 24b
3.2 Synthesis of 3-Substituted-5-pseudohalo-1,2,4-thiadiazoles 200 and 201
3.3 Model Heck Couplings
3.4 Synthesis of Spiro-5′,6′-dihydrothiopyran (±)-280
3.5 Alternative Routes to 24a/b from 5-Chloro-1,2,4-thiadiazole 217
3.6 Summary
Chapter Four: Biomimetic Total Synthesis of the Isatis indigotica-Derived Alkaloids 
4.1 Second Generation Approach to 24a and 24b using a Bioinspired Thio-Diels-Alder Reaction
4.2 Biomimetic Total Synthesis of Natural 24a/b and Putative 144a/b
4.3 Summary
4.4 Future Work
Chapter Five: Experimental Procedures 
5.1 General Details
5.2 Experimental Procedures Pertaining to Chapter Two
5.3 Experimental Procedures Pertaining to Chapter Three
5.4 Experimental Procedures Pertaining to Chapter Four .
References
GET THE COMPLETE PROJECT
TOTAL SYNTHESES OF 1,2,4-THIADIAZOLE NATURAL PRODUCTS