Synthesis and Synthetic Chemistry of Pleuromutilin

Despite renewed pleas by the world health organisat ion for action to be taken against antimicrobial resistance the development of new ant ibiotics has not kept pace with demand. The pleuromutilin class is vastly under explored an could hold a key weapon in the battle against antimicrobial resistance. This report explo res the synthesis and synthetic chemistry of this structurally intriguing, densely functionalise d and biologically important natural product.


Introduction
Kavanagh demonstrated in the early 1950's that several fungal species of the genus Pleurotus produced substances that inhibit the activity of Staphylococcus aureus. Further to this, the antibacterial substance produced by Pleurotus mutilus (Clitopilus scyphoides) and P. passeckerianus (Clitopilus passeckerianus) was isolated in crystalline form and named pleuromutilin. 1,2 Structural elucidation studies were first published in 1952, with Anchel correctly describing the molecular formula of pleuromutilin (C 22 H 34 O 5 ) and correctly assigning 3 of the 5 oxygen atoms as 2 non-phenolic hydroxyl groups and 1 hindered carbonyl group. 3 The remaining two oxygen atoms were tentatively assigned to a lactone ring; this was later corrected, and shown to be the ester functionality at C14. Independent work by Birch at the University of  The numbering of pleuromutilin described by Arigoni has become convention throughout pleuromutilin chemistry as shown in Figure 1. For simplicity this system will be used throughout this report.
The ever increasing emergence of multi-drug-resistant microorganisms has led to the World Health Organisation (WHO) issuing renewed pleas for action to be taken against M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 3 antimicrobial resistance. 7 The development of new antibiotics has not kept pace with demand; for example, new cases of multi-drug-resistant tuberculosis alone account for around 150,000 deaths annually. 7 Since the publication of the structure and activity of pleuromutilin, much synthetic effort has been focused on the synthesis of analogues with increased activity and a more favourable pharmacokinetic profile. A handful of compounds are widely used in veterinary medicine and a single analogue has been licensed for topical use on humans. It was recently announced that BC-3781 will enter Phase III clinical trials and may offer the first systemic pleuromutilin antibiotic for human use. [8][9][10][11][12] All of the compounds either licensed or in clinical trials vary only in the C14 side chain. Despite this a wealth of chemistry, including complex structural modifications of the core, has been investigated. Total synthesis, however, presents the only tool for major re-structuring of the pleuromutilin core and several groups have attempted to gain synthetic access to the pleuromutilin core, with three total syntheses now reported.

Synthetic manipulation of pleuromutilin
This section of the review catalogues the synthetic manipulation of pleuromutilin, which will be presented in the following sections: selective functionalization of the C14 hydroxyl, synthesis of metabolites and similar compounds, modification of the core, modification at C12 and modification at C13. Whilst this report does not aim to capture every chemical transformation performed on pleuromutilin or its derivatives, it will provide a useful guide to the reactivity of this structurally intriguing, densely functionalised and biologically important natural product.  Alternative analogues at C14 have been investigated including carbamate and acyl carbamate linkers. In this section the aim is to consider strategies and approaches to analogues of pleuromutilin and not repeat the excellent and extensive patent and structure activity relationship (SAR) reviews on the C14 side chain. [13][14][15][16][17][18][19][20][21][22][23][24] C14 Acyloxy derivatives of mutilin have been prepared, most commonly by activation of the primary alcohol of pleuromutilin (C22) as a tosylate 9a (R = p-tol), followed by substitution with a thiol (X = S),  amine (X = NH) 52,57,58 or alcohol (X = O) [43][44][45]59,60 to form 10 (Scheme 1. A.). This is the general procedure used for the synthesis of tiamulin (3) 61-63 and valnemulin (4) [64][65][66][67][68][69][70] (Figure 2. A.). Alternatively the mesylate 9b (R = Me) was also susceptible to substitution by a thiol (X = S) 59,[71][72][73][74] or amine (X = NH). 75 Retapamulin (5) and M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 5 a number of other analogues were synthesised by an alternate but similar approach: mesylpleuromutilin 9b (R = Me) was substituted with thioacetic acid to give 10 (XR ' = SAc) and treatment of 10 with tropine-3-mesylate in the presence of ethylenediamine cleaved the acetate and gave retapamulin (5) (Figure 2. B.). 76 The stepwise strategy has also been reported whereby treatment of 10 (XR ' = SAc) with ethylinediamine afforded pleuromutilin thiol 10 (XR' = SH), which could itself substitute a number of mesylates. 71,[77][78][79] In an approach to rapidly investigate a number of pleuromutilin nucleoside conjugates, click chemistry was employed; tosylpleuromutilin 9a was substituted with sodium azide to form 10 (XR' = N 3 ), which underwent click chemistry with a range of nucleoside derived alkynes to form 1,2,3-triazole analogues of pleuromutilin. 80,81 The azide 10 (XR' = N 3 ) has also been reduced to give aminopleuromutilin 10 (XR' = NH 2 ) which was converted to a range of analogues. 59 Further activation of the C22 position in tosylpleuromutilin 9a by conversion to the bromide or iodide 11 has also been reported (Scheme 1. B.). In these cases the halide has been substituted with a thiol (X = S), amine (X = NH 2 ) or alcohol (X = O) to give 10. 27,28,40,56,60,82 Scheme 1. Selective functionalization at C22 of pleuromutilin.
An alternate approach to C14 carbamates through 4-epi-mutilin 12 was reported in 2007: 89,90 Treatment of 12 with sodium cyanate and trifluoroacetic acid lead to the formation of 4-epimutilin-14-carbamate 16, which was acylated to give a range of acetylcarbamates. Yang has since reported a one-pot procedure for the synthesis of 4-epi-mutilin-14-acetylcarbamates using silver(I) cyanate and an acid chloride. 42,88,[91][92][93][94][95] Further optimization to a process route was also reported and in that case an acetyl isocyanate was preformed and coupled with 4epi-mutilin. 96 Alternatively, selective protection of the C11 hydroxyl followed by reaction of the C14 hydroxyl has been reported. Strategies involving bis-acylation (C11 and C14) followed by mild cleavage of the C14 acetate were first reported. 92 However, more recently the trifluoroacetyl group has been employed as an easily-cleaved protecting group that can be selectively installed at C11 (Scheme 3). Treatment of mutilin with TFAA and pyridine (Discovery route) or with trifluoroacetylimidazole (Process route) provided the C11 trifluoroacetyl mutilin 17 which has been shown to take part in several ester forming reactions. 58,97 The advantage of this strategy is that simple treatment of 18 with mild base will cleave the C11 trifluoroacetyl group and afford the desired pleuromutilin analogue 14. Intermediate 17 has also been used in a range of carbamate and acyl carbamate forming reactions, previously described with 4epi-mutilin (Scheme 2). 42,88,92,93,97 Yang has reported the synthesis and biological data for a series of water-soluble phosphate prodrugs of pleuromutilin analogues, the most promising example bears the phosphate at a phenolic hydroxyl on a C14 carbamate side chain; however Yang also reported an example of a C11 phosphate prodrug of a C14 carbamate pleuromutilin analogue. Both of these compounds were shown to be suitably soluble and stable in aqueous buffer, and the phenolic phosphate was metabolized to the active drug and exhibited comparable antibacterial activity to vancomycin in a mouse model. 42,96 In 2001 Takle and co-workers at SmithKline Beecham reported an approach to C14 amido mutilin derivative 22 (Scheme 4). 85 Treatment of diacetylmutilin 19 with potassium tertbutoxide afforded the cyclopropyl mutilin derivative 20, which proved largely resistant to further reaction. However, treatment of 20 with sodium azide in DMF at 150 °C led to the formation of 21, albeit in very low yield. Reduction of the azide in 21 proceeded with concomitant reduction of the C12 vinyl. This was followed by conversion of the C14 amine to an amide and saponification of the C11 acetate to give 22. The intermediate C14 amino mutilin has also been used in the synthesis of C14 acylurea analogues. Takle and co-workers have shown that these two C14 amino derived analogues have dramatically reduced antibacterial properties and therefore this analogue family was not investigated further.

Synthesis of metabolites and similar compounds
The metabolism of pleuromutilin has been mainly attributed to cytochrome P450 (CYP) dependant hydroxylation at either C2 (23) or C8 (24) of the tricyclic core (Scheme 5). These compounds are almost entirely devoid of biological activity. 98,99 Metabolite 25 was also isolated and is the product of de-ethylation of the tiamulin side chain. This compound exhibits around 25% of the activity of tiamulin, metabolites that are active, but exhibit activity at a lower level than the parent antibiotic would be expected to select for antimicrobial resistance. 100 Scheme 5. Metabolites formed in pig liver microsomes.
Several researchers have investigated synthetic access to the metabolites of pleuromutilin and also to compounds potentially blocked to metabolism. Hanson has investigated the biotransformation of pleuromutilin and mutilin using CYPs in whole cell cultures. 99,101 Treatment of (+)-pleuromutilin (1) with Cunninghamella echinulata SC 16162, a filamentous fungus, provided (2R)-hydroxy pleuromutilin 26 (Scheme 6. A.). Interestingly, the opposite epimer 27 is obtained upon treatment of (+)-mutilin (2) under the same conditions (Scheme 6. Several studies have investigated the products of biotransformation of mutilin: C8 hydroxy mutilin derivatives (28) have been used to synthesise a number of analogues including a C8 keto derivative. However, the vast majority of these analogues have been shown to exhibit no antibacterial activity. [102][103][104] Scheme 7. Synthesis of (2S)-fluoro and hydroxy mutilin derivatives by Yang. was formed in high yield and oxidized selectively at the allylic position using catalytic dirhodium(II) caprolactamate, to give 45 in 40% yield; the corresponding allylic alcohol was also isolated in 47% yield. Dissolving metal reduction afforded the desired cyclopentanone product in 64% yield but as a mixture of the diacetate (20%) and the monoacetates. 109 Scheme 9. Rhodium-mediated allylic oxidation to form C2-ketomutilin derivatives by Axten.
Simple and scalable access to 2-hydroxy mutilin was not available until 2008 when Wang reported a Rubottom oxidation procedure (Scheme 10). Treatment of mutilin with lithium hexamethdisilazide and TMSCl selectively afforded the kinetic silyl enol ether 46 which was oxidised using mCPBA. Commonly NaHCO 3 is employed to buffer the Rubottom oxidation however in this case issues of over-oxidation and irreproducibility were reported. This was partially due to the inhomogeneity of the reaction mixture. Buffering the reaction with acetic acid and pyridine offered improved yields and high reproducibility on kilogram scale. Acidic Reagents: a. TMSCl, LiHMDS; b. mCPBA, AcOH-py; c. HCl, 78% (3 steps).
As described above, several strategies aimed at mimicking nature in gaining oxidative access to C2 have been developed, however the seemingly unactivated positions of metabolism on pleuromutilin derivatives (C7 and C8) have provided more of a challenge. Biotransformation has provided preparative access to metabolites for the investigation of C7 and C8 oxidized analogues of pleuromutilin and more recently the group of White has accessed C7 oxidized analogues. Birch also observed the autoxidation of diacetoxymutilin using potassium tert-butoxide and tert-butanol to afford a ring cleaved ketoacid, similar to 64 (Scheme 16). Under the conditions employed by Birch the C11 acetate was hydrolysed, this was followed by a transannular hydride shift from C11 to C4 to afford an 11-keto-4-hydroxy derivative which proceeded to form a δ-lactone by cyclisation of the C4 hydroxyl with the pendant carboxylic acid. 5   Modification of the pleuromutilin core has offered access to alternative tricyclic and bicyclic analogues, some of which have been shown to possess interesting biological activity. This suggests that further investigation of pleuromutilin inspired tricyclic compounds with alternate ring sizes may offer breakthroughs in the search for improved pleuromutilininspired antibacterials.

Modification at C12
Another handle for the modification of pleuromutilin is the C19-20 olefin at C12. Indeed, one of the first reactions reported on pleuromutilin was the dissolving metal reduction with lithium in methanol and ammonia, resulting in reduction of the C3 ketone and the C12 vinyl Using dihydroxylation and then base mediated fragmentation, Takadoi provided an alternate route to a similar 12-devinyl mutilin derivative 83 (Scheme 20. B.). Takadoi installed a number of groups at C12, including the propargyl group by alkylation with propargyl bromide to give 84; whilst an interesting derivative in its own right, the C12 propargyl also offers opportunity for further functionalization. Takadoi also reported the addition of methyl para-toluenethiolsulfonate to form the 12-thiomethyl derivative 85. All of Takadoi's derivatives were then furnished with a suitable C14 side chain and the pleuromutilin stereochemistry restored (ZnCl 2 and HCl) before biological evaluation. [116][117][118] In an approach to serine protease inhibitors, Kaura at GlaxoSmithKline developed a Curtius rearrangement approach to C12 amino mutilin derivative 92 (Scheme 21). 119 The approach began with a 1,5-hydride shift dehydroxylation, first reported by Birch, 5 to give 87.
Dihydroxylation and oxidative cleavage was followed by oxidation under Jones conditions to give the carboxylic acid 88, which after activation as the acid chloride was treated with Exploitation of the reactivity of the C12 vinyl group of mutilin has provided access to a range of novel C12 modified derivatives. Furthermore an appreciation that the reduction product has largely unchanged bioactivity has allowed the use of C12 methyl, ethyl derivatives in chemistry that would be incompatible with the presence of an olefin.

Modification at C13
Introduction of nitrogen functionality at C13 of pleuromutilin was first reported by Berner; thermal decomposition of the saturated C14 carbonazidate 103 resulted in formation of 104 in modest yield as a single undefined diastereoisomer (Scheme 25). 123  Formation of the C13-14 fused oxazolidinone has been reported using two different routes.
Uccello at Pfizer has reported conditions compatible with the C12 vinyl group however a scalable procedure for the hydrolysis to C13 amino pleuromutilin has yet to be reported.

Synthetic approaches towards pleuromutilin
Alongside synthetic modification of pleuromutilin, intense effort has focused on access to synthetic cores. This has led to six distinct approaches to the core, three of which relate to complete total syntheses.

M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT
Kahn first investigated synthetic access to the core of pleuromutilin in 1980, 125   In the mid and late 1980s Paquette and co-workers identified advanced intermediates in a proposed asymmetric pleuromutilin synthesis through degradation studies (Scheme 28. A.). [127][128][129][130][131] The group identified the degradation product 122 and presented a synthetic approach to this lactone, which was then successfully converted into the bicyclic diketone    Independently of the work of Procter and co-workers, Sorensen reported a number of approaches to the core of pleuromutilin. Sorensen reported two RCM approaches, either forming the C12-13 or C11-12 bond of the 8-membered ring (Scheme 32). 139 The first approach converted 160 to the cis-hydrindanone 162 through a series of conjugate additions.
Protection of the ketone was followed by borylation of the alkene and palladium-catalysed cross-coupling to install the desired alkene tether. The nitrile 162 was converted to the α,β-

Total syntheses of pleuromutilin
Two racemic syntheses of pleuromutilin have been reported to date; the first by Gibbons was      Progression of 223 to (+)-mutilin (2) was described in three steps according to standard procedures that are shown in Scheme 36. Efficient conversion of mutilin (2) to pleuromutilin (1) was previously unreported in the literature. The Gibbons approach relied upon multiple iterations and the yield reported by Boeckman was low (39%). The use of C11 trifluoroacetate protection was, however, known for the synthesis of C14 analogues of pleuromutilin and the group of Procter utilised this methodology to finish the total synthesis.
Treatment of (+)-mutilin (2) with trifluoroacetylimidazole at reduced temperature gave C11 trifluoroacetylmutilin, which underwent efficient EDCI coupling with trifluoroacetoxyacetic acid. It was shown that quenching the reaction with methanol and treatment with triethylamine led to the cleavage of both trifluoroacetate groups and delivered (+)pleuromutilin (1) in 75% for the last two steps.
To date three strategically very different total syntheses of pleuromutilin have been reported, with surprisingly similar step counts. Each offers different insights into the construction and functionalization of this unique and beautiful natural product. Gibbons' fragmentation approach elegantly constructs a highly functionalized tricyclic core, which is further elaborated to the natural product in a total of 31 steps. Boeckman's route, by far the shortest

Conclusions
Industrialists and academics alike have been inspired by pleuromutilin since its first isolation 63 years ago, yet, only three pleuromutilin derivatives grace the market. A vast amount is now understood about the SAR surrounding the C14 side chain and semi-synthesis has provided some modifications to the core. However, these efforts have not resulted in new treatments. Semi-synthetic analogues are still under investigation and synthetic approaches to the core, reported over the last 5 years, promise to offer novel totally synthetic analogues which may become key weapons in the ongoing battle against antimicrobial resistance.

Abbreviations
Ac acyl