Biosynthesis and Total Synthesis Studies on The Jadomycin Family of Natural Products

Abstract

Jadomycins are unique angucycline polyketides, which are produced by soil bacteria Streptomyces venezuelae under specific nutrient and environmental conditions. Their unique structural complexity and biological activities have engendered extensive study of the jadomycin class of natural compounds in terms of biological activity, biosynthesis, and synthesis.

This review outlines the recent developments in the study of the synthesis and biosynthesis of jadomycins.

1. Introduction

Nature provides us with ample sources of therapeutically useful molecules. Often the natural products are derivatized to modulate their activity in order to achieve specificity and selectivity. In fact, more than half of the drugs on market today are derived from natural products.¹ The necessity to discover and develop new and potent therapeutic agents, to address a wide array of disease and health-related conditions, has encouraged scientists to discover versatile and efficient routes for the biosynthesis and synthesis of new biologically active natural products.

Jadomycins are novel angucycline polyketide natural products composed of an unusual 8H-benzo-[b]-phenanthridine backbone with variously fused nitrogen and oxygen containing heterocycles (e.g., oxazolidine, pyrimidine, oxazinane). With the exception of jadomycin A (aka, jadomycin B aglycon), all the jadomycins are glycosylated with a L-digitoxose sugar unit (Figure 1).² This sugar moiety plays a key role in the bioactivity exhibited by these compounds (e.g., jadomycin B exhibits anti-yeast activity whereas jadomycin A is inactive).³ Jadomycins are active against gram-positive and gram-negative bacteria and shows anti-cancer activity, which has tentatively been attributed to aurora-B kinase inhibition and/or DNA cleaving capacity.⁴ These secondary metabolites are produced by gram-positive soil bacteria Streptomyces venezuelae ISP5230 (ATCC10712) under nutrient limitation along with additional stress (e.g., heat shock, phase infection or exposure to toxic concentration of ethanol).^5,6 In contrast, chloramphenicol is the only antibiotic currently characterized from Streptomyces venezuelae under normal growth conditions.⁷

Figure 1.

Jadomycin A and its structural diversed congeners

2. Jadomycin Biosynthesis

2.1. Regulatory genes for jadomycin biosynthesis

Since the early isolation work of Ayer,⁵ the unique structure of the jadomycins has inspired extensive study of their biosynthesis. These initial studies ultimately led to the discovery of related jadomycin secondary metabolites. In 1994, Vining was the first to clone and characterize the polyketide synthase (PKS) genes responsible for jadomycin B biosynthesis in Streptomyces venezuelae ISP5230.⁸ Later, more detailed investigations of the gene clusters were carried out by Yang.⁹ A schematic diagram depicting an organizational map of the jadomycin gene cluster is shown in Figure 2. The genome sequence of Streptomyces venezuelae ISP5230 (ATCC10712) was recently reported.^6b AntiSMASH analysis of the genome sequence identified 26 natural product biosynthetic gene clusters along with four additional polyketide synthase clusters and an independent lantibiotic biosynthetic cluster. Although, production of the lantibiotic was not observed in Streptomyces venezuelae.¹⁰

Figure 2.

Organizational map of jadomycin gene cluster

2.2. Sequencing region upstream of PKS-II

The need for both nutrient depletion and environmental stress on S. venezuelae to induce jadomycin production is quite unique, which suggests the existence of a control mechanism.¹¹ In contrast, most other streptomycetes produce type-II PKS antibiotics under normal growth conditions⁷ and modulate the antibiotic production upon physiological imbalance.¹² Evidence for this control mechanism can be found in the genetic code for jadomycin. An Open Reading Frame (ORF) upstream of type-II PKS encodes for a 196 amino acid sequence (jadR₂). This amino acid sequence resembles known proteins (e.g., MtrR, AcrR, TetC and TcmR) that regulate resistance to xenobiotic exposure. Thus, when the nucleotide sequence of the jadR₂ repressor gene in S. venezuelae was disrupted,¹³ jadomycin B was produced under normal growth conditions, and overproduction was observed when these mutants were exposed to ethanol stress. This findings point to the conclusion that jadR₂ negatively regulates the expression of biosynthetic genes of jadomycin B.¹¹

In 2001, a second regulatory gene (jadR₁) with 234 amino acids was identified by Vining.¹⁴ The gene was located immediately upstream of the repressor gene (jadR₂) and resembles sequences for two component regulator systems.¹⁵ The mutants generated by disruption or deletion of jadR₁ did not produce jadomycin B, implying that the gene functions as a positive regulator in the antibiotic biosynthesis. When a disrupted chromosomal copy of jadR₁ was complemented with a cloned gene, recovery of the wild-type phenotype was observed. Also, a negative effect on the growth of S. venezuelae occurred upon increasing the copy number of jadR₁ in the plasmid. This investigation by Vining led to the conclusion that both jadR₁ and jadR₂ form an interacting stress-responsive regulatory system for the biosynthesis of jadomycin B.¹⁴

2.3. Sequencing region downstream of PKS-II

In 1996, Vining and co-workers cloned and sequenced DNA from a region downstream of an overlapping PKS gene cluster in S. venezuelae (ISP5230), which is responsible for jadomycin B biosynthesis.^6a Their analysis of the nucleotide sequence located one complete ORF (ORF6) and two incomplete ORFs (ORF4 and ORF7). The amino acid sequence for ORF6 and ORF7 resembles oxygenase. Transformation of S. venezuelae with an ORF6 disruption vector¹⁶ gave a mutant, which blocked jadomycin B biosynthesis. Instead, the mutant accumulated a non-nitrogenous secondary metabolite (rabelomycin 4), which is also produced by Streptomyces olivaceus (Scheme 1).¹⁷ This suggests that rabelomycin 4 is an intermediate¹⁸ in the biosynthesis of jadomycin, where ORF4, ORF6 and ORF7 encodes for oxygenases that are involved in the oxidative cleavage of the B ring in angucycline 4 to aldehyde 6 All that remains for the biosynthesis of jadomycin B is A-ring dehydration/aromatization, isoleucine incorporation, oxazoline ring formation and glycosylation. However, the specific order of these sequences is still in doubt. Left of jadR₂, a cluster of three other regulatory genes (jadW₁, jadW₂ and jadW₃) were identified by Vining et al. (2003).¹⁹

Scheme 1.

First hypothesis of the jadomycin biosynthetic pathway

Subsequently, another ORF was found (jadJ) between jadR₁ and jadI.²⁰ The gene jadJ encodes for proteins that are homologous to the biotin carboxylase (BC) and biotin carboxyl carrier protein (BCCP) components of acyl-coenzyme A carboxylase. The mutants produced by disruption of jadJ significantly reduced jadomycin production, without negative effects on the growth or morphology of the mutants. Thus Vining et al. concluded that jadJ is not essential for fatty acid biosynthesis, but rather responsible for supplying malonyl-coenzyme A for the synthesis of polyketide intermediate. Later, several more genes were identified for the biosynthesis of jadomycin, by Vining.²¹ Three complete ORF (jadM, N and X) and one incomplete ORF (jadO) were found. The later (jadO) relates to the biosynthesis of the sugar portion of jadomycin (Scheme 7). Although the role of jadX could not be identified, the roles jadM, N and O were assigned by homology to known proteins. The other two complete ORFs were assigned as a holo-ACP synthase (jadM), and acyl-coenzyme A decarboxylase (jadN), respectively. The incomplete ORF (jadO) showed strong similarities with to NDP-hexose-2,3-dihydratase, indicating it’s role in the synthesis of L-digitoxose sugar unit in jadomycin B.

Scheme 7.

Proposed biosynthesis of L-digitoxose sugar unit.

2.4. UWM6 a new intermediate in jadomycin biosynthesis

A new gene (jadI), upstream of jadA was characterized by Hutchinson et al.²² JadI has homology to a related cyclase responsible for the biosynthesis of the ABCD rings of jadomycin in combination with the minimal PKS genes (jadABC), a cyclase (jadD), and a keto reductase (jadE). When these genes are expressed in Streptomyces lividans, rabelomycin 4 is produced along with a new angucycline UWM6 12 (Scheme 2). When this experiment was run with an incomplete jadI shunt, products with out C-ring (SEK43 14 and UWM4 15 were obtained), thus suggesting the role of jadI in ABCD-ring construction and the potential of UWM6 12 and/or rabelomycin 4 as being an intermediate in the biosynthesis of jadomycin B.²³

Scheme 2.

Polyketides produced from jad PKS gene cassettes in S. lividans

2.5. Proposed ring opening mechanism of UWM6

Initial biosynthetic studies suggested that rabelomycin is the intermediate in jadomycin biosynthesis and jadF catalyzes the oxygenation of rabelomycin leading to B-ring opening.^6a However, subsequent work discovered other intermediates in the pathway, including UWM6 (12) and prejadomycin (16), which suggested that rabelomycin is not an actual intermediate in the pathway of jadomycin B biosynthesis.^24,25 The most recent studies by Yang et al.²⁶ suggests that jadF is a bifunctional enzyme that catalyzes the dehydration of UWM6 (12) to prejadomycin (16), and oxygenation at C-12 to form (17) which spontaneously air oxidizes to form dehydrorabelomycin (18) (Scheme 3). Although the exact identity of the enzyme responsible for the B-ring opening of dehydrorabelomycin (18) is still unknown, it is assumed that some combination of jadF, jadH and jadG is responsible for this reaction (18 to 19).²⁴ Rohr et al. have proposed a mechanism for the C-C bond cleavage of B-ring by a Baeyer-Villiger type mechanism (20 to 21). Subsequent hydrolysis of lactone intermediate 22 by jadK would give aldehyde 19. However, at that time, the C-12 oxygenation was incorrectly assigned as performed by jadG (Scheme 3).²⁵

Scheme 3.

Proposed ring opening of UWM6

3. Amino acid incorporation in jadomycin biosynthesis

Regardless of the pathway of its formation, aldehyde 19 is believed to be the intermediate that incorporates the amino acids. Initial amino-acid condensation followed by a series of decarboxylation and ring closing sequences to form the pentacyclic skeleton of jadomycin. Early on, Vining et al.^6a suggested that isoleucine incorporation begins with an intermolecular 1,4-addition across the quinone enone (6 to 7), which is followed by a decarboxylation (7 to 8). A subsequent intramolecular addition to the aldehyde and lactonization should give the pentacyclic core (i.e., 8 to 9, Scheme 1). Considering the presence of excess amino acid nucleophile (e.g., isoleucine) and the relative reactivity of the functional groups, Rohr et al. proposed an alternative mechanism. Their mechanism begins with the amine of isoleucine condensing with the aldehyde to form intermediate imine 23. A subsequent cascade cyclization (carboxylate addition to the imine and amine addition to the quinone)²⁷ and decarboxylation leads to the pentacyclic core of jadomycin (Scheme 4).

Scheme 4.

Proposed non-enzymatic incorporation of isoleucine

It is still an open question as to where the amino acid incorporation proceeds through an enzymatic or spontaneous process. To date, no enzyme has been isolated in the jadomycin gene cluster that is responsible for the process. Rohr et al. suggested the possibility of a non-enzymatic incorporation of amino acid based on the fact that jadomycin occur as an inseparable mixture of diastereomers.²⁷ They synthesized a range of jadomycin B analogues using different amino acids in the culture medium. A careful NMR analysis of the jadomycin analogues and molecular modelling studies suggested that the jadomycin is a mixture of diastereomers at the aminal position. They propose that the mixture of diastereomers equilibrate via an undetected zwitterionic intermediate (Scheme 5). A recent total synthesis of jadomycin A and carbasugar analogue of Jadomycin B by O’Doherty et al. (see synthetic approach to jadomycin) supports the proposal that amino acid incorporation is indeed a non-enzymatic process.²⁸

Scheme 5.

Proposed mechanism of diastereomeric interconversion

3.1. Alternative mechanism of ring opening in jadomycin B

Jakeman proposed an alternative mechanism for the inter-conversion of the jadomycin diasteromers.²⁹ His proposal was based on the isolation of an aldehyde intermediate (~pH 11) as a mixture of conformational isomers about the aldehyde C-C bond, 27 and 28. They suggested that these two conformational isomers can selectively cyclize into the two corresponding diastereomeric forms of jadomycin B upon acidification (~pH 4). Interestingly, Dalomycin T 30 (which has an oxazolidine ring and the same angucycline framework as of jadomycin B) was not responsive to the same acid-base treatment. This led to the conclusion that the carbonyl group of the oxazolone ring is essential for ring opening. They suggest that the aldehyde intermediate 27 is formed by nucleophilic attack of hydroxide ion at the carbonyl carbon of the oxazolone ring, subsequent opening of the oxazolone give aldehydes 27/28. Not surprisingly, Dalomycin T 30, which lacks the required carbonyl group, does not undergo a similar isomerization and exists as a single diastereomer (Scheme 6).

Scheme 6.

Role of carbonyl group in interconversation of 3a stereo center

3.2. Incorporation of non-natural/non-proteogenic amino acids

Jadomycin analogues can be obtained by using different amino acids in the growth medium. Doull et al. was the first to observe this phenomenon.^4a Differently colored pigments were observed in the culture medium when L-isoleucine was replaced with other nitrogen sources. But no effort was made to isolate or characterise these compounds. Rohr and Jakeman were the first to characterize and report novel jadomycin analogues by the fermentation of Streptomyces venezuelae in the presence of excess amino acids.^4d,27,30 The new jadomycin analogues that were produced using different amino acids are listed in Table 1. jadomycin N with a unique six-membered ring containing two nitrogens was obtained when L-asparagine was used as a nitrogen source. It is probably formed by the cyclization of the primary amine side chain of L-asparagine rather than ring closure through a carboxylic acid.^30b Jakeman et al. have reported spectroscopic evidence suggesting that, jadomycin analogues can also be obtained from the non-natural D-amino acids and non-proteogenic/synthetic amino acids. The size of the oxazolone ring in jadomycin analogues can also be expanded by using β-amino acids, thus, if β-alanine or racemic-3-aminoisobutyric acid is used as the only nitrogen source in Streptomyces venezuelae culture, derivatives containing six-membered oxazinanone rings are produced.^30a,c,d These efforts lead to the first systematic study of structure activity relationships associated with jadomycins. For instance, substitution of the alkyl side chain on the oxazolone ring of jadomycin B resulted in changes to the bioactivity (antibacterial and DNA cleaving properties), whereas the two analogues with no alkyl side chain on the oxazolone ring were significantly less active.^2,30b,d

Table 1.

Jadomycin analogues produced using different amino acids

4. Biosynthesis of the L-digitoxose sugar unit

The biosynthesis of the sugar portion of jadomycin has also attracted significant attention. In the jadomycin gene cluster, immediately downstream of jadN, eight structural genes, jadX, O, P, Q, S, T, U and V are located, which are involved in the biosynthesis of the carbohydrate portion of jadomycin B. The specific functions of the corresponding proteins from seven of these genes (jadO, P, Q, S, T, U and V) were assigned by sequence analysis and comparisons to the related genes involved in deoxy sugar biosynthesis (2,6-dideoxy sugar L-digitoxose). To carry out sequence analysis, the genes were cloned and propagated in Escherichia coli.³¹ Analysis of the intermediates produced by the mutants and the sequence similarities to genes described in other species producing deoxysugar derivatives confirmed that these seven genes (JadO, P, Q, S, T, U and V) were responsible for the dideoxy sugar biosynthesis and attachment to the aglycone (jadomycin A). In contrast to the other seven, jadX is not essential for jadomycin B biosynthesis. However, the presence of jadX improves the yield for the production of jadomycin B. Based on these results, Vining et al. have proposed a biosynthetic pathway for the production of L-digitoxose sugar unit from glucose-1-phosphale 32 (Scheme 7).³

jadQ encodes glucose-1-phosphate nucleotidyltransferase that activates glucose to nucleotide diphosphate (NDP) derivative 33. A 4,6-dehydratase encoded by jadT converts this NDP-glucose 33 to NDP-4-keto-6-deoxy-D-glucose 34. An NDP-hexose 2,3-dehydratase encoded by jadO converts 34 to a ketoenol intermediate 35, which exists in equilibrium with the diketone intermediate 36. This diketone intermediate is transformed to NDP-2,6-dideoxy-D-threo-4-hexulose 37 by an oxidoreductase encoded by jadP. Isomerization of this D-threo hexulose 37 to L-erythro form 39 occurs by the jadU product (NDP-4-keto-2,6-dideoxy-5-epimerase) via intermediate 38. Finally the ketone group of NDP-4-keto hexulose 39 is reduced by NDP-4-keto-2,6-dideoxyhexose 4-ketoreductase, encoded by jadV to give the L-digitoxose moiety 40 in jadomycin B biosynthesis. It is unclear whether glycosylation occurs on jadomycin A or other intermediates prior to the formation of the oxazolone ring by incorporation of amino acid. Comparison of the amino acid sequence encoded by jadS shows resemblance to deoxy-glycosyltransferase. So, it was proposed that jadS is involved in installing the L-digitoxose sugar unit onto the jadomycin B.

4.1. Substrate flexibility of 2,6-dideoxyglycosyl transferase

The function of jadO as a nucleoside diphosphate (NDP)-hexose 2,6-dehydratse was confirmed by Jakeman et al.³² The jadomycin analogue ILEVS 1080 was reported after a careful analysis of the natural products formed by using a previously developed S. venezuelae strain (VS 1080).³ NMR and MS/MS analysis showed that the sugar unit in ILEVS 1080 differs from the L-digitoxose in case of jadomycin B. The new sugar replaces the axial hydrogen at the C-2 position with an axial hydroxyl group. The disrupted jadO could not dehydrate at C-2 position of NDP-4-keto-6-deoxy-D-glucose 34 (Scheme 7), thus the biosynthetic mechanism produced the sugar unit having C-2-hydroxyl functionality. This unnatural 6-deoxy sugar is then transferred to phenanthroviridin by jadS (a 2,6-dideoxyhexosyltransferase) to produce ILEVS 1080 (Scheme 8). It is worth noting that this is the first report of a 2,6-dideoxysugar-O-glycosyltransferase with substrate flexibility at C-2-position. The yield of ILEVS 1080 production (2.6 mg/3L)³² is significantly lower than jadomycin B (12 mg/L).²⁷ The low yield of ILEVS 1080 is likely the result of alteration in kinetic parameters of the enzymes downstream of jadO as they are processing a non-natural substrate.

Scheme 8.

Effect of jadO on sugar unit

5. Biosynthetic improvemments for jadomycin B production

Effects of heat shock, ethanol treatment and phase infection on the production of jadomycin B by Streptomyces venezuelae have been studied extensively by the groups of Doull,^4a Rohr²⁷ and Jakeman.³³ Initial studies on S. venezuelae ISP5230 in galactose-isoleucine medium and corresponding media with different amino acids replacing isoleucine lead to isolation and characterization of jadomycin B analogues with modified oxazolone ring substituents.^2,3,4a Rohr’s fermentation procedure differs from that of Doull et al. only in the starting culture volume and the volume and time of ethanol shock.^4a,27 A more robust and efficient culture condition was developed by Vining and Jakeman where glucose was used as the carbon source instead of galactose, MOPS was employed as the buffer along with low concentration of phosphate and immediate ethanol shock. This altered media can render jadomycin B/analogue production in a time efficient and cost effective manner.³³

Yang et al.⁹ has reported a genetically modified strain of Streptomyces venezuelae that exhibits a two fold increase in the production of jadomycin B compared to the wild type. Four regulatory genes (3′ end 272 bp of jadW2, jadW3, jadR2 and jadR1) and the negative promoter upstream of jadJ (P_J) were replaced with promoter sequence ermEP*. The promoter sequence ermEP* can derive the expression of jadomycin biosynthetic genes from jadJ and is able to produce jadomycin without ethanol shock in contrast to the wild type.

6. Synthetic approaches

6.1. Benzo-[b]-phenanthridine synthesis

The benzo-[b]-phenanthridine containing natural products are structurally unique angucyclines isolated from different species of Streptomyces.³⁴ Despite many synthetic efforts, there has only been a limited number of successful total syntheses of this type of angucycline antibiotics. This lack of successful syntheses can be attributed to both the structural complexity and instability associated with these structures, particularly with the glycosidic bond. In contrast, there have been several successful approaches to the phenanthridine core.

Gould³⁵ reported the first synthesis of the naturally occurring benzo-[b]-phenanthridine aglycon in 1991. The A, C and D rings were constructed by coupling of cyanophthalide 42 with substituted bromocinnamate 41 followed by air oxidation to give quinone 43. This was then converted to amide 45 by reduction/methylation of quinone 43 followed by deprotection of TMSE. The C-5 carbon was installed by formylation to give desired product 46 along with dehalogenated product 47 and fluorenone 48. Finally the B ring was installed in phenanthridine 50 by Hoffmann reaction followed by hydrolysis/imine formation (Scheme 10).

Scheme 10.

Synthesis of phenanthridine ring.

In 1997, Valderrama et al. reported a synthesis of the benzo[b]phenanthridine framework by employing a hetero Diels-Alder reaction (Scheme 11).³⁶ The angular tetracyclic system was synthesized by an A plus CD to ABCD strategy. The B ring was constructed by a regioselective 4+2 cycloaddition of 5-hydroxy-1,4-naphthaquinone 51 and N,N-dimethylhydrazone 52 followed by spontaneous air oxidation/aromatization. The high regioselectivity of cycloaddition can be attributed to polarization of the azadiene 52 and directing effect of the C5 hydroxy group on juglone 51. The regioselectivity of cycloaddition of juglone and azadienes was previously established by Potts³⁷ and Fillion.³⁸ Although a higher degree of regioselectivity was achieved, the reaction was slow at room temperature and was associated with the formation of minor bi-products (e.g., regioisomer 54 and addition/oxidation products 55 and 56).

Scheme 11.

DA approach to phenanthridine ring

The most common method for constructing the phenanthridine B ring has been by the condensation of 3-amino quinones with aryl aldehyde at the 3-position of quinone.^39,40,41 In 1997, Snieckus reported the synthesis phenanthrovirdin aglycone 50. The ACD rings were constructed by Suzuki-Miyaura coupling of bromojuglone 59 and oxaborole 58. The 3-amino functionality was introduced by a 1,4-addition on to the quinone 60. The B ring was then constructed by condensation of the amine and the aldehyde after MnO₂ oxidation of allylic alcohol 61. Subsequent removal of methyl protection gave desired phenanthrovirdine aglycone 50 (Scheme 12).³⁹

Scheme 12.

Construction of B ring by condensation of amine and aldehyde

Echavarren et al. have also reported the synthesis of phenanthrivirdine aglycone 50 following the same cyclization procedure. They constructed the ACD ring system by Stille coupling of bromo juglone 59 with aryl stannane 63. A 1,4-addition of ammonia onto the quinone followed by subsequent deprotection of the acetal and subsequent condensation gave the B ring of the tetracycline 65. MOM and methyl protections were removed to obtain 50 (Scheme 13).⁴⁰

Scheme 13.

Construction of B ring by condensation of amine and aldehyde

Construction of the B ring of phenanthrivirdine by amine condensation with the aldehyde has also been achieved by Ishikawa (see total synthesis of dimethyl Jadomycn A).⁴¹ Although the B ring has been successfully constructed by the intramolecular condensation of an amine to the C3 aryl aldehyde of juglone, 1,4-addition of the aryl amine/imine at the C3 position of juglone does not give the desired B-ring of benzo-[b]-phenanthridine 68. Instead, it undergoes a 6-exo-dig cyclization by condensing with C1 carbonyl of juglone 67 to form benzo-[c]-phenanthridine 66 (Scheme 14).⁴⁰

Scheme 14.

Regioselectivity of B ring formation

A solution to this problem came in 2010 from the synthesis of jadomycin A by O’Doherty. They showed that the B ring in the benzo-[b]-phenanthridine skeleton could be constructed regioselectively by changing the mechanism to circumvent the issues associated with a 6-endo-dig cyclization. Specifically, they decided to use a 6-π-electrocyclic ring closure instead of 1,4-addition of an amine to the quinone (see total synthesis of jadomycin A and carbasugar analogue of jadomycin B).²⁸

6.2. Synthesis of L-digitoxose sugar unit

In contrast to syntheses of the aglycon, there have been many approaches to the synthesis of L-digitoxose sugar. In 1982, Fronza et al. reported a stereospecific synthesis of L-digitoxose by addition of diallylzinc on to chiral aldehyde 73. The stereochemistry of addition is controlled by steric interaction of the chiral α,β-dialkoxy aldehyde. Deprotection and ozonolysis gave L-digitoxose 75.⁴² Three years later, Fujisawa et al. discovered a highly stereospecific and scalable synthesis of the key intermediate (glyceraldehyde derivative 73) by diastereoselective reduction of α,β-diketodithiane with baker’s yeast and subsequent protection of the diol (Scheme 15).⁴³ Other related approaches to the sugar have been achieved by isoxazoline anion chemistry.^44,45

Scheme 15.

Synthesis of L-digitoxose employing baker’s yeast reduction

Braun et al. (1991) synthesized L-digitoxose from O-silylated lactal 77. Lactal 77 can be obtained from optically pure/commercially available S-lactate. The sugar C3 and C4 stereocenters were installed by the sequential diastereoselective addition of anions (78 to 79) and (79 to 71). Finally L-digitoxose 75 was obtained by reducing mixture of lactones 82 and 83 with disiamylborane followed by hydrolysis and silica gel chromatography (Scheme 16).⁴⁶

Scheme 16.

Synthesis of L-digitoxose from S-lactate

In the same year, Tatsuta et al. reported the synthesis of L-digitoxose 75 from a glucose sugar 84. Key to their approach was the use of a highly stereocontrolled hydride addition to a carbonyl group at the C-3 position of 2,6-anhydro-2-thiosugar 85. The structure of 2,6-anhydro-2-thiosugar and configuration of the anomeric methoxy group directs the incoming nucleophile in a stereoselective manner. The unfavorable 1,3-diaxial interaction overrules the lone-pair repulsion from the sulfur and thus the hydride is delivered from bottom face. The key intermediate 85 was prepared in 10 steps starting from methyl α-L-glucopyranoside 84 (Scheme 17).⁴⁷

Scheme 17.

Synthesis of L-digitoxose from L-glucose

In 2005, Wang et al. reported a synthetic strategy for optically pure L-digitoxose sugar unit by employing kinetic resolution followed by ring closing metathesis. The synthesis began with commercially available acyclic trans-4-phenyl-3-buten-2-one 88. Reduction of the α,β-unsaturated carbonyl followed by enzyme-metal catalysed dynamic kinetic resolution to obtain optically pure allylic alcohol 90. Mitsunobu esterification of 90 followed by Grubb’s ring closing metathesis and subsequent epoxidation/epoxide opening gave the key intermediate 94. This α,β-unsaturated lactone 94 (Osmundalactone) can be transformed to L-digotoxose 75 by hydroxyl directed epoxidation/reductive epoxide opening strategy.⁴⁸ Zhang et al. in 2007 has also reported the synthesis of the L-digitoxose sugar unit from Osmundalactone, following Wang’s procedure. However lactone 94 was synthesized from commercially available peracetyl L-rhamnal 95 by employing BF₃•OEt₂ induced epoxidation (Scheme 18).⁴⁹

Scheme 18.

Synthesis of L-digitoxose via Osmundalactone

In 2007, Jakeman reported synthesis of the L-digotoxose sugar unit based on a method developed by Brimacombe and co workers (1982). Their approach utilized a Klemer and Rodemeyer⁵⁰ elimination on fully protected L-rhamnopyranoside 102.⁵¹ The stereoselectivity was based on preferential deprotonation of the quasi-axial hydrogen which cleaves the benzylidene acetal protection forming 2-deoxy-3-ketosugar 101. The ketone of 101 was stereoselectively reduced with NaBH₄ to give the desired L-digitoxose stereochemistry. The major drawbacks to this method were the formation of side products from the S_N2 reaction of n-BuLi at the anomeric position and deprotonation at C-2 and the benzylidene ring. Jakeman and Timmons improved the yield of the Klemer-Rodemeyer elimination reaction by changing the base to s-BuLi and switching to an isopropylidene protecting group.⁵² They also reported a new method with improved yield to obtain the key intermediate, 2-deoxy-3-ketosugar 101 from L-rhamnose 98 (Scheme 19).

Scheme 19.

Synthesis of L-digitoxose from from L-rhamnose

In contrast to traditional approaches to carbohydrates from carbohydrates with pre-existing stereocenters, the O’Doherty group has been interested in the synthesis of carbohydrates from achiral starting materials.⁵³ These de novo asymmetric approaches have been developed from both achiral furans⁵⁴ and dienoates.⁵⁵ In 2010, they reported a de novo synthesis of L-digitoxose from the achiral and inexpensive acyl furan 104. Asymmetric Noyori reduction and Achmatowicz rearrangement gave pyranoside 105, The diastereomers 106 and 107 were separated after Boc-protection. Stereoselective palladium glycosylation of 106, followed by Luche reduction and NIS mediated iodo-carbonate formation installed the digitoxose stereochemistry in 110. Radical deiodination using tristrimethylsilylsilate (TTMSS) gave the desired L-digitoxose glycosyl donor 111 (Scheme 20).²⁸

Scheme 20.

de novo synthesis of L-digitoxose sugar

7. Total synthesis of jadomycin

To date, there have been only two synthetic efforts reported towards the total synthesis of jadomycins. In 2010, Ishikawa et al.⁴¹ reported the synthesis of dimethyl-protected jadomycin A Their routed used α-spirolactonydihydroquinone 113 as the key intermediate for construction of Jadomycin ring system. Retrosynthetically, they envisioned construction of jadomycin skeleton by imine cyclization followed by oxazolone formation of amino-aldehyde 112. This cyclization precursor was obtained from the spirolactone intermediate 113, which in turn was prepared from bromobenzene derivative 115 and tetralone 116 (Scheme 21).

Scheme 21.

Retrosynthetic analysis of jadomycin A

Palladium-catalyzed coupling of O-bromobenzoate 115 and 5-methoxy tetralone 116 gave enol-lactone 117. A subsequent dihydroxylation and rearrangement produced α-spirolactonyltetralone 118. Benzylic bromination and S_N2 displacement was followed by allylic oxidation to give the masked naphthoquinone 113. As a model study to form the phenanthroviridin skeleton, 113 was treated with methylamine to form 2-amino quinone 119, which upon thermolysis generated 120. Unfortunately, under identical conditions when methyl isoleucinate was used instead of methylamine as a nitrogen source, the desired dehydrative cyclization did not proceed. The most likely reason for this is the unfavorable steric interaction between the carboxyl and amino-acid fragment in 121. Because of the free rotation around C-C bond, 121 obtains energetically more stable conformation where carboxyl and amino functional groups are not in close proximity to undergo dehydration/cyclization (Scheme 22).

Scheme 22.

Synthesis of key intermediate α-spirolactonyltetralone 113

Ishikawa next decided to investigate the use of Snieckus’s oxidative cyclization procedure³⁹ involving the intramolecular condensation of an amine and aldehyde to construct the B ring. The carboxyl group on the masked naphthoquinone 113 was reduced with BH₃•THF complex and the isoleucine methyl ester was hydrolyzed by lithium hydroxide to give 123. Allylic oxidation of 123 with MnO₂ to aldehyde promotes condensation of the secondary amine with aldehyde forming iminium. This iminium intermediate is then trapped by carboxylate to form the pentacyclic ring system in jadomycin (124). Unfortunately, all attempts to remove the methyl protecting groups were unsuccessful (Scheme 23).

Scheme 23.

Synthesis dimethyl jadomycin A

That same year O’Doherty et al. reported a successful synthesis of jadomycin A and the carbasugar analogue of jadomycin B.²⁸ The key transformation utilized to construct the B-ring in jadomycin was a 6-π-electrocyclic ring closure. Retrosynthetically it was envisioned that the B-ring in the phenanthridine core could be obtained by a 6-π-electrocyclic ring closure and the oxazolone ring could be constructed by an acid promoted cyclization reaction (Figure 4). The connection between C-7a and C-7b could be formed by Stille coupling of aryl stannane 127 and bromojuglone 128. These stannane and bromojuglone can be obtained from commercially available starting materials 125 and 5-hydroxy-bromojuglone respectively.

Figure 4.

Retrosynthetic analysis of jadomycin

The stannane 127 was prepared by directed ortho-metallation of 126 using MOM or BOM ether as the O-directing group. These metallation precursors were prepared from commercially available phenol 125. The other coupling partner 128 was prepared by benzylation of commercially available bromojuglone. Stille coupling was carried out with stannane 127 and unprotected or protected bromojuglones, 128 to give tricyclic adduct 129 (Scheme 24).

Scheme 24.

Stille coupling; synthesis of intermediate 129

The acetal protection on 129 was selectively removed with aqueous THF to generate aldehyde 131. This aldehyde was then converted to 6-π-electrocyclic ring closure precursor 136 by condensation with t-butyl isoleucinate 133. Unfortunately 136 did not undergo the expected electrocyclic ring closure. This unexpected result was attributed to the unfavourable alignment of the π-electrons due to steric congestion of the nitrogen substituent, which could lead to unfavourable alignment for the electrocyclic ring closure. Fortunately, when the MOM or BOM group was removed, the 6-π-electrocyclic ring closure readily occurred to give jadomycin precursor 135. Finally acid promoted formation of oxazolone ring with simultaneous debenzylation yielded jadomycin A as a 6:1 diastereomeric mixture (Scheme 25).

Scheme 25.

Synthesis of jadomycin A

The authors assert that the removal of the offending protecting group benefited the cyclization beyond the removal of a negative steric interaction. They suggest that the presence of H-bonding between C-7 hydroxyl and the C-8 carbonyl of 134 preferentially aligned the cyclization precursor in proper orientation, thus promoting electrocyclic ring closure to 138. Once the B ring in 138 is formed, a subsequent 1,3-proton transfer forms intermediate 139, which undergoes hydration under the reaction condition to form quinol 140. Quinol 140 undergoes air oxidation to give 135 (Scheme 26).

Scheme 26.

Plausible mechanism of ring closure

After a successful synthesis of jadomycin A, total synthesis of jadomycin B was attempted by the direct installation (Schmidt glycosylation, S_N2 alkylation, Pd-catalyzed glycosylation and Mitsunobu conditions) of a suitably protected sugar moiety (106, 140, 141 and 142). The required L-digitoxose sugar units were synthesized by the route in Scheme 20. Despite numerous attempts of jadomycin A and 129, they were never able to isolate a product of glycosylation. The failure to successfully glycosylate the jadomycin aglycon was attributed to the acid sensitivity of the glycosylated products, both to the reaction conditions and silica gel chromatography. This acid sensitivity has also been seen in jadomycin B (Scheme 27). In fact, only tangential signs (a distinctive wine-red color) of glycosylation between 129 and 140 could be seen under the Mitsunobu glycosylation conditions.

Scheme 27.

Attempted glycosylations

With the acid sensitivity precluding the total synthesis of jadomycin B, they decided to investigate the synthesis of a less acid labile cyclitols (5a-carbasugar) analogue of jadomycin B. Following a previously reported method,⁵⁶ quinic acid 143 was converted to dihydroxy ketone 144 in nine steps. Enone 145 was obtained by Boc-protection of 144 followed by base mediated elimination. This enone was selectively reduced to 146 using LiAlH₄ at low temperature. Myers rearrangement of allylic alcohol gave alkene 147. Stereoselective dihydroxylation and diol protection gave 148. Removal of Boc-protection with LiAlH₄ to obtain the Mitsunobu glycosylation precursor 149 (Scheme 28).

Scheme 28.

Synthesis of cyclitol equivalent of L-digitoxose

Mitsunobu glycosylation was first attempted with cyclitol donor 149 and jadomycin A as the acceptor. Unfortunately, only a complex mixture was obtained, which was attributed to the sensitive oxazolone ring. Successful glycosylation occurred when cyclitol acceptor 150 was used forming stable acetal 151. Global deprotection with aqueous HCl removed all the protecting groups affording 152. This set the platform for introduction of amino acid and successive ring closure. Condensation of t-butylisoleucinate 133 with aldehyde 152 promotes 6-π-electrocyclic ring closure forming the B ring 153. Finally an acid catalyzed lactonization was used to install the oxazolone ring and the carbasugar analogue of jadomycin B was synthesized as 2.5:1 diastereomeric mixture (Scheme 29).

Scheme 29.

Synthesis of carbasugar analogue of jadomycin B

8. Conclusions

A detailed account of the jadomycin biosynthesis and synthetic studies has been reviewed. These biosynthetic studies have lead to a better understanding of the biosynthetic pathway, which has led to the discovery of new analogues and improved production of various jadomycins for further biological evaluation. Similarly, the synthetic contributions have shed light on the reactivity and stability of the jadomycin ring system and natural product. In this regard, the access to the cyclitol analogue of jadomycin B allows access to stably analogues for further structure activity relationship (SAR) studies.

Figure 3.

Phenanthridine core in natural products