Abstract

With incredible bioactivities and fascinating structural complexities, tetramic acid- (TA-) containing natural products have attracted favorable attention among the organic chemistry community. Although the construction of the TA core is usually straightforward, the intricate C3-side chain sometimes asks for some deliberative strategy so as to fulfill an elegant total synthesis. This review mainly covers some exceptional synthetic examples for each type of natural product in recent years, showcasing the great achievements as well as unsettled obstacles in this area, in the hope of accelerating the synthetic and biological investigations for this unique type of natural product.

In memory of Wen-Ju Bai’s grandmother, Yumei Zhong

1. Introduction

Natural products containing a tetramic acid motif (Figure 1, red) are commonly seen in nature. These compounds, usually isolated from assorted terrestrial and marine organisms, exhibit important antibiotic, antitumor, or antiviral activities [14]. Their sophisticated structural complexities have attracted significant attention in organic chemistry community for total synthetic studies [58]. The TA core (2,4-pyrrolidinedione) originates from a variety of amino acids, typically displaying chirality at C5 position (Figure 1, inserted). Among their 4-O-alkyl ether variations, only the O-methylated form is seen in nature. Although there is not a standard way to categorize the TA-containing natural products, it is quite convenient to use different C3-side chains (Figure 1, blue) to classify their types. A TA bearing a C3-hydrido substitute can be frequently seen, for example, the dolastatin 15 [9]. In contrast, the C3-Me TA was not found until 2002 when palau’imide was isolated from the apratoxin producing marine cyanobacterium Lyngbya sp. in Palau [10]. Besides palau’imide, the C-3 Me TA only exists in the tetrapetalone family natural products that were isolated from soil bacteria Streptomyces sp. USF-4727 in Japan [1113]. The C3-acyl TAs are the most commonly found natural derivatives (Figures 1(c)–1(g)). They can be further categorized as the C3-multienoyl TA (polycephalin B [14]), C3-decalinoyl TA (paecilosetin [15]), macrocyclic TA (cylindramide A [16]), and C3-spiro TA (azaspirofuran B [17]). Some TA-containing natural products may be difficult to classify due to their structural complexities (Figure 1, phaeosphaeride B [18] and aspergilline A [1]). Several reviews have paid particular attention to the isolation and biological activity studies of TA-containing natural products together with short description of some laboratory syntheses [1921]. Their total syntheses, especially the latest ones (2010−2016), are rarely covered. As a result, this review predominantly focuses on the synthetic efforts of TA-containing natural products in recent years. Examples for each type of TA can be seen below.

2. Selected Preparative Methods

Among the various preparative ways to build the TA core, some notable asymmetric methods are listed here (Figure 2). The chirality of the TAs generally arises from the amino acid building block. The acidity of the C5 position of the TA core (p ≈ 12−17) should be particularly noticed during any asymmetric synthesis to avoid possible racemization, because it is known that exposure of a chiral TA with KOtBu leads to partial racemization even at −30°C in less than one minute [33]. For C3-acyl TA, the Lacey-Dieckmann procedure is most often adopted, whereby an amino acid derivative is converted into a 1,3-dione intermediate and then subjected to cyclization (Figure 2(a)) [34]. Jouin et al. has showed that coupling an amino acid with Meldrum’s acid can afford a cyclic intermediate that undergoes cyclization and decarboxylation resulting in the expected C3-H TA (Figure 2(b)) [35]. An answer to prepare the unusual C3-Me TA had proved to be more challenging due to the limitation of the two aforementioned methods. Bai et al. eventually answered this call via a mild SmI2-mediated radical cyclization, whereby the chirality of the emerging TA core is retained from the initial amino acid (Figure 2(c)) [36]. Its application to palau’imide synthesis was also disclosed. Interestingly, Yoshii and Moloney reported the transformation of the C3-H TA to the corresponding C3-acyl TA via an O-acylation followed by an acyl migration (Figure 2(d)) [20, 37, 38]. Albeit appealing, it is better to define the limitation and scope of this method before it can be widely applied for total synthesis. In addition, Schlenk et al. discovered a practical way to construct C3-enoyl TAs using a selective 3-acylation of C3-H TA with ylide Ph3PCCO followed by a Wittig olefination (Figure 2(e)) [39]. Overall, the Lacey-Dieckmann cyclization and Jouin’s method represent the most popularly employed strategies, as can be seen from the following synthetic examples.

3. C3-H TA (Sintokamide A)

Sintokamide A was isolated from the marine sponge Dysidea sp. collected in Indonesia (Scheme 1) [40]. As the first small molecule known to selectively block transactivation of the N-terminus of the androgen receptor in prostate cancer cells, sintokamide A served as a promising drug leading us to evaluate and develop a new approach to treating castration-recurrent prostate cancer. Its chemical structure was unambiguously determined by X-ray crystallography. Sintokamide A contains both dichloromethyl and trichloromethyl groups, making it a challenging target for total synthesis because of the limited chemical methods for stereoselective di- or trichloromethylation. To solve this issue, Beaumont et al. first designed a method to achieve a highly diastereoselective haloalkylation of titanium enolates [41] and then successfully applied it to this total synthesis [22, 42]. The divergent route commenced with the preparation of the chlorinated amino acids 6 and 11. The known oxazolidinone 1 underwent a direct ruthenium-catalyzed tri-/dichloromethylation to afford the chlorinated product 2/7 that was further converted to nitrile 3/8. These nitriles were then transformed into sulfinimine 4/9 which proceeded with a diastereoselective scandium-catalyzed Strecker reaction to yield the advanced nitrile 5/10. Upon hydrolysis followed by the corresponding amine or carboxylic acid protection, the desired precursors 6 and 11 were obtained. Subsequent peptide coupling was straightforward. And the obtained dipeptide 12 underwent a methyl ester hydrolysis to release the free carboxylic acid that was ready to couple with Meldrum acid with Jouin’s procedure. Upon heating, the obtained crucial intermediate 13 was converted to a free tetramic acid that underwent an O-methylation to produce sintokamide A (14). Notably, an epimerization occurred at C5 stereocenter of the TA core during this process, demonstrating the easy racemization nature of the tetramic acid. Overall, the synthesis features an asymmetric chloroalkylation, an asymmetric Strecker reaction, and Jouin’s method to construct the C3-H TA core.

4. C3-Me TA (Tetrapetalone A)

The tetrapetalone family, currently containing four derivatives, represents a rare type of C3-Me tetramic acid (Scheme 2). They all share a highly decorated tetracyclic skeleton appended to a β-D-rhodinose. Because of these challenging structural motifs and the synthetic complexities, both alone and together, a variety of synthetic groups have pursued their syntheses for well over a decade. To date, no one has claimed to finish the total synthesis of any member of the tetrapetalone family. However, some notable approaches are worth of praise.

In 2010, Marcus and Sarpong reported his synthetic approach leading to the tetracyclic core of tetrapetalones (Scheme 2) [23]. A Nazarov cyclization of aryl dienone 15 using AlCl3 in toluene at ambient temperature provided the aryl indanone 16. Subsequent epimerization with base accomplished the trans relationship of the Me and isopropenyl in 4 : 1 dr. This indanone was transformed to the primary alcohol 17 that was then subjected to excess of Dess-Martin periodinane (DMP) to trigger the formation of the seven-membered ring. Likely, this process involves a three-step sequence, including a DMP oxidation of the primary alcohol 17 to aldehyde A, an intramolecular aldol reaction to form a secondary alcohol B, and a succeeding DMP oxidation to generate the ketone 18. After a reductive pyrrole ethylation and an oxidation, the α,β-unsaturated lactam 19 was obtained. The following methylation of the ene-lactam moiety was reached with LDA and MeI to give intermediate 20. The resulting conjugate addition with boron pinacolato ester and immediate oxidation afforded compound 21. Finally, Swern oxidation finished the installment of the tetramic acid moiety to give the racemic tetrapetalone intermediate 22. Further reformulation of the seven-membered cyclic ketone into the cyclic alkene likely proved arduous due to the known propensity of tetramate motifs to undergo retro-aldol reactions. Overall, the C3-Me TA core was constructed from a pyrrole ring in six steps, distinct from any aforementioned convenient TA preparative method.

Carlsen et al. pronounced progress by accomplishing the first racemic synthesis of tetrapetalone A-Me aglycon, a known degradation product of tetrapetalone A (Scheme 3) [24]. Similarly, the starting aniline 23 was prepared via an AlCl3-catalyzed Nazarov cyclization. Like Sarpong’s observation, Carlsen et al. also realized that using a Buchwald-Hartwig cross-coupling to install the C-N bond was problematic. It proved that only the nonsterically congested NH3 could deliver the expected aryl amine 24 that was further converted to a divinyl oxazolidinone 25. The resulting diastereoselective RCM reaction was carefully examined to obtain an effective seven-membered ring formation so as to yield the oxazolidinone 26. It was subsequently transformed to alcohol 27 that underwent a Ley oxidation to give the aldehyde 28. Its exposure to Stryker’s reagent followed by Swern oxidation smoothly furnished the chlorinated tetramate 29. Further dechlorination and O-methylation afforded the advanced intermediate 30 that proceeded with an atypical dearomatization [43] to complete the synthesis of tetrapetalone A-Me aglycon. A single-crystal X-ray diffraction experiment unambiguously secured the molecule’s structural assignment. Generally, the desired O-methylated tetramate derivative 30 was synthesized from the oxazolidinone 26 in ten steps.

Carlsen et al.’s synthetic effort represents a milestone for the tetrapetalone synthesis. But achieving the total synthesis of any member of the tetrapetalone family is still thought-provoking [4446]. Conditions for cleaving the methyl ether on the tetramic acid core and the glycosylation should be mild enough to tolerate the complexity of the tetracyclic intermediate. More importantly, developing asymmetric method for tetrapetalone synthesis is not easy or trivial.

5. C3-Multienoyl TA (Tirandamycin and Streptolydigin)

The novel dienoyl TA tirandamycin C was isolated from the marine environmental isolate Streptomyces sp. 307-9, displaying interesting bioactivity against vancomycin-resistance (Scheme 4) [47]. Obviously, the racemic tetramic acid fragment 33 could be easily accessed by subjecting phosphonate 31 to glycine methyl ester to provide the amide 32 that further proceeded with a typical Lacey-Dieckmann cyclization and a DMB protection (Schlessinger et al. had prepared similar tetramic acid core via the same method in his synthesis of tirandamycin A; see [48]) [49]. In contrast, the bicyclic ketal fragment 36 asks for precautious design to achieve the desired stereochemistry presented in tirandamycin C [25, 48, 5056]. The starting homoallylic alcohol 34, prepared from the mismatched double asymmetric stannylcrotylboration of an aldehyde, was treated with excess MeLi at low temperature to trigger the formation of the lactol intermediate C. Subsequent ketalization occurred with catalytic amount of pPTS, providing the bicyclic ketal 35 that underwent a cross metathesis with methacrolein to deliver the aldehyde 36 with excellent E/Z selectivity. Exposure of tetramic acid fragment 33 to KOtBu followed by addition of aldehyde 36 provided the N-DMB protected tirandamycin C. The resulting TFA deprotection released the desired natural product.

Another synthetic example can be seen in Pronin and Kozmin’s synthesis of streptolydigin (Scheme 5) [26]. Isolated from Streptomyces lydicus, streptolydigin was found to inhibit bacterial RNA polymerase but not eukaryotic RNA polymerase [57]. The starting chiral alcohol 37 proceeded with a TEMPO-catalyzed oxidative cyclization followed by a Boc-deprotection to give the cyclic amine 38. The following N-glycosylation with the rhodinose and Ag-promoted N-acylation with the thioester provided the phosphonate fragment 39. Meanwhile, a cross metathesis was deployed to obtain the lactone 41. Similarly, conversion of lactone 41 to the corresponding Weinreb amide followed by successive treatment with MeLi and TsOH triggered the intramolecular ketalization, providing the bicyclic ketal 42. It was then transformed to the aldehyde 43, which underwent a one-pot reaction with phosphonate 39 to cleanly yield streptolydigin, resulting from a Horner-Wadsworth-Emmons olefination, a Lacey-Dieckmann cyclization, and a desilylation.

As seen from the aforementioned two synthetic examples, it is the construction of the intricate C3-side chains rather than the tetramic acid cores that plays a pivotal role in total synthesis of natural products containing a C3-multienoyl TA motif.

6. C3-Decalinoyl TA (JBIR-22)

JBIR-22, specifically inhibiting proteasome assembling chaperone homodimerization in vitro, could be a useful primary compound for developing anticancer drugs [58, 59]. Due to its important bioactivity and unsolved stereochemical assignment, Healy et al. developed an elegant synthetic route (Scheme 6) [27]. One challenging problem is the lack of straightforward method to build the fragment 4,4-disubstituted glutamic acid derivative 46. This goal was achieved by reacting compound 44 with ethyl dimethylpyruvate to diastereoselectively give the lactone 45 which underwent a subsequent N-methylation and cleavage of the N-sulfinyl group. In contrast, the trans-fused decalin aldehyde 48 was constructed from trienal 47 via an organocatalytic intramolecular Diels-Alder (IMDA) reaction using MacMillan’s conditions. This aldehyde was then converted to the β-ketothioester 49 via an aldol reaction followed by a DMP oxidation. Its following Ag-mediated coupling reaction with the fragment 46 neatly delivered the intermediate 50, which underwent a Lacey-Dieckmann cyclization and microwave-assisted ester hydrolysis to provide JBIR-22. The absolute stereochemistry was finalized through total synthesis of both possible diastereomers. Overall, this synthesis relies on a practical way to build the unnatural 4,4-disubstituted glutamic acid as well as an IMDA reaction to construct the trans-fused decalin system.

7. Macrocyclic TA (Aburatubolactam A)

Aburatubolactam A was isolated from the culture broth of Streptomyces sp., SCRC-A20 in Japan (Scheme 7) [60]. One synthetic challenge lies in the construction of the bicycle[3.3.0]octane ring. This issue was easily answered by Henderson and Phillips through a tandem ring-opening/ring-closing metathesis of functionalized bicyclo[2.2.1]heptane to obtain the starting cyclopentanone 51 [28]. Subsequent enolate acylation with Mander’s reagent followed by a reduction of the ketone with NaBH4 gave an alcoholic intermediate, which underwent a mesylation with MsCl and an immediate elimination to safely yield the product 52. It was then transformed to the bicycle 53 by employing Majetich’s fluoride-mediated Sakurai allylation and subsequent epimerization to favor the desired syn-relationship. Further elaboration to compound 54 was realized by ester reduction with LAH and cross metathesis with Piv-protected butene-1,4-diol. The carbon side chains of 54 were further extended to deliver the dioxenone 55, which upon heating with amine 56 under reflux produced a sensitive β-ketoamide intermediate. This delicate intermediate then underwent Stille coupling with tert-butyl-β-iodoacrylate followed by the Lacey-Dieckmann cyclization that led to tetramic acid 57. Subsequent removal of the Boc and tert-butyl ester groups with TFA, macrolactamization, and cleavage of the TBS group with HF effectively provided aburatubolactam A.

8. C3-Spiro TA (Fusarisetin A)

Fusarisetin A, isolated from a soil fungus Fusarium sp. FN080326, displays potent inhibition of metastasis in MDA-MB-231 cells without any significant cytotoxicity [61]. Its intricate chemical structure possesses a 6,6,5,5,5-fused pentacyclic ring system and bears ten stereogenic centers. Due to the unprecedented complexity and remarkable bioactivity, fusarisetin A has attracted considerable attention in synthetic community (Schemes 811).

Deng et al. concisely achieved the first asymmetric total synthesis of fusarisetin A and also established its absolute configuration as (+)-fusarisetin A by spectroscopic and optical rotation analysis (Scheme 8) [29]. The trans-fused decalin derived β-ketothioester 59 was accomplished by a Lewis acid catalyzed IMDA reaction from the ketothioester 58 that originated from commercially available chiral building block (S)-(−)-citronellal. The cyclopentanone 61 was formed by an Ag-catalyzed O-allylation to produce the kinetically stable intermediate 60 followed by its Pd-catalyzed O to C rearrangement that ultimately gave the thermodynamically stable product. Subsequent peptide formation with amino acid 62 and Wacker oxidation provided the methyl ketone 63 that underwent a Luche reduction to deliver the alcoholic intermediate 64 with acceptable diastereoselectivity. Finally, a typical Lacey-Dieckmann cyclization concurrently with a hemiketalization formation afforded the (−)-fusarisetin A, even though partial racemization of the tetramic acid was observed during this process.

Unlike Deng et al.’s approach, Xu et al.’s synthesis featured an impressive one-pot radical cyclization/aminolysis, albeit employing a similar IMDA strategy to build the starting trans-fused decalin 66 (Scheme 9) [30]. The ensuing Reformatsky reaction produced a secondary alcohol that was further oxidized with IBX to β-ketoester 67. Its deprotonation with LiHMDS followed by addition of TEMPO and ferrocenium hexafluorophosphate led to the key intermediate 68 as inseparable diastereomers. Upon heating 68 in the presence of the amino acid 62, the radical cyclization smoothly occurred via the intermediacy of radical D to yield the tricyclic product 69 that then proceeded with an aminolysis with amino acid 62 to conveniently give the advanced intermediate 70. Subsequent cleavage of the N-O bond with Zn/AcOH followed by the classic Lacey-Dieckmann cyclization afforded the (−)-fusarisetin A.

Distinct from the two above-mentioned approaches, Yin et al.’s synthesis revealed a remarkable biomimetic oxidation process (Scheme 10) [31]. Aldehyde E, analogously produced through an IMDA reaction, proceeded with a Roskamp reaction to give the β-ketoester 72. Succeeding aminolysis with TBS-protected ent-62 followed by a Lacey-Dieckmann cyclization generated a tetramic acid intermediate that underwent a TBS-cleavage to release the free alcoholic equisetin, another TA-containing natural product. Further oxidation with Mn(OAc)3 formed the radical F, which was trapped by O2 to give a peroxy radical G. This radical then underwent a cyclization to yield peroxy radical H followed by intermolecular hydrogen abstraction to deliver the peroxyfusarisetin (73). The final Zn reduction in HOAc gave the naturally occurring (+)-fusarisetin A for the first time.

Extra notable synthesis came from Liu et al., highlighting a one-pot-four-step cascade reaction (Scheme 11) [32, 62]. The trans-fused decalin aldehyde I, prepared from the triene 74 via an IMDA reaction, underwent an intramolecular Mukaiyama aldol reaction to give the tetracyclic intermediate K, which proceeded through a TMS-deprotection and a PCC oxidation to conveniently provide the tetracyclic compound 75. Subsequent lactone opening followed by aminolysis with amino acid ent-62 produced the expected key intermediate 77, though a significant racemization occurred during this process. The final Lacey-Dieckmann cyclization finished the total synthesis of (+)-fusarisetin A.

9. Conclusion and Outlook

As rich sources for remarkable bioactivity, tetramic acid-containing natural products also provide the organic chemistry community with a variety of fascinating synthetic targets with their complicated molecular architectures. Overall, the Lacey-Dieckmann cyclization and Jouin’s method represent two robust and prevalent methods to asymmetrically construct the C3-acyl and C3-H tetramic acid core, respectively. The challenges to synthesize these natural products sometimes lie in the intricate C3-side chains rather than the tetramic acid cores, and overcoming these issues depends on the deliberative design of novel strategies enabling the rapid construction of these C3-side chains. Occasionally, biomimetic synthesis can significantly expedite the process of their total synthesis. However, there are still unresolved complications, such as tetrapetalones, calling for innovative and efficient approaches. Meanwhile, the total synthesis of TA-containing natural products examines current synthetic methods for their atom- and step-economic assembly. Accordingly, the shortages and limitations of existing methods can be realized, and in doing so they stimulate new methodology developments. In this review, we have summarized some outstanding synthetic routes in the past five years, in the hope that we might accelerate the synthetic and biological studies for this exceptional type of natural product.

Competing Interests

The authors declare that they have no competing interests.

Acknowledgments

Wen-Ju Bai is appreciative of the National Award for Outstanding Self-Financed Chinese Students Study Abroad provided by China Scholarship Council. Xiqing Wang appreciates municipal and provincial programs for innovative talents from Yangzhou ([2015]3) and Jiangsu ([2015]26).