Natural product anticipation through synthesis

Abstract

Natural product synthesis remains one of the most vibrant and intellectually rewarding areas of chemistry, although the justifications for pursuing it have evolved over time. In the early years, the emphasis lay on structure elucidation and confirmation through synthesis, as exemplified by celebrated studies on cocaine, morphine, strychnine and chlorophyll. This was followed by a phase where the sheer demonstration that highly complex molecules could be recreated in the laboratory in a rational manner was enough to justify the economic expense and intellectual agonies of a synthesis. Since then, syntheses of natural products have served as platforms for the demonstration of elegant strategies, for inventing new methodology ‘on the fly’ or to demonstrate the usefulness and scope of methods established with simpler molecules. We now add another aspect that we find fascinating, viz. ‘natural product anticipation’. In this Review, we survey cases where the synthesis of a compound in the laboratory has preceded its isolation from nature. The focus of our Review lies on examples where this anticipation of a natural product has triggered a successful search or where synthesis and isolation have occurred independently. Finally, we highlight cases where a potential natural product structure has been suggested as a result of synthetic endeavours but not yet confirmed by isolation, inviting further collaborations between synthetic and natural product chemists.

The total synthesis of natural products has always been a major goal in organic chemistry. The reasons for pursuing it have evolved as the field has progressed. In its early history, total synthesis mostly served to confirm the constitution and configuration of readily available natural products. With the advent of X-ray crystallography, NMR spectroscopy and mass spectrometry, this aspect has become less important, although numerous recent cases exist where the structure of a natural product was settled as a result of total synthesis^1,2. Consequently, the emphasis in the field has shifted more towards reaction development and the definition of efficient synthetic strategies. In some cases, the desire to achieve a particular transformation en route to a natural product has led to the invention of new reactions or reagents that did not exist previously³. If a total synthesis is suitably efficient, it might also be used to deliver a prized natural product on a scale that can other wise only be procured at great expense or by ignoring environmental concerns⁴. Many other motivations for total synthesis exist, ranging from its value as a training ground for medicinal chemists to the satisfaction that comes with solving the sheer intellectual challenge that it represents^5,6. In this account, we wish to highlight yet another reason to pursue it: natural product anticipation.

In the early days of organic synthesis, there were undoubtedly many cases where a compound was prepared in the laboratory and considered ‘synthetic’ that was subsequently identified as a natural product. For instance, using his eponymous method, Gabriel made aminoacetone (1) in 1893 (REFS^7,8). It was not until 1959 that the molecule was detected in nature, when Elliott identified it as a metabolite of Staphylococcus aureus⁹ (FIG. 1). Similarly, the amino acid proline (2) was first synthesized and characterized by Willstätter, in racemic form, during his studies on coca alkaloids in 1900 (REF.¹⁰). Shortly afterwards, and before Willstätter could publish his work, Fischer identified proline as a hydrolysis product of albumin and it was, thus, added to the canon of proteinogenic amino acids^11,12 (FIG. 1). A more recent, and considerably more complex, case is that of 13²,17³-cyclopheophorbide enol (3), a porphyrin synthesized in the Eschenmoser laboratory as early as 1971 and isolated as a natural product from the sponge Darwinella oxeata in 1986 (REFS^13,14) (FIG. 1). It was also found in 1999 as a ‘molecular fossil’ in various marine sediments¹⁵. In 1963, Bell and Ireland published synthetic studies towards the diterpene alkaloid (+)-atisine¹⁶. They arrived at the racemic hydrocarbon (±)-4, which contained the full carbon skeleton of their target. The exocyclic double bond of (±)-4 could also be isomerized to give the endocyclic alkene (±)-5 (REFS^16,17). Around the same time, Zalkow and Girota reported the first part of their synthetic efforts towards (+)-atisine (6), where they prepared intermediate (+)-4 in optically pure form^18,19. Two years later, Dev and colleagues found enantiomerically pure (−)-4 and (−)-5 in Erythroxylon monogynum and named the hydrocarbons (−)-atisirene and (−)-isoatisirene, respectively²⁰ (FIG. 1). Since complex terpenoids are generally biosynthesized from reduced precursors, one wonders whether Bell, Ireland and Zalkow suspected that their synthetic intermediates could be genuine natural products and, thus, how much of a surprise their subsequent isolation was.

Fig. 1 |. Notable examples of ‘unwitting’ natural product anticipation^7–20.

For these early examples, there is no evidence to suggest the synthetic chemists envisaged that these structures would be later identified as natural products. The year and corresponding author are highlighted in blue for the reported synthesis and in green for the subsequent isolation.

Numerous other examples of such ‘unwitting’ discoveries exist. In this account, however, we will focus on what we consider the most satisfying type of natural product anticipation, wherein synthetic compounds were first made in the laboratory, suspected to occur in nature and subsequently confirmed as genuine natural products. Such predictions are usually based on biosynthetic considerations, the existence of analogous compounds or on the reactivity of a natural product that was previously unrecognized. Hence, they often originate from biomimetic (or ‘bioinspired’) syntheses, which attempt to emulate certain patterns found in nature. These anticipated natural products are typically observed in the same natural source as the originally investigated natural product. In addition to confirmed cases, we will list compounds that have been anticipated but, for various reasons, not yet isolated. Our hope is that several of these compounds will be revealed as true natural products in the not-too-distant future. We do not cover natural product anticipation here that is based on genome mining and molecular network analysis. These powerful computational methods can predict not only the existence of natural products and their constitution but, in some cases, even their configuration and 3D structure^21–23. While fascinating, this aspect of anticipation is beyond the scope of this Review.

Unexpected products

In 2010, Lee and colleagues disclosed a biomimetic total synthesis of guajadial (11), the prototypical member of the caryophyllene-derived family of meroterpenoids isolated from Psidium guajava (the common guava)^24,25. The biosynthesis of guajadial (11) was proposed to involve a hetero-Diels–Alder reaction between ortho-quinone methide 10 and caryophyllene (7)²⁵ (FIG. 2a). To mimic this process in the laboratory, diformylphloroglucinol 9 was slowly added to an aqueous mixture of benzaldehyde (8) and caryophyllene (7). This gave guajadial (11), together with two unexpected diastereomers 12 and 13. It was known that caryophyllene (7) adopts two major conformations in solution, the βα and ββ conformers, which differ in the relative disposition of the methylene and alkenyl methyl group²⁶ (FIG. 2c). Therefore, Lee and colleagues rationalized that guajadial (11) and its diastereomer 12 originate from hetero-Diels–Alder reactions involving the major conformer βα−7, whereas diastereomer 13 results from the minor conformer ββ−7 (FIG. 2a). The isolated yields of 11 and 12 (from βα−7) compared with 13 (from ββ−7) are in good agreement with the reported βα/ββ conformation populations of caryophyllene (7) (FIG. 2c). Given the biomimetic nature of this reaction, Lee and colleagues already suspected that 12 and 13 might represent as-yet-undiscovered natural products. Indeed, while they were preparing the manuscript describing their synthesis, isomer 12 was isolated from P. guajava by Zhang and colleagues and named psidial A (REF.²⁷). In 2017, isomer 13 was also isolated from the leaves of P. guajava by the Yin group and named psiguajadial L (REF.²⁸) (FIG. 2a).

Fig. 2 |. Anticipation of caryophyllene-derived meroterpenoids from Psidium guajava.

a | Anticipation of psidial A (12) and psiguajadial L (13) through a multicomponent biomimetic reaction^24,25,27,28. b | Anticipation of the psiguajanones A–D (17–20) through a multicomponent biomimetic reaction, followed by reduction²⁹. c | The βα and ββ conformers of caryophyllene (7)²⁶. P. guajava, Psidium guajava.

In 2019, the Xie group reported the synthesis of the dimeric caryophyllene meroterpenoid psiguajdianone (21), which they had isolated from P. guajava²⁹ (FIG. 2b). Following similar biomimetic logic to Lee and colleagues, a Knoevenagel condensation of phloroglucinol derivative 15 with para-formaldehyde (14) led to the transient formation of a rapidly interconverting mixture of tautomeric ortho-quinone methides 16a and 16b (REF.²⁴) (FIG. 2b). Hetero-Diels–Alder reactions between these tautomers (16a and 16b) and caryophyllene (7), in either its ββ or βα form, gave four different cycloadducts. Following reduction of the formyl group using NaBH₃CN, four isomeric products 17–20 were isolated, one of which (20) was the proposed precursor towards psiguajdianone (21). Treatment of 20 with AgOAc then gave the desired dimer 21. Guided by their synthetic samples, they successfully isolated all four monomers 17–20 from P. guajava and named them psiguajanones A–D (FIG. 2b).

(±)-Incarviditone (24) is a racemic natural product isolated in 2009 from Incarvillea delavayi by Zhang and colleagues³⁰ (FIG. 3a). It formulates as a homochiral dimer — a combination of two ‘like’ enantiomers — of the co-isolated natural product (±)-rengyolone (22). Intrigued by homochiral selectivity in a presumably non-enzymatic biogenesis, Lawrence and colleagues investigated the feasibility of a proposed domino oxa-Michael/Michael dimerization of (±)-rengyolone (22)³¹ (FIG. 3a). Treatment of 21 with sub-stoichiometric K₂CO₃ in (CH₂Cl)₂ successfully gave (±)-incarviditone (24) in 19% yield. An even more complex dimer, (±)-25, originating from the coupling of two ‘unlike’ enantiomers of (±)-rengyolone (22), was isolated in 23% yield. This heterochiral dimerization follows the same oxa-Michael/Michael cascade proposed for (±)-incarviditone (24), but the heterochiral dimer (±)-27 undergoes a subsequent aldol reaction to give (±)-25 (FIG. 3a). During the preparation of a manuscript detailing this total synthesis, the Zhang group disclosed the isolation of the heterochiral product (±)-25, which they named incarvilleatone, from Incarvillea younghusbandii³². (±)-24 and (±)-25 belong to a growing number of natural products that have been isolated as true racemates³³.

Fig. 3 |. Anticipation of incarvilleatone, mesitylene and nagelamide E.

a | Homochiral dimerization of (±)-rengyolone (22) gives the intended target, (±)-incarviditone (24), whereas heterochiral dimerization gives the anticipated natural product, (±)-incarvilleatone (25)^30–32. b | Photochemical retro-[2+2] cycloaddition of the SNF-4435C and SNF-4435D (33 and 34) gives mesitylene (36) and orinocin (35)^34–38. c | Vinylcyclobutane rearrangement of sceptrin gives the intended target ageliferin (38) and the anticipated natural product nagelamide E (39)^40–42. I. younghusbandii, Incarvillea younghusbandii; S. orinoci, Streptomyces orinoci.

Spectinabilin (29) is an unusual nitroaryl-containing tetraene that was found in Streptomyces spectabilis, together with the two bicyclo[4.2.0]octadienes SNF-4435C (33) and SNF-4435D (34)^34–36 (FIG. 3b). The isomeric nature of these compounds led Trauner and Beaudry to suggest that irradiation with sunlight stimulates (E,E,E,Z)-configured spectinabilin (29) to undergo conversion into the (E,Z,Z,Z)-polyene 30, which then undergoes a thermal 8π–6π electrocyclization cascade³⁷. Subsequently, Hertweck investigated the fermentation broth of S. orinoci, a related bacterial species that produces spectinabilin (29)³⁸. When the fermentation was carried out in the dark, no SNF-4435C or SNF-4435D (33 or 34) was detected. However, when the culture was exposed to daylight and artificial light at room temperature, the bicyclo[4.2.0]octadienes were also formed in S. orinoci. Interestingly, irradiation of purified (+)-spectinabilin (29) gave SNF-4435C and SNF-4435D (33 and 34) and a truncated spectinabilin analogue 35, which was named ‘orinocin’ (FIG. 3b). It was proposed that, under irradiation, the 8π–6π electrocyclization cascade continues with a light-mediated retro-[2+2] cycloaddition, forming orinocin (35) via the extrusion of mesitylene (36). Reinvestigation of the fermentation broth with liquid chromatography and gas chromatography–mass spectrometry indeed led to the detection of orinocin (35), as well as mesitylene (36). This confirmed mesitylene as a polyketide natural product formed through a photochemical ‘polyene-splicing’ reaction.

Marine-derived dimeric pyrrole–imidazole alkaloids, such as ageliferin (38) and palau’amine, have attracted significant interest from the synthetic community³⁹. In 2004, Baran et al. reported a biomimetic vinylcyclobutane rearrangement of sceptrin (37) to give ageliferin (38)⁴⁰ (FIG. 3c). This synthetic evidence gave support to their hypothesis that ageliferin (38) was not the result of a [4+2] cycloaddition but was instead a rearrangement product of sceptrin (37), which can be seen as a [2+2] cycloadduct of hymenidin (40)⁴¹ (FIG. 3c). When the biomimetic vinyl cyclobutane rearrangement was conducted on a larger scale, a minor product, epi-ageliferin (39), was isolated as well. Whilst this synthetic work was ongoing, Kobayashi and colleagues investigated extracts from the Okinawan marine sponge Agelas sp. and found a new family of dimeric pyrrole–imidazole alkaloids, the nagelamides⁴². The structure of one of these metabolites, nagelamide E, matched epi-ageliferin (39). Notably, the ratio of nagelamide E to ageliferin was similar when isolated from the natural source (1:24) and when prepared synthetically (1:20) (FIG. 3c).

The unusual alkaloid exiguamine A (43) was isolated as a racemate from the marine sponge Neopetrosia exigua and was shown to be a potent indoleamine 2,3-dioxygenase inhibitor⁴³. Intrigued by its unusual structure, Trauner and colleagues embarked upon its total synthesis^44,45. Their biosynthetic hypothesis stipulated that the simple starting materials tryptophan, glycine and dopamine come together to yield ortho-quinone methide intermediate (42), which would undergo an oxa-6π electrocyclization to form (±)-exiguamine A (43) (FIG. 4). When catechol 41 was exposed to 10 equivalents of silver (II) oxide, under acidic conditions, (±)-exiguamine A (43) was formed (FIG. 4). However, when 20-fold excess of silver (II) oxide was used, a new, hydroxylated derivative (47) was isolated as a single diastereomer. After communication with the isolation chemist, Andersen, 47 was subsequently found in N. exigua and named exiguamine B (REF.⁴⁵). A biosynthetic pathway, supported by discrete Fourier transform calculations, was proposed that explains the formation of exiguamine A or B from the bis-quinone intermediate 44, a tautomer of 42. oxa-6π electrocyclization of 44 places an oxygen at the benzylic position of the ortho-quinone 45, which, in the presence of a large excess of oxidant, can be irreversibly intercepted through oxidation and tautomerization to give ortho-quinone methide 46, which undergoes a final oxa-6π electrocyclization to yield exiguamine B (47) (FIG. 4).

Fig. 4 |. Anticipation of exiguamine B.

Oxidation of catechol 41 with 10 equivalents of AgO gives the intended target exiguamine A (43), whereas the use of 20 equivalents gives the anticipated natural product exiguamine B (47)^43–45. N. exigua, Neopetrosia exigua.

‘Missing’ natural products

The xanthanolides are a large family of sesquiterpenoids that usually contain a γ-butyrolactone fused to a seven-membered ring. They include pungiolides A (49), B (50) and E (51), which evidently stem from Diels–Alder dimerization of 8-epi-xanthatin (48), followed by isomerizations and oxidations^46–50 (FIG. 5a). Tang and colleagues disclosed the total synthesis of various monomeric xanthanolides, including 8-epi-xanthatin (48) and its epimer xanthatin (52), which features a trans-fused butyrolactone^51,52. With 52 in hand, they investigated the formation of dimers analogous to pungiolide E (51), assuming that this epimer would undergo analogous dimerizations. Heating xanthatin (52) yielded the dimer 53 via a thermal ‘head-to-tail’, endo-selective Diels–Alder reaction (FIG. 5b). By contrast, under photochemical conditions, xanthatin (52) dimerized to give the ‘head-to-head’ dimeric xanthanolide 55. This outcome was rationalized by assuming that that irradiation led to isomerization of the C1–C5 double-bond to form a highly reactive trans-cycloheptene, which could then undergo a ‘head-to-head’ Diels–Alder homodimerization. The resulting intermediate 54 then underwent an intramolecular [2+2] cycloaddition to form 55 (FIG. 5b). Since no dimeric xanthanolides based on xanthatin (52) were known at the time of the investigations, the authors reinvestigated the natural source, Xanthium mogolium, a medicinal plant found in Northeast China. Remarkably, they isolated the predicted natural products 53 and 55, which they named mogolides A and B, respectively.

Fig. 5 |. Anticipation of ‘missing’ dimeric natural products.

a | Known dimeric xanthanolides natural products (49–51) and their biosynthetic monomer 8-epi-xanthatin (48)^46–49. b | Dimerization of xanthatin (52) leads to the anticipated natural products mogolides A and B (53 and 55)⁵². c | Known bisanthraquinone natural products (56–58)^53–57. d | Oxidative dimerization of monomer 59 leads to the anticipated natural product 2,2′-epi-cytoskyrin A (62)^58–60. e | Monomer 63, used by Nicolaou et al. to access rugulosin (56)^58,59. X. mogolium, Xanthium mogolium.

The bisanthraquinone natural products rugulosin (56), graciliformin (57) and cytoskyrin A (58) are of fungal and lichen origin and show marked bioactivities^53–57 (FIG. 5c). Rugulosin (56) and graciliformin (57) differ in their configurations at C2 and C2′ and are homodimers of a methyl-substituted anthraquinone. Cytoskyrin A (58) is a homodimer of a similar, methoxy-substituted anthraquinone, the configuration of which corresponds to graciliformin (57). By analogy to rugulosin, a second methoxy-substituted dimer could exist in nature. In 2005, the Nicolaou group reported the biomimetic synthesis of (+)-rugulosin (56) through an oxidative dimerization of 63 (REFS^58,59) (FIG. 5e). Using similar conditions, they also dimerized methoxy-anthraquinone 59, which led to the then unknown 2,2′-epi-cytoskyrin A (62) (FIG. 5d). A year later, the Shibuya group found (+)-62 in the fungus Diaporthe sp., confirming the suspected existence of the second dimer in nature⁶⁰. Interestingly, graciliformin (57) and cytoskyrin A (58), which bear the secondary hydroxy groups in an endo position, have not yet been synthesized in the laboratory.

Trauner and Miller pursued the biomimetic synthesis of pyrone natural products isolated from the sacogl ossan mollusc Placobranchus ocellatus^61,62. To this end, (E,Z,Z,E)-tetraene 66 was prepared using a Stille–Liebeskind coupling of alkenyl stannane 64 with alkenyl iodide 65 (FIG. 6a). The resulting tetraene underwent an in situ 8π–6π electrocyclization cascade to give a 1:9 mixture of racemic bicyclo[4.2.0]octadienes, (±)-67 (ocellapyrone A) and (±)-68 (REF.⁶³). The latter was subjected to singlet oxygen, which gave the endoperoxide ocellapyrone B (69). Ruthenium-catalysed isomerization then yielded bisepoxide 70, the 14-methyl homologue of the known bis-epoxide elysiapyrone A (71), which had been isolated from the ‘sap-sucking’ sacoglossan sea slug Elysia diomedea. The ease with which endoperoxides can be converted into bis-epoxides, even in the absence of a transition metal catalyst, suggested that 70 may also occur in nature. Indeed, in 2020, Li, Guo and Nay reported the isolation of 14-methylelysiapyrone from P. ocellatus⁶⁴ (FIG. 6a).

Fig. 6 |. Anticipation of 14-methylelysiapyrone A and psychotriadine.

a | Biomimetic total synthesis of the anticipated natural product, 14-methylelysiapyrone A (70)^63,64. b | Known bis(cyclotryptamine) alkaloids with different isomeric scaffolds^65–69. c | Total synthesis of a newly anticipated piperidinoindoline-type bis(cyclotryptamine) alkaloid, psychotriadine (80)⁷¹. P. colorata, Psychotria colorata; P. ocellatus, Placobranchus ocellatus.

The bis(cyclotryptamine) alkaloids have been of interest since the isolation of the first congener in 1888 and numerous total syntheses, biosynthetic studies and isolations have been reported^65–69. The natural products share a common carbon skeleton but feature different heterocyclic ring systems (72–75)⁷⁰ (FIG. 6b). Garg, Garcia-Garibay and colleagues were intrigued by the fact that no family members with a piperidinoindoline structure had been isolated, although this is conceivable based on a postulated common biosynthetic precursor⁷¹. Accordingly, they set out to synthesize this type using a photodecarbonylation strategy. Thus, ketone 76 was subjected to irradiation in the solid state, the product of which, following deprotection, afforded the bispyrrolidinone 77 (FIG. 6c). N-methylation followed by azidation then yielded C₂-symmetric precursor 78. Reduction of the aryl azides to the corresponding anilines with concomitant transamidation, cyclocondensation and reduction of one of the two amidines gave the unsymmetrical piperidinoindoline 79. This intermediate could be oxidized to C₂-symmetric piperidinoindoline 80, which, like its precursor, was suspected to be a natural product. Indeed, upon re-examination of an extract from Psychotria colorata, 80 could be identified as a genuine natural product and was named psychotriadine. Incidentally, meso-chimonanthine 75 itself is an anticipated natural product, as it was isolated from Calycanthus floridus in 1967, just a few days after its synthesis from N-methyl tryptamine⁷².

Unexpected reactivity

Marine gastropods produce a large variety of polypropionate-derived natural products. These compounds are suspected to act as a sunscreen to protect the molluscs from ultraviolet light in the shallow waters of their natural habitat. (−)-Tridachiahydropyrone (83) was isolated in 1996 from E. crispata⁷³ (FIG. 7a). In 2009, Moses and colleagues reported a biomimetic synthesis of racemic tridachiahydropyrone, which involved a photoinduced alkene isomerization of the linear all-E polyene chain in 81, followed by a photochemical conrotatory 6π electrocyclization⁷⁴. Interestingly, a side product, termed ‘phototridachiahydropyrone’ (84) was isolated that was presumably formed by a subsequent [1,3]-sigmatropic rearrangement of 83. Ultraviolet light was found to be necessary to promote the shift of the side chain. In 2015, Gavagnin et al., who originally isolated (−)-tridachiahydropyrone (83), reinvestigated the extract of E. crispata and successfully isolated (−)-phototridachiahydropyrone (84)⁷⁵ (FIG. 7a).

Fig. 7 |. Anticipation of phototridachiahydropyrone and atrop-abyssomicin C.

a | Biomimetic total synthesis of the anticipated natural product phototridachiahydropyrone (84) via photochemical [1,3]-sigmatropic rearrangement of tridachiahydropyrone (83)^74,75. b | Total synthesis of the anticipated natural product, atrop-abyssomicin C (86) and its isomerization into the target natural product, abyssomicin C (89). It was also found that reduction of atrop-abyssomicin C (86) gives abyssomicin D (88), whereas reduction of abyssomicin C (89) gives iso-abyssomicin D (91), which is not a known natural product^77–79. E. crispata, Elysia crispata; PIFA, phenyliodine(III) bis(trifluoroacetate).

The antibiotic polyketide abyssomicin C (89) was isolated in 2004 from the marine actinomycete Verrucosispora strain AB-18–032 by Süssmuth and colleagues⁷⁶ (FIG. 7b). Structurally, abyssomicin C (89) possesses a strained 11-membered ring that contains a reactive α,β-unsaturated ketone and a core tetronate motif. The Nicolaou group set out to achieve the biomimetic total synthesis of this intriguing natural product^77,78. Treatment of intermediate 85 with phenyliodine(III) bis(trifluoroacetate) (PIFA) to effect a dithioketal deprotection did not yield the desired product 89 but a compound that was identified as an atropisomer 86 (FIG. 7b). Upon exposure to acidic CDCl₃, atrop-abyssomicin C (86) underwent gradual isomerization to abyssomicin C (89), which could be separated by high-performance liquid chromatography. During an attempted biomimetic conversion of abyssomicin C (89) into abyssomicin D (88) via conjugate reduction of the enone, followed by intramolecular Michael addition, the authors exclusively isolated a diastereomer named iso-abyssomicin D (91). The latter slowly isomerized into abyssomicin D (88) upon standing in ethanol. By contrast, treatment of atrop-abyssomicin C (86) with l-selectride gave abyssomicin D (88) directly (FIG. 7b). These results suggest that atrop-abyssomicin C (86) could also be a natural product that is enzymatically reduced and converted into abyssomicin D (88). Indeed, in 2007, Süssmuth, Fiedler and colleagues isolated atrop-abyssomicin C (86) as the main component from the culture broth of Verrucosispora AB-18–032, along with abyssomicin C (89)⁷⁹ (FIG. 7b). Upon high-performance liquid chromatography purification with acidic solvents, atrop-abyssomicin C (86) was depleted and abyssomicin C (89) was formed. Interestingly, iso-abyssomicin D (91) has yet to be found in nature.

Conclusion and outlook

We hope to have shown that natural product anticipation adds another facet to total synthesis research, further increasing its intellectual intrigue and practical value. Synthetic studies can provide important insights into the formation and reactivity of natural products, whilst delivering valuable synthetic samples to help steer targeted spectroscopic/spectrometric identification and chromatographic separation of new natural products.

This Review cannot be comprehensive, since it relies mostly on our own experience in the field of biomimetic natural product synthesis and is limited by difficulties in finding diffuse information in the vast chemical literature. Anticipated natural products are often not identified as such in writing, perhaps reflecting an innate reluctance of scientists to publish speculations. The true story behind their prediction and isolation is sometimes buried in personal accounts. We are, therefore, indebted to many colleagues and friends for their suggestions and insights, which were invaluable in collating the examples we have presented. There are many more cases of natural product anticipation that could not be covered in detail in this Review and a selection thereof is shown in FIG. 8a. A sample of anticipated natural products awaiting confirmation of their natural product credentials is shown in FIG. 8b. Hopefully, our account will stimulate further collaborations between synthetic and natural product chemists and lead to new examples of anticipated natural products to add to this already impressive list.

Fig. 8 |. Additional examples for anticipated natural products and suspected natural products awaiting confirmation.

a | iso-Epicolactone (92)^80,81, (+)-brevianamide Y (93)^82–84, (±)-deoxyisobruceol (94)^85–88 and (−)-prehalenaquinone (95)⁸⁹ are additional examples for molecules that were synthesized in the laboratory prior to their isolation. b | Cases of anticipated natural products that await isolation from natural sources: 8-epi-isoaplydactone (96)⁹⁰, dia-angiopterlactone B (97)⁹¹, biyouyanagin C (98)⁹², epi-pycnanthuquinone C (99)⁹³, 8-epi-homodimericin A (100)⁹⁴, intricarene side product (101)⁹⁵, protected dia-millingtonine (102)⁹⁶, dia-incargranine B aglycone (103)⁹⁷, diastereomer towards neonectrolides (104)⁹⁸, preuisolactone precursor (105)⁹⁹, nuphar alkaloid isomer (106)¹⁰⁰, side product towards (+)-norcembrene 5 (107)¹⁰¹, monolomaiviticin A (108)¹⁰², 2-epi-lankacyclinol (109)¹⁰³, 3,7-epi-massadine (110)¹⁰⁴, santarubin S (111)¹⁰⁵, epi-guajadial B (112)¹⁰⁶, Δ^23,24-perovskone (113)¹⁰⁷, epi-pungiolide A (114)⁵⁰ and iso-aspergilasine A (115)¹⁰⁸.