FULL REVIEW | Chinese Chemical Society

liuherotao007 2019-04-09

展开全文

Introduction

The practice of total synthesis of natural products has a rich history of achievements and benefits to science and society. ¹ ^– ⁴ A branch of organic chemistry, total synthesis was born simultaneously as the more general discipline of organic synthesis did in 1828 by Friedrich Wöhler’s ⁵ synthesis of urea, a one-carbon-atom naturally occurring molecule. Its countless benefits to science and society range from replicating the molecules of nature in the laboratory to the preparation and production of pharmaceuticals and dyes, from confirming or revising previously assigned or proposed structures of natural products to discovering and developing new synthetic strategies and methods, and from synthesizing designed molecules for biology and medicine to educating students in the art and science of making molecules and training them to develop their creativity and problem-solving skills. In this review article, we summarize the contributions of select total synthesis endeavors from our laboratories under the following classification subtitles: (1) replicating the molecules of nature in the laboratory through novel synthetic strategies, (2) confirming, disproving, predicting, and revising molecular structures of natural products, (3) discovering and developing new synthetic methods for the advancement of organic synthesis, (4) designing, synthesizing, and biologically evaluating analogues for biology and medicine, and (5) educating and training students in the art and science of organic synthesis. The inputs and outputs of total synthesis endeavors are summarized in Figure 1.

Figure 1 | Inputs and outputs of total synthesis endeavors.

Contributions of Total Synthesis Endeavors

Replicating the molecules of nature in the laboratory through novel synthetic strategies

Total synthesis of natural products

The serendipitous first synthesis of a natural product, Wöhler’s synthesis of urea in 1828, had a profound impact on science. First, it demonstrated that the synthesis of naturally occurring molecules was possible by chemical means in the laboratory and not as previously believed as “the business of nature and nature’s alone” through the so-called “vital force.” ⁶ Wöhler’s work, together with earlier findings and collaborative work with Justus von Liebig and Jöns Jacob Berzelius, also brought to light the phenomenon of molecular isomerism by demonstrating that silver cyanate (AgOCN) has the same elemental composition as silver fulminate (AgCNO), yet exhibits different properties. ⁷ The eventual significance and evolution of these monumental discoveries and concepts into the fields of organic synthesis and isomerism, respectively, as we know them today, speak volumes for the enormous impact of Wöhler’s experiment. Hermann Kolbe’s synthesis of acetic acid (CH ₃COOH) in 1845 ⁸ represented a quantum and inspirational jump in total synthesis in that it demonstrated its power to create a carbon–carbon bond in the laboratory and extended its reach to a two-carbon–containing molecule, whereas the total synthesis of glucose by Emil Fischer ⁹ in 1890 provided a new measure of molecular complexity that could be reached by chemical synthesis at the time. While the 19th century witnessed the emergence of total synthesis and a steady growth of its reach, the 20th century saw an explosive increase in its power in terms of molecular complexity and diversity. ¹ ^– ⁴ With their challenging and widely varying structures, the myriad natural products discovered and structurally elucidated by isolation chemists around the world challenged and inspired synthetic organic chemists and gave further impetus to the field. Armed with the equally impressive number of new synthetic methods discovered and developed in the meantime, synthetic chemists responded admirably, synthesizing most of the biologically active and structurally challenging molecules unearthed through increasingly sophisticated synthetic strategies that became shorter, more efficient, and in many cases friendlier to the environment. Their endeavors were decisively facilitated by increasingly more powerful instrumentation and physical methods such as X-ray crystallography, NMR spectroscopy, mass spectrometry, and various chromatographic techniques.

Today, total synthesis stands as the flagship of organic synthesis, reflecting and symbolizing its powerful state of the art and providing synthetic avenues to the most complex and precious molecules found in living nature. Indeed, the practice of total synthesis of natural products provides challenges and opportunities that, more often than not, lead to advances in the field itself with regard to molecular complexity and diversity that can be reached, strategy design in terms of elegance, novelty, step-count, and delivery of scarce substances for biological investigations. These days, the selection of natural products as synthetic targets should be made carefully so as to maximize the significance of the resulting discoveries and inventions made during the total synthesis endeavor, often called a “campaign.” Important criteria for this selection include molecular architecture novelty and complexity, biological activity, scarcity, degree of and confidence in the full structural assignment of the natural product, and overall educational and training value for the students involved in the project. Often, and when the structure of the molecule appears unusual or too complex, the synthetic chemist stands in awe of the molecule, being unsure of whether he/she can actually accomplish the task of its seemingly impossible total synthesis. Hesitation sinks in and questions arise such as “What if I do not finish it?” or “How am I going to accomplish it?” That is the moment when courage is required to make the decision to go forward and exploit the opportunities to discover and invent, for it is not only the arrival at the destination of the journey that counts. Most importantly, what also counts are the riches and wisdom gathered along the way. And there are many treasures and knowledge awaiting to be discovered and acquired in the pursuit of every challenging and often seemingly impossible to synthesize natural product, or any designed molecule for that matter. To face such seemingly impossible challenges, creativity and imagination are required, and so are resourcefulness and experimental competence, not to mention a hard work ethic and persistence (see Figure 1).

The collective contributions of synthetic organic chemists to total synthesis over the last few decades are enormous and beyond the space limits of this article. Instead, herein, we provide highlights of some of our endeavors in total synthesis, representing only a fraction of the almost 200 natural products synthesized in our laboratories at the University of Pennsylvania; the University of California, San Diego; The Scripps Research Institute; the Chemical Synthesis Laboratory at ICES, A*STAR, Biopolis, Singapore (2004–2010); and Rice University (see Figure 2 for structures of select natural products synthesized). Collectively, the total synthesis endeavors directed toward the natural products shown in Figure 2 and the myriad others synthesized in our laboratories and other laboratories around the world reflect the advancement of the art and science of total synthesis to new levels of sophistication in terms of efficiency, strategy design, molecular complexity, and structural diversity. The target molecules included in the entire collection of natural products synthesized in our laboratories ¹⁰ also cover a wide range of biological activities, including antitumor, neurotoxic, antibiotic, and enzyme inhibition, all properties relevant to biology and medicine.

Figure 2 | Select natural products synthesized in the Nicolaou laboratories.

Figures 3 – 6 highlight in retrosynthetic format the total syntheses of four of the most sought- after challenging and biologically active natural products in the last decades of the 20th century: calicheamicin ( 7), ¹¹ ^– ¹⁴ paclitaxel (Taxol, 9), ¹⁵ brevetoxin B ( 10), ¹⁶ ^– ¹⁸ and vancomycin ( 17), ¹⁹ ^– ²² respectively. These syntheses charted new grounds in the art and science of total synthesis in terms of complexity and diversity. Most significantly, the complexity of some of these synthesized molecules supports the statement that, “no matter how complex, any natural product discovered in nature can, in principle, be replicated in the laboratory by chemical means given sufficient funding and time.” This notion came close to being experimentally verified with the case of maitotoxin ( 74, blue-ring fragments constructed, red rings remain to be cast, Figure 7), the largest secondary metabolite discovered from nature thus far. With all advanced segments dictated by our designed synthetic strategy toward this giant molecule synthesized, ²³ ^– ³⁰ all that remains is their planned couplings and elaboration of the resulting products to the final target. Unfortunately, the funding for this project was discontinued, bringing the completion of its experimental execution temporarily to a hold. However, the progress made, and the proven feasibility of the remaining steps as demonstrated in model systems and our total syntheses of maitotoxin’s cousin molecules, brevetoxins B ( 10, Figures 2 and 5) ¹⁶ ^, ¹⁷ and A ( 12, Figure 2), ³¹ ^– ³⁵ suggest that the art and science of total synthesis has reached a state that leaves no natural product, regardless of its molecular complexity, out of reach of chemical synthesis, provided sufficient time and resources.

Figure 3 | Total synthesis of calicheamicin ( 7). ¹¹ ^– ¹⁴ Retrosynthetic analysis and key building blocks.

Figure 4 | Total synthesis of paclitaxel (Taxol, 9). ¹⁵ Retrosynthetic analysis and key building blocks.

Figure 5 | Total synthesis of brevetoxin B ( 10). ¹⁶ ^– ¹⁸ Retrosynthetic analysis and key building blocks.

Figure 6 | Total synthesis of vancomycin ( 17). ¹⁹ ^– ²² Retrosynthetic analysis and key building blocks.

Figure 7 | Molecular structure of maitotoxin ( 74), confirmation of its original structure assignment (indicated with gold-colored oval), and remaining rings to be constructed ( P, B′).

Cascade reactions in total synthesis

Useful as they are, synthetic methods alone do not fulfill their potential in making molecules unless they are put together logically within a strategy toward a targeted molecule, more or less complex. Thus, not so unlike a composer putting together musical notes to produce a melody, the synthetic organic chemist has to arrange in the right sequence the appropriate reactions—the strategy—for the projected synthetic route toward the targeted molecule. And just as in music, the resulting total synthesis may, or may not, elicit feelings of awe and admiration from the audience, thereby achieving its desired objectives and receiving acclaim. Among the most important criteria for judging a total synthesis are elegance, efficiency, practicality, cost effectiveness, and environmental friendliness. Cascade reactions, ³⁶ also known as domino reactions and compared with “fireworks” (see Figure 8), are those in which the product of the first reaction becomes the starting material for the second reaction through the newly generated chemical reactivity, and the product of the second reaction becomes the starting material for the third and so on, until the sequence terminates due to chemical inactivity under the reaction conditions, to afford the final product without isolation of any intermediates (Figure 8). Needless to say, the success of such sequences demands insightful and clever synthetic designs through which the practitioner encodes within the first starting material the desired reactivity, and frequently the selectivity for stereochemical and other requirements of the expected cascade intermediates and final product. The design of such strategies requires deep knowledge of the principles and reactions of organic chemistry in general and organic synthesis in particular, as well as exquisite experimental skills for their laboratory execution. The total synthesis of tropinone by Robert Robinson ³⁷ in 1917 and that of (±)-progesterone by William S. Johnson ³⁸ in 1971 are notable examples of cation-mediated cascade reactions prior to our systematic engagement in the field. Below, we briefly highlight a select few of the numerous cascade-based total syntheses we devised in our laboratories, a theme that we embarked on in the early 1980s with the endiandric acid cascade and that we have promoted ever since. ³⁹ ^– ⁴¹

Figure 8 | Cascade reactions and their hallmarks.

Inspired by the novel molecular structures of the endiandric acids ⁴² ^, ⁴³ B ( 2) and C ( 3) and the hypothetical biosynthetic pathway through which they are supposed to be formed in nature, as suggested by David StC. Black, ⁴⁴ we designed and executed the cascade-based strategy for their total syntheses as shown in Figure 9. ³⁹ ^– ⁴¹ Defining acetylenic precursor 75 as the common and ultimate precursor, a Lindlar catalyst–promoted hydrogenation afforded the polyolefinic intermediate substrate 76, which underwent a series of spontaneous pericyclic reactions: an 8π-electrocyclization followed by a 6π-electrocyclization and an intramolecular [4+2]-cycloaddition (Diels–Alder reaction). This strategy was expected to produce, upon methyl ester hydrolysis, a racemic mixture of endiandric acids B ( 2) and C ( 3) through a process that required equilibration of the two initially formed cyclooctatriene intermediate conformers (i.e., 77 and 78) as shown in Figure 9. It was enormously pleasing to experience the successful implementation of this plan that led to synthetic and pure racemic (as found in nature) endiandric acids B ( 2) and C ( 3) upon chromatographic separation and hydrolysis of the two so-formed corresponding methyl ester products. ³⁹ ^– ⁴¹ The thermally induced processes involved, and the requirement of no reagents beyond hydrogen and the catalyst, symbolize the virtues of cascade reactions and their advantages in organic synthesis.

Figure 9 | Total synthesis of endiandric acids B ( 2) and C ( 3) through a cascade sequence. ³⁹ ^– ⁴¹

Another total synthesis notable for its aesthetically pleasing cascade strategy is that of the “alleged” structure of a natural product named rugulin ( 87, Figure 10). ⁴⁵ ^– ⁴⁸ Its central, cage-like structural motif comprised of two identical units of a tricyclic quinone system bridged by four C–C bonds made it an attractive, but challenging target for synthesis. Retrosynthetic disconnection of three of these C–C bonds led to the two monomeric units clinging to each other by a single C–C bond ( 83, Figure 10), whose further retrosynthetic rupture revealed monomeric unit 81 via 82 (Figure 10) as the starting material for the projected cascade sequence of reactions. This cascade was initiated and completed with two of the simplest reagents, MnO ₂ and Et ₃N. Thus, acting as mild oxidizing and basic agents, respectively, these reagents accomplished, at ambient temperature, the desired fusion of two molecules of 81 through the presumed cascade as shown in Figure 10 involving a two-bond oxidative coupling to afford heptacyclic system 82, whose newly formed C–O bond suffered facile rupture to give dimeric compound 83. The latter intermediate underwent two sequential intramolecular 1,4-conjugate additions (Michael reactions), the first spontaneously, the second initiated by Et ₃N, to form product 85 via intermediate 84. Exposure of the latter to additional amounts of MnO ₂ afforded projected, cage-type product 86, from which the targeted alleged natural product ( 87) was generated by simple cleavage of the two MOM protecting groups as shown in Figure 10. ⁴⁷ ^, ⁴⁸

The bis(anthraquinone) antibiotic BE-43472B ( 32, Figure 11) attracted our attention as a potential synthetic target because of its striking structure and potent antibacterial properties. ⁴⁹ ^, ⁵⁰ Upon close inspection of its molecular architecture and a brief retrosynthetic analysis, building blocks bis(quinone) 88 and hydroxy diene 89 (see Figure 11) were defined as possible starting materials for a synthetic strategy toward antibiotic BE-43472B that would proceed through an imagined cascade sequence of reactions. Through careful experimentation, the total synthesis of this novel target molecule was achieved as summarized in Figure 11. Thus, heating 88 and 89 actuated their regio- and stereoselective [4+2]-cycloaddition to afford Diels–Alder product 90a, which readily entered equilibrium with its lactol tautomer ( 90b), whose intramolecular ring-forming nucleophilic substitution under the thermal conditions employed was facilitated by the indicated remote intramolecular hydrogen bonding (quinone O with phenolic H) arrangement, leading to octacycle 91 as shown in Figure 11. Loss of MeOH from transient intermediate 91 then generated stable intermediate 92. The latter was converted to the also stable epoxyketone 93 through a short sequence involving epoxidation, opening of the resulting epoxide, further elaboration, and a second epoxidation. Upon photoirradiation, epoxyketone 93 entered a second cascade sequence of reactions leading to BE-43472B ( 32) through intermediates 94 and 95 as expected (see Figure 11). ⁴⁹ ^, ⁵⁰

Figure 10 | Total synthesis of the alleged structure of rugulin ( 87) through a cascade sequence. ⁴⁷ ^, ⁴⁸

Figure 11 | Total synthesis of antibiotic BE-43472B ( 32) through cascade sequences. ⁴⁹ ^, ⁵⁰

An impressive series of cascade reactions starting with a common precursor ( 96) was developed for the total synthesis of structurally novel and complex natural products (+)-bisorbicillinol ( 98), (+)-bisorbibutenolide ( 101), and trichodimerol ( 106) as highlighted in Figure 12. ⁵¹ ^, ⁵² Thus, upon sequential exposure to first KOH and then aq. HCl, precursor 96 suffered acetate hydrolysis to afford tertiary alcohol 97, two molecules of which reacted in a Diels–Alder reaction, one reacting as the diene and the other as the dienophile, to afford the corresponding dimeric product (not shown). The latter underwent spontaneous tautomerization under the conditions of the reaction leading to (+)-bisorbicillinol ( 98). On the other hand, sequential exposure of the so-formed (+)-bisorbicillinol ( 98), first to KHMDS and then to aq. HCl, furnished (+)-bisorbibutenolide ( 101) via intermediates 99 (deprotonation of the tertiary hydroxy group) and 100 (ring contraction), the latter undergoing spontaneous protonation under the conditions employed. A third, rather spectacular cascade sequence of reactions initiated by sequential exposure of common precursor 96 to basic (CsOH·H ₂O, MeOH) and acidic (NaH ₂PO ₄·H ₂O) conditions at ambient temperature led to the cage-like symmetrical structure of trichodimerol ( 106) ⁵³ via transient intermediates 97/ 102 (tautomers, formed upon acetate hydrolysis), 103 (formed through Michael reaction), 104 (formed upon intramolecular lactol formation), and 105 (formed through Michael reaction) as shown in Figure 12. ⁵¹ ^, ⁵² Featuring a minimal number of simple reagents and manipulations, these remarkable transformations are notable for their rapid and economical build-up of molecular complexity and diversity, underscoring once again the importance of cascade synthetic strategies in the art and science of total synthesis. A similar cascade sequence was employed by Corey and Barnes-Seeman in their total synthesis of trichodimerol. ⁵³

Figure 12 | Total syntheses of (+)-bisorbicillinol ( 98), (+)-bisorbibutenolide ( **101**), and trichodimerol ( **106**) through cascade sequences. ⁵¹ ^, ⁵²

As the final example of the many cascade sequences in total synthesis from our laboratories, the case of thiostrepton ( 20, Figure 13), the flagship and most complex of the thiopeptide class of antibiotics, is highlighted here. ⁵⁴ ^– ⁵⁸ Inspired by a biomimetic hypothesis, ⁵⁹ ^, ⁶⁰ we opted to explore the possibility of generating an azadiene system in situ from an easily accessible precursor (i.e., 107, Figure 13), expecting it to undergo dimerization through a [4+2]-cycloaddition as a means to construct thiostrepton’s unique and challenging dehydropiperidine structural motif carrying suitable appendages to serve as the scaffold onto which to build the rest of the molecule. Figure 13 summarizes the cascade sequence we designed toward this molecule and successfully executed in the laboratory. Thus, readily available thiazolidine 107 was converted to transient azadiene 108 (Ag ₂CO ₃, DBU, py, in the presence of BnNH ₂, −12 °C) which underwent spontaneous [4+2]-cycloaddition/dimerization through an endo transition state (see 109, Figure 13) to afford, initially, Diels–Alder product 110 (plus its diastereoisomer in ca. 1∶1 dr) and thence the desired dehydropiperidine 111 [plus its (5 S,6 R)-diastereoisomer in ca. 1∶1 dr] by sequential aminolysis (with BnNH ₂) and hydrolysis (upon quenching with H ₂O) of the resulting labile 1,1′-diamino intermediate (not shown) and further elaboration. Interestingly, in our initial attempts to coach precursor 107 toward the desired pathway in the absence of BnNH ₂, we encountered a diversion of the cascade at the junction of transient intermediate 110 toward an undesired pathway that led to an unwanted bicyclic imine product [not shown, plus its (5 S,6 R)-diastereoisomer in ca. 1∶1 dr] through initial imine/enamine tautomerization and subsequent aza-Mannich cyclization. ⁵⁶ Delivered in ample quantities through this cascade sequence of reactions through the fine-tuned experimental conditions as mentioned above, key building block 111 served well in our successful and first total synthesis of thiostrepton ( 20), achieved in 2004. ⁵⁴ ^– ⁵⁷

Figure 13 | Total synthesis of thiostrepton ( 20) through cascade sequences. ⁵⁴ ^– ⁵⁸

The examples of cascade reactions highlighted above and others like them emanating from our laboratories, and other groups, ³⁶ proved pivotal in elevating the art and science of total synthesis to higher levels of sophistication, elegance, and performance. They continue to play a paradigmatic and inspirational role for the young generation of practitioners of organic synthesis who are currently pushing its frontiers even further.

Confirming, disproving, predicting, and revising molecular structures of natural products

The science of assigning molecular structures of naturally occurring compounds began with the emergence of the concept of the structure of the molecule as introduced in 1857 by August Kekulé ⁶¹ ^, ⁶² and his contemporaries Joseph Le Bel, ⁶³ Jacobus van ‘t Hoff ⁶⁴ ^, ⁶⁵ and others, ⁶⁶ ^– ⁶⁸ who recognized the tetrahedral nature of carbon and the fundamentals of stereochemistry. ⁶ ^, ⁶⁹ These concepts ignited a surge of activities by chemists whose primary concerns centered around deciphering the connectivity and configuration of the atoms of naturally occurring molecules and other compounds derived from them or employed as intermediates to synthesize them. The pioneers of these endeavors were faced with enormously tedious challenges in the absence of separation and purification techniques and analytical instrumentation as we know them today. All they had at their disposal were the primitive tools of simple glassware, a balance, a Bunsen burner, and a thermometer. They relied on crystallization and the melting points of compounds—whenever that was possible—to determine purity (crystallization until a constant melting point signaled a pure sample) and identity by comparison with melting points of known compounds (identical melting points meant identity). The practitioner would then proceed to determine the molecular formula in terms of type and number of atoms within the molecule through elemental analysis. Once the empirical molecular formula was determined, the next task became the determination of the functional groups within the molecule. Toward this end, they would check Beilstein’s Handbuch der Organischen Chemie, and, later on, Chemical Abstracts to see if the compound with the determined formula was known. If not, they would embark on the arduous journey of its structural determination. The harsh conditions employed in their experiments included ozonolysis of olefinic bonds and oxidative rupture of carbonyl moieties by refluxing with HNO ₃, alkaline KMnO ₄, or CrO ₃ in acetic acid. As mentioned by Reinhard W. Hoffmann in his fascinating book Classical Methods in Structure Elucidation of Natural Products, ⁶⁹ “this approach was compared to an attempt to learn something about a Chinese porcelain figurine in a dark room by knocking it to pieces, collecting them and to examine them later by light.” Obviously, these archaic experimental techniques limited the number of compounds whose structures could be elucidated due to scarcity and sensitivity of many of them under the rather brutal conditions employed at that time. Be that as it may, the structural motifs so determined in the successful cases had to be put together to solve the complete structural puzzle at hand. That required the backbone of the molecule to be determined, which constituted another challenging problem to those courageous early chemists. The pending experiments also required severe conditions such as heating at 400 °C with Zn dust to remove any heteroatoms (a method introduced by Adolf von Baeyer ⁷⁰ in 1866) or/and selenium dehydration. ⁷¹ Once the backbone of the molecule was defined and the functional groups were placed on it in the suspected positions, the chemists were able to propose a structure that was only hypothetical until proven right. At that point, the synthesis of the molecule was required to prove its proposed structure, a task that challenged the synthetic organic chemists of the time. This challenge was, thankfully, taken up by the daring total synthesis practitioners of the 19th century who at the same time recognized the need to develop new methods to construct the required functional groups and make their molecules in predictable and selective ways in terms of regio- and stereochemical configurations. This double challenge—the proof of structure of the molecule and the discovery and development of new synthetic methods and strategies—gave a new impetus to organic synthesis that eventually bestowed the title of the “locomotive” that drove the field of organic synthesis forward, upon the science of total synthesis.

Going back to the early total synthesis practitioners whose primary purpose was to confirm the structure of the natural product, we should emphasize that they had at their disposal only a limited number of reactions with predictable performance and selectivity. In combination with the lack of analytical techniques and instrumentation, this predicament forced them often to confirm each synthetic step by going backward, a tedious but necessary process at the time. Struggling as they were, however, with the relatively primitive synthetic methods, these pioneers were steadily improving their art and enriching it with new and powerful methods, an outcome that allowed them to tackle more and more complex natural products. They advanced from crystallization and melting points to degradation of their target molecules into fragments, whose structural determination was easier than the original challenge, and, by identifying those fragments, they could put together their ultimate puzzle. Certain structural features and stereochemical configurations such as ( Z)- versus ( E)-double bond configurations and cis- versus trans- arrangements were still challenging to determine, unless the substituents were within or on a ring. Such assignments are, of course, routine today, due to the power of NMR spectroscopy and X-ray crystallographic analysis, but one can only imagine the magnitude of the task at the time. It is indeed remarkable how the organic chemists of that era managed to determine the structures of the most notable and abundant natural products of their era.

While the early period of the structural elucidation of natural products (1860–1930) can be characterized as inadequate in terms of analytical techniques and stereoselective synthesis, the field acquired significant momentum after that rather dark period. Thus, from 1930 onward, enabling analytical, spectroscopic, and X-ray crystallographic methods and techniques emerged and developed as powerful tools for structural elucidation, leaving behind the archaic original methods of the previous era, a period that can be called the “Stone Age” of structural elucidation. ⁶⁹ There was also a synchronous “liberation” of organic synthesis from the task of being the only means of proving a hypothetical structure of a natural product. As Albert Eschenmoser is quoted as saying, “Natural product synthesis was freed from the chains to be structure-proof and could develop to a creative field of chemistry in its own right.” ⁶⁹ In fact, total synthesis not only gained its independence to be creative in its own right, free from the “burden” of confirming the structures of naturally occurring molecules, but also benefited enormously from the same methods and techniques that liberated it from the task of being the essential proof provider of the true structures of natural products.

However, this liberation perception is not a universal truth, for even today, and despite the impressively powerful methods and instrumentation at our disposal, total synthesis is often the final proof of structure for a significant number of natural products. Indeed, despite the impressive advances in analytical, spectroscopic, and X-ray crystallographic analyses, it is rather surprising that even in the modern era of structural elucidation (1960–present) total synthesis still plays a role in this important endeavor, as demonstrated in ours and other laboratories. ⁷² These demonstrations come in a variety of ways including confirming a structure (the most common occurrence), fully assigning an incomplete structure (e.g., diastereoisomer, enantiomer, atropisomer), predicting a structure that has not been isolated as yet, disproving a structure, and revising a structure after proving it wrongly assigned, as occurred in our total synthesis endeavors over the last four decades. Figure 14 depicts a select number of natural product structures that were revised, fully assigned, predicted, proven wrong, or confirmed after been challenged in the literature. Thus, our total synthesis endeavors had an impact, one way or another, on the structural assignments of at least 24 natural products of the 184 synthesized thus far in our laboratories. This number represents 13% of the natural products we targeted and synthesized. Given the relatively large numbers involved, this percentile may be representative of the number of instances where total synthesis endeavors impacted structural assignments of natural products around the world during the last few decades. ⁷²

Figure 14 | Select structural revisions and assignments established in the Nicolaou laboratories.

However, this rather astounding observation should not be taken against isolation chemists, who so admirably collect samples from living creatures and isolate and structurally characterize precious molecules for chemistry, biology, and medicine, and thus provide invaluable knowledge and inspiration to other scientists around the world for further investigations in their own fields. The reasons for the errors or deficiencies in the structural elucidations of natural products are often due to their scarcity and the insufficient amounts of material isolated that prevents their full and proper investigation. However, it is ironic that with all our increasing power of analytical techniques and instrumentation, we are still unable to get all natural product structures right, often because of their rareness. It is in those cases that synthesis comes to the rescue of the true structure of the molecule lying there in disguise until discovered through total synthesis, the latter endeavor often turning into a “chemical detective story,” an odyssey of sorts. The numerous exemplary cases from our laboratories highlighted in Figure 14 fall in different types. The first category of structural elucidation by total synthesis in our laboratories are those of absolute configuration determination of natural products whose original structures were reported as racemic mixtures (e.g., 13, 14, 32, 119, Figure 14c, d, m). ⁴⁹ ^, ⁵⁰ ^, ⁷³ ^– ⁸² The second category is that of molecules synthesized before they were found in nature (e.g., 4, 22′, 115, 116, Figure 14a, b, g). ⁴⁰ ^, ⁴¹ ^, ⁸³ ^– ⁸⁷ The third category includes molecules whose originally reported structure was later challenged based on biosynthetic considerations ⁸⁸ [e.g., maitotoxin ( 74), Figure 7, stereogenic centers within gold-colored oval] and later confirmed by us as the correct one. ⁸⁹ The fourth, and most common, category of structural elucidation through total synthesis is that involving revision of one or more stereochemical configurations and/or structural motifs (functional groups) [e.g., 21, 120, and 121 (Figure 14e), ⁹⁰ ^– ⁹⁵ 22 (Figure 14g), ⁸⁵ ^, ⁸⁶ 24 (Figure 14h), ⁹⁶ ^– ⁹⁸ 30 and 31 (Figure 14n), ⁹⁹ ^– ¹⁰³ 33 (Figure 14j), ¹⁰⁴ 37 (Figure 14o), ¹⁰⁵ ^– ¹⁰⁷ 39 (Figure 14p), ¹⁰⁸ ^, ¹⁰⁹ 44 (Figure 14r), ¹¹⁰ 46 (Figure 14q), ¹¹¹ 47 (Figure 14s), ¹¹² 115 and 116 (Figure 14b), ⁸³ ^, ⁸⁴ 119 (Figure 14d), ⁷⁸ 127 (Figure 14i), ¹¹³ ^, ¹¹⁴ 130 and 132 (Figure 14k), ¹¹⁵ 134 (Figure 14l) ¹¹⁶ ].

One of the most interesting examples of a revision of an originally assigned structure of a natural product made by our group was that of azaspiracid-1 ( 21) and its congeners, azaspiracid-2 ( 120), and azaspiracid-3 ( 121). ⁹¹ ^– ⁹⁴ Thus, through a campaign, turned into an odyssey, we discovered 13 incorrectly assigned structural elements in one of the originally assigned structures and revised it from the originally assigned ( 122, see also 123 and 124, Figure 14e) to what we thought to be the true structure. ⁹¹ ^– ⁹⁴ Little did we know that we made an error ourselves in reassigning and depicting one of the stereocenters (C20). However, as it turned out, and as admirably noticed by Forsyth et al. ⁹⁵ in their final say on the subject, we had synthesized the correct structure of the natural azaspiracid acids, but assigned one stereocenter (i.e., C20, Figure 14e) wrongly. ⁹¹ ^– ⁹⁴ Another interesting example is that of the alleged rugulin structure ( 87, Figure 14f). Upon synthesizing this structure as described previously (see “Confirming, disproving, predicting, and revising molecular structures of natural products” section, Figure 10), we realized that it was wrong. Although X-ray crystallographic analysis confirmed the structure of the compound synthesized, we were not able to determine the true structure of the naturally occurring compound, which still remains a mystery. ⁴⁷ ^, ⁴⁸ Yet another interesting case was that of abyssomicin C ( 22, Figure 14g) and atrop-abyssomicin C ( 22′, Figure 14g) in which our total synthesis yielded two diasteroisomers (i.e., 22 and 22′, Figure 14g) ⁸⁵ ^– ⁸⁷ as opposed to the originally assigned structure (i.e., 125, Figure 14g) ¹¹⁷ ^, ¹¹⁸ that did not designate diastereoisomerism. The second diastereoisomer (i.e., 22′, Figure 14g) was later discovered in nature. ¹¹⁹ Endiandric acid D ( 4, Figure 14a) is another example of a natural product we synthesized ³⁹ ^– ⁴¹ before it was discovered in nature. ¹²⁰

Discovering and developing new synthetic methods for the advancement of organic synthesis

Total synthesis endeavors provide wonderful opportunities to discover and invent new synthetic reactions as a means to advance organic synthesis in general. Such discoveries and inventions can occur when the practitioner faces intransigent problems that cannot be solved by known methods and/or when method improvements are desired in terms of elegance, efficiency, cost-effectiveness, practicality, or environmental friendliness. They occur either by design or serendipity, but in both cases they require astuteness and creativity for their discovery and development into powerful and valuable processes. Below, we highlight a number of useful synthetic methods (see Figure 15) discovered and developed in our laboratories during total synthesis endeavors and in our attempts to fill in gaps in methodology as we recognized them while facing the challenges posed by our specific targeted natural products. This practice of blending method development with target-oriented total synthesis programs in our laboratories had its origins and inspiration from the time of the corresponding author’s stay in Professor E. J. Corey’s laboratories at Harvard in the 1970s when we discovered and developed the “double activation method for macrolactonization” ¹²¹ (see Figure 15a) as a prerequisite for the total synthesis of macrolide antibiotics, then considered out of reach of total synthesis due to the lack of such enabling methods.

Figure 15 | Select synthetic methods developed in the Nicolaou laboratories.

Among the most important and broadly employed methods we discovered in our laboratories are a series of useful single electron transfer oxidation reactions based on our newly discovered reactivity of hypervalent iodine reagents such as IBX and related compounds (see Figure 15b–e). ¹²² ^– ¹²⁵ These methods include the introduction of α,β-unsaturation into cyclic ketones (or their hydroxy precursors, which are converted to their ketone counterpart under the reaction conditions) to afford enones (Figure 15b left and Figure 15e), ¹²³ ^, ¹²⁴ benzylic oxidations to give aryl ketones (Figure 15b right), ¹²³ formation of various types of heterocycles and related systems from amides (Figure 15c left), ¹²² N-aryl carbamates (Figure 15c middle) ¹²² and cyclic glycals (Figure 15c right), ¹²² mild cleavage of dithioacetals and dithioketals (Figure 15d), ¹²⁵ and aromatization of functionalized N-heterocycles (Figure 15d upper right). ¹²⁵

Another set of powerful cyclizations developed in our laboratories are the phenylseleno (and phenylsulfeno) lactonizations and etherifications of olefinic acids and alcohols leading to lactones and cyclic ethers, respectively, employing PhSeX (or PhSX; X=Cl, Br). ¹²⁶ ^, ¹²⁷ Furthermore, our newly introduced at the time N-PSP ( N-phenyl-selenophthalimide) and N-PSS ( N-phenylselenosuccinimide) reagents were found useful in inducing phenylselenohydroxylation of olefins or phenylselenolactonizations and etherifications as shown in Figure 15f, g. ¹²⁶ ^– ¹²⁸ A number of these reagents and methods were proven enabling in our quest for prostacyclin analogues and related systems, ¹²⁹ ^, ¹³⁰ as well as in other laboratories for a variety of applications.

Another set of powerful reactions were pioneered in our laboratories as part of our forays toward the ladder-like marine neurotoxins brevetoxins B ¹⁶ ^– ¹⁸ and A ³¹ ^– ³⁵ and other challenging natural products. These methods included stereo- and ring-selective reactions for the synthesis of cyclic ethers ranging in size from common to medium and large. These reactions relied on a variety of ring closures such as intramolecular hydroxy epoxide openings (Figure 15h), ¹³¹ ^– ¹³³ hydroxy dithioketal cyclizations (Figure 15j), ¹³⁴ ^, ¹³⁵ oxidative couplings (Figure 15v), ¹³⁶ bridging of dithionolactones to form bicyclic systems (Figure 15y) ¹³⁷ —a method later employed for the synthesis of dithiatopazine, the first stable 1,2-dithietane (Figure 15y), ¹³⁸ copper-catalyzed intramolecular triazene-based synthesis of diaryl ethers, ¹³⁹ and olefin metathesis-based strategies (Figure 15m). ¹⁴⁰ Others relied on manipulation of the corresponding lactones by converting them to their corresponding thionolactones or ketene acetal phosphates (Figure 15i). The former were subjected to reductive desulfurization, ¹⁴¹ ^, ¹⁴² while the latter entered a range of palladium-catalyzed carbon–carbon bond-forming reactions to afford cyclic ethers. ¹⁴³

In addition to the ether-forming reactions, we also developed a number of macrolactonization methods, the most prominent of which was the intramolecular ketophosphonate–aldehyde coupling to form macrolactones (Figure 15k). ¹⁴⁴ ^, ¹⁴⁵ This method found many applications in organic synthesis, including in our total synthesis of amphotericin B ( 6, Figure 2). ¹⁴⁶ ^– ¹⁴⁹ Its macrocarbocyclic counterpart was also developed, in both solution and solid phase versions (Figure 15l), and applied by us to the synthesis of macrocarbocycles, including the naturally occurring muscone. ¹⁵⁰

As part of our total synthesis of carbohydrate-containing natural products, we had opportunities to discover and develop a number of glycoside-bond forming reactions including the aryl thioglycoside-based glycosylation (Figure 15n), ¹⁵¹ the solid-phase selenium-promoted formation of 2-deoxyglycosides, ¹⁵² ^, ¹⁵³ the generation of glycosyl fluorides and their exploration as glycosyl donors (Figure 15o), ¹⁵⁴ and the stereospecific synthesis of 2-deoxyglycosides from 2-hydroxy phenylthioglycoside donors through 1,2-migration of substituents followed by selective glycosylation and reductive cleavage of the phenylthiol residue (Figure 15o bottom). ¹⁵⁴ Notably, we also developed a solid-phase synthesis of oligosaccharides and applied it to the synthesis of a heptasaccharide phytoalexin elicitor (HPE). ¹⁵⁵ Collectively, these glycosylation reactions advanced organic synthesis and found numerous applications in laboratories around the world, including ours. ¹⁵⁶

In response to challenges we encountered in our thiostrepton total synthesis campaign, we developed a mild and selective ester hydrolysis method using Me ₃SnOH that performed exquisitely well in its intended purpose without epimerization of vulnerable stereocenters (Figure 15p) and other sensitive functionalities such as other esters (e.g., Me vs. iPr). ¹⁵⁷ This method found many applications in our and other laboratories around the world.

Included among the other methods developed in our laboratories are a number of asymmetric reactions, such as the catalytic asymmetric three-component 1,4-cycloaddition/aldol reaction for the synthesis of α,β-substituted cyclic ketones (not shown), ⁹⁹ the catalytic asymmetric intramolecular Friedel–Crafts type α-arylation of aldehydes (Figure 15q), ¹⁵⁸ the first asymmetric dichlorination of allylic and cinnamyl alcohols (Figure 15r), ¹⁵⁹ and the asymmetric alkylation of anthrones (Figure 15s). ¹⁰⁷ The latter reaction was developed as part of our program to synthesize enantioselectively the antibiotic viridicatumtoxin B ( 37, Figure 2), ¹⁰⁷ a task for which no satisfactory method existed at the time.

In response to a deficiency in existing methods for the sulfenylation of 2,5-diketopiperazines in pursuit of the corresponding naturally occurring 2,5-dithiodiketopiperazines, we developed a method to fill the void. Thus, it was discovered that starting with sulfur and NaHMDS, the powerful in situ generated reagent [(TMS) ₂NSSSSN(TMS) ₂] reacted sequentially with the formed enolates of the added substrate to produce, at will, the desired epidithiodiketopiperazine or bis-(methylthio)-diketopiperazine product as shown in Figure 15t. ¹⁶⁰ ^, ¹⁶¹ Another powerful and useful reaction discovered as part of another total synthesis endeavor was our selective synthesis of p- and o-amino- and methoxyphenolic anthraquinones (Figure 15u). ¹⁶² This practical method was applied to the total synthesis of the important enediyne antitumor antibiotic uncialamycin ( 24, Figure 2) and its analogues through a much more efficient and streamlined process ⁹⁸ than our earlier syntheses ⁹⁶ ^, ⁹⁷ employing a conventional approach.

In a collaborative effort directed toward the synthesis of natural and designed molecules of the benzopyran class, we invented a solid-phase, selenium-based method ¹⁶³ ^, ¹⁶⁴ as a rapid and practical means to produce a library of thousands of discrete and pure compounds for biological evaluations. Using radiofrequency encoded combinatorial chemistry with automation, ¹⁶⁵ this work was carried out at a biotechnology company and produced 10000 compounds, from which a number of biologically active lead compounds were discovered (Figure 15w, x). ¹⁶⁶ ^– ¹⁶⁸

Synthesizing molecules for biology and medicine

One of the joys and most important aspects of total synthesis endeavors is the opportunity to apply the developed synthetic strategies and methods to synthesize and biologically evaluate designed analogues of the targeted natural products. ¹⁶⁹ Such opportunities often lead to the discovery of useful biological tools and/or better than the naturally occurring lead compounds for further optimization as drug candidates. The principal author of this article was fortunate to have participated in such a project during his postdoctoral studies in the laboratories of Professor Corey at Harvard University in the 1970s. This experience involved the synthesis and biological evaluation of the designed diazo analogue of prostaglandin H ₂ (top, Figure 16a), which proved to be perfectly stable and more potent than its highly labile parent compound in aggregating platelets and releasing serotonin as revealed from a collaborative study with Bengt Samuelsson’s group at the Karolinska Institute. ¹⁷⁰ This thrilling experience provided lasting inspiration and encouragement to weave such an aspect of total synthesis within our own endeavors. Besides undertaking an endeavor of total synthesis as an exercise of developing novel synthetic strategies and technologies, and thus advancing organic synthesis for its own sake, there are often other compelling reasons for doing so. Included among them is to render rare naturally occurring molecules, and their analogues, readily available for biological investigations, especially when their properties appear promising with regard to their potential in biology and medicine. Even if the natural product is readily available from natural sources, certain desirable analogues may not be easily accessible by conventional manipulations of the natural product itself, in which case a synthetic route based on its total synthesis becomes indispensable, making the latter approach a priority. Designing and synthesizing natural product analogues for biological and pharmacological investigations is, therefore, a powerful means to establish research programs directed toward investigations in biology and drug discovery and development efforts through discovery of biological tools and drug candidates, respectively.

Figure 16 | Select designed synthetic analogues from the Nicolaou laboratories.

From our very early projects in the eicosanoid field we targeted, not only the naturally occurring molecules but also their analogues with the intention of extending the impact of our research into the domains of biology and medicine. Figure 16 includes some of our accomplishments in this area as highlighted with a select number of natural products we targeted, together with one or more examples of analogues we designed, synthesized, and evaluated in collaboration with biologists as part of our total synthesis endeavors. In a related collaborative research program with Charles Serhan of Harvard University directed toward the structural elucidation of lipoxins A and B [for the structures of 7- cis-11- trans-LXA ₄ ( 115) and lipoxin B ₄ ( 116), see Figure 14b], two arachidonic acid secondary metabolites processing important physiological activities, we synthesized a series of isomeric compounds from which we were able to assign the complete structures of the naturally occurring lipoxins. ⁸³ ^, ⁸⁴ ^, ¹⁷¹ ^– ¹⁷³ Most importantly, we were able to provide these molecules in sufficient quantities beyond the meager amounts isolated from natural sources, thereby enabling their biological investigation. These studies set the stage for further developments to occur, thus unraveling their physiological role and prompting drug discovery and development efforts. Another more recent example of a research program in our group with a major biological component is the one triggered by a report claiming the isolation and biological properties of Δ ¹²-prostaglandin J ₃ (Δ ¹²-PGJ ₃, 309, Figure 16e) that included potent activities against leukemic cancer stem cells. ¹⁷⁴ Prompted by these reports, we developed a short and practical synthesis of the natural product and numerous of its designed analogues leading to identification of potent antitumor agents, ¹⁷⁵ ^– ¹⁷⁷ including the monomeric and dimeric macrolactones as shown in Figure 16e. ¹⁷⁸

Numerous other total synthesis endeavors in our laboratories directed toward antitumor natural products were also extended to the design, synthesis, and biological evaluation of analogues of the target molecule as summarized in Figure 16. Included among them are paclitaxel ( 9, Figure 16f, g), ¹⁷⁹ ^, ¹⁸⁰ rapamycin ( 8, Figure 16h), ¹⁸¹ sarcodictyins A and B ( 315 and 316, Figure 16i), ¹⁸² epothilones A and B ( 318 and 11, Figure 16j), ¹⁸³ ^, ¹⁸⁴ and cortistatin A ( 25, Figure