Major Classes of Natural Product Scaffolds and Enzymatic Biosynthetic Machinery

1.1 Introduction

Natural products could be defined broadly as any molecules found in Nature. More traditionally in organic and medicinal chemistry communities natural products are defined as small organic molecules (molecular weight (MW) <1500 daltons) generated from conditional metabolic pathways (but see Vignette 1.1 in this chapter). That is the definition used here. Conditional metabolic pathways are also known as secondary pathways, not present in all organisms and not essential for life. Producer organisms include microbes such as bacteria, algae, fungi and also plants of every variety.

The natural products they generate from conditional pathways presumably confer some form of advantage or protection to the producers. The physiological functions may differ and are often not clear to the chemists who have done the isolation. On the other hand, many of the natural product classes isolated historically have either useful pharmacologic activities in human medicine or the reverse, showing mammalian toxicity through diverse mechanisms.

The adjective "natural" has a strong positive resonance in this era with consumers of food, cosmetics, medicines, nutraceuticals, even clothing and furnishings. In part that may be a reaction to the synthetic and abiotic materials that pervade our environments and in part is probably a connection back to humanity's past and a time when there was a closer dependence and harmony on what the natural world provided for carving out a simpler existence. One (incompletely examined) assumption is that humans have evolved with the plants and microbes that generate the natural materials and small molecules and have coadapted. This has led, over millennia, to a learned avoidance of toxic substances and conversely the utilization of natural extracts for treatment of health problems.

Starting some 200 years ago and continuing into the present, chemists have focused on isolating biologically and pharmacologically active substances, first from plants and then from fungi and bacteria, characterizing them molecularly, and producing useful molecules as pure compounds. At this point there are some 32 000 compounds tabulated from Chinese traditional medicine sources, including the antimalarial drug artemisinin (for which the 2015 Nobel Prize in chemistry was awarded). In parallel, the Dictionary of Natural Products database, which records information on purified natural molecules, contains some 210 000 compounds (Rodrigues, Reker et al. 2016). Natural products have been a continuing source of architectural and synthetic inspiration (Jurjens, Kirschning et al. 2015) to eight generations of chemists, since the first decades of the 19th century. It is estimated that 50% of natural products still have no synthetic counterparts and up to 80% of the natural product ring systems, which generate the constrained molecular architectures, are not mimicked by synthetic molecules (Rodrigues, Reker et al. 2016).

Figure 1.1 shows the structures of eight natural products isolated and characterized by their pharmacologic activities. Ergotamine, rebeccamycin, tubocurarine, and morphine have diverse biologic roles as foreign substances in humans. All four of these molecules have amino acid-derived scaffolds and can be broadly classified in the realm of alkaloid natural products, by virtue of one or more basic nitrogens embedded in a ring system. In structural terms morphine and the lysergic acid tetracyclic moiety of ergotamine are clearly related but the indolecarbazole framework of the antitumor rebeccamycin and the arrow poison tubocurarine bear no obvious overlap.

Ergotamine, in addition to the tetracyclic lysergic acid starter unit, is also built from the three amino acids l-alanine, l-proline, and l-phenylalanine on a nonribosomal peptide synthetase assembly line (Chapter 3). Similarly, the nitrogen atoms in the bicyclic antibiotic penicillin derive from a nonribosomally generated tripeptide aminoadipyl-cysteinyl-d-valine (Walsh and Wencewicz 2016).

The remaining three molecules in Figure 1.1 come from three additional distinct natural product classes. The anticancer microtubule blocking agent taxol (paclitaxel) is of diterpene origin. A late stage hydrocarbon intermediate taxadiene (Chapter 4) is subsequently heavily oxygenated and multiply acylated to yield taxol. Erythromycin is a venerable antibiotic with a 14-membered macrolactone core and a pair of deoxysugars. The substitution pattern on the macrolactone arises from a polyketide synthase assembly line (Chapter 2). The eighth natural product shown is rotenone, a mitochondrial respiratory blocker that is a member of the plant phenylpropanoid class of natural products (Chapter 7).

Taxol and rotenone lack any nitrogen atoms in their scaffolds, reflecting distinct building blocks and assembly logic from the alkaloids and penicillin, respectively. The presence or absence of nitrogens, particularly basic nitrogen atoms in natural product frameworks, affects physical and functional properties of the metabolite classes and is key factor in subclass definitions.

Natural products in general and dozens of particular compounds that have become therapeutic agents or inspired design of structural mimics have come to the attention of human investigators over the past 150–200 years on the basis of their diverse biologic activities. Figure 1.2 summarizes a gamut of pharmacologic activity of just 11 of the hundreds of thousands of known natural products. Among contemporary natural products of therapeutic interest, lovastatin, which lowers cholesterol by targeting the rate-determining enzyme in cholesterol biosynthesis, and the immunosuppressives rapamycin and cyclosporine have probably been the most significant human therapeutic leads. Lovastatin biogenesis is examined in Chapter 2 and cyclosporine and rapamycin in Chapter 3.

Estimates of the natural product inventory, defined as above, are in the range of 300 000 to 600 000 compounds. Three quarters have been isolated from plants, indicating their prodigious commitment to secondary metabolites; the remainder are microbial metabolites. There are no good estimates on the inventory yet to be discovered and whether many new molecular classes will be found. In the future as more plant genomes are sequenced better estimates may become available.

1.2 Primary Metabolites vs. Secondary Metabolites

Primary metabolites are the molecules that populate the pathways essential for life. At one limit they comprise the molecules in both the biosynthesis and degradation of the classes of biopolymers: nucleic acids, proteins, polysaccharides, and lipids. They also populate the pathways for generation and storage of energy, including glycolysis, the citrate cycle, aromatic biosyntheses, amino acid metabolism, the pentose phosphate pathways and others.

Secondary metabolites instead populate pathways that may only occur in some cells or in some organisms (Demain and Fang 2000) in some circumstances, for example when plants respond to predators by synthesis of defensive small molecules (phytoalexins and phytoanticipins) (Schenk, Kazan et al. 2000, War, Paulraj et al. 2012). They may represent specialized molecular scaffolds that are not found in primary metabolism. Often the natural products that sit as the end metabolites of secondary pathways have substantially more complex scaffolds than found in primary metabolites, reflecting C–C bond-forming reactions in their biosynthesis.

The boundaries between primary and secondary metabolic pathways often have a gate keeper enzyme which acts to shuttle some of the flux of a primary metabolite into the secondary pathways. For example, lignan (Chapter 7) is a key structural polymer in woody plants. After cellulose it is the most abundant form of plant biomass. The proteinogenic amino acid phenylalanine provides all the carbon framework for lignan polymers. The gate keeper enzyme, the first one committed to movingl-phenylalanine into phenylpropanoid metabolites, is phenylalanine deaminase. We will note in Chapter 7 that this enzyme has an unusual covalently attached cofactor that allows a low energy mechanistic path for elimination of the elements of NH3 across Ca and Cß to produce cinnamate.

Analogously, acetyl-CoA carboxylase and propionyl-CoA carboxylase, generating malonyl-CoA and 2S-propionyl-CoA, respectively, are on the border between primary metabolism and secondary pathways that lead to polyketide natural products. Malonyl-CoA can go either way in producer organisms, to fatty acids (primary pathway) or to polyketides (conditional pathway). 2S-Methylmalonyl-CoA is not used in fatty acid synthesis but is a key elongation substrate in erythromycin assembly.

Figure 1.3 tabulates a set of primary metabolites that are building blocks for many of the structural classes of natural products discussed in Chapters 2–7. Glucose is the most common sugar in cells and the glucose-1-phosphate derivative is the entry point for commitment of glucose flux to glycosylated natural products: this is the subject of Chapter 11.

The isomeric pair of isopentenyl diphosphates, the ?2- and ?3-isomers, in head to tail alkylative couplings are progenitors to >50 000 isoprenoid natural products. When such molecules are isolated from plants they have been known historically and even today as terpenoid molecules (Pichersky, Noel et al. 2006). The C30 isoprenoid squalene-2,3-oxide is a borderline primary/secondary metabolite (triterpene) that on directed enzymatic cyclizations gives rise to hundreds to thousands of sterol type natural products, as we shall note in Chapter 4.

The two aromatic amino acids l-tryptophan (Trp) and l-phenylalanine (Phe) are important building blocks for the thousands of proteins made in every free-living cell and organism. They are also utilized in nonribosomal peptide assemblies. As shown in Figure 1.3 they are also the building blocks for d-(+)-lysergic acid and the dimeric lignin (+)-pinoresinol, respectively (Chapters 5 and 7).

As noted above, the two carbon acetyl-CoA and its three carbon enzymatic carboxylation product malonyl-CoA are key acyl thioesters in primary metabolism but also in the genesis of the large and various natural product class of polyketides. Shown in Figure 1.3 is the antifungal ionophore monensin which is distinguished from other polyketide subclasses by the presence of furan and pyran cyclic ethers embedded in the molecular backbone.

Figure 1.3 contains two additional molecules in the primary metabolite column: molecular oxygen (O2) and S-adenosylmethionine (SAM). O2 is such a pervasive cosubstrate in the tailoring of all the major natural product classes of Chapters 2–8 that a separate chapter, Chapter 9, is devoted to the chemical logic and enzymatic catalysts that have evolved for its selective reductive activation.

S-Adenosylmethionine, with its trigonal sulfonium cation interspersed between a methionine residue and an adenosyl residue, is a crucial reactant in both primary and secondary metabolic pathways. We will note the iterative use of SAM as methyl donor to a diverse array of cosubstrate oxygen, nitrogen, and carbon nucleophiles of isoprenoid, polyketide, alkaloid, peptide, and phenylpropanoid frameworks. Most of these involve transfer of a [CH3+] equivalent. A significant set of methyl transfers go to substrates at unactivated carbon centers and these are dealt with in Chapter 10 where radical intermediates, including [CH3-] equivalents, are emphasized.

1.3 Polyketide Natural Products

Figure 1.4 shows five structurally distinct subclasses of polyketide natural products: polycyclic aromatics, macrolactones, decalin-containing scaffolds, polyenes, and polyethers containing furans and pyran cyclic ether rings. All of these are built with equivalent logic, as detailed in Chapter 2, that borrows the chemical and protein precepts from fatty acid biosyntheses. In this sense it is a good place to start, at the boundary of logic between a primary and a set of secondary metabolic pathways.

The three aromatic metabolites, oxytetracycline, xanthones, and urdamycin, represent the large subgroup of polycyclic fused aromatic ring polyketides. They are all made as polyketone-containing acyl chains (hence the class name of polyketides) covalently tethered to acyl carrier proteins. Such chains are reactive, a ready source of enolates acting as carbanions and intramolecular aldol condensations, followed by aromatizing dehydrations that lead to the characteristic aromatic polycyclic frameworks.

The antibiotic erythromycin and the antiparasitic drug ivermectin fall in the subclass of polyketide macrolactones, with deoxysugar moieties attached to oxygens of the macrocyclic core. Members of this class typically span 12-atom to 22-atom macrolactone rings that form a platform for display of the various substituents, including the sugar moieties, to biological targets.

Lovastatin is featured because of the presence of the bicyclic decalin ring system. This is a molecular signature in a small number of polyketide metabolites that suggest Diels–Alder [4+2] cyclization chemistry at some stage in the biosynthetic pathway.

Nystatin represents a number of polyketides with polyene moieties, in this case six olefins. The polyene rigidifies the macrocycle and contributes to its ability to insert into the membranes of fungal pathogens. The final structure in Figure 1.4 is the potassium ionophore lasalocid which has insecticidal properties. The distinguishing features of this subclass of polyketides are the cyclic ethers, in this instance a five ring furan and six ring pyran.

We will delve into the common strategy for carbon–carbon bond formation that constitutes the chain elongations that generate the core scaffold constructions in all these polyketide subclasses. The carbon nucleophile arises by thioclaisen decarboxylative condensation of malonyl thioesters, in all polyketide synthases and all fatty acid synthases. The carbonyl group of acyl thioesters are the electrophilic partners in those C–C bond formations. We will also observe how tailoring of the initial ß-ketoacyl-S-carrier proteins from such condensations lead to the array of –OH, CH and CH=CH functional groups in various types of mature polyketide frameworks.

Figure 1.5 indicates that two of the founding members of the most widely used types of polyketide antibiotics, the macrolides, represented by erythromycin, and the aromatics, represented by oxytetracycline, are conditional metabolites from soil bacteria. Saccharopolyspora erythraea gives its name to the macrolide antibiotic while Streptomyces aureofaciens and rimosus make the tetracycline scaffold.

1.4 Peptide Based Natural Products

The peptide bonds that form the backbone of proteins and smaller peptides are chemically stable in physiologic aqueous media. However, they are susceptible to the diverse set of proteases that recycle proteins and peptides back to constituent amino acids. A number of strategies are in play for producer organisms to turn proteins or peptides into long-lived low molecular weight natural products.

These strategies include the generation of cyclic peptides to thwart amino- and carboxypeptidase action. Modification of side chains and/or de novo utilization of nonproteinogenic amino acids can produce side chains that resist proteolytic hydrolysis of peptide bonds. We will note in Chapter 3 the combined heterocyclization of side chains to oxazoles and thiazole rings in cyanobactins, and the macrocyclization to create highly morphed cyanobactin scaffolds that behave as stable small molecules, despite their protein origin (Figure 1.6). The heterocyclizations in particular have converted protease-susceptible amide bonds into heterocyclic backbones that are protease-resistant linkages.

A complementary route to peptide-derived natural products involves molecules built on nonribosomal peptide synthetase assembly lines (Walsh and Wencewicz 2016). The logic is parallel to that of polyketide synthases. Growing peptidyl chains are covalently tethered as thioesters on peptidyl carrier proteins. In this universe the common chain elongation chemical step is C–N rather than C–C bond formation. The attacking nucleophile is the amine group of aminoacyl carrier proteins onto upstream peptidyl thioester carbonyl groups. Release of the full length chains is often via macrocyclization, in full analogy to the polyketide synthase release logic to give compact macrolactams and macrolactones. The third aspect of NRPS assembly lines that build in structural and functional diversity is the use of many dozens of nonproteinogenic amino acids as building blocks for the enzymatic assembly lines. In the nonribosomal natural product kutzneride A all six building blocks are nonproteinogenic amino acids (Figure 1.7) and the scaffold is a cyclic macrolactam. Both features impart protease resistance to this antifungal natural product produced by root-associated bacteria.