Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Curr Biol. Author manuscript; available in PMC 2010 June 9.
Published in final edited form as:
PMCID: PMC2731226

The natural history of antibiotics

Selman Waksman first used the word antibiotic as a noun in 1941 to describe any small molecule made by a microbe that antagonizes the growth of other microbes. From 1945–1955 the development of penicillin, which is produced by a fungus, along with streptomycin, chloramphenicol, and tetracycline, which are produced by soil bacteria, ushered in the antibiotic age (Figure 1). Today, the evolution of antibiotic resistance by important human pathogens has rendered these original antibiotics and most of their successors largely ineffective, and if replacements are not found, the golden age of antibiotics will soon come to an end.

Figure 1
Biosynthetic pathways for the antibiotics penicillin, tetracycline, streptomycin, and thiocillin

Understanding the success and failure of antibiotics requires understanding their natural history — the origins, evolution, and functions of the molecular medley that has played such an important role in human health. Studying their natural history could also result in new strategies to find novel antibiotics and delay resistance to existing ones.

Assembly from readily available parts

Antibiotics do not look like the familiar molecules in beginning biochemistry texts; they usually do not even resemble each other. In spite of these apparent differences, they are assembled from the same types of building block through enzyme catalysed reactions that closely resemble those used in making proteins, fatty acids, and polysaccharides. For example, penicillin is derived from a tripeptide of three amino acids, two of which are proteinogenic (cysteine and valine) and one of which is an intermediate in lysine metabolism (α-aminoadipate) (Figure 1). In conventional polypeptide biosynthesis, tRNAs bring the correct amino acid building block to a mRNA template and peptide bonds are formed to generate an amino-acid chain with the mRNA-encoded sequence. Some peptide precursors to antibiotics are biosynthesized this way, and the ribosomally encoded peptide undergoes enzyme catalysed post-translational modifications to produce the final antibiotic (Figure 1).

Amino acid-derived antibiotics are more frequently produced by modular metabolic pathways in which the templating function of the mRNA is embedded in the order of the modules, and specialized carrier proteins in each module perform the selection function of the tRNAs (Figure 1). The first module in the pathway selects the amino-terminal amino acid (α-aminoadipate in penicillin), and the last module selects the carboxy-terminal amino acid (valine). The last module also releases the tripeptide chain, so that auxiliary enzymes in the pathway can carry out the functional equivalent of posttranslational modifications, which are typically quite extensive.

While tRNAs select from a pool of standard (proteinogenic) amino acids, the nonstandard amino acids found in penicillin and other antibiotics need to be synthesized by specialized enzymes in a just-in-time fashion. The proteins of the penicillin pathway fall into various functional categories: enzymes that make nonstandard building blocks such as α-aminoadipate; enzymes that form the modules that select and stitch these building blocks together; enzyme that modify the peptide into the functional antibiotic; regulatory proteins that ensure the pathway is expressed under appropriate conditions; and resistance proteins that prevent the would-be producer from getting killed. In bacteria and fungi, the genes for all these proteins are usually found on a continuous stretch of DNA. Penicillin biosynthesis mimics protein biosynthesis in important ways, but the macromolecules that carry out the two processes are related only in function.

Tetracycline's biosynthesis is closely related to fatty acid biosynthesis, but in this case (distantly) related enzymes carry out both processes (Figure 1). Tetracycline's core is assembled from a starter unit (malonamyl-CoA) and eight two-carbon fragments. Streptomycin is assembled from sugar monomers by close relatives of the sugar linking enzymes that make the common polysaccharides. Of course, the sugar building blocks of streptomycin are more exotic than those of cellulose or chitin, and, as was the case for α-aminoadipate in penicillin biosynthesis, they require specialized pathways for their production from glucose. The enzymes that make these unusual sugars are also relatives of enzymes used in primary metabolism.

Antibiotic biosynthesis is modular, just like the biosynthetic pathways of more familiar biological molecules like proteins and DNA. The startling array of antibiotic structures arises from a more exotic set of starting materials and more extensive modifications of the polymeric core.

Modular biosynthetic pathways

The long linear pathways that nature uses to assemble antibiotics contrast with the laboratory syntheses chemists devise for similarly complex molecules. Efficient laboratory syntheses tend to avoid long linear reaction sequences, because they involve both the logistical difficulties inherent in having a long supply chain and the likelihood of low overall yields — a ten-step sequence with a 90% yield in every step results in a 35% overall yield. In a microbial cell, the catalytic proficiency of enzymes can push the yield of any step sufficiently close to 100% to tame the arithmetic demon that governs overall yields.

The modularity of the pathways — one module per subunit from the beginning to the end of the molecule — enforces a long linear sequence of reactions, and nature favors modularity to expedite the evolution of molecular diversity. Antibiotics are made by highly evolvable pathways. Consider the penicillin pathway. Each of its modules consists of three protein domains: one to select an amino-acid building block; one to activate it; and one to insert it into the growing chain. Three of these modules synthesize the tripeptide core of penicillin, and the sequence of the peptide chain is determined by the sequence of the modules. Vancomycin, which these days is invariably described as the ‘antibiotic of last resort’, has seven modules. There are many instances where two pairs of antibiotics produced by these modular pathways differ by the insertion, deletion, or replacement of a module.

A very similar analysis can be given for the pathways that produce antibiotics like tetracycline, although they are based on acetate-derived building blocks, not amino acids. After the starter unit, each module introduces a two-carbon building block to the growing linear chain (sometimes with a one-carbon side chain) (Figure 1). The two-carbon fragment can be processed in a variety of ways to provide distinguishing features to the originally identical building blocks. The result is a lipid-like molecule rather than a peptide, but the modularity that facilitates evolutionary molecular diversification persists. Antibiotics like streptomycin are also assembled in a modular fashion; except the building blocks are sugars and the coupling reactions are the same type that assemble glycogen and cellulose.

Because the modular assembly of most antibiotics mimics the modular assembly of biological macromolecules like proteins, all of the same general evolutionary strategies that provide for protein diversification — mutation, duplication, deletion, and rearrangement — are also used to evolve new antibiotics. The complex suite of antibiotics we see today result from rounds of alteration, selection, and amplification of simpler ancestors.

Antibiotic evolution is complicated by lineage and function

In principle, the evolutionary history of an antibiotic or an antibiotic family, say the beta-lactam family that includes penicillin, should be a wonderful model for the evolution of multigenic traits. There is a clear phenotype — a molecule — and the contribution of each of the gene products to forming the final molecule is increasingly understood. However, tracing a path back from two members of an antibiotic family to a common ancestor, not to mention the more difficult task of tracing the path from an early ancestor to today's family members, is complicated by our current ignorance of both the pedigree of the antibiotic producing genes and the ecological role of the antibiotic in the producer's natural community. The evolutionary history of bacterial genes, especially the genes involved in the biosynthesis of antibiotics and other secondary metabolites, is shaped by horizontal gene transfer. Horizontal gene transfer is undoubtedly the reason that the genes for regulation, resistance and biosynthesis are usually clumped together on a continuous stretch of DNA. As a result, a microbe's secondary metabolite repertoire probably depends more on its neighbors than its ancestors.

Some antibiotic gene clusters are cosmopolitan, while others have cameo roles. One analysis estimated that if 10,000 actinomycetes (the family of soil bacteria that has produced most of our antibiotics and other medically useful molecules) were screened, 2,500 would produce antibiotics. Of these, 2,250 would make streptothricin, 125 streptomycin, and 40 tetracycline. Vancomycin is predicted to be made by one in a hundred thousand; erythromycin, by one in a million; and daptomycin, our newest antibiotic, by one in ten million. Because the soil bacteria that produced so many of our antibiotics live in exceptionally complex multispecies environments, tracing both neighbors and ancestors will be a daunting task. Sequenced bacterial genomes are now appearing with increasing frequency, and it is likely that the genomes of antibiotic-producing microbes will be sequenced at an increased pace in the near future. If the past is any guide, they will reveal that these familiar microbes produce many more molecules than have been found using traditional methods, which will open up great opportunities to tease out the production of the cryptic antibiotics. These new genome sequences will also allow us to make some headway in tracing evolutionary histories, or at least suggest plausible models.

Another problem with tracing the evolutionary history of antibiotics is our current ignorance about their roles in the natural environment. We know what antibiotics can do for us, but what do they do for the producing organism? Without understanding the natural roles of antibiotics, we cannot understand the basis for their evolutionary selection. Most scientists assume that microbes produce antibiotic compounds to mediate interactions with other microbes in their neighborhood. The main evidence for this view is the wide distribution of antibiotic resistance genes: many microbes carry the resistance gene for antibiotics that they themselves cannot produce, from which it follows that resistance genes — and by extension the molecules to which they confer resistance — must have a function.

An appealing possibility is that antibiotics are made by microbes to kill competing microbes, but as early as 1961, Selman Waksman pointed out that the ability of a microbe to produce a small molecule with antibiotic properties when cultured under unnatural conditions in the laboratory, does not imply such a function for the molecule in nature. Recently, it has been shown that at concentrations well below those needed to inhibit the growth of other bacteria, antibiotics can modulate the transcriptional profiles of target bacteria. These revelations have caused several scientists to argue that what we call ‘antibiotics’ are actually signaling molecules that happen to kill bacteria when applied at unnaturally high concentrations. In this view, the products of resistance genes silence messages rather than provide protection. In short, we know little about the ecological role of the molecules we call antibiotics.

Advantages of studying less complicated microbial environments

If our ability to unravel the natural history of antibiotics is frustrated by the complexity of their producers' environment, a logical recourse is finding simpler systems in which microbes produce antibiotics. At a minimum, the likelihood of finding useful new molecules would increase by moving our search away from explored environments. For example, investigations of marine environments have provided many microbial-produced novel small molecules. While these molecules are likely to contribute new human therapeutic agents, the ecology of their marine habitats is not understood well enough to trace antibiotic phylogeny and/or function. In contrast, insect–bacteria mutualisms — a symbiotic association in which each of the participants receives a net benefit — appear quite tractable for functional and evolutionary analyses. An especially attractive system is the multilateral symbiosis among fungus-growing ants, the fungus they cultivate for food, and the bacterial symbionts that help protect the ants' fungal crops (Figure 2).

Figure 2
Exploring the natural history of antibiotics by focusing on ancient agriculture in ants

The relationship between fungus-growing ants and their food fungus first originated some 50 million years ago in the Amazon Basin. As the name suggests, these ants cultivate fungus for food in specialized gardens, typically underground. The relationship between ants and their food fungus is an obligate mutualism: the ants cannot survive without their fungal partner, and the fungal partner cannot survive without the ants. When new queens leave their parent colonies, they carry a fragment of the fungus with them to the site of the new colony. Both ant and fungus have prospered: from a single pair of founding species this initial symbiosis has evolved to include more than 230 species of ants and diverse fungal strains. In the leaf-cutter ant genus Atta, the most derived members of the fungus growers, a single colony can harbor millions of workers and persist for more than a decade. Leaf-cutter ants use fresh leaf substrate to cultivate their fungal partner, and their copious foraging activities make them one of the dominant herbivores of the Neotropics. The phylogeny of the ants and their fungal partners is largely known, and the evolutionary history of the food fungus broadly parallels the ant phylogeny — they have undergone diffuse co-evolution for tens of millions of years.

The ants engage in a second mutualism with bacteria that belong to the same order of bacteria (actinomycetes) that produce so many of our clinically used antibiotics (and anticancer agents). In this system, all of the known ant-associated bacteria belong to the genus Pseudonocardia, and the association between the ants and their bacterial symbionts appears to be an ancient one. The strongest evidence for their longstanding association is the elaborate morphological adaptations that the ants have evolved for housing their bacteria (Figure 2). Different ant genera have different types of modification, and the structures housing the bacteria are connected to glands, which are thought to produce nutrients that support the growth of the bacteria. Ants are highly specialized for their bacterial symbionts, and experiments to replace an ant's bacterial symbiont with that from another ant have not yet been successful. These ant-associated bacteria produce antibiotics, which are as yet poorly known, that protect the ants' fungal gardens from microbial pathogens. Experimental studies crossing the presence/absence of the bacteria with the presence/absence of a specialized garden pathogen — a fungus in the genus Escovopsis — have shown that ants with antibiotic-producing bacteria are better able to protect their fungal gardens from disease. These studies are among the best evidence that at least some antibiotics suppress infections in nature.

The ant–fungus–bacteria mutualism is an ancient system whose evolutionary histories can be deduced by traditional molecular phylogenetic studies. Once these histories have been established and the associated antibiotics have been identified, there will be a wealth of data to trace both the evolution of these small molecules and their function. These studies could also reveal how the ant-bacteria system has maintained itself over tens of millions of years without running out of antibiotics to combat the inevitable development of antibiotic resistance by their microbial pathogens. In short, we can learn a lot from bugs – both the six-legged and microbial varieties.

Further reading

  • Clardy J, Walsh C. Lessons from natural molecules. Nature. 2004;432:829–837. [PubMed]
  • Currie CR. A community of ants, fungi, and bacteria: A multilateral approach to studying symbiosis. Annu Rev Microbiol. 2001;55:357–380. [PubMed]
  • Currie CR, Poulsen M, Mendenhall J, Boomsma JJ, Billen J. Coevolved crypts and exocrine glands support mutualistic bacteria in fungus-growing ants. Science. 2006;311:81–83. [PubMed]
  • Currie CR, Scott JA, Summerbell RC, Malloch D. Fungus-growing ants use antibiotic producing bacteria to control garden parasites. Nature. 1999;398:701–704.
  • Donia MS, Ravel J, Schmidt EW. A global assembly line for cyanobactins. Nat Chem Biol. 2008;4:341–343. [PMC free article] [PubMed]
  • Fischbach MA, Walsh CT, Clardy J. The evolution of gene collectives: How natural selection drives chemical innovation. Proc Natl Acad Sci USA. 2008;105:4601–4608. [PubMed]
  • Linares JF, Gustafsson I, Baquero F, Martinez JL. Antibiotics as intermicrobial signaling agents instead of weapons. Proc Natl Acad Sci USA. 2006;103:19484–19489. [PubMed]
  • Schultz TR, Brady SG. Major evolutionary transitions in ant agriculture. Proc Natl Acad Sci USA. 2008;105:5436–5440. [PubMed]
  • Walsh CT. Polyketides and nonribosomal peptide antibiotics: modularity and versatility. Science. 2004;303:1805–1810. [PubMed]
  • Yim G, Wang HH, Davies J. Antibiotics as signaling molecules. Philos Trans R Soc Lond B. 2007;362:1195–1200. [PMC free article] [PubMed]