A considerable number of the recent series of posts have been concerned with molecules that can be referred to as drugs. It seems useful here to take a look at this from a semantic point of view.
Drugs at Different Levels
Most people carry around in their minds more than one specific meaning ascribed to the small word ‘drug’. If you hear “He’s a drug dealer” or, “She’s on drugs”, the references are not likely to be to antibiotics or blood-pressure medication. Conversely, the sentence, “My doctor prescribed a drug for me” is most unlikely to refer to crystal meth. As a result, the great majority of people realize (even if only intuitively) that the illegal band of drugs are but a subset of a much larger group, that includes substances both universally approved and potentially truly life-saving. It would then be reasonable to assume the definition of a ‘drug’ in this larger sense should be fairly straightforward.
Some standard dictionary references are along the lines of: “A substance used in the diagnosis, treatment, or prevention of a disease or as a component of a medication” or, “a chemical substance that affects the processes of the mind or body”. The US Food & Drug Administration (FDA) defines drugs with wording introduced by the Food, Drug, & Cosmetic Act of 1938, as “articles (other than food) intended to affect the structure or any function of the body of man or other animals” [Sec. 201(g)(1)]. These definitions are quite broad, especially the latter. But speak of ‘drugs’ with a pharmacologist, and small molecules are most likely to be the topic of conversation, and not just any small molecules. Indeed, the term ‘drug-like’ is frequently used in the general field of drug discovery to encapsulate (so to speak) the special features which a successful medicinal drug should embody. Obviously, a drug must have definable function(s), which means that it must be directed to a specific molecular target or a limited set of targets (very often, but by no means exclusively, proteins). But a number of additional properties are very important if the drug is to successfully survive in an active form long enough to be useful, and to find its way to the desired target when administered to a patient. For example, a simple set of guidelines for evaluating a candidate compound formulated by Lipinski has been termed the ‘rule of five’, owing to the recurrence of five (or multiples of it) in the definition of the useful range of properties to look for. Getting a drug to where it needs to go is the preoccupation of the burgeoning field of drug delivery, which now intersects in many cases with advances in nanotechnology.
These rules were designed expressly with reference to small molecules, since increasing molecular size is often associated with diminishing returns in terms of delivery, and sometimes physical properties such as solubility. But that is certainly not to say that large molecules cannot be useful pharmacological and therapeutic agents. In the previous post, it was noted that it has only been in very recent times (historically speaking) that large proteins (especially antibodies) could be garnered from the biosphere for useful human applications. In fact, (as also noted), antibodies have become a billion-dollar industry, especially where monoclonal or specifically engineered antibody variants are concerned. And these antibodies are often referred to as drugs, especially in lay usage. Though obviously moving beyond a tight pharmaceutical definition of ‘drug-likeness’, this is perfectly consistent with the above broad definitions, including that from the FDA. But there are some gray areas…..
Drugs and Category Overlap
It is quite clear that some classes of substances or preparations may have members which have dual drug and nondrug characteristics. One such case noted by the FDA is the field of cosmetics, where even more narrow types of products differ in this kind of duality. For example, shampoos may be seen as primarily hair cleansing, or cosmetic, preparations, and indeed many are simply that and no more. But certainly some have additional functions, such as treating fungal (dandruff) or louse infestations. In this case, specific compounds are added for the indicated medicinal specifications. Although it is obviously of practical significance for product safety and efficacy requirements that some preparations should be regulated and licensed as both drugs and cosmetics, here the relevant products are mixtures, and specific molecules are not doing functional double-duty. For example, an anti-dandruff shampoo may have many different components, but the most significant are the detergent (usually sodium lauryl sulfate, for the cosmetic washing function) and the dandruff inhibitor (specific compounds such as zinc pyrithione). In other words, we cannot speak of either of these individual compounds as having an overlapping drug / cosmetic function; it is only as mixtures (along with various other materials) that the product as a whole acquires this status.
Yet there are certainly well-defined cases where specific molecules share categorization as a drug in combination with other properties. A prime case in point is a contender for the title of oldest drug used by humans: ethyl alcohol, or ethanol. Its psychoactive and other physical effects clearly indicate its drug status, but ethanol can also be metabolized to yield specific calorific value. As such, it is then a food as well as a drug. This is straightforward, but some other areas of ‘foods’ are less so. One definition of a ‘food’ might focus on the ability of a substance to be digestible, or act as source of energy, but clearly this is not sufficient for a healthy diet. There are numerous nutritional ‘cofactors’ which are essential for human health, included among which are a number of inorganic elements (principally metals, but also some other trace elements), and a group of vitamins.
Where do vitamins stand with respect to drug classification schemes? As small molecules which act as organic enzymatic cofactors for catalysis (coenzymes), the defined vitamins are an essential human dietary requirement, owing to our inability to synthesize them. But since they are not directly digestible themselves, they are classified by the FDA as ‘dietary supplements’, which are encompassed within the broader area of foods, and not drugs. Vitamins, then, fall into the this category and therefore would escape labeling as drug materials, unless they were chemically altered from the natural forms. Most dictionary definitions of ‘food’ also include vitamins. Even so, in other quarters vitamins have been clearly depicted as drugs. One basis for doing so is that vitamins can clearly cure diseases – but since the relevant diseases are deficiencies of the vitamins themselves, this would seem to be a special case. As we have seen with alcohol, assignment of a specific compound as a food does not mean it cannot also be a drug. But clearly there is a difference here: vitamins are essential for life, while alcohol (whatever some people might say) is not.
The figure below depicts two separate classifications where the vitamins are considered either as drugs (A) or not (B):
Two depictions of drug categorization and its overlapping areas. These are not to scale in terms of the relative sizes of the respective groups, and are schematic only. ‘Food / nutrients’: This refers to subtances which directly provide energy, structural building blocks, or essential assistance with normal metabolic functioning. ‘Dietary cofactors’ in general include both vitamins and inorganic substances (such as essential metals). ‘Proteins / macromolecules’: Not all macromolecular therapeutic agents are proteins, as for example nucleic acid aptamers. A, Vitamins considered as drugs. Some cosmetic preparations contain vitamins, so vitamins are shown to intersect with the ‘drug-overlap’ region of all cosmetics. B, Vitamins excluded from classification as drugs.
Does this cover everything? Well, there is an additional broad grouping of substances termed ‘nutraceuticals’, a hybrid term from ‘nutrient’ and ‘pharmaceuticals’. A nutraceutical in principle can be any food source component with biological properties outside of direct nutrition, but many of the best-known examples are antioxidants. Included among these are phytoestrogen compounds, considered in an earlier post. Resveratrol in particular (see the relevant Figure from this same post ) has generated enormous interest for its observed anti-ageing effects.
Where do these compounds reside in the above figure? Although they are by definition associated with some kinds of foodstuffs, they are neither directly digested (as for proteins, and digestible carbohydrates and fats), nor required for essential metabolism (as for vitamins). Therefore, it is logical that they be considered a subset of the large drug category, outside of the macromolecular subregion, as shown above. These compounds can be identified, purified, synthesized, and administered independently of their original sources. In this respect, they are no different from any other small molecule natural products derived from the biosphere.
Drugs as Foreign?
Can drugs be thought of as molecules which are ‘foreign’ to the body to which they are administered? (In other words, compounds which are not synthesized by the human or animal which receives them). In a strict sense, this would include vitamins which are dietary essentials through the lack of synthetic machinery for their production by a host animal or person. But there are problems with this proposal, and vitamins themselves are a case in point. For example, although Vitamin C is essential for human health (scurvy resulting in its absence), rats, mice and numerous other species have no problem making their own. Is Vitamin C then a drug for humans (capable of curing scurvy) but not for rats?
And numerous human proteins can be administered under circumstances where they can be considered drugs. Antibodies are an interesting case in point. Originally, monoclonal antibodies were of murine origin owing to the technological requirements of their production. The xenogeneic (foreign) nature of these proteins resulted in the induction of immune responses against the monoclonal antibodies themselves, when they were given to patients. In more recent times, fully human monoclonal antibodies have been developed, in order to circumvent this very significant problem. Yet an antibody of this type is still not literally and totally ‘self’, since its specific combining site is generated by recombinational and mutational mechanisms such that its exact sequence is not directly encoded in the human germline.
But non-variable molecules both large and small also come into this picture. Think of human growth hormone, of value for treating some forms of dwarfism – and sometimes abused for the purposes of body-building. Numerous other proteins and small molecule hormones can also be cited – so the notion of ‘foreign-ness’ for drugs in general becomes untenable.
Drugs, Poisons, and Doses
Drugs have been termed ‘poisons that save lives’, which carries the implicit message of the importance of dosage. But stating that ‘all drugs are poisons’ may be technically correct at a broad enough level, yet not particularly useful, given the vast differences in dosage ranges for efficacy vs. safety seen with different therapeutic compounds. Here a balance or ‘window’ must be found between the two poles of beneficial drug activity and unacceptable toxicity. The old saying ‘the treatment was successful, but the patient died’ provides an ironic testament to this inherent dilemma of drug pharmacology.
As an example of the great range of drug therapeutic windows which can exist, consider the treatment of syphilis. The pioneer of chemotherapy, Paul Ehrlich, found an arsenical compound (Salvarsan) which became an effective treatment for syphilis through its activity against the bacterial causative agent Treponema pallidum. But its toxicity at therapeutic doses was a major problem encountered in a high percentage of patients, so it was clearly not ideal. When penicillin became available in the 1940s, it was not only highly effective but also associated with very low toxic side effects. Indeed, the rising problem of bacterial resistance was initially countered by simply increasing the dosage of penicillin (or its many derivatives) without problems – but of course this soon becomes ineffectual as resistance increases. (Penicillins can actually induce serious problems through allergies in a minority of people, but this is quite distinct from direct toxicity).
There is also a piece of folk-wisdom along the lines of ‘too much of anything can hurt you’, which is certainly true for some natural nutritional requirements as well as drugs. In a general sense, too much food is clearly bad through the development of obesity, but the ‘dosage’ effects of nutrients can be observed in a much more specific manner. We can look within the set of vitamins once more for useful comparisons, which also demonstrate similar variation in the ‘safety’ windows of dosage as seen within artificial drugs. Vitamin C has exceedingly low (if any) toxicity, for example, and some people have routinely taken very high doses of it for long periods as part of ‘megavitamin’ therapy. On the other hand, the fat-soluble Vitamins A and D are unquestionably highly toxic when taken in excess of recommended daily requirements.
The dosage effect can also be related to the above observation that drugs need not be alien to the biochemistry and physiology of the patient (or animal) undergoing treatment. A pathology caused by a deficiency in a specific molecule can be corrected through artificial intervention. Conversely, certain pathological states may benefit from the provision of ‘unnatural’ administration of normal bodily proteins, such that the circulating amounts of the factor of interest are maintained for therapeutic purposes at higher levels than would normally be the case.
After this short foray into some issues surrounding the meaning of drugs, I’ll conclude with references to ‘nutraceuticals’ once more, by means of a biopoly(verse) note.
By analyses really quite shrewd
On mixtures both complex and crude
Smart chemists have shown
(And now it is known)
Natural drugs exist in some food.
References & Details
(In order of citation, giving some key references where appropriate, but not an exhaustive coverage of the literature).
‘….. (FDA) defines drugs….’ For FDA definitions of both drugs and cosmetics, see the relevant page of the FDA site.
‘….set of guidelines for evaluating a candidate compound formulated by Lipinski…’ For a discussion of the basis of the Rule-of-five and moving beyond it, see Zhang & Wilkinson 2007.
‘…..field of drug delivery, which now overlaps with advances in nanotechnology.’ For a recent review of this topic in the cancer field, see Chidambaram et al. 2011.
‘….sodium lauryl sulfate….’ Also known as sodium dodecyl sulfate, this detergent also has wide application in laboratories as well as cosmetics.
‘…..a contender for the title of the oldest drug used by humans….’ Often alcohol is definitively cited as the oldest drug. I call it ‘a contender’ here since (as noted in an earlier post), the use of botanical medicines is also very old, and can even be linked with primate behavior (see a previous post on zoopharmacognosy). On the other hand, the use of alcohol (or abuse, depending on one’s views) is possible simply from natural cases of fruit fermentation, also seen with animals. So the origins of alcohol sampling by humans need not require any technology, and is undoubtedly of great antiquity.
‘……ethanol can also be metabolized……’ For an example of the differing influences of food vs. drug effects in an animal system, see Dole et al. 1985.
‘…..classified by the FDA as ‘dietary supplements’….’ For the FDA definitions of dietary supplements, see the relevant page of the FDA site.
‘…..vitamins can clearly cure diseases…..’ This was noted by Tulp et al. 2006, and that vitamins were thus ‘drugs by any definition’. Taken literally, this is clearly incorrect (one can simply exclude vitamins from a drug definition as dietary supplements, as for the FDA).
‘…..nucleic acid aptamers….’ (Figure footnotes). See a previous post for a brief consideration of RNA aptamers. From large libraries of variants, RNA molecules can be selected to bind desired ligands, and this can be used therapeutically. The first therapeutic aptamer (‘pegaptanib’) was directed against a specific form of vascular endothelial growth factor, for the treatment of ocular vascular diseases. See Ng & Adamis 2006.
‘……an additional broad grouping of substances termed ‘nutraceuticals’……’ For example, see Tulp et al. 2006.
‘……Resveratrol in particular …… has generated enormous interest for its observed anti-aging effects……’ See Pezzuto 2011.
‘……Vitamin C is essential for human health (scurvy resulting in its absence), rats, mice and numerous other species have no problem making their own.’ See Martí et al. 2009 for more information. The production of Vitamin C from glucose requires the enzyme L-gulonolactone oxidase, which humans, primates, and guinea pigs lack. Lachapelle & Drouin 2011 look at when this occurred in evolutionary time.
‘…..its exact sequence is not directly encoded in the human germline.’ Antibodies are composed of constant and variable regions, where the variation of the latter accounts for the vast range of different antibody binding specificities which can be induced by immunization. Particular variable region sequences allowing antigen recognition are specific to that immunoglobulin molecule, and are termed an ‘idiotype’. See Searching for Molecular Solutions Chapter 7 for a more detailed discussion of this.
‘ Numerous other proteins and small molecule hormones can also be noted…..’ Proteins such as interferons were noted in the previous post. Small molecules include adrenalin, thyroid hormones, and natural steroids. In all such cases, though, there is the potential (realized in many cases) for rendering such molecules ‘non-natural’ by various forms of artificial tinkering to improve their performances as drugs.
‘….pioneer of chemotherapy, Paul Ehrlich…’ See Thorburn 1983 for some biographical and other relevant information.
‘……certain pathological states may benefit from the provision of ‘unnatural’ administration of normal bodily proteins…..’ Again, see the reference to the example of interferons in the previous post.
‘ Vitamin C ….. ‘megavitamin’ therapy.’ The Nobel Prize-winning chemist Linus Pauling was a notable proponent for the efficacy of large Vitamin C (ascorbate) doses for conditions ranging from viral infections to cancer. (For example, see Pauling & Moertel 1986). However, no experimental evidence has validated these claims.
‘…..fat-soluble Vitamins A and D are unquestionably highly toxic….’ It is notable that the livers of certain polar animals (including bears and seals) are very rich in Vitamin A, and the eating of such livers by polar explorers has resulted in Vitamin A poisoning (hypervitaminosis A). See Rodahl & Moore 1943.
Next Post: Two Weeks from now.
It was noted in the previous post that most natural molecules of use to humans (‘chrestomolecules’ ) have traditionally been of low molecular weight, with the exception of simple biopolymers. Here it is time to have a brief look at the scope of large molecules and even larger biological systems which can be sampled from the biosphere and usefully applied to human activities.
Big Uses for Big Molecules….
Large biomolecules such as proteins can obviously act as nutrients, but this special category of ‘usefulness’ is not the focus of interest here. Indeed, in this context it should be emphasized that the benefits of consuming a protein as a foodstuff for any organism (humans included) result from the acquisition of its amino acid building blocks. ‘Digestion’ is intimately associated with breaking down consumed nutrients into components which an organism can employ for its own biosynthetic requirements. The original three-dimensional structure and function of the consumed protein is clearly irrelevant for such a nutritive purpose. But when we are considering the utility of large biomolecules for other human needs, the protein shape and function is crucial.
Biotechnological developments in the past century or so have extended the field of molecular utility to include a diversity of proteins (noted in part of in the ‘chrestomolecule’ Table of a previous post). The great majority of these can be grouped into the functional categories of natural biocatalysts (enzymes) and natural specific binding molecules (antibodies), but some additional cases can be noted:
Let’s take a quick look at some specific examples among these large protein sets:
Looking for Enzymes in All the Right Places
Many enzymes from a variety of biological sources are used in industrial processes, and several examples are of particular interest from a molecular biological point of view. There is at this time considerable interest in biological resources obtainable from ‘extremophile’ micro-organisms, or literally ‘lovers of extreme environments’. Certain bacteria thrive under amazing extremes of temperature, salinity, pH or pressure, and bacterial biosystems mediating metabolic processes and growth must in consequence be selected for tolerance of each type of extreme conditions. Only in quite recent times has it become recognized that the earth harbors an entire thermoresistant prokaryotic ecosystem which lives by chemical energy independent of solar radiation. These organisms may be found in geothermal hot springs (as in Iceland or Yellowstone National Park), at deep underground sites or at thermal vents in the deep oceans.
Thermophilic (heat-stable) enzymes in general have many applications, but probably the most famous thermophile application in molecular biology is the use of the DNA polymerase from the hot-spring bacterium Thermus aquaticus, or Taq polymerase. (DNA polymerases are of fundamental importance for the copying and replicating of an organism’s DNA during cellular growth). The hardy Taq polymerase has made the polymerase chain reaction (PCR) a practical proposition, which has had an enormous impact in all areas of this and related sciences and gained its originator (Kary Mullis) a Nobel Prize. PCR involves replication and amplification of a DNA duplex in the laboratory through successive cycles of strand denaturation, association on each strand of specific primer oligonucleotides (short DNA segments), and primer extension (strand replication) with the DNA polymerase. Since strand denaturation is achieved by heating to 95˚ C or greater, a conventional polymerase present in the reaction is ‘killed’ and must be (very tediously) re-added for each cycle. A thermostable DNA polymerase, however, can be maintained throughout the course of the reaction. Although many other such heat-resistant polymerases are compatible with PCR, Taq polymerase is still the most widely used. At the other extreme from thermophilic organisms are psychrophiles, which thrive at very low polar temperatures, and are also potential sources of novel enzymes for novel applications.
In the ‘chrestomolecule’ Table of a previous post, antibodies (immunoglobulins) were featured. These important effector molecules of the mammalian adaptive immune system can be regarded as natural binding molecules produced ‘to order’ in response to specific stimuli (usually threats from invading external organisms), by diversification and selection mechanisms. Their applications are many, but can be divided into in vivo tasks which relect the original biological role for antibodies, or in vitro roles. Even naturally-formed antibodies within an individual are obviously useful if they mediate recovery from microbial infections, and the production of antibodies within a host has long been artificially directed by means of vaccines. Relatively recently compared to the first use of vaccination (attributable to Edward Jenner and cowpox immunization against smallpox), passive immunity has emerged as a therapeutic tool. This can be effected by transfer of whole serum or serum components rich in immunoglobulins from immune animals or humans to non-immune recipients, for the purposes of their protection from sudden exposure to a toxic or infectious environmental agent. With passive immunization even with crude serum fractions, antibodies qualify as examples of large natural ‘chrestomolecules’. But in modern times, passive immunization goes far beyond this.
This essentially began with the discovery of monoclonal antibodies in 1975, and has led to the vast field of antibody engineering. The manipulation of antibodies has wide therapeutic applications, and products based on antibodies have become a billion-dollar earner for the biotechnology industry. But engineered (and conventional) antibodies have many uses in vitro as well, as noted above. These include diagnostics and sensors, and for a variety of purification purposes in research and industry. In its range, this topic really requires further attention in a later post. But before moving on, it’s worth noting that the fields of antibodies and enzymes are not mutually exclusive. For a couple of decades, it has been known that antibodies can be raised such that they catalyze specific chemical reactions. This is possible through clever immunization strategies, where host animals are immunized with analogs of chemical transition-state intermediates of enzymatic catalysis. Even so, the best artificially-generated catalytic antibodies fall short of the efficiencies observed with natural protein enzymes.
And Other Proteins….
Although antibodies remain the star players, the adaptive immune system can contribute other useful large molecules. The key adaptive receptor of the cellular arm of the vertebrate immune system is the T cell receptor, signaling through which activates various different classes of effector T cells. Engineered T cell receptors in a number of forms have application in some cutting-edge technologies for human therapies. Some other proteins which modulate the functions of immune and other cells also have therapeutic and commercial significance. Included among these are interferons, known initially for their antiviral properties, but also for their immunomodulatory activities.
Beyond the immune system, various other proteins are mostly important through their contributions to knowledge, which have enabled the subsequent design of other useful structures. For example, the zinc-finger protein motif mediating DNA-binding has been intensively studied, allowing the derivation of a ‘code’ for binding and the generation of artificial binding motifs towards virtually any desired DNA sequence of 18 base pairs. When teamed with a nuclease domain, such tailored ‘zinc finger nucleases’ have tremendous potential in a number of clinical contexts. In different areas, numerous proteins have contributed their structural frameworks themselves for useful applications. This principle has been used for development of protein frameworks which can replace immunoglobulins for the specific recognition of desired binding targets.
….and even Bigger Higher-Level Useful Systems
By analogy with the introduced term ‘chrestomolecule’, an economically useful biological system (involving numerous mutually interactive molecules) could be termed a ‘chrestosystem’. But if we introduce biosystems into this picture, there are many levels of complexity that might apply. Does a whole organism thus qualify? Is even a domestic mammal a high-level ‘chrestosystem’? (After all, such animals have had many uses over the millennia in addition to being raised as food sources). And downward from there, certain fungi, yeast, and bacteria have long had uses in the modification of foodstuffs, long before they were even recognized as living organisms.
Perhaps cows (and so on) might indeed be viewed as simply highly complex organized systems in the service of humanity, but for the present purposes it will be most constructive to restrict our focus onto viruses. Viruses are usually considered as life-forms, and often have sophisticated regulatory controls, but are nevertheless many orders of magnitude less complex than a large multicellular organism. In their simplest forms, they can be regarded as small nucleic acid genomes packaged within a protein coat, and thus have a relatively small set of macromolecules which collectively allow them to infect their target cells and reproduce. So it is worth considering how some viral systems have useful properties for artificial exploitation. These will be grouped into bacterial viruses (bacteriophages, often referred to simply as ‘phage’) and a select group of animal viruses:
The ‘other viral’ categories are relevant examples, and not intended to be exhaustively comprehensive. (Many other cases exist).
Phage For This Age
Viral parasites of bacteria and other prokaryotic organisms, phage in general are extremely diverse in structure and genetic organization. Phage must attach to their specific prokaryotic host and ensure that their genomes (single or double-stranded DNA or RNA) enter the host cell. Usually the result of this is viral hijacking of the host and its conversion into a phage factory, ultimately leading to host cell destruction (lysis) and liberation of a large number of new phage particles. In broad outline, or course, this is a universal propagation strategy used by viral parasites across all kingdoms of life. Sometimes phage can be carried by their hosts indefinitely by a process termed lysogeny, where the phage integrates into the host chromosome or is carried as an independent replicating genome (plasmid replicon). Study of these and other features of phage biology has yielded a wealth of productive information. Owing to the relative simplicity of phage genomes, it is in fact no exaggeration to state that the rise of molecular biology itself derives largely from phage studies.
Phage have been major contributors to numerous biotechnological advances, and providers of information of general molecular biological significance. These viruses of prokaryotes can also be used as specific information sources potentially leading to novel therapeutics. In the case of pathogenic bacteria, systematic study of their phage has been shown to be a useful approach towards identifying phage-expressed proteins or peptides directed against bacterial host cell components. Identification of both the phage products and their host targets allows screening systems to be established for finding low molecular weight chemical inhibitors of the host cell targets themselves. Owing to the specificity of phage interactions with bacterial hosts, drug-modified phage have also been considered as vehicles for drug delivery to desired target cells. Modified phage have also been tested as vectors for transfection of mammalian cells with protein or nucleic acid ‘payloads’.
The use of bacteriophage in the never-ending war against bacterial pathogens does not stop with phage protein or nucleic acid components. Perhaps inspired by the dictum, ‘the enemy of my enemy is my friend’, phage themselves have long been considered as therapies for a variety of bacterial diseases. The idea is simple enough: kill infecting bacteria with their own parasites. Since phage are specific to their bacterial hosts and cannot replicate in eukaryotes, safety also is not likely to be a problem. Yet the history of ‘phage therapy’ has seen conflicting claims of efficacy, although many successful treatments have been reported. In recent times there has been renewed interest in the potential of phage therapy, accompanied by the rise in antibiotic-resistant bacteria and a slow-down in the pipeline of new replacement drugs. The great diversity of phage has been suggested to be superior to the diversity of natural antibiotics themselves. The unique feature of phage used in this manner is that they are self-replicating, indeed the first therapeutic agents in this category. Their capacity for self-replication in turn is directly linked with the population of their target hosts, which gives the pharmacokinetics of phage therapy some unusual aspects. It has been suggested that inadequate understanding of these novel pharmacokinetics is a major source of the mixed results previously seen with trials of phage therapy, and due attention to this detail may well shift phage therapy into a mainstream alternative to antibiotics. This is especially likely in cases of bacterial infections which are resistant to the current antibiotic armamentarium. Apart from direct therapeutic applications, phage have also been considered for environmental use in the reduction of bacterial loads in human-derived wastewater.
Another interesting feature of the exploitation of phage as whole viruses is that they are not a one-molecule ‘magic bullet’, and are not categorizable as ‘molecular solutions’ in a the sense of involving a single molecular entity. (Certainly a phage could not be deemed a ‘chrestomolecule’, although its subunit molecules could if they were of value in isolation). Phage often use elegant genetic systems for their regulation, and also show highly efficient and compact biological organization. (For example, the phage ΦX174 provided the first natural demonstration of overlapping genes). But despite such layers of underlying organizational complexity, phage have simplicity at the level of the types of their molecular constituents, all consisting of their informational nucleic acids (genomes) protected by a protein coat (capsid), combined with an entry system into their specific hosts. It then follows that the use of phage (whether natural or engineered) as therapeutic agents is an early example of exploitation of a (replicating) supramolecular system.
A huge abundance of as-yet uncharacterized phage are still to be found in the natural environment, and if past experience is anything to go by, such phage will provide rich fields of novel biodiversity. In an extension of the earlier overview (in the previous post) of the bounty of marine bioproducts, marine bacteriophages have been cited as the largest reserve of untapped genomic information. In support of this contention is the observation that such phage have a very high proportion of unassignable open reading frames (predicted proteins which cannot be recognized as belonging to current known protein families). Current high-throughput sequencing technologies not only rapidly reveal new phage genes but allow high-altitude phage genomic comparisons, which are greatly facilitating studies of phage origins and evolutionary processes.
Other Viruses, or: What’s HIV Ever Done For Us?
Many viruses afflicting humans are infamous for the terrible toll they take on human lives, and for the resulting human suffering. Yet certain features of animal viruses can be harnessed towards desirable ends, especially their proficiency for targeting and invading specific host cells. As weapons against cancer, specific viruses have been modified and used to promote beneficial immune responses against cancers or infectious agents. Alphaviruses (a viral family which includes human encephalitis viruses) have been used as vectors to promote directed anti-tumor immune responses. Alternatively, viruses can be designed to directly kill (lyse) tumor targets, with viral examples including herpesviruses and poxviruses. (Obviously, in all such cases, it is of great importance that the introduced genetic modifications inactivate the ability of the virus to damage normal cells).
Even the human immunodeficiency virus (HIV), the causative agent of AIDS, has had something to offer. As a member of the lentivirus family, HIV has the useful property of infecting non-dividing cells, which many other viruses cannot accomplish. Vectors have been derived from ‘gutted’ versions of the HIV genome which are unable to complete the usual replication cycle, but retain the ability to insert a copy of the viral RNA as a DNA sequence in the host genome. These features render them an attractive means for transferring desirable gene sequences to recipient cells of interest. For example, such lentiviral vectors have been used to transfer specific T cell receptors to T cells from human cancer patients. In such cases, the novel engineered T cell receptors are designed to recognize tumor-specific antigens, and allow the transduced T cells to then recognize and kill tumor targets. The lentiviral vector-mediated gene transfer process is also useful for gene therapy applications.
It is interesting that a pathogen as serious as HIV, the destroyer of tens of millions of human lives since its full-scale emergence in the early 1980s, can ‘give something back’ by way of contributing a useful system for human application. Not that this would compensate its victims, of course, but the very fact that anything positive can be gleaned from the HIV experience says something about the human ability to recognize natural opportunities when they arise. There are many other examples of ‘positives’ accruing from dangerous viral and bacterial agents of disease, but these have only been made possible through technology. In many forms, modern biotechnology provides both the means to perceive useful biological opportunities (whether from pathogens or otherwise), and also pathways towards their beneficial application.
Homage to Nature, and then Moving Beyond It
In the molecular domain, Nature’s bounty is translatable into human benefit at several levels, as we have seen. There is firstly of course the direct discovery and application of low molecular weight bioproducts (detailed in the previous post), which has been done for thousands of years. The data provided by specific chrestobiomolecules can be used as a stepping-stone to even more effective agents, through successive rounds of human design intervention.
So, in the long run, the most important thing that large natural biomolecules and biosystems provide to us is information – information that acts as a spring-board towards the development of our own complex molecules and biosystems which are tailor-made towards specific applications. When this process is taken to its logical conclusion, we will reach a stage when we supersede the need to use such incredibly useful biological resources.
The use of higher-level systems in the future will thus inevitably become less a slavish derivation or copying of a natural precedent, and more and more an artificial improvement for specific design ends. This trend then increasingly overlaps with the field of synthetic biology, considered in several previous posts (30th May, 7th June, 20th June ).
Finally, a little comment on big molecules:
About natural products, one should recall:
They go far beyond molecules that are small
And think, if you please
That included in these
Are some considered the most useful of all.
References & Details
(In order of citation, giving some key references where appropriate, but not an exhaustive coverage of the literature).
‘ Certain bacteria thrive under amazing extremes of temperature, salinity, pH or pressure…..’ See Van den Burg 2003.
‘…..earth harbors an entire thermoresistant prokaryotic ecosystem…….’ See Gold 1992.
‘ Thermophilic (heat-stable) enzymes in general have many applications….’ See Haki & Rakshit 2003.
‘….the polymerase chain reaction…’ For more details, see the accompanying ftp site for Searching for Molecular Solutions, for Chapter 4, Section No. 6 (‘PCR’).
‘….natural binding molecules produced ‘to order’ …….by diversification and selection mechanisms…..’ In brief, somatic diversification within immune cells of an vertebrate with an adaptive immune system results from combinatorial genomic rearrangements and somatic hypermutation, coupled with selection and amplification processes for cells uniquely expressing receptors recognizing a target antigen. This is also described (and referenced) in Chapter 3 of Searching for Molecular Solutions.
‘…..passive immunity has emerged as a therapeutic tool…..’ Among a vast number of examples can be found ‘antivenoms’ produced in horses or goats against various natural toxins from snakes, spiders, jellyfish, and other toxic organisms.
‘…..discovery of monoclonal antibodies…..’ See Köhler & Milstein 1975.
‘…….products based on antibodies have become a billion-dollar earner…..’ See Reichert et al. 2005.
‘……Engineered T cell receptors in a number of forms have application….’ More details (and references) and provided in the Cited Notes for Searching for Molecular Solutions in the accompanying ftp site, for Chapter 7, Section 18, T Cell Receptor Technology and Applications.
‘….interferons, known initially for their antiviral properties, but also for their immunomodulatory activities….’ Modified isoforms of interferon-alpha have been widely used (in conjunction with antiviral drugs) for treatment of Hepatitis C (See Tsubota et al. 2011); interferon-beta has been used as a therapy for the autoimmune disease multiple sclerosis (See Rudick & Goelz 2011).
‘…..development of protein frameworks……’ This is considered in Chapter 7 of Searching for Molecular Solutions.
‘……the rise of molecular biology itself derives largely from phage studies…..’ See Cairns, J., Stent, G. & Watson, J. Phage and the Origins of Molecular Biology. (Cold Spring Harbor Laboratory Press, 40th Anniversary Edition, 2007).
‘…….screening systems to be established for finding low molecular weight chemical inhibitors of the host cell targets themselves…..’ See Liu et al. 2004.
‘….drug-modified phage have also been considered as vehicles for drug delivery….’ See Yacoby et al. 2007.
‘……the history of ‘phage therapy’ has seen conflicting claims of efficacy…..’ See Sulakvelidze 2005.
‘……The great diversity of phage has been suggested to be superior to the diversity of natural antibiotics themselves…..’ See Schnaitman 2002.
‘…..the pharmacokinetics of phage therapy [has] some unusual aspects…….’; ‘……inadequate understanding of these novel pharmacokinetics is a major source of the mixed results….’ See Payne & Jansen 2003; Levin & Bull 2004.
‘…….phage have also been considered for environmental use…….’ See Withey et al. 2005.
‘…….the phage ΦX174 provided the first natural demonstration of overlapping genes…..’ When genes ‘overlap’, portions of their coding sequence occupy the same DNA tract, but in different reading frames. For the ΦX174 precedent, see Barrell et al. 1976.
‘…….marine bacteriophages have been cited as the largest reserve of untapped genomic information…….’ See Paul & Sullivan 2005.
‘……Alphaviruses …… have been used as vectors to promote directed anti-tumor immune responses.…..’ See Avogadri et al. 2010; ‘…..viruses can be designed to directly kill (lyse) tumor targets (including herpesviruses…..’ See Todo 2008; ‘……and poxviruses…..’ See Kirn & Thorn 2009.
‘…..many other examples of ‘positives’ accruing from dangerous viral and bacterial agents of disease…..’ This general issue was discussed in the Searching for Molecular Solutions associated ftp site, in the Cited Notes for Chapter 3, Section No. 4 (DNA-binding protein design).
Note: If problems are encountered in directly accessing the above ftp site, go here, and then click on molecular-solutions.zip (8.11 Mb).
This post and the succeeding one are the culmination of a recent theme dealing with ‘Natural Molecular Space’ . This series has looked at the biology of natural products, and also their use by both animals and humans from the earliest of times. Here the same activity in the modern world is considered.
Natural ‘Chrestomolecules’ Now vs. Then
An earlier post discussed the significance of ‘chrestomolecules’, or useful molecules of economic significance (in its broadest sense). This survey did not discriminate on the basis of any molecular properties, and a range of different categories of such molecules was listed. Yet ‘natural molecular space’ as considered in several previous posts (19 July, 26 July, 9 Aug, 16 Aug) has been preoccupied with small molecules. How does the “old” (traditional) vs. “new” (modern) use of all molecules from the biosphere stack up if viewed through a size-based lens?
Notes for diagram: This compares traditional and modern use of all bioproducts. The traditional group is encompassed as a subset of the wider modern group, with Nutrients as a special category. Sources based exclusively on small molecules are shown in blue; those based on either large or small molecules are shown in red, and those based on large biomacromolecules (mostly proteins and nucleic acids) shown in orange. ‘Repetitive biopolymers’ refer to molecules of biological origin whose structures are usually highly repetitive polymers based on a limited number of subunits, such as certain polysaccharides. ‘Therapeutics’ is placed in both Traditional and Modern groups, owing to the use of many small natural biomolecules within each, but therapeutic antibodies only within modern times.
The above diagram uses the categories found in the chrestomolecule Table of a previous post, and sorts them into two broad levels based on their histories of usage. The main message is the preponderance of small molecules in traditional applications. With the exception of certain biopolymers of relatively regular (and often linear) structures, the traditional group does not feature complex proteins and other macromolecules which are well-represented in the modern set. Why should this be the case? Clearly, knowledge and technology (only available from relatively recent times) make a big difference, in allowing the exploitation of the properties of large and complex molecules. Also many small molecules have an inherent major advantage over potential protein therapeutics in terms of their oral delivery potential. Traditional societies could not provide intravenous delivery by hypodermic syringes, and protein or peptide-based substances usually fare poorly in making transit through the acidic digestive barrier of the stomach.
So, this size-wise breakdown will be used as a divide for the coverage the modern use of the biosphere: This post will feature the modern harvesting of natural small molecules, and the next will concentrate on the modern use of natural large molecules, and indeed entire useful molecular systems.
What other general differences might then exist between the (very) old and the (relatively) new, in this context? As a broad principle, one could note that modern science, technology, communications and mobility provide the ability to initiate systematic screens of a wide variety, with increasing refinements as to exactly how they are performed. One consequence of this is the ongoing exploitation of marine environments for natural products, which were almost untouched during ancient times. This theme is looked at in a little more detail below. Another area unique to modern times might be summed up by the phrase ‘getting more from what’s on offer’, or using natural drug precedents to the best advantage through technological developments.
Harvesting of Small-Molecule Natural Molecular Biodiversity in Modern Times
At this point we have considered natural molecular space from a number of viewpoints, including its functions, evolution, classification, and empirical exploitation by traditional tribal medical lore. This leads us to a central issue directly relevant to the theme of this post: How important do natural products as a whole remain for human needs, to what extent have they been replaced by other technologies, and what are future trends in this area of biotechnology? Or using a previously-introduced terminology, what fraction of chrestomolecules now and in the foreseeable future will derive from natural biomolecules?
Beyond two decades ago, a majority of pharmaceutical drugs derived from natural product sources, but by the 1990s the fraction of drugs directly or indirectly originating as natural biomolecules stood at about 50%, owing to advances in synthetic chemistry and combinatorial screening. Nevertheless, the fraction of naturally-derived pharmaceuticals varies considerably if specific therapeutic applications are considered. Over the time-frame 1981-2006, over 70% of cancer drugs have been cited as non-synthetic in origin, and naturally-derived drug percentages increase in all therapeutic categories if one includes semisynthetic analogs based on the original core structure of the biomolecule. In this area, the use of natural product molecular scaffolds for the future design of antibiotics has been promoted. Natural products remain highly regarded for their diversity and as source of novel structural motifs. Generalizations of the properties of natural product molecules have shown significant average differences to synthetic molecular libraries, including steric complexities, atomic contents, and ring structures. As considered in a previous post, the evolutionary origins of natural products may provide a positive bias in favor of their utility, at least as a source of novel molecular scaffolds. Even among plant sources alone there is still a huge range of material remaining to be investigated by systematic ‘bioprospecting’.
In modern times, conscious attempts have been made to harness traditional ethnic knowledge of therapeutic natural products, with the aim of accelerating drug discovery. As noted previously, these kinds of studies have been termed ‘ethnobotany’ (given the preponderance of plant products involved) or more generally ‘ethnopharmacology’, and the advent of a number of important drugs have been attributed to the transfer of ethnopharmacologic knowledge. (The example of quinine, and the complexities associated with Europeans’ awareness of it, was considered in the previous post). The continuing value of ethnopharmacology has been vigorously promoted by certain researchers. Throwing cold water on this enthusiasm to some extent, over a twenty year period the US National Cancer Institute (in the course of systematically screening very large numbers of plant extracts) did not find useful anti-cancer drugs specifically from ethnobotanical information, with the possible exception of the anti-cancer drug taxol.
Commercial interest in ethnopharmacology in recent times has led to the formation of companies dedicated to mining such potentially valuable knowledge as the basis for a drug discovery platform. One much-publicized example is the now defunct Shaman Pharmaceuticals, but numerous companies have had at least a passing interesting in drug acquisition by such means. A highly contentious issue has been the rights of indigenous peoples to compensation if their information led to a successful profit-making venture, and accusations of ‘biopiracy’ have been made towards many Western drug discovery activities in environments used by native people. These kinds of political, legal and ethical issues have clouded or retarded a number of relatively recent bioprospecting ventures. Despite the importance of these considerations, a more fundamental problem is the continuing destruction of rainforests and other natural habitats, which threaten to result in irreversible losses in biodiversity and rich sources of novel biomolecules. Associated with this, and the forces of cultural homogenization, loss of tribal languages and lore are also lamentable outcomes not only in their own right but as a potential source of ethnopharmacological information. This is an absolute loss if the ethnic group is pre-literate, but if old written records exist they may possibly be tapped for such irreplaceable knowledge.
A Natural Frontier in the Sea
In contrast to terrestrial environments, the seas and oceans have not yielded a large number of notable traditional medicines, or an associated rich ethnopharmacologic folklore. This is undoubtedly due to the relative inaccessibility of most marine environments without comparatively recent technological back-up, and marine bioproducts are thus greatly under-represented in traditional medicinal tool-kits. Even peoples with close access to the sea, such as the Samoans, appear to have derived most of their traditional ethnopharmacological lore from land plants. The environmental ‘diversity factor’ D noted in the previous post for traditional drug discovery is in turn reduced in practical terms by the inability to recognize and ‘fish’ the oceans for useful molecules. Thus, the marine environment has been poorly exploited as a source of drug discovery until recent times, despite it bearing a plethora of a potentially useful and highly diverse organisms with equally diverse biosynthetic capabilities. Or possibly even more; it has been claimed that at this juncture in history, searching marine natural molecular space is much more likely to yield novel biodiversity than land ecosystems.
Sponges alone have proven to be rich in a variety of bioproducts with promising applications. Useful metabolites, possibly with antimicrobial properties, may be obtained from seaweeds or marine cyanobacteria. Conotoxins (from cone shell molluscs) from approximately 500-700 Conus species are a highly diverse family of peptides with neurotoxic activities. These find many important applications in neurological research and possibly in a number of therapeutic contexts, at least as prototype molecules pointing towards pathways for future drug development. Sea hares, marine molluscs which have found productive application in research on memory mechanisms, are yet another source of useful products, including antibacterial proteins. Tunicates (sessile marine invertebrates) have yielded chemically diverse cytostatic and cytotoxic drugs with potential applicability in clinical oncology. Given the continuing unmet need for effective anti-cancer treatments, and the very large international market which successful cancer therapies can fill, it is not surprising that commercial interest has been stimulated towards marine natural products with potential for this kind of activity. Numerous clinical trials for the anti-cancer efficacy of certain marine bioproducts have been conducted and are continuing.
In a number of cases of apparent production of useful compounds by marine invertebrates, the true source may be commensal micro-organisms carried by the invertebrate organism. Combining this observation with the relatively poor knowledge base concerning marine microbes, and their ancient and robust variety, assessment of marine microbial populations accordingly deserves high priority for the analysis of oceanic biodiversity. A high-profile expedition launched with this end in mind has been the ‘Global Ocean Sampling’ voyage, under the aegis of J. Craig Venter, his eponymous Institute, and other participants.
More Bang for Your Natural Product Buck
At the present time conventional means for identifying and optimizing natural products has been supplemented by a number of different approaches, with some major examples shown in the Table below. Let’s now examine these in a little more detail.
Recent and ongoing advances in identifying, screening, processing, and developing natural products.
The first six categories of this ‘improvements’ Table deal with better ways for finding natural products of interest in the first place; categories 7 and 8 are more concerned with modifications and betterment of candidate molecules in hand.
Natural products are initially encountered as complex mixtures, and thus efforts have been made towards streamlining sample preparation and purification as much as practicable (No. 1 of the above Table) in advance of screening. Included within this category are important advances in the determination of the molecular structures of natural products from extremely small sample sizes. The capability of evaluating candidate samples for specific properties in very high volumes and with great rapidity (high-throughput screening) is also an important issue in modern natural product evaluation.
With the rise in bioinformatics and computer modeling, ‘virtual screening’ (No. 2 of the above Table; equivalent to computer-aided evaluation of possible candidate drugs by modeling their interactions with target receptors) has become an important adjunct in bioproduct testing, as well as general drug identification. Bioinformatics is also applied in computational searching for new members of specific gene families (such as novel biosynthetic genes), which may act as drug targets, and an important pre-requisite for this is the availability of complete genome sequences for an increasing number of key organisms which produce secondary metabolites.
Empirical screening of any bioproducts is fundamentally dependent on a specific assay and its read-out, whether it is high-throughput or not. If one is seeking a compound which can usefully modify a particular cellular system, a good understanding of the underlying biology of the system is likely to identify the specific molecule(s) which should be targeted. This in turn is a clear advantage for screening itself. Refined understanding of fundamental cellular processes involved with carcinogenesis, for example, will in general lead to assay improvement (No. 3 of the ‘improvements’ Table above). Ultimately, with a single-molecule target and structural information, a rational design strategy may become possible. Prior to such a point, a natural product (or any other) molecular library will be best put to use if it is well-focused on the appropriate target. While less efficient, a more complex assay system (such as whole cells) may on the other hand provide additional information about other side-effects of the tested compounds. The utility of therapeutic bioproducts is critically dependent on their abilities to approach the ideal of ‘magic bullets’ in complex biosystems, without deleterious side-effects.
The next item (No. 4) in the above Table refers to metagenomics, which has been briefly discussed in a previous post. In this context, the relevance of metagenomics centers on the large fraction of environmental bacterial species which are non-cultivatable in the laboratory, but which could potentially yield useful drugs. Amplification and assembly of environmental DNA samples into discrete genomes (in essence, the ambit of metagenomics) has the potential to accelerate the analysis and manipulation of important and novel metabolic pathways, through which new small molecules are synthesized. Accordingly, rapid high-throughput genomic sequencing technologies (No. 5) available only in past few years feed into this process. Further manipulation of whole pathways, ultimately with entire synthetic genomes (the domain of synthetic biology, item No. 6 of the ‘improvements’ Table; and discussed generally in a previous post) in turn will provide great control over both the production of new active secondary metabolites, and their specific chemical modifications as desired. These ambitions require an integrated ‘systems-level’ understanding of the pathways involved in the production of all small molecules by an organism of interest, or its metabolome. High-level metabolomic input thus has great potential for the engineering of organisms for either increased yields of specific secondary metabolites, or the production of novel ones.
Since early times in antibiotic research, natural products have contributed core molecular designs (‘scaffolds’) which have been modified artificially, either by complete or partial synthesis (No. 7 of the above Table). This effort towards bioproduct improvement is still a productive venture. The final (eighth) innovation of the ‘improvements’ Table above concerns attempts to improve yields of metabolites by chemical ‘elicitors’, and also efforts towards boosting metabolite chemical diversity by modifying culture conditions.
These modern technological developments, combined with the well-noted evolutionary advantages of ‘consulting’ the natural compendium of biological small molecules, suggests that there is still much practical value to be gained from them. The question of the ongoing future of the exploration of natural molecular space will be picked up again in the next post, particularly insofar as it may be eventually superseded (at least in part) by wholly artificial alternatives.
And finally, a general comment on the importance of technology in the modern science of natural bioproduct discovery and development, in a biopolyverse-like manner:
Bioprospectors don’t need to carry picks
Since they have a range of techniques in the mix
Using various means
And comprehensive screens
They can rely on a complete bag of tricks.
References & Details
(In order of citation, giving some key references where appropriate, but far from an exhaustive coverage of the literature).
‘…..small molecules have an inherent major advantage over potential protein therapeutics in terms of their oral delivery potential….’ It may be noted that this general issue raises the whole field of drug delivery, the science of devising the means for ensuring that useful drugs reach their intended in vivo targets in an efficient manner while retaining their functional properties. This is a huge modern field, of the most fundamental significance to pharmaceutical companies.
‘…..the fraction of drugs ….. originating as natural biomolecules stood at about 50%…’ See Li & Vederas 2009.
‘….over 70% of cancer drugs have been cited as non-synthetic in origin….’ See Newman & Cragg 2007.
‘…..the properties of natural product molecules have shown significant average differences to synthetic molecular libraries….’ See Koehn & Carter 2005.
‘…..Even among plant sources alone there is still a huge range of material…..’ See Phillipson 2003.
‘..…the advent of a number of important drugs have been attributed to the transfer of ethnopharmacologic knowledge….’ See Cox 1990.
‘……the US National Cancer Institute ….. has not found useful anti-cancer drugs specifically from ethnobotanical information…..’ See Cragg et al. 1994; ‘….with the possible exception of the anti-cancer drug taxol…’ See Cragg 1998.
‘…..accusations of ‘biopiracy’……..’ See Shiva, S. Biopiracy: the plunder of nature and knowledge (South End Press, Cambridge MA, 1997); ‘……..have clouded or retarded a number of relatively recent bioprospecting ventures……’ See Rosenthal 2002.
‘…..peoples with close access to the sea….appear to have derived most of their traditional ethnopharmacological lore from land plants…..’ See Cox 1993.
‘……marine natural molecular space is much more likely to yield novel biodiversity than land ecosystems….’ See O’Hanlon 2006.
‘ Sponges alone have proven to be rich in a variety of bioproducts with promising applications….’ Sample references for modern harvesting of marine bioproducts: From sponges, see Sipkema et al. 2005; seaweeds, see Kubanek et al. 2003; marine cyanobacteria, see Burns et al. 2005; conotoxins, see Alonso et al. 2003; sea hares, see Barsby 2006; tunicates, see O’Hanlon 2006, Simmons et al. 2005.
‘…..the ‘Global Ocean Sampling’ voyage….’ See Rusch et al. 2007.
‘…..advances in the determination of the molecular structures of natural products from extremely small sample sizes….’ See Murata et al. 2006.
‘……(high-throughput screening) is also an important issue…..’ See Koehn & Carter 2005.
‘…….‘virtual screening’ …..has become an important adjunct in bioproduct testing……’ See Rollinger et al. 2008.
‘……an integrated ‘systems-level’ understanding of the pathways involved in the ….. metabolome….’ See Weckwerth 2010. A good example of the economic significance and challenges faced with metabolic pathway manipulation can be found with efforts to engineer the synthesis of the anti-malarial drug artemisinin in microbes for large-scale production. Of Chinese origin, this compound has been effective against the deadly Plasmodium falciparum malaria species, but supplies of its natural plant source (Artemisia annua) are often limiting. To engineer ‘heterologous expression’ of the drug in microbial cells, an entire pathway of enzymes must be provided within the foreign host cells. To date, successes with producing artemisinin precursors in yeast and E. coli cells have been reported. See Arsenault et al. 2008 for an overview, and Tsuruta et al. 2009 for details on an E. coli expression system.
‘……to improve yields of metabolites by chemical ‘elicitors’…’ See Poulev et al. 2003.
‘…..efforts towards boosting metabolite chemical diversity by modifying culture conditions….’ See Bode et al. 2002.
‘…..suggests that there is still much practical value to be gained from them [small natural bioproducts]’ See Li & Vederas 2009.
Next Post: Two Weeks from now.
In the previous post, the current theme of ‘Natural Molecular Space’ was considered from the point of view of its sampling by animals for the purposes of self-medication, or zoopharmacognosy. Now, it is time to move beyond this, to the human use of the huge and highly diverse resource of natural biomolecules. Unlike other mammals, humans can use their intelligence to greatly expedite the profitable sampling of Natural Molecular Space, and in one form or another, this has been practiced from the earliest times across all cultures. The manner and nature of this ‘harvesting’ is the topic of the present post.
Early human pharmacognosy, and continuing
It is clear that since sampling and application of natural bioproducts do not require any awareness of the physical nature of drugs themselves, the process of useful molecular discovery can (in effect) take place long before the definition of a molecule. ‘Discovery’ here is thus defined in a loose sense as the identification of a natural source which contains a useful molecular species. Only within the last century can we apply the more rigorous definition which includes the purification and full characterization of the active constituent(s) in molecular terms. Obviously, early humans had no accurate picture of the nature of material reality, and even more obviously this applies as well to animals involved in zoopharmacognosy (as discussed in the previous post). Indeed, the very concept of a molecule as a precise group of atoms (held together in a very specific manner by the rules of chemical bonding) is of extremely recent vintage in historical terms.
It is unlikely that any human cultures have never used accessible environmental materials for health-related reasons, and many such natural sources contained active constituents which proved enduringly worthwhile. Examples of these are not hard to come by, particularly from local plant sources. Some well-known cases are the anti-malarial drug quinine, derived from the bark of South American trees of Cinchona species, anti-pyretic / analgesic salicylates from willow bark (related to aspirin), cardiotonic digitalis (digoxin) from foxglove, narcotics (morphine and derivatives) from the opium poppy, and anti-psychotics (such as reserpine) from Rauwolfia species. Apart from human disease management, other traditional uses for environmental pharmaceuticals include preparations for control of insect pests and animal diseases. This extensive fund of traditional knowledge about the usefulness of local biotas has been termed (in the most general terms) ethnobiology, or (when focusing on the active compounds found within traditional medicines), ethnopharmacology. In a majority of cases (albeit not exclusively), the sources of such medicines are plants, and the fund of long-standing plant-derived cultural medicinal lore is accordingly termed ethnobotany.
Obviously the biodiversity of the environment within reach of a specific culture will have a bearing on the range of potential tribal medicines, especially since most such treatments derive from plant sources. At extremes one can compare the scope of botanical sources available to polar Inuit peoples with rain forest dwellers; indeed jungle regions such as the Amazon have been particularly rich sources of biopharmaceuticals. So there is obviously no question that humans across the board have been skilled at finding medically useful materials from their environments. But how systematic has this been, and how comprehensive?
When it comes to finding edible and nutritious foodstuffs, it has been noted that human cultures which have long inhabited specific geographic regions have identified essentially all plants which can become dietary items. Apart from much practical field experience and observations in support of this proposition, if a tribal group have lived in a relatively small region for thousands of years, it would seem quite reasonable to accept ipso facto that their knowledge of the food value of local plants would be highly advanced, and close to completion, if not quite at a literal 100% level. (One factor complicating the sampling of foods is where certain plant materials possess initial toxicity, which can be removed by specific processing steps. Clearly, it will take longer to gain such knowledge than by simply tasting plant products for their agreeability).
But the acquisition of plants with medicinal value, as opposed to worthiness as foods, is not the same thing. Could we expect that a tribal people in a resource-rich environmental (such as a tropical rain forest) would have identified all potentially useful medical plant (or other biological) sources after thousands of years in a similar location? If ‘all potentially useful’ is taken literally to mean beneficial to all humanity, then the answer is clearly in the negative. This is obviously the case, given the simple observation that human medical afflictions are far from geographically homogeneous. While modern transport systems can now spread infectious diseases rapidly around the globe, infections and parasites were historically often relatively local in the extent of their reach. Although this is patently obvious, even potential treatments for genetic conditions would not be sought if such problems did not exist within a relatively small tribal population. (Genetic diseases occurring at a low frequency would only tend to be noted on average within large population groups). It then goes without saying that no tribal people could search for a treatment for any pathological condition outside their frame of reference, any more than even the most sophisticated modern screening and drug design methods can be used against a novel disease which has not yet ‘emerged’. So, natural landscapes may harbor drugs whose usefulness is not yet definable even in the present day.
But one can put this high-level issue of generality to one side, and restrict tribal medicine to matters which would be of direct concern to them. Yet here, too, there are problems. From first principles, it could not be automatically assumed that a molecular solution to a given medical problem will necessarily exist anywhere on Earth. And by the same token, a useful molecule within Natural Molecular Space might indeed exist, but not necessarily within the accessible environment of any given tribal group. (If it ain’t there, you can’t find it). This is a kind of counterpoint to the above point noting that a local biota might harbor drugs for as yet undefined purposes – it might also lack useful drugs for specific applications, which are naturally available elsewhere. Consider an example with respect to the ancient enemy malaria: Long endemic in Africa, one of the best natural solutions (at least until the acquisition of parasite resistance) was native to South America in the form of the Cinchona plant (more of which below). Thus, early African peoples obviously had no means for arming themselves against malaria with this particular weapon.
The other problem with identification of useful biological medications is the complexity of doing so in a systematic manner, when faced with knowledge limitations, a vast array of potential environmental sources, and the range and variation within human medical problems. Despite their familiarity with their biological surroundings gained over very long time periods, many tribal peoples, especially in tropical environments, certainly suffered from a high burden of infectious and parasitic disease before access to modern medicine. It can be accordingly inferred that evidently either no solutions to such pathologies existed in their environments, or potential solutions existed, but they had not yet found them (or they had not discovered how to use them in a productive way).
Sifting the Environment for Medicinals
What other factors might be involved in the growth of tribal medical lore? Is it possible to come up with some kind of formula as a rough guideline for the likelihood that a particular culture would develop, purely by trial-and error sampling from its available environment, a useful treatment for a specific medical condition? This question can be rephrased in the following manner: An early tribal people have lived in a forest area for many generations. During this time, some of them have become ill from a previously unknown disease. Their shamans try treating their sick patients with a whole range of available plant materials. What are the chances that they might find something genuinely useful? What factors influence whether there is no hope at all, or whether there is a real prospect of success, perhaps given an element of good luck? For the present purposes, ‘genuinely useful’ means a bioproduct which directly or indirectly alters the pathogenic state in a favorable manner. And such beneficial effects need to be objectively measurable.
Since time immemorial, human patients have been given real comfort from healers of any description who could deliver a placebo effect through their ministrations. Yet although shamanic medicine has a radically different view of reality to modern chemical and pharmaceutical science, there is no question that traditional healing practices have found many therapeutically useful biological materials. Indeed, this very fact has led to the development of ethnobotany as a science in its own right.
While it is not possible to produce a quantitative formula for the likelihood of finding useful local environmental bioproducts by traditional medicines, we can think about the relevant factors involved, and for convenience represent them as R, D, B, C, and t; explained as follows:
R denotes the ‘druggability’ of the specific biological need in question, in terms of the probability of ‘solutions’ available within Natural Molecular Space. In brief, this refers to how well-suited are target molecules (relevant to a specific disease process) to interaction with drug-like compounds found among the gamut of natural biomolecules. This is influenced by the complexity of the illness or disease symptom itself, and what kind of molecular targets it offers for outside intervention. A complex multicellular parasite will contain a large number of proteins vital to its own functioning (but foreign to the host) which are potential sites of action for therapeutic drugs, while a viral infection which hijacks the machinery of the host cell presents fewer targeting options. If the disease stems from some internal physiological malfunctioning, the susceptibility of relevant target molecules to drug action is uncertain in the absence of specific information or empirical evidence. For example, perturbation of protein-protein interaction surfaces has historically been regarded as a difficult proposition, in contrast to the ‘drugging’ of protein clefts and pockets (which usually include enzymatic active sites).
This ‘R factor’ can be viewed as a kind of general bias factor for small molecules in Natural Molecular Space towards the common range of protein folds. (This aspect of natural bioproducts as a very non-random selection of general chemical space has been considered in a previous post). But on the other hand, compounds modulating protein-protein interactions (as opposed to intraprotein folds) appear to be harder to find within the same natural set. To illustrate this further, consider the following hypothetical situation:
A series of human medical needs are compared, assuming complete knowledge of the underlying problem and the best protein molecule(s) which should be targeted for therapeutic improvement. These problems are grouped into two broad classes: In one type, the best target solutions involve clefts or pocket-like regions on a single protein; in the other, relatively flat protein-protein interaction surfaces present the best possible targets. Humans undertake a search of available biotas for natural products with therapeutic value for both types of pathological conditions. In principle, it does not matter whether tribal shamans or biotechnological operatives are involved, although of course the former act in complete ignorance of the underlying reality (biotechnologists may or may not possess full detailed target information in advance of their search). It can be then predicted that a preponderance of useful ‘hits’ for the cleft/pocket targets would be obtained from Natural Molecular Space, over the protein-protein surface type.
It should be noted though, that the R factor here is not at all intended to represent an absolute measure of druggability, only what Natural Molecular Space tends to define as ‘druggable’. Thus, the druggability of protein-protein interactions, once believed to be intractable, is now seen as solvable in many (if not all) cases through ingenious artificial design approaches. The central ‘take-home’ message is that not all diseases and not all molecular targets are created equal, and as a consequence some have a higher likelihood of being vulnerable to a therapy based on a naturally available low-molecular weight drug than others.
D refers to the diversity of the total potential set of biomolecules accessible in the human environment. (‘Accessible’ here means that which is available for the normal geographical range of the human group. Nomadic groups may often cover different environments and thereby increase their exposure to diversity, but possibly at the expense of the time factor as below). The assumption here is that the higher the diversity, the higher the probability that beneficial biomolecules will exist in the environment, and indeed a rainforest has far more potential as a pharmaceutical source than an arctic tundra. (But note that although arctic and other environments with extreme conditions have lower biodiversity, they are nonetheless useful for bioprospecting in specialist roles. The adverse conditions which restrict diversity in the first place also mean that life in these regions has special adaptations which can be useful for biotechnology, such as ‘antifreeze’ proteins. The same observations also apply to other extreme environments, including the opposite pole of very high temperatures.) How does one define and measure diversity, anyway? In molecular terms, one must have yardsticks for comparing molecules with each other, in order to assess their relative dissimilarities. A large number of such ‘descriptors’ have been devised, including molecular structures, shapes, and chemical and physical properties, which must be rendered into mathematical representations for formal modeling purposes. But a working definition of a diverse library of molecules is one which spans a chemical space in a non-redundant fashion, by means of covering a wide and continuous range of properties. A large group of molecules can thus be less diverse than a smaller set, and a set which only included members at extreme ends of property values would have reduced diversity in comparison with a set whose members possessed properties covering a wide dynamic range. Even so, no absolute standard for assessing diversity exists, although many sophisticated approaches have been developed.
But to return to out environmental considerations, an implicit assumption is that the higher the D value, the greater the odds that the total collection of environmental compounds will include one which is fortuitously useful for a human requirement, even if this is totally unrelated to the function for which the compound was originally derived by natural selection. But this D value is clearly not a random portion of chemical space, since the environmental biomolecules of inherent value are far from randomly generated (as discussed in a previous post). The ‘druggability’ factor R noted above already addresses a general bias of Natural Molecular Space towards a relatively limited set of protein folds found in nature (compared to the huge potential size of protein sequence space). But there are more direct factors which may also bias the prospect of success towards a given biomolecular target, and this is the basis for considering a specific bias element (the B factor raised above).
A full discussion of this would become complex and very lengthy, but for the present purposes, consider the notion of functional bias within Natural Molecular Space towards a human need, and consequently specific targets. ‘Secondary metabolites’ (raised in a previous post) are present as a result of a long history of natural selection processes in response to environmental pressures on the organism in question. In certain cases these influences may strongly bias the repertoire of available biomolecules in a positive manner from the point of view of human needs and desires. For example, a search for an anti-bacterial drug may have far greater prospects for success from sampling natural sources than from a random collection of molecules of comparable size. (Selective pressures have resulted in the generation of anti-bacterial products by a variety of organisms, and in this sense the ‘aims’ of the producer organisms and the human biologist coincide). Conversely, a search in a natural environment for a molecular function absent from the totality of environment’s biology (Such as a low-molecular weight organic catalyst of a non-natural chemical reaction ) will depend on chance alone for its success. To take this kind of functional bias into account, the B factor is therefore needed. (The higher this bias factor, the higher the probability that molecules of the type sought after exist in the natural environment).
Nevertheless, this ‘bias factor’ caveat is not necessarily as limiting as it may first seem, and can actually work in favor of the bioproduct ‘ocean’ as a source of useful molecules. The shared evolutionary origins of life on Earth means that many biochemical pathways of humans and their domestic animals also have analogs in environmental organisms. In some cases receptor proteins between widely evolutionarily separated organisms may have divergent functions but recognizably homologous structures and primary sequences. This means that a search for a biological modifier within a diverse natural environment in many cases may have a higher probability of success than within a random molecular collection of comparable size, and in turn stresses the worth and value of screening for active natural products.
As noted in a previous post, if it were possible to screen an utterly alien biological world, this positive bias factor would disappear, and such evaluation of the alien natural environment would be equivalent to a random molecular library screen. (Alien micro-organisms might generate secondary metabolites for similar selective reasons as for terrestrial organisms, but the chances of an alien ‘penicillin’ affecting Earth bacteria would be very low). Obviously the ‘molecular bias factor’ was not of great concern for tribal peoples, and in any case will often be difficult to predict in advance. But however one defines the chemical diversity of the total available environmental resources of biomolecules, in the context of ‘natural bioproduct space’ diversity alone is an insufficient guideline.
To return the above factors determining the likelihood of a human culture finding useful environmental molecules, C denotes ‘cultural factors’, probably impossible to quantitate but nonetheless real and important in the present context. Not all human groups will find and use biopharmaceuticals at the same rate, and the determining factor here is the shared set of social values referred to as ‘culture’. Within this set of values we will find such qualities as openness to enquiry and resistance to change. All human beings have the same fundamental genetic endowment, and it has even been suggested that the human talent for ‘folk biology’ may be based on an innate neural module facilitating recognition of differing plant forms or other biodiversity. These considerations aside, clearly human societies vary in their tolerance for behavioral experimentation and their willingness to implement new procedures, both of which can influence the rate of identification and adoption of pharmaceuticals from the bioenvironment. It was noted above that we must take care to define utility of a treatment at the outset, since it is also clear that some tribal medicines have little or no real effects on their supposed target diseases. No early pharmacognosy could pass judgements based on proper clinical trials of biomaterials, and the placebo effect would routinely be present on top of any real benefits deriving from administration of candidate preparations to ill individuals. (Given the demonstrated power of placebos in medicine, it is to the credit of tribal human groups that treatments of real effectiveness could nonetheless be identified). The cultural factor of passing on a time-honored tradition can effectively ensure that knowledge of genuinely useful medicines is retained by a group over time, but unfortunately bad memes can be just as transmissible as good ones. A case in point is the continued use of rhinoceros horns for medicinal purposes in some parts of the world, a false supposition whether for fever reduction or as an aphrodisiac.
Finally, t of the formula above is the time factor, on the logical assumption that the longer a group inhabits a particular environment, the greater the opportunities for biomaterial trial-and-error experimentation. We have already noted the potential for duration to have an impact on the diversity factor for nomadic groups, but time is also relevant to the C (cultural) factor, since even the most hidebound and anti-innovative group may eventually stumble on and adopt useful materials if they dwell for a long enough period in the same environment. And cultures, of course, are not static systems but are themselves mutable with time.
An Interesting Case Study – Cinchona and Malaria
The pathway towards finding that cinchona bark is an effective treatment for malaria serves to demonstrate the complexities that can impinge upon the time / culture factors, and an interesting example to consider in a little more detail. In historical terms, cinchona was identified as a malarial cure relatively soon after the arrival of Europeans in South America. Since it is generally accepted that the malarial parasite did not exist in the New World prior to the European invasions, cinchona may be viewed as an example of rapid acquisition of a natural medicine. Although Peruvian Indians are often credited with both the discovery of the anti-malarial effects of cinchona and its transmission to Jesuit or other European visitors, most available evidence suggests that it is more likely that Europeans themselves first came to this knowledge. But upon reflection, it is not probable that newcomers would have tested cinchona bark so rapidly without benefiting from the accumulated lore of the original inhabitants. Some sources indicate that native Peruvians used cinchona bark to relieve shivering, which can occur through pharmacological effects of active constituents of the crude bark. Quinine itself can act as a muscle relaxant, and also has inherent anti-pyretic activity, both of which could alleviate unwanted shivering depending on the cause of the problem. In any case, awareness of this on the part of Europeans could easily act as a prompt for testing the bark in malaria sufferers, owing to the intense fevered shivering occurring at certain stages during the disease cycle. Thus in this particular set of circumstances, the relatively rapid identification of the utility of cinchona bark (and thence quinine) is ironically likely to have resulted from a side-effect of the bark not directly related to its ability to kill malaria parasites. Also, it is most likely (though not provable) that it developed from interplay between traditional native Peruvian empirical lore and European activities. While the story of cinchona may be a special case (and it is certainly noteworthy in that it has occurred during historical times, even though much of the specific detail is obscure), the adoption of many other traditional pharmaceuticals may have similarly tangled origins.
All of the factors except C (‘cultural’ factors) are also relevant to modern empirical screening of natural sources. At least, cultural factors are not operative in the same sense as for isolated independent tribal groups. Potentially inhibitory cultural factors no longer apply since the acquisition of new drugs has become effectively a global enterprise, transcending national barriers. On the other hand, the uptake of a new drug in the modern world depends on regulatory agencies, marketing, and commercial competition with rival products, all of which can be considered as ‘cultural’ forces under a broad definition of the word.
By the arguments presented here, a traditional society would have maximal success for finding useful natural molecules for a specific need if all of the above relevant factors were optimal: Highly druggable ultimate target(s) • Rich natural local environment, (such as rainforest) • Highly positive specific functional bias (for example, sought-after treatment could be satisfied by metabolites of diverse fungi or other organisms) • Highly acquisitive culture, highly receptive to innovation • Extensive duration of time for conducting ad hoc empirical testing of environmental resources.
To finish up, a double-barreled offering of biopoly(verse) of relevant note:
Are shamans biology sages
Possessing the wisdom of ages?
Some shamanic insight
In a drug textbook write
Might add some significant pages
Some chemists consult tribal lore
Seeking botanical info, and more
From suitable plants
They’ll thus take a chance
That good drugs will pay them a score
But where Westerners (especially those with commercial motivations) “consult tribal lore”, the outcome may be an exploitative one-sided affair sometimes termed ‘biopiracy’ – and that will be taken up in the next post, among other things.
References & Details
(In order of citation, giving some key references where appropriate, but not an exhaustive coverage of the literature).
‘ Examples of these [natural bioproducts as medicinals] ….particularly from local plant sources’ See Rates 2001 for a general review, with specific examples.
‘…..jungle regions such as the Amazon have been rich sources of biopharmaceuticals…..’ See Schultes 1994. This author, the late Richard E. Schultes of Harvard University, was renowned for his ethnobotanical studies, including works on plant-derived hallucinogens. In the latter vein, he co-authored a well-known book in 1979 (Plants of the Gods: Origins of Hallucinogenic Use. New York: McGraw-Hill; with revised editions in 1987 and 2001) with Albert Hofmann, the discoverer of LSD.
‘…….human cultures …… have identified essentially all plants which can become dietary items.’ This point has been discussed by Jared Diamond in his well-known book, Guns, Germs and Steel (Vintage Books, 1997).
‘……where certain plant materials possess initial toxicity, but which can be removed by specific processing steps….’ A good example of this is the case of cycads of certain Pacific islands. The seeds of these ancient plants carry both nutrients and toxins, requiring pre-treatment by extensive washings and soakings in order to render the seeds edible for humans. This same example also raises another interesting general issue: humans screening their environments for edible plants will reject those that obviously poisonous (or finds ways to de-toxify them), but some plant sources may have toxins which produce subtle damage that is far more difficult to detect. Peoples of Guam and several other Pacific locales have suffered high rates of neurodegenerative diseases (with features of amyotrophic lateral sclerosis, Parkinson’s disease, and dementia, in varying combinations), and a long-standing hypothesis has attributed this disease focus to the local habits of cycad product consumption. (See the book The Island of the Colour Blind, by Oliver Sacks, Picador 1996, for an extended account of these issues). Yet this has been highly controversial, with varying forms of the ‘cycad hypothesis’ proposed and then discounted. It has become clear, though, that an environmental contribution is highly likely, as the disease incidence is in decline, in parallel with dietary changes (see Steele 2005; Steele & McGeer 2008).
‘…..historically infections and parasites were often relatively local….’ For more on this, one can refer to the same Jared Diamond book again as above (Guns, Germs and Steel ; Vintage Books, 1997).
‘…..even potential treatments for genetic conditions…’ It might be thought surprising that solutions to genetic problems could be found within low-molecular weight natural bioproducts, even in principle. A simple answer to this would be, ‘It depends on the nature of the genetic problem’. If a genetic lesion causes the loss of an entire genomic coding sequence for an essential large protein, clearly it is extremely unlikely that this defect will be ‘fixed’ by ingestion of a small natural molecule. But on the other hand, many genetic problems result from premature termination of protein synthesis, or mutations which cause protein misfolding. In the former case, mutations which result in the formation of abnormal stop codons in an mRNA molecule can be suppressed by drugs which induce translational ‘readthrough’, and thereby enable production of the formerly abnormally truncated protein. Although recent clinical successes have used artificially-obtained compounds for this purpose (see Welch et al. 2007), the original definition of the effect was provided by naturally-derived aminoglycoside antibiotics (see Howard et al. 2004). In the case of protein misfolding, certain compounds can interact with and stabilize misfolded proteins in their correct conformations. Natural products (and artificial compounds) have been found with this kind of ‘chemical chaperone’ potential in treating certain specific genetic diseases (for example, see Brumshtein et al. 2007)
‘……natural landscapes may harbor drugs whose usefulness is not definable even in the present day.’ This is a potent argument for the retention of natural biodiversity, and to resist increasing destruction of natural habitats – especially those within resource-rich tropical zones.
‘…..compounds modulating protein-protein interactions (as opposed to intraprotein folds) appear to be harder to find within the same natural set.’ Note that this is not stating that such compounds are entirely absent in Natural Molecular Space. A notable case in point are natural immunosuppressants: cyclosporins, rapamycin, and FK506. These molecules act by promoting the formation of ternary complexes which inhibit the activation of a key transcription factor for T cell activation. For some details, see Mann 2001; Lee & Park 2006.
‘……the druggability of protein-protein interactions, once believed to be intractable, is now seen as solvable.’ For progress in this area, see Dömling 2008.
‘…..a diverse library of molecules is one which spans a chemical space in a non-redundant fashion….’ For background on medicinal chemical diversity, see Gorse 2006.
‘…..no absolute standard for assessing diversity exists, although many sophisticated approaches have been developed….’ See Maggiora & Shanmugasundarum 2011, for a recent mathematical treatment of molecular similarity, which measures the flip side of diversity.
‘…….it has even been suggested that the human talent for ‘folk biology’ may be based on an innate neural module…..’ This was made by Steven Pinker in The Language Instinct (William Morrow & Co., 1994).
‘…..Peruvian Indians are often credited with both the discovery of the anti-malarial effects of cinchona and its transmission to Jesuit or other European visitors, most available evidence suggests that it is more likely that Europeans themselves first came to this knowledge….’ For much information on this on related topics, see Fiammetta Rocco’s excellent book; The Miraculous Fever Tree (HarperCollins, London, 2003).
This post continues the current theme of ‘natural molecular space’ and its usefulness. Here, and in the next two succeeding posts, the exploitation of natural biomolecules is considered as a history of sorts. Where does this old, old story begin? When looking at the acquisition of natural products as pharmaceuticals by humans, one might expect to go back to pre-historical times, or even to consider the question in the context of human evolutionary development. And yet it seems we need not stop there. …for the exploitation of natural molecular space can be framed as a general biological issue, and this is the topic of the day.
Zoopharmacognosy – Self-Protection from Threats, Large and Small
There is now a considerable literature describing the use of plant materials by various animals specifically for self-treating their health problems, usually parasitic infections. Many well-documented reports in this area concern apes and monkeys, although some suggested instances of primate self-medication are plausible but based on circumstantial evidence. More controversial findings have suggested that the same kind of phenomena exist in mammalian herbivores.
Since ‘pharmacognosy’ refers to the isolation of pharmaceuticals from natural sources, ‘zoopharmacognosy’ has not surprisingly been coined as a term for the phenomenon of directed animal self-medication. There is a sizable body of evidence, and a reasonable evolutionary rationale, to suggest that zoopharmacognosy is real and worthy of study. At the same time, given the human tendency to anthropomorphize the behavior of animals at the least opportunity, interpretation of animal actions and motivations should always proceed with caution.
If interpreted broadly, ‘self-medication’ can include the ingestion of environmental materials for defense against external predators as well as parasites, whether the latter are microbial or macroscopic. In fact, it has been suggested that the directed ingestion of plants for purposes not directly related to animal nutrition per se may be an ancient practice long predating vertebrates. Consider an insect eating a plant which is non-toxic to itself, but which renders the insect unpalatable to predators. Insofar as such behaviors are innately programmed, they should be subject to natural selection, where an individual with a specific innate eating preference gains a survival advantage. Therefore, variance in genes determining behavior may constitute the raw material for an evolutionary process modifying non-nutritive food ingestion. The patterns of such behavior, though, may be themselves complex. For example, plant self-medication in caterpillars of a specific insect (lepidopteran) species has been observed to occur only in parasitized animals, towards which the ingested plant materials are beneficial.
The use of environmental molecules (natural molecular space) for defense against macro-predators also exists in vertebrates. A very recent example has been reported which amply demonstrates this, and in fact is the only known case of such a phenomenon in placental mammals. The African crested rat (Lophiomys imhausi) has been found to chew roots and bark of a tree (Acokanthera schimperi) which make a compound with cardiotoxic effects on the rat’s large predators. (Chemically related compounds, such as ouabain from the same genus of plants are well-known for their effects on heart function, and have medical applications). This rat then transfers (slavers) the pulped plant material onto specialized hairs which soak up the added material, and these hairs thus become primed to act as toxic delivery systems for any unfortunate predator attempting to eat a Lophiomys individual. An obvious question here: why isn’t the rat itself bothered by the plant toxin? It was suggested by the same group that the rat may produce compounds in its copious saliva which bind and neutralize the toxic principle. Whatever the specific details, in these circumstances the provision for a self-protection mechanism while producing a defense against predators is exactly analogous to the process schematically depicted in a previous post for defense against local competitors.
A purist might argue that acquisition of environmental chemicals for defense against external predation is distinct from true zoopharmacognosy, where the ingested material is ‘aimed’ at fighting internal infections or parasites. There is obviously a distinction between these activities, but splitting the labeling only comes down to a semantic issue. But there is no question that all self-medication phenomena, whether directed against predation or parasites, involve animals sampling regions of natural molecular space accessible within their environments for their own benefits (technically, increased evolutionary fitness).
Innate vs. ‘Cultural’ Zoopharmacognosy
Still, there is one broad division within the whole field of animal self-medication / zoopharmacognosy: behaviors which are innate (as with self-medicating caterpillars), and those which are learned and transmitted by example, in a ‘horizontal’ fashion rather than ‘vertically’ by genetic inheritance. Horizontal transfer involves a ‘culture’, in that an different isolate of the same species (with the same genetic background) may not show the same behavior through lack of direct exposure to it. Certainly there are potential complications with this simple dichotomy. For example, a ‘plastic’ behavioral phenotype (variable outcome on behavior through specific gene action) may result in an adaptive (fitness-promoting) behavior being selected for, and subsequently becoming ‘fixed’ in a population through further genetic change. Also, in certain circumstances, an ‘innate’ behavior may not necessarily be manifested in an isolated individual unless it has been ‘primed’ by a degree of maternal or social interaction.
As a brief aside, the innate / ‘cultural’ divide can viewed through the lens of the ‘extended phenotype’ concept. This idea, first developed by Richard Dawkins, proposes that the action and expression of genes (in combination, the phenotype of an organism) do not necessarily stop at the boundary of the organism’s body. Classic examples are beaver dams or directed modulation of host behavior by parasites. But this general concept within evolutionary biology is very often abused and over-extended. A crucial criterion for a true extended phenotype by Dawkin’s definition is that there must be a correspondence between the success or failure of a putative phenotypic extension and the genes which influence the behavior or activity responsible. Gene variants (alleles) which direct alternate forms of a particular phenotypic extension are thus subject to Darwinian selection, according to the success (or not) of the ‘outer’ phenotypic effect. With this in mind, it can be seen that a building is not a human extended phenotype, since the building outcome has no effect on the frequency (relative genetic allele success) of relevant architect or builder’s genes in the total population.
Although innate behavior which drives evolutionarily useful self-medication does not build anything outside an organism’s body, it uses environmental materials (specific molecules, or sets of them) to provide a fitness benefit. Since neither these molecules nor the enzymes which make them are encoded in the genomes of such organisms, it follows that the gene-modified behavior directs the formation of a chemical extension to the organism’s phenotype. On the other hand, a truly ‘cultural’ transmission pattern violates the gene correspondence guideline, since adoption of such a cultural trend among a population does not favor the gene frequencies of the original innovators of the useful behavior. Therefore, this (somewhat simplistic) division of self-medication / zoopharmacognosy can indeed be used to illustrate both real instances of extended phenotypes and pseudotypes thereof.
But wait a minute, you might well ask. Culture is a human attribute. No one would dispute the importance of horizontal spread of information in Homo sapiens, and its continuation across generations by cultural propagation. But where does ‘cultural’ transmission of useful self-medication occur in the animal world? In fact, some work suggests that our primate cousins do share self-medicating activity in part through observation and imitation. (To invoke Dawkins once again, perhaps one could call this transmission of primate memes through aping). Where ‘cultural’ transmission of self-medication is possible, clearly the associated feeding behavior would have the potential to spread at explosive speed relative to natural selection. Clearly, a key factor here is intelligence, and perhaps a primate-level of cognition is necessary for true ‘cultural’ transmission of self-medication information to occur.
Another issue arises: We should also be careful to distinguish true zoopharmacognosy from any animal self-medicative behavior which is motivated by a direct positive reward from the consumed natural material. This is simply because a pleasure response elicited from eating (for example) a particular drug-bearing plant can produce a direct feedback behavioral loop – the consumption of the plant material and the reward ‘kick’ are relatively easy to connect on a cause-and-effect basis. Also, continued self-application of a pleasurable natural drug stimulus will almost always be neutral at best in terms of disease control, and may be generally deleterious for health if abused, or if addiction-related changes occur in the animal.
There is an extensive literature dealing with animal models for substance abuse and addictive behavior. For studying behavioral aspects of addiction, higher animals are usually required, but for the underlying effects of the drugs on neural systems, even invertebrates may do just fine. But where the result of consuming the natural material is alleviation of a health problem, the ‘reward’ (mitigation or elimination of feelings of ill-health) is unlikely to be so closely correlated in time with the original eating behavior, and indeed might often be preceded by even worse symptoms before improvement is noted.
Thus, where specific cases of zoopharmacognosy have been acquired through learning rather than being innate behavior, it might be presumed that the learning process involved may be sufficiently complex to restrict it to primates. (It is presumably easier to come up with an alcoholic rat than one with a true flair for zoopharmacognosy). On the other hand, rats can be quite sophisticated in their ability to use ‘delayed learning’ to determine if sampled foodstuffs are noxious or not. In such circumstances, a rat will sample a novel food and wait for a time (half an hour or so) to decide whether or not the ingestion of the food is associated with any negative outcome. But correlating a positive outcome (such as parasite reduction) with specific food consumption is a much taller order, since the time lag between ingestion and effect will usually be much longer.
Watch Animals and Learn?
Two issues relating to zoopharmacognosy are of special interest: did tribal humans in the past learn the value of certain plants from watching animals, and is it profitable at the present time to observe wild animal feeding behavior for obtaining new pharmaceuticals? Unfortunately, the biological source and mode of preparation of the majority of tribal medicinal preparations have been handed down through a long tradition which renders accurate knowledge of the origin of such practices impossible to obtain. Yet while we cannot be certain, it is certainly conceivable that at least some such lore was derived from animal observation, and some ‘living examples’ have been claimed. Is this relevant for drug development today? If self-medicative behavior is a significant factor in maintaining group health in primates, in theory the nature of the consumed plants could be identified by careful observation of the animals in the wild, with the possibility that the active constituent of the identified plant could be a useful pharmaceutical. Recently this has been claimed to be correct, with new candidate anti-malarials and other potentially useful drug candidates resulting from original observations of chimpanzees in Uganda.
To conclude, a relevant musing from biopoly(verse) once more:
Throughout the natural world, I surmise
Instinct can render an animal ‘wise’
And if self-learning can train
A higher animal’s brain,
Is ‘animal pharma’ any surprise?
References & Details
‘……a considerable literature describing the use of plant materials by various animals specifically for self-treating their health problems….’ Engel, C. Wild Health: How Animals Keep Themselves Well and What We Can Learn From Them (Houghton Mifflin, 2002); also Larkins & Wynn 2004; Raman & Kandula 2008.
‘…..some suggested instances of primate self-medication are plausible but based on circumstantial evidence.’ For example, see Carrai et al. 2003. This study of sifakas (prosimian primates from Madagascar) found a group where pregnant females eat a tannin-rich diet compared to other females and males. Females eating the tannin-rich plants had fewer pregnancy failures than those from another group with a diet lacking the tannin loading. Thus, it was inferred that the tannins assisted the pregnancies, possibly by acting as anti-parasitics. But it was not proven that the tannin-eating sifakas directly benefited from their diets, as other environmental factors (such as reduced stress) might have been the underlying cause. See also a New Scientist article on this topic.
‘…..it has been suggested that the directed ingestion of plants ….may be an ancient practice long predating vertebrates.’ See Huffman 2003.
‘…..recently plant self-medication in caterpillars of a specific insect (lepidopteran) species has been observed to occur only in parasitized animals……’ See Singer et al. 2009. This phenomenon has been suggested to be a special case of behavioral ‘adaptive plasticity’.
‘ The African crested rat …..has been found to chew roots and bark of a tree …..with cardiotoxic effects on the rat’s large predators.’ See Kingdon et al. 2011.
‘…..a ‘plastic’ behavioral phenotype …..subsequently becoming ‘fixed’ in a population….’ See Ghalambor et al. 2007 for a discussion of this general topic.
‘…… the lens of the ‘extended phenotype’ concept….’ See Dawkins’ book, The Extended Phenotype – The Long Reach of the Gene Oxford U. Press, 1982.
‘…..a building is not a human extended phenotype……’ This example paraphrased from Dawkins 2004. Obviously, the same genetic argument applies to any other human artifact, including computers, despite the latter often being portrayed as extended phenotypic examples.
‘…..an extensive literature dealing with animal models for substance abuse and addictive behavior…’ See Gardner 2008.
‘…..for the underlying effects of the drugs on neural systems, even invertebrates may do…’ See Wolf & Heberlein 2003.
‘…..easier to come up with an alcoholic rat than one with a true flair for zoopharmacognosy.’ The propensity to become dependent on alcohol in rats is genetically determined; for example, see Murphy et al. 2002.
‘…..rats can be quite sophisticated in their ability to use ‘delayed learning’..…’ This is based on the work of Paul Rozin (See Rozin, P. 1976. The Selection of Foods by Rats, Humans, and Other Animals. In Advances in the Study of Behavior, Vol 6, Eds J. Rosenblatt, R. A. Hide, C. Beer, and E. Shaw. (NY-Academic Press). PP 21-76.). For a lively discussion of this area, see also Michael Pollan, in his book The Omnivore’s Dilemma Bloomsbury Publishing 2007; (Start of Ch. 16).
‘……some ‘living examples’ [of tribal medicine learnt from animal behavior] have been claimed….’ See Huffman 2003.
In the previous post, we looked at the world of natural biomolecules, which can be considered as a ‘natural molecular space’. Before that, the notion was raised of ‘chrestomolecules’, or general molecular entities of economic significance in the most general sense. Now, it’s appropriate to bring these together and take a look at the relevance of natural molecular space to human beings, with a particular glance at a ‘why’ question: If molecules found within natural molecular space are especially useful to us, why should that be so?
Within the last few hundred years, certain observations can only have reinforced the supposition that the natural world was a source of cures for all humanity’s innumerable ailments. Let’s consider one case as an exemplar of this. In the early 18th century, the navies and merchant marines of European countries were plagued on oceanic voyages by the terrible affliction of scurvy, which often cut like a scythe through the hapless ranks of seafarers. We know today the exact problem and molecular details of the solution, which of course is correcting the deficiency of Vitamin C (ascorbic acid), a cofactor for enzymes involved in collagen formation and other functions. As such, scurvy (and other vitamin deficiency diseases) are in a very different category to infectious diseases and other human pathological conditions. But before this was clarified, to common sailors and ‘learned’ physicians alike, scurvy was bundled in with other diseases and just as mysterious (although no shortage of erroneous theories were proffered by the latter). It was eventually found that certain plant products and citrus fruits were ‘anti-scorbutic’, and could often completely reverse the shocking physical ravages of the disease. A ship’s surgeon witnessing near-miraculous recoveries of his sickening and dying charges, and ignorant of the underlying causes of the condition, might speculate that if some plant products could cure otherwise-terminal scurvy with such incredible efficacy, might not other diseases have a comparable natural cure somewhere in the world? If only one could find them…..
We then can jump from the former intuitive appreciation of ‘mother nature’ to current understanding of the tumultuous diversity of the Earth’s biosphere at the molecular level, which is a worthy and interesting intellectual pursuit in its own right. Humans being constituted as they are, however, such higher-level goals usually tend to take a back seat in favor of more pragmatic questions along the line of, ‘What’s in it for me?’. If the question concerns the bounty and application of natural molecules, the answer is ‘Plenty’. It might seem that asking why a situation exists is primarily driven by scientific curiosity, and should take a back seat in favor of more practical issues. Yet even a partial answer explaining the reason for observed circumstances can have potential practical pay-offs.. So let’s take a closer look….
Common origin, convergence, or co-incidence?
Why should a fungal biomolecule, or any other natural product for that matter, provide therapeutic benefit to a human being? Pragmatically, this ‘why’ question is often side-stepped, and emphasis given to studies on how potentially useful biomolecules work in mammalian cells. Obviously the latter focus is important for pharmacology. Yet a deeper understanding of why a molecule from the biological environment should have impact on human biology in the first place may also have practical pay-offs, as well as purely scientific advancement. A first basic point to note is the common origin of life on Earth, which ensures that many biomolecules and biomolecular systems have functional homologs across wide evolutionary gulfs, thereby raising the probability that metabolites from (for example) fungi will influence human biology. Furthermore, it has been pointed out that the diversity of protein structural folds is much less than protein sequence diversity alone would suggest, and that this has implications for the range of low-molecular natural bioproducts as well. The crux of this argument is that despite the great diversity of protein sequences, all types of natural protein folding shapes can be grouped into a much smaller set. This nevertheless remains a small number compared with corresponding set of primary protein sequences. Consequently, molecules selected by evolution to bind to such folds will tend to have special features promoting such interactions. For example, molecular volumes of most natural products are within the volume size range found within protein cavities. Through these constraints, natural biomolecules are a highly non-random representation of the totality of chemical space. To consider it from a different viewpoint, there is a much better chance of finding a binding molecule for a specific protein fold category in natural molecular space than in a totally random collection of chemical structures. This has been borne out by studies with artificial chemical libraries.
Beyond these general considerations lies the issue of the specific selective pressures determining the structural features of ecobiomolecules. These are simple to define if both the function of such a molecule in its natural context and its applicability to humans appears to coincide, as found in the case of antibiotics. Yet things are often not so clear-cut, and some bioproducts are in a gray area even from the viewpoint of their putative beneficial effects. Consider in this context the example of phytoestrogens. These are plant-derived molecules which mimic the action of specific mammalian steroidal sex hormones (estrogens), by binding to the estrogen nuclear receptor which transmits estrogen signaling. Phytoestrogens fall into four main different structural groups, including the isoflavonoids which are found within soybeans, a major food item in many cultures. A long-standing controversy has run concerning putative health benefits of dietary isoflavonoids (reducing breast and prostate cancer risk) and possible negative effects on human reproductive cycles or other areas. Although phytoestrogens may be viewed as ‘natural food additives’ as considered above, it may thus be premature to refer to isoflavonoids as ‘chrestomolecules’ (see an earlier post). In any case, let us get back to the ‘why’ question posed above. Why should plants produce chemicals which mimic mammalian sex hormones?
Note for Figure: Structures of an estrogen (17-β-estradiol) and three examples of different phytoestrogen classes (genistein [an isoflavanoid], resveratrol, and coumestrol). Also shown is a representative brassinosteroid (castasterone), with its steroid-like scaffold shaded in gray. Castasterone is metabolized into the more potent brassinolide, which has a seven-membered ring, indicated with an arrow. Also shown is the protein structure of estrogen receptor-β ligand binding domain in complex with genistein. (genistein ligand black; α-helical regions of protein red, β-strands green, loops light blue). The binding mode for genistein is similar but non-identical to that observed with 17-β-estradiol.
One possible answer, of course, is that plants do not gain anything from the mimicry itself. In this view, whatever the natural plant function of isoflavonoids (and other phytoestrogens), their effects on certain human nuclear hormone receptors is an accidental evolutionary cross-reaction. Or to rephrase this proposal, however important isoflavonoids might become to humans, it is of no evolutionary consequence to the plants which produce them. This in turn postulates that the selective pressures which resulted in plants synthesizing these compounds in the first place had no input from mammalian biological interactions. A somewhat more subtle version of this ‘pure chance’ viewpoint (along the lines of above discussions of natural molecular space) would hold that owing to the limited number of protein folds observed in nature, a certain level of spurious cross-binding between receptors and ligands from evolutionary divergent species is inevitable. Alternatively, it could be quite feasibly proposed that both plant and mammalian receptors for phytoestrogens happen to share a common (albeit distant) evolutionary origin, such that (despite functional divergence) they also share certain structural features which enable cross-binding of their respective ligands. Such proposed effects boost the attractiveness of natural molecular space as a source of useful bioactive compounds.
A defining chemical feature of estrogens and other steroids is their molecular ‘skeleton’, the perhydrocyclopentanophenanthrene fused ring structure (This is seen in the core ring composition of 17-β-estradiol in the above figure). Although (as noted above) several structural classes of phytoestrogens exist, none possess this basic steroidal feature (as in the above Figure, with coumestrol, resveratole, and genistein as examples). Structural studies with the phytoestrogen genistein have revealed that its binding to the estrogen receptor-β (one of the estrogen-binding nuclear receptors) shares many features to that seen for the natural ligand 17-β-estradiol, although genistein’s binding is suboptimal for the induced structural changes necessary for activation of the receptor. This finding is consistent with the observation that genistein (like other phytoestrogens) is estrogenic, but more weakly than the natural estrogen hormones themselves. Structural data thus show that compounds without a complete steroid ring system can recapitulate many of the appropriate binding contacts with the estrogen receptor, but cannot in themselves inform us as to the ultimate source of this mimicry.
Interestingly, despite its absence in phytoestrogens, the characteristic steroid ring is found in certain plant sources, including useful drugs such as cardiac glycosides. In fact, some plant steroid-like compounds, such as hecogenin, have frequently been used as source material for beginning synthesis of steroidal drugs. Most significantly, though, an important class of plant hormones possess the steroid ring system or a very close analog of it (See castasterone in the above Figure). These ‘brassinosteroids’ are synthesized by enzymes which have been shown to be homologous between plants and animals, pointing to a common (albeit distant) evolutionary origin between plant and animal steroid hormones. (Looking at the enzymes involved in natural product formation is the best way to trace the evolutionary ancestries of the natural small molecules themselves). So the existence of bona fide plant steroids is not at all a coincidence, yet these brassinosteroids have not been described as phytoestrogens, while several other distinct molecular structural classes (lacking the steroid ring system) have proven estrogenic activity (as in the above Figure). This seems at least superficially surprising, and might be taken as supporting the chance-mimicry hypothesis for phytoestrogens. But this is not quite the end of the story….
At this point we must compare the natural functions of phytoestrogens with their estrogenic activities in mammals. One ascribed phytoestrogen natural role beneficial to their plants of origin is antimicrobial activity. Genistein is the precursor of other known plant antimicrobial compounds, and has some antibacterial effects itself. But this does not readily explain its activation of estrogen receptors. Other ecological functions for phytoestrogens have been postulated, such that their effect on mammalian hormone receptors is not coincidental. Certainly the reproductive systems of ruminant herbivores can be seriously affected by consumption of phytoestrogen-bearing plants, and this has been proposed to have selective benefit to such plants by control of herbivore population size. For this to occur, though, one would have to invoke an evolutionary group selection mechanism. (If herbivore populations were reduced through consumption of a plant species where a minority of variant individuals produced toxic or inhibitory phytoestrogens, all members of the plant group would gain a survival advantage, not just those bearing the genes enabling phytoestrogen synthesis). Evolutionary selection at group levels has been a problematic and contentious issue. But in any case, there is an alternative explanation.
Relatively recently, information has emerged suggestive of a real ecological role for phytoestrogens in chemical communication between plants and beneficial bacterial symbionts. There is good evidence that at least isoflavonoids are involved in signaling between nitrogen-fixing bacteria (Rhizobium and related genera) and their leguminous plant hosts, by means of direct interaction with the bacterial transcriptional regulator NodulationD1 (NodD1). It then follows that it is of considerable interest to compare NodD1 and estrogen receptors. Although they share no significant sequence homology suggestive of a common evolutionary antecedent, folding (shape) homologies between the two cannot be ruled out until further structural information is available. It has been suggested that NodD1 and estrogen receptors have undergone ‘convergent evolution’ in response to estrogen-like molecules as signaling agents. Such a contention does not in itself explain what selective forces drove both systems towards a related end-point, and why some other receptor-ligand system could not have served equally well in either case (obviously other receptor-ligand systems exist in both organisms). Chance again at a different level, or were additional selective pressures operative (as with the possible role of herbivore control for the phytoestrogens)? These considerations serve to demonstrate the complexities which can arise in attempting to unravel the evolutionary origins of ecobiomolecules.
So what is the take-home message at the end of this rumination (so to speak) on phytoestrogens, herbivores, and more? The phytoestrogen story also shows that one must be wary in ascribing any apparent molecular cross-reactivity purely to chance, given the possibility of signaling mechanisms between ecologically associated organisms of diverse evolutionary lineage. Based on bacterial-eukaryotic relationships, ‘inter-kingdom’ signaling has been proposed to extend beyond responses to pathogens alone, and the symbiotic relationship between nitrogen-fixing bacteria and certain plants (noted in the context of phytoestrogens) is a case in point. An enthusiastic view of the importance of ecology would hold that most (if not all) biological chrestomolecules owe their usefulness to specific inter-relationships ultimately traceable to the molecular level. In other words, such an opinion would hold little place for coincidence in the provenance of biomolecular human utility. Yet even if the estrogenic activity of molecules such as genistein is explained through structural similarities between its natural target bacterial receptors and mammalian estrogen receptors, we should remember that many phytoestrogens possess multiple activities in animal cells. Genistein is something of a champion in this regard, and it becomes unlikely that all of these effects have an underlying evolutionary rationale.
To find clear-cut cases of fortuitous interactions, we can look beyond biological molecules. Further inspection of the biology of estrogen receptors themselves reveals examples of molecular interactions which can only be attributed to chance. Since the advent of artificial chemical pesticides, many concerns have been raised about their safety. One important side-effect of many chlorinated hydrocarbon insecticides (including DDT) was revealed to be estrogenic activity, or mimicking of estrogens. DDT, for example, can directly bind and activate the estrogen receptor (albeit much more weakly than normal hormone), and DDT can even support the growth of estrogen-dependent tumor cells in culture. Now, the ‘selection’ for artificial insecticides was empirical searching for insect-killing abilities, and certainly not as ‘xenoestrogens’. It follows in turn that the interaction of such artificial compounds with estrogen receptors is an undirected chance event. (Although, as we have noted above, the probability for such events may be increased through the relatively limited number of protein folds in biological protein space). Interestingly, in an equivalent manner as for phytoestrogens, xenoestrogens can also interact (and interfere) with symbiotic nitrogen-fixing bacteria, another potential environmental problem associated with the use of such compounds. It is true of course that these artificial xenoestrogens are not derived from natural molecular space, but they do serve to demonstrate unanticipated molecular cross-binding, which could equally well occur in principle with a natural compound (if not more so, via their ‘pre-validation’ for natural protein folds). By the same token, completely artificial chemical libraries can yield ‘hits’ for modulation of biological processes, even if at low frequency.
Considering the impact of natural molecular space on humans in general, biases in favor of the activity of natural products thus exist both from a fundamental biological perspective, and in specific cases of ecological interaction. When chance plays a role, it may be in form of a negative side-effect, analogous to the xenobiotic DDT / estrogen receptor example. There are many examples of biological toxins deadly for humans which were never designed by evolution with humans as specific targets. The explanation for this, of course, is the sharing of a wide variety of biochemical pathways between humans and other organisms, especially our fellow mammals. Yet some quirks to this general rule illustrate the role of ‘molecular chance’ in the outcome of the interaction between humans and environmental biomolecules. For example, the venom of the Australian funnel web spider, one of the deadliest in the world towards humans, evolved for the purpose of quickly immobilizing its insect prey. Despite lethal effects of funnel web neurotoxins in humans and other primates, however, most other mammals are little bothered by them, apparently through neutralizing serum factors rather than at the level of the direct neural targets of the toxins themselves.
Having examined both artificial and natural undesirable chance-based molecular interactions, one could logically argue that bad results are likely to be much more frequent than beneficial ones. After all, (as with genetic mutations and many other things) it is far easier to damage an organized system by randomly plugging things into it than improve it. Still, beneficial random mutations do occur, and likewise a chance biomolecular interaction could have useful consequences. An obvious follow-up at this point is to ask for specific examples of valued biomolecules whose human utility is completely unlinked to the selective pressures by which such molecules originated. As we have seen with the phytoestrogens, judging this can be a complex undertaking, and it depends to some extent on what level of ‘co-incidence’ one is referring to. For example, marine organisms in general are hardly likely to be subject to direct selective pressures towards synthesizing effective anti-tumor drugs for humans, yet some such bioproducts are currently in clinical trials. But at another level, these cytotoxic or cytostatic anti-tumor effects are explicable owing to the evolution of chemical defenses in the sessile marine organisms from which they have been derived. Then it is only a matter of the correspondence between the biological systems and pathways of humans vs. those of the natural predators of the marine organisms against which the compounds are targeted.
Yet some aspects of the specific uses of natural products may indeed be chance-based, such as the effects of useful plant compounds called alkaloids (included within the ‘Therapeutics’ of the ‘chrestomolecule’ Table of an earlier post). As secondary metabolites, they are not essential for plant growth. The provision of an intensely bitter taste and toxicity may provide a selective advantage against herbivores in plants which produce them, although other functions (again analogously with phytoestrogens) cannot be ruled out. Useful effects of alkaloids on human cardiovascular or nervous systems (such as with reserpine or atropine) may nonetheless be hard to account for if not through fortuitous interactions. This may apply even more so to the dramatic effects some plant or fungal-derived psychotropic drugs have on human nervous systems. At a deeper level again, however, the ‘pre-validation’ of such molecules as protein-binding agents may greatly increase the likelihood that they will cross-react with other proteins (with related folds). As an example, to proceed from the premise that the intensely bitter taste of plant alkaloids was a selective factor in their evolutionary origins (providing a selectable advantage, through immediate deterrence of potential herbivores), a molecule selected for its binding to any of the mammalian bitter taste receptors might also fortuitously bind to other proteins with related structural features, with accompanying physiological effects. In fact, taste receptors are but one class among a large protein group called G Protein-Coupled Receptors, a huge target set for pharmacology.
‘Side-effects’ in general are unanticipated molecular interactions which are usually negative. Nevertheless, as we have seen, a useful property of an empirically-discovered natural product in principle might result from a beneficial ‘side-effect’ rather than its normal role. Yet even where the natural role and human application of a substance appear to perfectly coincide (as with antibiotics), it must be kept in mind that the usefulness of any biomolecule (especially in a therapeutic context) is also dependent on what it doesn’t do. In other words, irrespective of the origin of the useful activity of a molecule, we also hope for it to avoid a host of potentially troublesome side-effects. (‘Collateral damage’ and ‘magic bullets’ are mutually exclusive). And given the enormous complexity of living organisms, avoidance of side-effects will be fortuitous much of the time. To return to the example of antibiotics, the utility of penicillins depends not only on their anti-bacterial efficacy but also their low background toxicity. Clearly, fungal secondary metabolites of any description are not selected in favor of human biochemistry. Prediction of specific molecular side effects, from whole cell systems to whole animals, is dauntingly difficult and requires multiple rounds of empirical safety tests.
Natural Molecular Space: All Good?
We have paid attention to the advantages provided by natural bioproducts as a result of their great diversity. Are there any inherent disadvantages in harvesting nature’s molecular bounty, as opposed to using completely artificial compounds occupying regions of chemical space not present in the natural world? Perhaps not in general, but there is such an issue in the specific (but important) field of natural drugs against pathogens. Since many microbial organisms and humans have a common ‘need’ to control bacterial competitors or pathogens, then the existence of environmental antibiotics seems (with hindsight) entirely reasonable. Yet at the same time the seeds of future problems may exist from the same evolutionary logic. This is because (as we have seen in the previous post) the origin of antibiotics and their counter-measures are closely linked: organisms producing a potentially self-toxic molecule must simultaneously have the means for neutralizing the toxic activity towards itself. If counter-measures against an identified anti-bacterial drug already exist in the environment, there is also a high probability that they can be rapidly marshaled and spread. A compound which is derived by entirely non-natural means would not be associated with this problem, and counter-measures in the target organisms would need to newly evolve. This can occur by co-opting existing proteins or pathways into novel uses; gene duplication is a major route towards this end. From another viewpoint, binding of a cellular target protein by any small molecule inhibitor can be regarded as an extra ‘promiscuous’ (albeit non-natural) function of the protein itself, which usually involves intermolecular contacts outside of those which are essential for activity. Evolution towards resistance can thus arise by modifying promiscuous contacts rather than functionally vital ones.
So resistance to a natural anti-bacterial will tend to arise more rapidly than to a completely artificial one, and in general this is why natural metabolites tend to be more biodegradable and ‘eco-friendly’ than artificially-derived compounds. But a drug (whether natural or not) must have a target molecule of some kind, and the latter may in itself serve as the starting-point for evolved resistance. Thus there is evidence that β-lactamases, enzymes breaking down β-lactam antibiotics, themselves evolved from the original penicillin-binding protein targets of the antibiotics themselves, probably by gene duplication processes. Certainly there are many precedents for the development of microbial enzymatic activity against non-natural compounds. So in the end, resistance to natural vs. artificial molecules may be distinguished only by the speed of its onset.
To conclude, one can find an intersection of sorts between this post and as earlier one on astrobiology. How so? A completely alien biosphere would have its own ‘natural molecular space’, and there is no reason for assuming that it would have structural overlap with the natural molecular space with which we are familiar. Even if the alien biochemistry used proteins (perhaps an unwarranted assumption in itself), different amino acid alphabets might be used, and the range of protein folds might be radically different. The small molecules acting as drugs within such an alien ecosystem would therefore have no tendency to ‘fit’ our bio-world at all, although alien pharmacologists themselves would be well-advised to search their own specific biological backyard for molecules compatible with their biochemistry, mirroring the situation that we have found on Earth. The total set of alien biomolecules would then be no more beneficial for functional screening on our planet than a random chemical library– certainly an occasional promising hit would be found, but only through the vagaries of chance.
And on that general theme, a biopoly-verse comment:
Most people offer only shrugs
If questioned why nature yields drugs
Yet the secret is told
Via ways many-fold
In proteins, from humans to bugs.
References & Details
(In order of citation. In most cases, cited references are examples from numerous possible sources)
‘…..the navies and merchant marines of European countries were plagued …. by the terrible affliction of scurvy….’ For a good history of scurvy and its solution, see Bown, S. R. Scurvy : how a surgeon, a mariner and a gentleman solved the greatest medical mystery of the age of sail. (Viking Press, Melbourne, 2003).
‘…..all types of natural protein folding shapes can be grouped into a much smaller set……’ / ‘…….a small number compared with corresponding set of primary protein sequences….’ Protein ‘sequence space’ is hyperastronomically vast. Consider that even a chain of 100 amino acids (small by protein standards) can have 20100 combinations, given the 20 usual biological amino acids. But the numbers of sequences which are actually used in biology (capable of folding into specific structures) is a tiny subset of these. A fundamental protein subunit is called a domain, often defined as an autonomous folding unit which can be combined with other domains on a modular basis. Several thousand domain folding types are known, and from combinations of these at least 10,000 protein families have been defined. For basic information about protein domains and families, see Koehn & Carter 2005; Orengo & Thornton 2005.
‘….molecular volumes of most natural products are within the volume size range found within protein cavities…..’ See Koch et al. 2005.
‘……borne out by studies with artificial chemical libraries……’ Initial screening of artificial libraries (generated through combinatorial chemical methods were disappointing, since many primary hits lacked ‘drug-like’ character which would see them through later stages of screening. (for example, see Gribbon & Sewing 2005) Methods which integrate natural scaffolds and chemical diversification have been developed. See Krier et al. 2005; Bauer et al. 2010.
‘…….by binding to the estrogen nuclear receptor…..’ Other effects of estrogen mimics not operating via the conventional estrogen receptors have also been reported (See Rosselli et al. 2000). In particular, some act as inhibitors of protein kinases, which transfer phosphate groups to substrate proteins or peptides. The isoflavone phytoestrogen genistein (Shown in the above figure) is a tyrosine kinase inhibitor, along with other distinct functions in mammalian cells (See Dixon & Ferreira 2002).
‘Phytoestrogens fall into four main different structural groups…..’ See Cos et al. 2003.
Additional notes and references for above structural figure: ‘Castasterone is metabolized into the more potent brassinolide, which has a seven-membered ring…’ See Fujioka & Yokota 2002. ‘……..the protein structure of estrogen receptor-beta ligand binding domain in complex with genistein….The binding mode for genistein is similar but non-identical to that observed with 17-β-estradiol….’ See Pike et al. 1999. Source: Protein Data Bank (See Berman et al. 2003) 1QKM. Images generated with Protein Workshop (See Moreland et al. 2005).
‘……genistein’s binding is suboptimal for the induced structural changes necessary for activation of the [estrogen] receptor.’ See Pike et al. 1999.
‘……drugs such as cardiac glycosides…..’ The non-sugar (aglycone) cores of drugs such as digoxin (from the Digitalis [foxglove] plant) have the same perhydrocyclopentanophenanthrene ring system.
‘…….some plant steroid-like compounds…….have frequently been used as source material for beginning synthesis of steroidal drugs.’ See Heusler & Kalvoda 1996.
‘…..the reproductive systems of ruminant herbivores can be seriously affected by consumption of phytoestrogen-bearing plants….’ See Adams 1995 ‘……this has been proposed to have selective benefit to such plants by control of herbivore population size….’ See Wynne-Edwards 2001.
‘Evolutionary selection at group levels has been a problematic and contentious issue….’ For a discussion of this (among numerous other issues), see Richard Dawkins’ The Extended Phenotype; Oxford University Press 1982.
‘There is good evidence that at least isoflavonoids are involved in signaling between nitrogen-fixing bacteria ……and their leguminous plant hosts….’ See Peck et al. 2006.
‘….homologies between the two [NodD1 and estrogen receptors] cannot be ruled out until further structural information is available….’ See Fox 2004.
‘……have undergone ‘convergent evolution’ in response to estrogen-like molecules…..’ See Fox 2004. Convergence in evolutionary terms occurs when organisms lacking a recent common evolutionary origin independently acquire comparable phenotypes in response to similar environmental pressures. This can be observed at the morphological level, as (for example) the gross similarities between certain marsupial and placental mammalian carnivores (the extinct Tasmanian ‘tiger’ or thylacine and placental canids).
‘….‘inter-kingdom’ signaling has been proposed to extend beyond responses to pathogens alone….’ See the previous post for a brief discussion of quorum sensing and its extension to inter-kingdom communication. Also see Sperandio 2004; Shiner et al. 2005.
‘ Genistein is something of a champion in this regard [for multiple activities in animal cells]…..’ See Dixon & Ferreira 2002.
‘……the ‘selection’ for artificial insecticides was empirical searching for insect-killing abilities…..’ DDT was subsequently shown to be an insect neurotoxin by binding to the voltage-gated sodium channel in insect neurons. It appears that only a three amino acid residue difference in the human vs, insect sodium channel is the determinant of the differential toxicity of DDT to insects (See O’Reilly et al. 2006).
‘……xenoestrogens can also interact ….. with symbiotic nitrogen-fixing bacteria…….’ See Fox et al. 2004.
‘…..taste receptors are but one class among a large protein group called G Protein-Coupled Receptors, a huge target set for pharmacology…..’ For some information about human G Protein-Coupled Receptors (GPCRs), see Fredriksson et al. 2003.
‘……artificial chemical libraries yield ‘hits’ for modulation of biological processes…..’ See Golebiowski et al. 2001.
‘…….the venom of the Australian funnel web spider…….evolved for the purpose of quickly immobilizing its insect prey….’ See Tedford et al. 2004.
‘ Despite lethal effects of funnel web neurotoxins in humans ….. most other mammals are little bothered by them…..’ See Sheumack et al. 1991.
‘…..cytotoxic or cytostatic anti-tumor effects are explicable owing to the evolution of chemical defenses……’ See O’Hanlon 2006. It must be inferred that such marine organisms have self-protective mechanisms against their own toxins, in an analogous manner depicted in a Figure of a previous post. .
‘……the ‘pre-validation’ of such molecules as protein-binding agents….’ See Koehn & Carter 2005.
‘ Evolution towards resistance can thus arise by modifying promiscuous contacts…..’ For a discussion of the role of general protein promiscuity as an evolutionary opportunity, see Aharoni et al. 2005; Tokuriki & Tawfik 2009.
‘……why natural metabolites tend to be more biodegradable….’ See Saxena & Pandey 2001.
‘…….there is evidence that β-lactamases ….. evolved from the original penicillin-binding protein targets of the antibiotics……’ See Meroueh et al. 2003.
Next post: Two weeks from now.
Many large sets of physical or theoretical entities can be modeled as ‘spaces’, through mapping of values for their specific properties as multi-dimensional arrays. Such modeling is useful to examine networks of similarities and inter-relationships between members of the comprehensive set of interest, and ‘chemical space’ for small molecules is one such conception. When we make use of natural products for specific purposes, we are in effect sampling from a local natural molecular space, which in turn is a tiny subset of an enormous space of all chemically possible molecules. This ‘local’ set of molecules is but a tiny corner of the larger universe, but the special conditions on Earth where life flourishes render its molecular endowment vastly more complex than what has been found in the universe elsewhere to date. And this will remain the case until truly independent alien life is identified, if it indeed exists. (See an earlier post on the subject of astrobiology.)
The previous post looked at ‘chrestomolecules’, or molecules classifiable as economically useful in some way or another, irrespective of their sizes or origins. The present post will primarily focus on small biologically-derived molecules (less than ~3500 Daltons, based on the largest known naturally-made examples of this type. As we will see at the end, though, a size-based distinction can become somewhat arbitrary when certain classification criteria are applied.
The Nature of Biomolecules on Earth
At this point we should consider in more detail some special features of natural molecular space as a whole. It is possible to classify all biomolecules by many schemes, including their chemical natures (protein, lipid, carbohydrate, nucleic acid, and so on). One way is to sort them by function, and two very broad functional categories are molecules which are required (directly or indirectly) for all biochemical activities of an organism relating to normal survival and growth, and those involved with some form of interaction with the external environment. The latter products constitute the subject material for the burgeoning fields of ‘molecular ecology’ and ‘ecological biochemistry’, and most natural products with medicinal value to human beings fall within this territory. The great majority of biomolecules employed by humans (especially until recent times) are of relatively low molecular weight and (as noted previously) often categorized as secondary metabolites (products of metabolism). Although still widely used, the term ‘secondary metabolites’ (as distinct from the products of ‘primary’ metabolism) does not meet with universal approbation, and has ‘no simple one-line definition’. Traditionally, secondary metabolites are products of certain plants, fungi and bacteria, but it is clear that biomolecules of demonstrated or potential medicinal use are found in a range of both vertebrate and invertebrate animals as well.
A fundamental aspect of these compounds embodied in the ‘secondary’ tag is that they are not essential for the growth and survival of the organism per se, and are found in very variable levels between different taxonomic groups. When present, however, secondary metabolites are synthesized from universally available biological precursor molecules, by specialized enzyme protein molecules, often in a series of sequential steps. Each stage in such a ‘biosynthetic pathway’ involves a separate enzyme operating on successively modified metabolic products until the final version(s) are made. The functions of secondary metabolites have been controversial, but they have usually been presumed to provide a selective advantage under some conditions, acting as toxins against predators, competitors, or parasites. The evolutionary origins of secondary metabolites have also been a long-standing source of debate, especially given that their specific functions are often poorly defined. Organisms producing these compounds are generally associated with competitive environments which may promote rapid diversification in response to strong selective pressures.
Notes for Figure: Utility of specific host modification for allowing broad-spectrum inhibition of targets of competing organisms. Following target molecule A modification (by an enzyme present in organism A but not competing organisms), a secondary metabolite binding to a conserved region can be produced, and will bind to the conserved sites of homologous targets X, Y and Z.
Microbial populations (especially among soil organisms) may show great diversity in secondary metabolite production, but gene homologies between different organisms suggest common origins for some biosynthetic enzymes, probably by gene transfer. Gene duplication and divergence is acknowledged as a probable important driver of diversity in the enzymatic machinery of secondary metabolite production. Genes encoding enzymes which control the production of antibiotic secondary metabolites tend to cluster and to be selected as a group. A good example of this is biosynthesis of the antibiotic erythromycin in the bacterial organism Saccharopolyspora erythrea, which involves a complex series of reactions mediated by many proteins all encoded within a single large contiguous segment of its genomic DNA. In order to avoid self-toxicity, this organism also has a gene which confers resistance to erythromycin itself. (This works by the action of the gene’s protein product, which specifically modifies the organism’s ribosomal RNA where protein synthesis occurs, such that it is not susceptible to the antibiotic. The general principle of this is depicted in the above Figure). This situation is reminiscent of bacterial restriction-modification systems, where enzymes targeting specific DNA sequences are produced in order to destroy incoming foreign viral or other DNAs, with host DNAs protected by a specific chemical modification (usually methylation). Restriction-modification enzymes of bacteria confer an ‘immunity’ to invading DNA, and secondary metabolite production has also been compared (at least in some respects) with an immune system against competitors.
One observation which must be reconciled with any evolutionary hypothesis is that the diversity of the metabolites (the ‘metabolome’) produced by specific organisms appears at first glance to exceed their genetic capacity to encode the relevant diversity of biosynthetic enzymes. With the increasing number of microbial and other genomes which have been completely sequenced this issue has become well-defined if not comprehensively resolved. The metabolome / genome size discrepancy has been attributed to relatively low substrate or catalytic specificities (‘catalytic promiscuity’) of the enzymes involved in the pathways of metabolite biosynthesis. This means simply that the product molecules resulting from the enzymes’ actions are not precisely defined, although in practice they will usually be related molecules. Low substrate specificity indicates that an enzyme is less choosy about the starting molecule that it will act on (its ‘substrate’) than in the case of a high-specificity enzyme. Low catalytic specificity refers to the range of enzyme action itself. A biosynthetic enzyme with low catalytic specificity might be capable of transferring multiple types of chemical groups to its substrate molecule(s), or might have the ability to transfer the same chemical group to different sites on such substrate(s). The end result is that a limited number of enzymes with reduced specificity can synthesize many different metabolite molecules. Although these products may be closely related chemically, small differences in chemical structure can radically change their biological effectiveness. Also, provision of alternative precursor substrates or alteration of culture conditions can induce dramatic changes in the types of product yielded from the same organism (‘one strain, many compounds’). In fact, deliberate inhibition of specific enzymes in a multi-enzymatic pathway can be a useful route towards artificial end-product modification.
Many secondary metabolites have shown at best only moderate measurable activity in their presumptive biochemical roles. An interesting hypothesis accounts for the diversity of apparently sub-optimal metabolites from a single organism by postulating that a selective advantage is conferred by maintaining such diversity, provided it is accomplished with minimal metabolic cost. This is based upon the supposition that the likelihood of inhibiting a competing organism (with a multitude of potential target molecules) is maximized by producing a wide range of compounds, some of which may have moderate but useful affinity for a target. In contrast, a high-affinity interaction between a metabolite ligand and a protein is a relatively rare event. One problem with this model is that failure to pinpoint a selectable (evolutionary fitness-conferring) function of any specific metabolite of an organism does not prove that such a function is absent. This is especially so given the complexity of environments in which secondary metabolites are typically formed, and the difficulty of screening all possible competing organisms. (Only a quite small component of all soil organisms appear to be amenable to in vitro cultivation at present). Also, it has been suggested that secondary metabolites may have different effects at low concentrations (for example, by altering behavior of competitors rather than a direct toxicity), calling for more ‘ecologically relevant’ assays. Moreover, it’s not all about competition. In relatively recent times, the importance of bacterial ‘quorum sensing’ has been increasingly appreciated, and intensively investigated. In this ‘social’ effect, some aspects of gene regulation of specific bacterial species are controlled by their cell densities, through the action of secreted ‘autoinducers’. Chemical ‘cross-talk’ between different bacterial species has also been implicated, and even between bacteria and eukaroytic organisms (‘cross-kingdom’ communication). And at least in some circumstances, synergism between different bacterial species in the production of secondary metabolites has been documented. So the roles of metabolites produced by highly diverse bacterial communities may be very diverse as well.
While acknowledging these complexities, competition is still a central ecological principle. Many competitors of any given organism may be closely related, with associated conservation of many potential target proteins. In fact, a great many targets which could be potentially bound by a metabolite will be shared by the host organism itself, especially at highly conserved active sites such as those in essential enzymes. So any metabolites modulating the activities of such shared conserved target proteins will be selected against (an organism producing a compound which is even slightly toxic to itself will be at a strong competitive disadvantage and rapidly eliminated). One way around this problem, as we have seen with erythromycin-producing organisms or the conceptually related restriction-modification systems, is for the host to specifically modify itself to avoid self-damage (as in the above figure). In this case, choosing a highly conserved target (protein synthesis in the case of erythromycin) is beneficial by maximizing the sweep of the counter-competitor response. An alternative solution with the same outcome is for the host to synthesize a third-party gene product which neutralizes the activity of the toxic bioproduct only while it is internalized, leaving it free to inhibit conserved targets in the environment. For example, some snake species appear to cope with their own venom production in this manner. The problem of avoiding host damage while targeting a pathogen (or distinguishing self from not-self) is also fundamental to the vertebrate immune system, which has evolved sophisticated (albeit not universally successful) mechanisms for eliminating or suppressing self-reactive immune responses. There is evidence that selection can positively drive the co-production of even structurally unrelated secondary metabolites, by means of synergy between the compounds at the functional level. Thus, chemically unrelated clavulanic acid (an inhibitor of β-lactamase enzymes which break down penicillins and related antibiotics) and β-lactam (penicillin-family) antibiotics themselves are co-produced with a frequency and pattern highly suggestive of strong selective pressures rather than chance.
In a complex environment, the dynamics of interactions between mutually competing organisms will also result in complex selective pressures on secondary metabolite production, with the potential for ‘arms races’ between offensive strategies and defensive countermeasures. An example of the latter can be found in adaptations of herbivorous insects for circumventing the action of repellent or toxic plant metabolites. In microbial populations, acquisition of resistance to counteract the effects of toxic secondary metabolites can be via mutation of target proteins themselves or through the modifying activity of another gene product, acquired through evolution ‘from scratch’ or (more commonly) lateral gene transfer from a different organism. Such resistance genes can operate in a variety of ways, including direct destruction or chemical modification of a secondary metabolite to render it inactive (as with the b-lactamases we noted above), prevention of the uptake of metabolites into cells or active export of them, or modification of a cellular target of the toxic metabolite. By way of example of the latter, the gene protecting erythromycin-synthesizing bacteria from self-toxicity will also confer erythromycin resistance if expressed alone in otherwise sensitive organisms.
We noted above the comparison of defensive / offensive secondary metabolite production with an ‘immune system’. Generation of ‘response diversity’ by degenerate syntheses of metabolites with promiscuous enzymes has some parallels with innate immune systems of invertebrates, although not with adaptive immune responses which have elaborate systems for selecting, amplifying and fine-tuning molecular solutions for recognition of a foreign antigen. A specific secondary metabolite can only be fine-tuned towards optimal target recognition by repeated rounds of selection for organisms with improved fitness as a result of suppressing competitors or predators. Where inhibition of a specific single target confers a strong fitness advantage, successive generations will be selected towards production of compounds with increasingly improved target-interactive properties. As we have seen, however, in complex environments selection may also tend to favor production of a wider range of structurally-related compounds.
The Molecular Hardware of Ecology
Since we have noted that natural product therapeutic molecules generally fall within the category of bioproducts with an ecologically-based function, a different and more general term than secondary metabolites would be useful. ‘Ecobiomolecules’ is a reasonably self-descriptive word (an abbreviation of ‘ecological biomolecules’) which is defined in this context as any biological molecule whose function involves direct interaction with molecules deriving from an organism’s environment. This term therefore encompasses all secondary metabolites with known functions, but also many more biomolecules beyond the generally-accepted secondary metabolite scope. (Here we use the definition of ‘ecology’ as the interactions of an organism with other organisms and the environment in general).
If we assume that all secondary metabolites originate from various environmental selective pressures, then it is automatically true that all secondary metabolites are ecobiomolecules, but not vice versa. (It could be pointed out that all biomolecules (or more accurately, the genes that directly encode them, or genes which encode the enzymes which synthesize them) can be considered as molded by the environment over time in response to selective pressures, to maximize the fitness of the lineage of organisms in which they are expressed. So when referring to ecobiomolecules, we can further restrict the definition to biological molecules whose major normal function involves an interaction with a physical or chemical environmental factor. Within the total ecobiomolecular set, we can find molecules whose function is concerned with sexual or other forms of communication (such as pheromones or molecules mediating the above-mentioned phenomenon of quorum-sensing), defense (plant products selected to reduce herbivore activity; anti-bacterial products; sprayed or ejected chemical deterrence as with skunks, etc.), or offense (various animal toxins). A vast array of arthropod pheromones and defense / offence toxins are included here. Indeed, the success of arthropods among all animal groups (‘phyletic dominance’) has been attributed at least in part to their virtuosity in the production of such molecules, with associated improved levels of survival. In some cases, rather than synthesizing useful compounds themselves, insects can acquire a useful property (such as acquired toxicity towards predators) from ingesting host plant material. This phenomenon of ‘chemical sequestration’ involves the evolution of specific transport systems for host plant toxins.
We can see that an organism which possesses a toxic phenotype, whether through an endogenous toxin or an actively acquired one, gains a survival advantage if the presence of the toxin is a deterrent to predators. As a modulator of the interaction between two or more organisms, the associated toxin can be thus termed an ecobiomolecule, and its presence confers an ‘immunity’ of sorts. But what of the intricate immune systems of higher organisms? It would seem that we must technically also classify all components of both the innate and adaptive vertebrate immune systems as ecobiomolecules, as they have evolved to cope with the environmental pressures of a wide variety of pathogenic organisms, and are not intrinsically required for growth and metabolic survival. From this stance, the size of any participating molecule is irrelevant (Antibodies, for example, are large multimeric proteins). It should be noted, though, that some functions of immune systems are concerned with internal homeostasis (such as surveillance against tumors), so not all aspects of immune systems can be viewed as ecologically based. But antibodies in particular have been extremely useful to humans in both basic research and therapeutic applications, and we therefore should group them in the broad category of ecobiomolecules which also includes small molecule secondary metabolites.
To conclude with another glance at chemical space, through a biopoly-verse lens:
Consider a figurative notion:
The Earth as a molecular ocean
- Yet still but a trace
Of chemical space
Whose vastness might inspire emotion.
…But we’ll see in subsequent posts that as far as the human usefulness of natural product molecules on Earth is concerned, it is very definitely a case of quality over quantity (even though the quantity in practical terms is still a huge and largely untapped resource, whatever its size with respect to the general potential of chemical space).
References & Details
(In order of citation. In most cases, only sample references are provided from many existing in the field.)
‘…….modeled as ‘spaces’…… and ‘chemical space’ for small molecules is one such conception.’ See Lipinski & Hopkins 2004; Rojas-Ruiz et al. 2011. Estimates of the number of chemically possible members of small-molecule chemical space have been made, with figures of up to 10200 compounds postulated (See Fink et al. 2005).
‘…….the largest known naturally-made ‘small’ molecules…’ Currently the marine bioproduct maitotoxin (molecular weight 3422) holds this record. (See Nicolaou et al. 2008). Such molecules are ‘non-alphabetic’ in the sense that they are sequentially built by a series of enzymatic reactions, and are not macromolecules such as proteins or nucleic acids composed of specific sequences of distinct subunit ‘alphabets’. Neither are they long-chain carbohydrates built by enzymatic linking of repetitive monosaccharides.
‘…..the term ‘secondary metabolites’ (as distinct from the products of ‘primary’ metabolism) does not meet with universal approbation….’ See Firn & Jones 2009; ‘……and has ‘no simple one-line definition’ See Challis & Hopwood 2003.
‘…….they [secondary metabolites] are found in very variable levels between different taxonomic groups…’ See Wink 2003.
‘………functions of secondary metabolites have been controversial, but they have usually been presumed to provide a selective advantage under some conditions….’ See Cavalier-Smith 1992; Maplestone et al., 1992.
‘………gene homologies between different organisms suggest common origins for some biosynthetic enzymes, probably by gene transfer. ’ See Vining 1992.
‘…..Gene duplication and divergence is acknowledged as a probable important driver of diversity in the enzymatic machinery of secondary metabolite production. ‘ See Cavalier-Smith 1992; Kliebenstein et al. 2001.
‘…….biosynthesis of the antibiotic erythromycin….’ See Donadio et al., 1991.
‘……secondary metabolite production has also been compared ….with an immune system against competitors…’ See Firn & Jones 2003.
‘…..provision of alternative precursor substrates or alteration of culture conditions can induce dramatic changes in the types of product…. a useful route towards artificial end-product modification…’ See Bode et al. 2002.
‘An interesting hypothesis accounts for the diversity of apparently sub-optimal metabolites….’ See Firn & Jones 2003.
‘Only a quite small component of all soil organisms appear to be amenable to in vitro cultivation at present. ’ This has been partially overcome by metagenomic techniques, discussed in a previous post’s References & Details.
‘…….more ‘ecologically relevant’ assays…’ See Engel et al. 2002.
‘Chemical ‘cross-talk’ …….even between bacteria and eukaroytic organisms……’ See Williams 2007.
‘…….synergism between different bacterial species in the production of secondary metabolites has been documented.’ See Angell et al. 2006.
‘……some snake species appear to cope with their own venom production in this manner….’ See Smith et al. 2000.
‘There is evidence that selection can positively drive the co-production of even structurally unrelated secondary metabolites….’ See Challis & Hopwood 2003.
‘…..adaptations of herbivorous insects for circumventing the action of repellent or toxic plant metabolites…’ See Wittstock et al. 2004.
‘……the gene protecting erythromycin-synthesizing bacteria from self-toxicity will also confer erythromycin resistance if expressed alone in otherwise sensitive organisms…’ See Teuber et al. 2003.
‘……the success of arthropods …….has been attributed at least in part to their virtuosity in the production of such molecules…’ See Meinwald & Eisner 1995.
‘…….insects can acquire a useful property …..from ingesting host plant material….’ See Termonia et al. 2002.
‘…..phenomenon of ‘chemical sequestration’ involves the evolution of specific transport systems for host plant toxins……’ See Kuhn et al. 2004.