Skip to content

Bioorthogonal Chemistry and Recognition

February 24, 2013

This post further extends some themes raised in previous posts, concerning xenorecognition  and its corresponding biological repertoires in relation to bioorthogonality (first noted  in the context of the possibility of ‘weird life’ in a shadow biosphere on Earth, and further considered later  in the context of the recognition of xenobiotics). Previously, the concept of true biological orthogonality was posed as a hypothetical state of complete ‘invisibility’ of a chemical entity and a biological system with respect to each other. Here, we are concerned with an artificial application of growing importance, where orthogonality refers to the specificity of chemical reactions of interest. This particular topic, noted briefly in earlier posts (see 30 May 2012), is thus further developed as the theme of this post. At the same time, it should be noted that it is not the intention to provide extensive chemical detail, for which many excellent sources are already available. Primarily, this post provides an overview with some additional views which are relevant to xenobiotic molecular recognition and the general concept of orthogonality in biology. .

Chemistry and Life

Curious parallels and contrasts can be made between the modern concept of chemical bioorthogonality and non-scientific (pre-molecular) notions regarding the nature of life. In early times it seemed self-evident that living things were so fundamentally different from non-living matter that life must be constituted of radically different stuff, whatever that might be. This supposition was frequently interlinked with the doctrine of vitalism, or the need for a ‘vital spark’ to animate anything that can be called ‘alive’.  Nevertheless, the old belief in the ‘otherness’ of life would result in ‘bioorthogonality’ (in a literal sense) following on as a natural corollary. This statement carries the assumption that the supposed living stuff would not only be distinct from non-living matter, but also would remain aloof from it. By this reasoning, any conventional inorganic reaction (insofar as they were known in earlier times) would fail to impact on a living organism.

But this viewpoint could be shot down very rapidly even in the earliest of times, with just a little thought. A great many natural substances can burn, damage or kill a living organism, ranging from acids and alkalis to toxic metals. For example, natural deposits of arsenic or mercury ores could easily be prepared and administered as poisons – the point  being that such deadly agents come from the non-living world,  irrespective of their chemical natures. Therefore, the lifeless material world can obviously impact dramatically on ‘life-stuff’, and ‘extreme bioorthogonality’ (the belief in the complete independence of life from non-living matter) could never have been a tenable proposition. (This might seem obvious as well from the need take in water and air, but these facts might be more easily rationalized away than the dramatic action of inorganic poisons).

But having conceded this, to many early thinkers it still seemed ‘obvious’ that living things, though influenced by the non-living world,  could not be made of the same stuff. It was well known from observing the effects of combustion that carbon was present in anything alive, and as a primordial discipline of chemistry gradually emerged from nonscientific alchemy, it became clear that many pure materials could be obtained from specific living sources. These were accordingly viewed as ‘organic’ in the literal sense, but such products were held to be special by virtue of their origins within things that (at least before their demise) possessed the mysterious property of life. Clearly it was possible to influence a living organism with the products of another. Any number of natural medicines or poisons from (mainly) plants and (less commonly) animals could easily make this point. Yet a pre-modern ‘bio-philosopher’ could simply claim that all such products were made by life, and thus possessed the enigmatic ‘organic’ quality which set them apart from everything else. True, the above-noted action of many never-living materials proved that inorganic matter could modulate life, but apparently in a crude and battering-ram sort of way.

This mental picture of organic material as partitioned from everything else received its first major challenge when the German chemist Friedrich Wöhler prepared urea in the laboratory in 1828, without any recourse to the urine of animals. (Urea, of course, derives its name from the excretory source from which it was first isolated, and it therefore stood as a classic organic substance ‘impossible’ to synthesize, at least until Wöhler showed otherwise). Since that time, the science of organic chemistry has seen a series of increasingly intricate synthetic triumphs, where all manner of complex natural compounds have yielded to the ingenuity of synthetic chemists. Obviously, all notions of the ‘special’ nature of organic compounds and their dependence on a ‘life-force’ have long since been consigned to the proverbial trash-can of history. We retain, though, an echo of this through the retention of the word ‘organic’ in reference to covalent carbon compounds in general, whether or not they derive from living biosystems.

The demolition of the special status of the chemistry of life, and its integration with chemistry in general, might seem at first glance to remove the possibility of ‘orthogonality’ from biology.  If life is, in essence, nothing more than souped-up chemistry, how could it stand aside in any aspect from general chemical reactivity? Would not the incredible complexity of living organisms demand an equally complex repertoire of chemical reactions? And if the latter was true, would not virtually any chemical reaction engineered by humans already have its biological counterpart? Clearly, were that to be the case, then the concept of a ‘bioorthogonal’ chemical reaction would be illusory, since any reaction component would be able to find a biological target with which to react. Nevertheless, this supposition is emphatically not the case at all; it is demonstrably incorrect.

It is undeniable and obvious fact that life is a phenomenon of consummate complexity (often described as the most complex known system in the universe, with the emergence of  intelligence at its apex). Yet while an enormous number of distinct chemical transactions exist across the biosphere, the vast majority of these are catalyzed by protein enzymes composed of a relatively small ‘alphabet’ of amino acids (usually the canonical 20) linked together by peptide bonds in highly specific sequences. Many such reactions have an absolute requirement for additional cofactors as well as the relevant protein enzymes, in the way of metal ions, small organic compounds (vitamins) or metal-sulfur clusters. Even so, a ‘take home message’ from biology is that incredible complexity can emerge from combinations and permutations of a relatively small set of building blocks.

A corollary of this biological insight is that it should not be surprising to find that there are many precedents for chemical reactions which have no known biological parallels, although an important caveat here is the need to add the qualifier “under normal physiological conditions”. It is obvious that chemistry only operating under conditions of temperature and pressure far removed from those tolerated by life cannot be applicable within living systems. Nevertheless, even after this restriction is enforced, chemical candidates for non-biological reaction status still remain. A pair of such interacting chemical groups could thus potentially ‘only have eyes for each other’, in a metaphorical sense, as long as they did not react with the plethora of functional groups commonly encountered in biological systems. This notion is shown schematically in the Figure below:

Biooorthog1

Figure 1: Depiction of a bioorthogonal reaction pair, in the presence of examples of common biological functional groups (not exhaustive), which do not interact with the participating non-biological groups. Geometric shapes represent any biological molecules to which the bioorthogonal groups have been appended in vitro, prior to introduction to the intracellular environment.

_______________________________________________________________________

Many applications of bioorthogonal reactions have been envisaged, such as specific labeling of biocomponents and tracing of the intracellular traffic of specific molecules. But for such systems to work in practice, bioorthogonality of the reactions is necessary but not sufficient: the reactions must also be highly efficient under intracellular conditions. The required reaction properties of high yield and product specificity are satisfied by a class of reactions termed ‘click chemistry’, envisaged as ‘spring loaded’ to rapidly and efficiently proceed down a specific reaction pathway. Some such reactions require catalysis to operate under normal ambient conditions, but this need can be obviated within certain molecular arrangements. A classic case of this is the reaction between alkynes (bearing a carbon-carbon triple bond) and the dipolar azide group. While this reaction involving a linear alkyne group requires a cuprous ion (Cu[1]) catalyst to work well at normal temperatures and pressures, if the triple bond is placed under strain within a cyclic structure (an eight-membered cyclooctyne ring) then the reaction is greatly accelerated. These observations are presented in the following Figure:

BiooorthogRxn

Figure 2: Alkyne-azide bioorthogonal reaction pairs, for (A) both linear-alkynes (Cu[1]-catalyzed) or (B) strained-alkyne octyl ring (uncatalyzed).

_______________________________________________________________________

But the rate at which any intermolecular chemical reactions proceed is necessarily driven by the respective concentrations of reactants. Within the intracellular environment, the total concentration of solutes (all macromolecules and small molecules included) is very high, and at relatively low concentrations, even highly self-reactive and bioorthogonal groups may have slow kinetics of interaction. A useful approach which solves this issue is to engineer the reaction process such that a pair of bioorthogonal participants are brought into close spatial proximities with respect to each other, as a direct consequence of the targeting strategy. This is illustrated in the Figure below:

BiooorthogProxim

Figure 3: Bioorthogonal click reaction engendered through spatial proximity produced through the targeting choice. Two related proteins (1 and 2) are depicted with two proximal sites (in principal, either for natural ligands, or sites for which artificial ligands can be designed). If two such ligands equipped with mutually interactive bioorthogonal click groups are introduced into the intracellular environment, then only Protein 1 can bind the correct pair such that desired labeling product is formed.

_______________________________________________________________________

Spatial proximity for promotion of specific reactions can be engineered in other ways, most notably via nucleic acid templating, a very large field beyond the scope of this post. But having given this very brief taste of how bioorthogonal reactions are practically engineered, let us move on to some general issues of the nature of chemical bioorthogonality.

Chemically Possible Bio-reactions

Chemical reactions that are ascribed as bioorthogonal could in principle be compromised in two broad ways: by the discovery of biocatalysts capable of accelerating the reaction between the ‘orthogonal’ pair of chemical groups, or capable of promoting reactivity between either (or both) of the foreign functional groups and one or more host molecules. Either possibility would require recognition by an enzyme of at least one of the ‘foreign’ functional groups.

Here it is appropriate to reiterate a point noted in a previous post, to the effect that orthogonality in biological systems is a relative concept. This referred to the observation that while a xenobiotic compound might be completely orthogonal within an isolated biological system, that status certainly is not necessarily the case within the biosphere as a whole. With respect to the potential biorthogonality of chemical reactions, a particular contrast can be drawn between the relatively limited repertoire of most eukaryotes in comparison with the astonishing catalytic diversity of the prokaryotic world. A good example of this is the observation that the complex core structure of Vitamin B12, an essential nutritional cofactor for many (although not all) multicellular eukaryotes, is made solely by prokaryotes.

If a pair of inter-reacting functional groups attained ‘pan-biospheric bioorthogonality’, then all known biological systems would necessarily lack this kind of chemical functionality. But the absence of a reaction from biology certainly does not mean that it could never exist. A case in point also noted in the same previous post  has been the apparent extremely rapid evolution of the bacterial enzyme phosphotriesterase. In fact, such apparent large jumps in an imaginary ‘catalytic space’ are not necessarily difficult in terms of the parallel evolutionary moves within protein sequences. A recent demonstration of this comes from the artificial adaptation of a specific cytochrome P450 enzyme (one of a large class involved with xenobiotic recognition and modification) from its normal catalytic transfer mechanism into one involving a hitherto non-biological process. The adapted cytochromes in this case were selected from a relatively small set of artificial enzyme variants.

Another useful precedent is that of the catalysis of the Diels-Alder reaction, a chemical cycloaddition of great importance in synthetic chemistry. By rational design and directed evolutionary approaches, both protein enzymes (catalytic antibodies) and ribozymes have been derived which are capable of catalyzing this reaction, albeit not with very high efficiency. Yet although there is suggestive evidence from certain biosynthetic pathways, there is as at present no definitive evidence that a natural precedent exists for this particular catalytic activity. While this is a significant evolutionary question, the artificial generation of the specific catalytic activity in diverse macromolecular frameworks already shows its inherent biological feasibility.

So, with these precedents in mind, the assignation of a reaction process as ‘bioorthogonal’ should always be defined with specific limits of space and time. Spatially, in the sense that the orthogonality of a reactive pair can be assigned for a specific cell type, whole organism, phylogenetic group, or even the entire existing biosphere (‘pan-bioorthogonality’ once more). Temporally, in the sense that a ‘hole’ in biological rendition of a chemical reaction may be plugged through evolution (either natural or artificial), where the required time can be surprisingly short in at least some instances. Of course, the evolution of other challenges might require more fundamental protein adaptation. A specific example can be found from the above Figure 2A (showing the Cu(1) catalyzed linear alkyne-azide cycloaddition reaction), if an enzyme evolved which could accomplish both catalysis (possibly via a copper ion itself bound near the reactive site, or through some other cofactor) and engendering spatial proximity (through binding of reactants within an active-site pocket). This is portrayed below:

 

AlkyneAzideCat

Figure 4: Schematic depiction of a hypothetical protein catalyst for an alkyne-azide cycloaddition reaction.

_______________________________________________________________________

In fact, the alkyne (‘acetylenic’) triple bond is not at all absent from biological systems, being found in metabolic products of diverse organisms, ranging from prokaryotes to plants and certain animal lineages. Organic azides, on the other hand, are believed to be totally absent from known biosystems. Thus, although the hypothetical catalysis of Fig. 4 might be artificially engineered, in the absence of biological azides, its prospects for arising naturally would be remote. So any theoretical threat to the bioorthogonality of such reactions as a pair is most unlikely to be realized by natural agency. In principle, though, if organic azides were artificially introduced into specific cells synthesizing certain natural products bearing strained acetylenic bonds in ring structures (members of the enediyne natural product family), then an azide-associated reaction with the enediyne product could occur. In such circumstances, the resulting cycloaddition reaction, deriving from a natural host product and a non-biological reactant, would lack complete orthogonality.

Levels of Orthogonality

It should be kept in mind that the above considerations regarding bioorthogonality are at the level of chemical reactions, which is a distinct concept from biorthogonality at the level of molecular recognition of initial reactive groups or the products by xenobiotic recognition systems. Certainly these two categories have regions of overlap, as when a natural enzyme must bind (show molecular recognition for) a substrate as well as enact a catalytic change upon it. But binding per se is not necessarily linked with catalysis, as demonstrated by a vast array of ligand / receptor interactions, as well as antibodies and their cognate antigens. It was in fact noted in a previous post  that the physiological action of some xenobiotics arises from non-covalent interactions with normal host receptors, with concomitant aberrant signaling induction.

Nevertheless, in terms of recognition of xenobiotics by natural defense mechanisms, binding and catalysis are usually linked, since catalytic modification of foreign compounds is an important part of detoxification processes (as discussed in another previous post. In any case, if we return to the example of Fig. 2, then assignment of the reaction as essentially bioorthogonal is acknowledged (with the above caveat of rare possible acetylenic host product interactions). Yet orthogonality may not necessarily exist at the level of recognition of either the reactants or the product(s) in terms of xenobiotic recognition, even in the same host organism. And at both reactant and product levels, recognition precedents exist for the components within Fig. 2.

In some organisms, alkyne groups are produced and recognized through biosynthetic machinery, as noted above for metabolites such as the enediynes. But more generally, substituted alkynes are clearly interactive with specific cytochrome P450 enzymes (important for xenobiotic processing, also noted above). In fact, certain compounds with carbon-carbon triple bonds have been well-characterized as P450 inhibitors, through the formation of covalent adducts. The action of a P450 enzyme on a simple substituted alkyne, propargyl alcohol, has been found to be responsible for converting it into the toxic propiolaldehyde, another example of the self-activation of a deleterious effect of a xenobiotic (this general phenomenon was discussed by way of an analogy to autoimmunity in a previous post).

While the azide functional group appears to be absent from biology, it can certainly have profound physiological consequences when artificially introduced as a moiety in certain molecular contexts. Probably the best-known example of this is the compound AZT (3’-azido-3’deoxythymidine), formerly used as a mainstay of treatment for HIV patients, although now superseded by more effective drugs. A major limitation of AZT was its tendency to produce often serious side-effects in certain patients. While ‘off-target’ activities are a bane to patients and clinicians, for the present purposes they help make the point that azido-compounds are far from orthogonal when introduced into complex (human) biosystems. Side-effects do not arise randomly; they result from unwanted interactions of drugs with otherwise-irrelevant specific host structures. Such interactions themselves surely provide a back-up demonstration of ‘non-bioorthogonality’ at the level of molecular recognition.

In addition, AZT is known to be recognized and metabolized by P450 enzymes. The final component of the reactions of Fig. 2, the 1,2,3-triazole group, is likewise recognized in a variety of biological contexts, including P450 enzymes as well. The 1.2.3-triazole isomer is also a central moiety of a variety of drug candidates for a wide range of potential targets.

All of these pieces of information are proffered as evidence that each functional group involved as reactants or products in a well-known bioorthogonal reaction are not at all orthogonal at the level of molecular recognition within an entire complex multicellular organism (where xenobiotic processing enzymes are expressed), as opposed to an isolated single cellular system. But the mutual interaction of bioorthogonal chemical components (such as azides and alkyne groups) and their avoidance of interaction with host functional groups in the vast majority of cases classifies the chemical reaction itself as bioorthogonal in almost all useful contexts.

The correlation of reactivity per se and orthogonality was considered in a previous post, where the remarkable fact was noted that highly unreactive compounds (such as saturated fluorocarbons) and even inert gases cannot be considered to be bioorthogonal, through their potent effects in inducing anesthesia. Therefore, even the complete preclusion of possible covalent bond formation does not necessarily exclude a chemical agent from the capability of strongly perturbing a biological system, the complete antithesis of a generalized notion of bioorthogonality. At the level of designing bioorthogonal chemical reactions, on the other hand, covalent bond formation is of central and defining importance. And new bioorthogonal reactions are being actively sought and found.

To finish up, a biopoly(verse) comment on an ideal biorthogonal reactive situation:

Make a molecule to enter a cell

Yet trigger no metaphorical bell

(Meaning bio-inert –

And thus causing no hurt)

While reacting with designed targets well.

References & Details

(In order of citation, giving some key references where appropriate, but not an exhaustive coverage of the literature).

‘…….it is not the intention to provide extensive chemical detail, for which many excellent sources are already available.’    Some reviews from the lab of a major pioneer of bioorthogonal chemistry (and related areas of click chemistry), Carolyn Bertozzi, are very useful. See Sletten & Bertozzi 2009; Jewett & Bertozzi 2010.  For another good recent review, see Best 2009.

‘…..Friedrich Wöhler prepared urea in the laboratory in 1828…..’    This occurred serendipitously while he was attempting to make ammonium cyanate (the ammonium salt of cyanic acid), which can undergo a conversion rearrangement from an ionic salt to a simple covalent organic compound:

Urea

A good source of further information is the relevant site from the Chemical Heritage Foundation. Here, a quote from Wöhler states,” the ammonium salt of cyanic acid is urea”. Of course, this is not literally correct as such, since the compounds have the same atomic compositions but very distinct structures and bonding states. But this discovery helped to frame the concept of chemical isomers, as coined by the Swedish chemist Berzelius, a contemporary of Wöhler.

‘………chemistry only operating under conditions of temperature and pressure far removed from those tolerated by life cannot be applicable within living systems.’     An example of this with a biological reply is the artificial Haber process for the reaction of atmospheric nitrogen with hydrogen to form ammonia, which requires both catalysts and high temperatures. But natural evolution has found a way to ‘fix’ nitrogen using specific enzymes (nitrogenases) and essential metal (typically iron and molybdenum, but sometimes vanadium) – sulfur clusters (See Hu and Ribbe 2011 for a recent review). For details on the Haber process itself, see Modak 2002.

‘………a class of reactions termed ‘click chemistry’…..’      For reviews on this topic, See Kolb et al. 2001; Best 2009.

‘……if the triple bond is placed under strain within a cyclic structure (an eight-membered cyclooctyne ring)……’  The smaller the ring group, the more strain that is placed on the normally linear alkyne bond, and in fact an eight-membered ring is the smallest that is stable while retaining a carbon-carbon triple bond under normal conditions.

Spatial proximity for promotion of specific reactions can be engineered in other ways, most notably via nucleic acid templating…..’      For a review, see Li & Liu 2004.

‘……the complex core structure of Vitamin B12, an essential nutritional cofactor for many (although not all) multicellular eukaryotes, is made solely by prokaryotes.’      And this also demonstrates but one aspect of the dependence of ‘higher’ multicellular organisms upon the activities of prokaryotes. For details on Vitamin B12 biosynthesis, see Martens et al. 2002.

‘…..the artificial adaptation of a specific cytochrome P450 enzyme ………. from its normal catalytic transfer mechanism into one involving a hitherto non-biological process.’     For the recent paper, see Coelho et al. 2013. Their work resulted in the normal oxene transfer (equivalent of atomic oxygen) mechanism of the P450 enzyme of interest converted into an isoelectronic (but non-biological) carbene (equivalent to a substituted methylene) transfer.

‘…….the Diels-Alder reaction a chemical cycloaddition reaction of great importance…..’      The eponymous title of this reaction (named for the two German chemists who originated it) has also lent itself towards the naming of the insecticides Dieldrin and Aldrin.     ‘…..both protein enzymes (catalytic antibodies) and ribozymes have been derived which are capable of catalyzing this [Diels-Alder] reaction.’       For catalytic antibody Diels-Alderase, see Heine et al. 1998; for an example of ribozyme Diels-Alderase work see Agresti et al. 2005.     ‘…..there is at present no definitive evidence that a natural precedent exists for this particular catalytic activity [Diels-Alderases]….’      See Kim et al. 2012.

‘……alkyne (‘acetylenic’) triple bond is not at all absent from biological systems, being found in metabolic products of diverse organisms……’      A note on terminology: strictly speaking, an ‘alkyne’ would only refer to a hydrocarbon containing at least one carbon-carbon triple bond. Here it denotes substituted alkynes, where one or both of the hydrogen atoms of an acetylene (C2H2) are replaced by other chemical groups, which could be as large as a protein molecule, or which could join cyclicly to form ring structures. Triple-bonded carbons within larger molecules are also often referred as ‘acetylenic’ after the prototypical alkyne molecule. For a good review of natural molecules containing such alkyne groups, see Minto & Blacklock 2008.

‘  Organic azides, on the hand, are believed to be totally absent from known biosystems.’     See Sletten & Bertozzi 2009.

‘……….if organic azides were artificially introduced into specific cells synthesizing certain natural products ….. (members of the enediyne natural product family), then an azide-associated reaction with the enediyne product could occur.’     The natural enediynes have 9- or 10-membered ring structures, less strained then cyclooctynes (as in Fig. 2 above), but still significantly strained relative to a linear alkyne. See Kim at al. 1993 for the structure of an enediyne antibiotic when complexed with its specific carrier protein.

‘……….. the resulting cycloaddition reaction, deriving from a natural host product and a non-biological reactant, would lack complete orthogonality.’      This hypothetical example is another illustration of how bioorthogonality may exist within a specific closed and defined biosystem, but not necessarily for the biosphere as a whole.

‘……alkyne groups are recognized through biosynthetic machinery……’      See Van Lanen & Shen 2008.

‘……certain compounds with carbon-carbon triple bonds have been well-characterized as P450 inhibitors….’     See Lin et al. 2011.

‘…..the best-known example of this is the compound AZT (3’-azido-3’deoxythymidine)…..’      AZT is a nucleoside analog, in common with numerous other antiviral compounds, aimed at impeding the function of virally-encoded polymerases.

‘…..limitation of AZT was its tendency to produce often serious side-effects…..’      One such side-target of AZT is human mitochondrial RNA polymerase II; see Arnold et al. 2012.

‘…..1,2,3-triazole group, is likewise recognized in a variety of biological contexts, including P450 enzymes….’    See Conner et al. 2012.

‘……1.2.3-triazole isomer is also a central moiety of a variety of drug candidates……’     For example, See Röhrig et al. 2012 (a dioxygenase target in cancer contexts), Manohar et al. 2011 (malaria), and Jordão et al. 2009; Jordão et al. 2011 (antiviral).

‘…….new bioorthogonal reactions are being actively sought and found.’     As an example, see Sletten and Bertozzi 2011.

Next post: April.

Advertisements
One Comment leave one →
  1. February 24, 2014 2:33 pm

    Incredible points. Great arguments. Keep up the great morale.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s