Combinatorial chemistry with biological help
Michael Gross investigates the ways in which nature can be used to help in the quest for new molecules
- Combinatorial chemistry offered great promise as a way to help drug discovery but success has been limited
- New approaches have used nature as an inspiration to find different ways to join synthetic molecules with DNA fragments
- The information from DNA tagging could help decipher the origins of the first nucleic acids in the evolving Earth
Imagine you had all the different organic molecules that can be made, up to a size of, say, 500 Dalton, which is about the size range of monomers used in the cell. If you had just one molecule of each kind, how big a test tube would you need to hold them all? Estimates suggest that the galaxy might not be large enough for this purpose.
From this consideration, the idea of a vast 'chemical space' was born. Like the real space in our universe, most of the chemical space is empty (in that the corresponding molecules haven't been made) and even more of it is unexplored. A living cell can get by with producing less than 100 different kinds of small molecules, thus it uses only an infinitesimally small proportion of the chemical space.
Similarly, all humanity's needs for drugs and chemistry-based materials could probably be satisfied with a very small fraction of chemical space. The only problem is that researchers don't know their way around chemical space, so they don't know where to find the right molecules.
Thus they often take the approach of a drunkard looking for his key under a street light, where visibility is better, rather than in the dark alley where he probably lost it. That is, research tends to explore the immediate neighbourhood of those areas that are already illuminated by previous research.
A more recent strategy is to explore larger parts of the space more or less at random, but with high efficiency, an approach known as high throughput screening. The combined progresses made in sensor, microarray, and robotic technology allow researchers to sift through thousands of different molecules and test them for desirable properties in parallel, implying that millions can be tested in a reasonably short time. But where can one find millions of new, previously uncharacterised molecules to feed the screening automata?
In the quest to produce large numbers, multiplication tends to be more useful than addition. This is where combinatorial chemistry comes in. If you can get a library of m molecules to react with another one of n different molecules, you might get up to m*n products to screen from.
Based on this idea, combinatorial chemistry became a buzzword in the mid-1990s. The approach was first tried in drug discovery and then also in inorganic, organic and materials chemistry. Although this development has produced some success stories, the hit rate generally remained low, such that even from vast libraries, only very few new compounds emerged.
Therefore, researchers have continued to modify combinatorial approaches in a quest to improve the outcome. In recent years, many have looked to nature for inspiration.
Nature has used combinatorial chemistry for billions of years. Combining a small number of building blocks (four DNA bases, 20 amino acids), the cell can create an astronomical diversity of different molecules.
The 30 000 or so proteins encoded in the human genome are just a tiny fraction of the chemical space that would in principle be accessible via protein synthesis.
Some branches of combinatorial chemistry have directly benefited from the combinatorial nature of the DNA/protein system of the cell. Manfred Reetz at the Max-Planck-Institut für Kohlenforschung, in Mülheim an der Ruhr, Germany, and others have, for example, used in vitro evolution to create novel enzymes for chiral synthesis.
RNA researchers have used guided in vitro evolution to produce new ribozymes and gain insight into what kinds of molecules might have populated the RNA world at the dawn of life. And yet, if one sticks with the biologically relevant monomers, the resulting combinations will cover only a small fraction of chemical space. While this fraction is known to contain biologically relevant substances, it might miss attractive solutions coming from entirely non-biological areas of chemical space.
Therefore, researchers have started to combine new sets of building blocks with nature's powers to combine, synthesise and select. Adding new amino acids or new DNA bases to the set only broadens the accessible range.
In a much more radical approach, known as DNA-templated synthesis, David Liu and his co-workers at Harvard University in the US have hooked up small synthetic molecules with DNA oligomers to direct their reactions and to be able to select the products.
In a much noted pair of papers that appeared within weeks of each other in Nature and Science, Liu's group harnessed the power of this combination for different purposes, namely for the selection of a macrocycle with desirable binding properties, and for the discovery of a new kind of chemical reaction.
The latter work demonstrates how the combinatorial power of DNA can benefit chemistry. Liu and his co-workers created two libraries (A and B) each containing 12 small molecules representing different chemical functionalities (A1-A12, B1-B12).
Each of the 144 possible combinations is represented by a unique DNA sequence composed of a region specific for one of the A molecules, and one for the B molecules. This means there is a set of 12 different DNA sequences coding for each specific A molecule, say A1, and all its possible combinations with B molecules.
Nanomolar amounts of the A1 molecule will be linked to each of these 12 DNA strands at the end coding for the B molecules. All B molecules will be linked with shorter DNA strands complementary to the B-specific strands next to the A molecule, such that the DNAs of A and B molecules will form duplexes representing all 144 combinations and thus bring the small molecules into close proximity (high local concentration) of each other, all in the same solution.
While the A molecules are directly bound to their DNA strands, the B molecules are labelled with biotin and then attached to the DNA via a disulphide bridge that can be easily dissolved. So after the libraries have been incubated together to allow the molecules time to react, the researchers break the disulphide bonds between B molecules and their DNA and select for biotin-labelled molecules by fishing with avidin-coated magnetic beads.
The resulting solution will contain unreacted B molecules without their DNA plus any newly formed AB reaction products attached to the A-linked DNA, which encodes the identity of the pair.
So by using standard molecular biology methods (PCR, fluorescent labelling, DNA array), researchers can rapidly find out which pairs of molecules have formed a covalent bond. Using this approach, Liu's group has discovered a new reaction generating an enone from an alkyne and an alkene using a palladium catalyst at 25°C.
While the synthesis of all the different DNA combinations is a significant starting investment, the power of the system lies in the fact that the entire selection experiment can be easily and rapidly repeated for many different reaction conditions. If none of your reagents have formed a bond by themselves, you could try out 100 different catalysts or 20 different temperatures. Liu estimates that a nanomole preparation of the DNA pool should suffice for over 1000 experiments, such that well over 100 000 combinations of reactants and reaction conditions can be tested.
The way in which Liu tags small molecules to nucleic acids could shed light on how the first nucleic acids (presumably RNA) recruited amino acids to create protein synthesis and ultimately DNA/RNA/protein life as we know it today. Like his DNA constructs, transfer RNAs (tRNAs) contain two separate coding regions, one for the codon (the base triplet encoding the amino acid) and one for the tRNA synthetase, which attaches the correct amino acid to the tRNA. Research into artificial coding systems like Liu's could clarify how the complex tRNA-based translation system might have come into being.
How chemistry can turn into life is one of the more ambitious questions that Liu wants to tackle. 'A major long-term goal of my research programme will be the creation of synthetic biotic systems with the potential to evolve in unpredictable ways,' he says.
'This will involve the integration of nucleic acid-templated synthesis together with other systems not yet developed. From this new line of research, I hope to gain insight into the minimal chemical requirements for living, evolving systems ...' but not content with that, Liu also hopes for further spinoffs for chemistry, '... as well as to discover functional small molecules and macromolecules that would be difficult to find using existing methods.'
When nature started experimenting with combinatorial chemistry, some 3.5 billion years ago, the resulting living systems soon became encapsulated in cell membranes, suggesting that containers are important. Some researchers have even suggested that other containers such as rock pores have hosted the very first dabblings of biochemistry.
Consequently, some researchers have made use of natural containers to constrain the movements of reagents in combinatorial chemistry. One approach of obvious usefulness is to use the active site cavity of an enzyme as a test tube for the reaction that will form a tight-binding inhibitor for this enzyme.
Recent work from the laboratories of Barry Sharpless and Hartmuth Kolb at the Scripps Institute, La Jolla, California, has demonstrated the combination of a non-biological reaction mechanism, namely the cycloaddition of azides with alkynes, with an enzyme template to yield powerful inhibitors for enzymes including carbonic anhydrase and acetylcholinesterase (see Chemistry World February 2005, p12).
It appears that our voyage into chemical space has only just begun. Using a little help from the combinatorial chemistry of nature, chemists will be able to explore not only the biological parts of that space, but even new galaxies that nobody has dared to dream of.
Michael Gross is science writer in residence at the school of crystallography, Birkbeck College, University of London.
- C M Dobson, Nature, 2004, 432, 824
- M T Reetz, Angew. Chem. Int. Ed., 2001, 40, 284
- Z J Gartner et al, Science, 2004, 305, 1601
- M W Kanan et al, Nature, 2004, 431, 545
- S K Silverman, Org. Biomol. Chem., 2004, 2, 2701