The earliest envisioned applications of any technology are not always the ones that end up being the most powerful or popular application. A good example would be RFID tags, the predecessor of which developed in 1945 in the USSR as a covert listening device. Now it’s used anywhere where large numbers of items need to be individually tracked, such as toll roads, mass transit systems, casino chips, and most famously every item in a Wal-Mart.
Fluorous technology is no different. As first described by Horváth and Rábai in 1994 it was seen as a green technology for catalyst recovery and recycle. By the late 90′s, Curran had pioneered “light” fluorous chemistry using fluorous solid phase extraction (FSPE) for the removal of tin residues from free radical reactions. This led to the realization that fluorous separations could be a powerful and highly attractive alternative to solid-phase for the synthesis of small molecule libraries. As the 2000′s progressed, fluorous techniques have seen increasing application in biomolecule and life sciences applications as a sample enrichment method for classes of compounds rather than a purification method for individual compounds. This has led to applications in proteomics, metabolomics, and microarraying. Just based on the peer-reviewed literature, the use of fluorous methods for the synthesis of small molecule libraries is no longer the preeminent use, but one of many varied applications.
Based on FTI sales and website traffic, however, small molecule library synthesis still remains the major application of the technology. The use of fluorous methods in library production is dependent of course on the state of library production or combinatorial chemistry in general. So what is the state of combinatorial chemistry? Certainly it isn’t nearly as popular as it was ten years ago and many view it as a complete failure of a strategy in terms of drug discovery. A new “Highlight” publication in Chemical Communications written by Prof. Tom Kodadek, currently at the Scripps Research Institute in Jupiter, FL, provides his individual perspective on combinatorial chemistry. The paper is entitled “The rise, fall, and reinvention of combinatorial chemistry“.
Prof. Kodadek correctly points out that combinatorial chemistry has in many ways been a victim of its own hype which he likened to Wall Street’s “irrational exuberance” of the same era. This resulted in a predictable and widespread backlash which has caused people to distance themselves from the entire field. He uses the Combinatorial Chemistry Gordon Conference changing it’s name to High Throughput Chemistry and Chemical Biology as an example. (One could also cite the Journal of Combinatorial Chemistry changing its name also.)
The history of combinatorial chemistry started with large (tens of thousands) unbiased libraries which could serve as sources of lead compounds. Critics pointed that such an exercise was relatively meaningless when considering the vast number of possible compounds with MW<500. Since the early 2000′s combinatorial chemistry has, therefore, primarily been used in the preparation of small focused libraries (50-200 compounds or so) for lead optimization. Fluorous methods have been used extensively and quite successfully in this type of library synthesis. (Another example of this was recently published by Zhang, W. et al in the production of a small substituted tetrahydrofuran library, see below.)
The problem with using combichem only for lead optimization is that it relies on other methods for the lead discovery which usually mean HTS of existing compound collections many of which number in the millions. The major criticism of most collections however is that they are dominated by flat, arene rich, hydrophobic compounds. (These are the same type of compounds that are usually made by the small focused libraries also, so one gets stuck in a local minima.) That’s all well and good, but unfortunately it misses out on a lot of chemical space occupied by more natural product-like compounds with greater 3-dimensional richness that some have argued have historically been much better sources of hits and leads that result in drugs.
Kodadek posits that combinatorial chemistry should be applied to large unbiased libraries for de novo screening efforts as was originally intended, in order to overcome the limitations of current compound collections. Basically he cites two ways to accomplish this; 1) diversity-oriented synthesis (DOS) which prepares libraries of compounds with varied structural motifs and 2) big, and I mean BIG, compound libraries with numbers described in scientific notation.
For DOS, he specifically cites Nelson’s use of fluorous tags with olefin metathesis as an elegent example of high molecular diversity within a library. These types of libraries, however, cannot lead to massive libraries due to the need for purification of each individual compound, which becomes the rate limiting step. The design of the Nelson library led to molecules that underwent cascade olefin metathesis with concommitant detagging, thus any molecules which did not undergo full reaction remained fluorous tagged. These could be easily separated from the products by FSPE. Since one can purify intermediates and products easily with little method development this leads to crude products which are more pure thus simplifying and streamlining final HPLC purification. Compounds of higher purity, higher quantity, higher diversity = better library.
Kodadek, however, clearly favors massive libraries with hundreds of millions of compounds. Some of the examples use split-and-pool strategies with DNA encoding. These libraries can then be evaluated in binding screens where library members bound to proteins of interest are detected or separated through a number of means. The problem with these type of libraries is that only very high yielding reactions can be used since it’s critical that the DNA encryption is for a single compound. So the examples are mostly peptoids. For general small molecule synthesis this is going to be tough.
Of course, nothing says you can’t combine fluorous tagging with DNA encoding. Considering that fluorous tags have been used with oligonucleotide synthesis it should work quite well. A properly designed split-and-pool reaction sequence which incorporates encryption along with fluorous tagging could enable bigger DOS-type libraries. After fluorous detagging binding assays could be used in a similar fashion. For example, one could imagine a vast library of cyclic peptides being fashioned where cyclization would result in fluorous detagging. Any non-productive cyclizations would retain the fluorous tag, much like in Nelson’s olefin metathesis, to provide acceptable purity for a reaction that otherwise would be unusable using such a strategy.
Prof. Kodadek concludes by stating that despite the “roller coaster history” of combinatorial chemistry, he believes that it is poised for a rebirth due to advancements in both synthetic and screening technologies. If so then fluorous should be able to play a major role whether in DOS libraries as has been conducted or in giant encoded libraries. Either way the quality and diversity of the compounds can hopefully be enhanced leading to new therapeutics.
I’m sure that readers will have their own opinions on the future of combinatorial chemistry.