On-array Synthesis

Chemical diversity for chemical microarrays can be accessed in two ways importing compounds onto the chip surface or synthetically creating molecules directly on the support. In the pioneering work by Fodor and coworkers, a combination of solid-phase organic synthesis and photolithographic techniques was applied to the in siiu creation of oligonucleotide and peptide microarrays [54, 55]. 

A different way to produce chemical microarrays in situ is spot synthesis of combinatorial libraries on cellulose sheets [56]. Spot synthesis is configured as an open system to be operated at room temperature. Despite attempts to replace cellulose with polypropylene as a synthesis support [57], cellulose is still the support of choice for spot synthesis, and reaction conditions have to be compatible with the restricted chemical stability of cellulose. Due to the acid labihty of such membranes, the diversity content of these arrays was initially restricted to the synthesis of peptides. Recently, a method was described that could widen the scope of spot synthesis arrays. Germeroth and coworkers [57] succeeded in the assembly of a library of 8000 cellulose-bound 1,3,5-triazines under mild reaction conditions. They employed a strategy that took advantage of a temperature-dependent, successive displacement of cyanuric chlorides by different nucleophiles in a first report of the synthesis of smaU organic compounds on ceUulose sheets. 

In addition to the limited range of cellulose-compatible synthesis protocols, two further drawbacks remain that are inherent to on-array approaches. Because synthesis takes place directly on the surface that is subsequently used for screening, quality control of the synthesis products is restricted to the surface-bound molecules. Standard cleave-and-characterize procedures involving LC-MS analytical techniques are not practical. This is a severe problem even for the synthesis of peptides where well-established protocols are available and becomes more pronounced when novel chemistries have to be employed. Secondly, each array produced is unique, which renders the production rather costly and prevents the generation of numerous copies of the same array for high-throughput applications. 

---