Split synthesis

Split synthesis is a completely different approach to library generation from parallel synthesis. Furthermore, split synthesis can be performed only on a resin solution phase is not an option. Split synthesis involves dividing, or splitting, a pile of loaded resin into batches for reaction with a set of building blocks. After the reaction, all the resin is pooled back together, thoroughly mixed, and then resplit for reaction with the next set of building blocks. Based on this protocol, split synthesis is often also called mix-split or pool-split synthesis. 

SCHEME 9.12 Split synthesis and screening of an eight compound library 

An important idea for split synthesis is that each well must contain a large number of individual resin beads. The large number of beads is important for two reasons. One, each well must include a representative sampling of beads from the previous reactions. Two, each reaction must form enough reacted beads to provide samples for each well in the next reaction. Since the mixing and splitting processes are not perfect, the use of large numbers of beads ensures that each desired library member will be prepared on at least one bead. 

The number of wells needed for each reaction step is the same as the number of building blocks in the step. This is a reduction in workload compared to parallel synthesis. Once the synthesis is complete, the molecules are cleaved from the resin into wells, one well per bead. 

At the end of a split synthesis, because the beads have been pooled and mixed, the exact identity of a molecule on a given bead is unknown. Likewise, the identity and structure of compounds in wells is unknown. Split synthesis is not a spatially addressable method. Fortunately, the exact structure does not need to be known unless a compound shows activity in a screen. If active, the structure of the compound in the well will need to be elucidated through a process called deconvolution. Deconvolution is generally accomplished through one of two methods recursive deconvolution21 or binary encoding.22 

---

Two main approaches to combinatorial chemistry are used—parallel synthesis and split synthesis. In parallel synthesis, each compound is prepared independently. Typically, a reactant is first linked to the surface of polymer beads, which are then placed into small wells on a 96-well glass plate. Programmable robotic instruments add different sequences of building blocks to tfie different wells, thereby making 96 different products. When the reaction sequences are complete, the polymer beads are washed and their products are released. 

Due to the statistical nature of split synthesis, an average 20 beads per member must be used to ensure that all library members are prepared successfully This requirement results in increased reagent consumption and extra effort in screening and structural determination. One of the most powerful aspects of the rf memory chips is the ability to perform automated sorting of library members. Thus a complete library may be prepared with each member synthesized only once... 

A library containing several million beads can be screened in a single afternoon. Furthermore, the library is reusable, as it may be washed in 8 M guanidine hydrochloride and then re-screened using a different probe. This split synthesis approach displays the ability to generate peptide libraries of incredible variety, variety that can be further expanded by incorporation of, for example, D-amino acids or rarely occurring amino acids. 

Figure 2.6. The use of the split synthesis technique to generate a dipeptide library. After the two coupling steps shown are completed, the four possible (2 ) dipeptide sequence combinations are found. Note that any single bead will have multiple copies of the (identical) dipeptide attached and not just a single copy as shown above. This synthesis technique is economical and straightforward to undertake and requires no sophisticated equipment. The coupling steps, etc. utilized are based upon standard peptide S5mthesizing techniques, such as the Merrilield method (Box 2.1)...

Split synthesis of the library using four amino acids, four aldehydes, and five olefins in the presence of four mercaptoacyl chlorides (Scheme 3.110) generated the required proline library that was screened, after TEA cleavage of the products from the solid support, for inhibition of angiotensin converting enzyme ACE. 

Although Schreiber did not discover highly active compounds from his library, the power of split synthesis to make a library of over 2 million compounds is nonetheless impressive. 

Because each compound occupies its own well, parallel synthesis libraries require less effort for deconvolution than split libraries. Deconvolution in parallel synthesis arises from reactions that generate mixtures of diastereomers, enantiomers, or regioisomers. Split synthesis libraries require deconvolution to determine the synthetic pathway as well as any isomeric compounds that may be present. 

The amount of a single compound generated by split synthesis is limited to the amount that can be loaded onto a single bead, typically a few hundred picomoles (1 pico-mole = 10 12 moles). Specialized linkers can increase the amount,25 but the final quantity is still small relative to parallel synthesis. A parallel solid-phase synthesis uses many beads per well and therefore generates more product. 

A 27-member library was prepared by split synthesis in three steps. Each step involved three different building blocks A, B, C 1, 2, 3 and a, b, c. The library afforded one hit. In a recursive deconvolution of the library, how many compounds would need to be made and tested to determine the identity of the hit For a hypothetical three-step library of steps involving x, y, and z building blocks, how many compound syntheses would be required for recursive deconvolution How many compounds would be required to deconvolve a four-step library of m, n, o, and p building blocks ... 

Split synthesis enables the rapid generation of large numbers of related compounds (a library ) on a solid support. The procedure entails three steps that can be repeated many times. First, the resin is divided into equal portions (step 1). Each resin portion is then subjected separately to different reactions (step 2). The modified resin portions are combined and mixed (step 3) and are ready for the next cycle of the synthesis (Figure B.ll.l). Successive cycles of distributing, reacting, and mixing of the beads lead to a combinatorial increase of the diversity of products - for example four cycles with 15 different reactions per cycle lead to 154 = 50625 different com-... 

Fig. 6. Split synthesis strategy using the OntoCODE system.

---