Combinatorial Chemistry PDF

146 CH5 COMBINATORIAL CHEMISTRY of a library for activity using...

146 CH5 COMBINATORIAL CHEMISTRY of a library for activity using high-throughput screening techniques enables the development team to select suitable compounds for a more detailed investigation by either combinatorial chemistry or other methods. 5 The basic concept of combinatorial chemistry is best illustrated by an example. Consider the reaction of a set of three compounds (A1–A3) with a set of three building blocks (B1–B3). In combinatorial synthesis, A1 would simultaneously undergo separate reactions with compounds B1, B2 and B3, respectively (Fig. 5.2). At the same time compounds A2 Combinatorial chemistry Stage 1 Stage 2 R' R' R' H2NCHCOOR" R"'NH2 RCOCl RCONHCHCOOR" RCONHCHCONHR"' 5.1 Introduction A1-B1 -C1 A1-B1 A1 -B1-C2 The rapid increase in molecular biology technology has resulted in the development of A1 -B1-C3 rapid, efficient, drug testing systems. The techniques used by these systems are collecti- A1-B2-C1 vely known as high-throughput screening (HTS). High-throughput screening methods A1 A1-B2 A1-B2 -C2 give accurate results even when extremely small amounts of the test substance are A1 -B2 -C3 available (see section 5.6). However, if it is to be used in an economic fashion as well as A1-B3 -C1 efficiently, this technology requires the rapid production of a large number of substances A1 -B3 A1 -B3-C2 for testing that cannot be met by the traditional approach to organic synthesis, which is A1 -B3-C3 usually geared to the production of one compound at a time (Fig. 5.1). Using this slow, A2 -B1-C1 labour-intensive traditional approach, a medicinal chemist is able to produce about 25 test A2-B1 -C2 A2-B1 compounds a year. Consequently, the production of the large numbers of compounds required A2-B1 -C3 by HTS would be expensive, both economically and in time, if this approach was used. A2 -B2-C1 A2 A2-B2 A2-B2 -C2 HOCH2 CH2 N(C2 H5 )2 A2-B2 -C3 C4 H9 Br + H2N COOH C 4 H 9 NH COOH C 4 H 9 NH COOCH2 CH2 N(C2 H5)2 A2-B3-C1 N-Alkylation Esterification A2 -B3 A2-B3 -C2 Tetracaine A2 -B3 -C3 Figure 5.1 A traditional stepwise organic synthesis scheme illustrated by the synthesis of the local anaesthetic tetracaine A3 -B1-C1 A3 -B1 A3-B1 -C2 A3-B1 -C3 Combinatorial chemistry was developed to produce the large numbers of compounds A3 -B2-C1 required for high-throughput screening. It allows the simultaneous synthesis of a large A3 A3-B2 A3-B2 -C2 number of the possible compounds that could be formed from a number of building blocks. A3-B2 -C3 The products of such a process are known as a combinatorial library. Libraries may be a A3-B3-C1 collection of individual compounds or mixtures of compounds. Screening the components A3 -B3 A3 -B3 -C2 A3-B3 -C3 Medicinal Chemistry, Second Edition Gareth Thomas Figure 5.2 The principle of combinatorial chemistry illustrated by a scheme for synthesis of a hypothetical # 2007 John Wiley & Sons, Ltd polyamide using three building blocks at each stage  5.1 INTRODUCTION 147 148 CH5 COMBINATORIAL CHEMISTRY and A3 would also be undergoing reactions with compounds B1, B2 and B3. These 1. The reactions should be specific, relatively easy to carry out and give a high yield. simultaneous reactions would produce a library of nine products. If this process is repeated by reacting these nine products with three new building blocks (C1–C3), a combinatorial 2. The reactions used in the sequence should allow for the formation of as wide a range of library of 27 new products would be obtained. structures for the final products as possible, including all the possible stereoisomers. The reactions used at each stage in such a synthesis normally involve the same functional groups, that is, the same type of reaction occurs in each case. Very few libraries have been 3. The reactions should be suitable for use in automated equipment. constructed where different types of reaction are involved in the same stage. In theory this approach results in the formation of all the possible products that could be formed, but in 4. The building blocks should be readily available. practice some reactions may not occur. However, the combinatorial approach does mean that normally large libraries of many thousands of compounds can be formed rapidly in the 5. The building blocks should be as diverse as possible so that the range of final products same time that it takes to produce one product using the traditional approach to synthesis. includes structures that utilise all the types of bonding (see section 8.2) to bind to or A number of the techniques used in combinatorial chemistry to obtain this number of react with the target. products are dealt with in sections 5.2 and 5.4. 6. It must be possible to accurately determine the structures of the final products. 5.1.1 The design of combinatorial syntheses In practice it is not always possible to select reactions that meet all these criteria. However, criterion 6 must be satisfied otherwise there is little point in carrying out the synthesis. One of two general strategies may be followed when designing a combinatorial synthesis The degree of information available about the intended target will also influence the (Fig.5.3a). In the first case the building blocks are successively added to the preceding selection of the building blocks. If little is known a random selection of building blocks is structure so that it grows in only one direction (linear synthesis, see section 15.2.3). It used in order to identify a lead. However, if a there is a known lead, the building blocks are usually relies on the medicinal chemist finding suitable protecting groups so that the selected so that they produce analogues that are related to the structure of the lead. This allows reactions are selective. This design approach is useful if the product is a polymer the investigator to study the SAR/QSAR and/or determine the optimum structure for potency. (oligomer) formed from a small number of monomeric units. Alternatively, the synthesis can proceed in different directions from an initial building block known as a template provided that the template has either the necessary functional groups or they can be 5.1.2 The general techniques used in combinatorial synthesis generated during the course of the synthesis (Fig. 5.3b). Both routes may require the use of protecting groups (see section 15.2.4). Combinatorial synthesis may be carried out on a solid support (see section 5.2) or in solution (see section 5.4). In both cases, synthesis usually proceeds using one of the strategies outlined in Figure 5.3. Both solid support and solution synthetic methods may be B C D used to produce libraries that consist of either individual compounds or mixtures of A A–B A–B–C A – B – C –D compounds. Each type of synthetic method has its own distinct advantages and (a) D disadvantages (Table 5.1). A – B –C D A C D B A–B A–B–C A – B – C –D 5.2 The solid support method D D – A – B –C (b) The solid support method originated with the Merrifields (1963) solid support peptide synthesis. This method used polystyrene–divinylbenzene resin beads as a solid support for Figure 5.3 (a) Linear synthesis. The sequential attachment of building blocks. (b) Template Method. The the product of each stage of the synthesis. Each bead had a large number of non-sequential attachment of building blocks using B as a template monochlorinated methyl side chains. The C-terminal of the first amino acid in the peptide chain was attached to the bead by an SN2 displacement reaction of these chloro groups by a The reactions used when designing a combinatorial sequence should ideally satisfy the suitable amino acid (Fig. 5.4). The large number of chlorinated side chains on the bead following criteria: meant that one bead acts as the solid support for the formation of a large number of peptide  5.2 THE SOLID SUPPORT METHOD 149 150 CH5 COMBINATORIAL CHEMISTRY Table 5.1 A comparison of the advantages and disadvantages of the solid support and in solution R1 techniques of combinatorial chemistry HOOCCHNH 2 {(CH3)3COCO}2 The amino group of the first amino acid is On a solid support In solution Di-t-butyl carbonate protected by a t-butyloxycarbonyl (Boc) Reagents can be used in excess in order to Reagents cannot be used in excess, unless addition group. drive the reaction to completion purification is carried out (see section 5.4.6) R1 Resin bead CH2Cl + HOOCCHNH CO O C(CH3)3 Purification is easy, simply wash the support Purification can be difficult Merrifield bead (C2 H5)3N Attachment of the C-terminal of the Automation is easy Automation may be difficult N-protected amino acid to the resin Fewer suitable reactions In theory any organic reaction can be used R1 Scale up is relatively expensive Scale up is relatively easy and inexpensive A CH2 OOCCHNH CO O C(CH3 )3 Not well documented and time will be required Only requires time for the development of the Removal of the protecting group to find a suitable support and linker for a chemistry Acid and purification of the product specific synthesis by washing R1 CH 3 B CH2 OOCCHNH 2 + CO2 + C CH 2 CH 3 molecules of the same type. Additional amino acids were added to the growing peptide R2 NH Addition of the next N-protected chain using the reaction sequence shown in Figure 5.4. This sequence, in common with ( CH 3 ) 3 COCONHCHCOOC amino acid and purification of the other peptide syntheses, uses protecting groups such as t-butyloxycarbonyl (Boc) to control N product by washing the position of amino acid coupling. To form the amide peptide link the N-protected amino NHCONH acids were converted to a more active acylated derivative of dicyclohexylcarbodiimide (DCC), which reacted with an unprotected amino group to link the new amino acid residue R R 1 2 Dicyclohexylurea to the growing peptide (Fig. 5.4). At the end of the synthesis the peptide was detached from C CH2 OOCCHNHCOCHNH-CO O C(CH3)3 the bead using a mixture of hydrogen bromide and trifluroethanoic acid. Elongation of the peptide by A Boc group R 2 repetition of steps A to C. 2 NH (CH3)3C O CONHCHCOOH R R 1 R R Boc-protected amino acid residue (CH3)3 C O CONHCHCOOC CH2 OOCCHNH OCCHNH COCHNH2 + N n N C N HBr / CF3 COOH Release of the peptide from the resin. Active acyl derivative of DCC Dicyclohexylcarbodiimide (DCC) R 1 R R CH2 Br + HOOCCHNH COCHNH COCHNH2 Merrifields’ original resin bead has been largely superseded by other beads, such as the n TentaGel resin bead. This bead is more versatile as it can be obtained with a variety of Peptide functional groups (X) at the end of the side chain (Fig. 5.5). These functional groups are Figure 5.4 An outline of the Merrifield peptide synthesis where R is any amino acid side chain separated from the resin by a polyethylene glycol (PEG) insert. As a result, the reacting groups of the side chains are further from the surface of the bead, which makes reactant access and subsequent reaction easier. Like the Merrifield bead, each TentaGel bead 5.2.1 General methods in solid support combinatorial chemistry contains a large number of side chain functional groups. For example, the number of amine groups per bead is about 6  1013. This means that in theory each bead could act as Solid support combinatorial chemistry has been carried out on a variety of supports that the support for the synthesis of up to 6  1013 molecules of the same compound. In peptide include polymer beads, arrays of wells, arrays of pins, glass plates, spatial arrays on synthesis the amount of peptide found on one bead is usually sufficient for its structure to microchips and cellulose sheets. However, most syntheses are performed using polymer be determined using the Edman thiohydantoin microsequencing technique. beads. The group that anchors the compound being synthesised to the bead is known either  5.2 THE SOLID SUPPORT METHOD 151 152 CH5 COMBINATORIAL CHEMISTRY Resin X The reactions used in a solid support combinatorial syntheses are, of necessity, those that bead PEG Key X is follow a relatively simple procedure. They are often developed from existing traditional Swells in aqueous NH2, OH, SH, organic chemistry reactions. This can take considerable time and effort to adapt a solution Br or COOH traditional organic reaction to use under solid phase conditions. Furthermore, many modified reactions cannot be used as they give too low a yield under solid phase conditions. Figure 5.5 TentaGel resin beads This is a serious limitation for solid phase combinatorial chemistry. Reactions are normally carried out by mixing the reagents with the solid support to bring about reaction and, after as a handle or a linker (Fig. 5.6). As well as modifying the properties of the bead they move reaction, washing the support with reagents and solvents to purify the product. Any the point of substrate attachment further from the bead, making reaction easier. The choice heating required is usually carried out by placing the solid support in an oven. Recently, of linker will depend on the nature of the reactions used in the proposed synthetic pathway. microwave heating has also been employed. However, multistep reactions and reactions For example, an acid-labile linker, such as HMP (hydroxymethylphenoxy), would not be that involve the use of high or low pressure, extreme temperature and inert atmospheres are suitable if the reaction pathway contained reactions that were conducted under strongly avoided. acidic conditions. Consideration must also be given to the ease of detaching the product from the linker at the end of the synthesis. The method employed must not damage the required product but must also lend itself to automation. 5.2.2 Parallel synthesis This technique is normally used to prepare combinatorial libraries that consist of separate Resin Linker compounds. It is not suitable for the production of libraries containing thousands to bead Synthesis site millions of compounds. In parallel synthesis the compounds are prepared in separate OH RCOOH O C R TFA Final reaction vessels but at the same time, that is, in parallel. The array of individual reaction O O O Product vessels often takes the form of either a grid of wells in a plastic plate or a grid of plastic The Wang linker for rods called pins attached to a plastic base plate (Fig. 5.7) that fits into a corresponding set carboxylic acids of wells. In the former case the synthesis is carried out on beads placed in the wells whilst O O ROH O O OR TFA in the latter case it takes place on so-called plastic ‘crowns’ pushed on to the tops of the FinalProduct pins, the building blocks being attached to these crowns by linkers similar to those found The tetrahydropyranyl (THP) on the resin beads. Both the well and pin arrays are used in the same general manner; the linker for alcohols position of each synthetic pathway in the array and hence the structure of the product of O Cl Amines O Amine TFA that pathway is usually identified by a grid code. C C FinalProduct O O A benzyloxycarbonyl chloride linker for amines Figure 5.6 Examples of linkers and the reagents used to detach the final product. TFA is trifluoroacetic acid Combinatorial synthesis on solid supports is usually carried out either by using parallel synthesis (see section 5.2.2) or Furka’s mix and split procedure (see section 5.2.3). The precise method and approach adopted when using these methods will depend on the nature of the combinatorial library being produced and also the objectives of the investigating team. However, in all cases it is necessary to determine the structures of the components of the library by either keeping a detailed record of the steps involved in the synthesis or giving beads with a label that can be decoded to give the structure of the compound attached to that bead (see section 5.3). The method adopted to identify the components of the library will depend on the nature of the synthesis. Figure 5.7 Examples of the arrays used in combinatorial chemical synthesis  5.2 THE SOLID SUPPORT METHOD 153 154 CH5 COMBINATORIAL CHEMISTRY bead. The amino acids are deprotected by hydrogenolysis and 12 isocyanates (Y1, Y2  Y8) 3 R 1 1 1 R T FA R 3 R O N O added to the wells so that each numbered row at right angles to the lettered rows contains R NCO 3 HCl OCOCNHBOC OCOCNH 2 OCOCNHCONHR only one type of isocyanate. In other words, compound Y1 is only added to row one, compound 2 or 2 Heat 2 N R R R 2 R Resin bead Piperidine R 1 H Y2 is only added to row two, and so on (Fig. 5.9b). The isocyanates are allowed to react to A substituted urea Hydantoins form substituted ureas. Each well is treated with 6M hydrochloric acid and the whole array heated to simultaneously form the hydantoins and release them from the resin. Although Figure 5.8 The reaction of amino acids with isocyanates to form hydantoins it is possible to simultaneously synthesise a total of 96 different hydantoins (Z1–Z26, Fig.5.9c) by this technique, in practice it is likely that some of the reactions will be unsuccessful and The technique of parallel synthesis is best illustrated by means of an example. Consider a somewhat smaller library of compounds is normally obtained. the general theoretical steps that would be necessary for the preparation of a combinatorial A well array combinatorial synthesis can consist of any number of stages. Each stage is library of hydantoins by the reaction of isocyanates with amino acids (Fig. 5.8) using a carried out in the general manner described for the previous example. However, at each 96-well array. At each stage in this synthesis the product would be purified by washing with stage only the numbered or lettered rows are used, not both, unless a library of mixtures suitable reagents. is required. Finally, the products are liberated from the resin by the appropriate linker Eight N-protected amino acids (X1, X2  X8) are placed in the well array so that only cleavage reaction (see Fig. 5.6) and the products isolated. The structures of these products one type of amino acid occupies a row, that is, row A will only contain amino acid X1, row are usually determined by following the history of the synthesis using the grid references of B will only contain amino acid X2, and so on (Fig.5.9a). Beads are added to each well and the wells and confirmed by instrumental methods (mainly NMR, GC, HPLC and MS). the array placed in a reaction environment that will join the X compound to the linker of the The pin array is used in a similar manner to the well array except that the array of crowns is inverted so that the crowns are suspended in the reagents placed in a corresponding array A B C D E F G H A B C D E F G H of wells (Fig. 5.7b). Reaction is brought about by placing the combined pin and well unit in 1 X1 X2 X3 X4 X5 X6 X7 X8 1 X1-Y1 X2-Y1 X3-Y1 X4-Y1 X5-Y1 X6-Y1 X7-Y1 X8-Y1 a suitable reaction environment. The loading of the wells follows the pattern described in 2 X1 X2 X3 X4 X5 X6 X7 X8 2 X1-Y2 X2-Y2 X3-Y2 X4-Y2 X5-Y2 X6-Y2 X7-Y2 X8-Y2 3 X1 X2 X3 X4 X5 X6 X7 X8 3 X1-Y3 X2-Y3 X3-Y3 X4-Y3 X5-Y3 X6-Y3 X7-Y3 X8-Y3 Figure 5.9. 4 X1 X2 X3 X4 X5 X6 X7 X8 4 X1-Y4 X2-Y4 X3-Y4 X4-Y4 X5-Y4 X6-Y4 X7-Y4 X8-Y4 5 X1 X2 X3 X4 X5 X6 X7 X8 5 X1-Y5 X2-Y5 X3-Y5 X4-Y5 X5-Y5 X6-Y5 X7-Y5 X8-Y5 6 X1 X2 X3 X4 X5 X6 X7 X8 6 X1-Y6 X2-Y6 X3-Y6 X4-Y6 X5-Y6 X6-Y6 X7-Y6 X8-Y6 Fodor’s method for parallel synthesis 7 X1 X2 X3 X4 X5 X6 X7 X8 Deprotection 7 X1-Y7 X2-Y7 X3-Y7 X4-Y7 X5-Y7 X6-Y7 X7-Y7 X8-Y7 In theory almost any solid material can be used as the solid support for parallel 8 X1 X2 X3 X4 X5 X6 X7 X8 of the amino 8 X1-Y8 X2-Y8 X3-Y8 X4-Y8 X5-Y8 X6-Y8 X7-Y8 X8-Y8 9 X1 X2 X3 X4 X5 X6 X7 X8 acid combinatorial synthesis. Fodor et al. (1991) produced peptide libraries using a form of 9 X1-Y9 X2-Y9 X3-Y9 X4-Y9 X5-Y9 X6-Y9 X7-Y9 X8-Y9 10 X1 X2 X3 X4 X5 X6 X7 X8 10 X1-Y10 X2-Y10 X3-Y10 X4-Y10 X5-Y10 X6-Y10 X7-Y10 X8-Y10 parallel synthesis that could be performed on a glass plate. The plate is treated so that its 11 X1 X2 X3 X4 X5 X6 X7 X8 11 X1-Y11 X2-Y11 X3-Y11 X4-Y11 X5-Y11 X6-Y11 X7-Y11 X8-Y11 surface is coated with hydrocarbon chains containing a terminal amino group. These amino 12 X1 X2 X3 X4 X5 X6 X7 X8 12 X1-Y12 X2-Y12 X3-Y12 X4-Y12 X5-Y12 X6-Y12 X7-Y12 X8-Y12 groups are protected by the UV-labile 6-nitroveratryloxycarbonyl (NVOC) group. (a) The placement of the first building (b) The placement of the isocyanate building blocks blocks, the Boc protected amino acids Y1 to Y8 CH 3 O NO 2 X1 to X12 and their attachment to the resin CH 3 O NO 2 A B C D E F G H CH 3 O 1 Z1 Z2 Z3 Z4 Z5 Z6 Z7 Z8 NH 2 O O CH 3 O UV 2 Z9 Z10 Z11 Z12 Z13 Z14 Z15 Z16 C NVOC O O 3 Z17 Z18 Z19 Z20 Z21 Z22 Z23 Z24 Cl C R1 4 Z25 Z26 Z27 Z28 Z29 Z30 Z31 Z32 NVOC NHCHCOOR The hydantoins 5 Z33 Z34 Z35 Z36 Z37 Z38 Z39 Z40 NH 2 NH Z1 to Z96 6 Z41 Z42 Z43 Z44 Z45 Z46 Z47 Z48 7 Z49 Z50 Z51 Z52 Z53 Z54 Z55 Z56 Step A (c) Reaction, by placing the array in a Glass plate 8 Z57 Z58 Z59 Z60 Z61 Z62 Z63 Z64 suitable reaction environment, to form the 9 Z65 Z66 Z67 Z68 Z69 Z70 Z71 Z72 substituted urea and subsequent treatment 10 Z73 Z74 Z75 Z76 Z77 Z78 Z79 Z80 with hot 6M hydrochloric acid to form the 1 Repeat steps A R1 R 11 Z81 Z82 Z83 Z84 Z85 Z86 Z87 Z88 hydantoins Z1 to Z96 and B until the NHCOCHNH 2 UV NHCOCHNH NVOC 12 Z89 Z90 Z91 Z92 Z93 Z94 Z95 Z96 peptide synthesis Step B Figure 5.9 The pattern of well loading for the formation of a combinatorial library of 96 hydantoins is complete  5.2 THE SOLID SUPPORT METHOD 155 156 CH5 COMBINATORIAL CHEMISTRY Mask Mask Mask libraries. Large libraries are possible because the technique produces one type of M1 M2 M3 compound on each bead, that is, all the molecules formed on one bead are the same but X X X X (1) UV X G X G (1) UV S G S G different from those formed on all the other beads. Each bead will yield up to 6  1013 X X X X (2) G X G X G (2) S S G S G product molecules, which is sufficient to carry out high-throughput screening procedures. The technique has the advantage that it reduces the number of reactions required to produce UV UV UV a large library. For example, if the synthetic pathway required three steps, it would require 30,000 separate reaction vessels to produce a library of 10,000 compounds if the reactions M1 M2 M3 were carried out in separate reaction vessels using orthodox chemical methods. The Furka XX X X XX X X XXGG XXGG SSGG SSGG mix and split method reduces this to about 22 reactions. The Furka method produces the library of compounds on resin beads. These beads are divided into a number of equally sized portions corresponding to the number of initial building Further masking and blocks. Each of the starting compounds is attached to its own group of beads using the (1) UV S G S G (1) UV S-F G-F amino acid coupling appropriate chemical reaction (Fig.5.11). All the portions of beads are now mixed and (2) A S-A G-A S-A G-A (2) F S-A G-A as required separated into the number of equal portions corresponding to the number of different starting compounds being used for the first stage of the synthesis. A different reactant building block is UV added to each portion and the reaction is carried out by putting the mixtures of resin beads and A A FA A F reactants in a suitable reaction vessel. After reaction, all the beads are mixed before separating A A M4 SSGG SSGG SSGG Figure 5.10 A schematic representation of the Fodor approach to parallel synthesis. X represents an NVOC- protected amino group attached to the glass plate. The other letters correspond to the normal code used for Resin beads amino acids. Each of these amino acids is in its NVOC-protected form A B C The bonding of the initial building blocks A B C A photolithography mask (M1) is placed over the plate so that only a specific area of the to the resin beads plate can be irradiated with UV light (Fig. 5.10). This results in removal of the NVOC COMBINE, MIX and SPLIT into three new portions protecting group from the amino groups in the irradiated area. The entire plate is exposed Step ONE in the D E F to the first activated NVOC-protected amino acid. However, it will only bond to the amino synthetic pathway A-D A-E A-F groups exposed in the irradiated area (Step A). The process is repeated using a new mask B-D B-E B-F C-D C-E (M2) and a second activated NVOC-protected amino acid attached to the exposed amino C-F groups (Step B). This process is repeated using different masks (M3, etc.) until the desired COMBINE, MIX and SPLIT into three new portions library is obtained, the structure of the peptide occupying a point on the plate depending on Step TWO in the G H I the masks used and the activated NVOC-protected amino acid used at each stage in the synthetic pathway A-D-G A-E-H A-F-I synthesis. The technique is so precise that it has been reported that each compound B-D-G B-E-H occupies an area of about 50 mm  50 mm. A record of the way in which the masks are C-D-G B-F-I C-E-H C-F-I used will determine both the order in which the amino acids are added and, as a result, the A-E-G A-D-H A-E-I structures of each of the peptides at specific coordinates on the plate. B-E-G B-D-H B-E-I C-E-G C-D-H C-E-I A-F-G A-F-H A-D-I 5.2.3 Furka’s mix and split technique B-F-G B-F-H B-D-I C-F-G C-F-H C-D-I The Furka method was developed by Furka and co-workers from 1988 to 1991. It uses resin Figure 5.11 An example of the Furka approach to combinatorial libraries using a two-step synthesis beads and may be used to make both large (thousands) and small (hundreds) combinatorial involving three building blocks at each stage  5.3 ENCODING METHODS 157 158 CH5 COMBINATORIAL CHEMISTRY them into the number of equal portions corresponding to the number of building blocks being produced in a sufficient amount to enable it to be decoded to give the history and hence the used in the second stage of the synthesis. A different second stage building block is added to possible structure of the library compound. each of these new portions and the mixture is allowed to react to produce the products for this Compounds used for tagging must satisfy a number of criteria: stage in the synthesis. This process of mix and split is continued until the required library is synthesised. In peptide and similar polymer library formation where the same building blocks (1) The concentration of the tag should be just sufficient for its analysis, that is, the are used at each step, the maximum possible number of compounds that can be synthesised for majority of the linkers should be occupied by the combinatorial synthesis. a given number of different building blocks (b) is given by: (2) The tagging reaction must take place under conditions that are compatible with those Number of compounds ¼ bx ð5:1Þ used for the synthesis of the library compound. where x is the number of steps in the synthesis. Unlike in parallel synthesis the history of the bead cannot be traced from a grid (3) It must be possible to separate the tag from the library compound. reference, it has to be traced using either a suitable encoding method (see section 5.3) or deconvolution (see section 5.5). Encoding methods use a code to indicate what has (4) Analysis of the tag should be rapid and accurate using methods that could be automated. happened at each step in the synthesis. They range from putting an identifiable tag compound on to the bead at each step in the synthesis to using computer-readable Many peptide libraries have been encoded using single-stranded DNA oligonucleotides silicon chips as the solid support. If sufficient compound is produced, its identity may as tags. Each oligonucleotide acts as the code for one amino acid (Table 5.2). Furthermore, also be confirmed using a combination of analytical methods such as NMR, MS, HPLC a polymerase chain reaction (PCR) primer is usually attached to the tag site so that at the and GC. end of the combinatorial synthesis the concentration of the completed DNA oligonucleo- tide tag may be increased using the Taq polymerase procedure. This amplification of the yield of the tag makes it easier to identify the sequence of bases, which leads to a more 5.3 Encoding methods accurate decoding. A wide variety of encoding methods have been developed to record the history of beads used in the Furka mix and split technique. This section outlines a selection of these Table 5.2 The use of oligonucleotides to encode amino acids in peptide methods. synthesis Amino acid Structure Oligonucleotide code Glycine NH2 CACATG 5.3.1 Sequential chemical tagging  (Gly) CH2COOH Methionine NH2  Sequential chemical tagging uses specific compounds (tags) as a code for the individual (Met) CH3SCH2CH2CHCOOH ACGGTA steps in the synthesis. These tag compounds are sequentially attached in the form of a polymer-like molecule to the same bead as the library compound at each step in the synthesis, usually by the use of a branched linker (Fig. 5.12). One branch is used for the library synthesis and the other for the encoding. At the end of the synthesis both the library At each stage in the peptide synthesis a second parallel synthesis is carried out on the compound and the tag compound are liberated from the bead. The tag compound must be same bead to attach the oligonucleotide tag (Fig. 5.13). In other words, two alternating parallel syntheses are carried out at the same time. On completion of the peptide synthesis Key: the oligonucleotide tag is isolated from the bead and its base sequence is determined and A-B-C-B-C-etc. Library compound Building block Code compound decoded to give the sequence of amino acid residues in the peptide. Linker A R B S Peptides and individual amino acids have also been used to code for the building blocks Resin R-S-T-S-T-etc. Code compound C T in a synthesis because they can be sequentially joined. Syntheses are usually carried out bead using a branched linker so that the synthesis of the encoding molecule can be carried out in Figure 5.12 Chemical encoding of resin beads. Branched linkers, with one site for attaching the library parallel to that of the combinatorial library molecule it encodes. For example, the compound and another for attaching the tag, are often used for encoding Zuckermann approach uses a diamine linker protected at one end by an Fmoc group and at  5.3 ENCODING METHODS 159 160 CH5 COMBINATORIAL CHEMISTRY bead. It should be remembered that each bead will yield up to 6  1013 product molecules, Synthesis site Linker Primer (P) Tag site which is sufficent to carry out high-throughput screening procedures. In addition, each bead will produce sufficient of the tagging compound to deduce the structure of the product (1) Gly Divide into two aliquots (1) Met molecule. The detached product and tag are separated and the sequence of amino acids in (2) CACATG (2) ACGGTA the encoding peptide is determined using the Edman sequencing method. This sequence is Gly (P)-CACATG Met (P)-ACGGTA used to determine the history of the formation and hence the structure of the product found Mix and split into two aliquots on that bead. (1) Gly (1) Met (2) CACATG (2) ACGGTA Gly-Met (P)-ACGGTACACATG Met-Met (P)-ACGGTAACGGTA Gly-Gly (P)-CACATGCACATG Met-Gly 5.3.2 Still’s binary code tag system (P)-CACATGACGGTA Mix and split into two aliquots and A unique approach by Still was to give each building block its own chemical equivalent of repeat the previous processes until the required library is obtained a binary code for each stage of the synthesis using inert aryl halides (Fig. 5.15a). One or more of these tags are directly attached to the resin using a photolabile linker at the Figure 5.13 The use of oligonucleotides to encode a peptide combinatorial synthesis for a library based on appropriate points in the synthesis. They indicate the nature of the building block and two building blocks the stage at which it was incorporated into the solid support (Table 5.3). Aryl halide tags are used because they can be detected in very small amounts by GC. They are selected on the other end by a Moz group (Fig. 5.14). The Fmoc group was cleaved under basic the basis that their retention times were roughly equally spaced (Fig. 5.15b). At the end of conditions and the suitably protected building block was joined to the linker. The Moz the synthesis all the tags are detached from the linker and are detected by GC. The gas group was removed under acid conditions and a suitably protected peptide was attached. chromatogram is read like a bar code to account for the history of the bead. Suppose, for The process was repeated for the coupling of each building block to each portion of example, that the formation of a tripeptide using six aryl halide tags allocated as shown beads as the mix and split procedure progressed. At the end of the synthesis each bead is in the tagging scheme outlined in Table 5.3 gave the tag chromatogram shown in separated from its fellows and the product and its encoding peptide were liberated from that Figure 5.15c. The presence of T1 shows that in the first stage of the synthesis the first amino acid residue is glycine. This residue will be attached via the C-terminal of the NO 2 O O O C C O (CH2 )n O Ar RES I N O Photolabile Aryl halide 0 carbonate T1 T2 T3 T4 T5 T6 tag (b) linker Cl H H Cl Cl Cl Cl Cl F Cl H H 0 Cl Cl Cl T1 T3 T5 T6 Tag retention time (a) (c) Figure 5.15 (a) Molecular tags used by Still; indicates the point at which the tag is attached to the linker. (b) A hypothetical representation of the GC plots obtained for some aryl halide tags. (c) The tag Figure 5.14 An outline of the Zuckermann approach using peptides for encoding chromatogram for a hypothetical tagging scheme  5.4 COMBINATORIAL SYNTHESIS IN SOLUTION 161 162 CH5 COMBINATORIAL CHEMISTRY Table 5.3 A hypothetical tagging scheme for the preparation of tripeptides 3. Requires especially modified reactions with high yields (>98 per cent) if multistep using binary combinations of six tags syntheses are attempted. Tag 4. Requires additional synthesis steps to attach the initial building block to and remove the Stage Glycine (Gly) Alanine (Ala) Serine (Ser) product from the support. 1 T1 T2 T1 þ T2 2 T3 T4 T3 þ T4 5. The final product is contaminated with fragments (truncated intermediates) of the 3 T5 T6 T5 þ T6 product formed by incomplete reaction at different stages and often needs additional purification steps. Many of these disadvantages are eliminated or reduced when combinatorial syntheses are peptide if a linker with an amino group was used or via its N-terminal if a linker with an carried out in solution. For example, solution phase combinatorial chemistry does not have acid group was used. The presence of T3 shows that the second residue is also glycine, to have a common functional group at the position corresponding to the one used to link the whilst the presence of T5 and T6 indicates that the third amino acid in the peptide is serine. synthesis substrate to the linker or bead. Both the linear, template and convergent synthesis routes (see sections 5.1.1 and 15.2.3) can be followed. Unmodified traditional organic reactions may be used but multistep syntheses will still require very efficient reactions. 5.3.3 Computerised tagging However, it does not require additional synthetic steps to attach the initial building block to and remove the product from a solid support. The product is not likely to be contaminated Nicolaou has devised a method of using silicon chips to record the history of a synthesis. with truncated intermediates but unwanted impurities will still need to be removed at each Silicon chips can be coded to receive and store radio signals in the form of a binary code. stage in a synthesis. Disadvantages of solution phase combinatorial chemistry are given in This code can be used as a code for the building blocks of a synthesis. The silicon chip and Table 5.1. beads are placed in a container known as a can that is porous to the reagents used in the Solution phase combinatorial chemistry can be used to produce libraries that may synthesis. Each can is closed and treated as though it were one bead in a mix and split contain single compounds or mixtures (see section 5.4.2). Their production is usually by synthesis. The cans are divided into the required number of aliquots corresponding to the parallel synthetic methods (see sections 5.4.1 and 5.4.2). The main problem in their number of building blocks used in the initial step of the synthesis. Each batch of cans is preparation is the difficulty of removing unwanted impurities at each step in the synthesis. reacted with its own building block and the chip is irradiated with the appropriate radio Consequently, many of the strategies used for the preparation of libraries using solution signal for that building block. The mix and split procedure is followed and at each step the chemistry are directed to purification of the products of each step of the synthesis (see chips in the batch are irradiated with the appropriate radio signal. At the end of the sections 5.4.3–5.4.8). This and other practical problems have often limited the solution synthesis the prepared library compound is cleaved from the chip, which is interrogated to combinatorial syntheses to short synthetic routes. determine the history of the compound synthesised on the chip. The method has the advantage of producing larger amounts of the required compounds than the normal mix and split approach because the same compound is produced on all the beads in a can. 5.4.1 Parallel synthesis in solution Combinatorial synthesis in solution using the technique of parallel synthesis (see section 5.2.2) is used to prepare libraries of single compounds. Reactions are usually, carried out in 5.4 Combinatorial synthesis in solution microwave vials and 96-well plates. They are normally relatively simple, having one or two steps. For example, in 1996 Bailey et al. produced a library of 20 2-aminothioazoles by Solid phase combinatorial synthesis has a number of inbuilt disadvantages: means of a Hantzsch synthesis. They used a 5  4 grid of glass vials (Fig 5.16). Five different substituted thioureas, one per row, were treated with four different a- 1. All the libraries have a common functional group at the position corresponding to the bromoketones. Each of the a-bromoketones, only one per row, was added to a separate one used to link the initial building block to the linker or bead. row. After reaction the products were isolated before being characterised by high- resolution MS and NMR. One of the compounds synthesised by this method was the anti- 2. Syntheses are usually carried out using the linear approach. inflammatory fanetizole.  5.4 COMBINATORIAL SYNTHESIS IN SOLUTION 163 164 CH5 COMBINATORIAL CHEMISTRY S O 4 RCOCl + R'NH2 RCONHR' + Cl 1 R 1 R S R Br 70 ºC + 3 N The acid chloride-based set: N NH2 N R 5 hours R 2 R 4 R 2 N R 3 A 1 + 1 2 3 4 (B ,B ,B ,B ,B ,B ,B ,B ,B ,B ) 5 6 7 8 9 10 Mixture 1 containing all the possible A1__B compounds. (a) α-Bromoketones (BK) A 2 + 1 2 3 4 (B ,B ,B ,B ,B ,B ,B ,B ,B ,B ) 5 6 7 8 9 10 Mixture 2 containing all the possible A2__B compounds............... BK1 BK2 BK3 BK4.............. SU1 Ph S A 5 + 1 2 3 4 (B ,B ,B ,B ,B ,B ,B ,B ,B ,B ) 5 6 7 8 9 10 Mixture 5 containing all the possible A5__B SU2 compounds. Substituted N thioureas SU3 The amine-based set: N Mixture 6 containing all the possible B1__A H Ph 1 1 2 3 4 5 (SU) B + (A ,A ,A ,A ,A ,) SU4 (c) compounds............................. SU5 (b) B10 + (A1,A2,A3,A4,A5,) Mixture 15 containing all the possible B1__A Figure 5.16 (a) The Hantzsch synthesis of 2-aminothioazoles. (b) The reaction grid. Each square corre- compounds. sponds to a glass vial. (c) Fanetizole Figure 5.17 A Schematic representation of the index approach to identifying active compounds in libraries formed in solution 5.4.2 The formation of libraries of mixtures higher than that exhibited by the individual compounds responsible for activity after they Libraries of mixtures are formed by separately reacting each of the members of a set of have been isolated from the mixture. similar compounds with the same mixture of all the members of the second set of compounds. Consider, for example, a combinatorial library of amides formed by reacting a 5.4.3 Libraries formed using monomethyl polyethylene glycol (OMe-PEG) set of five acid chlorides (A1–A5) with ten amines (B1–B10). Each of the five acid chlorides is reacted separately with an equimolar mixture of all ten amines and each of the amines is Polyethylene glycols are polymers with hydroxy groups at each end of the chain reacted with an equimolar mixture of all the acid chlorides (Fig. 5.17). This produces two (Fig. 5.18a). These polymers are soluble in both water and organic solvents, the degree of sub-libraries, one consisting of a set of five mixtures based on individual acid halides and solubility depending on the length of the polymer chain. Combinatorial syntheses in the other consisting of ten mixtures based on individual amines. This means that each solution are carried out using monomethyl polyethylene glycol (OMe-PEG-OH), which compound in the main library is prepared twice, once from the acid chloride set in the acid tends to precipitate in diethyl ether. The synthesis is started by reacting the acid group of an chloride sub-library and once from the amine set in the amine sub-library. Consequently, acidic building block to the hydroxy group of OMe-PEG. The product is precipitated by determining the most biologically active of the mixtures from the acid halide set will define adding diethyl ether and the excess reagent and other impurities are removed by washing the acyl part of the most active amide and, similarly, identifying the most biologically active of (Fig. 5.18b). The solid product is redissolved in fresh solvent and the second stage of the the amine-based set of mixtures will identify the amine residue of that amide. Libraries synthesis is carried out using a similar reaction and washing procedure. At the end of the prepared and used in this manner are often referred to as indexed libraries. They are restricted synthesis the product may be cleaved from the OMe-PEG, purified and assayed. In some to two points of diversity. The approach is most successful when used to prepare small libraries. cases the product is assayed when it is still attached to the OMe-PEG. This approach may This method of identifying the structure of the most active component of combinatorial be carried out using either the parallel synthesis or split and mix methods, the latter being libraries of mixtures assumes that inactive compounds in the mixture will not interfere with carried out while the products of a stage are in solution. the active compounds in the bioassay used for determining activity. It depends on both of the mixtures containing the active compound giving a positive result for the assay procedure. However, it is not possible to identify the active structure if one of the sets of 5.4.4 Libraries produced using dendrimers as soluble supports mixtures gives a negative result. Furthermore, complications arise if more than one mixture is found to be active. In this case all the active structures have to be synthesised and tested Dendrimers are branched oligomers (small polymers) with regular structures that have been separately. However, it is generally found that the activity of the library mixture is usually used as soluble supports in a similar fashion to OMe-PEG. The synthesis initially involves  5.4 COMBINATORIAL SYNTHESIS IN SOLUTION 165 166 CH5 COMBINATORIAL CHEMISTRY HO-CH2 CH 2 O-(CH 2 CH 2 O) n -CH2 CH 2 -OH CH 3 O-CH2 CH 2 O-(CH 2 CH 2 O) n -CH 2 CH 2 -OH X X X- A X- A X X A- X X- A (a) (b) Reaction with A, the Removed by initial building block, washing plus any necessary Impurities reagents and solvents