Drug Design: Optimizing Target Interactions PDF

4. Drug design: optimizing target interactions Once a lead compound has been discovered, it can be used as the starting point for drug design. Various properties are aimed in drug design. The eventual drug:  should have a good selectivity and level of activity for its target;  should ha...

4. Drug design: optimizing target interactions Once a lead compound has been discovered, it can be used as the starting point for drug design. Various properties are aimed in drug design. The eventual drug:  should have a good selectivity and level of activity for its target;  should have minimal side effects;  should be easily synthesized and chemically stable  should be non-toxic and have acceptable pharmacokinetic properties. During drug optimization, the pharmacodynamic and pharmacokinetic properties of the drug are tackled together. It would be irrational to spend months or even years perfecting a drug that interacts perfectly with its target, but has no chance of reaching that target because of adverse pharmacokinetic properties. 1 4.1. Structure-activity relationships After determining the structure of the lead compound, the medicinal chemist has to study its structure-activity relationships (SAR). The aim is to identify those parts of the molecule that are important to biological activity and those that are not. In case it is possible to crystallize the lead compound binding to the binding site of the target, X-ray crystallography can be used to solve the crystal structure of the complex and then modelling software can be used to identify those binding interactions that are important. 2  However, this is not possible if the target structure has not been determined or cannot be crystallized. In this case, traditional methods are used which involve synthesizing a selected number of analogues of the lead compound and then studying what effect that has on the biological activity. It is important to recognize different functional groups present in a drug and the types of intermolecular interactions that they may be involved in with the target. 3 4.1.1. Binding role of alcohols and phenols Alcohol and phenol functional groups are commonly found in drugs and they are often involved in hydrogen bonds. Hydrogen plays the role of a hydrogen bond donor (HBD) whereas oxygen plays the role of a hydrogen bond acceptor (HBA) as shown in figure 1. One, or all, of these interactions may be important in binding the drug to the binding site. Synthesizing a methyl ether or an ester analogue would be important in testing this since it is very likely that the hydrogen bonding would be disrupted in either analogue. In actual fact, hydrogen bond is lost by removing the hydrogen of the hydroxyl of the original alcohol or the phenol if that original hydrogen acted as a hydrogen bond donor. 4  If we suppose that oxygen is acting as a hydrogen bond acceptor, the extent to which it acts would be very likely diminished:  The methyl group of the ether and the alkyl group of the ester are bulky and they are very likely to hinder the close approach that was previously attainable and this is very likely to disrupt hydrogen bonding. For the case of an ester, there is also a difference between the electronic properties of an ester and an alcohol:  The lone pair of the oxygen is involved in resonance and consequently, it will be a less effective hydrogen bond acceptor. 5 6 4.1.2. Binding role of aromatic rings and alkenes Aromatic rings and alkenes are planar and hydrophobic and therefore they can interact with the hydrophobic regions of the binding site through van der Waals interactions. The activity of the equivalent saturated analogue would be worth testing since the saturated alkyl region is bulkier and cannot approach the relevant region of the binding site so closely. In the case of an aromatic ring, the ring is no longer flat, and although the axial protons of a cyclohexane can interact weakly, they also serve as buffers to keep the rest of the cyclohexane ring at a distance. Saturated analogues of alkenes may be prepared from the leads as alkenes are generally easier to reduce than aromatic rings, as for analogues of aromatic rings, they would generally need a full synthesis because it is generally difficulty to reduce aromatic rings to cyclohexane rings. 7  4.1.3. The binding role of ketones and aldehydes A ketone group is planar and it can form two hydrogen bonds with the binding site using the two lone pairs on oxygen of the carbonyl. The carbonyl group can also interact with the binding site through a dipole-dipole interaction because the group has a significant dipole moment. It is relatively easy to reduce a ketone to an alcohol and it may be possible to carry out this reaction on the lead compound. This changes the geometry of the functional group from planar to tetrahedral. Such an alteration in geometry may weaken any existing hydrogen bonds and will certainly weaken any dipole-dipole interactions, as both the magnitude and orientation of the dipole moment will be altered (Figure 2). If it was suspected that the oxygen present in the alcohol analogue might still be acting as a hydrogen bond acceptor, then the ether or ester analogue could be synthesized and studied as described above. 8  Aldehydes are less common in drugs because they are more reactive and are susceptible to metabolic oxidation to carboxylic acids. However, they could interact as ketones, and similar analogues could be studied. 9  4.1.4. Binding role of amines Amines are extremely important functional groups in drug design and they are present in many drugs. They may be involved in hydrogen bonding, either as a hydrogen bond donor or a hydrogen bond acceptor. In many cases, the amine may be protonated when it interacts with its target binding site, which means that it is ionized and cannot act as a hydrogen bond acceptor. However, it can still act as a hydrogen bond donor and will form stronger hydrogen bonds than if it was not ionized. Alternatively, a strong ionic interaction may take place with a carboxylate ion in the binding site. 10  To test whether ionic or hydrogen bonding interactions are taking place, an amide analogue could be studied. This will prevent the nitrogen acting as a hydrogen bond acceptor, as the nitrogen’s lone pair will interact with the neighbouring carbonyl group instead. This interaction also prevents protonation of the nitrogen and rules out the possibility of ionic interactions. It is relatively easy to form secondary and tertiary amides from primary and secondary amines, respectively, and it may be possible to carry out this reaction directly on the lead compound. A tertiary amide lacks the N-H group of the original secondary amine and would test whether this is involved as a hydrogen bond donor. The secondary amide formed from a primary amine still has a N-H group present, but the steric bulk of the acyl group should hinder it acting as a hydrogen bond donor. 11  Tertiary amines cannot be converted directly to amides, but if one of the alkyl groups is a methyl group, it is often possible to remove it with vinyloxycarbonyl chloride (VOC-Cl) to form a secondary amine, which could then be converted to the amide (Scheme 1). This demethylation reaction is extremely useful and has been used to good effect in the synthesis of morphine analogues. 12 4.1.5. Binding role of amides Many of the lead compounds currently studied in medicinal chemistry are peptides or polypeptides consisting of amino acids linked together by peptide or amide bonds. Amides are likely to interact with binding sites through hydrogen bonding. The carbonyl oxygen atom can act as a hydrogen bond acceptor and has the potential to form two hydrogen bonds. Both the lone pairs involved are in sp2-hydbridized orbitals which are located in the same plane as the amide group. The nitrogen cannot act as a hydrogen bond acceptor because the lone pair interacts with the neighbouring carbonyl group. Primary and secondary amides have a N-H group, which allows the possibility of this group acting as a hydrogen bond donor. 13  The most common type of amide in peptide lead compounds is the secondary amide. Suitable analogues that could be prepared to test out possible binding interactions are shown in figure 3. 14  All the analogues, apart from the primary and secondary amines, could be used to check whether the amide is acting as a hydrogen bond donor. The alkenes and amines could be tested to see whether the amide is acting as a hydrogen bond acceptor. However, there are traps for the unwary. The amide group is planar and does not rotate because of its partial double bond character. The ketone, the secondary amine, and the tertiary amine analogues have a single bond at the equivalent position which can rotate. This would alter the relative positions of any binding groups on either side of the amide group and lead to a loss of binding, even if the amide itself was not involved in binding. Therefore, a loss of activity would not necessary mean that the amide is important as a binding group. With these groups, it would only be safe to say that the amide group is not essential if activity is retained. 15  Similarly, the primary amine and carboxylic acid may be found to have no activity, but this might be due to the loss of important binding groups in one half of the molecule. These particular analogues would only be worth considering if the amide group is peripheral to the molecule (e.g. R-NHCOMe or R-CONHMe) and not part of the main skeleton. The alkene would be a particularly useful analogue to test because it is planar, cannot rotate, and cannot act as a hydrogen bond donor or hydrogen bond acceptor. However, the synthesis of this analogue may not be simple. In fact, it is likely that all the analogues described would have to be prepared using a full synthesis. Amides are relatively stable functional groups and, although several analogues described might be attainable directly from the lead compound, it is more likely that the lead compound would not survive the forcing conditions required. 16 4.1.6. Binding role of quaternary ammonium salts Quaternary ammonium salts are ionized and can interact with carboxylate groups by ionic interactions (Figure 4). Another possibility is an induced dipole interaction between the quaternary ammonium ion and any aromatic rings in the binding site. The positively charged nitrogen can distort the p electrons of the aromatic ring such that a dipole is induced, whereby the face of the ring is slightly negative and the edges are slightly positive. This allows an interaction between the slightly negative faces of the aromatic rings and the positive charge of the quaternary ammonium ion. This is also known as a p-cation interaction. 17 18  The importance of these interactions could be tested by synthesizing an analogue that has a tertiary amine group rather than the quaternary ammonium group. Of course, it is possible that such a group could ionize by becoming protonated and then interact in the same way. Converting the amine to an amide would prevent this possibility. The neurotransmitter acetylcholine has a quaternary ammonium group which is thought to bind to the binding site of its target receptor by ionic bonding and/or induced dipole interactions. 19 4.1.7. Binding role of carboxylic acids The carboxylic acid group is reasonably common in drugs. It can act as a hydrogen bond acceptor or as a hydrogen bond donor. Alternatively, it may exist as the carboxylate ion. This allows the possibility of an ionic interaction and/or a strong hydrogen bond where the carboxylate ion acts as the hydrogen bond acceptor. The carboxylate ion is also a good ligand for metal ion cofactors present in several enzymes, for example zinc metalloproteinases. In order to test the possibility of such interactions, analogues such as esters, primary amides, primary alcohols, and ketones could be synthesized and tested (Figure 5). None of these functional groups can ionize, so a loss of activity could imply that an ionic bond is important. 20 21  The primary alcohol could shed light on whether the carbonyl oxygen is involved in hydrogen bonding, whereas the ester and ketone could indicate whether the hydroxyl group of the carboxylic acid is involved in hydrogen bonding. It may be possible to synthesize the ester and amide analogues directly from the lead compound, but the reduction of a carboxylic acid to a primary alcohol requires harsher conditions and this sort of analogue would normally be prepared by a full synthesis. The ketone would also have to be prepared by a full synthesis. 22 4.1.8. Binding role of esters An ester functional group has the potential to interact with a binding site as a hydrogen bond acceptor only (Figure 6). The carbonyl oxygen is more likely to act as the hydrogen bond acceptor than the alkoxy oxygen, as it is sterically less hindered and has a greater electron density. The importance, or otherwise, of the carbonyl group could be judged by testing an equivalent ether, which would require a full synthesis. Esters are susceptible to hydrolysis in vivo by metabolic enzymes called esterases. This may pose a problem if the lead compound contains an ester that is important to binding, as it means the drug might have a short lifetime in vivo. Having said that, there are several drugs that do contain esters and are relatively stable to metabolism thanks to electronic factors that stabilize the ester or steric factors that protect it. 23  Esters that are susceptible to metabolic hydrolysis are sometimes used deliberately to mask a polar functional group, such as a carboxylic acid, alcohol, or phenol, in order to achieve better absorption from the gastrointestinal tract. Once in the blood supply, the ester is hydrolysed to release the active drug. This is known as the prodrug strategy. 24 4.1.9. Binding role of alkyl and aryl halides Alkyl halides involving chlorine, bromine, or iodine tend to be chemically reactive as the halide ion is a good leaving group. As a result, a drug containing an alkyl halide is likely to react with any nucleophilic group that it encounters and become permanently linked to that group by a covalent bond – an alkylation reaction. This poses a problem, as the drug is likely to alkylate a large variety of macromolecules which have nucleophilic groups, especially amine groups in proteins and nucleic acids. It is possible to moderate the reactivity to some extent, but selectivity is still a problem and leads to severe side effects. These drugs are, therefore, reserved for life-threatening diseases, such as cancer. Alkyl fluorides, however, are not alkylating agents because the C-F bond is strong and not easily broken. Fluorine is commonly used to replace a proton as it is approximately the same size, but has different electronic properties. It may also protect the molecule from metabolism. 25  Aryl halides do not act as alkylating agents and pose less of a problem in that respect. As the halogen substituents are electron-withdrawing groups, they affect the electron density of the aromatic ring and this may have an influence on the binding of the aromatic ring. The halogen substituents chlorine and bromine are hydrophobic in nature and may interact favourably with hydrophobic pockets in a binding site. Hydrogen bonding is not important. Although halide ions are strong hydrogen bond acceptors, halogen substituents are poor hydrogen bond acceptors. Aliphatic and aromatic analogues lacking the halogen substituent could be prepared by a full synthesis to test whether the halogen has any importance towards the activity of the lead compound. 26 4.1.10. Binding role of thiols and ethers The thiol group (S-H) is known to be a good ligand for d-block metal ions and has been incorporated into several drugs designed to inhibit enzymes containing a zinc cofactor, for example the zinc metalloproteinases. If the lead compound has a thiol group, the corresponding alcohol could be tested as a comparison. This would have a far weaker interaction with zinc. An ether group (R′OR) might act as a hydrogen bond acceptor through the oxygen atom. This could be tested by increasing the size of the neighbouring alkyl group to see whether it diminishes the ability of the group to take part in hydrogen bonding. Analogues where the oxygen is replaced with a methylene (CH2) isostere should show significantly decreased binding affinity. The oxygen atom of an aromatic ether is generally a poor hydrogen bond acceptor. 27 4.1.11. Binding role of other functional groups A wide variety of other functional groups may be present in lead compounds that have no direct binding role, but could be important in other respects. Some may influence the electronic properties of the molecule (e.g. nitro groups or nitriles). Others may restrict the shape or conformation of a molecule (e.g. alkynes). Functional groups may also act as metabolic blockers (e.g. aryl halides). 28 4.1.12. Binding role of alkyl groups and the carbon skeleton The alkyl substituents and carbon skeleton of a lead compound are hydrophobic and may bind with hydrophobic regions of the binding site through van der Waals interactions. The relevance of an alkyl substituent to binding can be determined by synthesizing an analogue which lacks the substituent. Such analogues generally have to be synthesized using a full synthesis if they are attached to the carbon skeleton of the molecule. However, if the alkyl group is attached to nitrogen or oxygen, it may be possible to remove the group from the lead compound as shown below. 29 30 4.1.13. Binding role of heterocycles A large diversity of heterocycles are found in lead compounds. Heterocycles are cyclic structures that contain one or more heteroatoms, such as oxygen, nitrogen, or sulphur. Nitrogen-containing heterocycles are particularly prevalent. The heterocycles have the potential to interact with binding sites through a variety of bonding forces.  For example, the overall heterocycle can interact through van der Waals and hydrophobic interactions, while the individual heteroatoms present in the structure could interact by hydrogen bonding or ionic bonding. As far as hydrogen bonding is concerned, there is an important directional aspect. 31  The position of the heteroatom in the ring and the orientation of the ring in the binding site can be crucial in determining whether or not a good interaction takes place. For example, adenine can take part in six hydrogen bonding interactions: three as a hydrogen bond donor and three as a hydrogen bond acceptor. The ideal directions for these interactions are shown in figure 7. 32  Van der Waals interactions are also possible to regions of the binding site above and below the plane of the ring system. If the lead compound contains a heterocyclic ring, it is worth synthesizing analogues containing a benzene ring or different heterocyclic rings to explore whether all the heteroatoms present are really necessary. A complication with heterocycles is the possibility of tautomers. Knowing the preferred tautomers of heterocycles can be important in understanding how drugs interact with their binding sites. 33 4.1.14. Isosteres Isosteres are atoms or groups of atoms which share the same valency and which have chemical or physical similarities.  For example, SH, NH2, and CH3 are isosteres of OH, whereas S, NH, and CH2 are isosteres of O. Isosteres can be used to determine whether a particular group is an important binding group or not by altering the character of the molecule in as controlled a way as possible. Replacing O with CH2, for example, makes little difference to the size of the analogue, but will have a marked effect on its polarity, electronic distribution, and bonding. Replacing OH with the larger SH may not have such an influence on the electronic character, but steric factors become more significant. Isosteric groups could be used to determine whether a particular group is involved in hydrogen bonding.  For example, replacing OH with CH3 would completely eliminate hydrogen bonding, whereas replacing OH with NH2 would not. 34  The b-blocker propranolol has an ether linkage.  Replacement of the OCH2 segment with the isosteres CH=CH, SCH2, or CH2CH2 eliminates activity, whereas replacement with NHCH2 retains activity (though reduced). These results show that the ether oxygen is important to the activity of the drug and suggests that it is involved in hydrogen bonding with the receptor. 35 4.2. Structure-activity relationships in drug optimization: strategies in drug design In the previous section, we have focused on SAR studies aimed at identifying important binding groups in a lead compound. SAR studies are also used in drug optimization, where the aim is to find analogues with better activity and selectivity.  This involves further modifications of the lead compound to identify whether these are beneficial or detrimental to activity. 36  Once the important binding groups and pharmacophore of the lead compound have been identified, it is possible to synthesize analogues that contain the same pharmacophore. But why is this necessary? If the lead compound has useful biological activity, why bother making analogues? The answer is that very few lead compounds are ideal. Most are likely to have low activity, poor selectivity, and significant side effects. They may also be difficult to synthesize, so there is an advantage in finding analogues with improved properties. We look now at strategies that can be used to optimize the interactions of a drug with its target in order to gain better activity and selectivity. 37 4.2.1. Variation of substituents Varying easily accessible substituents is a common method of fine tuning the binding interactions of a drug. a) Alkyl substituents Certain alkyl substituents can be varied more easily than others.  For example, the alkyl substituents of ethers, amines, esters, and amides are easily varied as shown in figure 8. 38 39  In these cases, the alkyl substituent already present can be removed and replaced by another substituent. Alkyl substituents which are part of the carbon skeleton of the molecule are not easily removed, and it is usually necessary to carry out a full synthesis in order to vary them. If alkyl groups are interacting with a hydrophobic pocket in the binding site, then varying the length and bulk of the alkyl group (e.g. methyl, ethyl, propyl, butyl, isopropyl, isobutyl, or t-butyl) allows one to probe the depth and width of the pocket. Choosing a substituent that will fill the pocket will then increase the binding interaction. 40  Larger alkyl groups may also confer selectivity on the drug.  For example, in the case of a compound that interacts with two different receptors, a bulkier alkyl substituent may prevent the drug from binding to one of those receptors and so cut down side effects.  For example, isoprenaline is an analogue of adrenaline where a methyl group was replaced by an isopropyl group, resulting in selectivity for adrenergic b-receptors over adrenergic a-receptors. 41 a) Aromatic substituents If a drug contains an aromatic ring, the positions of substituents can be varied to find better binding interactions, resulting in increased activity.  For example, the best anti-arrythmic activity for a series of benzopyrans was found when the sulphonamide substituent was at position 7 of the aromatic ring (Figure 9). 42  Changing the position of one substituent may have an important effect on another.  For example, an electron-withdrawing nitro group will affect the basicity of an aromatic amine more significantly if it is at the para position rather than the meta position.  At the para position, the nitro group will make the amine a weaker base and less liable to protonate. This would decrease the amine’s ability to interact with ionic binding groups in the binding site, and decrease activity. If the substitution pattern is ideal, then we can try varying the substituents themselves. Substituents have different steric, hydrophobic, and electronic properties, and so varying these properties may have an effect on binding and activity.  For example, activity might be improved by having a more electron-withdrawing substituent, in which case a chloro substituent might be tried in place of a methyl substituent. 43 4.2.2. Extension of the structure The strategy of extension involves the addition of another functional group or substituent to the lead compound in order to probe for extra binding interactions with the target. Lead compounds are capable of fitting the binding site and have the necessary functional groups to interact with some of the important binding regions present. However, it is possible that they do not interact with all the binding regions available. For example, a lead compound may bind to three binding regions in the binding site but fail to use a fourth. Therefore, why not add extra functional groups to probe for that fourth region? 44  Extension tactics are often used to find extra hydrophobic regions in a binding site by adding various alkyl or arylalkyl groups. These groups can be added to functional groups, such as alcohols, phenols, amines, and carboxylic acids should they be present in the drug, as long as this does not disrupt important binding interactions that are already present. Alternatively, they could be built into the building blocks used in the synthesis of various analogues. By the same token, substituents containing polar functional groups could be added to probe for extra hydrogen bonding or ionic interactions. 45  Extension strategies are used to strengthen the binding interactions and activity of a receptor agonist or an enzyme inhibitor, but they can also be used to convert an agonist into an antagonist. This will happen if the extra binding interactions results in a different induced fit from that required to activate the receptor. As a result, the antagonist binds to an inactive conformation of the receptor and blocks access to the endogenous agonist. The extension tactic has been used successfully to produce more active analogues of morphine. 46 4.2.3. Chain extension/contraction Some drugs have two important binding groups linked together by a chain, in which case it is possible that the chain length is not ideal for the best interaction.  Therefore, shortening or lengthening the chain length is a useful tactic to try (Figure 10). 47 4.2.4. Ring expansion/contraction If a drug has one or more rings that are important to binding, it is generally worth synthesizing analogues where one of these rings is expanded or contracted.  The principle behind this approach is much the same as varying the substitution pattern of an aromatic ring. Expanding or contracting a ring may put other rings in different positions relative to each other, and may lead to better interactions with specific regions in the binding site (Figure 11). 48 49  Varying the size of a ring can also bring substituents into a good position for binding.  For example, during the development of the anti-hypertensive agent cilazaprilat (an ACE inhibitor), the bicyclic structure I in figure 12 showed promising activity.  The important binding groups were the two carboxylate groups and the amide group.  By carrying out various ring contractions and expansions, cilazaprilat was identified as the structure having the best interaction with the binding site. 50 51 4.2.5. Ring variations A popular strategy used for compounds containing an aromatic or heteroaromatic ring is to replace the original ring with a range of other heteroaromatic rings of different ring size and heteroatom positions.  For example, several non-steroidal anti-inflammatory agents (NSAIDs) have been reported all consisting of a central ring with 1,2-biaryl substitution.  Different companies have varied the central ring to produce a range of active compounds (Figure 13). 52 53  Admittedly, a lot of these changes are merely ways of avoiding patent restrictions (‘me too’ drugs), but there can often be significant improvements in activity, as well as increased selectivity and reduced side effects (‘me-better’ drugs).  For example, the antifungal agent (I) in figure 14 acts against an enzyme present in both fungal and human cells. Replacing the imidazole ring of structure (I) with 1,2,4-triazole ring to give UK 46245 resulted in better selectivity against the fungal form of the enzyme. 54  One advantage of altering an aromatic ring to a heteroaromatic ring is that it introduces the possibility of an extra hydrogen bonding interaction with the binding site, should a suitable binding region be available (extension strategy). 55 4.2.6. Ring fusions Extending a ring by ring fusion can sometimes result in increased interactions or increased selectivity. One of the major advances in the development of the selective b-blockers was the replacement of the aromatic ring in adrenaline with a naphthalene ring system (pronethalol). 56  This resulted in a compound that was able to distinguish between two very similar receptors – the a- and b-receptors for adrenaline. One possible explanation for this could be that the b-receptor has a larger van der Waals binding area for the aromatic system than the a-receptor, and can interact more strongly with pronethalol than with adrenaline. Another possible explanation is that the naphthalene ring system is sterically too big for the a-receptor, but is just right for the b-receptor. 57 4.2.7. Isosteres and bioisosteres Isosteres have often been used in drug design to vary the character of the molecule in a rational way with respect to features such as size, polarity, electronic distribution, and bonding. Some isosteres can be used to determine the importance of size towards activity, whereas others can be used to determine the importance of electronic factors.  For example, fluorine is often used as an isostere of hydrogen as it is virtually the same size. However, it is more electronegative and can be used to vary the electronic properties of the drug without having any steric effect. 58  The presence of fluorine in place of an enzymatically labile hydrogen can also disrupt an enzymatic reaction, as C-F bonds are not easily broken.  For example, the antitumour drug 5-fluorouracil is accepted by its target enzyme because it appears little different from the normal substrate – uracil.  However, the mechanism of the enzyme-catalysed reaction is totally disrupted, as the fluorine has replaced a hydrogen which is normally lost during the enzyme mechanism. 59  Several non-classical isosteres have been used in drug design as replacements for particular functional groups. Non-classical isosteres are groups that do not obey the steric and electronic rules used to define classical isosteres, but which have similar physical and chemical properties.  For example, the structures shown in figure 15 are non-classical isosteres for a thiourea group. They are all planar groups of similar size and basicity. 60  The term bioisostere is used in drug design and includes both classical and non-classical isosteres. A bioisostere is a group that can be used to replace another group while retaining the desired biological activity. Bioisosteres are often used to replace a functional group that is important for target binding, but is problematic in one way or another.  For example, the thiourea group was present as an important binding group in early histamine antagonists, but was responsible for toxic side effects.  Replacing it with bioisosteres allowed the important binding interactions to be retained for histamine antagonism but avoided the toxicity problems. It is important to realize that bioisosteres are specific for a particular group of compounds and their target. Replacing a functional group with a bioisostere is not guaranteed to retain activity for every drug at every target. 61  In some situations, the use of a bioisostere can actually increase target interactions and/or selectivity.  For example, a pyrrole ring has frequently been used as a bioisostere for an amide.  Carrying out this replacement on the dopamine antagonist sultopride led to increased activity and selectivity towards the dopamine D3-receptor over the dopamine D2-receptor (figure 16).  Such agents show promise as antipsychotic agents that lack the side effects associated with the D2-receptor. 62 4.2.8. Simplification of the structure Simplification is a strategy which is commonly used on the often complex lead compounds arising from natural sources. Once the essential groups of such a drug have been identified by SAR, it is often possible to discard the non-essential parts of the structure without losing activity. Consideration is given to removing functional groups which are not part of the pharmacophore, simplifying the carbon skeleton (for example removing rings), and removing asymmetric centres. 63  This strategy is best carried out in small stages.  For example, consider a hypothetical natural product glipine (Figure 17).  The essential groups have been highlighted and we might aim to synthesize simplified compounds in the order shown.  These still retain the essential groups making up the pharmacophore. 64  Chiral drugs pose a particular problem.  The easiest and cheapest method of synthesizing a chiral drug is to make a racemate.  However, both enantiomers then have to be tested for their activity and side effects, doubling the number of tests that have to be carried out.  This is because different enantiomers can have different activities.  For example, compound UH-301 is inactive as a racemate, whereas its enantiomers have opposing agonist and antagonist activity at the serotonin receptor (5-HT1A).  Another notorious example is thalidomide, where one of the enantiomers is teratogenic. 65  The use of racemates is discouraged and it is preferable to use a pure enantiomer. This could be obtained by separating the enantiomers of the racemic drug or carrying out an asymmetric synthesis. Both options inevitably add to the cost of the synthesis and so designing a structure that lacks some, or all, of the asymmetric centres can be advantageous and represents a simplification of the structure.  For example, the cholesterol-lowering agent mevinolin has eight asymmetric centres, but a second generation of cholesterol-lowering agents has been developed which contain far fewer (e.g. HR 780). 66  The advantage of simpler structures is that they are easier, quicker, and cheaper to synthesize in the laboratory. 67  Usually, the complex lead compounds obtained from natural sources are impractical to synthesize and have to be extracted from the source material – a slow, tedious, and expensive business. Removing unnecessary functional groups can also be advantageous in removing side effects if these groups interact with other targets or are chemically reactive. There are, however, potential disadvantages in oversimplifying molecules. Simpler molecules are often more flexible and can sometimes bind differently to their targets compared with the original lead compound, resulting in different effects. It is best to simplify in small stages, checking that the desired activity is retained at each stage. Oversimplification may also result in reduced activity, reduced selectivity, and increased side effects. 68 4.2.9. Rigidification of the structure Rigidification has often been used to increase the activity of a drug or to reduce its side effects. In order to understand why this tactic can work, let us consider a hypothetical neurotransmitter in figure 18. 69  This is quite a simple, flexible molecule with several rotatable bonds that can lead to a large number of conformations or shapes. One of these conformations is recognized by the receptor and is known as the active conformation. The other conformations are unable to interact efficiently with the receptor and are inactive conformations. However, it is possible that a different receptor exists which is capable of binding one of these alternative conformations. If this is the case, then our model neurotransmitter could switch on two different receptors and give two different biological responses, one which is desired and one which is not. 70  The body’s own neurotransmitters are highly flexible molecules, but, fortunately, the body is efficient at releasing them close to their target receptors, then quickly inactivating them so that they do not make the journey to other receptors. This is not the case for drugs. They have to be sturdy enough to travel throughout the body and will interact with all the receptors that are prepared to accept them. The more flexible a drug molecule is, the more likely it will interact with more than one receptor and produce other biological responses (side effects). Too much flexibility is also bad for oral bioavailability. The strategy of rigidification is to make the molecule more rigid, such that the active conformation is retained and the number of other possible conformations is decreased. This should reduce the possibility of other receptor interactions and side effects. This same strategy should also increase activity. 71  By making the drug more rigid, it is more likely to be in the active conformation when it approaches the target binding site and should bind more readily. This is also important when it comes to the thermodynamics of binding. A flexible molecule has to adopt a single active conformation in order to bind to its target, which means that it has to become more ordered. This results in a decrease in entropy and, as the free energy of binding is related to entropy by the equation DG = DH – TDS, any decrease in entropy will adversely affect DG. In turn, this lowers the binding affinity (Ki), which is related to DG by the equation DG = -RTlnKi. A totally rigid molecule, however, is already in its active conformation and there is no loss of entropy involved in binding to the target. If the binding interactions (DH) are exactly the same as for the more flexible molecule, the rigid molecule will have the better overall binding affinity. 72  Incorporating the skeleton of a flexible drug into a ring is the usual way of locking a conformation – for our model compound the analogue shown in figure 19 would be suitably rigid. 73  A ring was used to rigidify the acyclic pentapeptide in figure 20.  This is a highly flexible molecule that acts as an inhibitor of a proteolytic enzyme.  It was decided to rigidify the structure by linking asparagine residue with the aromatic ring of the phenylalanine residue to form a macrocyclic ring.  The resulting structure showed a 400-fold increase in activity. 74  Locking a rotatable bond into a ring is not the only way a structure can be rigidified.  A flexible side chain can be partially rigidified by incorporating a rigid functional group such as a double bond, alkyne, amide, or aromatic ring. Rigidification also has potential disadvantages.  Rigidified structures may be more complicated to synthesize.  These is also no guarantee that rigidification will retain the active conformation; it is perfectly possible that rigidification will lock the compound into an inactive conformation.  Another disadvantage involves drugs acting on targets which are prone to mutation.  If a mutation alters the shape of the binding site, then the drug may no longer be able to bind, whereas a more flexible drug may adopt a different conformation that could bind. 75 4.2.10. Conformational blockers We have seen how rigidification tactics can restrict the number of possible conformations for a compound. Another tactic that has the same effect is the use of conformational blockers. In certain situations, a quite simple substituent can hinder the free rotation of a single bond.  For example, introducing a methyl substituent to the dopamine (D3) antagonist (I in figure 21) gives structure II and results in a dramatic reduction in affinity. 76  The explanation lies in a bad steric clash between the new methyl group and an ortho proton on the neighbouring ring which prevents both rings being in the same plane.  Free rotation around the bond between the two rings is no longer possible and so the structure adopts a conformation where the two rings are at an angle to each other.  In structure I, free rotation around the connecting bond allows the molecule to adopt a conformation where the aromatic rings are co-planar – the active conformation for the receptor.  In this case, a conformational blocker rejects the active conformation. Examples of a conformational blocker favouring the active conformation also exist. 77  Rigidification is also possible through intramolecular hydrogen bonding, which may help to stabilize particular conformations (Figure 22). 78 4.2.11. Structure-based drug design and molecular modelling So far we have discussed the traditional strategies of drug design. These were frequently carried out with no knowledge of the target structure, and the results obtained were useful in providing information about the target binding site. Clearly, if a drug has an important binding group, there must be a complementary binding region present in the binding site of the receptor or enzyme. If the macromolecular target can be isolated and crystallized, then it may be possible to determine the structure using X-ray crystallography. Unfortunately this does not reveal where the binding site is, and so it is better to crystallize the protein with a known ligand bound to the binding site. X-ray crystallography can then be used to determine the structure of the complex and this can be downloaded to a computer. Molecular modelling software is then used to identify where the ligand is and thus identify the binding site. 79  Moreover, by measuring the distances between the atoms of the ligand and neighbouring atoms in the binding site, it is possible to identify important binding interactions between the ligand and the binding site. Once this has been done, the ligand can be removed from the binding site in silico and novel lead compounds can be inserted in silico to see how well they fit. Regions in the binding site which are not occupied by the lead compound can be identified and used to guide the medicinal chemist as to what modifications and additions can be made to design a new drug that occupies more of the available space and binds more strongly. The drug can then be synthesized and tested for activity. If it proves active, the target protein can be crystallized with the new drug bound to the binding site, and then X-ray crystallography and molecular modelling can be used again to identify the structure of the complex to see if binding took place as expected. This approach is known as structure-based drug design. 80  A related process is known as de novo drug design. This involves the design of a novel drug structure, based on a knowledge of the binding site alone. This is quite a demanding exercise, but there are examples where de novo design has successfully led to a novel lead compound which can then be the starting point for structure- based drug design. Structure-based drug design cannot be used in all cases. Sometimes the target for a lead compound may not have been identified and, even if it has, it may not be possible to crystallize it. This is particularly true for membrane-bound proteins. One way round this is to identify a protein which is thought to be similar to the target protein, and which has been crystallized and studied by X-ray crystallography. The structural and mechanistic information obtained from that analogous protein can then be used to design drugs for the target protein. 81  Molecular modelling can also be used to study different compounds which are thought to interact with the same target. The structures can be compared and the important pharmacophore identified, allowing the design of novel structures containing the same pharmacophore. Compound databanks can be searched for those pharmacophores to identify novel lead compounds. There are many other applications of molecular modelling, however, a word of caution is worth making at this stage.  Molecular modelling studies tackle only one part of a much bigger problem – the design of an effective drug.  True, one might design a compound that binds perfectly to a particular enzyme or receptor in silico, but if the drug cannot be synthesized or never reaches the target protein in the body, it is useless. 82 4.2.12. The elements of luck and inspiration It is true to say that drug design has become more rational, but the role of chance or the need for hard-working, mentally alert medicinal chemists has not yet been eliminated. Most of the drugs currently on the market were developed by a mixture of rational design, trial and error, hard graft, and pure luck. There are a growing number of drugs that were developed by rational design, such as ACE inhibitors, thymidylate synthase inhibitors, HIV protease inhibitors, neuraminidase inhibitors, pralidoxime and cimetidine, but they are still in the minority. 83  Frequently, the development of drugs is helped by reading the literature to see what works on related compounds and what doesn’t, then trying out similar alterations to one’s own work. It is often a case of groping in the dark, with the chemist asking whether the addition of a group at a certain position will have a steric, electronic, or interactive effect. Even when drug design is carried out on rational lines, good fortune often has a role to play, for example the discovery of the b-blocker propranolol. Finally, there are some cases where the use of logical step-by-step modifications to a structure fails to result in significant improved activity. In such cases, there may be some advantage in synthesizing a large range of structures with different substituents or modifications in the hope of striking lucky. 84