Computational Techniques in the Drug Design

Computational Techniques in the Drug Design Process

David Young
Cytoclonal Pharmaceutics Inc.

The purpose of this document is to outline the drug design process and specifically the role of computational modeling techniques. This is not meant to be a comprehensive review. It is meant to list the most important techniques currently in use.

The process of designing a new drug and bringing it to market is very complex. According to a 1997 government report, it takes 12 years and 350 million dollars for the average new drug to go from the research laboratory to patient use. Pieces of this process are often repeated to create successively better drugs for the same condition. In the case of antibiotics, drugs loose effectiveness as an immunity is built up, thus leading to a continuing "arms race". The major steps in the drug design process "from scratch" are.

1. FIND WHAT IS KNOWN

   Find out all that is known about the disease and existing or traditional remedies. It is also important to look at very similar afflictions and their known treatments.

2. DEVELOP AN ASSAY

   Develop an assay technique to test drug effectiveness. An ideal assay is one in which a compound can be added to tissue samples or micro-organism colonies and there will be a visible indication of an effective treatment. At worst, there must be a way to test the drug on a laboratory animal that is susceptible to the disease. If the only way to test the effectiveness of a trial compound is to inject an untested compound into a human subject then there is no way to proceed in finding a pharmaceutical treatment.

3. CONSIDER FINANCIAL ISSUES

   The next step is to make a financial decision about whether to proceed with the development process. The assay technique will determine the cost of testing compounds. If there are existing chemical treatments, it will be a refinement effort which saves the expense of finding lead compounds. All drugs must go through extensive testing so this is a fairly fixed cost. There may be governmental grants or tax incentives associated with certain diseases. The number of patients requiring treatment and merits of existing treatments will determine the long term profitability of producing a drug.
Steps 4 and 5 of this procedure are often performed simultaneously.

4. FIND LEAD COMPOUNDS

   Lead compounds are compounds that have some activity against a disease. These may be only marginally useful and may have severe side effects. However, the lead compounds provide a starting point for refinement of the chemical structures. Lead compounds may come from many sources, including
1. The isolation of active compounds from traditional remedies.
2. The testing of natural materials followed by an isolation effort.
3. Drugs effective against similar diseases.
4. Use of combinatorial chemistry techniques which produce large numbers of related chemical compounds. This allows testing a large number of compounds at once. When a mixture that is useful is found, a separation must be done to determine which of the related structures has some drug activity. This has been one of the most promising and rapidly growing techniques in recent years.
5. Searching chemical databases to find compounds similar to those found by the above means. This is the only part of the lead finding process that is considered to be a computational technique. There are many different measures of molecular similarity and ways of efficiently handling large databases, so this is not yet a trivial step.

5. ISOLATE THE MOLECULAR BASIS FOR THE DISEASE

   If it is known that a drug must bind to a particular spot on a particular protein or nucleotide then a drug can be tailor made to bind at that site. This is often modeled computationally using any of several different techniques. Traditionally, the primary way of determining what compounds would be tested computationally was provided by the researchers' understanding of molecular interactions. A second method is the brute force testing of large numbers of compounds from a database of available structures.

   More recently a set of techniques, called rational drug design techniques or De Novo techniques have been used. These techniques attempt to reproduce the researchers' understanding of how to choose likely compounds built into a software package that is capable of modeling a very large number of compounds in an automated way. Many different algorithms have been used for this type of testing, many of which were adapted from artificial intelligence applications. No clear standard has yet emerged in this area so it is impossible to say what is best the best technique at this time.

   These techniques have seen quite a bit of active development in recent years. Unfortunately, the complexity of biological systems makes it very difficult to determine the structures of large biomolecules. Ideally a x-ray chrystallography structure is desired, but biomolecules are very difficult to chrystalize. Another very useful technique, called "distance geometry" is to find some of the internuclear distances using NMR Nuclear Overhauser Effect experiments then find molecular geometries that have these distances. If only a protein sequence is known, there are many techniques for predicting how that protein will fold, but none has yet been shown to be 100% reliable. Even once a structure has been determined, identifying the site where a drug must bind is not a trivial task.

   The difficulty in find geometries makes it possible to bring first generation drugs to market by refinement of lead compounds without ever knowing the target site for the drug in the body. As such, these techniques are being used primarily for designing improved treatments for diseases that have already been characterized extensively.

6. REFINE DRUG ACTIVITY

   Once a number of lead compounds have been found, computational and laboratory techniques have been very successful in refining the molecular structures to give a greater drug activity and fewer side effects. This is done both in the laboratory and computationally by examining the molecular structures to determine which aspects are responsible for both the drug activity and the side effects.

   Synthetically, functional groups are removed in order to find out which must be present to give a useful drug and which are not necessary. The back bone of the structure is made more flexible or more rigid. A rigid back bone may hold the functional groups in the exact alignment necessary for the drug to bind. A flexible back bone may be necessary to allow the drug to get into the binding site. Adding bulky groups at other points on the molecule is often done in the hopes that these new groups may hinder the molecule from binding at unwanted sites which are responsible for the side effects.

   Computationally, the technique used is known as QSAR (Quantitative Structure Activity Relationships). It consists of computing every possible number that can describe a molecule then doing an enormous curve fit to find out which aspects of the molecule correlate well with the drug activity or side effect severity. This information can then be used to suggest new chemical modifications for synthesis and testing.

   Another important aspect of the molecular structure is its solubility. Whether the molecule is water soluble or readily soluble in fatty tissue will affect what part of the body it becomes concentrated in. The ability to get a drug to the correct part of the body is an important factor in its potency.

   Ideally there is a continual exchange of information between the researchers doing QSAR studies, synthesis and testing. These techniques are frequently used and often very successful since they do not rely on knowning the biological basis of the disease which can be very difficult to determine.

7. DRUG TESTING

   Once a drug has been shown to be effective by an initial assay technique, much more testing must be done before it can be given to human patients. Animal testing is the primary type of testing at this stage. The scientists doing the testing must be particularly observant of many little details since this is where unexpected side effects can be found. Another question to be answered is whether the drug will work well or poorly with other drugs. This is also where initial data necessary to determine correct dosages is obtained.

   Eventually, the compounds which are deemed suitable at this stage are sent on to clinical trials. In the clinical trials, additional side effects may be found and human dosages are determined. The typical testing process goes like this.

   1. Preclinical testing in animals and test tubes. This takes an average of 6.5 years. Only one compound in 1000 is sent on to clinical testing.
   2. Phase I clinical trials in a few human volunteers. This typically takes a year and a half. Seventy percent of the compounds are sent on to the next step. This is primarily a safety test.
   3. Phase II clinical trials in a few hundred patients. This takes two years and a third of the compounds are passed on to the next step. This is further safety testing and an initial examination of the ability of the drug to have the intended effect in humans.
   4. Phase III clinical trials in a few thousand patients. This step collects more data on safety, dosage, drug activity and side effects. About a quarter of the compounds pass this phase.
   5. An advisory panel of doctors reviews the data and makes recommendations to the FDA.
   6. FDA approval or rejection.
   7. The FDA continues to monitor drug performance long after approval has been given.

8. FORMULATION

   Before a drug can be produced, there must be a means to administer it. Ideally, a tasteless or bland tablet can be created. Alternatively, an oral liquid, intravenous injection or directly applied cream may be created.

   Tablets are created by adding other compounds to minimize stomach upset and control timed release of the drug. A tablet may also have a compound which is a matrix that helps it hold it's shape without crumbling into a powder.

   Oral liquids are often combined with strong flavors and alcohol to mask the taste of the drug and prevent throat irritation.

   A cream may have to be thickened or have a component that the skin will absorb readily.

9. PRODUCTION

   The large scale production of complex molecules can be very difficult. Compounds originally isolated from natural products may continue to be harvested. Often natural products are found in nature only in extremely small quantities necessitating a complex synthesis. One route that has been under development more recently is to have compounds produced by genetically engineered micro-organisms or plants.

   Drugs have a high value per gram. As such production techniques can be viable even though they are far more inefficient than those used by bulk chemical producers. Often all possible production techniques are researched even though only one will be put into practice. This is done so that there are no openings for competing corporations to get around a manufacturers patents by using a different technique.

   Manufacturing regulations have become much more stringent in recent years. It is now also important to determine what by-products will result from production and what environmental impact there will be. It is possible to have a case in which a less efficient manufacturing process is more profitable due to the value of side products and reduced waste disposal costs.

10. MARKETING

    If there is only one available treatment for a disease, it is only necessary to see that physicians know about it. If there are several competing treatments, there may be quite a bit of marketing done so that physicians will understand the relative merits of each.

11. NON-PERSCRIPTION SALES

    After a large amount of experience under a physicians supervision, a drug may be approved for over-the-counter sales. This is often the biggest profit making end of the pharmaceutical industry.

12. GENERIC PRODUCTION

    Once the chemical patents have expired, a drug can be produced by any manufacturer. Generic drugs are often less expensive for the consumer and yield a low profit margin for the producer. The production of generic drugs favors the most cost effective production process.

REFERENCES

A good book over all, and chapter 7 in particular, is
G. L. Patrick "An Introduction to Medicinal Chemistry" Oxford (1995)

A recent review is

L. M. Balbes, S. W. Mascarella and D. B. Boyd, in "Reviews in Computational Chemistry, Vol. 5" K. B. Lipkowitz, D. B. Boyd, Eds., VCH, 337 (1994)

An introduction to computational techniques is

G. H. Grant, W. G. Richards "Computational Chemistry" Oxford (1995)

A more detailed description of computational techniques is

A. R. Leach "Molecular Modelling Principles and Applications" Longman (1996)

L. Balbes' "Guide to Rational (Computer-aided) Drug Design" is at

gopher://www.ccl.net/00/documents/drug.design.guide

There are many links on Soaring Bear's web page at

http://ellington.pharm.arizona.edu/%7Ebear/

An introduction to structure-based techniques is

I. D. Kuntz, E. C. Meng, B. K. Shoichet Acct. Chem. Res. 27 (5), 117 (1994)

An introduction to De Novo techniques is

S. Borman Chemical and Engineering News 70 (12), 18 (1992)

There is more information about clinical testing at

http://rarediseases.info.nih.gov/ord/ct-info-patient.html

and http://rarediseases.info.nih.gov/ord/ct-about.html

An expanded version of this article will be published in "Computational Chemistry: A Practical Guide for Applying Techniques to Real World Problems" by David Young, which will be available from John Wiley & Sons in the spring of 2001.

Biocomputing and Drug Design

Proteins as Cellular Targets of Medical Drugs

The cellular targets (or receptors) of many drugs used for medical treatment are proteins. By binding to the receptor, drugs either enhance or inhibit its activity. Basically there are two major groups of receptor proteins: proteins that "float" around in the cytoplasm of the cell, and proteins that are incorporated into the cell membrane. In the latter case, a drug does not even need to enter the cell, it can bind simply to an extracellular binding site of the protein and control intracellular reactions from the outside.


Two representations of a drug or "ligand" which is bound to a specific region ("domain") of a "receptor" protein. The first image shows caves and cavities of the binding domain (visualised by the yellow grid). The ligand is shown in purple. It is buried in an internal cave of the protein. The second image displays the folding pattern of the protein and shows the structural organisation of the binding site. The protein backbone is drawn in blue, the ligand in purple.

Drug Specificity and Side Effects

An important criterion to determine the medical value of a drug is specificity: the physiological effect of the drug should be as clearly defined as possible. It has to specifically bind to the target protein in order to minimise undesired side-effects. The idea that molecules can interact in a highly specific manner has a long history in medical chemistry. A century ago, Fischer and Ehrlich already used a "lock-and-key" analogy. Undesired side-effects, however, are not always an indication for insufficient specificity of drugs as these effects might also result from a reaction of our body to the desired and therefore successful regulation of the malfunctioning biochemical process.

The Molecular Basis of Drug Specificity

On the molecular level specificity includes two more or less independent mechanisms: first the drug has to bind to its receptor site with a suitable affinity (better binding means lower doses) and second it has to either stimulate or inhibit certain movements of the receptor protein in order to regulate its activity. Both mechanisms are mediated by a variety of interactions between the drug and its receptor site. Usually tens of thousands of compounds have to be screened to find a promising new drug and only very few of these candidates will make their way through the final clinical tests. Looking for help from powerful computers seems straightforward. So how can they help?


A more detailed image of the ligand binding cavity shown in the beginning. Again, the ligand is displayed in purple, and the yellow grid shows the contour of the protein at the binding site.

"Rational" Drug Design

The input of biocomputing in drug discovery is twofold: firstly the computer may help to optimise the pharmacological profile of existing drugs by guiding the synthesis of new and "better" compounds. Secondly, as more and more structural information on possible protein targets and their biochemical role in the cell becomes available, completely new therapeutic concepts can be developed. The computer helps in both steps: to find out about possible biological functions of a protein by comparing its amino acid sequence to databases of proteins with known function, and to understand the molecular workings of a given protein structure. Understanding the biological or biochemical mechanism of a disease then often suggests the types of molecules needed for new drugs.

Techniques Applied

In all cases, the aim of using the computer for drug design is to analyse the interactions between the drug and its receptor site and to "design" molecules that give an optimal fit. The central assumption is that a good fit results from structural and chemical complementarity to the target receptor. The techniques provided by computational methods include computer graphics for visualisation and the methodology of theoretical chemistry. By means of quantum mechanics the structure of small molecules can be predicted to experimental accuracy. Statistical mechanics permits molecular motion and solvent effects to be incorporated. Basically statistical mechanics is a three-dimensional equivalent of describing the position of billiard balls using Newton's law of motion.


The ligand (purple) is shown together with possible interaction partners of the binding site (blue). Residues within 10 Angstrom (10^-10meters) of the ligand are considered as putative interaction partners.

Drug Design Using Known Receptor Structures

The best possible starting point is an X-ray crystal structure of the target site. If the molecular model of the binding site is precise enough, one can apply docking algorithms that simulate the binding of drugs to the respective receptor site. In a first step these programs create a negative image of the target site, as shown above, place the putative ligands into the site, as shown down below , and finally they evaluate the quality of the fit.

What if we don't Know the Structure of the Receptor ?

Even if the structure of the receptor site is unknown the computer may help to figure out how it might look by comparing the chemical and physical properties of drugs that are known to act at a specific site. Moreover, if the amino acid sequence of the receptor site is known, one can try to predict the structure of the unknown site. This can either be done "from scratch" (which still is rather an adventure...) or by using a known structure of a related protein as template. If about 25 to 30 % of the amino acid residues are identical in two proteins, one may assume, that the three-dimensional structure of these two proteins is very similar, see also Joelle Thonnard's contribution. The technique used for this approach is called "homology modelling": the folding pattern of the template protein is maintained and the side chain atoms of the template protein are replaced by the side chain atoms of the unknown protein. Basically, the three-dimensional structure of a protein is represented by the three-dimensional organisation of the backbone atoms. The side chain atoms, which are different for all 20 amino acids, define the specific interactions with ligands or other protein domains. Replacing the side chains while maintaining the backbone therefore allows to keep the general structure of the protein and to evaluate the specific properties of the unknown protein with respect to ligand interactions.


These two images show possible orientations and conformations of the ligand in the binding pocket (same as above) in side view (first image) and top view (second image). The different conformations are generated and evaluated using computer programs. Similar calculations are performed for a set of different molecules. Those molecules giving the highest interaction energies are considered for further experimental tests. The calculations were done using FlexX.

Future Perspectives

During the last years, we have witnessed a huge progress in the development of new methods in the field of molecular biology and computer science. This has improved the tools for rational drug design significantly. More and more new drugs are developed with the help of computer techniques. A prominent example is the design of potent HIV protease inhibitors (Science, 263, 1994, 380). The design was based on knowledge of the target structure.
The computational power is yet insufficient to simulate complex biochemical reactions or even to monitor conformational changes of proteins caused by the binding of ligands. There are, however, examples where the opposite process, the unbinding of a ligand, was monitored (Science, 271, 1996, 997). These results confirm that we are able to successfully simulate and predict the basic principles of ligand-receptor interactions.

QSAR and Drug Design

David R. Bevan

Introduction

Quantitative structure-activity relationships (QSAR) represent an attempt to correlate structural or property descriptors of compounds with activities. These physicochemical descriptors, which include parameters to account for hydrophobicity, topology, electronic properties, and steric effects, are determined empirically or, more recently, by computational methods. Activities used in QSAR include chemical measurements and biological assays. QSAR currently are being applied in many disciplines, with many pertaining to drug design and environmental risk assessment.

The Early Years

QSAR date back to the 19th century. In 1863, A.F.A. Cros at the University of Strasbourg observed that toxicity of alcohols to mammals increased as the water solubility of the alcohols decreased [1]. In the 1890's, Hans Horst Meyer of the University of Marburg and Charles Ernest Overton of the University of Zurich, working independently, noted that the toxicity of organic compounds depended on their lipophilicity[1,2].

Linear Free Energy Relationships

Little additional development of QSAR occurred until the work of Louis Hammett (1894-1987), who correlated electronic properties of organic acids and bases with their equilibrium constants and reactivity. Consider the dissociation of benzoic acid:

Hammett observed that adding substituents to the aromatic ring of benzoic acid had an orderly and quantitative effect on the dissociation constant. For example,

a nitro group in the meta position increases the dissociation constant, because the nitro group is electron-withdrawing, thereby stabilizing the negative charge that develops. Consider now the effect of a nitro group in the para position:

The equilibrium constant is even larger than for the nitro group in the meta position, indicating even greater electron-withdrawal.

Now consider the case in which an ethyl group is in the para position:

In this case, the dissociation constant is lower than for the unsubstituted compound, indicating that the ethyl group is electron-donating, thereby destabilizing the negative charge that arises upon dissociation.

Hammett also observed that substituents have a similar effect on the dissociation of other organic acids and bases. Consider the dissociation of phenylacetic acids:

Electron-withdrawal by the nitro group increases dissociation, with the effect being less for the meta than for the para substituent, just as was observed with benzoic acid. The electron-donating ethyl group decreases the equilibrium constant, as would be expected.

Data for these equilibria typically are graphed as illustrated below:

Figure 1: Example of a graph for a linear free energy relationship. K0 or K0' represent equilibrium constants for unsubstituted compounds and K or K', for substituted compounds. Values for the abscissa are calculated from the dissociation constants of unsubstituted and substituted benzoic acid. Values for the ordinate are obtained from another organic acid or base with identical patterns of substitution, in this case phenylacetic acid.

Because this relationship is linear, the following equation can be written:

where is the slope of the line. The values for the abscissa in Figure 1 are always those for benzoic acid and are given the symbol, . Therefore, we can write:

, the slope of the line, is a proportionality constant pertaining to a given equilibrium. It relates the effect of substituents on that equilibrium to the effect of those substituents on the benzoic acid equilibrium. That is, if the effect of substituents is proportionally greater than on the benzoic acid equilibrium, then > 1; if the effect is less than on the benzoic acid equilibrium, < 1. By definition, for benzoic acid is equal to 1.

is a descriptor of the substituents. The magnitude of gives the relative strength of the electron-withdrawing or -donating properties of the substituents. is positive if the substituent is electron-withdrawing and negative if it is electron-donating.

These relationships as developed by Hammett are termed linear free energy relationships. Recall the equation relating free energy to an equilibrium constant:

That is, the free energy is proportional to the logarithm of the equilibrium constant. These linear free energy relationships are termed "extrathermodynamic". Although they can be stated in terms of thermodynamic parameters, no thermodynamic principle states that the relationships should be true.

To develop a better understanding of these relationships, it is instructive to consider some values of and . Values of are provided below:

In the aniline and phenol equilibria, the hydrogen ion that is dissociating is one atom removed from the phenyl ring, whereas in the benzoic acid equilibrium it is two atoms removed. Thus, substituents are able to exert a greater effect on the dissociation in aniline and phenol than in benzoic acid and the value of > 1. In phenylacetic and phenylpropionic acids, the hydrogen ion dissociating is three and four atoms removed, respectively, from the phenyl ring. Substituents are able to exert a lesser effect on the equilibrium than on the benzoic acid equilibrium and < 1.

Some illustrative values of for substituents in the meta and para positions are given below:

By definition, for hydrogen is 0. The positive values of for the nitro group indicate that it is electron-withdrawing. In understanding the magnitudes of the values for the nitro group in meta vs. para positions, consider the mechanisms of electron withdrawal or donation. For a nitro group in the meta position, electron-withdrawal is due to an inductive effect produced by the electronegativity of the constituent atoms. If only induction were operative, one would expect the electron-withdrawing effect of a nitro group in the para position to be less than in the meta position. The larger value for a para-substituted nitro group results from the combination of both inductive and resonance effects. For chlorine, the electronegativity of the atom produces an inductive electron-withdrawing effect, with the magnitude of the effect in the para position being less than in the meta position. For chlorine, only the inductive effect is possible. The methoxy group can be electron-donating or -withdrawing, depending on the position of substitution. In the meta position, the electronegativity of the oxygen produces an inductive electron-withdrawing effect. In the para position, only a small inductive effect would be expected. Moreover, an electron-donating resonance effect occurs for the methoxy group in the para position, giving an overall electron-donating effect. Tables of values for numerous substituents have been published[3,4]. In some cases, the sigma values are generally applicable to many different equilibria. In other cases, sigma values have been derived for specific equilibria, which is particularly true when one considers sigma values for ortho substituents.

Applications of the Hammett Equation

Illustrative examples of the application of the Hammett relationship will be presented. The first is the prediction of the pKa of ionization equilibria. Recall the relationship

Therefore,

which for benzoic acid is

Consider the substituted benzoic acid

Given = 0.71 for nitro groups and = -0.13 for methyl groups, we calculate pKa=2.91, which compares favorably with the experimental value of 2.97.

The second example illustrates the applicability of Hammett's electronic descriptors in a QSAR relating the inhibition of bacterial growth by a series of sulfonamides,

where X represents various substituents [5,6]. A QSAR was developed based on the values of the substituents,

where C is the minimum concentration of compound that inhibited growth of E. coli. From this relationship, we see that electron-withdrawing substituents favor inhibition of growth.

Hansch Analysis

QSAR based on Hammett's relationship utilize electronic properties as the descriptors of structures. Difficulties were encountered when investigators attempted to apply Hammett-type relationships to biological systems, indicating that other structural descriptors were necessary.

Robert Muir, a botanist at Pomona College, was studying the biological activity of compounds that resembled indoleacetic acid and phenoxyacetic acid, which function as plant growth regulators. In attempting to correlate the structures of the compounds with their activities, he consulted his colleague in chemistry, Corwin Hansch. Using Hammett sigma parameters to account for the electronic effect of substituents did not lead to meaningful QSAR. However, Hansch recognized the importance of the lipophilicity, expressed as the octanol-water partition coefficient, on biological activity [7]. We now recognize this parameter to provide a measure of the bioavailability of compounds, which will determine, in part, the amount of the compound that gets to the target site.

Relationships were developed to correlate a structural parameter (i.e., lipophilicity) with activity. In some cases, a univariate relationship correlating structure and activity was adequate. The form of the equation is:

where C is the molar concentration of compound that produces a standard response (e.g., LD50, ED50). With other data, it was observed that correlations were improved by combining Hammett's electronic parameters and Hansch's measure of lipophilicity using an equation such as

where is the Hammett substituent parameter and pi is defined analogously to . That is,

In yet other cases, parabolic relationships between biological response and hydrophobicity were observed that could be fit by including a (log P)**2 term in the QSAR. One interpretation to account for this term is that many membranes must be traversed for compounds to get to the target site, and those with greatest hydrophobicity will become localized in the membranes they encounter initially. Thus, an optimum hydrophobicity may be found in some test systems.

QSAR are now developed using a variety of parameters as descriptors of the structural properties of molecules. Hammett sigma values are often used for electronic parameters, but quantum mechanically derived electronic parameters also may be used. Other descriptors to account for the shape, size, lipophilicity, polarizability, and other structural properties also have been devised. A QSAR database has been established at Pomona College that summarizes over 6000 datasets of biological and chemical QSAR.

Drug Design

Researchers have attempted for many years to develop drugs based on QSAR. Easy access to computational resources was not available when these efforts began, so attempts consisted primarily of statistical correlations of structural descriptors with biological activities. However, as access to high-speed computers and graphics workstations became commonplace, this field has evolved into what is often termed rational drug design or computer-assisted drug design.

We will discuss the application of QSAR to drug design, some examples of which relied primarily on statistical correlation and some, on computer-based visualization and modeling. An early example of QSAR in drug design involves a series of 1-(X-phenyl)-3,3-dialkyl triazenes.

These compounds were of interest for their anti-tumor activity, but they also were mutagenic. QSAR was applied to understand how the structure might be modified to reduce the mutagenicity without significantly decreasing the anti-tumor activity. Mutagenic activity was evaluated in the Ames test, and from those data, the following QSAR was developed:

where C is the molar concentration required to give 30 revertants per 10**8 bacteria and is a "through resonance" electronic parameter [8,9]. From the equation, it is seen that factors that favor mutagenicity are increased lipophilicity and electron-donating substituents.

Studies of the anti-tumor activity were done against L1210 leukemia in mice. From the data, the following QSAR was developed:

where C is the molar concentration of compound producing a 40% increase in life span of mice, MR is molar refractivity, which is a measure of molecular volume, and EsR is a steric parameter for the R group [10]. Based on these equations, mutagenicity is more sensitive than anti-tumor activity to the electronic effects of the substituents. Thus, electron-withdrawing substituents were examined, as illustrated in the example below:

By substituting a sulfonamide group at the para position, the anti-tumor activity was reduced 1.2-fold, whereas the mutagenicity was reduced by about 400-fold.

Computer-Assisted Design

Computer-assisted drug design (CADD), also called computer-assisted molecular design (CAMD), represents more recent applications of computers as tools in the drug design process. In considering this topic, it is important to emphasize that computers cannot substitute for a clear understanding of the system being studied. That is, a computer is only an additional tool to gain better insight into the chemistry and biology of the problem at hand.

In most current applications of CADD, attempts are made to find a ligand (the putative drug) that will interact favorably with a receptor that represents the target site. Binding of ligand to the receptor may include hydrophobic, electrostatic, and hydrogen-bonding interactions. In addition, solvation energies of the ligand and receptor site also are important because partial to complete desolvation must occur prior to binding.

This approach to CADD optimizes the fit of a ligand in a receptor site. However, optimum fit in a target site does not guarantee that the desired activity of the drug will be enhanced or that undesired side effects will be diminished. Moreover, this approach does not consider the pharmacokinetics of the drug.

The approach used in CADD is dependent upon the amount of information that is available about the ligand and receptor. Ideally, one would have 3-dimensional structural information for the receptor and the ligand-receptor complex from X-ray diffraction or NMR. The ideal is seldom realized. In the opposite extreme, one may have no experimental data to assist in building models of the ligand and receptor, in which case computational methods must be applied without the constraints that the experimental data would provide.

Based on the information that is available, one can apply either ligand-based or receptor-based molecular design methods. The ligand-based approach is applicable when the structure of the receptor site is unknown, but when a series of compounds have been identified that exert the activity of interest. To be used most effectively, one should have structurally similar compounds with high activity, with no activity, and with a range of intermediate activities. In recognition site mapping, an attempt is made to identify a pharmacophore, which is a template derived from the structures of these compounds. It is represented as a collection of functional groups in three-dimensional space that is complementary to the geometry of the receptor site.

In applying this approach, conformational analysis will be required, the extent of which will be dependent on the flexibility of the compounds under investigation. One strategy is to find the lowest energy conformers of the most rigid compounds and superimpose them. Conformational searching on the more flexible compounds is then done while applying distance constraints derived from the structures of the more rigid compounds. Ultimately, all of the structures are superimposed to generate the pharmacophore. This template may then be used to develop new compounds with functional groups in the desired positions. In applying this strategy, one must recognize that one is assuming that it is the minimum energy conformers that will bind most favorably in the receptor site. In fact, there is no a priori reason to exclude higher energy conformers as the source of activity.

The receptor-based approach to CADD applies when a reliable model of the receptor site is available, as from X-ray diffraction, NMR, or homology modeling. With the availability of the receptor site, the problem is to design ligands that will interact favorably at the site, which is a docking problem.

An Example of CADD: Carbonic anhydrase

Carbonic anhydrase catalyzes the reaction ,the hydration of some aldehydes and ketones, and the hydrolysis of alkyl and aryl esters. It is a zinc-containing enzyme of about 30,000 daltons, and the three-dimensional structure has been characterized by X-ray diffraction. Physiologically, carbonic anhydrase is involved in gastric, urinary, pancreatic, lacrimal, and cerebrospinal secretions. Inhibitors of carbonic anhydrase include aromatic and heterocyclic sulfonamides, and some of these compounds have found application as diuretics.

Both traditional QSAR and computer graphical methods have been applied to the development of sulfonamides and other compounds as inhibitors of carbonic anhydrase. For example, Hansch et al. [11]developed a QSAR based on the binding constants of 29 phenylsulfonamides to the enzyme. The equation that was derived was the following:

where K is the binding constant, I1=1 if X is meta and 0 otherwise, and, I2 = 1 if X is ortho and 0 otherwise.

The negative coefficients of I1 and I2 suggest that they account for unfavorable steric effects when substituents are in the meta or ortho positions. Binding is favored by electron-withdrawing substituents, which is consistent with the hypothesis that the ionized form of -SO2NH2 binds to the zinc in the active site of carbonic anhydrase [12].

Interactive computer graphics also was applied to understand better the interaction of carbonic anhydrase inhibitors with the enzyme as illustrated in Figure 2 [11].

Figure 2: Active site of carbonic anhydrase containing the inhibitor MTS [(4S-trans)-4-(methylamino)-5,6-dihydro-6-methyl-4H-thieno (2,3-B)thiopyran-2-sulfonamide-7,7-dioxide]. The image was prepared from the PDB file 1cin.pdb.

The active site is a cavity approximately 12 Angstroms deep with a zinc atom (magenta) near the bottom of the cavity. The active site is divided into a hydrophilic half (blue) and a hydrophobic half (red). In the complex, the inhibitor appears to be bound such that the sulfonamide moiety occupies the fourth coordination site of the zinc atom, with the other three sites being occupied by histidine residues. For subsequent discussion, note that the active site is much larger than is required to accommodate an inhibitor of this size.

Receptor-based drug design incorporates a number of molecular modeling techniques, one of which is docking. The Kuntz research group [13] applied their DOCK program to the identification of compounds that may inhibit carbonic anhydrase. Structures of two of the candidates are shown below.

These molecules are considerably larger than the arylsulfonamides that traditionally are used as carbonic anhydrase inhibitors. In fact, no arylsulfonamides were identified as potential inhibitors in this study. These results probably arise because scoring of candidates was based on the size and shape of the molecules. These large candidates can engage in a greater number of favorable interactions within the large carbonic anhydrase active site than can the smaller arylsulfonamides. More recent versions of DOCK allow scoring based on force fields, which include both van der Waals and electrostatic interactions [14]. These results with DOCK illustrate the potential for programs such as this one to search objectively for ligands than are complementary to receptor sites, thereby assisting researchers in identifying potential drugs than may be considerably different from existing drugs. As yet, the efficacy as drugs of these candidates identified by DOCK has not been demonstrated.

Applications of Other Modeling Techniques

Once potential drugs have been identified by the methods described above, other molecular modeling techniques may then be applied. For example, geometry optimization may be used to "relax" the structures and to identify low energy orientations of drugs in receptor sites. Molecular dynamics may assist in exploring the energy landscape, and free energy simulations can be used to compute the relative binding free energies of a series of putative drugs.

References

1. Borman, S. (1990) New QSAR Techniques Eyed for Environmental Assessments. Chem. Eng. News, 68: 20-23.
   
   
2. Lipnick, R.L. (1986) Charles Ernest Overton: Narcosis Studies and a Contribution to General Pharmacology. Trends Pharmacol. Sci., 7: 161-164.
   
   
3. Hansch, C., Leo, A., and Taft, R.W. (1991) A Survey of Hammett Substituent Constants and Resonance and Field Parameters. Chem. Rev., 91: 165-195.
   
   
4. Hansch, C., Leo, A., and Hoekman, D. (1995) Exploring QSAR - Hydrophobic, Electronic, and Steric Constants. American Chemical Society, Washington, D.C.
   
   
5. Seydel, J.K. (1966) Prediction of in Vitro Activity of Sulfonamides, Using Hammett Constants or Spectrophotometric Data of the Basic Amines for Calculation. Mol. Pharmacol., 2: 259-265.
   
   
6. Hansch, C. (1974) Drug Research or the Luck of the Draw. J. Chem. Ed., 51: 360-365.
   
   
7. Hansch, C. (1969) A Quantitative Approach to Biochemical Structure-Activity Relationships.Acct. Chem. Res. 2: 232-239.
   
   
8. Venger, B.H., Hansch, C., Hatheway, G.J., and Amrein, Y.U. (1979) Ames Test of 1-(X-Phenyl)-3,3-dialkyltriazenes. A Quantitative Structure-Activity Study. J. Med. Chem., 22: 473-476.
   
   
9. Hansch, C. (1984-85) The QSAR Paradigm in the Design of Less Toxic Molecules. Drug Metab. Rev., 15: 1279-1294.
   
   
10. Hatheway, G.J., Hansch, C., Kim, K.H., Milstein, S.R., Schmidt, C.L., Smith, R.N., and Quinn, F.R. (1978) Antitumor 1-(X-Aryl)-3,3-dialkyltriazenes. 1. Quantitative Structure-Activity Relationships vs. L1210 Leukemia in Mice. J. Med. Chem., 21: 563-574.
    
    
11. Hansch, C., McClarin, J., Klein, T., and Langridge, R. (1985) A Quantitative Structure-Activity Relationship and Molecular Graphics Study of Carbonic Anhydrase Inhibitors. Mol. Pharmacol.,27: 493-498.
    
    
12. Kumar, K., King, R.W., and Carey, P.R. (1974) Carbonic Anhydrase - Aromatic Sulfonamide Complexes, A Resonance Raman Study. FEBS Lett. 48: 283-287.
    
    
13. DesJarlais, R.L., Sheridan, R.P., Seibel, G.L., Dixon, J.S., Kuntz, I.D., and Venkataraghavan, R. (1988) Using Shape Complementarity as an Initial Screen in Designing Ligands for a Receptor Binding Site of Known Three-Dimensional Structure. J. Med. Chem., 31: 722-729.
    
    
14. Meng, E.C., Shoichet, B.K., and Kuntz, I.D. (1992) Automated Docking with Grid-Based Energy Evaluation. J. Comput. Chem., 13: 505-524.