Rational drug design — designing a molecule that binds to a target
In the second part of our science article on rational drug design, AdamTodd, Roz Anderson and Paul W. Groundwater describe how a molecule thatbinds to a target is designed, using angiotensin-converting enzyme as aworked example to illustrate the process
In the second part of our science article on rational drug design, Adam Todd, Roz Anderson and Paul W. Groundwater describe how a molecule that binds to a target is designed, using angiotensin-converting enzyme as a worked example to illustrate the process
Following on from the first article (PJ, 4 July 2009, p19), in which the role of pharmacists in drug design was discussed, the development of rational drug design software using molecular modelling techniques has helped revolutionise the development of new drugs.
Software-aiding rational drug design has been around since the early 1980s but, since it takes around 10 to 15 years to get a drug from concept to market, the results have only appeared recently.
Examples of this approach include donepezil, an anticholinesterase inhibitor, used in the treatment of Alzheimer’s disease and developed using quantitative structure activity relationship studies (QSAR),1 and indinavir and nelfinavir, both HIV-1 protease inhibitors, developed using structure-based drug design.2
There are generally two methods used in structure-based drug design to generate a lead compound — de novo design and the in silico screening of databases. De novo design involves creating a molecule from scratch to fit a given target, such as a receptor or enzyme, and this can be a challenging and time consuming process.
The in silico screening of databases, however, uses databases consisting of drug-like molecules, which are screened against a particular target. There are several large readily available commercial databases that can be downloaded free of charge from the internet.
These databases consist of thousands of drug-like compounds that can be purchased directly and this approach offers the advantage that potential lead compounds can be identified, sourced, and tested quickly.
Angiotensin-converting enzyme was chosen as the target in this article because it is well characterised and there is much information available on its binding site interactions for inhibitors. In this case, the crystal structure of ACE in complex with captopril, the first orally active ACE inhibitor, was used.3
When a compound is designed to bind to a given target, it is essential to establish which residues are required for binding, so the first step in the design process was to identify the key ACE residues for interaction with the bound ligand, in this case, captopril.
ACE is a metalloprotease enzyme whose catalytic mechanism is dependent on a zinc ion. In common with previous studies, this work shows there are six key interactions between captopril and ACE, the most significant being the favourable interaction between the thiol group of captopril and the zinc ion of ACE. These molecular interactions “lock” captopril into the active site of ACE, thereby inhibiting its activity.
In agreement with the current literature, the other residues of ACE found to interact with captopril are glutamine-A281, histidine-A353, lysine-A511, histidine-A513 and tyrosine-A520.3
Interestingly, the binding modes of the other ACE inhibitors, lisinopril and enalaprilat, have also been elucidated and indicate that captopril, lisinopril and enalaprilat all bind to ACE in a similar way, by interacting with the same six key residues, as outlined above.
Lisinopril and enalaprilat, however, interact with the essential zinc ion of ACE through a carboxylate group, rather than a thiol group, as is observed with captopril.
The pharmacophore model
The next step in the process was to develop a model from the key-interacting residues that contains important structural information about the drug target. The resulting pharmacophore is considered to be an important tool in rational drug design.
Using Catalyst, a specialist drug design software developed by Accelrys, the six residues of ACE were used to generate a pharmacophore model, which was then used as a tool to generate lead compounds with the potential to inhibit ACE.
The pharmacophore model acts as a three-dimensional map and outlines interaction sites where potential compounds can bind (Figure 1). This model was then screened against an in silico database of drug-like molecules from the Maybridge screening collection. Compounds that fitted the 3D map were categorised as hits and these were stored and selected for further investigation.
|Figure 1: A pharmacophore model generated for ACE using Catalyst. The green features represent interaction sites derived from captopril while the black features represent exclusion volumes (regions that are already occupied by ACE and therefore cannot accommodate an inhibitor)|
Since the pharmacophore model was generated from the six key molecular interactions between ACE and the bound ACE inhibitors, it would be expected that any compound fitting the model would have ACE-inhibiting properties.
Filtering of hits
Once the hit compounds were identified, they were filtered for “drugability”, usually with reference to Lipinski’s rule of five.4
Pharmacists can play an integral part in this selection process because a broad understanding of factors influencing drug bioavailability is required. In addition, some functional groups have limited aqueous solubility, while other groups have implications for toxicity, and pharmacists are uniquely able to provide information on all these important aspects.
The rule of five was developed by chemist Christopher Lipinski during a research fellowship with Pfizer. He looked at four key features of around 5,000 clinically successful drugs at that time: hydrogen bond acceptor and hydrogen bond donor ability (for receptor binding), molecular weight and log P (for ease of passive absorption from the gastrointestinal tract).
Although the rule of five is a useful guide, to which around 80 per cent of successful drugs adhere, it is not absolute and there are many examples of drugs currently in use that do not comply. Nevertheless, Lipinski’s rule of five is a useful filter to focus on the discovery process.
The rule of five is named accordingly because the cut-off point for each parameter is a multiple of five. It predicts that a compound may have poor oral bioavailability if it meets any of the following criteria:
- More than 10 hydrogen bond acceptors
- More than five hydrogen bond donors
- A molecular weight above 500
- A log P of more than five
Therefore, any hit compound that met any of Lipinski’s criteria was removed from the screen since there was a possibility that it would have poor oral bioavailability.
Once hit compounds were filtered, the remaining compounds were then inspected in silico to see how well they matched the 3D pharmacophore model (Figure 2).
|Figure 2: Compound 1 matched to the pharmacophore model of ACE. This compound appears to fit the 3D interaction map generated from captopril|
Compounds that fit the interaction sites of the pharmacophore were then chosen as candidates for biological testing in order to evaluate their activity. The worked example in Figure 2 with ACE showed that compound 1 fits the pharmacophore model and would, therefore, be a good candidate to investigate inhibitory activity against ACE.
Finally, compound 1 was compared in silico to the ACE inhibitor captopril to allow evaluation of possible binding modes (Figure 3), showing that compound 1, identified using structure-based drug design, could bind to ACE in a similar manner to captopril and supporting the idea that compound 1 may have good ACE inhibiting properties.
|Figure 3: The binding modes of captopril (yellow) and compound 1 (blue) identified using structure-based drug design. The essential catalytic zinc ion of ACE is represented as a purple sphere |
At the end of this drug design process, a set of drug-like molecules, which interact with the key residues of ACE, were identified. One of these structures, compound 1, binds to ACE in a similar way to captopril, suggesting that it may have good ACE-inhibiting properties and so is a good candidate for biological testing.
Structure-based drug design is a common method used by the pharmaceutical industry to discover a lead compound. Its use in drug discovery should help the move away from serendipity into a more structured and focused environment but, having said that, luck still certainly has an important part to play in drug discovery.
In this particular part of drug design, pharmacists’ almost unique understanding of factors influencing drug bioavailability allows them to be an integral part of the drug discovery team.
In conclusion, structure-based drug design has been used to design a compound with the potential to inhibit ACE. Discovering a lead compound is, however, only the first step in the long and complicated process of getting a drug from concept to the market.
Many more steps, such as biological testing, toxicity, pharmacokinetics, pre-formulation and formulation studies, need to be overcome before a drug moves into the clinical phase in preparation for the market, and pharmacists can also play a key role in all of these processes.
1. Kawakami Y, Inoue A, Kawai T, Wakita M, Hachiro Sugimoto H, Hopfinger AJ. The rationale for E2020 as a potent acetylcholinesterase inhibitor. Bioorganic and Medicinal Chemistry 1996;4:1429–46.
2. Kaldor SW. Viracept (nelfinavir mesylate, AG1343): a potent, orally active bioavailable inhibitor of HIV-1 protease. Journal of Medicinal Chemistry 1997;40:3979–85.
3. Natesh R, Schwager SLU, Evans HR, Sturrock ED, Acharya KR. Structural details on the binding of antihypertensive drugs captopril and enalaprilat to human testicular angiotensin I-converting enzyme. Biochemistry 2004;43:8718–24.
4. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Advanced Drug Delivery Review 1997;23:3–25.
Citation: The Pharmaceutical Journal URI: 10973553
Recommended from Pharmaceutical Press
Pharmacy Registration Assessment Questions features over 400 closed book and calculation questions. With the registration exam having gone through a complete transformation in 2016, this volume has been developed around the new General Pharmaceutical Council (GPhC) guidelines.£33.00Buy now