Advantages of using High-Resolution Structures in Drug Development SBDD has been used successfully to develop a number of drugs that are currently marketed, as shown in Table 1. The number of different diseases and enzyme families represented here underlines the versatility of SBDD in developing drugs.

Table 1. Examples of drugs developed using SBDD in the market

Inhibitor/Drug	Disease	Company(s)	Protein targeted	Enzyme Family	Reference
STI-571/Gleevec	Chronic Myeloid Leukemia	Novartis	c-Abl kinase	Tyrosine kinase
Fluoroquinolone/Ciprofloxacin	Bacterial infection	Bayer	Gyrase	ATP Hydrolase	{Wise, 1983 #7}
Saquinavir/Invirase, Ritonavir/Norvir, Indinavir/ Crixivan, Nelfinavir/Viracept, Amprenavir/Agenerase, Fosamprenavir/Lexiva,	AIDS	Roche, Abbott, Agouron, Merck, Vertex	HIV-1 Protease	Aspartylprotease	{Roberts, 1990 #4;Kaldor, 1997 #3;Kempf, 1990 #5;Thompson, 1992 #6}
Trusopt	Glaucoma	Merck	Carbonic Anhydrase	Lyase
Thymitaq	Cancer	Agouron	Thymidylate synthase	Methyl transferase
Celecoxib/Celebrex, Rofecoxib/Vioxx	Inflammation, rheumatoid arthritis	Searle, Merck	Cox-2	Oxidoreductase
AG3340/Prinomastat	Cancer	Agouron	Matrix metalloprotease	Metalloprotease
Oseltamivir phosphate/Tamiflu, Zanamivir/Relenza	Influenza	Roche	Neuraminidase	Glycosidase	{Hayden, 1996 #2}

How does SBDD work?

Several excellent reviews of SBDD are available in the literature and online (Stevens, 2003; Anderson, 2003). Beginning with Target Identification and Validation, the causative nature of a drug target (usually a protein) is proven using biological or biochemical methods. Then, X-ray crystallography is used to determine the high-resolution three-dimensional structure of the protein. During Lead Discovery, the structure of the ligand-binding site (if the location is known) is then used to computationally screen potential inhibitors. Programs such as DOCK (fit prospective inhibitor compounds into the ligand-binding site and score them based on the number of hydrogen bond or van der Waals interactions they will make in the site) are often used for computational screening. This process has been traditionally considered an alternative to high-throughput screening with combinatorial chemistry, but recent reviews indicate that the two techniques can be used synergistically. The compounds which score highly during computational screening are tested in the laboratory to determine whether they will inhibit at micromolar concentrations. If micromolar inhibitors are found, SBDD is then used during the next stage, Lead Optimization, to improve the inhibitor so that it binds at nanomolar levels. Methods to improve the inhibitor can be manual or computational, with the goals of maximizing the number of interactions between the protein and inhibitor and minimizing cross-reaction with similar enzymes. Visual inspection of the protein-inhibitor 3D structure, diagrams depicting protein-inhibitor interactions, and comparison of the interactions with known substrates or inhibitors are used to predict the changes that can be made to improve the inhibitor. The new inhibitor is then synthesized and tested, and the process is repeated until a nanomolar inhibitor is obtained.

SBDD for designing specific Protein Kinase inhibitors     

An emerging area of SBDD is in developing drugs to target protein kinases, which are the cause of many types of cancer. Until recently, it was thought that inhibitors specific for a single kinase were impossible to achieve due to the fact that the human protein kinase family consists of over 500 enzymes with very similar active sites. However, the success in SBDD design of Gleevec, which specifically targets c-Abl kinase (Figure 1), has made the prospect more attainable.

Phosphate-binding loop or P-loop (green, residues 248-253): Contains a Glycine-rich motif that binds to phosphates of ATP and is flexible in absence of substrate.

aC helix (yellow, residues 280-291): The position of this helix is an important determinant for the conformation and activity of protein kinases.   Activation Loop (orange/red,residues 381-402): The conformation of this loop, found in all protein kinases, indicates whether the kinase is active or inactive. In this conformation, the kinase is inactive. The DFG motif, which is conserved in all kinases, is shown in orange and is important for coordinating a Magnesium ion needed for catalysis.

Figure 1. The 3-dimensional structure of the catalytic domain of Abl-kinase complexed with Gleevec, a small molecule designed using SBDD and currently on the market for the treatment of chronic myeloid leukemia. Over 500 kinases are present in humans, and they all possess a similar catalytic domain (shown here) consisting of two domains with the ligand bound between them. SBDD is especially useful in designing kinase inhibitors because knowledge of the active site structure allows the design of an inhibitor that is highly complementary to it, leading to specific inhibition and reduced inhibition of anti-targets (see text for more details).

The Success of Gleevec

Specificity is crucial for a kinase inhibitor to be a successful drug. The drug Gleevec specifically inhibits Abl kinase. The leftmost portion of Gleevec (as seen in Figure 1) is positioned similarly to the ATP substrate, but Tyr 253, Leu 370, and Phe 382 form a hydrophobic cage causing the inhibitor to be displaced further from the interdomain hinge than ATP (Nagar, Bornmann et al. 2002) . The rightmost part of the compound binds between the C helix and the activation loop, making interactions that are only possible in the inactive form of the kinase. In the inactive form, the conformation of the activation loop is different than that of the Src kinases, leading to Gleevec’s specificity for Abl.

The success of Gleevec has fueled interest in small-molecule inhibitors for many different kinases (Table 2). It has been estimated that 20-30% of current pharmaceutical efforts are directed towards kinase drug targets.

Table 2. A partial list of protein kinase drug targets

Kinase*	Type	Disease	Drugs developed
p38	MAP ser/thr kinase	Rheumatoid arthritis	SCIO-469 (Phase II)
EGFR	Receptor Tyrosine Kinase	Lung cancer	Gefinitib/Iressa
EGFR	Receptor Tyrosine Kinase	Lung Cancer	Erlotinib/Tarceva (Phase II)
VEGFR-2	Receptor Tyrosine Kinase	Cancer (several types), angiogenesis	IND (Phase I,II)
Raf-1	Ser/thr kinase	Liver cancer	Bay 43-9006 (Phase II)
Protein Kinase C	Ser/thr kinase	Breast cancer	Tamoxifen (Market)
PDK1	AGC Ser/Thr Kinase	Cancer (several types)	UCN-01

*Note that in some cases the compounds have been shown to inhibit related kinases.

Kinase inhibitor development

Inhibitors such as Gleevec, which bind to the inactive form of the kinase, are part of a group of non-competitive, or allosteric, kinase inhibitors. They are attractive because they do not compete with ATP, which is present at a high concentration in the cell, allowing them to be effective even at lower binding affinities. An additional benefit of targeting the inactive form is that selectivity is easier to obtain because the conformation of inactive kinases is much less conserved than the active form. The corollary to this is that compounds that bind the inactive form of the kinase are more susceptible to resistance since the inactive form is not evolutionarily selected for. For these reasons, a new generation of kinase inhibitors are being sought that:

1. Are non-competitive for ATP (improve potency/optimize selectivity)
2. Bind to regions of the enzyme that are crucial for function (limit resistance)
3. Bind to regions which are not conserved between kinase families (optimize selectivity)

Analysis of SBDD information

SBDD information can be visualized in a number of ways to determine how to improve the inhibitor’s binding affinity. 3D visualizations with important interactions (Figures 2,3) can be viewed statically or in a viewer that allows movement of the complex. 2D representations showing the protein-inhibitor interactions, or Ligplots, show the salient features of the complex. In addition, it is useful to present information regarding overall features of the active site and comparisons to relevant complexes. An analysis of the Abl kinase/Gleevec complex revealed that the compound bound tightly to a hydrophobic “cage” in the active site that caused it to bind in a different region than other small molecule inhibitors.

Figures 2,3. A 3D representation of Gleevec bound to Abl-kinase. During SBDD static pictures such as these are inspected, as well as the use of computer programs that allow movement of the complex.

Figure 4. A Ligplot, or 2D representation of the interactions between Gleevec and Abl-kinase

References

• Anderson, A. C. (2003). “The Process of Structure-Based Drug Design.” Chemistry & Biology 10 787-797.
• Hayden, F. G., J. J. Treanor, et al. (1996). "Safety and efficacy of the neuraminidase inhibitor GG167 in experimental human influenza." Jama 275(4): 295-9.
• Kaldor, S. W., V. J. Kalish, et al. (1997). "Viracept (nelfinavir mesylate, AG1343): a potent, orally bioavailable inhibitor of HIV-1 protease." J Med Chem 40(24): 3979-85
• Kempf, D. J., D. W. Norbeck, et al. (1990). "Structure-based, C2 symmetric inhibitors of HIV protease." J Med Chem 33(10): 2687-9.
• Nagar, B., W. G. Bornmann, et al. (2002). "Crystal structures of the kinase domain of c-Abl in complex with the small molecule inhibitors PD173955 and imatinib (STI-571)." Cancer Res 62(15): 4236-43.
• Roberts, N. A., J. A. Martin, et al. (1990). "Rational design of peptide-based HIV proteinase inhibitors." Science 248(4953): 358-61.
• Stevens, R. C. (2003). "The cost and value of three-dimensional protein structure." Drug Discovery World 4: 35-48 .
• Thompson, W. J., P. M. Fitzgerald, et al. (1992). "Synthesis and antiviral activity of a series of HIV-1 protease inhibitors with functionality tethered to the P1 or P1' phenyl substituents: X-ray crystal structure assisted design." J Med Chem 35(10): 1685-701.