Page 1
WelcometoWeek6
StartingweeksixvideoPlease watch the online video (52 seconds).
OPTIONAL‐Please participate in the online discussion forum.
Chapter10‐LeadDiscovery
IntroductiontoChapter10Chapter 10 contains seven subsections.
In Vitro Screening
Fragment Based Screening
Filtering Hits Pt 1
Filtering Hits Pt 2
Filtering Hits Pt 3
Selective Optimization of Side Activities
Natural Products
Upon completing this chapter, you understand the different methods for screening libraries of
molecules for hits. You should also understand how to prioritize the hits as leads for their potential
for ultimately becoming drugs. Finally, you should realize that some leads are discovered through
non‐screening approaches.
OPTIONAL‐Please participate in the online discussion forum.
10.1InVitroScreening
FindinghitsvideoPlease watch the online video (6 minutes, 34 seconds).
A condensed summary of this video can be found in the Video summary page.
OPTIONAL‐Please participate in the online discussion forum.
Page 2
InsilicoscreeningBackground: The most common method for finding hits involves searching through libraries of
molecules using high‐throughput screening to reveal compounds with promising activity.
Instructions: Read the passage below on screening molecules through computer modeling.
Learning Goals: To understand the advantages and limitations of screening molecules through
computer simulations.
The rise in computing power, especially in the area of protein modeling, allows library compounds
to be “flown” into a binding pocket that is modeled on a computer. Binding energies can then be
estimated to approximate the Ki of each library member. The process of matching a compound to a
binding site is called docking. Estimating the binding energy is called scoring. The overall process of
testing for biological activity using a computer simulation is called in silico screening or virtual
screening.
A very attractive aspect of in silico screening is that the library does not to be real. As long as one
can draw the molecules in a computer, the computer will handle docking the molecule to the target
protein. The logistics of obtaining, maintaining, and dispensing compounds for testing are
unnecessary. The library can potentially be far larger than the one or two million compound libraries
held by major pharmaceutical companies.
While there are many attractive aspects to in silico screening, the method is still developing. One
problematic aspect is scoring. Current scoring methods are somewhat inaccurate, and many
compounds that are not strong binders are predicted to be hits. The number of false hits, or false
positives, can be reduced by using multiple different scoring methods. Only compounds that are
predicted to be active through multiple methods are selected as hits. This approach is called
consensus scoring. The activity of any virtual hits must be confirmed by synthesizing a sample of the
molecule and testing the compound in an in vitro screen.
OPTIONAL‐Please participate in the online discussion forum.
10.2FragmentBasedScreening
Fragment‐basedscreeningvideoPlease watch the online video (7 minutes, 58 seconds).
A condensed summary of this video can be found in the Video summary page.
OPTIONAL‐Please participate in the online discussion forum.
Page 3
RevisitingstromelysininhibitorsBackground: Fragment based drug discovery involves screening for small, weakly active compounds
in an assay and then tethering them together to form hit‐level compound.
Instructions: Read the passage below concerning a fragment based search for inhibitors of
stromelysin, a topic that was first presented in Chapter 9. Use the ideas in the passage to answer
the questions that follow.
Learning Goal: To understand better the types of molecules used as fragments and how hits are
generated from the fragments.
Back in Chapter 9 we discussed the development of stromelysin inhibitors to highlight the
relationship between binding energies and the structure of a molecule. As it so happens, the
stromelysin study is also an example of fragment based drug discovery.
The study started with two fragments that were found to bind to stromelysin. One was
acetohydroxamic acid (1) and the other was 4‐hydroxybiphenyl (2). Note that both are small,
fragment‐sized molecules and have weak Ki values close to 1 mM and binding energies of −2.4 and
−4.8 kcal/mol, respectively.
The fragments were then joined together. In this particular study, the binding sites for 1 and 2 were
known to be close together and the approximate orientations of the two fragments were also
known. These two details greatly helped the research team in designing tethers to connect the two
fragments. The team reported four different tethers through the addition of between one and four
CH2 units.
The discussion becomes more complicated because the activities are reported as IC50 values instead
of Ki values. Through the Cheng‐Prussoff equation, if we know the concentration of the substrate
Page 4
([S]) and Km of the substrate for stromelysin, we can convert the IC50 values to Ki values, which can in
turn be used to calculate ΔGobind of the tethered fragments.
The original stromelysin report gives [S] as 200 µM. Km is not provided, but a very similar enzyme
has a Km of 4,000 µM for stromelysin. Using these values, we can determine Ki and ΔGobind for the
best hit formed by tethering fragments. The most potent hit is compound 4.
An interesting thing about fragment binding is that Ki and ΔGobind of hit can be determined based on
the fragments that were combined. Specifically, if fragments 1 and 2 are correctly combined to
form a new hit, then Ki of the hit should be equal to the product of the Ki values of the two
fragments.
Ki (hit) = Ki (fragment 1) × Ki (fragment 2)
Similarly, ΔGobind of the hit should be equal to the sum of the ΔGo
bind of the two fragments.
ΔGobind (hit) = ΔGo
bind (fragment 1) + ΔGobind (fragment 2)
Under this logic, Ki of compound 4 should be 4.8×10−6 M (17×10−3 × 0.28×10−3 = the product of the
two fragment Ki values). Instead, the actual value is 0.30×10−6 M. ΔGobind of compound 4 should be
−7.2 kcal/mol (−2.4 + −4.8). Instead, the actual value is −8.9 kcal/mol. Compound 4 (the hit) binds
more strongly than we would predict based on its fragments.
Why do the predictions (which are theoretically sound) differ from the experimental value? The
discrepancy is the tether. Compound 4 has two more CH2 groups than the individual
fragments. These CH2 groups lie within a channel in the protein and generate binding energy
through the hydrophobic effect. In Chapter 9, the binding energy of a CH2 group through the
hydrophobic effect was listed as 0.8 kcal/mol. The energy difference between 4 and the fragments
Page 5
is 1.7 kcal/mol, essentially equal to 2 × 0.8. Once we consider the effect of the additional
CH2 groups, the strong binding of hit 4 is more reasonable.
Please complete the online exercise.
OPTIONAL‐Please participate in the online discussion forum.
10.3FilteringHitsPt1
VisualinspectionvideoPlease watch the online video (8 minutes, 33 seconds).
A condensed summary of this video can be found in the Video summary page.
OPTIONAL‐Please participate in the online discussion forum.
AnotherstructuralalertBackground: Compounds that contain functionality that may cause toxicity problems raise a
structural alert. Anilines are perhaps the most common functional group that causes structural
alert.
Instructions: Read the passage below about arylacetic acids, which also trigger a structural alert.
Learning Goal: To gain exposure to another functional group that forms reactive metabolites and is
associated with structural alerts.
Please continue on the following page…
Page 6
In addition to anilines, arylacetic acids (1) frequently form reactive metabolites. Like most carboxylic
acids, arylacetic acids often undergo phase II metabolism and are conjugated with glucuronic acid
(2). Glucuronides of arylacetic acids can rearrange from the 1‐glucuronide (3) to the 3‐
glucuronide (4). The 3‐glucuronide exists in equilibrium with its open‐chain form (5). The open‐
chain form is important because it can react with the NH2 groups on lysine residues of proteins, and
ultimately the glucuronide becomes covalently bound to the protein through a multistep
process. Modified proteins can trigger an immune response and cause tissue damage. Tissue
damage in the liver is particularly common because glucuronidation occurs primarily in the
liver. Skin rashes can also indicate immune response problems.
OPTIONAL‐Please participate in the online discussion forum.
PickingoutproblemcompoundsBackground: Functional groups in a hit can cause the hit to be less attractive to a drug discovery
program. Being able to visually identify problematic compounds can help a medicinal chemistry
group to advance the correct compounds to the lead optimization stage.
Instructions: Look at the structures below and answer the questions that follow.
Learning Goal: To practice identifying compounds that contain less desirable functional groups.
Page 7
Below are six hit‐like compounds.
Please complete the online exercise.
OPTIONAL‐Please participate in the online discussion forum.
10.4FilteringHitsPt2
MolecularindexesvideoPlease watch the online video (5 minutes 15 seconds).
A condensed summary of this video can be found in the Video summary page.
OPTIONAL‐Please participate in the online discussion forum.
LigandefficiencycalculationsBackground: Of the many molecular indexes, ligand efficiency (LE) is among the most widely used.
Instructions: Use the equation above to calculate the typical LE values of hits and
drugs. Furthermore, use the equation to calculate the LE value of two drugs, duloxetine and
sildenafil.
Learning Goal: To gain a sense of typical values for LE for different types of compounds relevant to
drug discovery.
Page 8
Please complete the online exercise.
OPTIONAL‐Please participate in the online discussion forum.
LigandlipophilicityefficiencyBackground: Molecular indexes and metrics allow members of a drug discovery team to quickly
prioritize one hit above another.
Instructions: Read the passage below on another metric, ligand lipophilicity efficiency, by which hits
may be prioritized. Use the information on ligand lipophilicity efficiency to answer the questions
that follow.
Learning Goal: To learn another molecular index by which the quality of a hit can be gauged.
A frequently encountered molecular index is ligand lipophilicity efficiency, or LLE.1 (Do not confuse
LLE with LE, the visually similar but very different ligand efficiency.) LLE is the difference between a
drug's activity in the form of −log IC50 or −log Ki and a drug's lipophilicity in the form of log P. A more
highly positive value for LLE is more favorable.
LLE acknowledges the reality that additional activity (a larger value for −log IC50.) is obtained by
adding molecular weight and likely increasing lipophilicity (a larger value for log P). According to
Lipinski's rule, log P has a maximum value of 5. Therefore, hits with a value for log P closer to 5 and
only hit‐like activity have low values for LLE leave little room in terms of lipophilicity for growth of
the molecule. LLE creates a direct relationship between activity and lipophilicity. A more active
molecule will have a larger, positive value for −log IC50.
1. Edwards, M. P.; Price, D. A. Role of Physiochemical Properties and Ligand Lipophilicity Efficiency
in Addressing Drug Safety Risks. Annu. Rep. Med. Chem. 2010, 45, 2615‐2623.
Please complete the online exercise.
OPTIONAL‐Please participate in the online discussion forum.
10.5FilteringHitsPt3
OtherleadselectioncriteriavideoPlease watch the online video (7 minutes 59 seconds).
A condensed summary of this video can be found in the Video summary page.
Page 9
OPTIONAL‐Please participate in the online discussion forum.
LeadsarebetterthanhitsBackground: Hits are often defined as compounds with moderate activity (Ki = 1 µM) against a
target. The most promising hits will be advanced as leads.
Instructions: Read the passage below on how a hit is structurally modified and improved in activity
during the lead selection process.
Learning Goal: To make clear that a hit is typically tweaked and improved before it becomes a lead.
Back in the 1970s, Saturday morning cartoons for children on ABC (a United States television
network) were broadcast with occasional educational programming called Schoolhouse Rock. The
short Schoolhouse Rock cartoons taught about topics like grammar and history. One particular
cartoon was called I'm Just a Bill and followed a congressional bill through the long and winding
approval process of becoming a federal law. Only after significant modification and editing can a bill
be approved. At the close of the cartoon, the US President approves the bill. The bill responds by
shouting, "Oh yes!"
A hit is very much like a congressional bill. A hit merits investigation because it shows activity
against a target, but that credential alone does not make the hit worthy of being a lead. The hit is
put through a battery of tests ‐ the filtering process that we discussed in sections 10.3, 10.4, and
10.5 ‐ to make sure that only the most promising compounds are advanced as leads. Once a lead, a
molecule will require more and more resources from the pharmaceutical company.
One aspect of lead discovery that we have not discussed is how the structure of a hit is explored
during the lead discovery stage. Chemists are not idle during this process, and simple structural
modifications to the hit will be made to confirm that there is room for improving the binding of the
molecule to its intended target. The outcome of these modifications is a series of molecules very
similar to the original hit. It is likely one of these modified hits that will ultimately be advanced as
the lead compound. Therefore, it is common for a lead molecule to bind more strongly to the target
than the original hit. Leads will often have Ki values of 100 nM or even lower ‐ a marked
improvement over the typical value of 1 µM of a hit.
Just as a bill, if it becomes a law, is modified and crafted as it is marched through the approval
process, so also a hit is incrementally improved as it is becoming judged to be a lead. (Oh yes!)
OPTIONAL‐Please participate in the online discussion forum.
Page 10
10.6SelectiveOptimizationofSideActivities
SOSAvideoPlease watch the online video (6 minutes 41 seconds).
A condensed summary of this video can be found in the Video summary page.
OPTIONAL‐Please participate in the online discussion forum.
Casestudy:viloxazineBackground: SOSA ‐ selective optimization of side activities ‐ is an alternative to traditional lead
discovery techniques. In SOSA, a compound that is intended to bind to one target is modified so
that the compound instead binds to a different target.
Instructions: Read the passage below about viloxazine, a drug that was discovered through the
SOSA approach.
Learning Goal: To gain exposure to another SOSA drug.
Many treatments for high blood pressure have been developed. β‐Blockers are a class of blood
pressure medication developed in the 1960s and antagonize (block) β‐receptors, which play a role in
regulating heart rate and blood vessel dilation. Propranolol (1) is among the most frequently
prescribed β‐blockers. β‐Blockers also display sedative and anticonvulsant activity.
In order to exploit the secondary sedative effects of β‐blockers, researchers modified the
pharmacophore of β‐blockers (2) by forming a new ring (3) with the flexible side chain. The
rationale was that restricting the conformation of the molecule would enhance binding to secondary
targets will minimizing binding to β‐receptors. The formation of rings is a standard approach for
controlling the conformation of a molecule. The ultimate outcome of the research program was
viloxazine (4), an antidepressant that blocks the uptake of certain neurotransmitters from synaptic
junctions.
Image credit: Pearson Education
Page 11
Viloxazine is therefore an example of SOSA drug. Viloxazine was designed by the modification of an
existing compound. The modification were specifically intended to improve side activities and
diminish binding to the original target.
OPTIONAL‐Please participate in the online discussion forum.
SOSAandmolecularindexesBackground: SOSA is potentially excellent method for discovering leads by screening established
drugs against new targets. SOSA hits are typically safe and show good cell permeability.
Instructions: Answer the questions below about how SOSA hits might fare when evaluated against
the molecular indexes discussed in this chapter.
Learning Goal: To apply the ideas of filtering hits to the active molecules identified in a SOSA based
screen.
Please complete the online exercise.
OPTIONAL‐Please participate in the online discussion forum.
10.7NaturalProducts
NaturalproductsvideoPlease watch the online video (6 minutes 50 seconds).
A condensed summary of this video can be found in the Video summary page.
OPTIONAL‐Please participate in the online discussion forum.
MorenaturalproductsBackground: Natural products are an established source for hits, leads, and drugs in a drug
discovery program.
Instructions: Read the passage below on three drugs that were developed as a result of insights
gained from related natural products.
Learning Goal: To recognize the structural similarities between certain drugs and the natural
products that resulted in their discovery.
Page 12
atorvastatin
The synthesis of cholesterol by the body requires conversion of 3‐hydroxy‐3‐methylglutaryl‐CoA (1,
HMG‐CoA) to mevalonic acid (2) by HMG‐CoA reductase. This fact was known in the early 1970s.
During the 1970s, scientists at several pharmaceutical companies investigated different fungi for
compounds that might inhibit HMG‐CoA reductase and therefore shut down cholesterol
biosynthesis.
The first marketed compound from this search was lovastatin (3), which is found in oyster
mushrooms. Once hydrolyzed in the body, the resulting acid (4) bears a clear resemblance to
mevalonic acid (2) (see boxed part of molecule). Armed with the idea of attaching a non‐polar group
to mevalonic acid as a lead compound (5), drug companies sought to develop synthetic versions of
lovastatin. The first synthetic version was atorvastatin (6), which is better known as Lipitor.
zidovudine
Just as amino acids are linked together to make proteins, nucleosides are connected to make the
biological molecules DNA and RNA. The enzymes that make DNA or RNA are called nucleotide
polymerases. Nucleosides consist of two parts. One part is a sugar, either ribose (7) or 2‐
Page 13
deoxyribose (8). The other part is a nucleobase. Nucleobases can vary in structure but consist of
either a bicyclic purine core (9) or a monocyclic pyrimidine core (10). Specific examples of
nucleosides are shown below.
Nucleosides play a vital role in certain diseases, including viral infections. A cell infected by a virus is
converted into a virus factory. One thing the cell produces is the DNA or RNA required to make a
new virus. Naturally, to make viral DNA or RNA, a cell needs a steady supply of nucleosides. A
method for treating viral infections is to dose a patient with nucleoside analogues. Nucleoside
analogues are molecules that are similar enough to natural nucleosides to serve as a substrate for
nucleotide polymerases but lack complete functionality to form active DNA or RNA. The overall
effect of nucleoside analogues is that they slow the rate of virus production by infected cells.
The first nucleoside analogue developed to treat HIV infections was zidovidine (13). Zidovudine very
closely resembles the structure of 2'‐deoxythymidine (12), a naturally occurring nucleoside in the
body. Other antiviral nucleoside analogues include aciclovir (14) and telbivudine (15). Nucleoside
analogues walk a very fine line for their activity. The compounds must resemble natural nucleosides
closely enough to bind the active site of nucleotide polymerases. The compounds must also be
different enough not to engage in the chain elongation process of making DNA or RNA.
Page 14
propranolol
β‐adrenergic receptors affect a person's heart rate and vasoconstriction. Epinephrine (16) is an
endogenous ligand for the β‐adrenergic receptors. When pharmaceuticals started researching β‐
adrenergic receptor antagonists as potential blood pressure therapies, epinephrine was a natural
starting point.
The first β‐adrenergic receptor antagonist, called a β‐blocker, was dichloroisoprenaline
(17). Dichloroisoprenaline was not sufficiently active in humans, but continued research in the field
afforded propranolol (18), the first successful β‐blocker. Both dichloroisoprenaline and propranolol
have obvious structural similarities to epinephrine.
OPTIONAL‐Please participate in the online discussion forum.