Drug Target Lifetime

1491 ISSN 1756-8919 Future Med. Chem. (2011) 3(12), 1491–1501 10.4155/FMC.11.112 © 2011 Future Science Ltd
PERSPECTIVE
The pharmacologic basis of drug action almost
always involves modulation of the physiologi-
cal activity of macromolecules (e.g., enzymes,
receptors and ribosomes) by binding of drug
molecules (e.g., small organic molecules and
biologics) to these targets. Thus, pharmacology
is based on the formation of a drug–target com-
plex and, in turn, the duration of pharmacologic
effect is often dictated by the temporal persis-
tence of target occupancy by the drug. Although
drug–target interactions are commonly illus-
trated in terms of structurally static binding and
dissociation events, in which the conformation
of the drug molecule and that of its target mac-
romolecule are fixed, such a description is inad-
equate to explain the impact of conformational
dynamics on drug–target interactions. Both the
association phase of drug binding to a target,
and the subsequent dissociation of the binary
drug–target complex, are often controlled by
conformational changes, especially involving
structural changes in the immediate vicinity of
the drug-binding pocket [1–3].
Historically, the effectiveness of a drug’s
interaction with a target has been quantified by
measuring the concentration of drug required
to achieve a specific level of target occupancy
under equilibrium conditions (e.g., the K
d
or
IC
50
value). In recent years, however, there has
been increased recognition that drug–target
interactions in vivo are not defined by equilib-
rium conditions. In particular, the importance
of stabilizing the binary drug–target complex
in vivo for sustained pharmacologic effect
has been highlighted, and the term residence
time has been coined to describe the temporal
duration of the drug–target complex under dif-
ferent conditions [4,5]. In addition, a differentially
long target residence time provides a mechanism
for temporal target selectivity, hence a cogent
approach to the mitigation of off-target based
toxicity in vivo; this topic has been covered in
considerable detail in previous reports [4–6].
Residence time (t) is commonly quanti-
fied through experimental measurements of
the reciprocal of the dissociation rate con-
stant (t = 1/k
off
) or the dissociative half-life
(t
1/2
= 0.693/k
off
), and it has been argued that
the residence time provides an important metric
for compound optimization through medicinal
chemistry. The residence time of the drug–target
complex is very clearly dependent on the confor-
mational stabilization of the drug–target com-
plex; hence, conformational adaptation plays a
key role in drug binding and stabilization of the
final structure of the drug–target complex, as
will be described later.
While the concept of conformational adapt-
ation in drug–target interactions has been pre-
sented previously [1–7], it is not widely appreciated
throughout the drug-discovery and medicinal
chemistry communities. In this brief perspec-
tive, we review some aspects of conformational
adaptation in drug–target interactions as they
relate to drug-discovery efforts.
The static view of
drug–target interactions
The conventional view of drug–target inter-
actions was first formulated by H Emil Fischer,
to describe enzyme–substrate interactions and
has been dubbed the lock-and-key model [8].
Conformational adaptation in drug–target
interactions and residence time
Although drug–target interactions are commonly illustrated in terms of structurally static binding and dissociation
events, such descriptions are inadequate to explain the impact of conformational dynamics on these processes. For
high-affnity interactions, both the association and dissociation of drug molecules to and from their targets are
often controlled by conformational changes of the target. Conformational adaptation can greatly infuence the
residence time of a drug on its target (i.e., the lifetime of the binary drug–target complex); long residence time can
lead to sustained pharmacology and may also mitigate off-target toxicity. In this perspective, the kinetics of
drug–target association and dissociation reactions are explored, with particular emphasis on the impact of
conformational adaptation on drug–target residence time.
Robert A Copeland
Epizyme, Inc., 325 Vassar Street,
Cambridge, MA 02139, USA
Tel.: +1 617 500 0707
Fax: +1 617 349 0707
E-mail: [email protected]
For reprint orders, please contact [email protected]
PERSPECTIVE | Copeland
Future Med. Chem. (2011) 3(12) 1492 future science group
In this model, the target macromolecule con-
tains a binding pocket that is complementary
to the drug molecule (or other ligand), in both
steric and electronic ways, such that a network
of favorable interactions between the drug and
recognition elements within the binding pocket
is established upon binding; thus the binary
drug–target complex is stabilized relative to
the free reactants (i.e., receptor and ligand) [2,3].
This conventional view further considers the
recognition elements of the binding pocket
to be held static in the most complementary
arrangement with respect to ligand interactions.
Hence, drug association and dissociation each
occur in a single kinetic step, and the efficiency
of interaction may be quantified by familiar,
mathematically related parameters such as IC
50

values, K
d
values and DG
binding
(FIGURE 1) [3].
Although K
d
and DG
binding
are thermodynamic
constants, they can be readily related to the
kinetic rate constants for drug association and
dissociation as follows [2].
K
k
k
d
on
off
=
EQUATION 1
G RTln K RTln
k
k
binding d
on
off
= = D 6 ; @ E
EQUATION 2
For weak binding interactions (i.e., K
d
val-
ues in the µM to mM range), association and
dissociation are usually rapid, with half-lives
on the µs timescale. This rapidity of binding
and dissociation can be important for physio-
logical reactions, such as enzymes binding to
substrates and cofactors. As binding affinity
increases, however, it is often the case that the
rate of association and, especially, of dissociation
slow down to timescales of seconds, minutes and
sometimes longer; hence, these reactions may
be conveniently measured in vitro by a number
of biochemical and biophysical methods [1–4,9].
Using convenient experimental methods,
one can mix a macromolecular target (let us
refer to these universally as receptors and use
the symbol R to represent them) with a drug
or other ligand (we will use the symbol L to
universally represent these molecules) and
measure the amount of binary complex (RL)
formed as a function of time after mixing. In
most experimental approaches to measuring
receptor–ligand binding, the receptor concen-
tration is held constant at a very low, limiting
concentration relative to that of the ligand.
Under these conditions, binding of ligand to
the receptor follows pseudo-first order kinet-
ics [3] and the approach to equilibrium can
therefore be described by a pseudo-first order
rate constant k
obs
.
The value of k
obs
depends on the concentra-
tion of ligand in characteristic ways that can be
mechanistically informative [1,5]. For the simple,
static binding mechanism illustrated in FIGURE 1,
the value of k
obs
is a linear function of ligand
concentration for which the slope is equal to
the value of k
on
and the intercept is equal to the
value of k
off
[5].
There are indeed examples of drug–target
complexes for which this type of binding mea-
surement yields a linear plot of k
obs
as a func-
tion of ligand concentration. Hence, in these
cases the experimental data are consistent with
single-step binding and dissociation, and there-
fore a static drug binding pocket. These cases
are not common, however, and often it turns out
that there are conformational adjustments to
the binding pocket that attend ligand binding.
In these cases the thermodynamic stability of
the protein conformers are similar and there-
fore interconversion among conformers occurs
too rapidly to be observed in standard binding
experiments. This was the case, for example, in
studies of piperidine inhibitors of the aspartyl
protease pepsin [10].
Conformational adaptation in
drug–target interactions
Despite being commonly found in textbooks,
the static model of drug–target interactions
(see earlier) is seldom adequate to describe fully
the association and dissociation of high-affin-
ity drugs with their targets. For the majority
of drugs that bind with nanomolar or lower
K
d
values, it is common to find conformation
adjustments of the drug binding pocket that
attend complex formation [2–6,11]. This type of
conformational adaptation can result from two
+
L
R
R RL
k
on
k
off
Figure 1. Static lock-and-key model of receptor–ligand binding in which
k
on
and k
off
each occur in a single kinetic step.
k
on
: Complex association rate constant; k
off
: Complex dissociation rate constant.
Key Term
Temporal target
selectivity: Refers to the
degree of target occupancy,
relative to occupancy of
collateral off-target proteins, by
a drug as a function of time over
the course of in vivo dosing. The
temporal target selectivity can
be quantifed as the ratio of
residence times for the
off-target protein and that for
the target protein.
Conformational adaptation in drug–target interactions & residence time | PERSPECTIVE
www.future-science.com 1493
future science group
kinetic pathways for drug association that have
been referred to as the conformational selection
and induced-fit models of binding (FIGURE 2) [1].
In the conformational selection model the
receptor exists in an ensemble of conformers
in the absences of ligand; only some of these
conformers are capable of binding ligand. For
simplicity let us say that the ensemble of con-
formers is composed of only two states that are
in equilibrium with one another: a state that
is unable to bind ligand (R) and an alterna-
tive conformer that does bind ligand (R*). In
the absence of ligand, the equilibrium strongly
favors the R state over the R* state, and inter-
conversion between these states is relatively
slow. Upon addition of ligand, those receptor
molecules in the R* conformer will bind ligand
(R*L) and therefore be removed from the equi-
librium between the free forms R and R*. This
leads to a shift in the equilibrium position to
favor formation of more R*, which can then bind
more ligand until, at infinite ligand concentra-
tion, the entire system has shifted to the R*L
state (FIGURE 2). In this model the rate limiting
step in binding is assumed to be interconver-
sion between the two free forms R and R*; once
formed, R* binds ligand rapidly.
The induced-fit model (FIGURE 2) results in
the same final form of the drug–target complex,
R*L, but arrives at this state through a differ-
ent kinetic pathway. Here the unbound receptor
exists in a single conformational state, R, that is
capable of binding ligand to form an encounter
complex, RL. The recognition elements within
the binding pocket are not optimally comple-
mentary to ligand in the RL state. The act of
ligand binding causes a conformational read-
justment of the target to form a new conforma-
tion (R*L) in which optimal complementarity
between ligand and binding pocket is achieved.
In this model, ligand binding to the initial
encounter state, R, is considered rapid and the
rate-limiting step is a slow conformational tran-
sition (i.e., isomerization) from the RL state to
the final R*L state.
An important point to bear in mind is that,
for both models, each target conformer (R,
R*, RL and R*L) represents an ensemble of
C
o
n
f
o
r
m
a
t
i
o
n
a
l

s
e
l
e
c
t
i
o
n
I
n
d
u
c
e
d

f
i
t
k
iso
k
rev
+ +
R R*
R*L RL
L L
B
R
Interatomic distance
P
E
Interatomic distance
R*
P
E
Interatomic distance
Interatomic distance
P
E
R*L
P
E
RL
K
d
K
d
k
iso
k
rev
A
Figure 2. (A) Thermodynamic cycle for two-step L binding to R. The reaction scheme for the conformational selection mechanism
starts with the unliganded receptor state R (top left corner) and proceeds along the clockwise direction indicated by the arrow.
The induced-fit mechanism also starts with unliganded receptor state R, but proceeds along the trajectory indicated by the
counterclockwise arrow. In both models R and R* refer to distinct conformational states of the same receptor molecule. In the
conformational selection model, the interconversion between states R and R* is slow and occurs prior to rapid ligand binding to state
R*. In the induced-fit model, ligand binds rapidly to state R after which there is a slow conversion (i.e., receptor isomerization) to the
bound state R*L. (B) Thermodynamic cycle for two-step ligand binding to a receptor, as in (A), illustrating the changes in
conformational microstate ensembles associated with each overall state of the receptor. Note that the stability of the system is defined
by the depth of the potential energy well(s) associated with each state.
K
d
: Concentration of drug required to achieve a specific level of target occupancy under equilibrium conditions; k
iso
: Rate constant for
forward isomerization from state R to R*; k
rev
: Rate constant for the reverse isomerization from R* to R; PE: Potential energy.
PERSPECTIVE | Copeland
Future Med. Chem. (2011) 3(12) 1494 future science group
conformational microstates that may interconvert
through vibrational, rotational and translational
excursions, depending on the energy barrier to
interconversion. Hence, stabilization of a particu-
lar state, such as a ligand-bound state, depends
on populating a deep, narrow potential energy
well that creates a substantial energy barrier to
escape and thus to interconversion. This con-
cept is pictorially illustrated in FIGURE 2B. It is
also important to realize that the conformational
selection and induced-fit models merely represent
opposite poles of a continuum of conformational
adaptation mechanisms that nature may use to
promote molecular recognition between binding
partners [12,13]. A final point that is worth mention
is that the conformational selection and induced-
fit models are thermodynamically indistinguish-
able. That is, the overall binding affinity in both
models will depend on the free energy difference
between the starting and ending states of the sys-
tem. By what kinetic process the system arrives at
the final state is irrelevant from a thermodynamic
perspective (i.e., the free energy of binding is a
path-independent parameter).
The conformational selection and induced-
fit models may be experimentally distinguished
by measurements of the pseudo-first order rate
constant for approach to equilibrium as a func-
tion of ligand concentration [1–5,12]. Equations
describing the dependence of k
obs
on ligand con-
centration for the two mechanisms have been
independently derived by multiple investigators
and are well established in the biochemical lit-
erature. One finds that the quantitative value of
k
obs
varies with ligand concentration in opposing
fashions for the two mechanisms [3,5,14,15]. For
the conformational selection, mechanism k
obs

depends on ligand concentration as described by
the following equation:
k k k
K L
K
obs iso rev
d
d
= +
+
6 @
EQUATION 3
where k
iso
is the rate constant for forward isomer-
ization from state R to R* and k
rev
is the rate
constant for the reverse isomerization from R*
to R (FIGURE 2A). We can define the limits of
k
obs
at zero and infinite ligand concentrations by
inspection of EQUATION 3. When [L] is zero, k
obs

reduces to (k
iso
+ k
rev
), and when [L] is infinite,
k
obs
reduces to k
iso
. Thus the value of k
obs
decreases
curvilinearly with increasing ligand concentra-
tion from an intercept value of (k
iso
+ k
rev
) to a
final value of k
iso
at infinite ligand concentration
(FIGURE 3).
For the induced-fit model, k
obs
depends on
ligand concentration as follows:
k k
K L
L
k obs iso
d
rev =
+
+
6
6
@
@
EQUATION 4
The limits at zero and infinite concentration
from EQUATION 4 are k
rev
and (k
iso
+ k
rev
), respec-
tively. Thus, for the induced-fit model, k
obs
is a
saturable, hyperbolic function of ligand concen-
tration, increasing from an intercept value of k
rev

to a final value of (k
iso
+ k
rev
) at infinite ligand
concentration (FIGURE 3).
This type of experiment provides a clear and
unambiguous basis for defining the mechanism of
interaction that is germane to a specific drug–target
pair. In this manner, the two mechanisms of con-
formational adaptation in drug binding are readily
distinguished from one another.
Most high-affinity drugs bind to their targets
through a conformational adaptation mecha-
nism [11]. Hence, one may ask which of the two
conformational adaptation models presented
above is most germane for drugs binding to their
macromolecular targets. In fact, examples of both
mechanisms can be found in the literature, based
on the type of kinetic ana lysis just described.
TABLE 1 provides examples of drugs or drug-related
compounds for which either a conformational
selection or induced-fit model may be invoked
on the basis of kinetic data. Note that the entries
in TABLE 1 for conformational selection represent
all examples of this mechanism that are known to
the author from survey of the literature. In con-
trast, the entries for induced fit represent a sam-
pling of a much larger pool of known examples
of this mechanism.
Reviewing the information summarized in
TABLE 1, all of the examples of conformational
selection are for enzymes, mainly binding to
natural substrates or cofactors. In the over-
whelming majority of cases of a drug molecule
binding to a macromolecular target, the bind-
ing appears to conform to an induced-fit mecha-
nism. Thus, while both mechanisms appear to be
applicable to receptor-ligand binding in general,
pharmacologic modulation of targets appears
to often involve an induced-fit mechanism of
conformational adaptation.
Structural changes associated with
conformational adaptation
The difference in affinity between the states RL
and R*L can be quite significant for some drug–
target pairs; it is not uncommon to see the K
d
Conformational adaptation in drug–target interactions & residence time | PERSPECTIVE
www.future-science.com 1495
future science group
value change from micromolar to nanomolar or
picomolar during this transition [16]. A common
question raised in such cases is what are the struc-
tural alterations to the target that result in such
dramatic changes in compound affinity? Recently,
Garvey reviewed this topic and concluded that
three broad mechanisms account for the affinity
changes observed during conformational adapta-
tion [16]: protein conformational changes, cova-
lent adduct formation and compound ionization.
Of the three, protein conformational changes are
perhaps of the most interest to medicinal chemists
and others focused on lead optimization. In his
survey of the literature, Garvey found many cases
in which the conformational changes that attend
two-step binding of compounds to targets were
quite subtle in nature [16]. Garvey concluded that
most of the recognition elements that resulted in
high-affinity interactions between compound
and target are formed within the context of the
initial encounter complex (RL) and these are rein-
forced through small conformational adjustments
that lead to the final binary complex state (R*L).
The conclusions of Garvey are based largely on
comparison of crystallographic structures of the
ligand-free target with that of compound-bound
forms of the target. In fact, there is a dearth of
structural information in the literature compar-
ing different ligand-bound conformers of targets
(e.g., RL and R*L) as a basis for formulating a
structural hypothesis for explaining affinity dif-
ferences. Certainly, there are many exceptions
to the generalization that structural differences
between RL and R*L states are small. Among the
proteins kinases, for example, there are several
examples of significant loop movement and other
structural rearrangements that attend inhibitor
binding [17,18]. Likewise, in the case of the aspartyl
proteases, closure of a flap region to occlude the
enzyme active site from bulk solvent is a common
feature of substrate and inhibitor binding [19].
Structural changes that stabilize the R*L state
also lead to extended residence times for the over-
all drug–target complex. Indeed, while thermo-
dynamic affinity (i.e., K*
d
for the R*L state) and
residence time can be independent parameters,
it is often the case that the same structural ele-
ments of recognition are involved in optimization
of both. In many cases, the optimization of K*
d
is
actually achieved by inadvertent optimization of
residence time [4–6]. Hence, understanding struc-
ture–activity relationships (SARs) with a view
towards maximizing the stability of the R*L state
should be an important goal of lead optimization
activities [11].
Although, as discussed, the literature is scant on
this topic, some generalization can nevertheless be
made with respect to SARs. First, conformational
changes that attend the RL to R*L transition tend
to lead to greater occlusion of the binding pocket
from bulk solvent; hence, hydrophobic–hydro-
phobic interactions favor stabilization of the R*L
state. Often, this occlusion involves ordering of
loops and other structural elements of proteins
to form a ‘lid’ over the drug binding site [19].
Second, while similar recognition elements tend
to be engaged in the RL and R*L states (e.g.,
hydrogen bonds and salt bridges), these tend to
be strengthened in the R*L state. Finally, the R*L
state can provide additional recognition elements
for compound engagement that are not available
in the RL state. Examples of this include engage-
ment of hydrogen bond networks between bound
inhibitors and flap elements in aspartyl protease
inhibitors [2,19], ‘back pocket’ engagement by
ATP-competitive inhibitors of protein kinases [18],
and ‘side-pocket’ engagement by selective inhibi-
tors of cyclooxygenase-2 [2]. A particularly com-
mon mechanism of binding pocket occlusion for
1
2
3
4
5
6
7
8
0 20 40 60 80 100
Ligand concentration
k
o
b
s

(
s
-
1

×

1
0
3
)
Figure 3. Plot of k
obs
as a function of ligand concentration (arbitrary units)
for a receptor–ligand complex conforming to the conformational
selection mechanism (closed circles) or one conforming to the induced-fit
mechanism (open circles).
Key Term
Drug–target residence
time: Lifetime of the binary
complex between a drug
molecule and a macromolecular
target. Experimentally, the
residence time is measured as
the reciprocal of the rate
constant for drug–target
complex dissociation (1/koff).
PERSPECTIVE | Copeland
Future Med. Chem. (2011) 3(12) 1496 future science group
targets that display induce-fit inhibitor binding is
the folding (or ordering) of unstructured, flexible
loops within the polypeptide sequence, over the
solvent-exposed surface of the binding site [19]. In
this manner, proteins form ‘lids’ over the inhibi-
tor-bound pocket to block the escape trajectory of
ligands from the protein (FIGURE 4B; vide infra).
This mechanism is seen, for example, upon potent
inhibitor binding to a variety of kinases, HIV
protease, HIV integrase [20], methionine adeno- methionine adeno-
syltransferase [21], ribulose-bisphosphate carbox-
ylase [22], Hepatitis NS3 protease [23], enoyl-ACP
reductase [24] and many other protein targets.
Given the commonality of this mechanism, it
seems reasonable to suggest medicinal chemistry
efforts focused on engaging specific intermolecu-
lar interactions between drug molecules and rec-
ognition elements within such flexible loop ‘lids’
as a concrete approach to systematic optimization
of both overall target affinity and drug–target
residence time.
A retrograde induce-ft model of
drug–target complex dissociation
Regarding drug interactions with pharmaco-
logic targets, it seems clear that the induced-fit
mechanism is relevant to a large number of medi-
cally important systems. As described previously,
formation of the drug–target binary complex is
a bimolecular process that can be mediated by
the induced-fit mechanism, the conformational
selection mechanism or other mechanisms that
incorporate features of both extreme models.
Regardless of the sequence of events that lead
to drug–target binary complex formation, dura-
ble pharmacologic action is determined by the
residence time of drug occupancy on the receptor.
We [4–6] and others [25–27] have made the case
that in vivo, the duration of drug–target occu-
pancy is determined mainly by the rate of drug
dissociation (i.e., dissociative half-life and resi-
dence time [4]). Drug dissociation from the binary
drug–target complex is kinetically a unimolecular
process (i.e., the observed rate constant for the
process is dependent only on the concentration
of binary complex and not on the concentrations
of total [or free] receptor and ligand). Thus, any
conformational changes that must accompany
drug dissociation most likely occur through the
equivalent of a retrograde induced-fit mecha-
nism (i.e., operating in the reverse sequence of
conformational events leading to association).
As described above, formation of the final R*L
state likely includes conformational changes that
occlude the drug binding site (hence the drug)
Table 1. Some examples of receptor–ligand binding interactions for
which a conformational selection or induced-fit mechanism has
been demonstrated.
Target Ligand(s) Ref.
Conformational selection
Human glucokinase Glucose
[32]
Rat liver glucokinase Glucose
[33]
a-chymotrypsin Proflavin
[34]
Escherichia coli
alkaline phosphatase
2,4-dinitrophenyl phosphate
[35]
Ribonuclease T
1
Guanosine 3’-GMP
[36]
Protein kinase A PLN
1–20
[37]
Induced-fit
Cyclooxygenase-2 DuP697
NS-398
[38]
Cyclooxygenase-1 Indomethacin
[39]
Purine nucleoside
phosphorylase
DADMe-ImmH
DADMe-IMMG
[40]
Xanthine oxidase Allopurinol
[41]
Mycobacterium tuberculosis
enoyl reductase
Isoniazid
[42]
Dihydrofolate reductase Methotrexate
[43]
Hepatitis C virus NS3 protease ITMN-191
VX-950
[44,45]
HIV-1 protease Darunivir
[46]
Prostate-specific antigen Phosphoramidate peptidomimetics
[47]
Hsp90 Geldenamycin
[29]
Bacterial β-ketoacyl-acyl carrier
protein synthases
Thiolactomycin
[48]
HIV-1 Integrase Elitegravir
Raltegravir
GSK364735
[20]
Aurora B GSK1070916
[30]
AKT GSK690693
[49]
Steroid 5a-reductase Finasteride
Dutasteride
[50]
Bacterial ribosome Erythromycin
Retapamulin
[51]
HIV reverse transcriptase Efavirenz
[52]
Glu-tRNA
Gln
amidotransferase Boronate peptidomimetics
[53]
Polypeptide deformylase Actinonin
[54]
Kinesin motor protein Ispinesib
[55]
Bacterial deacetylase LpxC Ciprofloxacin
[56]
HMG CoA reductase Rosuvastin
[57]
Lipoxygenase-1 Amidrazine
[58]
Calcineurin L-732531
[59]
Xylanose ABTI
[60]
Nitric oxide synthase 1400W
[61]
BACE Statine peptidomimetic
[62]
p38 MAP kinase BIRB796
[63]
Dialkylglycine decarboxylase Aminophosphonates
[64]
Conformational adaptation in drug–target interactions & residence time | PERSPECTIVE
www.future-science.com 1497
future science group
from exposure to bulk solvent [19]. A well-known
example of such a conformational change is the
flap closing that occurs after ligand binding to the
active site of aspartyl proteases, such as the HIV
protease. It is difficult to imagine how a drug mol-
ecule could escape from such an occluded binding
pocket without first opening up an escape trajec-
tory by displacement of the occluding flap region.
Hence, a retrograde induced-fit mechanism seems
likely to be a necessary component of drug disso-
ciation in cases such as this. Thermodynamically,
an energetically equal path to ligand dissociation
is afforded by the reverse trajectory associated with
the conformational selection model, as illustrated
in FIGURE 2A. From a physical structure perspec-
tive, however, this latter dissociation path would
require the ligand to diffuse out of the binding
pocket through the occlusion barrier imposed by
the protein lid, flap, or other conformation transi-
tions resulting in the R*L state. Ligand ‘tunneling’
through proteins has been invoked to describe the
diffusion of protons and diatomic gaseous ligands
(e.g., O
2
, CO and NO) out of metalloproteins
(e.g., heme proteins) [28], but this seems a very
unlikely possibility for a large, organic compound,
such as a drug molecule.
The retrograde induced-fit mechanism requires
the conversion of the R*L complex back to the
RL complex before dissociation of the drug and
recovery of the free R state of the receptor. As
illustrated in FIGURE 4, both of these conversions
require the system to surmount a significant
energy barrier to transiently attain two sequential
transition states: R*L
‡
and RL
‡
. Once the system
has reached the RL state, it can again surmount
the R*L
‡
transition state to return to the R*L state
or surmount the RL transition state to complete
the ligand escape process. Thus, the residence
time of a drug–target complex relates directly to
the relative stabilities of the R*L and RL states,
which in turn relate directly to the depth of the
potential energy wells associated with each state of
the system and the energetic height of the accom-
panying transition states [27]. The value of the free
energy differences between the states R, RL and
R*L are experimentally measurable through a
variety of biochemical and biophysical methods,
as described previously [2]. For example, among
the drug–target pairs summarized in TABLE 1, the
ratio of K
i
/K
i
* vary from 3.5-fold to more than
2300-fold, representing differences in binding free
energy (DDG
binding
) between R*L and RL of 0.7 to
more than 4.6 kcal/mol [15]. Consideration of this
type of retrograde induced-fit mechanism pro-
vides a useful framework for drug optimization
A
Escape trajectory
∆
G
b
i
n
d
i
n
g
B
RL
‡
R
RL
R*L
R*L
‡
Escape trajectory
∆
G
b
i
n
d
i
n
g
R*L
RL
R
Figure 4. Reaction coordinate diagram of the escape trajectory for ligand
dissociation following a retrograde induced-fit mechanism. (A) Free-energy
reaction coordinate diagram. The system starts off in the R*L state and must
overcome the energy barrier to attain the first transition state R*L
‡
. From there, the
system decays to the intermediate state RL. The system next must overcome another
energy barrier to attain a second transition state, RL
‡
, before decaying to the final,
unliganded form of the receptor. (B) Representation of the reaction coordinate
diagram of the escape trajectory for ligand dissociation following a retrograde
induced-fit mechanism, illustrating the conformational changes required to open up
an escape trajectory to bulk solvent for the bound ligand.
PERSPECTIVE | Copeland
Future Med. Chem. (2011) 3(12) 1498 future science group
activities. Thus, improvements in overall com-
pound affinity and residence time may, in some
cases, be achieved by optimization of compound
interactions with both states RL and R*L. In
other cases, destabilization of the RL state that is
accompanied by stabilization of the R*L state may
be most optimal for prolonged residence time.
Ultimately, it is the overall stabilization of the
R*L state that has the greatest impact on affinity
and residence time.
Future perspective
There is growing appreciation for the importance
of understanding the kinetics of drug interactions
with their macromolecular targets. In particular,
the drug-discovery community has begun to
consider drug–target residence time as an impor-
tant factor for sustained pharmacologic impact
in patients. Hence, there is growing interest in
measuring dissociation rates of compound–target
complexes during lead optimization activities, to
identify clinical candidates that may demonstrate
long residence time in vivo. This is a significant
change from the exclusive reliance on thermo-
dynamic parameters (e.g., IC
50
) that has domi-
nated drug-discovery efforts for much of the 19th
and 20th Centuries. Yet, treating residence time as
a phenomenological measurement is unsatisfying
in the context of hypothesis-driven SAR. Hence,
medicinal chemists rightly ask questions about the
elements of molecular recognition that bear on
prolonged residence time and how these recogni-
tion elements may be most optimally engaged by
small-molecule drugs.
The key theme of this article has been that
recognition elements are not static, and that con-
formational adaptation is an important aspect of
drug–target interactions that must be considered
carefully during lead optimization. We have seen
how conformational adjustments can lead to
changes in drug–target affinity that contribute
directly to prolongation of residence time. We have
introduced the concept of a retrograde induced-fit
mechanism for drug dissociation in the common
situation of two-step, conformationally gated
interactions between drugs and their targets. This
concept highlights the importance of conforma-
tional adaptation for enhanced residence time and
the need to take this into consideration in drug
discovery. While not stated explicitly above, it is
clear that failure to properly consider the role of
binding kinetics and conformational adaptation
in the evaluation of drug–target interactions can
lead to significant errors in SAR that can mislead
medicinal chemistry efforts. For example, failure
to account for slow compound association and/
or dissociation during binding assays can grossly
underestimate the affinity and residence time
of a compound [2]. An excellent example of this
is provided by the evaluation of Hsp90 inhibi-
tors, such as geldenamycin [29]. For some time
researchers were puzzled by the low affinity of
such compounds, determined by in vitro Hsp90
binding assays (IC
50
~1 µM), when contrasted
to the nanomolar effects of such compounds in
cellular assays. This apparent discordance was
resolved by Gooljarsingh et al. by carefully mea-
suring the time required to reach full equilibrium
in the binding assays [29]. Geldenamycin and
similar compounds turn out to be slow binding
and very slow dissociating compounds with nano-
molar affinity for Hsp90. The true affinity was
not previously realized because the binding assays
failed to account fully for the kinetics of com-
pound association and dissociation. Surely, there
are many other unknown examples of such mis-
informed SAR due to a failure to properly account
for the kinetics of drug–target interactions.
The residence time concept is now fairly well
established within the medicinal chemistry and
pharmacology communities. In many, but cer-
tainly not all cases, prolonged residence time is
seen as a cogent mechanistic underpinning for
durable pharmacology and mitigation of off-target
mediated toxicity for drugs in vivo (see [4–6] how-
ever, for examples where long residence time is
contraindicated). What remains to be developed
over the next 5 to 10 years, is a detailed under- next 5 to 10 years, is a detailed under-
standing of the structural determinants that medi-
ate prolonged drug–target residence time. In this
article we have made the general statement that
longer residence time is facilitated by stabiliza-
tion of the more closed, solvent occluded R*L state
(see earlier), within the context of the retrograde
induced-fit mechanism of drug–target dissocia-
tion. We have further suggested that stabilization
of the R*L state might be optimized by engage-
ment of recognition elements within flexible loops
of the target macromolecule, that form lid-like
gates to compound exodus. Yet, these generaliza-
tions provide little direction to medicinal chemists
in their efforts to optimize residence time.
Overall, this article should be viewed as a call-
to-action for the medicinal chemistry, biochemis-
try and structural biology communities. As these
topics have not yet received the experimental
efforts that they deserve, we have not been able
to address the questions of residence time SAR in
any systemic fashion. This remains a challenge
for the drug-discovery community to address in
Conformational adaptation in drug–target interactions & residence time | PERSPECTIVE
www.future-science.com 1499
future science group
Executive summary
„
Conformational dynamics of target macromolecules significantly affect the binding and dissociation of drug molecules.
„
Drug–target residence time relates to the lifetime of a drug–target complex.
„
It is the lifetime of the drug–target complex, rather than the affinity, that determines the duration of pharmacologic effect of drugs
in vivo. Long target residence time can also provide a mechanism for mitigating off-target mediated toxicity by limiting systemic
exposure of drugs.
„
Many potent drugs bind their targets through a two-step, induced-fit mechanism.
„
Drug–target dissociation is often mediated by a retrograde induced-fit mechanism, requiring surmounting of multiple energy barriers
along the drug escape trajectory.
„
Stabilization of the final drug–target complex state provides a mechanism for prolonging drug–target residence time.
a prospective way. Recently, a number of more
detailed reports of residence time SAR, coupled
with crystallographic data, have been submitted to
various journals for publication. This is an encour-
aging development, and it is hoped that more such
reports will begin to appear in the scientific litera-
ture. Basic questions also remain to be answered
regarding the relative contributions of enthalpic
and entropic forces in drug binding and release,
and in surmounting the multiple transition states
associated with retrograde induced-fit drug disso-
ciation. Similarly, the question of how heat capac-
ity differences among conformational states of the
target might influence drug dissociation remains
to be addressed systematically. Again, some pre-
liminary reports, based largely on calorimetric
studies, are beginning to appear in the literature
to address these questions. A final area for future
exploration is the influence of auxiliary proteins
and other intracellular binding partners of targets,
on drug–target residence time. It is reasonable to
consider that binding partners could influence the
conformational dynamics of targets and in this
way indirectly influence drug residence time. For
example, in a recent set of studies, Anderson et al.
demonstrated that substrate and inhibitor bind-
ing to Aurora kinases could be significantly influ-
enced by the presents of auxiliary proteins, such
as TPX2 and INCEPE [30,31].
As additional work in all of these areas con-
tinues and begins to populate the literature, it
is hoped that more definitive answers to these
important questions may soon emerge. As more
clear descriptions of the structural determinants
of drug–target residence time arise, the pharma-
ceutical community will be in a much better posi-
tion to fully exploit the residence time concept
for improved development of safe, long-lasting
therapies against currently unmet medical needs.
Acknowledgements
I wish to thank Roderick Hubbard, of the University of
York, and Roman Hillig, of Bayer Schering Pharma AG,
for encouraging me to write this article. I also wish to
thank my colleagues at Epizyme, especially Mikel Moyer,
Richard Chesworth, Robert Gould and Jason Rhodes for
helpful suggestions and Caroline Hill for help in preparing
the manuscript.
Financial & competing interests disclosure
The author is an employee and stockholder of Epizyme, Inc.
The author has no other relevant affiliations or financial
involvement with any organization or entity with a finan-
cial interest in or financial conflict with the subject matter
or materials discussed in the manuscript apart from
those disclosed.
No writing assistance was utilized in the production of
this manuscript.
Bibliography
1 Morrison JF, Walsh CT. The behavior and
significance of slow-binding enzyme inhibitors.
Adv. Enzymol. Relat. Areas Mol. Biol. 61,
201–301 (1988).
2 Copeland RA. Evaluation of Enzyme Inhibitors
in Drug Discovery. A Guide for Medicinal
Chemists and Pharmacologists. Wiley & Sons,
NY, USA (2005).
3 Copeland RA. Enzymes: A Practical
Introduction to Structure, Mechanism and Data
Analysis. 2nd Edition. Wiley & Sons, NY, USA
(2000).
4 Copeland RA, Pompliano DL, Meek TD.
Drug–target residence time and its
implications for lead optimization.
Nat. Rev. Drug Discov. 5(9), 730–739
(2006).
5 Tummino PJ, Copeland RA. Residence
time of receptor-ligand complexes and
its effect on biological function.
Biochemistry 47(20), 5481–5492
(2008).
6 Copeland RA. The dynamics of drug–target
interactions: drug–target residence time and
its impact of efficacy and safety. Expert
Opin. Drug Discov. 5(4), 305–310 (2010).
7 Boehr DD, Nussinov R, Wright PE. The role of
dynamic conformational ensembles in
biomolecular recognition. Nat. Chem. Biol.
5(11), 789–796 (2009).
8 Fischer E. Uber die optischen isomeren des
traubenzuckers, der gluconsaure und der
zuckersaure [Concerning the optical isomers of
the grape sugars, gluconic acid and saccharic
acid]. Ber. Dtsch. Chem. Ges. 23, 2611–2620
(1890).
9 Hulme EC, Trevethick MA. Ligand binding
assays at equilibrium: validation and
interpretation. Br. J. Pharmacol. 161(6),
1219–1237 (2010).
PERSPECTIVE | Copeland
Future Med. Chem. (2011) 3(12) 1500 future science group
10 Marcinkeviciene J, Kopcho LM, Yang T et al.
Novel inhibition of porcine pepsin by a
substituted piperidine. Preference for one of the
enzyme conformers. J. Biol. Chem. 2002.
277(32), 28677–28682 (2010).
11 Swinney DC. Biochemical mechanisms of drug
action: what does it take for success? Nat. Rev.
Drug Discov. 3(9), 801–808 (2004).
12 Hammes GG, Chang YC, Oas TG.
Conformational selection or induced fit: a flux
description of reaction mechanism. Proc. Natl
Acad. Sci. USA 106(33), 13737–13741 (2009).
13 Bhabha G, Lee J, Ekiert DC et al. A dynamic
knockout reveals that conformational
fluctuations influence the chemical step of
enzyme catalysis. Science 332(6026), 234–238
(2011).
14 Auzat I, Gawlita E, Garel JR. Slow ligand-
induced transitions in the allosteric
phosphofructokinase from Escherichia coli.
J. Mol. Biol. 249(2), 478–492 (1995).
15 Weikl TR, von Deuster C. Selected-fit versus
induced-fit protein binding: kinetic differences
and mutational ana lysis. Proteins 75(1),
104–110 (2009).
16 Garvey EP. Structural mechanisms of
slow-onset, two-step enzyme inhibition. Curr.
Chem. Biol. 4, 64–73 (2010).
17 Wood ER, Truesdale AT, McDonald OB et al.
A unique structure for epidermal growth factor
receptor bound to GW572016 (lapatinib):
relationships among protein conformation,
inhibitor off-rate, and receptor activity in
tumor cells. Cancer Res. 64(18), 6652–6659
(2004).
18 Cowan-Jacob SW, Mobitz H, Fabbro D.
Structural biology contributions to tyrosine
kinase drug discovery. Curr. Opin. Cell. Biol.
21(2), 280–287 (2009).
19 Sullivan SM, Holyoak T. Enzymes with
lid-gated active sites must operate by an
induced fit mechanism instead of
conformational selection. Proc. Natl Acad. Sci.
USA 105(37), 13829–13834 (2008).
20 Garvey EP, Schwartz B, Gartland MJ et al.
Potent inhibitors of HIV-1 integrase display a
two-step, slow-binding inhibition mechanism
which is absent in a drug-resistant T66I/M154I
mutant. Biochemistry 48(7), 1644–1653 (2009).
21 Gonzalez B, Pajares MA, Hermoso JA et al.
The crystal structure of tetrameric methionine
adenosyltransferase from rat liver reveals the
methionine-binding site. J. Mol. Biol. 300(2),
363–375 (2000).
22 Pearce FG, Andrews TJ. The relationship
between side reactions and slow inhibition of
ribulose-bisphosphate carboxylase revealed by a
loop 6 mutant of the tobacco enzyme. J. Biol.
Chem. 278(35), 32526–32536 (2003).
23 Liu Y, Stoll VS, Richardson PL et al. Hepatitis
C NS3 protease inhibition by peptidyl-a-
ketoamide inhibitors: kinetic mechanism and
structure. Arch. Biochem. Biophys. 421(2),
207–216 (2004).
24 Kapoor M, Mukhi PL, Surolia N, Suguna K,
Surolia A. Kinetic and structural ana lysis of
the increased affinity of enoyl-ACP
(acyl-carrier protein) reductase for triclosan in
the presence of NAD+. Biochem. J. 381(Pt 3),
725–733 (2004).
25 Vauquelin G. and I. Van Liefde. Slow
antagonist dissociation and long-lasting
in vivo receptor protection. Trends Pharmacol
Sci, 27(7), 356–359 (2006).
26 Gabrielsson J, Dolgos H, Gillberg PG,
Bredberg U, Benthem B, Duker G. Early
integration of pharmacokinetic and dynamic
reasoning is essential for optimal development
of lead compounds: strategic considerations.
Drug Discov. Today 14(7–8), 358–372 (2009).
27 Lu H, Tonge PJ. Drug–target residence time:
critical information for lead optimization.
Curr. Opin. Chem. Biol. 14(4), 467–474
(2010).
28 Alberding N, Austin RH, Beeson KW.
Tunneling in ligand binding to heme
proteins. Science 192(4243), 1002–1004
(1976).
29 Gooljarsingh LT, Fernandes C, Yan K et al.
A biochemical rationale for the anticancer
effects of Hsp90 inhibitors: slow, tight
binding inhibition by geldanamycin and its
analogues. Proc. Natl Acad. Sci. USA 103(20),
7625–7630 (2006).
30 Anderson K, Lai Z, McDonald OB et al.
Biochemical characterization of GSK1070916,
a potent and selective inhibitor of Aurora B
and Aurora C kinases with an extremely long
residence time. Biochem. J. 420(2), 259–265
(2009).
31 Anderson K, Yang J, Koretke K et al. Binding
of TPX2 to Aurora A alters substrate and
inhibitor interactions. Biochemistry 46(36),
10287–10295 (2007).
32 Storer AC, Cornish-Bowden A. Kinetic
evidence for a ‘mnemonical’ mechanism for
rat liver glucokinase. Biochem. J. 165(1),
61–69 (1977).
33 Kim YB. Kalinowski SS, Marcinkeviciene J.
A pre-steady state ana lysis of ligand binding
to human glucokinase: evidence for a
preexisting equilibrium. Biochemistry 46(5),
1423–1431 (2007).
34 Fersht AR, Requena Y. Equilibrium and rate
constants for the interconversion of two
conformations of a-chymotrypsin: the
existence of a catalytically inactive
conformation at neutral pH. J. Mol. Biol.
60(2), 279–290 (1971).
35 Halford SE. Escherichia coli alkaline
phosphatase. an ana lysis of transient kinetics.
Biochem. J. 125(1), 319–327 (1971).
36 Walz FG Jr. Relaxation kinetics of
ribonuclease T1 binding with guanosine and
3’-GMP. Biochim. Biophys. Acta 1159(3),
327–334 (1992).
37 Masterson LR, Cheng C, Yu T et al.
Dynamics connect substrate recognition to
catalysis in protein kinase A. Nat. Chem. Biol.
6(11), 821–828 (2010).
38 Copeland RA, Williams JM,
Giannaras J et al. Mechanism of selective
inhibition of the inducible isoform of
prostaglandin G/H synthase. Proc. Natl Acad.
Sci. USA 91(23), 11202–11206 (1994).
39 Callan OH, So OY, Swinney DC. The kinetic
factors that determine the affinity and
selectivity for slow binding inhibition of
human prostaglandin H synthase 1 and 2 by
indomethacin and flurbiprofen. J. Biol. Chem.
271(7), 3548–3554 (1996).
40 Basso LA, Santos DS, Shi W et al. Purine
nucleoside phosphorylase from Mycobacterium
tuberculosis. Analysis of inhibition by a
transition-state analogue and dissection by
parts. Biochemistry 40(28), 8196–8203
(2001).
41 Cha S, Agarwal RP, Parks RE Jr. Tight-
binding inhibitors-II. Non-steady state nature
of inhibition of milk xanthine oxidase by
allopurinol and alloxanthine and of human
erythrocytic adenosine deaminase by
coformycin. Biochem. Pharmacol. 24(23),
2187–2197 (1975).
42 Rawat R, Whitty A, Tonge PJ. The
isoniazid-NAD adduct is a slow, tight-binding
inhibitor of InhA, the Mycobacterium
tuberculosis enoyl reductase: adduct affinity
and drug resistance. Proc. Natl Acad. Sci. USA
100(24), 13881–13886 (2003).
43 Williams JW, Morrison JF, Duggleby RG.
Methotrexate, a high-affinity pseudosubstrate
of dihydrofolate reductase. Biochemistry
18(12), 2567–2573 (1979).
44 Rajagopalan R, Misialek S, Stevens SK et al.
Inhibition and binding kinetics of the
hepatitis C virus NS3 protease inhibitor
ITMN-191 reveals tight binding and slow
dissociative behavior. Biochemistry 48(11),
2559–2568 (2009).
45 Perni RB, Almquist SJ, Byrn RA et al.
Preclinical profile of VX-950, a potent,
selective, and orally bioavailable inhibitor of
hepatitis C virus NS3–4A serine protease.
Antimicrob. Agents Chemother. 50(3),
899–909 (2006).
46 Dierynck I, De Wit M, Gustin E et al.
Binding kinetics of darunavir to human
immunodeficiency virus type 1 protease
Conformational adaptation in drug–target interactions & residence time | PERSPECTIVE
www.future-science.com 1501
future science group
explain the potent antiviral activity and high
genetic barrier. J. Virol. 81(24), 13845–13851
(2007).
47 Liu T, Toriyabe Y, Kazak M, Berkman CE.
Pseudoirreversible inhibition of prostate-
specific membrane antigen by
phosphoramidate peptidomimetics.
Biochemistry 47(48), 12658–12660 (2008).
48 Machutta CA, Bommineni GR, Luckner
SR et al. Slow onset inhibition of bacterial
β-ketoacyl-acyl carrier protein synthases by
thiolactomycin. J. Biol. Chem. 285(9),
6161–6169 (2010).
49 Rhodes N, Heerding DA, Duckett DR et al.
Characterization of an Akt kinase inhibitor
with potent pharmacodynamic and antitumor
activity. Cancer Res. 68(7), 2366–2374
(2008).
50 Frye SV. Discovery and clinical development
of dutasteride, a potent dual 5a-reductase
inhibitor. Curr. Top. Med. Chem. 6(5),
405–421 (2006).
51 Davidovich C, Bashan A, Auerbach-Nevo T,
Yaggie RD, Gontarek RR, Yonath A.
Induced-fit tightens pleuromutilins binding to
ribosomes and remote interactions enable
their selectivity. Proc. Natl Acad. Sci. USA
104(11), 4291–4296 (2007).
52 Braz VA, Holladay LA, Barkley MD.
Efavirenz binding to HIV-1 reverse
transcriptase monomers and dimers.
Biochemistry 49(3), 601–610 (2010).
53 Harpel MR, Horiuchi KY, Luo Y et al.
Mutagenesis and mechanism-based inhibition
of Streptococcus pyogenes Glu-tRNAGln
amidotransferase implicate a serine-based
glutaminase site. Biochemistry 41(20),
6398–6407 (2002).
54 Van Aller GS, Nandigama R, Petit CM et al.
Mechanism of time-dependent inhibition of
polypeptide deformylase by actinonin.
Biochemistry 44(1), 253–260 (2005).
55 Lad L, Luo L, Carson JD et al. Mechanism of
inhibition of human KSP by ispinesib.
Biochemistry 47(11), 3576–3585 (2008).
56 McClerren AL, Endsley S, Bowman JL.
A slow, tight-binding inhibitor of the
zinc-dependent deacetylase LpxC of lipid A
biosynthesis with antibiotic activity
comparable to ciprofloxacin. Biochemistry
44(50), 16574–16583 (2005).
57 Holdgate GA, Ward WH, McTaggart F.
Molecular mechanism for inhibition of
3-hydroxy-3-methylglutaryl CoA (HMG-
CoA) reductase by rosuvastatin. Biochem. Soc.
Trans. 31(Pt 3), 528–531 (2003).
58 Clemens F, Drutkowski G, Wiese M,
Frohberg P et al. The inactivation of
lipoxygenase-1 from soybeans by amidrazones.
Biochim. Biophys. Acta 1549(1), 88–98 (2001).
59 Salowe SP, Hermes JD. Competitive and
slow-binding inhibition of calcineurin by drug
x immunophilin complexes. Arch. Biochem.
Biophys. 355(2), 165–174 (1998).
60 Dash C, Vathipadiekal V, George SP,
Rao M et al. Slow-tight binding inhibition of
xylanase by an aspartic protease inhibitor:
kinetic parameters and conformational
changes that determine the affinity and
selectivity of the bifunctional nature of the
inhibitor. J. Biol. Chem. 277(20),
17978–17986 (2002).
61 Garvey EP, Oplinger JA, Furfine ES et al.,
1400W is a slow, tight binding, and highly
selective inhibitor of inducible nitric-oxide
synthase in vitro and in vivo. J. Biol. Chem.
272(8), 4959–4963 (1997).
62 Marcinkeviciene J, Luo Y, Graciani NR,
Combs AP, Copeland RA. Mechanism of
inhibition of β-site amyloid precursor
protein-cleaving enzyme (BACE) by a
statine-based peptide. J. Biol. Chem. 276(26),
23790–23794 (2001).
63 Pargellis C, Tong L, Churchill L et al.
Inhibition of p38 MAP kinase by utilizing a
novel allosteric binding site. Nat. Struct. Biol.
9(4), 268–272 (2002).
64 Rogers CJ, Fisher AJ, Toney MD.
Aminophosphonate inhibitors of
dialkylglycine decarboxylase: structural basis
for slow binding inhibition. Biochemistry
41(41), 12320–12328 (2002).