Co-Conserved Features Associated with cis Regulation of
ErbB Tyrosine Kinases
Amar Mirza
1
, Morad Mustafa
1
, Eric Talevich
2
, Natarajan Kannan
1,2
*
1 Department of Biochemistry and Molecular Biology, University of Georgia, Athens, Georgia, United States of America, 2 Institute of Bioinformatics, University of Georgia,
Athens, Georgia, United States of America
Abstract
Background: The epidermal growth factor receptor kinases, or ErbB kinases, belong to a large sub-group of receptor
tyrosine kinases (RTKs), which share a conserved catalytic core. The catalytic core of ErbB kinases have functionally diverged
from other RTKs in that they are activated by a unique allosteric mechanism that involves specific interactions between the
kinase core and the flanking Juxtamembrane (JM) and COOH-terminal tail (C-terminal tail). Although extensive studies on
ErbB and related tyrosine kinases have provided important insights into the structural basis for ErbB kinase functional
divergence, the sequence features that contribute to the unique regulation of ErbB kinases have not been systematically
explored.
Methodology/Principal Findings: In this study, we use a Bayesian approach to identify the selective sequence constraints
that most distinguish ErbB kinases from other receptor tyrosine kinases. We find that strong ErbB kinase-specific constraints
are imposed on residues that tether the JM and C-terminal tail to key functional regions of the kinase core. A conserved
RIxKExE motif in the JM-kinase linker region and a glutamine in the inter-lobe linker are identified as two of the most
distinguishing features of the ErbB family. While the RIxKExE motif tethers the C-terminal tail to the N-lobe of the kinase
domain, the glutamine tethers the C-terminal tail to hinge regions critical for inter-lobe movement. Comparison of the
active and inactive crystal structures of ErbB kinases indicates that the identified residues are conformationally malleable
and can potentially contribute to the cis regulation of the kinase core by the JM and C-terminal tail. ErbB3, and EGFR
orthologs in sponges and parasitic worms, diverge from some of the canonical ErbB features, providing insights into sub-
family and lineage-specific functional specialization.
Conclusion/Significance: Our analysis pinpoints key residues for mutational analysis, and provides new clues to cancer
mutations that alter the canonical modes of ErbB kinase regulation.
Citation: Mirza A, Mustafa M, Talevich E, Kannan N (2010) Co-Conserved Features Associated with cis Regulation of ErbB Tyrosine Kinases. PLoS ONE 5(12):
e14310. doi:10.1371/journal.pone.0014310
Editor: Darren R. Flower, University of Oxford, United Kingdom
Received May 19, 2010; Accepted November 8, 2010; Published December 13, 2010
Copyright:  2010 Mirza et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was supported by funds to NK from the American Cancer Society (RSG-10-188-01-TBE), Georgia Cancer Coalition (GCC) and University of
Georgia. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: kannan@bmb.uga.edu
Introduction
The epidermal growth factor receptor (EGFR) and related
kinases, ErbB2, ErbB3, and ErbB4 (collectively called the ErbB
family) [1], are key components of our cellular machinery that
control signaling pathways associated with cell migration,
proliferation, and differentiation. Understanding how ErbB
kinases respond to specific activation and regulatory signals in
these pathways is essential for the development of new therapies
for human cancers that are associated with abnormal regulation of
ErbB kinase activity [2]. The domain architecture of ErbB kinases,
like most receptor tyrosine kinases (RTKs), is characterized by an
extracellular ligand-binding domain, a transmembrane domain, a
juxtamembrane (JM) segment, a kinase domain, and a COOH-
terminal tail (C-terminal tail). The kinase domain adopts a bi-lobal
structure consisting of an N-terminal ATP-binding lobe (N-lobe)
and a C-terminal substrate-binding lobe (C-lobe) [3,4].
Extensive studies on the extracellular ligand-binding domain
[5,6,7,8,9] and more recently on the intracellular kinase domain of
EGFR [10,11,12] have provided key insights into how EGFR
converts an extracellular signal into an intracellular response.
Ligand binding to the extracellular receptor induces dimerization
and activation of the intracellular kinase domain, which, upon
activation, autophosphorylates conserved tyrosine residues in the C-
terminal tail [13,14]. Autophosphorylation of the tyrosine residues
activates downstream signaling pathways by recruiting signaling
and docking proteins to the C-terminal tail [2,15,10]. In the absence
of an extracellular ligand, EGFR is maintained in an inactive
dimeric form [11], which prevents formation of the active dimer.
ErbB3 is believed to diverge from this canonical mechanism of
action because of its inability to catalyze phosphoryl transfer [16].
Recent studies, however, have challenged this view by showing low,
but detectable levels of ErbB3 autophosphorylation [17].
A key step in ErbB kinase signaling is the activation of the
intracellular kinase domain, which is achieved by an intermolecular
interaction between two kinase molecules in an asymmetric dimer
[10]. In the asymmetric dimer, the C-lobe of one kinase molecule
(the ‘‘activator’’) allosterically activates the other (‘‘receiver’’) by
PLoS ONE | www.plosone.org 1 December 2010 | Volume 5 | Issue 12 | e14310

inducing conformational changes in key regions of the receiver [10].
In particular, the regulatory C-helix in the N-lobe of the receiver
kinase switches from an inactive ‘‘out’’ to an active ‘‘in’’
conformation upon dimerization, and the flexible activation loop
in the C-lobe of the kinase switches from a substrate-inaccessible
conformation to a substrate-accessible conformation [10,11,12,18].
Also, the N-lobe of the kinase domain moves from an ATP-
accessible ‘‘open’’ conformation to an ATP-inaccessible ‘‘closed’’
conformation [15,19,20]. These conformational changes, which
occur upon activation of many protein kinases [21,22,23,24], are
tightly regulated to avoid physiological catastrophes [25,26].
The catalytic activity of EGFR is also regulated by conforma-
tional changes in the JM and C-terminal tail—two sequence
segments flanking the kinase domain. The JM segment functions
as an activation domain [12] by facilitating the formation of the
asymmetric dimer [11]. Specifically, the JM segment of the
receiver docks to the C-lobe of the activator to stabilize the
asymmetric dimer [11,12]. This docking interaction is prevented
in the inactive dimer of EGFR [11], where the JM docking surface
on the C-lobe is shielded by the C-terminal tail. Presumably, the
conformational changes associated with the JM and C-terminal
tail during kinase activation are closely coupled with the
conformational changes in the kinase core (described above) for
the tight regulation of kinase activity [27,28]. The atomic details of
how this coupling is achieved are not fully understood.
Receptor tyrosine kinases (RTKs) outside of the ErbB family
also contain flexible JM and C-terminal tail segments that play
important regulatory roles. In c-KIT and Ephrin receptor tyrosine
kinases, for example, the JM segment plays an autoinhibitory role,
in contrast to its activating role in EGFR, by interacting with the
active site [29] and the substrate-binding regions of the kinase
domain [30]. Likewise, the C-terminal tail in Tie2 inhibits
catalytic activity by an autoinhibitory mechanism [31], which is
distinct from EGFR [27]. Thus, individual RTKs have evolved
unique mechanisms to regulate catalytic activity by the JM and C-
terminal tails. Information regarding these family-specific regula-
tory mechanisms are encoded in the protein sequences—the cell’s
own medium for specifying molecular mechanisms. However,
despite the availability of RTK sequences from diverse organisms
(,3000 sequences), the sequence features that contribute to the
unique modes of regulation in individual RTKs have not been
systematically delineated.
We have shown using several case studies that Bayesian analysis
of the evolutionary constraints distinguishing functionally diver-
gent kinases is a viable approach for investigating the functional
specificity of kinases in signaling pathways [32,33]. Using this
approach, we recently identified that a conserved C-terminal tail
which wraps around the kinase core of AGC kinases is a
distinguishing feature of the AGC group [34]. This study also
revealed novel AGC kinase-specific motifs in the C-terminal tail
that were experimentally shown to be important for AGC kinase
functions [35,36]. In this study, we compare the functional
constraints acting on ErbB and related RTKs to identify the key
residues/motifs that contribute to ErbB kinase functional diver-
gence. We show, for the first time, that nearly all the residues that
distinguish ErbB kinases from other RTKs are involved in
tethering the JM and C-terminal tail to key functional regions of
the kinase core. Analysis of these tethering interactions in light of
the wealth of structural and functional data available on the ErbB
kinases suggests a model in which the identified residues contribute
to ErbB kinase functional specialization by facilitating a unique cis
interaction between the kinase core and the flanking JM and C-
terminal tails. Our analysis provides new testable hypotheses
regarding the cis regulation of the kinase core by the JM and C-
terminal tails, and provides new insights into cancer mutations that
alter this mode of regulation.
Results and Discussion
A co-conserved sequence pattern characteristic of the
ErbB kinase domain
To identify which sequence features most distinguish ErbB
kinases from other RTKs, we measured and analyzed the selective
constraints imposed on ErbB kinase sequences from diverse
organisms (see Methods). These constraints generally correspond
to residues that are highly conserved within the ErbB family, but
strikingly different in RTKs outside of the ErbB family (Figure 1).
Within the catalytic core, these residues correspond to W731,
P733, G735, E736, V738, K739, P741 in the b2-b3 loop; Y764
and S768 in the C-helix; S784 in the b4-b5 loop; Q791, P794,
C797 in the inter-lobe linker; V802 in the D-helix; G810,
N816,W817 in the E-helix; P848 in the b8 strand (not shown);
L861 in the activation loop (not shown); I/V904 in the F-helix (not
shown); I938 in the aG-aH loop; I/L941 and D942 in the H-helix.
Among these residues, Q791 in the inter-lobe linker contributes
the most (indicated by the height of the histogram in Figure 1) to
ErbB functional divergence, since none of the RTKs outside of the
ErbB family conserve a glutamine at the 791 position (Background
alignment in Figure 1). The residues described above also
distinguish ErbB kinases from non-receptor tyrosine kinase
(NRTK) sequences, as NRTKs also conserve strikingly different
residues at these positions. The only exceptions are W731 in the
b2-b3 loop and L861 in the activation loop. These two residues
are conserved in ErbB’s as well as in some NRTKs.
The JM and C-terminal tail contribute to ErbB kinase
functional divergence
In addition to the kinase domain, strong ErbB-specific
constraints are also imposed on residues flanking the kinase core,
namely, the Juxtamembrane segment, the JM-kinase linker and
the COOH-terminal tail.
ErbB-specific constraints in the JM and JM-kinase
linker. The JM segment is conserved across diverse organisms
within the ErbB family (Figure 1). However, across RTKs, the JM
segment displays little or no detectable sequence similarity. This
indicates that the JM segment is unique to the ErbB family and
likely contributes to its functional divergence. Some of the most
distinguishing residues in the JM region include: E690, P694, S695
and N700. Unlike the JM, the JM-kinase linker of ErbB kinases
share significant structural similarity with the JM-kinase linker of
other RTKs [37], despite very low sequence similarity. This is
indicated by the shared hydrophobic residues (L/I 707, L712)
between ErbB kinases (Foreground alignment in Figure 1) and
RTKs (Background alignment in Figure 1). The JM-linker region
also contains several residues that distinguish ErbB kinases
(Foreground) from other RTKs (Background). These include:
R705, I706, K708, E709, E711 and K714 (Figure 1).
The C-terminal tail is a distinguishing feature of ErbB
kinases. The C-terminal tail is also a distinctive feature of the
ErbB family. In particular, the sequence segment immediately
following the kinase domain (residues 971-1020 in Fig 1) is highly
conserved in ErbB kinases, but strikingly different in RTKs outside
of the ErbB family. The C-terminal tail segment is also co-conserved
with key regions of the kinase domain (see below). Some of the
distinctive residues/motifs in the C-terminal tail segment include:
[MF][AC][RK]DPxR[YF]LVI motif in the beginning of the C-tail,
D/E994 and F/L997 in the middle, and [DE]x[DE]xYL motif at
the C-terminal end (Figure 1).
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 2 December 2010 | Volume 5 | Issue 12 | e14310

Lineage and sub-family-specific variations within the
ErbB family
The ErbB prototypic features, described above, are generally
well conserved across diverse eukaryotic phyla (Figure 1).
However, some lower eukaryotes and parasitic worms diverge
from the canonical ErbB features, and display correlated sequence
changes in the JM, kinase, and C-terminal tail regions (Figure 1).
For example, a distinctive glutamine (Q791) in the inter-lobe
linker region is conserved as a glutamate (E) in sponges (Figure 1).
This change is correlated with the absence of the C-terminal tail
[DE]x[DE]xYL motif (Figure 1), which typically interacts with
Q791 in mammalian ErbB’s (Figure 2). Thus, the correlated
Figure 1. Contrast Hierarchical Alignment showing sequence patterns that most distinguish ErbB kinases (foreground alignment) from
other receptor tyrosine kinase (RTK) sequences (background). The residues identified by the Bayesian pattern partitioning procedure (see
Methods) as distinctive of the ErbB family are indicated by black dots above the alignment. The histograms on top of the alignment indicate the degree to
which residue composition in the foreground alignment (ErbB sequences) contrast with residues observed at the corresponding position in the
background alignment (other RTKs). The foreground set includes the sequences shown in the alignment and others whose conservation levels are
denoted by the consensus pattern and corresponding weighted residue frequencies (wt_res_freqs) below the alignment. Residue frequencies are
indicated in integer tenths where, for example, a ‘6’ indicates that the corresponding residue directly above it occurs 60-70% of the weighted sequences.
The number of weighted sequences and the total number of alignment sequences are indicated in parentheses next to ‘wt_res_freqs’ and ‘conserved’,
respectively. The background alignment and the corresponding residue frequencies are shown directly below the foreground alignment. The structural
location of the ErbB kinase-conserved residues and the overall domain organization of the ErbB family are shown above the alignment. The numbering
used in the alignment and in the text is according to the pre-mature EGFR numbering, which includes the 24 amino acid signaling sequence. A
background alignment for the C-terminal tail region is not shown because the C-terminal tail of ErbB kinases shares no detectable sequence similaritywith
the C-terminal tail of other RTKs. Thus, a standard background alignment consisting of protein sequences from NCBI-nr database was used to quantify the
constraints acting on the C-terminal tail residues. The NCBI sequence identifiers used in the query display alignment are: EGFR-human: 134104655;
ERBB2_human: 119533; ERBB3_human: 119534; ERBB4_human: 3913590; EGFR-fruitfly: 4588511; EGFR-worm: 212645651; EGFR-sponge: 18146642.
doi:10.1371/journal.pone.0014310.g001
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 3 December 2010 | Volume 5 | Issue 12 | e14310

sequence change observed in the kinase and C-terminal tail
suggests possible co-evolution of these two regions during
mammalian ErbB kinase evolution.
ErbB3, an atypical member of the ErbB family, also displays
significant variations in some of the canonical ErbB motifs/
residues. For example, a phosphorylatable tyrosine (Y1016) within
the C-terminal tail [DE]x[DE]xYL motif is conserved as an
aspartate (D) in ErbB3. Likewise, ErbB3 and ErbB4 replace a
canonical tyrosine (Y764) in the C-helix by a leucine (L) (Figure 1).
The structural and functional implications of these family-specific
variations are discussed in the sections below.
ErbB conserved residues are frequently mutated in
human cancers
Since ErbB kinases are one of the most frequently mutated gene
families in human cancers, we investigated whether any of the
identified ErbB conserved residues are among those known to be
associated with human cancers. Mapping of somatic mutations
identified in the ErbB family (see Methods) to available crystal
structures indicates that several of the ErbB kinase conserved
residues are indeed mutated in human cancers. S768 in the C-
helix, and L861 in the activation loop, are two of the most
frequently mutated residues in EGFR (Table 1). In addition, ErbB
conserved residues in the JM-kinase linker and b2-b3 loop are also
frequently mutated in lung, esophagus and upper digestive track
cancers (Table 1). The structural/functional impacts of these
mutations, however, are not fully understood.
Structural analysis of ErbB kinase-conserved residues and
proposed roles
To understand how the identified ErbB kinase conserved
residues contribute to ErbB kinase functional specialization, and
how mutations of these residues contribute to disease, we
performed crystal structure analysis of the identified residues (see
methods). As shown in Figure 2, nearly all the ErbB conserved
residues, although widely dispersed in sequence, spatially interact
Figure 2. A schematic of the ErbB kinase domain showing the distinctive ErbB residues and associated interactions (based on PDB:
2J5F). These residues are broadly classified into three categories based on their structural location and their role in tethering the JM and C-terminal
tail: (i) N-lobe tether (NLT), (ii) Active site tether (AST) and (iii) C-lobe tether (CLT). The residues are shown in stick representation. ErbB kinase-
conserved residues are colored in cyan and kinase conserved residues are colored in magenta. The sites of homo/hetero-dimerization are shown by
dark arrows.
doi:10.1371/journal.pone.0014310.g002
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 4 December 2010 | Volume 5 | Issue 12 | e14310

with the flexible JM and C-terminal tail segments to ‘‘tether’’ them
to three regions of the kinase core, namely, the N-lobe, the C-lobe
and the active site (Figure 2). Because these interactions are
malleable in crystal structures (Table S1), we use the term ‘‘tether’’
[34] to describe these interactions. Broadly, ErbB conserved
residues can be classified into three categories based on their
structural location and interaction: (i) N-Lobe Tether (NLT):
residues that tether the JM and C-terminal tail to the kinase N-
lobe; (ii) Active Site Tether (AST): residues that tether the C-
terminal tail to the ATP binding site; and (iii) C-Lobe Tether
(CLT): residues that tether the JM and C-terminal tail to the
kinase C-lobe (Figure 2; Figure 3).
The NLT, AST and CLT residues are highly conserved in ErbB
kinases, but strikingly different in RTKs outside of the ErbB
family. This selective conservation is likely to be important for
ErbB kinase functions, rather than for maintaining ErbB kinase
structure or fold, because RTKs that lack these residues essentially
adopt the same fold as ErbB kinases [30,31,38]. Indeed, recent
studies on the activation mechanism of EGFR and related ErbB
kinases support the functional importance of some of the identified
residues. For example, the C-terminal tail [MF][AC][RK]DPxR
[YF]LVI motif, which is part of the CLT, has been shown to play
an important role in ErbB3-ErbB2 hetero-dimerization [4,39],
and EGFR activation [11] (Figure 3A). Similarly, D984, a
distinctive aspartate (D984) in the C-terminal tail, was recently
shown to control C-terminal tail movement and kinase activation
[40]. Likewise, N816 and W817 in the CLT have been noted to
provide a malleable docking surface for the JM and C-terminal tail
in the active [12], and inactive states [11] of EGFR, respectively
(Figure 3B-C) [22]. We note that the JM and C-terminal tail
docking surface on the C-lobe is coupled to the substrate binding
aD-helix [41] via hydrophobic interactions between W817 and
V802 in the CLT (Figure 3C).
Whereas the role of CLT residues is well understood, little is
known about the role of the NLT and AST residues in ErbB
kinase functions. In particular, the selective conservation of
residues in the b2-b3 loop and JM-kinase linker are largely
mysterious. To obtain insights into these mysterious residues, we
performed crystal structure analysis of NLT and AST residues,
and interpreted our observations in light of the wealth of
functional data available on ErbB kinases. Our analysis suggests
important functional roles for the NLT and AST residues, and
provides new clues to cancer mutations that alter these residues.
NLT: A structural framework for coupling C-helix and
inter-lobe movement in ErbB kinases
As mentioned earlier, activation of EGFR kinase by dimeriza-
tion involves conformational changes in key regions of the N-lobe,
including repositioning of a regulatory C-helix from an inactive to
active conformation, and movement of the N-lobe relative to the
C-lobe [11,42]. We find that these flexible regions of the N-lobe
are tethered to the JM and C-terminal tail via ErbB kinase-
conserved interactions described below.
Interactions tethering the JM and regulatory C-
helix. Tethering of the JM to the C-helix is mediated through
ErbB conserved residues in the C-helix and JM segment. In
particular, a conserved asparagine (N700) in the JM segment
hydrogen bonds to the side-chain of Y764 and S768 in the C-helix
(Figure 4A). While these interactions are stable in the active state
of EGFR (see Table S1), in the inactive state these interactions are
disrupted, in part, because of repositioning of the C-helix in an
inactive ‘‘out’’ conformation [11,42] (Figure 4B). In particular,
Y764 in the C-helix moves away from N700 in the inactive state to
interact with hydrophobic residues in the b4 strand. This
malleable tethering of the JM to the C-helix is likely to be
functionally significant, as this may provide a framework for the
JM and the activating monomer to dynamically control C-helix
movement [11,12]. Consistent with this view, mutation of Y764 to
a phenylalanine [43], or N700 to an alanine [12], have been found
to significantly impair EGFR kinase activity. Notably, ErbB3 and
ErbB4 conserve a leucine at the Y764 position (Figure 4C). This
variation may reflect the unique ability of ErbB3 and ErbB4 to
form inactive N-lobe-N-lobe dimers, as the leucine, which replaces
Y764, is part of the N-lobe-N-lobe dimer interface in ErbB3 and
ErbB4 [15,16,17].
Oncogenic mutations, S768I and L861Q, may alter the
canonical interactions at the JM-C-helix interface. S768I:
S768I is a frequently occurring mutation (Table 1) in the C-helix
of EGFR that increases basal kinase activity [44]. S768 is located
Table 1. Somatic mutations in EGFR targeting the canonical
ErbB kinase-specific residues.
Mutation Cancer Primary Site Structural location
P694L(1)
P694S(1)
Lung JM
S695G(1) Thyroid JM
I706T(1) Lung JM-kinase linker
K708M(1) Lung JM-kinase linker
E709A(8)
E709G(4)
E709H(2)
E709K(9)
E709V(2)
Lung, prostate JM-kinase linker
E711K(1) Lung JM-kinase linker
W731R(1) Lung b2-b3 loop
P733L(1)
P733S(1)
P733T(1)
Lung b2-b3 loop
G735S(3) Prostate, Lung b2-b3 loop
V738G(2) Prostate b2-b3 loop
P741H(1)
P741L(2)
Thyroid
Central-nervous-system
b3 strand
aa
S768C(1)
S768I(21)
S768I(23)
S768N(1)
Lung, Oesophagus,
Central-nervous-system
C-helix
S784F(2)
S784P(1)
S784Y(1)
Lung,
Upper_aerodigestive_tract
b4-b5 loop
V802F(2)
V802I(1)
Lung,
Upper_aerodigestive_tract
aD-helix
G810D(2)
G810S(1)
Upper_aerodigestive_tract,
Lung
aE-helix
P848L(3)
P848L(4)
Upper_aerodigestive_tract,
Lung
b7-b8 loop
L861F(1)
L861P(1)
L861Q(26)
L861Q(30)
L861R(5)
L861R(3)
L861V(1)
Lung,
Central-nervous-system
Activation loop
F968L(1) Central-nervous-system I-helix
D1012H(1) Lung C-terminal Tail
doi:10.1371/journal.pone.0014310.t001
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 5 December 2010 | Volume 5 | Issue 12 | e14310

at the asymmetric dimer interface and is known to get
phosphorylated by Calcium calmodulin-dependent kinase II
(CAMK2), which modulates EGFR autokinase activity by
phosphorylation of S768 and C-terminal tail serine residues
[45]. Thus, mutation of S768 to isoleucine can contribute to
abnormal EGFR kinase activity by impacting one or more of the
following functions: (i) altering the tethering interactions between
the JM and C-helix, (ii) changing the dimer interface [46], and (iii)
preventing CAMK2 phosphorylation.
L861Q: L861Q is a frequently occurring activating mutation in
the activation loop of EGFR [47]. L861 is specific to ErbB kinases
(Figure 5A) and is typically conserved as an aspartate (D) in RTKs
outside of the ErbB family (Figure 5A). In the inactive state of
EGFR, L861 packs up against hydrophobic residues in the C-helix
[20], and this observation previously led to the suggestion that the
L861Q mutation may activate EGFR by destabilizing the
hydrophobic interactions in the inactive state [10,46,48]. Howev-
er, the structural interactions that stabilize the active form of the
L861Q mutant have not been proposed before. Modeling of a
glutamine in the active form of EGFR indicates that a glutamine at
the L861 position can potentially form a hydrogen bond with
Y764 (in the C-helix) in the active form, but not in the inactive
form of EGFR (Figure 5B). Furthermore, molecular dynamics
studies on the L861Q mutant (Figure S1) indicates that the
hydrogen bond between Q861 and Y764 is stable during the
course of the simulation (Figure 5C), and can likely prevent Y764
Figure 3. Role of CLT residues in tethering the JM and C-terminal tails. A) Role of the PxR[YF]LVI motif in tethering the C-terminal tail to the
C-lobe (PDB:2J5F). B) Tethering of the JM segment to the C-lobe in the active state (PDB: 3GOP). C) Tethering of the C-terminal tail to the C-lobe in the
inactive state (PDB: 3GT8). EGFRa and EGFRb correspond to monomer A and monomer B in the asymmetric dimer, respectively. The PxR[YF]LVI motif
in the C-terminal tail is labeled as LVI motif. The structure images were generated using Pymol (www.pymol.org). ErbB kinase-specific residues are
colored in cyan and hydrogen bonds are depicted as black dotted lines. The disordered segment of the C-terminal tail in the active state (Figure 3B) is
shown in dotted representation. Note that R973 is colored in green in Figures 3B and C because this residue is shared by some tyrosine kinases
outside of the ErbB family.
doi:10.1371/journal.pone.0014310.g003
Figure 4. Interactions tethering the JM to the C-helix in the N-lobe. A) Tethering interactions in the active state of EGFR (PDB: 2J5F). The
modeled nucleotide (ANP-PNP) is shown in sticks representation. Key hydrogen bonds mediated by ErbB kinase conserved residues are shown. B)
Conformational changes associated with ErbB kinase conserved residues in the active (PDB: 2J5F) and inactive state of EGFR (PDB: 2RFE). C) A multiple
sequence alignment showing the replacement of Y764 by a leucine in ErbB3 and ErbB4. The foreground alignment in Figure 4C corresponds to ErbB3
and ErbB4 sequences (62 sequences) and the background alignment corresponds to ErbB1 (EGFR) and ErbB2. The NCBI sequence identifiers numbers
for the sequences used in the query alignment of Figure 4C are as follows: ErbB4-human: 167745042; ErbB4-dog: 74005688; ErbB4-frog: 147906005;
ErbB3-cattle: 156718140; ErbB3-chicken: 113206130.
doi:10.1371/journal.pone.0014310.g004
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 6 December 2010 | Volume 5 | Issue 12 | e14310

from switching to an inactive conformation (Figure 5B-D). Thus,
in addition to destabilizing the inactive state, the L861Q mutation
may activate EGFR by stabilizing the C-helix tyrosine (Y764) in
an active conformation. We also predict that the L861Q mutation
in ErbB3 and ErbB4 may not have the same functional impact as
in EGFR because ErbB3 and ErbB4 conserve a leucine at the
Y764 position (Figure 4C).
Interactions tethering the C-terminal tail and inter-lobe
hinge. The opening and closing of the N-terminal ATP binding
lobe relative to the C-terminal substrate-binding lobe is an
essential part of catalysis [23]. Inter-lobe movement in eukaryotic
protein kinases is facilitated by the inter-lobe linker [49], which
connects the N and C lobes, and lobe-spanning salt bridges, which
serve as hinge points for domain movements [50].
In ErbB kinases, the hinge regions of the kinase domain are
tethered to the C-terminal tail via ErbB kinase-conserved residues
(Figure 6A). In particular, an ErbB kinase-conserved glutamine
(Q791) tethers the C-terminal tail to the inter-lobe linker by
Figure 5. Selective conservation and modeling of L861Q mutation (based upon PDB: 2JIU and 2ITN). A) Selective conservation of L861
in the ErbB family. B) Local structural interactions mediated by the L861Q lung cancer mutation in EGFR and snapshots of this interaction at various
time points (0, 5 and 10 ns) during the MD run. RMSD plots of the backbone atoms during the course of the simulation for the active (dimer) and
inactive (monomer) are shown in Figure S1.
doi:10.1371/journal.pone.0014310.g005
Figure 6. Interactions tethering the C-terminal tail to hinge regions of the kinase core. A) Structural location of the hinge tether. B-D) A
close-up view of the tethering interactions in the (B) active dimer (PDB: 2ITN), (C) inactive monomer (PDB: 2RFD), and (D) inactive dimer (PDB: 3GT8).
ANP-PNP is shown as green sticks and hydrogen bonds are depicted as black dotted lines. Some hydrogen bonds have been omitted for clarity.
doi:10.1371/journal.pone.0014310.g006
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 7 December 2010 | Volume 5 | Issue 12 | e14310

hydrogen bonding to two conserved asparates (D1012 and D1014)
in the C-terminal tail (Figure 6B). In EGFR, one of the aspartates
(D1014) in the C-terminal tail also hydrogen bonds to a kinase
conserved lysine (K852 in the C-lobe), which has been noted to
serve as a pivot point for inter-lobe movement [50]. Thus, ErbB
kinases have diverged from other RTKs to uniquely tether the C-
terminal tail to hinge regions of the kinase domain critical for
inter-lobe movement. Why would such tethering be important for
ErbB functions? One possibility is that this may provide an
additional layer of regulation by allowing the C-terminal tail to
internally control inter-lobe movement, and consequently kinase
activity. Notably, in the inactive structure of EGFR, where the two
lobes are in a closed conformation, the lobe-spanning salt bridge
between the glutamine (Q791) and the lysine (K852) is lost, in part
because of the movement of C-terminal tail away from the lysine
(K852) (Figure 6C) [42]. Also, the C-terminal tails in the inactive
dimer [11] prevent the formation of inter-lobe salt bridge by
engaging Q791 and K852 in different interactions (Figure 6D).
The control of inter-lobe movement by the C-terminal tail is
likely to differ in parasitic worms and sponges because the C-
terminal tail residues that interact with the inter-lobe hinge are
different in these organisms. Sponges lack the C-terminal tail
aspartates (D1012 and D1014), and parasitic worms contain a
glutamine at the D1014 position. Notably, both sponges and
parasitic worms replace the glutamine (Q791) in the inter-lobe
linker by an isoleucine (I) and glutamate (E), respectively (Figure 1).
Although the functional implication of this lineage specific
variation is unclear, it is likely that EGFR orthologs in sponges
and parasitic worms do not require regulation of catalytic activity
by the C-terminal tail. We note that ErbB3 differs from other
ErbB members in the inter-lobe hinge. In particular, the kinase
conserved lysine (K852), which forms a lobe spanning salt bridge
with Q791, is conserved as a glutamine (Q) in ErbB3. This ErbB3-
specific variation may contribute to the low levels of kinase activity
[17] by preventing opening and closing motion during catalysis.
Interactions coupling the C-terminal tail and the JM
segment. As mentioned earlier, some of the strongest ErbB
kinase-specific constraints are imposed on residues in the JM-
kinase linker and b2-b3 loop. ErbB-conserved residues in these
two regions structurally couple a phosphorylatable tyrosine
(Y1016) in the C-terminal tail to the JM and N-lobe regions
involved in dimerization (Figure 7). Some of the distinctive
residues involved in this coupling include, R705 in the JM-kinase
linker, W731 in the b2-b3 loop, and E736 in the b2-b3 loop.
Specifically, R705 and E736 form hydrogen bonds with the
hydroxyl group of Y1016, and W731 provides a favorable docking
surface for the aromatic ring of Y1016. These interactions are
further coupled to the C-helix and dimerization sites in the N-lobe
by I706 and E711 in the JM-kinase linker (Figure 7). Specifically,
I706 packs up against hydrophobic residues in the C-helix, and
E711 hydrogen bonds to the side-chain of W731, as well as to the
backbone of K708, which is involved in the asymmetric dimer
interface [10,12] (Figure 7).
The canonical ErbB kinase-conserved interactions between the
C-terminal tail, JM-kinase linker, and b2-b3 loop are altered in the
inactive dimer [11] (Figure 7). In particular, a conserved C-
terminal tail phenylalanine (F997) occupies a position analogous to
Y1016 in the inactive state (Figure 7). ErbB kinase-conserved
residues in JM-kinase linker and b2-b3 loop also undergo
concerted changes in the inactive dimer. In particular, R705,
which hydrogen bonds to Y1016 in the active state, moves away to
interact with E711 in the JM-kinase linker. Likewise, E736, which
Figure 7. Interactions tethering the C-terminal tail, JM-kinase linker and the b2-b3 loop (labeled as beta2-beta3). The top panel shows
structural interactions associated with ErbB kinase-conserved residues in the active asymmetric dimer (PDB: 2ITN) and the bottom panel shows the
inactive dimer (PDB: 3GT8). The right panel shows a close-up view of the interactions.
doi:10.1371/journal.pone.0014310.g007
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 8 December 2010 | Volume 5 | Issue 12 | e14310

typically interacts with Y1016 in the active state, moves towards a
phosphorylatable tyrosine (Y801 in the D-helix) in the inactive
state [11]. Together, these concerted changes appear to dynam-
ically couple the C-terminal tail with the JM and C-helix regions
involved in dimerization. Notably, a similar coupling between the
SH2-kinase linker, b2-b3 loop and SH3 domain have been noted
for Src tyrosine kinase [51,52], where the SH3 domain performs a
function analogous to the C-terminal tail of EGFR [11].
Our analysis suggests that the structural coupling between the
C-terminal tail and JM-kinase linker (described above) is likely to
differ in ErbB3, and EGFR orthologs in sponges and parasitic
worms because ErbB3 conserves an aspartate at the Y1016
position, and EGFR orthologs in sponges and parasitic worms lack
some of the canonical ErbB residues in the JM-kinase linker and
b2-b3 loop (Figure 1). Our analysis also suggests that cancer
mutations in the b2-b3 loop and the JM-kinase linker (Table 1)
may contribute to abnormal regulation by altering the conforma-
tional coupling between these two regions.
AST: A hypothetical mechanism for regulating ATP
binding by the C-terminal tail
The AST is largely formed by a helical segment (residues 997–
1001) in the C-terminal tail (Figure 8A), also referred to as the AP-
2 helix [11]. The AST is typically disordered in most ErbB
structures; however, in two structures of EGFR (PDB:1XKK and
2JIU) [19,53], the AST segment adopts two distinct conforma-
tions. In one conformation, it protrudes into the ATP binding
pocket, thereby tethering the C-terminal tail to the ATP binding
site (via two hydrophobic residues, F/L997 and L1001), while in
the other conformation the AST swings away from the ATP
binding pocket to become solvent-exposed (Figure 8A–B) [19,53].
This mode of dynamically tethering the C-terminal tail to the ATP
binding pocket is remarkably similar to PKA (Figure 8C), where a
conserved phenylalanine (F327
PKA
) in the C-terminal tail moves in
and out of the ATP binding pocket to serve as a gate for nucleotide
binding [23,54]. An analogous role for F997 in EGFR would
suggest a similar gating mechanism, wherein nucleotide binding is
controlled in a conformation dependent manner by the AST. Such
a function may also explain the paradoxical experimental
observations, where mutations in the AST both increases [11]
and decreases catalytic activity [10,55].
Concluding Remarks
Bayesian analysis of the evolutionary constraints acting on
receptor tyrosine kinase sequences has revealed a co-conserved
pattern characteristic of the ErbB family. Analysis of this co-
conserved pattern, in light of the wealth of structural and
functional data available on ErbB kinases, suggests a model in
which the identified residues contribute to ErbB kinase functional
divergence by providing a structural framework for the JM and C-
terminal tail to uniquely regulate ErbB kinase activity (Figure 9). A
compelling aspect of this model is that it readily explains the
Figure 8. Active site tether (AST) and its role in ATP binding. A) Superposition of the C-terminal tail in the active (PDB: 2JIU), inactive
monomer (PDB: 1XKK) and inactive dimer (PDB: 3GT8). B) A close up view of F997 and its proposed role in ATP binding. The modeled ANP-PNP is
shown in sticks representation. C) Structural location of F327 (PKA numbering) in the C-terminal tail of PKA (PDB: 1ATP [49]). D) Conformational
changes associated with F997 in the inactive symmetric dimer (PDB: 3GT8).
doi:10.1371/journal.pone.0014310.g008
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 9 December 2010 | Volume 5 | Issue 12 | e14310

inhibitory and activating functions of the C-terminal tail segment,
and provides new testable hypotheses for experimental studies. For
example, the hypothesis that the ErbB kinase-conserved residues
contribute to the cis regulation of the kinase core by the JM and C-
terminal tail can be tested experimentally. Likewise, the hypothesis
that the activation mechanism of sponges and roundworms differs
from their mammalian counterparts can also be tested experi-
mentally. Finally, by identifying a potential role for the C-terminal
tail in ATP binding (AST), our study provides new avenues for
designing selective ErbB kinase inhibitors.
Materials and Methods
Identification of ErbB-specific selective constraints
ErbB and related receptor tyrosine kinase (RTK) sequences
from diverse organisms were identified within NCBI nr, env_nr,
and translated EST databases using PSI-BLAST and motif-based
search procedures. These sequences were multiply aligned using
the MAPGAPS program [56] by building curated alignments/
profiles for each of the 18 families within the RTK sub-group.
These alignments included the Juxtamembrane region, the kinase
domain and ,50 residue segments C-terminal to the kinase
domain. The aligned sequences (,3,170 sequences) were
partitioned into two functionally divergent subgroups using a
Bayesian partitioning with pattern selection (BPPS) procedure
[57]. This identified a distinctive pattern that most optimally
distinguishes ErbB kinase sequences from other receptor tyrosine
kinase sequences (Figure 1). The extent to which these residues
contribute to the divergence of ErbB kinases was quantified using
a ball-in-urn statistical model [58], and indicated by the height of
the histogram above the alignments in Figures 1, 4C, and 8A.
Visualization of ErbB-specific selective constraints
The residues that contribute the most to ErbB kinase
evolutionary divergence are shown using a ‘‘Contrast Hierarchical
Alignment (CHA)’’ [58] (Figure 1). A CHA is based on three
categories of related sequences: (i) a query set, (ii) a foreground set
and (iii) a background set. In the Figure 1 alignment, represen-
tative ErbB sequences from diverse organisms constitute the query
set, all ErbB kinase sequences (419 sequences) correspond to the
foreground set, and receptor tyrosine kinases outside of the ErbB
family (2,751 sequences) correspond to the background set. The
Figure 9. Mechanistic model of EGFR kinase activation involving ErbB kinase-conserved residues. Three functional states of EGFR are
used (active dimer based upon PDB: 3GOP), inactive dimer (based upon PDB: 3GT8) and inactive monomer (based upon PDB: 1XKK) to illustrate the
coordinated functions of the identified residues in kinase activation. The JM and C-terminal tails are colored in green and red, respectively. The
residue numbers for the labeled residues are shown. The opening and closing motion of the two lobes are shown to illustrate how the C-tail may
regulate catalytic activity in a conformation dependent manner.
doi:10.1371/journal.pone.0014310.g009
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 10 December 2010 | Volume 5 | Issue 12 | e14310

residues that contribute to ErbB kinase evolutionary divergence, as
identified by the BPPS procedure, are shown by block dots above
the alignment. Notably, the residues identified by the BPPS
procedure are highly conserved in the ErbB family (foreground
alignment) and strikingly different in receptor tyrosine kinases
outside of the ErbB family (background alignment) (Figure 1).
ErbB3 is considered an atypical member of the ErbB family. To
determine to what extent ErbB3 contributes to the divergence of
the ErbB family, we ran the BPPS procedure by removing ErbB3
sequences from our alignments. Removing ErbB3 sequences did
not significantly alter the pattern-partitions created by the BPPS
procedure.
Identification of ErbB3-ErbB4 shared patterns
Representative sequences from ErbB3 and ErbB4 (query set) were
multiply aligned against mammalian ErbB kinase sequences (,164
sequences). The BPPS procedure was applied on this alignment to
identify sequence patterns that most distinguish ErbB3 and ErbB4
from other ErbB members (ErbB1 and ErbB4). Among other
residues, L794 in the C-helix (Figure 4C) was identified as one of the
most contributing residues to ErbB3-ErbB4 functional divergence.
Structural analysis of ErbB kinase conserved residues
Crystal structures of ErbB kinases solved in various function
states (see below) were obtained from the PDB database (http://
www.rcsb.org). Protein hydrogen bonds were added to structural
coordinates using the Reduce program [59]. Hydrogen bonds, van
der Waals interaction and CH-p interaction were calculated using
the CHAIN suite of programs [58]. The identified interactions
were further quantified by calculating the frequency of occurrence
of each interaction across multiple crystal structures (Table S1).
The structural interactions were visualized using PyMOL (http://
www.pymol.org). The symmetry related molecules in Figs 3b, 4a,
5b and 7a were generated using the ‘‘symmetry mates’’ utility in
PyMOL. The following PDB files were used in our analysis:
Active State: 1M14 [4]; 1M17 [4]; 2EB2(not published);
2EB3(not published); 2GS2 [10]; 2GS6 [10]; 2ITN [60]; 2ITO
[60]; 2ITP [60]; 2ITQ [60]; 2ITT [60]; 2ITU [60]; 2ITV [60];
2ITW [60]; 2ITX [60]; 2ITY [60]; 2ITZ [60]; 2J5E [61]; 2J5F
[61]; 2J6M [60]; 2JIT [53]; 2JIU [53]; 3GOP [12]
Inactive State: 1XKK [18]; 2GS7 [10]; 2JIV [53]; 2RF9 [42];
2RFD [42]; 2RFE [42]; 2RGP [62]; 3BEL [63]; 3GT8 [11]
Retrieval of cancer mutations, modeling and molecular
dynamics simulations of the mutant forms
Cancer-associated mutations in EGFR were identified by mining
the COSMIC database [64], a repository for somatic mutations in
human cancers. A structural model of the lung cancer-associated
L861Q mutant was built using two crystal structures of active EGFR
kinase domain (PDB ID: 2ITN and 2JIU) as templates. Because all
available crystal structures of EGFR have disordered regions, we
chose structures with non-overlapping disordered regions (PDB:
2ITN and 2JIU) for our modeling studies. ANP-PNP was modeled in
some of the nucleotide unbound structures to show the proximity of
ErbB conserved residues to the ATP binding pocket. ANP-PNP was
modeled by superimposing the nucleotide unbound structure to the
ANP-PNP bound structure of EGFR (PDB: 2ITN). Superposition
was done using the ‘‘cealign’’ plugin in Pymol.
For molecular dynamics studies, water molecules, bound
inhibitors, and other heteroatom’s were removed. The missing
residues were modeled using MODELLER [65]. Missing hydrogen
and heavy atoms were added using the LEaP program in the Amber
software suite [66]. Each protein was solvated with TIP3P water
model [67] and counterions were added for neutralization.
Molecular dynamics (MD) simulations were done using NAMD
software [68], version 2.7b1, and all-atom ff03 force fields from the
Amber package. Prior to the regular MD production run, a
smoothing function was applied to both the electrostatic and the van
der Waals forces at a distance of 10 A
˚
, and a pair list distance of
14 A
˚
with a switching cutoff distance of 12 A
˚
. All bonds with
hydrogen were kept rigid by applying the ShakeH algorithm, and
the protein backbone atoms were restrained with a harmonic
restraint (k
f
= 10 kcal/(mol ? A
˚
2
)). Conjugate-gradient energy min-
imization was performed on the solvated protein for 10,000 steps,
followed by heating from 0 to 298.15 K. The restraints on the
protein backbone atoms over multiple stages of equilibration under
NPT ensemble (P = 1 atm, T = 298.15 K) were released to obtain a
relaxed protein. The unrestrained MD productions were run for
10 ns using a time step of 2 fs and the NPT ensemble. Root-mean-
square deviation (RMSD) calculations were performed and
monitored to ensure that the simulation was stable during the
10ns time scale. Hydrogen bonding analysis was done using the ptaj
program in Amber suite of programs.
Supporting Information
Table S1
Found at: doi:10.1371/journal.pone.0014310.s001 (0.09 MB
DOC)
Figure S1
Found at: doi:10.1371/journal.pone.0014310.s002 (0.07 MB
DOC)
Acknowledgments
We thank Dr. Claiborne Glover and Dr. Zachary Wood for helpful
comments, and members of NK lab for helpful discussions. A summer
CURO fellowship to AM by the University of Georgia is acknowledged.
Author Contributions
Conceived and designed the experiments: NK. Performed the experiments:
AM NK. Analyzed the data: AM NK. Contributed reagents/materials/
analysis tools: MM ET NK. Wrote the paper: AM NK.
References
1. Manning G, Whyte DB, Martinez R, Hunter T, Sudarsanam S (2002) The
protein kinase complement of the human genome. Science 298: 1912–1934.
2. Yarden Y, Sliwkowski MX (2001) Untangling the ErbB signalling network. Nat
Rev Mol Cell Biol 2: 127–137.
3. Knighton DR, Zheng JH, Ten Eyck LF, Ashford VA, Xuong NH, et al. (1991)
Crystal structure of the catalytic subunit of cyclic adenosine monophosphate-
dependent protein kinase. Science 253: 407–414.
4. Stamos J, Sliwkowski MX, Eigenbrot C (2002) Structure of the Epidermal
Growth Factor Receptor Kinase Domain Alone and in Complex with a 4-
Anilinoquinazoline Inhibitor. J Biol Chem 277: 46265–46272.
5. Cho H-S, Leahy DJ (2002) Structure of the Extracellular Region of HER3
Reveals an Interdomain Tether. Science 297: 1330–1333.
6. Ferguson KM, Berger MB, Mendrola JM, Cho H-S, Leahy DJ, et al. (2003)
EGF Activates Its Receptor by Removing Interactions that Autoinhibit
Ectodomain Dimerization. Molecular Cell 11: 507–517.
7. Garrett TPJ, McKern NM, Lou M, Elleman TC, Adams TE, et al. (2002)
Crystal Structure of a Truncated Epidermal Growth Factor Receptor
Extracellular Domain Bound to Transforming Growth Factor [alpha]. Cell
110: 763–773.
8. Ogiso H, Ishitani R, Nureki O, Fukai S, Yamanaka M, et al. (2002) Crystal
Structure of the Complex of Human Epidermal Growth Factor and Receptor
Extracellular Domains. Cell 110: 775–787.
9. Schlessinger J (2002) Ligand-Induced, Receptor-Mediated Dimerization and
Activation of EGF Receptor. Cell 110: 669–672.
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 11 December 2010 | Volume 5 | Issue 12 | e14310

10. Zhang X, Gureasko J, Shen K, Cole PA, Kuriyan J (2006) An Allosteric
Mechanism for Activation of the Kinase Domain of Epidermal Growth Factor
Receptor. Cell 125: 1137–1149.
11. Jura N, Endres NF, Engel K, Deindl S, Das R, et al. (2009) Mechanism for
Activation of the EGF Receptor Catalytic Domain by the Juxtamembrane
Segment. Cell 137: 1293–1307.
12. Red Brewer M, Choi SH, Alvarado D, Moravcevic K, Pozzi A, et al. (2009) The
Juxtamembrane Region of the EGF Receptor Functions as an Activation
Domain. Molecular Cell 34: 641–651.
13. Schreiber A, Libermann T, Lax I, Yarden Y, Schlessinger J (1983) Biological
role of epidermal growth factor-receptor clustering. Investigation with
monoclonal anti-receptor antibodies. J Biol Chem 258: 846–853.
14. Yarden Y, Schlessinger J (1987) Self-phosphorylation of epidermal growth factor
receptor: evidence for a model of intermolecular allosteric activation.
Biochemistry 26: 1434–1442.
15. Qiu C, Tarrant MK, Choi SH, Sathyamurthy A, Bose R, et al. (2008)
Mechanism of Activation and Inhibition of the HER4/ErbB4 Kinase. Structure
16: 460–467.
16. Jura N, Shan Y, Cao X, Shaw DE, Kuriyan J (2009) Structural analysis of the
catalytically inactive kinase domain of the human EGF receptor 3. Proc Natl
Acad Sci U S A 106: 21608–21613.
17. Shi F, Telesco SE, Liu Y, Radhakrishnan R, Lemmon MA (2010) ErbB3/HER3
intracellular domain is competent to bind ATP and catalyze autophosphoryla-
tion. Proc Natl Acad Sci U S A 107: 7692–7697.
18. Wood ER, Truesdale AT, McDonald OB, Yuan D, Hassell A, et al. (2004) 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: 6652–6659.
19. Wood ER, Shewchuk LM, Ellis B, Brignola P, Brashear RL, et al. (2008) 6-
Ethynylthieno[3,2-d]- and 6-ethynylthieno[2,3-d]pyrimidin-4-anilines as tunable
covalent modifiers of ErbB kinases. Proceedings of the National Academy of
Sciences 105: 2773–2778.
20. Yun CH, Boggon TJ, Li Y, Woo MS, Greulich H, et al. (2007) Structures of lung
cancer-derived EGFR mutants and inhibitor complexes: mechanism of
activation and insights into differential inhibitor sensitivity. Cancer Cell 11:
217–227.
21. Jeffrey PD, Russo AA, Polyak K, Gibbs E, Hurwitz J, et al. (1995) Mechanism of
CDK activation revealed by the structure of a cyclinA-CDK2 complex. Nature
376: 313–320.
22. Sicheri F, Moarefi I, Kuriyan J (1997) Crystal structure of the Src family tyrosine
kinase Hck. Nature 385: 602–609.
23. Narayana N, Cox S, Xuong N-h, Eyck LFT, Taylor SS (1997) A binary complex
of the catalytic subunit of cAMP-dependent protein kinase and adenosine further
defines conformational flexibility. Structure 5: 921–935.
24. Huse M, Kuriyan J (2002) The conformational plasticity of protein kinases. Cell
109: 275–282.
25. Ferguson KM (2008) Structure-based view of epidermal growth factor receptor
regulation. Annu Rev Biophys 37: 353–373.
26. Cheng HC, Johnson TM, Mills RD, Chong YP, Chan KC, et al. (2010)
Allosteric networks governing regulation and catalysis of Src-family protein
tyrosine kinases: implications for disease-associated kinases. Clin Exp Pharmacol
Physiol 37: 93–101.
27. Landau M, Fleishman SJ, Ben-Tal N (2004) A Putative Mechanism for
Downregulation of the Catalytic Activity of the EGF Receptor via Direct
Contact between Its Kinase and C-Terminal Domains. Structure 12:
2265–2275.
28. Walton GM, Chen WS, Rosenfeld MG, Gill GN (1990) Analysis of deletions of
the carboxyl terminus of the epidermal growth factor receptor reveals self-
phosphorylation at tyrosine 992 and enhanced in vivo tyrosine phosphorylation
of cell substrates. J Biol Chem 265: 1750–1754.
29. Mol CD, Dougan DR, Schneider TR, Skene RJ, Kraus ML, et al. (2004)
Structural basis for the autoinhibition and STI-571 inhibition of c-Kit tyrosine
kinase. J Biol Chem 279: 31655–31663.
30. Wybenga-Groot LE, Baskin B, Ong SH, Tong J, Pawson T, et al. (2001)
Structural basis for autoinhibition of the Ephb2 receptor tyrosine kinase by the
unphosphorylated juxtamembrane region. Cell 106: 745–757.
31. Shewchuk LM, Hassell AM, Ellis B, Holmes WD, Davis R, et al. (2000)
Structure of the Tie2 RTK domain: self-inhibition by the nucleotide binding
loop, activation loop, and C-terminal tail. Structure 8: 1105–1113.
32. Kannan N, Neuwald AF (2004) Evolutionary constraints associated with
functional specificity of the CMGC protein kinases MAPK, CDK, GSK, SRPK,
DYRK, and CK2alpha. Protein Sci 13: 2059–2077.
33. Kannan N, Neuwald AF (2005) Did protein kinase regulatory mechanisms
evolve through elaboration of a simple structural component? J Mol Biol 351:
956–972.
34. Kannan N, Haste N, Taylor SS, Neuwald AF (2007) The hallmark of AGC
kinase functional divergence is its C-terminal tail, a cis-acting regulatory module.
Proceedings of the National Academy of Sciences 104: 1272–1277.
35. Gould CM, Kannan N, Taylor SS, Newton AC (2009) The Chaperones Hsp90
and Cdc37 Mediate the Maturation and Stabilization of Protein Kinase C
through a Conserved PXXP Motif in the C-terminal Tail. J Biol Chem 284:
4921–4935.
36. Robert ARomano, NatarajanKannan, Alexandr PKornev, Craig JAllison,
Susan STaylor (2009) A chimeric mechanism for polyvalent trans-phosphory-
lation of PKA by PDK1. Protein Science 18: 1486–1497.
37. Hubbard SR (2004) Juxtamembrane autoinhibition in receptor tyrosine kinases.
Nat Rev Mol Cell Biol 5: 464–471.
38. Hubbard SR, Wei L, Ellis L, Hendrickson WA (1994) Crystal structure of the
tyrosine kinase domain of the human insulin receptor. Nature 372: 746–754.
39. Schaefer G, Akita RW, Sliwkowski MX (1999) A Discrete Three-amino Acid
Segment (LVI) at the C-terminal End of Kinase-impaired ErbB3 Is Required for
Transactivation of ErbB2. J Biol Chem 274: 859–866.
40. Yang KS, Macdonald-Obermann JL, Piwnica-Worms D, Pike LJ (2010) Asp-
960/Glu-961 controls the movement of the C-terminal tail of the epidermal
growth factor receptor to regulate asymmetric dimer formation. J Biol Chem
285: 24014–24022.
41. Wu J, Yang J, Kannan N, Madhusudan, Xuong NH, et al. (2005) Crystal
structure of the E230Q mutant of cAMP-dependent protein kinase reveals an
unexpected apoenzyme conformation and an extended N-terminal A helix.
Protein Sci 14: 2871–2879.
42. Zhang X, Pickin KA, Bose R, Jura N, Cole PA, et al. (2007) Inhibition of the
EGF receptor by binding of MIG6 to an activating kinase domain interface.
Nature 450: 741–744.
43. Walker F, Hibbs ML, Zhang HH, Gonez LJ, Burgess AW (1998) Biochemical
characterization of mutant EGF receptors expressed in the hemopoietic cell line
BaF/3. Growth Factors 16: 53–67.
44. Chen YR, Fu YN, Lin CH, Yang ST, Hu SF, et al. (2006) Distinctive activation
patterns in constitutively active and gefitinib-sensitive EGFR mutants. Oncogene
25: 1205–1215.
45. Feinmesser RL, Wicks SJ, Taverner CJ, Chantry A (1999) Ca2+/calmodulin-
dependent kinase II phosphorylates the epidermal growth factor receptor on
multiple sites in the cytoplasmic tail and serine 744 within the kinase domain to
regulate signal generation. J Biol Chem 274: 16168–16173.
46. Landau M, Ben-Tal N (2008) Dynamic equilibrium between multiple active and
inactive conformations explains regulation and oncogenic mutations in ErbB
receptors. Biochimica et Biophysica Acta (BBA) - Reviews on Cancer 1785:
12–31.
47. Kumar A, Petri ET, Halmos B, Boggon TJ (2008) Structure and clinical
relevance of the epidermal growth factor receptor in human cancer. J Clin
Oncol 26: 1742–1751.
48. Papakyriakou A, Vourloumis D, Tzortzatou-Stathopoulou F, Karpusas M
(2009) Conformational dynamics of the EGFR kinase domain reveals structural
features involved in activation. Proteins 76: 375–386.
49. Padyana AK, Qiu H, Roll-Mecak A, Hinnebusch AG, Burley SK (2005)
Structural Basis for Autoinhibition and Mutational Activation of Eukaryotic
Initiation Factor 2{alpha} Protein Kinase GCN2. J Biol Chem 280:
29289–29299.
50. Lamers MBAC, Antson AA, Hubbard RE, Scott RK, Williams DH (1999)
Structure of the protein tyrosine kinase domain of C-terminal Src kinase (CSK)
in complex with staurosporine. Journal of Molecular Biology 285: 713–725.
51. Gonfloni S, Williams JC, Hattula K, Weijland A, Wierenga RK, et al. (1997)
The role of the linker between the SH2 domain and catalytic domain in the
regulation and function of Src. EMBO J 16: 7261–7271.
52. Young MA, Gonfloni S, Superti-Furga G, Roux B, Kuriyan J (2001) Dynamic
coupling between the SH2 and SH3 domains of c-Src and Hck underlies their
inactivation by C-terminal tyrosine phosphorylation. Cell 105: 115–126.
53. Yun C-H, Mengwasser KE, Toms AV, Woo MS, Greulich H, et al. (2008) The
T790M mutation in EGFR kinase causes drug resistance by increasing the
affinity for ATP. Proceedings of the National Academy of Sciences 105:
2070–2075.
54. Yang J, Kennedy EJ, Wu J, Deal MS, Pennypacker J, et al. (2009) Contribution
of non-catalytic core residues to activity and regulation in protein kinase A. J Biol
Chem 284: 6241–6248.
55. Xia L, Wang L, Chung AS, Ivanov SS, Ling MY, et al. (2002) Identification of
Both Positive and Negative Domains within the Epidermal Growth Factor
Receptor COOH-terminal Region for Signal Transducer and Activator of
Transcription (STAT) Activation. J Biol Chem 277: 30716–30723.
56. Neuwald AF (2009) Rapid detection, classification and accurate alignment of up
to a million or more related protein sequences. Bioinformatics 25: 1869–1875.
57. Neuwald AF, Kannan N, Poleksic A, Hata N, Liu JS (2003) Ran’s C-terminal,
basic patch, and nucleotide exchange mechanisms in light of a canonical
structure for Rab, Rho, Ras, and Ran GTPases. Genome Res 13: 673–692.
58. Neuwald AF (2007) The CHAIN program: forging evolutionary links to
underlying mechanisms. Trends in Biochemical Sciences 32: 487–493.
59. Word JM, Lovell SC, Richardson JS, Richardson DC (1999) Asparagine and
glutamine: using hydrogen atom contacts in the choice of side-chain amide
orientation. J Mol Biol 285: 1735–1747.
60. Yun C-H, Boggon TJ, Li Y, Woo MS, Greulich H, et al. (2007) Structures of
Lung Cancer-Derived EGFR Mutants and Inhibitor Complexes: Mechanism of
Activation and Insights into Differential Inhibitor Sensitivity. Cancer Cell 11:
217–227.
61. Blair JA, Rauh D, Kung C, Yun C-H, Fan Q-W, et al. (2007) Structure-guided
development of affinity probes for tyrosine kinases using chemical genetics. Nat
Chem Biol 3: 229–238.
62. Xu G, Abad MC, Connolly PJ, Neeper MP, Struble GT, et al. (2008) 4-Amino-
6-arylamino-pyrimidine-5-carbaldehyde hydrazones as potent ErbB-2/EGFR
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 12 December 2010 | Volume 5 | Issue 12 | e14310

dual kinase inhibitors. Bioorganic & Medicinal Chemistry Letters 18:
4615–4619.
63. Xu G, Searle LL, Hughes TV, Beck AK, Connolly PJ, et al. (2008) Discovery of
novel 4-amino-6-arylaminopyrimidine-5-carbaldehyde oximes as dual inhibitors
of EGFR and ErbB-2 protein tyrosine kinases. Bioorganic & Medicinal
Chemistry Letters 18: 3495–3499.
64. Bamford S, Dawson E, Forbes S, Clements J, Pettett R, et al. (2004) The
COSMIC (Catalogue of Somatic Mutations in Cancer) database and website.
Br J Cancer 91: 355–358.
65. Sali A (1995) Comparative protein modeling by satisfaction of spatial restraints.
Mol Med Today 1: 270–277.
66. Case DA, Darden TA, Cheatham TEI, Simmerling CL, Wang J, et al. (2008)
AMBER 10. San Francisco, California, USA: University of California.
67. Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML (1983)
Comparison of Simple Potential Functions for Simulating Liquid Water. J Chem
Phys 79: 926–935.
68. Phillips JC, Braun R, Wang W, Gumbart J, Tajkhorshid E, et al. (2005) Scalable
Molecular Dynamics with NAMD. J Comput Chem 26: 1781–1802.
ErbB Kinase Evolution
PLoS ONE | www.plosone.org 13 December 2010 | Volume 5 | Issue 12 | e14310