Kinetic mechanism and rate-limiting steps of focal adhesion kinase-1

Article abstract of DOI:10.1021/bi100824v

Steady-state kinetic analysis of focal adhesion kinase-1 (FAK1) was performed using radiometric measurement of phosphorylation of a synthetic peptide substrate (Ac-RRRRRRSETDDYAEIID-NH₂, FAK-tide) which corresponds to the sequence of an autophosphorylation site in FAK1. Initial velocity studies were consistent with a sequential kinetic mechanism, for which apparent kinetic values k_cat (0.052 ± 0.001 s^-1), K_MgATP (1.2 ± 0.1 μM), K_iMgATP (1.3 ± 0.2 μM), K_FAK-tide (5.6 ± 0.4 μM), and K _iFAK-tide (6.1 ± 1.1 μM) were obtained. Product and dead-end inhibition data indicated that enzymatic phosphorylation of FAK-tide by FAK1 was best described by a random bi bi kinetic mechanism, for which both E-MgADP-FAK-tide and E-MgATP-P-FAK-tide dead-end complexes form. FAK1 catalyzed the βγ-bridge:β-nonbridge positional oxygen exchange of [γ-¹⁸O₄]ATP in the presence of 1 mM [γ- ¹⁸O₄]ATP and 1.5 mM FAK-tide with a progressive time course which was commensurate with catalysis, resulting in a rate of exchange to catalysis of k_x/k_cat = 0.14 ± 0.01. These results indicate that phosphoryl transfer is reversible and that a slow kinetic step follows formation of the E-MgADP-P-FAK-tide complex. Further kinetic studies performed in the presence of the microscopic viscosogen sucrose revealed that solvent viscosity had no effect on k_cat/K_FAK-tide, while k_cat and k_cat/K_MgATP were both decreased linearly at increasing solvent viscosity. Crystallographic characterization of inactive versus AMP-PNP-liganded structures of FAK1 showed that a large conformational motion of the activation loop upon ATP binding may be an essential step during catalysis and would explain the viscosity effect observed on k_cat/K_m for MgATP but not on k_cat/K _m for FAK-tide. From the positional isotope exchange, viscosity, and structural data it may be concluded that enzyme turnover (k_cat) is rate-limited by both reversible phosphoryl group transfer (k_forward ≈ 0.2 s^-1 and k_reverse ≈ 0.04 s^-1) and a slow step (k_conf ≈ 0.1 s^-1) which is probably the opening of the activation loop after phosphoryl group transfer but preceding product release.

Full text of DOI:10.1021/bi100824v

Biochemistry 2010, 49, 7151–7163 7151
DOI: 10.1021/bi100824v
Kinetic Mechanism and Rate-Limiting Steps of Focal Adhesion Kinase-1
Jessica L. Schneck,‡ Jacques Briand,§ Stephanie Chen,‡ Ruth Lehr,‡ Patrick McDevitt,‡
Baoguang Zhao, Angela Smallwood, Nestor Concha, Khyati Oza,‡ Robert Kirkpatrick,‡ Kang Yan,^
James P. Villa,‡ Thomas D. Meek,*,‡ and Sara H. Thrall‡
‡Departments of Biological Reagents and Assay Development, §Analytical Chemistry, and Computational and Structural Chemistry,
GlaxoSmithKline Pharmaceuticals, 1250 South Collegeville Road, Collegeville, Pennsylvania 19426-0989, and ^Chemizon, Inc.,
Minzhuang Road 3, Haidian District, Beijing, 100195 China
Received May 24, 2010; Revised Manuscript Received June 24, 2010
ABSTRACT: Steady-state kinetic analysis of focal adhesion kinase-1 (FAK1) was performed using radiometric
measurement of phosphorylation of a synthetic peptide substrate (Ac-RRRRRRSETDDYAEIID-NH2,
FAK-tide) which corresponds to the sequence of an autophosphorylation site in FAK1. Initial velocity studies
were consistent with a sequential kinetic mechanism, for which apparent kinetic values kcat (0.052 (
0.001 s-1), KMgATP (1.2 ( 0.1 μM), KiMgATP (1.3 ( 0.2 μM), KFAK-tide (5.6 ( 0.4 μM), and KiFAK-tide (6.1 (
1.1 μM) were obtained. Product and dead-end inhibition data indicated that enzymatic phosphorylation of
FAK-tide by FAK1 was best described by a random bi bi kinetic mechanism, for which both E-MgADP-
FAK-tide and E-MgATP-P-FAK-tide dead-end complexes form. FAK1 catalyzed the βγ-bridge:β-non-
bridge positional oxygen exchange of [γ-18O4]ATP in the presence of 1 mM [γ-18O4]ATP and 1.5 mM FAK-
tide with a progressive time course which was commensurate with catalysis, resulting in a rate of exchange to
catalysis of kx/kcat = 0.14 ( 0.01. These results indicate that phosphoryl transfer is reversible and that a slow
kinetic step follows formation of the E-MgADP-P-FAK-tide complex. Further kinetic studies performed in
the presence of the microscopic viscosogen sucrose revealed that solvent viscosity had no effect on kcat/KFAK-tide
,
while kcat and kcat/KMgATP were both decreased linearly at increasing solvent viscosity. Crystallographic
characterization of inactive versus AMP-PNP-liganded structures of FAK1 showed that a large conforma-
tional motion of the activation loop upon ATP binding may be an essential step during catalysis and would
explain the viscosity effect observed on kcat/Km for MgATP but not on kcat/Km for FAK-tide. From the
positional isotope exchange, viscosity, and structural data it may be concluded that enzyme turnover (kcat) is
rate-limited by both reversible phosphoryl group transfer (kforward ≈ 0.2 s-1 and kreverse ≈ 0.04 s-1) and a slow
step (kconf ≈ 0.1 s-1) which is probably the opening of the activation loop after phosphoryl group transfer but
preceding product release.
Focal adhesion kinase-1 (FAK1)1 is a nonreceptor tyrosine
kinase which is virtually ubiquitous in cells (1). The primary role
of FAK1 is the conductance of signals from activated integrin
receptors to intracellular protein kinase cascades by either
phosphorylation or localization of its downstream effectors,
which ultimately leads to activation of the ERK and JNK/
MAPK pathways (2). Association of FAK1 and several signaling
proteins, such as Src-family tyrosine kinases, p130Cas, Shc, Grb2,
PI3-kinase, paxillin, and tallin, has been demonstrated (2).
Signal transduction through FAK1 indirectly plays a critical role
in many cellular processes, such as cell cycle regulation, cadherin-
dependent cell adhesion, migration, invasion, and metastasis,
and apoptosis (3). Specifically, FAK1 is involved in the
assembly/disassembly of cellular adhesion sites, which are vital
in the control of cell migration. Increased expression of FAK1
has been observed in a variety of human cancers, including
thyroid, prostate, cervical, colon, rectal, oral epithelial, melanoma,
and ovarian cancers (1, 3). Amplified FAK1 expression and
activity are correlated with malignant or metastatic disease and at
stages of the diseases which comprise poor prognosis (3, 4).
FAK1 is a homodimeric protein consisting of 119-kDa sub-
units (2). The catalytic domain of FAK1 contains a SH2 (Src
homology 2) motif and is flanked by a large N-terminal FERM
(band four-point-one, erzin, radixin, moesin homology) domain
and a C-terminal domain containing the focal adhesion targeting
(FAT) sequence (5). When FAK1 is in its inactive form, the FERM
domain bridges the N- and C-lobes of the kinase domain and
extends across the active site cleft, blocking access to substrate and
thereby ultimately hindering catalysis (5). The FAT domain
consists of a four-helix bundle (6, 7), and this domain is critical
for targeting to focal adhesions via association with paxillin (5).
Src kinase activates FAK1 by phosphorylating Tyr576 and
Tyr577 residues in the activation loop of the enzyme (amino acids
564-584), which triggers autophosphorylation of the FAK1
protein at Tyr397 (8). This autophosphorylation, which can be
*To whom correspondence should be addressed. Phone: (610) 917-
6786. Fax: (610) 917-7385. E-mail: Thomas.D.Meek@gsk.com.
1Abbreviations: AMP-PNP, adenosine 50-(β,γ-imido)triphosphate;
DTT, dithiothreitol; EDTA, ethylenediaminetetraacetic acid; FAK1,
focal adhesion kinase-1; FAK-tide, Ac-RRRRRRSETDDYAEIID-
NH2; G6PDH, glucose-6-phosphate dehydrogenase; GST-Src, glu-
tathione transferase N-terminally tagged Src kinase; HEPES, 4-(2-hydro-
xyethyl)-1-piperazineethanesulfonic acid; HK, yeast hexokinase; LC/MS/
MS, liquid chromatography, tandem mass spectrometry; LDH, porcine
heart lactate dehydrogenase; P-FAK-tide, H-SETDD(phosphoY)AEIID-
OH; PBS, phosphate-buffered saline; PEP, phosphoenolpyruvate; PIX,
positional isotope exchange; PK, rabbit muscle pyruvate kinase; TCEP,
tris(2-carboxyethyl)phosphine.
r
2010 American Chemical Society
Published on Web 07/02/2010
pubs.acs.org/Biochemistry

7152 Biochemistry, Vol. 49, No. 33, 2010
Schneck et al.
cis or trans phosphorylation, is required for full activation of the
kinase domain (9). FAK1 is unique among kinases in that Tyr397
is not located in the activation loop but rather in a linker region
between the N-terminal FERM domain and the catalytic do-
main (9). Natural substrates for FAK1 include paxillin and
p130Cas (2).
protease cleavage site. FLAG-6ꢀHis-TEV-FAK1 was generated
using the Bac-to-Bac (Invitrogen), and the protein was expressed
in a proprietary Sf9 insect cell line. Cells were grown to a density
of 4 ꢀ 106 cells/mL in 10 L and infected with Flag-6ꢀHis-TEV-
FAK1 virus at a multiplicity of infection factor of 2.8. At 48 h
postinfection, cells were harvested by centrifugation at 1200 rpm
for 20 min. The baculovirus-expressed protein was solubilized
from 60 g of cells by sonication in 50 mM Tris buffer (pH 7.5)
containing 300 mM NaCl, 0.1 mM TCEP, and 1% Triton (v/v)
(lysis buffer) and purified by batch binding to 7.5 mL of NiNTA
agarose resin (Qiagen) equilibrated in lysis buffer. The resin was
washed with 10 mM imidazole in lysis buffer with 0.1% Triton,
and the FAK1 protein was eluted with 10 column volumes of a
linear gradient from 10 to 300 mM imidazole in wash buffer. The
eluted protein was concentrated and subjected to size-exclusion
chromatography on a 300 mL SDX200 column (2.6 ꢀ 60 cm)
(GE Healthcare) equilibrated with 50 mM Tris (pH 7.5), 200 mM
NaCl, and 0.1% Triton, to remove aggregated protein. All steps
were performed at 4 ꢀC.
The collected fractions corresponding to FAK1 were activated
with recombinant HisSBP-Src (SBP is streptavidin binding
protein) at a 1:20 (w/w) ratio of Src to FAK1 in the presence
of 20 mM MgCl2, 2 mM ATP, and 2 mM DTT. The reaction was
incubated at room temperature for 3 h and then overnight at
4 ꢀC, after which Src protein was removed from the reaction by
binding of its SBP tag to streptavidin resin. The activated FAK
protein was chromatographed on a SDX200 column as above to
remove any aggregated protein resulting from the activation.
This procedure resulted in almost complete phosphorylation of
Tyr-576 and Tyr-577 as determined by LC/MS/MS analysis.
Phosphorylation of these two residues is required for fully active
FAK1.
The gene that was synthesized for the coding sequence of the
FAK1 kinase domain (411-686) was optimized in terms of
codon usage and also contains a 50 fused sequence encoding an
N-terminal Tev cleavage site (Supporting Information Figure 1S)
and was transferred by Gateway BP recombination into
pDONR221 (Invitrogen). The Tev-FAK1(411-686) coding se-
quence was subsequently transferred by Gateway LR recombi-
nation into pDEST8His6, a modified version of the Invitrogen
pDEST8 baculovirus destination vector containing an N-terminal
His6 tag module in frame with the Gateway cassette. The
resultant construct encodes His6TevFAK1(411-686) with the
following N-terminal His6tev linker preceding serine-411 of
FAK1 (MGHHHHHHTSTSLYKKAGFENLYFQG). This
destination plasmid was used to generate recombinant baculo-
virus expressing His6TevFAK1(411-686) using the Bac-to-Bac
baculovirus expression system (Invitrogen). The protein was
expressed in a high cell density optimized Sf9 insect cell line
(S9). Cells were infected at 1 ꢀ 106 cells/mL viable cell concen-
tration with frozen baculovirus infected insect cells (BIIC) and
harvested at 65 h postinfection.
Frozen baculovirus-infected S9 cells expressing His6Tev-
FAK1(411-686) were dispersed (1 g of cells/10 mL of buffer)
in buffer A (50 mM Tris-HCl, 0.5 M NaCl, 10 mM imidazole,
5 mM 2-mercaptoethanol, 5% glycerol, pH 7.6) containing
protease inhibitors (Roche complete, no EDTA) and 1 mM
PMSF. Cells were lysed by two passes at 10000 psi through a
M110Y fluidizer (Microfluidics Inc., Newton, MA) and clarified
by centrifugation for 1 h at 35000g. The supernatant was applied
to a NiNTA column (Qiagen Inc.) equilibrated in buffer
A, washed with 10 column volumes of buffer A, and eluted with
The detailed kinetic mechanisms of several protein kinases
involved in signal transduction have been elucidated. Two
examples include Csk (10) and ERK2 (11, 12). Csk and ERK2
both employ a random bi bi reaction mechanism for which
release of the MgADP product is mostly rate-determining. A
thorough understanding of the kinetic mechanism of FAK could
greatly augment the proper design and execution of high-
throughput screening assays intended to find mechanistically
novel inhibitors which target the rate-limiting step(s) of catalysis.
In this study initial velocity, dead-end and product inhibition,
positional isotope exchange of MgATP, solvent viscometric, and
structural studies were undertaken to characterize the catalytic
mechanism of FAK1 using FAK-tide, a septidecapeptide of the
sequence Ac-RRRRRRSETDDYAEIID-NH2, which is a mimic
of the autophosphorylation site of FAK1 containing Tyr397.
MATERIALS AND METHODS
Materials. HEPES, DTT, ATP, and ADP were purchased
from Invitrogen. Magnesium chloride, manganese chloride,
Tween-20, KCl, NADH, PK, PEP, glucose, and sucrose were
purchased from Sigma-Aldrich. AMP-PNP was purchased from
Roche. NADþ, HK, and G6PDH were purchased from Calbio-
chem. LDH was purchased from USB Corp. Ficoll 400 was
purchased from Alfa Aesar. FAK-tide (Ac-RRRRRRSETD-
DYAEIID-NH2) was purchased from 21st Century Biochem-
icals, Inc. (purity g95%). Phospho-FAK-tide (H-SETDD-
(phosphoY)AEIID-OH) and Phe/Tyr-FAK-tide (H-SETDD-
FAEIID-OH) were purchased from California Peptide at
>99% purity. Phosphocellulose filter paper plates (96-well) were
purchased from Millipore. [γ-33P]ATP (10 μCi/μL specific activ-
ity, diluted to 2-120 nCi/μL for assays) and Microscint-20 were
obtained from Perkin-Elmer. Reactions were conducted in
96-well half-area, nonbinding surface plates purchased from
Corning, Inc.
Preparation of [γ-18O]ATP. [γ-18O]ATP was provided by
WuXi (Shanghai, China) by coupling of the mono(tri-n-
butylammonium) salt (13) of [18O4]phosphate, prepared from
PCl5 and [18O]H2O (14), with ADP-morpholidate (15). The
resulting tris(triethylammonium salt) of [γ-18O4]ATP was found
to be >96% pure by high-performance liquid chromatography
and with a molecular mass of 818.75 amu as determined by mass
spectrometry. Further quantification of the product [γ-18O]ATP
was ascertained from measurement of a sample by its absorbance
at 259 nm for total adenine content and also by enzymatic
conversion of ATP to ADP by use of the hexokinase/glucose-
6-phosphate dehydrogenase coupling enzyme system as de-
scribed (17). Samples were found to consist of g93% ATP.
Analysis of the integrals of γ-phosphate resonance of the
[γ-18O4]ATP by 31P NMR indicated that the substitution pattern
of 18O atoms in the γ-phosphate group was as follows: [γ-18O4],
74.4%; [γ-18O316O], 25.6% (16).
Enzyme Preparation. Full-length human FAK1
(NM_153831) was cloned into a baculovirus expression vector
derived from pDest8 (Invitrogen) containing an N-terminal
FLAG epitope tag followed by a hexahistidine tag and a TEV

Article
Biochemistry, Vol. 49, No. 33, 2010 7153
0.25 M imidazole in buffer A. Fractions containing FAK1 kinase
domain were pooled and treated with Tev protease (1:10 ratio) to
remove hexaHis tag. The reaction mixture was fractionated on a
Superdex 200 column in buffer B (40 mM Tris-HCl, 0.15 M
NaCl, 5% glycerol, 1 mM DTT, pH 7.6). FAK1 kinase domain
was eluted as a monomer from a Superdex 200 column. LC/MS
mass determination indicated unphosphorylated FAK1. FAK1
kinase domain was activated with His-GST-cSrc (constitutively
active cSrc, 86-537, Y530F) in buffer B containing 10 mM ATP,
20 mM MgCl2, 10 mM DTT, and 10 mM β-glycerophosphate
(1:30 ratio) at 30 ꢀC, monitoring phosphorylation by LC/MS.
His-GST-cSrc was removed by passage through a NiNTA
column. FAK1 solution was further fractionated on Source
15S in buffer C (50 mM sodium acetate, 1 mM TCEP, 1 mM
glycerophosphate, 1 mM EDTA, 5% glycerol, pH 5.5) with a 20
mM to 0.5 M NaCl gradient. The major peak containing a
diphosphorylated form of FAK1 (observed mass of 31982 Da
against calculated mass of 31815 Da) was pooled, dialyzed
against the final storage buffer (50 mM Tris-HCl, 0.25 M NaCl,
2 mM TCEP, 1 mM EDTA, 5% glycerol, pH 7.5), and snap-
frozen in liquid N2 for storage at -80 ꢀC.
Steady-State Kinetic Assays. Steady-state kinetic para-
meters were obtained using a filter-binding assay which measured
phosphorylation of the tyrosyl residue of FAK-tide peptide from
[γ-33P]ATP. Initial rate studies of full-length FAK1 were per-
formed in 40 μL solutions containing 40 mM HEPES (pH 7.2),
10 mM MgCl2, 2 mM MnCl2, 1 mM DTT, 0.01% (v/v) Tween-20
buffer, 800 pM FAK1, FAK-tide (0.5-250 μM), and [γ-33P]ATP
(2-120 nCi/μL, 0.1-10 μM ATP) in half-area, 96-well plates (21
( 1 ꢀC). Reactions were quenched at 0-60 min with 40 μL of 1%
(v/v) H3PO4, and reaction mixtures were transferred to 96-well
plates containing phosphocellulose filter papers. After 20 min
incubation, filter wells were washed three times with 0.5% (v/v)
H3PO4 and then incubated at 50 ꢀC for 30 min. MicroScint-20
(50 μL) was added to each dried well, and the amount of [γ-33P]-
labeled peptide was quantified by scintillation counting on a
TopCount instrument (Perkin-Elmer). Total input picomoles of
[γ-33P]ATP, assuming quantitative capture of radioactivity, was
determined for each filter plate by generation of a calibration
curve comprised of enzyme-free reaction mixtures. For all data at
all combinations of substrate concentrations tested, the linearity
of the reaction time course was determined via individual time
points taken at 10 min intervals at 0-80 min. All combinations of
substrate concentrations tested appeared linear for at least
60 min, and initial rates were obtained from analysis of the initial
linear phases.
Likewise, initial rates for the reverse reaction, or for quanti-
fication of MgATP concentrations within reaction mixtures for
PIX, were determined using the HK-G6PDH enzyme-coupling
system. In a 384-well clear bottom Greiner plate, 48 μL reaction
mixtures comprised of 100 mM HEPES (pH 7.2), 100 mM KCl,
20 mM MgCl2, 2 mM NADþ, 10 units/mL G6PDH, and 25
units/mL HK were used to detect the remaining concentrations
of [γ-18O]MgATP with 2 or 25 μL aliquots taken from reaction
mixtures (21 ( 1 ꢀC). Reactions were initiated by addition of 1 μL
of 50 mM glucose, and as above, [MgATP] was determined from
a calibration curve.
For viscosity studies, solutions of 0-40% (w/v) sucrose or
0-10% (w/v) Ficoll 400 in 40 mM HEPES (pH 7.2), 10 mM
MgCl2, 2 mM MnCl2, 1 mM DTT, and 0.01% (v/v) Tween-20
buffer were prepared, and the resulting viscosity was measured by
an Oswalt viscometer. Initial velocity data were then determined
to obtain kinetic parameters for the two substrates in the presence
of 0-40% (ηrel = 1-3.8) sucrose or 0-10% (ηrel = 1-4.5) Ficoll
400 using the filtration binding assay at saturating fixed substrate
concentrations (12 μM MgATP or 25 μM FAK-tide).
Positional Isotope Exchange in [γ-18O]ATP. In a 1.0 mL
reaction mixture, 100 nM full-length FAK was incubated with
40 mM HEPES (pH 7.2), 10 mM MgCl2, 1 mM DTT, 1 mM
[γ-18O]ATP, and 1.5 mM FAK-tide at 21 ( 1 ꢀC for 70-250 min.
The reaction mixtures were quenched by adding 0.25 mL of 0.5 M
EDTA (pH 8.0), followed by vigorous mixing. The extent of βγ-
bridge:β-nonbridge positional oxygen exchange in [γ-18O]ATP
was determined from the peak intensities of 31P NMR signals of
the [γ-18O4]-P and [γ-18O316O]-P and [β-18O16O3]-P and [β-16O4]-
P species, owing to the ≈0.02 ppm change in chemical shift of the
31P species resulting from substitution of an 16O ligand by
18O (16). NMR samples were prepared by mixing together
450 μL of the quenched reaction mixtures, 60 μL of D2O, and
90 μL of 0.5 M EDTA yielding the final sample conditions:
0.6 mM combined [γ-18O]ATP þ [β-18O]ADP, 24 mM HEPES
(pH 7.2), 6 mM MgCl2, 60 nM FAK1, 0.9 mM FAK-tide, 10%
(v/v) D2O, and 150 mM EDTA in a volume of 600 μL. The 31
P
NMR spectra were measured by using a Varian Unity INOVA
600 operating at 242.76 MHz. All of the spectra were recorded
using a 5 mm probe regulated at 25 ꢀC, 32K data points, a sweep
width of 20000 Hz, an acquisition time of 1.6384 s, 16384 scans,
and a relaxation delay of 0.1 s. Prior to Fourier transformation,
the resolution was enhanced by applying a shifted sine bell (36ꢀ)
over the first 32K data. In addition to the spectral quantification,
ADP formation and residual ATP in each of the quenched
samples were also quantified enzymatically to assess the extent
of the FAK reaction using respectively the PK-LDH and
HK-G6PDH coupling systems as described above and in
Meek et al. (17).
Crystallization and Structural Characterization of
AMP-PNP-Bound FAK1. Crystals of the catalytic domain
of human FAK1 (residues 411-686) in complex with the
nonhydrolyzable ATP analogue, 2þAMP-PNP (adenosine
50-(β,γ-imido)triphosphate), and Mg were grown by sitting-
drop vapor diffusion at 20 ꢀC from a solution of 18-21%
polyethylene glycol 3350 and 0.2 M lithium sulfate buffered with
0.1 M Tris-HCl (pH 8.5). MgCl2 and AMP-PNP were added to
the FAK1 enzyme solution immediately before crystallization to
give a final concentration of 3 mM. Crystals started to appear
after 2 weeks and attained full size after 6 weeks, by which time
they were harvested, allowed to equilibrate in a similar crystal-
lization solution, now containing 18% ethylene glycol, and quick
Initial rates were also assessed by use of the PK-LDH coupled-
enzyme system to measure amounts of MgADP formed from
FAK-tide and MgATP as catalyzed by FAK1. Briefly, 2-25 μL
aliquots were withdrawn at intervals of 0-80 min from 1.0 mL
reaction mixtures (21 ( 1 ꢀC) containing 40 mM HEPES (pH
7.2), 10 mM MgCl2, 3 mM DTT, 0-1 mM MgATP, 0-1.5 mM
FAK-tide, and 0.1 μM FAK1. The concentrations of MgADP
were quantified by measurement of the decrease in NADH
absorbance (340 nm) on a Spectramax spectrophotometer in a
48 μL reaction mixture (384-well clear-bottom Greiner plate)
containing 100 mM HEPES (pH 7.2), 100 mM KCl, 20 mM
MgCl2, 2 mM PEP, 200 μM NADH, 25 units/mL LDH, and
25 units/mL PK, and resulting MgADP concentrations were
determined from a calibration curve. This coupled-enzyme assay
was also used to quantify the MgADP produced in the samples
for PIX.

7154 Biochemistry, Vol. 49, No. 33, 2010
Schneck et al.
˚
frozen in liquid nitrogen. X-ray diffraction data to 1.45 A
resolution were collected using the MAR-USA CCD165 detector
at the Industrial Macromolecular Crystallography Association
beamline 17-ID in the Advanced Photon Source at Argonne
National Laboratory (Chicago, IL). Diffraction data were in-
dexed and scaled using the HKL2000 software (18) (see Table 4).
The crystals belong to the monoclinic space group P21, with two
molecules of the binary complex FAK/AMP-PNP in the asym-
metric unit. The structure of the complex was determined by the
molecular replacement method using the published coordinates
of FAK with ADP (PDB: 1MP8) (19) using the program
Phaser (20). The graphic program Coot (21) was used to make
manual adjustments to the model. The structure has been refined
eqs 3 and 4, respectively
v ¼ VA=½Kað1 þ I=KisÞ þ Aꢁ
ð3Þ
v ¼ VA=½Kað1 þ I=KisÞ þ Að1 þ I=KiiÞꢁ
ð4Þ
where Kis and Kii are respectively the apparent slope and
intercept inhibition constants.
The effect of viscosogens on the initial velocity data of FAK1
in double-reciprocal form was fitted to either eq 5 or 6
1=v ¼ ð1=kaÞð1=AÞηrel þ 1=kc þ ηrel=kd
ð5Þ
1=v ¼ ð1=ka þ ηrel=kbÞð1=AÞ þ 1=kc þ ηrel=kd
ð6Þ
˚
to 1.45 A resolution using REFMAC5 (22) to a final crystallo-
for which A is either of the substrates, ηrel is the relative viscosity,
V is the maximal velocity, and ka, kb, kc, and kd are respectively
collections of individual rate constants found in the expressions
for Ka/V and 1/V, which are dependent upon the mechanism and
substrate.
graphic R-value of 0.23 and a free R-value of 0.24. The final
structure includes two FAK1 molecules (chain A residues from
414 to 686 and chain B residues from 415 to 686) each bound with
one AMP-PNP molecule, one sulfate ion, and 201 water mole-
cules (Table 4).
Data Analysis. Initial velocity data for FAK1 were fitted to
eq 1 or 2, for single substrate and bisubstrate kinetics, respec-
tively, using Grafit 5.0.8 (Erithacus Software Ltd.)
RESULTS AND DISCUSSION
Initial Velocity Studies. Initial rate data for the FAK1-
catalyzed reaction conformed to standard Michaelis-Menten
kinetics with either MgATP or FAK-tide as the variable sub-
strate, which were evaluated by fitting of data to eq 1 over a wide
range of concentrations (0.2-5Km) (data not shown). An initial
velocity study was then performed, for which the double-
reciprocal plot is shown (Figure 1). The resulting intersecting initial
velocity pattern of FAK-tide vs MgATP indicates that FAK1
operates via a sequential kinetic mechanism. Fitting of the initial
velocity data to eq 2 resulted in the kinetic parameters in Table
1. The turnover number for FAK1 using FAK-tide as substrate
was very small (kcat = 0.052 ( 0.001 s-1), indeed, significantly
smaller than the turnover numbers of other protein kinases using
peptide substrates of similar size [Csk, kcat = 0.8 s-1 (10); S6K1,
v ¼ VA=ðKa þ AÞ
ð1Þ
v ¼ VAB=ðKiaKb þ KbA þ KaB þ ABÞ
ð2Þ
for which V is the maximal velocity, A is the concentration of
MgATP, B is concentration of FAK-tide, Ka and Kb are
apparent Michaelis constants, and Kia is the apparent dissocia-
tion constant for MgATP.
Product and dead-end inhibition studies apparently conform-
ing to competitive and noncompetitive inhibition were fitted to
kcat = 0.2 s-1 (23); PDK1, kcat = 0.4 s-1 (24); and AKT, kcat
=
0.4 s-1 (25)]. The values of the Michaelis and dissociation
constants for both substrates were equal within experimental
error; that is, KMgATP = 1.2 ( 0.1 μM, KiMgATP = 1.3 ( 0.2 μM
and KFAK-tide = 5.6 ( 0.4 μM, KiFAK-tide = 6.1 ( 1.1 μM. This
equality would be expected for an enzyme for which catalysis is
rate-limiting and suggests a rapid-equilibrium random mechanism.
Values for kcat/KMgATP and kcat/KFAK-tide of 4.3 ꢀ 104 M-1 s-1
and 9.3 ꢀ 103 M-1 s-1 (Table 1), respectively, also reflect the
modest catalytic properties of FAK1. The curvature observed in
the plot in Figure 1 at 0.08 and 0.16 μM MgATP and at the two
lowest concentrations of FAK-tide is most likely due to experi-
mental error, as the signal/background at these lowest concentra-
tions was minimal. The apparent curvilinear nature of the initial
velocity plot could also result from a steady-state random bi bi
mechanism, although if this were the case, one expects to observe
similar curvature in double-reciprocal plots for either substrate in
F
IGURE 1: Initial velocity of the FAK1-catalyzed reaction. Experi-
mental initial rates were fitted to eq 2 as described in Materials and
Methods, and shown in double-reciprocal format, where v represents
picomoles of product per minute, and with variable concentrations of
FAK-tide at changing fixed concentrations of MgATP.
Table 1: Kinetic Parameters from Initial Velocity Studies
apparent kinetic parameters
KMgATP
KiMgATP
kcat/KMgATP
(M-1 s-1
KFAK-tide
KiFAK-tide
kcat/KFAK-tide
(M-1 s-1
varied substrate
(μM)
(μM)
)
(μM)
(μM)
)
kcat (s-1
)
MgATP
FAK-tide
1.2 ( 0.1
1.3 ( 0.2
(4.3 ( 0.4) ꢀ 104
5.6 ( 0.4
6.1 ( 1.1
(9.3 ( 0.7) ꢀ 103
0.052 ( 0.001

Article
Biochemistry, Vol. 49, No. 33, 2010 7155
Table 2: Dead-End and Product Inhibition Patterns of FAK1
inhibition pattern
fixed substrate
pattern type
Kisa (μM)
Kiib (μM)
Kmc (μM)
AMP-PNP vs MgATP
AMP-PNP vs peptide
Phe/Tyr-FAK-tide vs MgATP
Phe/Tyr-FAK-tide vs peptide
MgADP vs MgATP
peptide, 5 μM
MgATP, 2 μM
peptide, 5 μM
MgATP, 2 μM
peptide, 5 μM
MgATP, 2 μM
peptide, 5 μM
MgATP, 2 μM
Cd
4.5 ( 0.4
12 ( 4
1.0 ( 0.1
5.2 ( 0.4
1.2 ( 0.1
7.4 ( 0.3
1.2 ( 0.04
5.3 ( 0.3
1.2 ( 0.1
2.6 ( 0.2
NCe
NC
C
12 ( 1
700 ( 100
900 ( 300
700 ( 100
1.4 ( 0.1
6.3 ( 3.0
530 ( 180
220 ( 20
C
MgADP vs peptide
NC
NC
C
4.3 ( 0.5
640 ( 100
P-FAK-tide vs MgATP
P-FAK-tide vs peptide
aApparent slope inhibition constant. bApparent intercept inhibition constant. cMichaelis constant for the varied substrate. dDenotes competitive
inhibition. eDenotes noncompetitive inhibition.
the product and dead-end inhibition data, which was not observed
(Figures 2S and 3S in Supporting Information). Data in Figure 1
could not be successfully fitted to an initial velocity expression for a
steady-state random mechanism (v = [(aB þ cB2)A þ bBA2]/
[d þ fB þ hB2 þ (e þ jB þ mB2)A þ (g þ kB)A2], where a-m are
collections of kinetic parameters (26)). From all considerations
above, we conclude that FAK1 does not operate via a steady-state
random mechanism.
Product and Dead-End Inhibition Studies. The results of
product inhibition and dead-end inhibition experiments are
shown in Table 2 (more details in Supporting Information Table
2S and Figures 2S and 3S). MgADP is a competitive inhibitor vs
MgATP (Kis = 1.4 ( 0.1 μM) and noncompetitive vs FAK-tide.
Phospho-FAK-tide is a noncompetitive inhibitor vs MgATP and
competitive vs FAK-tide (Kis = 220 ( 20 μM). These product
inhibition patterns are consistent with either a Theorell-Chance
mechanism or a random bi bi mechanism for which the observed
noncompetitive patterns of P-FAK-tide vs MgATP and
MgADP vs FAK-tide indicate the formation of the two possible
substrate-product dead-end complexes, E-MgADP-FAK-tide
(EBQ) and E-MgATP-P-FAK-tide (EAP). From the dead-end
inhibition patterns, however, AMP-PNP is a competitive inhi-
bitor of MgATP (Kis = 4.5 ( 0.4 μM) and a noncompetitive
inhibitor vs FAK-tide (Table 2). The phenylalanine-substituted
peptide analogue of FAK-tide, Phe/Tyr-FAK-tide, exhibited
noncompetitive inhibition vs MgATP and competitive inhibition
vs FAK-tide (Kis = 700 ( 100 μM) (Table 2). These dead-end
inhibition patterns are most consistent with a random bi bi
mechanism and rule out a Theorell-Chance mechanism for which
an uncompetitive inhibition would be expected for AMP-PNP vs
FAK-tide if MgATP were the second substrate to bind.
F
IGURE 2: 31P NMR spectra at 242.76 MHz of [γ-18O]ATP and
[β-18O]ADP for various incubation times of FAK-catalyzed posi-
tional isotope exchange as outlined in Table 3. The chemical shifts are
relativetoanexternalsampleof85% H3PO4/15%D2O measuredina
separate tube. (A) Experiment 1: [γ-18O]ATP under assay buffer
conditions in the absence of FAK. The number of atoms of 18O in
each of the peaks of a doublet of the γ-phosphorus group is noted.
(B-D) Experiments 2-4: FAK-catalyzed positional isotope ex-
change of [γ-18O]ATP for incubation times of 73, 146, and 249
min, respectively.
A pair of doublets corresponding to the
[β-18O]ADP formed in the reaction is seen upfield from the
γ-phosphorus resonances. Additional details are provided in Table 3
and in the text.
The apparent values of Kis obtained from the competitive
inhibition patterns of Table 2 correspond to the true inhibition
constants for the respective binding of AMP-PNP and Phe/Tyr-
FAK-tide to free enzyme (KiAMP-PNP = 4.5 ( 0.4 μM and
KiPhe/Tyr-FAK-tide = 700 ( 100 μM). In the case of a rapid-
equilibrium random bi bi mechanism, these Ki values may also be
afforded by the apparent Kis and Kii values of the two non-
competitive inhibition patterns, which provides a useful valida-
tion of the kinetic mechanism. For the AMP-PNP vs FAK-tide
pattern, KiAMP-PNP = Kis/(1 þ [MgATP]/KiMgATP) = Kii /(1 þ
[MgATP]/KMgATP), from which we may obtain respective values
of KiAMP-PNP = 5 ( 3 μM and 4.7 ( 0.8 μM, which is in excellent
agreement with the inhibition constant obtained from the
competitive inhibition data (KiAMP-PNP = 4.5 ( 0.4 μM). From
the noncompetitive dead-end inhibition data of Phe/Tyr-
FAK-tide vs MgATP, KiPhe/Tyr-FAK-tide = Kis/(1 þ [FAK-tide]/
KiFAK-tide) = Kii/(1 þ [FAK-tide]/KFAK-tide), from which one
may solve respective values of KiPhe/Tyr-FAK-tide = 500 ( 300 μM
and 400 ( 100 μM, which are in reasonable agreement with the
value KiPhe/Tyr-FAK-tide = 700 ( 100 μM obtained from the
competitive inhibition data for the peptide inhibitor.
The dead-end substrate-product complexes E-MgADP-
FAK-tide (EBQ) and E-MgATP-phospho-FAK-tide (EAP)
must form to account for the noncompetitive product inhibition
patterns observed. The formation of the MgATP-P-FAK-tide
(EAP) complex is surprising, because intuitively one would
expect untoward steric interactions between the γ-phosphate of
MgATP and the phosphorylated peptide in the active site. It is
unlikely that this dead-end complex has physiological signifi-
cance. Despite this, E-MgATP-phosphocreatine and E-MgADP-
creatine complexes have precedence in creatine kinase (27). The
E-MgADP-peptide (EBQ) complex has been observed for other
kinases, specifically for the E-MgADP-Ser-peptide complex of

7156 Biochemistry, Vol. 49, No. 33, 2010
Schneck et al.
Table 3: Positional Isotope Exchange of [γ-18O]ATP Catalyzed by Human FAKa
fraction of
fraction of
catalytic
rate of MgATP rate of MgADP
incubation
time (min)
exchange
exchange
formation
(vcat) μM/s)f
d
)
g
expt
% of [γ-18O4]ATPb % of [γ-18O316O]ATPb reaction (Fx)c reaction (Fcat
(vex) (μM/s)e
kx/kcat
1
2
3
4
0
73
74.4
72.8
72.3
71.1
25.6
27.2
27.7
28.9
0.0
0.0
0.0
0.0
0.0
0.040
0.052
0.080
0.25
0.35
0.47
0.0082
0.005
0.0037
0.056
0.04
0.027
0.15
0.12
0.14
146
249
0.14 ( 0.01
a40 mM HEPES (pH 7.2), 10 mM MgCl2, 1 mM DTT, 1 mM [γ-18O]ATP, 1.5 mM FAK-tide, and 100 nM FAK (21 ( 1 ꢀC). For expt 2-4, quantification of
MgADP concentrations as determined by coupled enzyme assay indicated fractions of catalysis of 0.25, 0.41, and 0.56, respectively. bSums of peak intensities
of data in Figure 2. cFraction of [γ-18O]ATP exchanged; Fx = ([γ-18O4]t - [γ-18O4]0)/([γ-18O4]0 - [γ-18O4]eq), where [γ-18O4]0 and [γ-18O4]eq = 74.4% and
24.8%, respectively. dSums of β-phosphate peak intensities/total β- and γ-phosphate peak intensities, which were corrected for MgADP (7%) in the
[γ-18O]ATP sample. eDetermined from the method of Litwin and Wimmer (47), wherein vex = [X/ln(1 - X)](ATP0/t)[ln(1 - Fx)] wherein X = Fcat, t =
reaction time, and ATP0 = 1000 μM. fDetermined from ratios of peak intensities of sums of β-phosphate peaks divided by total peak intensities of γ- and
g
β-phosphate signals. vex/vcat
.
the cAMP-dependent protein kinase (28). From the apparent
product inhibition constants in Table 2, the equilibrium constants
describing the formation of the [E-MgATP][P-FAK-tide]/[E-
MgATP-P-FAK-tide] and [E-MgADP][FAK-tide]/[E-MgADP-
FAK-tide] complexes are respectively KIap = 190 ( 60 μM and
KIbq = 1.6 ( 0.2 μM (Table 1S in Supporting Information). The
E-MgATP-P-FAK-tide complex is therefore 100-fold less stable
than the E-MgADP-FAK-tide complex, and moreover, the
E-MgADP-FAK-tide is slightly more stable than the E-MgADP
complex (Kiq = 2.5 ( 1.0 μM). This latter result demonstrates
the relative stability of the ternary E-MgADP-FAK-tide complex
compared to E-MgADP.
both decrease as reaction times are elongated. The observed PIX
most likely arises from the reverse of the E-MgADP-P-FAK-tide
complex to substrates rather than the re-formation of exchanged
MgATP from the dissociated MgADP and P-FAK-tide pro-
ducts. This is evident from two observations: (1) under the
experimental conditions of PIX, the ratio of the rate of the
reverse:forward reactions, found to be 0.048,2 was far lower than
the average value of kx/kcat = 0.14 ( 0.01, and (2) the formation
of the dead-end E-MgADP-P-FAK-tide complex, which con-
tributes to the observed diminution of values of vex and vcat as the
MgADP product accumulates (Fcat > 0.25), would further
retard, if not ablate, the formation of MgATP from MgADP
and P-FAK-tide. The value of kx/kcat demonstrates that the
regeneration of ATP from the E-MgADP-P-FAK-tide enzyme
form occurs 14% as frequently as the formation of MgADP
product. This indicates that despite the low value of kcat, the
ability of the E-MgADP-P-FAK-tide complex to re-form sub-
strates suggests that a slow step follows the formation of this
ternary enzyme-product complex and that the phosphoryl
group transfer reaction is reversible.
Effects of Viscosogens on the Steady-State Kinetic
Parameters. The steady-state kinetic parameters for FAK1-
catalyzed phosphorylation of FAK-tide were measured in buffers
of varying relative viscosity (ηrel = 1.0-3.8). It is expected that
increasing concentrations of a microviscogenic agent, such as
sucrose, will diminish the rate constants, specifically, if the
substrate-binding steps, product-release steps, or a protein con-
formational change that is affected by solvent, comprises slow
kinetic steps in the respective kcat/Km or kcat expressions (30).
However, if phosphoryl group transfer is the sole rate-determin-
ing step, then no viscosity effects would be expected (30, 31).
Macroviscogenic agents, such as Ficoll 400, affect the viscosity of
the solution without altering small molecule diffusional rates (32),
such that viscosity effects arising from sucrose, but not Ficoll 400,
would result from changes to the diffusional rates of substrates
and products.
Positional Isotope Exchange of [γ-18O4]ATP. The ability
of FAK1 to catalyze βγ-bridge:β-nonbridge positional oxygen
exchange (29) of [γ-18O]ATP would provide kinetic information
about the partitioning of the E-MgADP-P-FAK-tide ternary
complex to either form products or to regenerate its MgATP and
FAK-tide substrates. FAK1-catalyzed PIX of [γ-18O]ATP was
determined at saturating concentrations of both substrates and
analyzed by 31P NMR using the method of Cohn and Hu (16).
PIX would be evident by the progressive enrichment of the peak
intensity of the [γ-18O316O]ATP species at the “expense” of the
parent [γ-18O4]ATP species, which arises as a result of the
scrambling of the single βγ-bridge 18O atom into the nonbridge
position of the β-phosphate of the enzyme-bound [β-18O]-
MgADP prior to the re-formation of [γ-18O]ATP.
From 31P NMR spectra, the substitution pattern of 18O atoms
in the γ-phosphate of the synthetic [18O]ATP was found to be
γ-18O4, 74.4%, and 8O316O, 25.6% (Figure 2, Table 3). As shown
in Figure 2, FAK1 catalyzes the time-dependent decrease and
increase, respectively, of the [γ-18O4]ATP and [18O316O]ATP
forms, indicating that PIX has occurred. A tabulation of these
data is found in Table 3. The rate of PIX of [γ-18O4]ATP is
12-15% of the rate of MgATP hydrolysis throughout the time
course of the study, and this ratio of rates of exchange to catalysis
is invariant as the reaction progresses (average of kx/kcat = 0.14
( 0.01), although the rates of exchange (vex) and catalysis (vcat
)
Shown in Figure 3 are the initial velocity curves for variable
FAK-tide at fixed saturating MgATP (10 μM). It is evident that
kcat decreases as sucrose concentration increases and that this
effect is not observed at the highest concentration of Ficoll 400
(10%), despite its higher relative viscosity than 40% sucrose.
From initial velocity data for each substrate at changing-fixed
concentrations of sucrose, replots of the apparent values of kcat vs
relative viscosity (ηrel) for either substrate and kcat/KMgATP were
linear with a negative slope, while the plot of kcat/KFAK-tide vs ηrel
2By use of the two coupled-enzyme systems, we compared in separate
reaction mixtures the FAK1-catalyzed production of MgATP from 500
μM MgADP and 500 μM phospho-FAK-tide and MgADP from 1000
μM MgATP and 1500 μM FAK-tide with 100 nM FAK1 (40 mM
HEPES (pH 7.2), 10 mM MgCl2, 3 mM DTT, 21 ( 1 ꢀC, 50-120 min
incubation). Under these conditions, the ratio of the initial rate of the
reverse reaction was much less than that of the forward reaction (vrev
vfor = 0.048).
/

Article
Biochemistry, Vol. 49, No. 33, 2010 7157
F
IGURE 3: Effect of increasing viscosity (both microscopic and
macroscopic) on the FAK-catalyzed reaction at 12 μM MgATP
conditions. Dataare from fitsofinitial velocitydatato eq 1, where v is
in picomoles of product per minute.
had an apparent slope of zero (data not shown). As expected,
values of kcat and kcat/Km for either substrate were not affected by
increasing concentrations of the macroviscogen Ficoll 400. The
linear diminution in values of kcat and kcat/KMgATP, which were
not observed for kcat/KFAK-tide, indicated that the kinetic steps
involving the binding of MgATP, but not FAK-tide, and one or
more postcatalytic steps are affected by solvent viscosity.
Further analysis of the viscosity dependence of these initial
velocity data in double-reciprocal form, that is, a plot of 1/v vs 1/
MgATP at changing-fixed values of ηrel, conformed to an
intersecting pattern as both slope and intercept effects were
observed as values of ηrel were increased (Figure 4A). Fitting
of the data in Figure 4A to eq 5 resulted in a slope factor of ka =
0.036 ( 0.003 μM-1 s-1 and two intercept factors of kc = 0.05 (
0.01 s-1 and kd = 0.08 ( 0.02 s-1. The plot of 1/v vs 1/FAK-tide
at changing-fixed values of ηrel conformed to an apparent parallel
pattern for which an intercept effect was observed as values of ηrel
were increased (Figure 4B). From fitting of these data to eq 6, two
slope factors of ka = 40 ( 20 mM-1 s-1 and kb = 0.4 ( 1 μM-1 s-1
and two intercept factors of kc = 0.2 ( 0.1 s-1 and kd = 0.083 (
0.007 s-1 were obtained. That the values of kd in both plots are
equal suggests that the viscosogen exerts an equivalent effect on
one or more rate constants within the expression for kcat. Also, the
lack of an apparent slope effect from the changing-fixed levels of
the viscosogen on the plot of 1/v vs 1/FAK-tide reflects the small,
and poorly determined, value of the viscosity-dependent term kb,
which is obscured by the viscosity-independent slope factor, ka.
Combined, these data show that solvent viscosity had no effect on
kcat/KFAK-tide, while kcat and kcat/KMgATP were both decreased
linearly at increasing solvent viscosity. From the PIX and viscosity
data, it may be concluded that enzyme turnover (kcat) is rate-limited
by both reversible phosphoryl group transfer and a slow step that
follows phosphoryl group transfer but precedes product release.
The Crystal Structure of AMP-PNP-Complexed FAK1.
We have determined the crystal structure of the catalytic domain
of active, phosphorylated human FAK1 from crystals grown
from FAK1 incubated with AMP-PNP. Crystallographic data
are shown in Table 4 and details summarized in the Supporting
Information section. As shown in Figure 5A, there are two kinase
domains of phosphorylated FAK1/AMP-PNP binary complex
molecules in the asymmetric unit. Overall, as observed previously
for the FAK1-ADP complex (19), the FAK1 catalytic core
has the typical bilobal kinase fold (33, 34), comprised of an
N-terminal lobe (residues 414-500) formed by a β-sheet and
single R-helix, helix C, and a C-terminal R-helical lobe domain
(residues 506-686) (Figure 5A, Supporting Information
F
IGURE 4: (A) Double-reciprocal plot of initial velocity data of FAK
with changing fixed levels of sucrose (0-40%; ηrel = 1.0-3.8),
variable concentrations of MgATP (0.1-12 μM), and 25 μM
FAK-tide. The lines drawn through the experimental data points
were fitted to eq 5. (B) Double-reciprocal plot of initial velocity data
of FAK with changing fixed levels of sucrose (0-40%; ηrel
=
1.0-3.8), variable concentrations of FAK-tide (0.39-50 μM), and
12 μM MgATP. The lines drawn through the experimental data
points were fitted to eq 6.
Table 2S). The two molecules interact mostly via two identical
loop regions (440-450) that comprise essential parts of both
active sites. The active site is situated at the interface between the
lobes and includes the ATP (AMP-PNP) binding site and the
kinase activation loop segment (residues 564-584; Figure 5B).
The AMP-PNP is clearly identified by electron density in the
ATP-binding pocket and interacts with the protein both directly
and through water-mediated contacts,3 although the γ-phos-
phate group is not evident (Supporting Information Table 2S).
The R- and β-phosphates of AMP-PNP interact with the activa-
˚
˚
tion loop residues Lys581 (2.7 A to O2A) and Lys 583 (2.7 A
to N3B).
Comparison of this AMP-PNP-FAK1-complexed structure to
the published structures of inactive, unphosphorylated
FAK1 (5, 19) reveals one significant conformational difference.
As shown in the AMP-PNP-complexed structure in Figure 5B,
the activation loop is mostly ordered, with residues visibly
interacting with the R- and β-phosphates of AMP-PNP (amino
acids 564-584), and apparently covers the ATP-binding site.
However, the activation loop in the inactive FAK1 structure is
oriented away from the ATP-binding site and is essentially part of
3Crystallography details about the AMP-PNP-FAK1 interactions are
summarized in the Supporting Information section.

7158 Biochemistry, Vol. 49, No. 33, 2010
Schneck et al.
and dead-end inhibition, PIX, and viscometric studies and is
consistent with the structural evidence for activation loop move-
ment is shown in Scheme 1.
Table 4: Data Collection and Refinement Statistics for Crystallographic
Data for the Kinase Domain of Focal Adhesion Kinase-1 (411-686) in
Complex with AMP-PNP
The rate constants for the formation and breakdown of all
binary complexes are expected to be much more rapid than
catalysis, that is, k1, k2, k7, k8, and k19-k27 are faster than k13.
Enzyme forms denoted with an asterisk indicate steps that follow
a conformational change resulting in a “closed” activation loop.
Closure or reopening of the kinase activation loop (such as EPQ*
f EPQ), described by the k4, k12, and k15 steps in Scheme 1, may
be similar in rate to the catalytic steps and are likely to be affected
in the presence of the sucrose viscosogen.
Based on the mechanism shown in Scheme 1, expressions for
the initial velocity of FAK1 (eq 7), and by using the method of net
rate constants (36), expressions for Michaelis and dissociation
constants of substrates, kcat, viscosity effects, and the rates of PIX
to turnover were derived below (eq 7-16). It is assumed that all
binary enzyme-substrate, enzyme-product, and the two dead-
end enzyme-substrate-product complexes are in rapid equilib-
rium and that the expression for V/Et is described by three slow
kinetic steps k13, k14, and k15 when k3 and k11 . k12, k13, k14, and
k15, and k16 is negligible:
Data Collection
space group
P21
cell dimensions
˚
a, b, c (A)
39.7, 176.8, 39.8
90.0, 107.9, 90.0
1.45 (1.51-1.45)
0.049 (0.287)
31.0 (2.53)
R, β, γ (deg)
resolution (A) (last shell)
Rsyma (last shell)
˚
I/σI (last shell)
completeness (%) (last shell)
redundancy (last shell)
86.4 (53.0)
3.3 (2.6)
Refinement
˚
resolution (A)
25-1.45
72951
no. of reflections
c
Rworkb/Rfree
0.231/0.241
no. of atoms
protein
4204
54
ligand
water/solvent
B-factors
protein
206
v ¼ VABEt=ðKiaꢂKbð1 þ P=Kip þ Q=KiqÞ þ KaꢂBð1 þ Q=KIbq
Þ
23.9
52.2
53.1
ligand
þ KbAð1 þ P=KIapÞ þ ABÞ ð7Þ
water/solvent
rms deviations
for which V/Et = k13k15/(k13 þ k14 þ k15), Kia* = k2k4/k1k3, Kib =
k8/k7, Kip = k23/k24, Kiq = k19/k20, KIap = k26/k25, KIbq = k28/k27,
and Ka* and Kb will be defined below.
˚
bond lengths (A)
0.005
0.878
bond angles (deg)
The effects of the viscosogen sucrose on the kinetic parameters
kcat, kcat/KMgATP, and kcat/KFAK-tide were performed at saturat-
ing concentrations of the fixed substrate, such that the expression
for kcat/KMgATP describes the lower pathway in Scheme 1 (E-FAK-
tide combines with MgATP), while kcat/KFAK-tide describes the
upper pathway (E-MgATP* combines with FAK-tide). Expres-
sions for these three kinetic parameters are found in eq 8-10, with
the assumptions that k13, k14, and k15 are considerably slower than
the steps representing closing of the activation loop (k3 and k11) and
that the substrate desorption step k6 is much faster than reopening
of the activation loop (k12 and k15) and phosphoryl transfer (k13
and k14).
the C-terminal R-helical lobe domain, leaving the ATP-binding
site open to solvent. Interestingly, only one major conformational
change, that of the activation loop, has been observed for FAK1.
The structural analyses of most protein kinases (35) have
identified two major conformational changes upon ATP binding:
the activation loop movement as seen for FAK1 and an addi-
tional significant conformational change of the C-helix. The
movement of the C-helix toward the active site promotes the
formation of the salt bridge between the active site residues
Lys454 and Glu471 that characterizes the active form of protein
kinases. This movement of the C-helix is not observed in FAK1
structures; the C-helix conformation of FAK1 structures remains
unchanged (5). This suggests that the conformational change
undergone by the activation loop, presumably upon ligand
binding, is the major conformation change to be expected during
the catalytic cycle of the isolated FAK1 kinase domain. It is
tempting to speculate that this single conformational change
which attends the binding of MgATP, without a conformational
contribution from the C-helix, is reflected in the viscosity
dependence of MgATP binding but lack of same with the larger
FAK-tide substrate. Furthermore, movement of the C-helix in
other protein kinases which employ ordered mechanisms may
help to elaborate a peptide binding site in the E-MgATP form
and thereby militate an ordered kinetic mechanism. Also of note
in the AMP-PNP-complexed structure reported here is that
although the protein is phosphorylated (data not shown) and
fully active, the two phosphorylated tyrosines (Tyr576, Tyr577),
along with five residues from 573 to 578 in the activation segment,
are disordered and not resolved in our structure, which suggests
that phosphorylation alone may not be sufficient for formation
of the active, “closed” conformation.
kcat=KMgATPꢂ ¼ ½k9k11ðk13k15=ðk14 þ k15ÞÞꢁ=½k10ðk12 þ k13k15=
ðk14 þ k15ÞÞ þ k11ðk13k15=ðk14 þ k15ÞÞꢁ ꢃ k9k11=ðk10 þ k11Þ
ð8Þ
kcat=KFAK-tide ¼ ½k5ðk12 þ k13k15=ðk14 þ k15ÞÞðk3=k4Þꢁ=
ð½k6 þ k12 þ k13k15=ðk14 þ k15Þꢁð1 þ k3=k4ÞÞ ꢃ ðk3k5=k4Þ=
½ðk6ðk14 þ k15Þ=k13k15 þ 1Þð1 þ k3=k4Þꢁ
ð9Þ
kcat ¼ 1=½1=k3 þ 1=k11 þ ðk14 þ k15Þ=k13k15 þ 1=k15ꢁ
ꢃ k13k15=ðk13 þ k14 þ k15Þ
ð10Þ
As one expects that the viscosogen will lower the rates of
substrate binding, product desorption steps, and, possibly, con-
formational movements of the activation loop by the amount of
relative viscosity (for example, k9o = k9/ηrel, where k9 is the rate
constant at no viscosogen), the expressions for the effect of
sucrose on the three kinetic parameters in eq 8-10 are found in
eq 11-13.
A Comprehensive Model of FAK1 Catalysis. A mechan-
istic model that accounts for the initial velocity, product
kcat=K
¼ k9k11=ðk10 þ k11Þηrel
ð11Þ
MgATPꢂðηrel
Þ

Article
Biochemistry, Vol. 49, No. 33, 2010 7159
k11) = 0.036 ( 0.003 μM-1 s-1 was similar to the apparent value
of kcat/KMgATP* = 0.04 ( 0.01 μM-1 s-1 (at [sucrose] = 0). It is
noteworthy that if the expression for kcat/KMgATP* found in eq 8
was merely the reciprocal of the dissociation constant for
MgATP binding to E-FAK-tide (k9/k10), there would be no
effect of viscosity on this expression as sucrose would be expected
to lower values of both k9 and k10 equally. The observed slope
effects seen in Figure 4A likely arise because of the additional
effect of the viscosogen on the closing of the activation loop as
described by k11, giving rise to the expression in eq 11. The
resulting values of kc = k13 = 0.05 ( 0.01 s-1 and kd
=
k13k15/(k13 þ k14) = 0.08 ( 0.01 s-1 reflect low values of the rate
constants of trans-phosphorylation and opening of the activation
loop following catalysis.
The plot of 1/v vs 1/FAK-tide at changing-fixed values of ηrel
conformed to an apparent parallel pattern for which an intercept
effect was observed as values of ηrel were increased (Figure 4B).
For FAK-tide, a higher value for kc = k13 = 0.2 ( 0.1 s-1 was
obtained, with the same value found for kd = k13k15/(k13 þ k14) =
0.083 ( 0.007 s-1. The parameter ka = k6/k5k13 = Kib/k13
40 ( 20 μM-1 s-1, which from the experimental value of Kib
=
=
6.1 ( 1.1 μM from the initial velocity data, upon solution also
results in a value of k13 = 0.2 s-1. The absence of a slope effect in
the plot of 1/v vs 1/FAK-tide arises from the comparatively
modest size of the bimolecular rate constant of FAK-tide
combining with E-MgATP (kb = k5 = 40000 M-1 s-1) and
reflects the small, and poorly determined, value of the viscosity-
dependent term, kb.
Importantly, the results of the effects of the viscosogen sucrose
on the initial velocity data, while in accord with a random bi bi
kinetic mechanism, do not support a kinetic mechanism for
which phosphoryl transfer is the sole rate-limiting step in
catalysis. This is because changes in solvent viscosity are not
expected to affect the rates of binding and desorption for
substrates and products that are in true rapid equilibrium with
the enzyme. That is, for kinetic events that are far more rapid
than the rate of phosphoryl transfer, the observation of signifi-
cant viscosogenic effects on both kcat and kcat/KMgATP indicates
that one or more steps involved in the binding of MgATP or a
postbinding conformational change and product release are slow
enough to influence the values of these kinetic parameters under
the assumption that catalysis is also very slow.
F
IGURE 5: (A) Crystal structure of the kinase domain of human focal
adhesion kinase (chain A, amino acids 414-686; chain B, amino
acids 415-686) in complex with AMP-PNP. The N-terminal lobe is
formed by residues 414-500 and the C-terminal lobe by residues
506-686. The two independent molecules interact mainly through
residues 448-450 in β-strands 3. (B) Structural overlay of inactive,
unphosphorylated FAK1 (2J0J) and active, phosphorylated FAK1.
The activation segment in the active, AMP-PNP complex is close to
the N-lobe and covers the ATP-binding site while in the inactive,
unphosphorylated FAK1 the activation loop is removed from the
active site. Colored orange is the modeled trajectory of the section of
the activation loop not seen in the structure.
The observation of FAK1-catalyzed PIX of [γ-18O]ATP is also
consistent with the lack of a purely rapid-equilibrium mechanism,
since a rapid release of [β-18O]ADP from the E-MgADP-P-FAK-
tide complex would not support the re-formation of scrambled
[γ-18O]ATP by facile reversal of this complex. From Scheme 1, one
may write an expression for the ratio of the rate of PIX to turnover
by the use of the method of net rate constants (36) in order to
integrate these data with the outcome of the viscosity effect data to
provide solutions for the rate constants k13, k14, and k15:
0
kx=kcat ¼ k140=k15
kcat=K
¼ ðk3k5=k4Þ=½ðk6ðk14ηrel þ k15Þ=
k13k15 þ 1Þð1 þ k3=k4Þꢁ
FAK-tideðηrel
Þ
¼ k14k12k10=½ðk10ðk12 þ k13Þ þ k11k13Þk15ꢁ
¼ ðk14=k15Þ=½1 þ ðk13=k12Þð1 þ k11=k10Þꢁ
ð12Þ
ð13Þ
kcatðη ¼ k13k15=½ðk13 þ k14Þη þ k15ꢁ
ꢃ ðk14=k15Þ=½1 þ ðk13=k12Þꢁ
ð14Þ
relÞ
rel
Accordingly, the parameters ka-kd obtained from fitting of
the double-reciprocal data in Figure 4 to either eq 5 or eq 6 may
now be related to the individual rate constants of Scheme 1 and in
eq 11-13. From the double-reciprocal analysis of viscosity effects
We assume that k10 > k11 and that k12 ≈ k15 as both steps comprise
the opening of the activation loop from the respective E-MgATP-
FAK-tide and E-MgADP-P-FAK-tide complexes.
From the average experimental value of kx/kcat = (k14/k15)/
[1 þ (k13/k12)] = 0.14 ( 0.01, wherein k12 ≈ k15, and from kinetic
on variable MgATP (Figure 4A), the value of ka = k9k11/(k10
þ

7160 Biochemistry, Vol. 49, No. 33, 2010
Schneck et al.
Scheme 1
Scheme 2: Kinetics of the Ternary Complexes of FAK1a
aIn this scheme, EAB* represents the MgATP-FAK-tide-bound complex with the activation loop “closed”, EPQ* represents the
MgADP-phospho-FAK-tide-bound complex with the activation loop “closed”, and EPQ represents the MgADP-phospho-FAK-tide-
bound complex with the activation loop in the “open” confirmation.
constants obtained from the viscosity data for MgATP above
(kc = k13 = 0.05 ( 0.01 s-1 and kd = k13k15/(k13 þ k14) = 0.08 (
0.01 s-1), we may solve for values of these three rate constants:
k13 = 0.05 ( 0.01 s-1, k14 = 0.02 ( 0.05 s-1, and k15 = 0.12 (
Normally, one expects that for a purely rapid-equilibrium
random bi bi mechanism for which catalysis is completely rate-
limiting that Michaelis and dissociation constants for substrates
will be identical (37), as observed experimentally for FAK1. The
availability of solved rate constants for the slow kinetic steps
represented by k13, k14, and k15 now allows evaluation of the
virtual identity of the Michaelis and dissociation constants of the
substrates for FAK1 for which we have demonstrated that the
rate of a postcatalytic conformational change step is slower than
phosphoryl transfer. The expressions for the Michaelis constants
under conditions that the fixed substrate is saturating are found
in eqs 15 and 16.
0.04 s-1. By use of eq 14, we may calculate kcat = 0.03 ( 0.02 s-1
,
from these three solved rate constants, which compares favorably
with the apparent value of kcat = 0.04 ( 0.004 s-1 obtained from
the viscosity data with MgATP above. A similar assessment with
values from the viscosity data with FAK-tide where k13 = 0.2 (
0.1 s-1, k13k15/(k13 þ k14) = 0.083 (0.007 s-1, andk14/(k13 þ k15) =
0.14 ( 0.01, we calculate values of k14 = 0.04 ( 0.04 s-1 and k15
=
,
0.1 ( 0.01 s-1 from which we may obtain kcat = 0.06 ( 0.04 s-1
from these three solved rate constants, which compares favorably
with the apparent value of kcat = 0.06 ( 0.001 s-1, obtained from
the viscosity data with FAK-tide. It is likely that this value of
kcat = 0.06 ( 0.001 s-1 is more reflective of the true kcat of FAK1
because (a) a limiting value of kcat = 0.052 ( 0.001 s-1 was
obtained from the initial velocity data and (b) the changing-fixed
level of MgATP ([MgATP] = 10KMgATP) in this study is a more
saturating level than the counterpart value of changing-fixed
FAK-tide ([FAK-tide] = 5KFAK-tide) in the data of Figure 4A.
Accordingly, we propose that the limiting values of the rate
constants of k13 = 0.2 ( 0.1 s-1, k14 = 0.04 ( 0.04 s-1, and
k15 = 0.1 ( 0.01 s-1 are the more representative from our studies.
From this it is evident that phosphoryl transfer is faster in the
forward reaction than the reverse reaction and is twice the rate of
the postcatalytic reopening of the activation loop. Scheme 2
summarizes the relative rates of the ternary complexes vs the
putative conformational change in the k15 step determined by the
viscosity and PIX experiments.
KMgATPꢂ ¼ ½k13k15=ðk13 þ k14 þ k15Þꢁ½ðk10 þ k11Þ=k9k11ꢁ
¼ 1:2 ( 0:1 μM
ð15Þ
KFAK-tide ¼ ½k6ðk15 þ k14Þð1 þ k3=k4Þꢁ=
½ðk13 þ k14 þ k15Þðk3k5=k4Þꢁ ¼ 5:4 ( 0:5 μM
ð16Þ
These expressions may now be compared with their counter-
part substrate dissociation constants of KiMgATP* = k2k4/k1k3 =
1.3 ( 0.2 μM and KiFAK-tide = k8/k7 = 6.1 ( 1.1 μM. When
k
13 > k15 > k14 and k10 > k11, KMgATP* approximates k10k15/
k9k11, which would be equal to KiMgATP* = k4k2/k3k1 when the
rate constants for equivalent kinetic steps in the two pathways are
all equal (k1 = k9, k2 = k10, k3 = k11, and k4 = k12 ∼ k15, as both
of these latter rate constants describe opening of the activation
loop). In kind, the expression for KFAK-tide equals (k6/k5)[(k14þ
k15)/(k13 þ k14 þ k15)] when k3/k4 . 1 and k13 > k15 > k14 and
will be nearly equal to KiFAK-tide when k5 = k7 and k6 = k8. The

Article
Biochemistry, Vol. 49, No. 33, 2010 7161
equality of the Michaelis and dissociation constants suggests that
the rates of the commensurate kinetic steps in the two branches of
the random pathway are also equal.
postcatalytic step (product release) was slightly slower than the
rate of phosphoryl transfer. Both the Cdk5/p25 kinase (41) and
p21-activated protein kinase (Pak2 (42)) have been characterized
as rapid-equilibrium random kinetic mechanisms using respec-
tively the tau protein and maltose-binding protein (MBP) and an
oligopeptide (LIMKtide) as substrates. For Pak2, analysis of
kinetic parameters by viscosity effects indicated that for MBP the
rate of phosphoryl transfer was twice as fast as desorption of the
protein product, giving rise to an effect on kcat, while for the
LMNtide substrate, no viscosity effect was evident on kcat as
product release was 5-fold greater than the rate of catalysis. An
important analogue to the FAK1 mechanism is found in the
kinetic data of the extracellular-regulated kinase-2 (Erk2 kinase)
using the protein substrate Ets(delta)138, for which a solvent
viscosity effect is found for both values of kcat/Km and kcat and
from which the authors cite the existence of a random bi bi
mechanism which contains a rate-limiting conformational
change following substrate binding (43).
Through a series of elegant studies, Cook and colleagues have
shown that the cAMP-dependent protein kinase operates via a
steady-state random mechanism using a serine-containing pep-
tide substrate (28), for which isotope partitioning studies revealed
that, at high concentrations of Mg2þ, MgATP and serine-peptide
desorb from central complexes more slowly than catalysis (44).
Interestingly, the kinetic mechanism of the reverse reaction of
cAMP-dependent protein kinase was shown to vary with changes
of pH, as an equilibrium-ordered mechanism (phosphopeptide
binding to free enzyme) at low pH converts to a steady-state
random mechanism for which MgADP now preferentially binds
first as the pH was raised (45).
The random mechanism elucidated herein for FAK1 is con-
cordant with that of other protein kinases that play similarly
crucial roles both in phosphorylation of other cellular proteins
that induce signal transduction events and by autoactivation via
phosphorylation of endogenous sequences within the kinase. The
low value of kcat for FAK1 using FAK-tide as substrate
undoubtedly contributes to the observed random mechanism in
that for the low rate of phosphoryl transfer (k13 = 0.2 s-1) the
desorption rates of the substrates and products are significantly
faster such that the kinetics approximate a rapid-equilibrium
model, as reflected in the equality of the Michaelis and dissocia-
tion constants. These kinetic findings may have a physiological
significance. First, since the FAK-tide substrate is a mimic of the
site of autophosphorylation in full-length FAK1, the binding and
processing of FAK-tide may be competing with the high local
concentration of the analogous, endogenous substrate sequence
(phosphorylated or not) within FAK1. This could account for
the poor kinetics of the “free” peptide substrate. Second, the
random mechanism observed for FAK1 and other protein
kinases with poor turnover numbers may constitute a mechanism
of substrate fidelity. One would expect that a more rapid, and
therefore less selective, rate of phosphorylation of cellular
proteins would give rise to adventitious, and possibly cata-
strophic, signaling pathways. Accordingly, a poor turnover and
non-“sticky” substrates, as observed here, guarantee a judicious
selection by FAK1 for the appropriate peptide sequence. Such
substrate fidelity, coupled with an apparent rapid-equilibrium
mechanistic component, is manifest in the remarkable specificity
of the T7 DNA polymerase, for which substrate misincorpora-
tions result in disastrous genetic consequences (46).
The internal equilibrium constant of phosphoryl transfer, Kint
=
/
k13/k14 (38), of FAK1 may now be calculated to be 0.2 ( 0.1 s-1
0.04 ( 0.04 s-1 = 5 ( 5, which is similar to the values of Kint= 1
obtained for rabbit muscle creatine kinase (39) and yeast hexoki-
nase (40). By use of the Haldane expression for a rapid-
equilibrium random bi bi mechanism (Keq = V1KpKiq/V2Kia*Kb
with the approximate ratio2 of V1/V2 = 20, and with values of
Kip = 220 μM (Supporting Information Table 1S), Kq ∼ Kiq
=
=
2.5 μM (Table 2 and Supporting Information Table 1S), Kia*
1.3 μM and Kb = 5.4 μM (Table 1)), one may calculate an
equilibrium constant for FAK1 of 1600. The observed disparity
between the “internal” and “external” equilibrium constants is
found for other kinases, including creatine kinase (Keq = 0.08)
and hexokinase (Keq ∼ 1000). The ratio V1/V2 ∼ [k13k15(k14 þ k15
þ k16)]/[k14k16(k13 þ k14 þ k15)] = 20 for the values of k13-k15
above and when k16 = 0.01, so that the finding that Keq . Kint
holds for FAK1 when the reopening of the activation loop is far
slower than its closing in both directions, such that the substrates
are “locked” into the central complexes.
Mechanistic Implications. Common to other protein ki-
nases, the binding of MgATP to either free enzyme or E-FAK-
tide likely induces a conformational change tantamount to the
movement of the activation loop of FAK1 (31). It is understood
that the activation segment samples a range of conformations
that results in the loops being disordered in many instances where
the structures have been determined (32). It is conceivable that
this range of conformations includes an extended conformation
away from the active site in a relatively “open” form, as well as a
more “closed” conformation that is most often accessed after
activation by phosphorylation (31) such that, for example, in the
case of FAK1, Tyr576 in the activation loop in the presence and
absence of the nonhydrolyzable MgATP analogue, AMP-PNP,
˚
may traverse as much as 10 A or more upon binding of the
nucleotide analogue. This suggests that this conformational
change in the formation of the binary E-MgATP complex is also
found in the ternary complex, that is, the binding of MgATP to
E-FAK-tide, which is then poised for catalysis. Moreover, it is
anticipated that the activation loop must again open (k15 ≈ 0.1
s
-1) to allow the release of the MgADP product from the ternary
E-MgADP-phospho-FAK-tide complex. Interestingly, the ob-
servation that pTyr575 and pTyr576 and other residues in their
vicinity of the activation loop are disordered in the structure of
the AMP-PNP- FAK1 complex suggests that phosphorylation of
Tyr575 and Tyr576 is not sufficient to promote ordering of the
activation loop into a closed conformation; rather, phosphoryla-
tion appears to introduce sufficient dynamics into this region of
the polypeptide chain that is consistent with it undergoing
substantial displacements depending partly on the presence or
absence of ligands in the active site.
The kinetic mechanisms utilized by the protein kinases con-
stitute a broad spectrum ranging from fully ordered, where
nucleotides are the first and last to bind, to rapid-equilibrium
random, as appears to be true for FAK1. Differences in these
mechanisms often reflect the range of values of kcat and on the
type of substrate used. Kinetics of the C-terminal Src kinase
(Csk (10)) using a poly(Glu,Tyr) substrate are consistent with a
rapid-equilibrium random mechanism. Other findings for Csk
were highly similar to FAK1; that is, values of Michaelis and
dissociation constants were identical for both substrates, and a
The current findings contribute to the idea that protein kinases
utilize a spectrum of kinetic mechanisms with a variety of

7162 Biochemistry, Vol. 49, No. 33, 2010
Schneck et al.
predominant enzyme forms present during catalysis, thus yield-
ing different transitory enzyme forms to target for inhibitor
design. An example would be that if desorption of MgADP from
a binary kinase complex were rate-limiting in the mechanism,
drug discovery efforts could be well served by seeking to find small
molecules that specifically bind to the E-MgADP form in order to
further retard the regeneration of fresh, unliganded enzyme. For
FAK1, the resolution of the kinetics of activation, catalysis, and
product release deduced by this work shows that drug discovery
efforts for this or kinetically similar protein kinases could focus on
developing inhibitors that “trap” the enzyme in the “locked
down” closed form that has ATP or ADP bound.
16. Cohn, M., and Hu, A. (1978) Isotopic (18O) Shift in 31P Nuclear
Magnetic Resonance Applied to a Study of Enzyme-Catalyzed
Phosphate-Phosphate Exchange and Phosphate (Oxygen)-Water
Exchange Reactions. Proc. Natl. Acad. Sci. U.S.A. 75, 200–203.
17. Meek, T. D., Karsten, W. E., and DeBrosse, C. W. (1987) Carbamoyl-
Phosphate Synthetase II of the Mammalian CAD Protein. Kinetic
Mechanism and Elucidation of Reaction Intermediates by Positional
Isotope Exchange. Biochemistry 26, 2584–2593.
18. Otwinowski, Z. M., and Minor, W. (1997) Processing of X-ray
Diffraction Data Collected in Oscillation Mode. Methods Enzymol.
276, 307–326.
19. Nowakowski, J., Cronin, C. N., McRee, D. E., Knuth, M. W.,
Nelson, C. G., Pavletich, N. P., Rogers, J., Sang, B. C., Scheibe,
D. N., Swanson, R. V., and Thompson, D. A. (2002) Structures of the
Cancer-Related Aurora-A, FAK, and EphA2 Protein Kinases from
Nanovolume Crystallography. Structure 10, 1659–1667.
20. McCoy, A. J. (2007) Solving Structures of Protein Complexes by
Molecular Replacement with Phaser. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 63, 32–41.
21. Emsley, P., and Cowtan, K. (2004) Coot: Model-Building Tools for
Molecular Graphics. Acta Crystallogr. D60, 2126–2132.
22. Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Refine-
ment of Macromolecular Structures by the Maximum-Likelihood
Method. Acta Crystallogr. D53, 240–255.
23. Keshwani, M. K., and Harris, T. K. (2008) Kinetic Mechanism
of Fully Activated S6K1 Protein Kinase. J. Biol. Chem. 283,
11972–11980.
24. Gao, X., and Harris, T. K. (2006) Steady-State Kinetic Mechanism of
PDK1. J. Biol. Chem. 281, 21670–21681.
25. Zhang, X., Zhang, S., Yamane, H., Wahl, R., Ali, A., Lofgren, J. A.,
and Kendall, R. L. (2006) Kinetic Mechanism of AKT/PKB Enzyme
Family. J. Biol. Chem. 281, 13949–13956.
ACKNOWLEDGMENT
The authors thank Dr. Richard Gontarek and Dr. Ryan
Kruger for thoughtful comments and WuXi AppTec Co., Ltd.
(Shanghai, China), for the preparation of [γ-18O]ATP.
SUPPORTING INFORMATION AVAILABLE
Figures 1S-3S and Tables 1S and 2S showing codon optimi-
zation of FAK(411-686), product and dead-end inhibition
results, and structural details of FAK1 interactions with AMP-
PNP. This material is available free of charge via the Internet at
http://pubs.acs.org.
26. Cook, P. F., and Cleland, W. W. (2007) Initial Velocity Studies in the
Absence of Added Inhibitors, in Enzyme Kinetics and Mechanism, pp
76-77, Garland Science, New York, NY.
27. Morrison, J. F., and James, E. (1965) The Mechanism of the Reaction
Catalysed by Adenosine Triphosphate-Creatine Phosphotransferase.
Biochem. J. 97, 37–52.
28. Cook, P. F., Neville, M. E., Jr., Vrana, K. E., Hartl, F., T., and
Roskoski, R., Jr. (1982) Adenosine Cyclic 30,50-Monophosphate
Dependent Protein Kinase: Kinetic Mechanism for the Bovine Ske-
letal Muscle Catalytic Subunit. Biochemistry 21, 5794–5799.
29. Midelfort, C. F., and Rose, I. A. (1976) A Stereochemical Method for
Detection of ATP Terminal Phosphate Transfer in Enzymatic Reac-
tions. Glutamine Synthetase. J. Biol. Chem. 251, 5881–5887.
30. Brouwer, A. C., and Kirsch, J. F. (1982) Investigation of Diffusion-
Limited Rated of Chymotrypsin Reactions by Viscosity Variation.
Biochemistry 21, 1302–1307.
31. Adams, J. A. (2003) Activation Loop Phosphorylation and Catalysis
in Protein Kinases: Is There Functional Evidence for the Autoinhi-
bitor Model? Biochemistry 42, 601–607.
32. Blacklow, S. C., Raines, R. T., Lim, W. A., Zamore, P. D., and
Knowles, J. R. (1988) Triosephosphate Isomerase Catalysis Is Diffu-
sion Controlled. Biochemistry 27, 1158–1167.
33. Bossemeyer, D. (1995) Protein Kinases;Structure and Function.
FEBS Lett. 369, 57–61.
34. Hanks, S. K., Quinn, A. M., and Hunter, T. (1988) The Protein Kinase
Family: Conserved Features and Deduced Phylogeny of the Catalytic
Domains. Science 241, 42–52.
35. Huse, M., and Kuriyan, J. (2002) The Conformational Plasticity of
Protein Kinases. Cell 109, 275–282.
36. Cleland, W. W. (1975) Partition Analysis and the Concept of Net
Rate Constants as Tools in Enzyme Kinetics. Biochemistry 14,
3220–3224.
37. Cook, P. F., and Cleland, W. W. (2007) Isotope Effects as a Probe of
Mechanism, in Enzyme Kinetics and Mechanism, pp 318-383,
Garland Science, New York, NY.
38. Burbaum, J., and Knowles, J. R. (1989) Internal Thermodynamics of
Enzymes Determined by Equilibrium Quench: Values of Kint for
Enolase and Creatine Kinase. Biochemistry 28, 9306–9317.
39. Nageswara Rao, B. D., and Cohn, M. E. (1979) 31P NMR of Enzyme-
bound Substrates of Rabbit Muscle Creatine Kinase. Equilibrium
Constants, Interconversion Rates, and NMR Parameters of Enzyme-
Bound Complexes. J. Biol. Chem. 256, 1716–1721.
REFERENCES
1. van Nimwegen, M. J., and van de Water, B. (2007) Focal Adhesion
Kinase: A Potential Target in Cancer Therapy. Biochem. Pharmacol.
73, 597–609.
2. Schlaepfer, D. D., Hauck, C. R., and Sieg, D. J. (1999) Signaling
through Focal Adhesion Kinase. Prog. Biophys. Mol. Biol. 71, 435–478.
3. McLean, G. W., Carragher, N. O., Avizienyte, E., Evans, J., Brunton,
V. G., and Frame, M. C. (2005) The Role of Focal-Adhesion Kinase
in Cancer;A New Therapeutic Opportunity. Nat. Rev. 5, 505–515.
4. Kornberg, L. J. (1998) Focal Adhesion Kinase and Its Potential
Involvement in Tumor Invasion and Metastasis. Head Neck 20,
745–752.
5. Lietha, D., Cai, X., Ceccarelli, D. F. J., Li, Y., Schaller, M. D., and
Eck, M. J. (2007) Structural Basis for the Autoinhibition of Focal
Adhesion Kinase. Cell 129, 1177–1187.
6. Arold, S. T., Hoellerer, M. K., and Noble, M. E. (2002) The Structural
Basis of Localization and Signalling by the Focal Adhesion Targetting
Domain. Structure 10, 319–327.
7. Hayashi, I., Vuori, K., and Liddington, R. C. (2002) The Focal
Adhesion Targeting (FAT) Region of Focal Adhesion Kinase Is a
Four-Helix Bundle That Binds Paxillin. Nat. Struct. Biol. 9, 101–106.
8. Calalb, M. B., Polte, T. R., and Hanks, S. K. (1995) Tyrosine
Phosphorylation of Focal Adhesion Kinase at Sites in the Catalytic
Domain Regulates Kinase Activity: A Role for Src Family Kinases.
Mol. Cell. Biol. 15, 954–963.
9. Toutant, M., Costa, A., Studler, J. M., Kadare, G., Carnaud, M., and
Girault, J. A. (2002) Alternative Splicing Controls the Mechanisms of
FAK Autophosphorylation. Mol. Cell. Biol. 22, 7731–7743.
10. Cole, P. A., Burn, P., Takacs, B., and Walsh, C. T. (1994) Evaluation
of the Catalytic Mechanism of Recombinant Human Csk (C-terminal
Src Kinase) Using Nucleotide Analogs and Viscosity Effects. J. Biol.
Chem. 269, 30880–30887.
11. Prowse, C. N., Hagopian, J. C., Cobb, M. H., Ahn, N. G., and Lew, J.
(2000) Catalytic Reaction Pathway for the Mitogen-Activated Protein
Kinase ERK2. Biochemistry 39, 6258–6266.
12. Waas, W. F., Rainey, M., Szafranska, A., and Dalby, K. (2003) Two
Rate-Limiting Steps in the Kinetic Mechanism of the Serine/Ther-
onine Specific Protein Kinase ERK2: A Case of Fast Phosphorylation
Followed by Fast Product Release. Cell. Mol. Biol. Lett. 8, 516–518.
13. Moffatt, J. G. (1964) A General Synthesis of Nucleosides-50 Trios-
phosphates. Can. J. Chem. 42, 599–604.
40. Wilkinson, K. D., and Rose, I. A. (1979) Isotope Trapping Studies of
Yeast Hexokinase during Steady-State Catalysis. J. Biol. Chem. 254,
12567–12572.
14. Hackney, D., Stempel, K. E., and Boyer, P. D. (1980) Oxygen-18
Probes of Enzymic Reactions of Phosphate Compounds. Methods
Enzymol. 64, 60–83.
15. Wehlri, W. E., Verheyden, D. L. M., and Moffatt, J. G. (1965) J. Am.
Chem. Soc. 87, 2265–2277.
41. Liu, M., Choi, S., Cuny, G. D., Ding, K., Dobson, B. C., Glicksman,
M. A., Auerbach., K., and Stein, R. L. (2008) Kinetic Studies of Cdk5/

Article
Biochemistry, Vol. 49, No. 33, 2010 7163
p25 Kinase: Phosphorylation of Tau and Complex Inhibition of Two
Indicates a Steady-State Random Kinetic Mechanism. Biochemistry 27,
4795–4799.
Prototype Inhibitors. Biochemistry 47, 8367–8377.
42. Wu, H., Zheng, Y., and Wang, Z-x (2003) Evaluation of the Catalytic
Mechanism of the p21-Activated Protein Kinase PAK2. Biochemistry
42, 1129–1139.
43. Wang, Z-x., and Wu, J-w. (2007) The Complete Pathway for ERK2-
Catalyzed Reaction. Evidence for an Iso Random Bi Bi Mechanism.
J. Biol. Chem. 282, 27678–27684.
45. Qamar, R., and Cook, P. F. (1993) pH Dependence of the Kinetic
Mechanism of the Adenosine 30,50-Monophosphate Dependent Pro-
tein Kinase Catalytic Subunit in the Direction of Magnesium Ade-
nosine 50-Diphosphate Phosphorylation. Biochemistry 32, 6802–6806.
46. Tsai, Y-c., and Johnson, K. A. (2006) A New Paradigm for DNA
Polymerase Specificity. Biochemistry 45, 9675–9687.
47. Litwin, S., and Wimmer, M. J. (1979) Correction of Scrambling Rate
Calculation for Loss of Substrate. J. Biol. Chem. 254, 1859.
44. Kong, C.-T., and Cook, P. F. (1988) Isotope Partitioning in the
Adenosine 30,50-Monophosphate Dependent Protein Kinase Reaction