The structure of the proline utilization a proline dehydrogenase domain inactivated by N-propargylglycine provides insight into conformational changes induced by substrate binding and flavin reduction

Article abstract of DOI:10.1021/bi901717s

Proline utilization A (PutA) from Escherichia coli is a flavoprotein that has mutually exclusive roles as a transcriptional repressor of the put regulon and a membrane-associated enzyme that catalyzes the oxidation of proline to glutamate. Previous studies have shown that the binding of proline in the proline dehydrogenase (PRODH) active site and subsequent reduction of the FAD trigger global conformational changes that enhance PutA-membrane affinity. These events cause PutA to switch from its repressor to its enzymatic role, but the mechanism by which this signal is propagated from the active site to the distal membrane-binding domain is largely unknown. Here, it is shown that N-propargylglycine irreversibly inactivates PutA by covalently linking the flavin N(5) atom to the ε-amino of Lys329. Furthermore, inactivation locks PutA into a conformation that may mimic the proline-reduced, membrane-associated form. The 2.15A resolution structure of the inactivated PRODH domain suggests that the initial events involved in broadcasting the reduced flavin state to the distal membrane-binding domain include major reorganization of the flavin ribityl chain, severe (35°) butterfly bending of the isoalloxazine ring, and disruption of an electrostatic network involving the flavin N(5) atom, Arg431, and Asp370. The structure also provides information about conformational changes associated with substrate binding. This analysis suggests that the active site is incompletely assembled in the absence of the substrate, and the binding of proline draws together conserved residues in helix 8 and the β1-αl loop to complete the active site. 2009 American Chemical Society.

Full text of DOI:10.1021/bi901717s

560 Biochemistry 2010, 49, 560–569
DOI: 10.1021/bi901717s
The Structure of the Proline Utilization A Proline Dehydrogenase Domain Inactivated by
N-Propargylglycine Provides Insight into Conformational Changes Induced by Substrate
Binding and Flavin Reduction†,‡
Dhiraj Srivastava,§ Weidong Zhu,^ William H. Johnson Jr.,@ Christian P. Whitman,@ Donald F. Becker,^ and
John J. Tanner*,§,
§Department of Chemistry, and Department of Biochemistry, University of Missouri, Columbia, Missouri 65211, ^Department of
Biochemistry, University of Nebraska, Lincoln, Nebraska 68588, and @Division of Medicinal Chemistry, College of Pharmacy,
The University of Texas, Austin, Texas 78712
Received October 5, 2009; Revised Manuscript Received December 7, 2009
ABSTRACT: Proline utilization A (PutA) from Escherichia coli is a flavoprotein that has mutually exclusive
roles as a transcriptional repressor of the put regulon and a membrane-associated enzyme that catalyzes the
oxidation of proline to glutamate. Previous studies have shown that the binding of proline in the proline
dehydrogenase (PRODH) active site and subsequent reduction of the FAD trigger global conformational
changes that enhance PutA-membrane affinity. These events cause PutA to switch from its repressor to its
enzymatic role, but the mechanism by which this signal is propagated from the active site to the distal
membrane-binding domain is largely unknown. Here, it is shown that N-propargylglycine irreversibly
inactivates PutA by covalently linking the flavin N(5) atom to the ε-amino of Lys329. Furthermore,
inactivation locks PutA into a conformation that may mimic the proline-reduced, membrane-associated form.
˚
The 2.15 A resolution structure of the inactivated PRODH domain suggests that the initial events involved in
broadcasting the reduced flavin state to the distal membrane-binding domain include major reorganization of
the flavin ribityl chain, severe (35ꢀ) butterfly bending of the isoalloxazine ring, and disruption of an
electrostatic network involving the flavin N(5) atom, Arg431, and Asp370. The structure also provides
information about conformational changes associated with substrate binding. This analysis suggests that the
active site is incompletely assembled in the absence of the substrate, and the binding of proline draws together
conserved residues in helix 8 and the β1-Rl loop to complete the active site.
The oxidative conversion of proline to glutamate requires
the coupled activities of proline dehydrogenase (PRODH)1
and Δ1-pyrroline-5-carboxylate (P5C) dehydrogenase (P5CDH)
(Scheme 1). In some bacteria, these activities are combined in a
single polypeptide known as proline utilization A (PutA). The
polypeptide chain lengths of PutAs exceed 1000 residues, with the
PRODH and P5CDH domains appearing in the N-terminal and
C-terminal halves of the protein, respectively (1, 2). Some PutAs,
including Escherichia coli PutA, also serve as transcriptional
repressors of the put regulon, which comprises the divergently
transcribed genes encoding PutA and the high-affinity Naþ/
proline transporter, PutP, with an intervening 419 bp control
DNA region (1, 3). PutA represses transcription of putA and putP
by recognizing, via an N-terminal ribbon-helix-helix domain, a
consensus binding sequence that appears at five different loca-
tions in the put control DNA region. In the absence of proline,
PutA occupies the five operator sites, thereby blocking transcrip-
tion of the put genes (3). In the presence of proline, PutA
associates with the membrane where it couples the oxidation of
proline to P5C with reduction of the electron transport chain
(Scheme 1). Following the nonenzymatic hydrolysis of P5C, the
P5CDH domain catalyzes the NADþ-dependent oxidation of
γ-glutamate semialdehyde to glutamate (Scheme 1) (4-6). Thus,
the association of PutA with the inner membrane not only is
essential for efficient catalytic turnover but also increases
the level of expression of putA and putP by preventing
PutA from engaging its operator sites in the put control DNA
region (1).
†This research was supported by National Institutes of Health (NIH)
Grants GM065546 (to J.J.T.), GM061068 (to D.F.B.), and GM041239
(to C.P.W.). This research is a contribution of the University of
Nebraska Agricultural Research Division, supported in part by funds
provided through the Hatch Act. This publication was also made
possible by NIH Grant P20 RR-017675-02 from the National Center
for Research Resources.
‡Coordinates and structure factors have been deposited in the Protein
Data Bank as entry 3ITG.
*To whom correspondence should be addressed. Telephone: (573)
884-1280. Fax: (573) 882-2754. E-mail: tannerjj@missouri.edu.
1Abbreviations: PutA, proline utilization A; PRODH, proline dehy-
drogenase; P5C, Δ1-pyrroline-5-carboxylate; P5CDH, Δ1-pyrroline-5-
carboxylate dehydrogenase; PPG, N-propargylglycine; TtPRODH,
Thermus thermophilus proline dehydrogenase; PutA86-630, protein
corresponding to residues 86-630 of E. coli PutA; PutA86-669, protein
corresponding to residues 86-669 of E. coli PutA; MES, 2-(N-morpho-
The process by which PutA is transformed from a DNA-
bound repressor to a membrane-associated enzyme, known as
functional switching, involves conformational changes triggered
by proline binding and reduction of the flavin cofactor (1, 7-9).
Recently, conformational changes in the FAD were deduced
from a structure of the dithionite-reduced E. coli PutA PRODH
domain with hyposulfite bound in the proline binding site (7).
lino)ethanesulfonic acid; THFA, L-tetrahydro-2-furoic acid; SDS, so-
dium dodecyl sulfate; HEPES-N, N-(2-hydroxyethyl)piperazine-N0-2-
ethanesulfonic acid; rmsd, root-mean-square deviation; PDB, Protein
Data Bank.
r
pubs.acs.org/Biochemistry
Published on Web 12/08/2009
2009 American Chemical Society

Article
Biochemistry, Vol. 49, No. 3, 2010 561
Scheme 1
Upon reduction, the FAD was observed to adopt a new
conformation characterized by a significant “butterfly” bend
(22ꢀ) of the isoalloxazine ring and rotation of the 20-OH group of
the ribityl chain, resulting in formation of a new hydrogen bond
between the 20-OH group and the FAD N(1) atom (7). The 20-OH
group of the FAD was subsequently demonstrated to act as a
redox-sensitive toggle switch that helps control association of
PutA with the membrane (7).
Scheme 2
A second key interaction involves the FAD N(5) atom and
Arg431. Arg431 is strategically placed within hydrogen bonding
distance of the flavin N(5) atom and is thus potentially able to
sense changes in electron density across the N(1)-N(5) enedi-
amine system attendant to FAD reduction. Indeed, mutation of
Arg431 to Met, or replacement of FAD with 5-deaza-FAD,
abrogated the ability of proline to activate binding of PutA to the
membrane (7). Thus, Arg431 was proposed to have an important
role in transmitting redox signals out of the FAD active site (7),
although the dithionite-reduced enzyme structure did not indi-
cate any movement of this residue.
The conformational changes in the FAD deduced from the
structure of the dithionite-reduced PRODH domain (bending
of the isoalloxazine ring and rotation of the 20-OH group)
presumably represent the initial step in transmission of the flavin
redox status to a remote membrane-binding domain. Signal
transmission likely involves Arg431, but other residues impor-
tant for broadcasting the flavin redox state remain largely
unknown.
EXPERIMENTAL PROCEDURES
Materials. N-Propargylglycine (PPG) was synthesized as
reported previously (10). E. coli polar lipid extracts, phosphati-
dylglycerol, and 1,2-dioleoyl-sn-glycero-3-ethylphosphocholine
were purchased from Avanti Polar Lipids Inc. and used
without purification. Full-length E. coli PutA and an E. coli
PutA PRODH domain construct containing residues 86-630
(PutA86-630) were expressed and purified as previously repor-
ted (8, 11, 12). All chemicals and buffers were purchased from
Fisher Scientific and Sigma-Aldrich, unless stated otherwise. All
experiments used Nanopure water.
Crystallization. Crystallization experiments were performed
with PutA86-630, which was purified by Ni-iminodiacetic acid
chromatography and HiTrapQ anion exchange chromatography
as described previously (12). After ion exchange, the protein was
dialyzed into 50 mM Tris-HCl buffer, 50 mM NaCl, 0.5 mM
EDTA, and 5% glycerol (pH 7.5). The sample was then
concentrated to ∼12 mg/mL as estimated by the bicinchoninic
acid method (Pierce kit). The protein was incubated with PPG in
a ratio of 1 mg of enzyme/mg of PPG (∼560-fold molar excess of
PPG) for at least 30 min at 4 ꢀC. Crystallization experiments were
performed at room temperature using the method of vapor
diffusion. Initial crystallization conditions were identified using
sitting drops and commercially available crystal screen kits.
These experiments revealed promising crystals in solutions con-
taining PEG 5000 monomethyl ether, ammonium sulfate, and
2-(N-morpholino)ethanesulfonic acid (MES) buffer. Following a
few rounds of optimization, further improvement was obtained
using an additive screen approach in which crystal screen
reagents served as additive solutions. The base condition con-
sisted of 26% (w/v) PEG 5000 monomethyl ether, 0.15 M
Here, we have used mechanism-based inactivation of E. coli
PutA by N-propargylglycine (PPG, 1 in Scheme 2) to elicit
conformational changes similar to those involved in functional
switching. This work was motivated by our recent study of the
inactivation of PRODH from Thermus thermophilus (TtPRO-
DH) by PPG (10). TtPRODH is a 307-residue monofunc-
tional PRODH that is 27% identical in sequence to residues
235-569 of the 1320-residue E. coli PutA protein. The crystal
structure of PPG-inactivated TtPRODH revealed a three-
carbon link between the ε-amine of Lys99 and the N(5) atom
of FAD. In one scenario, the mechanism of inactivation is
initiated by the oxidation of PPG to N-propargyliminoglycine
(2 in Scheme 2) and concomitant reduction of the FAD. It is
proposed that attack of 2 by N(5) of the reduced flavin,
rearrangement of a putative allenic species (3) to an imminium
intermediate (4), and nucleophilic attack by an active site
lysine residue (Lys99) on the imminium intermediate lead to
the covalent FAD-Lys99 adduct (5). As a result, the enzyme
is, effectively, locked into the reduced state. Sequence align-
ments show that Lys99 is highly conserved among mono-
functional PRODHs and PutAs, with Lys329 being the
corresponding residue in E. coli PutA. Since PPG locks
TtPRODH in the reduced state, we investigated the possibility
of using PPG to stabilize the reduced conformation of PutA to
facilitate examination of this important conformation. Results
of these studies are described herein.

562 Biochemistry, Vol. 49, No. 3, 2010
Srivastava et al.
ammonium sulfate, and 0.1 M MES buffer (pH 5.8). Reservoir
solutions were created by mixing the base condition and crystal
screen reagent in a ratio of three parts base condition to one part
crystal screen solution. The best results were obtained in hanging
drops using Index reagent 82 as the additive [0.2 M MgCl2, 0.1 M
Bis-Tris buffer, and 25% (w/v) PEG 3350 (pH 5.5)]. In prepara-
tion for low-temperature data collection, the crystals were soaked
in 22% (w/v) PEG 5000 monomethyl ether, 112.5 mM ammo-
nium sulfate, 50 mM MgCl2, 9% (w/v) PEG 3350, 100 mM MES
buffer (pH 5.8), and 20% (v/v) PEG 200, picked up with
Hampton mounting loops, and plunged into liquid nitrogen.
X-ray Diffraction Data Collection, Processing, and Re-
finement. Diffraction data were collected at beamline 4.2.2 of
the Advanced Light Source. The data set used for structure
determination and refinement consisted of 200 frames collected
with an oscillation width of 0.5ꢀ per frame and a detector distance
of 160 mm. The data were integrated with MOSFLM (13)
Table 1: Data Collection and Refinement Statisticsa
space group I212121
˚
unit cell lengths (A)
a = 127.5, b = 133.4, c = 133.6
˚
wavelength (A)
1.0000
˚
diffraction resolution (A)
no. of observations
no. of unique reflections
redundancy
31.5-2.15 (2.25-2.15)
243987
60418
4.0 (3.3)
97.9 (98.9)
0.094 (0.489)
0.053 (0.303)
10.3 (3.3)
26
completeness (%)
Rmerge(I)
R
pim(I)
average I/σ
2
˚
Wilson B-factor (A )
no. of protein chains
no. of atoms
2
6480
no. of protein residues
no. of FAD atoms
no. of water molecules
Rcryst
800
106
241
0.193 (0.250)
0.234 (0.292)
˚
b
through the iMosflm graphical interface and scaled to 2.15 A
resolution with SCALA (14) using the CCP4i interface (15). The
space group is I212121 with the following unit cell dimensions:
Rfree
rmsdc
˚
bond lengths (A)
0.006
0.98
bond angles (deg)
Ramachandran plotd
favored (%)
˚
˚
˚
a = 127.5 A, b = 133.4 A, and c = 133.6 A. There are two
molecules in the asymmetric unit, which implies a solvent content
98.1
1.9
3
˚
of 47% and a Vm of 2.3 A /Da (16). Data processing statistics are
allowed (%)
listed in Table 1. We note that this crystal form differs from the
one used in previous crystallographic studies of the E. coli PutA
PRODH domain, which is in space group I222 with one molecule
in the asymmetric unit and the following unit cell dimensions:
outliers (%)
0.0
2
˚
average B-factor (A )
protein
34
FAD
26
modified Lys329
27
˚
˚
˚
a = 73 A, b = 141 A, and c = 146 A (7, 12, 17-19).
water
31
The structure was determined using molecular replacement as
implemented in MOLREP (20). The search model was derived
from the coordinates of the E. coli PutA PRODH domain
e
˚
coordinate error (A)
0.52
3ITG
PDB entry
aValues for the outer resolution shell of data are given in parentheses.
bA 5% random test set. cCompared to the Engh and Huber para-
meters (30). dThe Ramachandran plot was generated with RAM-
PAGE (31). eMaximum likelihood-based coordinate error reported by
PHENIX.
complexed with the inhibitor L-tetrahydro-2-furoic acid [THFA;
PDB entry 1TIW (19)]. The structure was refined with PHE-
NIX (21). Model building was performed with Coot (22).
The final model includes two protein chains, a covalently
modified FAD cofactor bound to each protein chain, and 241
water molecules. Each protein chain includes residues 87-137,
258-555, and 561-610. As with all other structures of the E. coli
PutA PRODH domain (7, 12, 18, 19), electron density for resi-
dues 138-257 is weak, and thus, these residues were not modeled.
As described in Results, inactivation by PPG induces substantial
disorder in R8 (residues 545-562). Because of the weak electron
density in this region of the structure, residues 549-555 were
modeled as polyalanine and residues 556-660 were not modeled.
The two protein chains have identical conformations within
60-90 min. After PPG treatment, the extent of inactivation
exceeded 95%. PPG-inactivated PutA was denatured via addi-
tion of sodium dodecyl sulfate (SDS) to a final concentration of
0.2% [in 50 mM sodium phosphate buffer and 10% glycerol (pH
7.5)] as described elsewhere (23), and the resulting flavin spec-
trum was recorded. Free FAD was then separated from dena-
tured PutA by centrifugation using a 10 kDa cutoff membrane.
The molar extinction coefficients for the flavin compounds with
maximum absorbance at 378 and 397 nm were estimated using
˚
experimental error (root-mean-square deviation of 0.12 A for
the flavin concentration determined for native PutA (ε451 =
12700 M-1 cm-1) prior to inactivation.
CR atoms). Likewise, the active site structures of the two chains
are identical. Refinement statistics are listed in Table 1.
The kinetic parameters for PRODH activity of full-length
PutA were determined as described previously for TtPRODH
(10). Accordingly, a fresh 50 mM stock solution of PPG dissolved
in 50 mM potassium phosphate buffer (pH 7.5) and 10% glycerol
was prepared. E. coli PutA (5 μM) was incubated at 25 ꢀC with
PPG at concentrations of 0.25, 0.5, 0.75, 1.0, and 2.5 mM.
Aliquots were removed at 15 min intervals over a period of
120 min and assayed at 25 ꢀC using a previously described
PRODH assay based on the reduction of dichlorophenolindo-
Spectroscopic and Kinetic Characterization of the In-
activation of Full-Length E. coli PutA. Spectral changes
associated with inactivation of full-length PutA by PPG were
monitored by a Cary-100 spectrophotometer using the wave-
length range of 300-600 nm at 25 ꢀC. The absorbance spectrum
of oxidized PutA was measured in a quartz cuvette at 10 μM in
50 mM potassium phosphate buffer (pH 7.5) with 10% glycerol.
A quartz cuvette with buffer served as the reference. A fresh
50 mM stock solution of PPG was added to the cuvette at a final
PPG concentration of 500 μM. The solution was then briefly
mixed via cuvette inversion. Spectra were recorded imme-
diately after the cuvette was returned to the cuvette holder
and at subsequent 20 s intervals for 0-3 min, 30 s intervals for
3-10 min, 60 s intervals for 10-60 min, and 120 s intervals for
phenol with L-proline (80 mM) as the substrate. Inactivation
parameters were estimated using the standard method of extract-
ing half-lives from plots of the natural logarithm of activity versus
time and making a Kitz and Wilson replot (24, 25).
Conformational Changes of Full-Length PutA Accompa-
nied by N-Propargylglycine Inactivation. Limited proteolysis

Article
Biochemistry, Vol. 49, No. 3, 2010 563
of full-length PutA by chymotrypsin was performed as previously
described with the following modifications (26). In all digests,
PutA was preincubated with 5 mM proline or 5 mM THFA, for
5 min prior to addition of chymotrypsin except for PPG, which
was incubated with PutA for 30 min. Proline and THFA were
dissolved in 50 mM potassium phosphate buffer (pH 7.5). PutA
(1 mg/mL) was then digested with chymotrypsin (10 μg/mL) for
1 h at 25 ꢀC in 50 mM potassium phosphate buffer (pH 7.5) con-
taining 10% glycerol. The reactions were stopped via addition of
phenylmethanesulfonyl fluoride (final concentration of 2 mM)
and hot SDS-PAGE sample buffer at a 1:2 (v/v) ratio of SDS
buffer to sample. The samples (7.5 μg each) were analyzed by
SDS-PAGE using 8% acrylamide separating gels. As a control,
oxidized PutA (2.5 μg) without digestion was also run on the gel.
Gels were visualized after Coomassie Blue G-250 staining and
imaged using a Bio-Rad Gel Doc 2000 System. The molecular
weight marker was the Precision Dual Color reagent (Bio-Rad).
Assays of Full-Length PutA Binding to E. coli Polar
Lipids. The binding buffer used in these experiments consisted of
10 mM N-(2-hydroxyethyl)piperazine-N0-2-ethanesulfonic acid
containing 150 mM NaCl (pH 7.4) (HEPES-N buffer). E. coli
polar lipid vesicles were prepared as previously described using a
LiposoFast microextruder with a 100 nm polycarbonate filter at a
final concentration of 5 mg/mL (8). After preincubation of full-
length PutA (0.25 mg/mL) with proline (5 mM, 5 min), THFA
(5 mM, 5 min), or PPG (5 mM, 30 min) at 25 ꢀC in 150 mM
HEPES-N buffer, the freshly prepared E. coli polar lipid vesicles
were added to the reaction mixtures to a final concentration of
0.8 mg/mL and incubated at room temperature for 1 h with
slow rotation. The lipid-bound fraction was then separated from
the solution phase using an air-driven ultracentrifuge (Airfuge,
Beckman) at the maximum speed as previously described (27).
The supernatant (soluble phase) was removed and denatured
with SDS sample buffer and labeled as soluble fractions. The lipid
pellet was resuspended in an equal volume of HEPES-N buffer,
denatured with SDS sample buffer, and labeled as the lipid
fraction. Nonspecific sedimentation of the PutA proteins was not
observed in the absence of the lipids. All the soluble and lipid
fractions were analyzed with 10% SDS-polyacrylamide gels,
which were stained with Coomassie Blue G-250. Quantitative
analysis of gels was performed with a Bio-Rad Gel Doc 2000
System.
F
IGURE 1: Spectroscopic analysis of the inactivation of full-length
E. coli PutA by PPG. (A) Flavin spectral changes of E. coli PutA
induced by inactivation by PPG. The spectrum of oxidized PutA is
represented as the thick curve and has maxima (λ) at 451 and 378 nm.
Spectra were acquired immediately after addition of 500 μM PPG
and were continually collected for up to 90 min as described in
Experimental Procedures. For the sake of clarity, only 14 spectra are
shown, with the dashed curve indicating the last spectrum. Note that
the maximum at λ = 451 nm disappears (decreases) and a new maxi-
mum centered at λ = 378 nm appears (increases) as time advances.
The inset shows the absorbance recorded at 451 nm (b) and 378 nm
(O) during the 90 min inactivationprocess. (B) Denaturation of PPG-
inactivated PutA. Spectra are as follows: PutA (7.35 μM) before
inactivation (;), PutA inactivated with 500 μM PPG for 90 min (---),
and the flavin species isolated after treatment of PPG-inactivated
PutA with 0.2% SDS ( ).
3 3 3
RESULTS
Spectral Analysis of the Inactivation of Full-Length
E. coliPutA by PPG. Similar to that observed with TtPRODH,
PPG inactivates full-length E. coli PutA with a concomitant
spectral change (Figure 1A). The absorption spectrum of PutA-
bound FAD decreases at 451 nm after the addition of PPG.
Coinciding with the decrease at 451 nm is a dramatic increase in
absorbance at 378 nm (ε ∼ 20800 M-1 cm-1). The decrease at
451 nm indicates that the FAD is fully reduced, similar to
reduction of the flavin by sodium dithionite or the substrate
proline. The significant increase at 378 nm, which is not observed
with proline or dithionite, suggests a new reduced flavin species is
formed. The new species is most likely the Lys329-FAD cova-
lent adduct shown in Scheme 2 (5). Formation of this adduct was
confirmed by X-ray crystallography (vide infra).
maximum at 397 nm (ε ∼ 22750 M-1 cm-1) [Figure 1B ( )],
3 3 3
which is similar to the spectrum of PPG-inactivated PutA
[Figure 1B (---)] and distinctly different from the spectrum of
oxidized PutA [Figure 1B (;)]. This result suggests that the
isolated flavin is covalently modified, albeit separated from the
enzyme. Presumably, the Schiff base in 5 of Scheme 3 undergoes
hydrolysis to produce 6 of Scheme 3 and the free enzyme as a
result of treatment with SDS buffer.
Kinetic Analysis of the Inactivation of Full-Length E.
coli PutA by PPG. The inactivation of full-length PutA by PPG
was monitored in a time- and concentration-dependent manner,
as shown in Figure 2. Inactivation profiles of PutA with five
different PPG concentrations were used to estimate inactivation
parameters. The standard method of extracting half-lives from
plots of the natural logarithm of activity versus time was used to
construct a Kitz and Wilson replot (Figure 2, inset). Note that the
Kitz and Wilson replot intersects the ordinate above the origin,
which suggests that inactivation proceeds with saturation, i.e.,
The covalently modified flavin could be separated from the
inactivated enzyme by treating PPG-inactivated PutA with a
0.2% SDS solution and isolating the flavin component by mem-
brane filtration. The isolated flavin exhibits a strong absorbance

564 Biochemistry, Vol. 49, No. 3, 2010
Srivastava et al.
F
IGURE 2: Kinetics of PutA inactivation by PPG at 25 ꢀC using a
dichlorophenolindophenol-based activity assay. The percent activity
remaining after incubation with PPG is plotted as a function of time
for four inactivator concentrations (from 0.25 to 2.5 mM). The inset
shows the replot of the half-life of inactivation as a function of the
reciprocal concentration of PPG. The following inactivation para-
meters were obtained from fitting: kinact = 0.13 ( 0.05 min-1, and
KI = 1.5 ( 0.2 mM.
Scheme 3
F
IGURE 3: Two views of PPG-inactivated PutA86-630. Strands and
helices are colored pink and blue, respectively. The dashed curves
denote disordered segments of the polypeptide chain. The covalent
FAD-Lys329 adduct is colored green. Red coloring denotes
regions displaying large conformational differences from other PutA
PRODH domain structures (i.e., R8 and the β1-R1 loop).
that a dissociable complex forms prior to inactivation (25). This
complex of enzyme and PPG is analogous to the binding of
enzyme and substrate in the E S complex. The estimated values
3
of kinact and KI are 0.13 min-1 and 1.5 mM, respectively.
Structural Characterization of the Enzyme-FAD Ad-
duct. The structure of an E. coli PutA PRODH domain construct
(PutA86-630) inactivated by PPG was determined to confirm
formation of a Lys-FAD covalent adduct, i.e., 5 in Scheme 2.
˚
The structure was determined at 2.15 A resolution using mole-
cular replacement (Table 1). As expected, the enzyme displays the
distorted (βR)8 barrel fold characteristic of the PRODH family
(Figure 3). The FAD cofactor is bound at the C-termini of the
strands of the barrel, with the re face of the isoalloxazine packed
against strands 4-6 of the barrel.
F
IGURE 4: Electron density maps showing the (A) overall FAD
conformation and covalent attachment to Lys329 and (B) curvature
of the isoalloxazine ring. The cage represents a simulated annealing
σA-weighted Fo - Fc map contoured at 3σ. Prior to calculation of the
map, the FAD and side chain of Lys329 were omitted, and simulated
annealing refinement was performed with PHENIX.
Electron density maps clearly indicated that the flavin is
covalently modified by PPG (Figure 4). Inactivation with PPG
results in covalent attachment of Lys329 to the flavin N(5) atom
via a three-carbon linkage (Figure 4A). Also, the flavin isoallox-
azine ring is distinctly nonplanar, being bent by ∼35ꢀ around the
N(5)-N(10) axis (Figure 4B). Such deviations from planarity are
termed butterfly conformations and are associated with reduced
states of the flavin. We note that the isoalloxazine ring of the
dithionite-reduced enzyme also exhibits a butterfly confor-
mation (bend angle of 22ꢀ), and the isoalloxazine ring of
the oxidized enzyme is planar. Both the covalent attachment of
an active site lysine residue to the flavin N(5) atom and the
butterfly conformation were observed in PPG-inactivated
TtPRODH (10).
full-length E. coli PutA have unique proteolytic fingerprints (26).
Therefore, the PPG-inactivated enzyme was also subjected to
limited proteolysis using chymotrypsin. As shown in Figure 5,
oxidized PutA displays major proteolysis bands at 135 and 111
kDa, whereas the proline-reduced enzyme displays bands at 119
and 111 kDa. PutA inhibited by the nonreducing proline ana-
logue THFA exhibits a predominant band at ∼90 kDa. These results
are in excellent agreement with previous studies (26). PutA inacti-
vated with PPG displays prominent bands at 119 and 111 kDa, a
pattern similar to that observed with proline-reduced PutA. These
results suggest that the global structure of the PPG-inactivated
Limited Proteolysis Studies of Full-Length E. coli PutA.
As reported previously, oxidized, reduced, and inhibited forms of

Article
Biochemistry, Vol. 49, No. 3, 2010 565
F
IGURE 5: Conformational changes during PutA inactivation by
PPG detected by limited proteolysis. Purified PutA (1 mg/mL) was
incubated for 5 min eachwithproline (5 mM) orTHFA(5 mM) orfor
30 min with PPG (5 mM) in 50 mM potassium phosphate (pH 7.5)
followed by digestion with chymotrypsin (10 μg/mL) for 1 h at 23 ꢀC .
The reactions were quenched with phenylmethanesulfonyl fluoride
and hot SDS sample buffer after 1 h. After complete denaturation in
SDS buffer, 7.5 μg of each digestion was loaded onto an 8%
polyacrylamide denaturing gel for analysis. As a control, undigested
ligand-free PutA was also loaded (2.5 μg). The gel was stained with
Coomassie Blue to visualize the major products. A molecular mass
marker (M) was loaded in the first lane.
F
IGURE 6: Physical binding of PPG-inactivated PutA to E. coli polar
lipids. (A) PutA (0.25 mg/mL) was preincubated with HEPES-N
buffer with 5 mM proline, 5 mM THFA, or 5 mM PPG for 5, 5, or
30 min, respectively, and then incubated with freshly prepared E. coli
polar lipids (0.8 mg/mL) for 1h atroomtemperature. The soluble and
lipid fractions were then separated by Air-fuge ultracentrifugation,
denatured with SDS buffer, and analyzed via SDS-PAGE. As a
control, an equal amount of denatured PutA without treatment was
also loaded. (B) Quantitative analysis of PutA bands for the soluble
(gray bars) and lipid (gray hatched bars) fractions normalized as a
percentage of total PutA using a Bio-Rad 2000 system.
enzyme resembles that of the proline-reduced enzyme, and that
PPG and proline trigger similar conformational changes in PutA.
Lipid Associations of PPG-Inactivated Full-Length
E. coli PutA. Physical lipid binding assays were performed in
which full-length PutA was incubated with E. coli polar lipid
vesicles, and the soluble and lipid-bound protein fractions were
separated by centrifugation. Results of this assay for the oxidized,
proline-reduced, THFA-bound, and PPG-inactivated forms of
PutA are shown in Figure 6. As expected, oxidized PutA is detec-
ted almost entirely in the soluble fraction, indicating that it does
not associate with the lipid vesicles (Figure 6A). The correspond-
ing soluble/lipid partitioning ratio is 98/2 (Figure 6B). In con-
trast, proline-reduced PutA exhibits a soluble/lipid distribution
of 35/65, indicating that reduction of the FAD by the substrate
induces a dramatic increase in the level of lipid binding. THFA
also strengthens PutA-lipid associations relative to the inhibitor-
free, oxidized enzyme, but to a lesser extent than proline. The
soluble/lipid ratio for THFA-bound PutA is 57/43. These results
are consistent with surface plasmon resonance studies (8) show-
ing that oxidized PutA displays no binding to E. coli polar lipid
vesicles, proline induces very tight binding (KD < 0.01 nM), and
THFA stimulates moderately strong binding (KD = 34 nM).
PPG-inactivated PutA displays a soluble/lipid ratio of 34/66,
which is comparable to that of the proline-reduced enzyme
(Figure 6B). These results suggest that PPG causes PutA-lipid
associations that are similar to those induced by proline.
Comparison to Other PRODH Structures. Since PPG
mimics the effects of proline on PutA conformation and PutA
lipid binding, the structure of PPG-inactivated PutA86-630 was
compared to other PutA PRODH structures to gain insight into
conformational changes involved in functional switching. Also,
because this structure is the only one of a PutA PRODH domain
without a ligand occupying the proline binding site, such a
comparison potentially provides clues about conformational
changes associated with substrate binding. The relevant struc-
tures in the PDB are those of oxidized PutA86-669 (residues
86-669 of E. coli PutA) in a complex with the proline analogue
THFA (PDB entry 1TIW) and dithionite-reduced PutA86-669
in a complex with hyposulfite (PDB entry 2FZM).
The structure of PPG-inactivated PutA86-630 reveals a new
conformation for the FAD ribityl chain of PutAs (Figure 7A).
The 20-OH group is tucked below the pyrimidine ring and forms
hydrogen bonds with the FAD N(1) atom and Gly435. The
30-OH group is rotated 66ꢀ from the 20-OH group and forms a
hydrogen bond with the FAD ribose. The 40-OH group is
oriented trans to the 30-OH group and forms a hydrogen bond
with a conserved water molecule buried on the re side of the
FAD. We note that this ribityl conformation is very similar to
that of the FAD in oxidized and PPG-inactivated TtPRODH.
The ribityl hydroxyl groups are arrayed much differently in
other structures of the PutA PRODH domain (Figure 7B,C). In
the THFA complex, the 20-OH and 30-OH groups point into the
proline binding site where they form hydrogen bonds with
Arg556 and Glu559, respectively (Figure 7B). Both residues are
conserved in PutAs and are located on helix R8. As described
below, these critical residues are disordered in the PPG-inacti-
vated enzyme. The 40-OH group of the THFA complex is trans to
the 30-OH group and forms hydrogen bonds with Gly435 and the
FAD ribose. The ribityl chain of the dithionite-reduced enzyme is
identical to that of the THFA complex, except that the 20-OH
group sits below the flavin N(1) atom, as in the PPG-inactivated
enzyme (Figure 7C).
Significant differences are also observed in the protein com-
ponents of the active site. Since the THFA complex and the
dithionite-reduced enzyme structures have identical protein con-
formations, we will restrict the following analysis to a compar-
ison with the THFA complex.
The PPG-inactivated enzyme differs from the other PutA
PRODH structures in the conformations of two residues near
the FAD N(5) atom, Arg431 and Asp370 (Figure 8). In the

566 Biochemistry, Vol. 49, No. 3, 2010
Srivastava et al.
F
IGURE 8: Close-up view of structural differences between PPG-
inactivated PutA86-630 (green) and PutA86-669 in a complex with
THFA (white), highlighting the movements of Arg431 and Asp370.
The occupancy values for the two conformations of Asp370 in the
inactivated enzyme are given in parentheses. Orange dashed lines
represent hydrogen bonds in the PPG-inactivated enzyme. Black
dashed lines represent hydrogen bonds in the THFA complex. The
cage represents a simulated annealing σA-weighted Fo - Fc map
contoured at 3σ. Prior to calculation of the map, the side chains of
Asp370 and Arg431 were omitted, and simulated annealing refine-
ment was performed with PHENIX.
373, and 375). When Asp370 is in conformation A, the ion pair
with Arg431 and hydrogen bond with Asn368 are not formed.
Profound structural differences are also observed in two
regions important for proline binding, R8 (residues 545-562)
and the loop connecting β1 to R1 (residues 287-292). As shown
˚
in Figure 9, R8 is shifted by 3 A away from the isoalloxazine ring
and substantially disordered in the inactivated enzyme. Electron
density in the region expected for R8 was quite weak, and
although the density allowed the modeling of residues 545-555
in an R-helical conformation, residues 549-555 were modeled as
polyalanine due to a complete absence of density for the side
chains. Furthermore, density for residues 556-660 was not
apparent, and these residues were omitted entirely from the final
model (dashed curve in Figure 9). The conformation of the
β1-R1 loop is also quite different in the inactivated enzyme. In
the PPG-inactivated enzyme, this loop is shifted away from the
˚
active site by 4 A, and the side chain of Glu289 is disordered
(Figure 9). As a result, the Glu289-Arg555 ion pair, which is
observed in all other PutA PRODH structures, is broken in the
PPG-inactivated enzyme.
Interestingly, the conformations of R8 and the β1-R1 loop are
reminiscent of those in the structure of oxidized TtPRODH [PDB
entry 2G37 (28)], the only other PRODH structure without a
ligand bound in the proline binding site. A comparison of these
F
IGURE 7: FAD conformations observed in three crystal structures
of the E. coli PutA PRODH domain: (A) PPG-inactivated PutA86-
630, (B) oxidized PutA86-669 in a complex with THFA (cyan),
and (C) dithionite-reduced PutA86-669. Helix R8 is colored red
for the latter two enzymes; this helix is disordered in PPG-inacti-
vated PutA86-630. Note that the reduced flavins are colored green
(A and C), whereas the oxidized flavin is colored yellow (A). The
three flavins are oriented such that the AMP moieties are super-
imposed.
two ligand-free enzymes with the PutA86-669 THFA complex
3
is shown in Figure 10. As in PPG-inactivated PutA86-630, R8 of
TtPRODH is shifted away from the isoalloxazine ring, the
β1-R1 loop is withdrawn from the active site, and the conserved
ion pair (Glu65-Arg288) is broken. Consequently, the active
sites of the ligand-free enzymes are more open than in the
THFA complex, Arg431 donates a hydrogen bond to the FAD
˚
N(5) atom (3.1 A) and forms an ion pair with Asp370 (also 3.1
˚
A). In the PPG-inactivated enzyme, the Arg431-N(5) distance
PutA86-669 THFA complex.
˚
3
has increased to 4.2 A, indicating disruption of this key interac-
tion. Furthermore, electron density maps indicated that Asp370
occupies two conformations in the PPG-inactivated enzyme
(Figure 8). One of these conformations corresponds to the one
observed in the THFA complex (labeled B in Figure 8). In the
other conformation (labeled A in Figure 8), Asp370 is rotated
180ꢀ away from Arg431 and toward a cluster of glutamate
residues located on the surface of the protein (residues 372,
DISCUSSION
Mechanism-Based Inactivation of PRODH. Mechanism-
based inactivation of PRODH was first reported in 1993, but this
area of research is nascent nonetheless. In the first study, Trisch
et al. found that 4-methylene-
activity of rat mitochondrial extracts (29). This area of research
L-proline decreased the PRODH

Article
Biochemistry, Vol. 49, No. 3, 2010 567
F
IGURE 9: Stereographic view of a superposition of the active sites of PPG-inactivated PutA86-630 (green) and PutA86-669 in a complex with
THFA (white). The THFA ligand is colored cyan. The thin dashed lines indicate hydrogen bonds of the THFA complex. The thick dashed curve
denotes the disordered section of R8 of the PPG-inactivated enzyme. For the sake of clarity, conformation B of Asp370 of the inactivated enzyme
has been omitted.
F
IGURE 10: Stereographic view of a comparison of PPG-inactivated PutA86-630 (green), PutA86-669 in a complex with THFA (white), and
TtPRODH (orange). The THFA ligand is colored cyan. The thin dashed lines indicate the Arg555-Glu289 ion pair observed in the THFA
complex. Note that Glu289 is disordered in the PPG-inactivated enzyme (modeled as Ala), and the corresponding residue of TtPRODH (Glu65)
points out of the active site and into the solvent. The thick dashed curve denotes the disordered section of R8 of the PPG-inactivated enzyme.
lay dormant until our report of the inactivation of TtPRODH by
PPG in 2008 (10). In the study presented here, we have further
shown that PPG is also an inactivator of the PRODH domain
within PutA (Figure 2). The kinetic constants for PutA inactiva-
tion (kinact = 0.13 min-1, and KI = 1.5 mM) are similar to those
of TtPRODH (kinact = 0.43 min-1, and KI = 0.8 mM).
Furthermore, as with TtPRODH, a dissociable complex is likely
formed between PutA and PPG before inactivation of the
enzyme. The structures of the PPG-inactivated active sites of
PutA and TtPRODH are also similar. In both cases, the flavin is
highly bent and covalently attached to an active site lysine by a
three-carbon linkage. These similarities suggest that PutA and
TtPRODH share a common mechanism of inactivation by PPG.
The fact that PPG inactivates both monofunctional PRODH and
PutA suggests that it could be a broad-based inactivator of the
PRODH family. A critical test of this idea will be whether PPG
inactivates monofunctional eukaryotic PRODHs.
E. coli PutA exhibits spectral features, proteolysis patterns, and
lipid binding properties similar to those of proline-reduced PutA,
implying that PPG and proline induce analogous conformational
changes in PutA. Therefore, we suggest that the structure of
PPG-inactivated PutA86-630 is relevant for understanding the
conformational changes involved in functional switching of
PutA. In particular, the structure may provide insight into the
changes occurring in the vicinity of the flavin that help to
broadcast the reduced flavin redox state to the distal mem-
brane-binding domain.
The PPG-inactivated enzyme structure may also be relevant
for understanding conformational changes associated with sub-
strate binding or, equivalently, product release, because it is the
first one of a PutA PRODH domain without a ligand in the
proline binding site. All structures of the oxidized PutA PRODH
domain have a competitive inhibitor, such as THFA, L-lactate, or
acetate, bound in the active site. These structures are good models
of the Michaelis complex between the oxidized enzyme and the
substrate, and they show that proline is completely buried in the
Insight into Conformational Changes of PutA Induced
by Substrate Binding and Flavin Reduction. PPG-inactivated

568 Biochemistry, Vol. 49, No. 3, 2010
Srivastava et al.
active site, implying that protein conformational changes must
accompany substrate binding and product release. The structure
of the dithionite-reduced enzyme did not provide insight into
such conformational changes because a hyposulfite ion, an
oxidation product of dithionite, occupies the proline site (7).
Because the structure of the empty active site has proven to be
elusive, the nature of the conformational changes attendant to
substrate binding and product release is currently unknown.
Our analysis of the available structures indicates four regions
of interest with regard to conformational changes associated with
flavin reduction and substrate binding: the FAD cofactor,
residues near the FAD N(5) atom (Arg431 and Asp370), R8,
and the β1-Rl loop.
Involvement of the FAD ribityl chain in functional switching is
consistent with previous biochemical and structural studies.
Membrane binding studies with mutant PutA proteins and PutA
reconstituted with 2-deoxy-FAD showed that the 20-OH group
acts as a redox-sensitive switch that controls membrane associa-
tion (7). The structures of the PPG-inactivated and dithionite-
reduced enzymes indicate that reduction of the flavin induces a
90ꢀ rotation in this group. Additionally, the former structure
reveals a more substantial reorganization of the ribityl chain
involving all three hydroxyl groups and a larger distortion of the
isoalloxazine ring (butterfly bend). Although it seems clear that
the bending of the isoalloxazine ring and rotation of the 20-OH
group are associated with functional switching, whether the
conformational changes of the 30-OH and 40-OH groups are
consequences of the reduced flavin state or empty proline binding
site remains to be determined.
It should be noted that the experimental details of crystal
preparation might explain some of the differences between the
flavin conformations observed in the structures of the PPG-
inactivated and dithionite-reduced enzymes. The crystal used to
determine the former structure was grown from inactivated
enzyme, and thus, conformational changes induced by inactiva-
tion were allowed to proceed in solution. In contrast, the latter
structure was determined from an oxidized crystal that was
soaked in dithionite prior to being freeze-trapped in liquid
nitrogen. In this case, the preformed crystal lattice likely pro-
hibited extensive conformational changes. Furthermore, the use
of a high dithionite concentration resulted in a hyposulfite ion
binding in the active site.
The structure of PPG-inactivated PutA86-630 also implicates
Arg431 in functional switching. In the THFA complex, Arg431 is
poised above the hydride acceptor of the FAD and is thus
positioned to sense the redox state of the flavin. The PPG-
inactivated enzyme structure implies that reduction induces
rupture of the N(5)-Arg431 hydrogen bond. This result is
satisfying because it provides an explanation for previously
reported results showing that replacement of Arg431 with Met
or reconstitution of PutA with 5-deaza-FAD eliminated reduc-
tive activation of PutA-membrane binding (7). It was suggested
at that time that Arg431 plays an important role in transmitting
redox signals out of the PRODH active site that lead to activation
of membrane binding and transcription of the put regulon,
although the structure of the dithionite-reduced enzyme did not
indicate any change in this residue. The structure reported here
confirms the importance of Arg431 in functional switching and
suggests that rupture of the N(5)-Arg431 hydrogen bond might
be among the early events that occur in transmitting the
membrane association signal from the flavin to the membrane-
binding domain.
The PPG-inactivated enzyme structure suggests that Asp370
might also have a role in functional switching. The structure
shows that flavin reduction weakens the interaction between
Asp370 and Arg431, allowing the former residue to sample two
side chain conformational states that are related by a rotation of
180ꢀ around χ1. This result suggests the new hypothesis that
transmission of the functional switching signal out of the active
site involves correlated movement of Asp370 and Arg431.
Finally, the structure of the PPG-inactivated enzyme is unique
among PutA PRODH domain structures in that helix 8 is
substantially unfolded, the conserved Arg555-Glu289 ion pair
is broken, and the β1-Rl loop is withdrawn from the active site.
Consequently, the active site is partially disassembled, and the si
face of the isoalloxazine ring, i.e., the proline binding site, is
exposed to solvent. We suggest that the disassembled active site
reflects the absence of a bound proline analogue, rather than the
reduced state of the flavin, because R8 and the β1-Rl loop are
critical for substrate binding. In particular, Tyr552, Arg555, and
Arg556 of R8 directly contact the substrate, Glu559 of R8
stabilizes Arg556 through ion pairing, and Glu289 of the
β1-Rl loop ion pairs with Arg555 (Figure 9). Furthermore, the
open active site of the PPG-inactivated enzyme bears a striking
resemblance to that of oxidized TtPRODH, the only other
PRODH structure with an empty proline binding site (Figure 10).
We thus suggest that the collective structural information
about the PutA PRODH domain and TtPRODH provides clues
about conformational changes attendant to substrate binding.
The structures imply that the active site might be incompletely
assembled in the absence of the substrate, and the binding of
proline stabilizes residues of R8 and draws the β1-Rl loop into
the active site to connect the loop with R8 via the conserved ion
pair. In this scenario, the ion pair might function as a gate that
closes and opens in response to substrate binding and product
release, respectively. Further experiments are needed to test these
ideas. In this regard, structures of the ligand-free, oxidized PutA
PRODH domain and a monofunctional PRODH in a complex
with THFA would be especially enlightening.
ACKNOWLEDGMENT
We thank Dr. Jay Nix of Advanced Light Source beamline
4.2.2 for help with data collection. The Advanced Light Source is
supported by the Director, Office of Science, Office of Basic
Energy Sciences, of the U.S. Department of Energy under
Contract DE-AC02-05CH11231.
REFERENCES
1. Zhou, Y., Zhu, W., Bellur, P. S., Rewinkel, D., and Becker, D. F.
(2008) Direct linking of metabolism and gene expression in the proline
utilization A protein from Escherichia coli. Amino Acids 35, 711–718.
2. Tanner, J. J. (2008) Structural biology of proline catabolism. Amino
Acids 35, 719–730.
3. Zhou, Y., Larson, J. D., Bottoms, C. A., Arturo, E. C., Henzl, M. T.,
Jenkins, J. L., Nix, J. C., Becker, D. F., and Tanner, J. J. (2008)
Structural basis of the transcriptional regulation of the proline
utilization regulon by multifunctional PutA. J. Mol. Biol. 381,
174–188.
4. Abrahamson, J. L., Baker, L. G., Stephenson, J. T., and Wood, J. M.
(1983) Proline dehydrogenase from Escherichia coli K12. Properties of
the membrane-associated enzyme. Eur. J. Biochem. 134, 77–82.
5. Brown, E. D., and Wood, J. M. (1992) Redesigned purification yields
a fully functional PutA protein dimer from Escherichia coli. J. Biol.
Chem. 267, 13086–13092.
6. Menzel, R., and Roth, J. (1981) Enzymatic properties of the purified
putA protein from Salmonella typhimurium. J. Biol. Chem. 256, 9762–
9766.

Article
Biochemistry, Vol. 49, No. 3, 2010 569
7. Zhang, W., Zhang, M., Zhu, W., Zhou, Y., Wanduragala, S.,
Rewinkel, D., Tanner, J. J., and Becker, D. F. (2007) Redox-induced
changes in flavin structure and roles of flavin N(5) and the ribityl
20-OH group in regulating PutA-membrane binding. Biochemistry 46,
483–491.
proline dehydrogenase domain in complex with competitive inhibi-
tors. Biochemistry 43, 12539–12548.
20. Vagin, A., and Teplyakov, A. (2000) An approach to multi-copy
search in molecular replacement. Acta Crystallogr. D56 (Part 12),
1622–1624.
8. Zhang, W., Zhou, Y., and Becker, D. F. (2004) Regulation of PutA-
membrane associations by flavin adenine dinucleotide reduction.
Biochemistry 43, 13165–13174.
9. Wood, J. M. (1987) Membrane association of proline dehydrogenase
in Escherichia coli is redox dependent. Proc. Natl. Acad. Sci. U.S.A.
84, 373–377.
10. White, T. A., Johnson, W. H., Jr., Whitman, C. P., and Tanner, J. J.
(2008) Structural basis for the inactivation of Thermus thermophilus
proline dehydrogenase by N-propargylglycine. Biochemistry 47,
5573–5580.
21. Zwart, P. H., Afonine, P. V., Grosse-Kunstleve, R. W., Hung, L. W.,
Ioerger, T. R., McCoy, A. J., McKee, E., Moriarty, N. W., Read, R.
J., Sacchettini, J. C., Sauter, N. K., Storoni, L. C., Terwilliger, T. C.,
and Adams, P. D. (2008) Automated structure solution with the
PHENIX suite. Methods Mol. Biol. 426, 419–435.
22. Emsley, P., and Cowtan, K. (2004) Coot: Model-building tools for
molecular graphics. Acta Crystallogr. D60, 2126–2132.
23. Macheroux, P. (1999) UV-Visible Spectroscopy as a Tool to Study
Flavoproteins. In Flavoprotein Protocols (Chapman, S. K., and Reid, G.
A., Eds.) pp 1-7, Humana Press, Totowa, NJ.
11. Zhu, W., and Becker, D. F. (2005) Exploring the proline-dependent
conformational change in the multifunctional PutA flavoprotein by
tryptophan fluorescence spectroscopy. Biochemistry 44, 12297–12306.
12. Ostrander, E. L., Larson, J. D., Schuermann, J. P., and Tanner, J. J.
(2009) A conserved active site tyrosine residue of proline dehydro-
genase helps enforce the preference for proline over hydroxyproline as
the substrate. Biochemistry 48, 951–959.
13. Leslie, A. G. (2006) The integration of macromolecular diffraction
data. Acta Crystallogr. D62, 48–57.
14. Evans, P. (2006) Scaling and assessment of data quality. Acta Crystal-
logr. D62, 72–82.
15. Potterton, E., Briggs, P., Turkenburg, M., and Dodson, E. (2003) A
graphical user interface to the CCP4 program suite. Acta Crystallogr.
D59, 1131–1137.
16. Matthews, B. W. (1968) Solvent content of protein crystals. J. Mol.
Biol. 33, 491–497.
17. Nadaraia, S., Lee, Y. H., Becker, D. F., and Tanner, J. J. (2001)
Crystallization and preliminary crystallographic analysis of the
proline dehydrogenase domain of the multifunctional PutA flavo-
protein from Escherichia coli. Acta Crystallogr. D57, 1925–1927.
18. Lee, Y. H., Nadaraia, S., Gu, D., Becker, D. F., and Tanner, J. J.
(2003) Structure of the proline dehydrogenase domain of the multi-
functional PutA flavoprotein. Nat. Struct. Biol. 10, 109–114.
19. Zhang, M., White, T. A., Schuermann, J. P., Baban, B. A., Becker, D.
F., and Tanner, J. J. (2004) Structures of the Escherichia coli PutA
24. Kitz, R., and Wilson, I. B. (1962) Esters of methanesulfonic acid as
irreversible inhibitors of acetylcholinesterase. J. Biol. Chem. 237,
3245–3249.
25. Silverman, R. B. (1995) Mechanism-based enzyme inactivators.
Methods Enzymol. 249, 240–283.
26. Zhu, W., and Becker, D. F. (2003) Flavin redox state triggers
conformational changes in the PutA protein from Escherichia coli.
Biochemistry 42, 5469–5477.
27. Brown, E. D., and Wood, J. M. (1993) Conformational change and
membrane association of the PutA protein are coincident with
reduction of its FAD cofactor by proline. J. Biol. Chem. 268, 8972–
8979.
28. White, T. A., Krishnan, N., Becker, D. F., and Tanner, J. J. (2007)
Structure and kinetics of monofunctional proline dehydrogenase
from Thermus thermophilus. J. Biol. Chem. 282, 14316–14327.
29. Tritsch, D., Mawlawi, H., and Biellmann, J. F. (1993) Mechanism-
based inhibition of proline dehydrogenase by proline analogues.
Biochim. Biophys. Acta 1202, 77–81.
30. Engh, R. A., and Huber, R. (1991) Accurate bond and angle para-
meters for X-ray protein structure refinement. Acta Crystallogr. A47,
392–400.
31. Lovell, S. C., Davis, I. W., Arendall, W. B., III, de Bakker, P. I., Word,
J. M., Prisant, M. G., Richardson, J. S., and Richardson, D. C. (2003)
Structure validation by CR geometry: φ, ψ and Cβ deviation. Proteins
50, 437–450.