Vrije Universiteit Brussel
The crystal structure of phenylpyruvate decarboxylase from Azospirillum brasilense at 1.5 Åresolution. Implications for its catalytic and regulatory mechanism.Versees, Wim; Spaepen, S; Vanderleyden, J; Steyaert, Jan
Published in:The FEBS Journal
Publication date:2007
Document Version:Final published version
Link to publication
Citation for published version (APA):Versees, W., Spaepen, S., Vanderleyden, J., & Steyaert, J. (2007). The crystal structure of phenylpyruvatedecarboxylase from Azospirillum brasilense at 1.5 Å resolution. Implications for its catalytic and regulatorymechanism. The FEBS Journal, 274, 2363-2375.
General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portalTake down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Download date: 05. Oct. 2021
The crystal structure of phenylpyruvate decarboxylasefrom Azospirillum brasilense at 1.5 A resolution
Implications for its catalytic and regulatory mechanism
Wim Versees1,2, Stijn Spaepen3, Jos Vanderleyden3 and Jan Steyaert1,2
1 Department of Ultrastructure, Vrije Universiteit Brussel, Brussels, Belgium
2 Department of Molecular and Cellular Interactions, VIB, Brussels, Belgium
3 Centre of Microbial and Plant Genetics, INPAC, Katholieke Universiteit Leuven, Heverlee, Belgium
Azospirillum is a genus of plant growth promoting bac-
teria that colonize the rhizosphere of certain tropical
and subtropical crop plants [1,2]. The production of
phytohormones such as the auxin indole-3-acetic acid
(IAA) by the bacterium has been recognized as an
important factor in its plant growth promoting abilities
[3]. The ipdC gene of Azospirillum brasilense was iden-
tified as a key gene in the production of IAA in the
presence of tryptophan [4–6]. The same gene was later
also found to be implicated in the production of phe-
nylacetic acid (PAA), an auxin-like molecule with anti-
microbial activity [7]. In A. brasilense, IAA and PAA
originate primarily from the amino acids l-tryptophan
and l-phenylalanine, respectively, and are synthesized
in a three-step pathway (Fig. 1) consisting of (a)
conversion of l-phenylalanine or l-tryptophan to
Keywords
allosteric enzyme activation; enzyme
mechanism; nonoxidative decarboxylation;
protein crystallography; thiamine
diphosphate
Correspondence
W. Versees, Department of Ultrastructure,
Vrije Universiteit Brussel, Oefenplein,
Gebouw E, Pleinlaan 2, 1050 Brussel,
Belgium
Fax: +32 2629 19 63
Tel: +32 2629 19 49
E-mail: [email protected]
(Received 27 November 2006, revised 22
February 2007, accepted 6 March 2007)
doi:10.1111/j.1742-4658.2007.05771.x
Phenylpyruvate decarboxylase (PPDC) of Azospirillum brasilense, involved
in the biosynthesis of the plant hormone indole-3-acetic acid and the anti-
microbial compound phenylacetic acid, is a thiamine diphosphate-depend-
ent enzyme that catalyses the nonoxidative decarboxylation of indole- and
phenylpyruvate. Analogous to yeast pyruvate decarboxylases, PPDC is sub-
ject to allosteric substrate activation, showing sigmoidal v versus [S] plots.
The present paper reports the crystal structure of this enzyme determined
at 1.5 A resolution. The subunit architecture of PPDC is characteristic for
other members of the pyruvate oxidase family, with each subunit consisting
of three domains with an open a ⁄ b topology. An active site loop, bearing
the catalytic residues His112 and His113, could not be modelled due to
flexibility. The biological tetramer is best described as an asymmetric dimer
of dimers. A cysteine residue that has been suggested as the site for regula-
tory substrate binding in yeast pyruvate decarboxylase is not conserved,
requiring a different mechanism for allosteric substrate activation in PPDC.
Only minor changes occur in the interactions with the cofactors, thiamine
diphosphate and Mg2+, compared to pyruvate decarboxylase. A greater
diversity is observed in the substrate binding pocket accounting for the dif-
ference in substrate specificity. Moreover, a catalytically important glutam-
ate residue conserved in nearly all decarboxylases is replaced by a leucine
in PPDC. The consequences of these differences in terms of the catalytic
and regulatory mechanism of PPDC are discussed.
Abbreviations
AbPPDC, Azospirillum brasilense PPDC; BFD, benzoylformate decarboxylase; EcIPDC, Enterobacter cloacae IPDC; IPDC, indolepyruvate
decarboxylase; KlPDC, Kluyveromyces lactis PDC; PDC, pyruvate decarboxylase; POX, pyruvate oxidase; PP domain, pyrophosphate domain;
PPDC, phenylpyruvate decarboxylase; PYR domain, pyrimidine domain; R domain, regulatory domain; ScPDC, Saccharomyces cerevisiae
PDC; ThDP, thiamine diphosphate; ZmPDC, Zymomonas mobilis PDC.
FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS 2363
phenyl- or indolepyruvate, catalysed by an amino-
transferase, (b) decarboxylation to the corresponding
aldehydes, and (c) oxidation of the aldehyde to phenyl-
acetic acid or indole-3-acetic acid [4,8,9]. The second
step in this pathway is catalysed by the ipdC gene
product, phenylpyruvate decarboxylase (PPDC;
EC 4.1.1.43), a thiamine diphosphate (ThDP)-depend-
ent enzyme that catalyses the nonoxidative decarboxy-
lation of indole- and phenylpyruvate [10,11].
The A. brasilense PPDC (AbPPDC) is a homo-
tetramer of about 58 kDa subunits. Steady-state kinetic
analysis showed that the enzyme converts phenyl-
pyruvate with a 10 times higher specificity than indole-
pyruvate despite a slightly higher affinity for the latter,
and is completely inactive toward benzoylformate
(S. Spaepen, W. Versees, D. Gocke, M. Pohl, J. Steyaert
& J. Vanderleyden, unpublished results). This clearly dis-
tinguishes AbPPDC from indolepyruvate decarboxylase
(IPDC; EC 4.1.1.74) of Enterobacter cloacae (EcIPDC),
which showed a high activity toward indolepyruvate
and benzoylformate, while phenylpyruvate was found to
function as a competitive inhibitor [12,13]. Moreover, in
contrast to the EcIPDC, the PPDC enzyme is subject to
substrate activation, showing sigmoidal v versus [S]
plots as often observed in pyruvate decarboxylases
(PDC, EC 4.1.1.1), with the exception of most bacterial
PDCs (e.g., Zymomonas mobilis PDC) [14–18].
The amino acid sequence of A. brasilense PPDC
shows relatively low identity with the sequence of
other well studied ThDP-dependent decarboxylases
with, for instance, 26% identity with EcIPDC, 23%
and 25% identity with PDCs from Saccharomyces
2-oxoketoglutarate
glutamate
aminotransferase
RCH2
O
O
O-
phenyl/indole- pyruvatedecarboxylase
H+
CO2
RCH2
O
H
RCH2
O
O-
RCH2
NH3+
O
O-
oxidase
O2
H2O
N
H
R1=R1: L-phenylalanine
R2: L-tryptophan
R1: phenylpyruvate
R2: indole-3-pyruvate
R1: phenylacetaldehyde
R2: indole-3-acetaldehyde
R1: phenylacetate
R2: indole-3-acetate
R2=
Fig. 1. Postulated indole-3-acetate (IAA) and
phenylacetate (PAA) biosynthesis route in
A. brasilense, starting from L-tryptophan and
L-phenylalanine, respectively. The involved
enzymes are indicated in blue. This pathway
is proposed on the basis of Koga et al. [70]
and Somers et al. [7].
Structure of Azospirillum brasilense PPDC W. Versees et al.
2364 FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS
cerevisiae (ScPDC) and Zymomonas mobilis (ZmPDC)
and 20% identity with benzoylformate decarboxylase
(BFD, EC 4.1.1.7) from Pseudomonas putida.
This paper reports the first crystal structure of a
PPDC, with the structure of A. brasilense PPDC being
solved at 1.5 A resolution. The general fold of the sub-
units is similar to this of prototypical yeast PDCs
[19,20], while the biological tetramer is present in an
asymmetric conformation. The cofactor-binding site is
fairly conserved compared to IPDC and PDCs. How-
ever, some important differences occur in active site
and proposed regulatory site residues. The potential
consequences of these changes will be discussed in the
light of the currently proposed catalytic and regulatory
mechanisms.
Results and Discussion
Structure determination and subunit description
The Azospirillum brasilense PPDC crystal structure
was solved to 1.5 A resolution using molecular
replacement with the E. cloacae IPDC monomer (PDB
1OVM) as a search model [21]. The asymmetric unit
contains two polypeptide chains, two ThDP molecules,
two Mg2+ ions, two Cl– ions, three glycerol molecules
from the cryo solution, and 1337 water molecules (see
Table 1 for refinement statistics). Clear electron den-
sity is present for the cofactors and for most of the
polypeptide backbone. However, neither the residues
of the N-terminal histidine-tag, nor the seven to eight
C-terminal residues of both subunits were included in
the model due to a lack of electron density. Also a
large peptide region spanning residues 104–119 has
very poorly defined electron density in both subunits.
Hence, this region could only be modelled partially
(residues 106–107 and 112–118 in subunit A, and
105–108 and 111–118 in subunit B were omitted from
the model).
Each subunit consists of three domains with an open
a ⁄ b class topology (Fig. 2). The N-terminal PYR
domain (pyrimidine or a domain, residues 1–160) con-
sists of a central six-stranded parallel b-sheet surroun-ded by six a-helices. The topology of the R domain
(regulatory or b domain, residues 185–330) differs
from the PYR domain and consists of a mixed six-
stranded b-sheet (five parallel and one antiparallel
strand) surrounded by six a-helices, while the C-ter-
minal PP domain (pyrophosphate or c domain, resi-
dues 350–539) is clearly homologous to the PYR
domain and contains a six-stranded parallel b-sheetsurrounded by eight a-helices. Only relatively small
differences are observed between the two subunits in
the asymmetric unit (rmsd of 0.63 A for superposition
of 516 Ca atoms of subunit A on B).
The subunit architecture of PPDC is conserved
among other ThDP-requiring enzymes of the pyruvate
oxidase group (POX group, named after the first estab-
lished structure) [19,22] of which the structure has been
solved, including pyruvate decarboxylase [23,24],
indolepyruvate decarboxylase [21], benzoylformate
decarboxylase [25], oxalyl-CoA decarboxylase [26],
pyruvate oxidase [22] and benzaldehyde lyase [27].
Superposition of the PPDC subunit onto its closest
neighbours yields rmsd values of 2.7 A for 511 Caatoms of ScPDC, 2.8 A for 514 Ca atoms of ZmPDC
and 2.9 A for 501 Ca atoms of EcIPDC.
An asymmetric ‘open’ dimer of dimers
In the asymmetric unit, two subunits of PPDC (desig-
nated A and B) are tightly packed to form a dimer
related by noncrystallographic two-fold symmetry
(Fig. 3A). This dimer interface buries a surface area of
about 2821 A2 and is exclusively formed by interactions
between residues of the PYR and PP domains. The
Table 1. Data collection and refinement statistics. Rsym ¼ S | Ii (hkl)
– ÆIi (hkl)æ | ⁄ S Ii (hkl), where Ii (hkl) are the intensities of multiple
measurements and ÆIi (hkl)æ is the average of the measured intensi-
ties for the ith reflection. Rcryst ¼ 100 · (Sh |Fobs, h – Fcalc, h | ⁄ Sh
Fobs, h), where Fobs and Fcalc are observed and calculated structure
amplitudes, respectively. Rfree ¼ Rcryst calculated for the test set of
reflections not used in refinement. a.u., asymmetric unit.
Diffraction Data
Space group C2221
a (A) 99.4
b (A) 179.0
c (A) 120.9
Resolution range (A) 50.0–1.5 (1.55–1.50)a
Rsym (%) 8.4 (38.2)a
I ⁄ rI 21.4 (3.5)a
Completeness (%) 99.5 (99.1)a
Redundancy 5.6 (4.7)a
Structure Refinement
Resolution range (A) 50.0–1.5
Rcryst (%) 17.5
Rfree (%) 19.6
rmsd for bond lengths (A) 0.0081
rmsd for bond angles (deg) 1.406
Ramachandran plot (% in favoured,
allowed, outlier regions)
98.58, 99.7, 0.3
Number of atoms per a.u. 9305
Average B-factors (A2)
(protein, water, ligands)
19.0, 34.2, 17.6
a Values in parentheses are for the highest resolution shell.
W. Versees et al. Structure of Azospirillum brasilense PPDC
FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS 2365
cofactors ThDP and Mg2+ are also bound at this inter-
face. ThDP interacts with its aminopyrimidine moiety
to the PYR domain of one subunit and with its diphos-
phate moiety to the PP domain of another subunit.
A tetrameric assembly is expected for AbPPDC from
elution profiles on size exclusion chromatography per-
formed at pH 6.5 in the presence of the cofactors
ThDP and Mg2+ (S. Spaepen, W. Versees, D. Gocke,
M. Pohl, J. Steyaert & J. Vanderleyden, unpublished
results). This tetramer can be generated from two di-
mers, related by a two-fold crystallographic axis (with
the two-fold rotation axis transforming subunit A onto
C, and B onto D; see Fig. 3B for subunit nomencla-
ture). However, the tetramer has an architecture far
from 222 symmetry, caused by the nonperpendicular
arrangement between the noncrystallographic axis rela-
ting the two subunits in the dimer and the crystallo-
graphic axis relating the two dimers. As illustrated in
Fig. 3B, the noncrystallographic axis does not intersect
the crystallographic axis but is offset by about 6.5 A.
The noncrystallographic axis is moreover inclined to
the crystallographic two-fold axis by 80.4�, instead of
the 90� expected for 222 symmetry. The asymmetric
character of the PPDC tetramer is illustrated by the
difference in distance between corresponding residues
of the A and C subunits on the one hand and the
Fig. 2. Subunit of the A. brasilense PPDC crystal structure. The
PYR, R and PP domains are coloured blue, green and red, respect-
ively, with intervening loops in yellow. The two cofactors, ThDP
and Mg2+, are shared by two noncrystallographic symmetry-related
subunits and are shown as space-fill models. The positions of the
N- and C-termini, as well as the edges of a flexible active site loop
(residues 105–119) are also indicated.
A
BA C
DB
A C
DB
Fig. 3. Quaternary structure of A. brasilense
PPDC. (A) Stereo picture of a tight dimer of
AbPPDC as found in the asymmetric unit of
the crystal. The subunits are colour coded
red (subunit A) and green (subunit B). The
cofactors are shown as space-fill models,
with the carbon atoms of ThDP coloured
gold, and the Mg2+ ion depicted as a grey
sphere. The position of the flexible loop
(amino acids 105–119) of subunit B and the
C-terminus of subunit A are indicated. (B)
Stereo picture of an asymmetric dimer of
dimers of AbPPDC. The colour codes for
the A and B subunits of the tight dimer are
the same as above. The subunit nomencla-
ture is indicated. The crystallographic two-
fold axis relating the two dimers (A to C, B
to D) is indicated as an orange bar. The two
non-crystallographic symmetry axes relating
the subunits in the asymmetric unit are
indicated as blue and red bars. Note the
nonperpendicular arrangement of crystallo-
graphic and non-crystallographic symmetry
axes resulting in a nonsymmetric arrange-
ment of the tetramer.
Structure of Azospirillum brasilense PPDC W. Versees et al.
2366 FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS
B and D subunits on the other. For example, the dis-
tance between the Mg2+ ions bound to the A and C
subunits is 68 A, while the corresponding distance
between the B and D subunits amounts to 83 A. The
asymmetry moreover imposes very different dimer–
dimer interfaces on both subunits of the dimer in the
asymmetric unit. The dimer–dimer interface involving
the B and D subunits is mainly formed by an exten-
sion of the six-stranded b-sheet across the R domains,
forming a 12-stranded twisted b-sheet. The interface
involving the A and C subunits is different, with resi-
dues from different secondary structure elements
involved. The total accessible surface area buried in
the dimer–dimer interface is only 876 A2, much less
than in the interface within the dimer so that the entire
tetramer assembly is best described as an asymmetric
dimer of dimers. A similar asymmetric tetramer was
also observed in pyruvamide-activated ScPDC (PDB
1QPB) and in Kluyveromyces lactis PDC (KlPDC;
PDB 2G1I) [28,29]. Apart from a relatively small rota-
tion, of about 6�, of one dimer vis-a-vis the other, the
PPDC tetramer has the same overall organization as
these asymmetric PDC tetramers.
Substrate activation and signal transduction in
PPDC: comparison to the existing models
The AbPPDC enzyme displays pronounced substrate
activation with indolepyruvate as a substrate, typified
by sigmoidal plots of the reaction rate versus substrate
concentration (Hill coefficient of 1.85; with phenylpyru-
vate only marginal cooperativity is observed). This type
of regulation of enzyme activity has been observed in
nearly all PDCs characterized to date with the excep-
tion of most bacterial PDCs such as the ZmPDC
[30,31]. In contrast, most other ThDP-dependent de-
carboxylases, such as BFD and IPDC seem to exhibit
normal hyperbolic Michaelis–Menten kinetics without
any indication of allostery [13,32]. Although substrate
activation was first described by Davies in 1967 for
PDC from wheat germs [33], and has since been subject
of detailed kinetic investigations, the exact molecular
mechanism of this phenomenon remains still enigmatic.
Essentially two models have been proposed.
In the first model Baburina et al. proposed Cys221
in yeast PDC as the covalent binding site of the regu-
latory substrate molecule and the starting point of a
signal transduction pathway to the active site [34–37].
In this for the moment most generally accepted activa-
tion mechanism, Cys221 (in the R domain) would
undergo hemiketal formation with the pyruvate sub-
strate. The negative charge of the bound pyruvate is
transferred to His92 (PYR domain) via a salt bridge.
The information is than proposed to be transmitted
via Glu91 onto Trp412 (PP domain) and the loop
comprising residues 410–415. This loop provides two
conserved hydrogen bonds with the aminopyrimidine
ring of ThDP (Gly413 and Ile415) and its rearrange-
ment could induce the observed increase in exchange
rate of the proton on the C2 of ThDP, leading to an
increase in catalytic rate [38,39]. Evidence has been
accumulating in favour of this mechanism as muta-
tions in Cys221 or other residues on the proposed
information transfer route abolishes substrate activa-
tion or decreases the Hill coefficient significantly. The
pivotal Cys221 residue of yeast PDC is, however, not
conserved in AbPPDC and is replaced by a glutamate
residue (Glu212). The histidine (His92), which was
thought to interact with the Cys221-bound pyruvate in
PDC, is not conserved either (Fig. 4) [35,38]. In
AbPPDC this residue structurally aligns with Lys90.
Analogous to the situation in PPDC, an acidic residue
(Asp210) was found on the position corresponding to
Cys221 in the nonallosterically regulated BFD from
Pseudomonas putida [25]. In this enzyme a direct
hydrogen bond occurs between Asp210 and His89
(homologous to His92 in PDC). This direct hydrogen
bond is considered as a mechanism for permanent
activation of BFD, without the need for a bridging
regulatory pyruvate molecule. In AbPPDC a similar
‘permanent’ hydrogen bond occurs between Glu212
and Lys90, without losing the need for a regulatory
substrate molecule. These arguments seem to disfavour
this pocket as the primary site of regulatory substrate
binding in PPDC, although it would not a priori rule
out the use of common residues on the proposed infor-
mation transfer route. An alternative trigger position
in PPDC could as a first site be provided by a cysteine
residue at position 210, only two residues separated
from the position corresponding to Cys221 in PDC.
This residue could in theory fulfil a similar role as
Cys221 in PDC. A closer look, however, shows that
this cysteine is not readily accessible from solvent.
A second model for substrate activation proposed
by Konig and coworkers involves large scale conform-
ational changes and tetramer reassembly [28,40]. This
model was proposed on the basis of the crystal struc-
ture of ScPDC in complex with the substrate analogue
and allosteric activator pyruvamide, showing pyruva-
mide bound in a pocket 10 A away from Cys221. A
transition was observed from a symmetrical tetramer
(obeying 222 symmetry) in the unactivated state (called
form A) to an asymmetrical tetramer (called form B)
in complex with pyruvamide. In going from the nonac-
tivated to the activated form of this enzyme, one dimer
has to rotate by about 30� relative to the other. In
W. Versees et al. Structure of Azospirillum brasilense PPDC
FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS 2367
form B ScPDC this structural asymmetry leads to a
disorder-order transition in two active site loops (resi-
dues 106–115 and 292–303) in half of the active sites,
leading to a half-site closed, half-site open asymmetric
tetramer. Both loops also contribute to the new dimer–
dimer interface created in the asymmetric tetramer,
leading to an increase in the dimer–dimer interface
(from about 900 A2)1550 A2). From this observation
a relationship was inferred between activator binding,
tetramer asymmetry and active site loop reorganization
in half of the sites, leading ultimately to catalysis. A
recent site-directed mutagenesis study involving several
residues of loop 292–303 of ScPDC has shown that
this loop is involved in both catalysis and substrate
activation. However, although the different mutants
lower the Hill coefficient, none of the mutants is able
to completely abolish substrate activation [41]. Also,
the other active site loop of ScPDC (residues 106–115)
harbours two catalytically important residues, His114
and His115. Replacing His114 (but not His115) with a
phenylalanine has been described to reduce the Hill
coefficient, also implying a role for this flexible loop in
the substrate activation mechanism of yeast PDC [42].
As already described in the previous paragraph, the
AbPPDC also adopts a highly asymmetrical tetramer
with relative loose dimer–dimer interactions, which
might seem in agreement with this last model of sub-
strate activation. However, as previously also observed
in the KlPDC [29], the AbPPDC is present as the
asymmetric tetramer without any activator present.
This might indicate an equilibrium in solution between
the asymmetric and symmetric tetramer, shifted toward
the asymmetric architecture in AbPPDC and KlPDC.
Moreover, the active site loop comprising residues
104–119, remains flexible in all subunits, indicating no
strict relationship between tetramer asymmetry and
loop closure in AbPPDC. In AbPPDC a glycerol mole-
cule originating from the cryo solution is bound at the
site corresponding to the observed pyruvamide binding
site in yeast PDC. This glycerol, however, doesn’t seem
to be tightly bound, because apart from an interaction
with Ser156 it only interacts with the protein via brid-
ging water molecules. Whether this or yet another site
corresponds to the regulatory substrate binding site in
AbPPDC and how this information can be transferred
to the active site is currently being investigated
through cocrystallizations with substrate (analogues)
and mutational studies combined with enzyme kinetic
measurements.
Apart from the communication between a regulatory
site and the active site as described above, recent kin-
etic and structural data on certain ThDP-dependent
enzymes also suggested communication between the
active sites of these multimeric enzymes, resulting in a
half-of-the-sites reactivity. In the case of ScPDC and
BFD this type of regulation has mainly been inferred
from complex kinetic behaviour not being consistent
with independently working active sites [43]. For the
E1 component of the pyruvate dehydrogenase complex
(PDH) of both bacterial and human origin, proof for
such a regulatory mode has been accumulating from
kinetics, dissymmetry of active sites in proteolytic
Fig. 4. Putative regulatory binding site of
PDC versus PPDC. The central frame shows
a superposition of the proposed regulatory
substrate binding site of yeast PDC (PDB
1PVD), shown in grey, with the correspond-
ing site in AbPPDC, shown in green (resi-
dues from the R domain) and in blue
(residues from the PYR domain). Cys221
has been proposed as the central residue in
the substrate activation mechanism of
ScPDC [34,35]. In AbPPDC this residue is
replaced by a glutamate (Glu212) forming a
direct hydrogen bond with Lys90 (see text
for further details). In the left corner the
location of this site within the PPDC mono-
mer is indicated by a black square.
Structure of Azospirillum brasilense PPDC W. Versees et al.
2368 FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS
patterns and X-ray structures, and from direct obser-
vation of significantly differing ThDP C2 ionization
rates in different active sites of the multimer [44–46].
On the basis of the X-ray structure of the E1 compo-
nent of the PDH of Bacillus stearothermophilus, Per-
ham and coworkers proposed the existence of a proton
wire mediated by a tunnel of water molecules lined by
acidic residues that connect one cofactor to another.
This proton shuttle serves to reversibly transfer a pro-
ton between the active sites leading to an alternating
site mechanism with both active sites working out of
phase [46]. While waiting for further detailed kinetic
measurements on the AbPPDC, no definite conclusion
about the existence of a half-of-the-sites mechanism in
this enzyme can be drawn for the moment. From a
structural point of view, however, evidence for such a
mechanism in AbPPDC is lacking. No significant
asymmetry between the two active sites in the tight
dimer (AB) is observed. Such a structural asymmetry
would of course not be an absolute requirement for
the ‘half of the sites’ reactivity in the case of a proton
wire mechanism, where the only difference between the
active sites would be their protonation state. More sig-
nificantly no tunnel-like cavity filled with water mole-
cules that links both cofactors seems to be present.
Rather, the region spanning the two cofactors (20 A
between the N1 atoms of both ThDPs) is occupied by
the residues Glu48 (which interacts with N1 of ThDP),
His47, Pro49 and Asn78 of both subunits (not shown).
The distances between these residues and their nonio-
nisable nature (Pro and Asn) preclude this route as a
proton wire. Possibly the active sites of AbPPDC work
independently of each other as also observed for cer-
tain other ThDP-dependent enzymes such as bacterial
PDCs, transketolase and POX [44]. Alternatively, more
complex signal transduction routes would have to
underlie an active centre communication mechanism in
PPDC.
The active site: variations on a common theme
The catalytic cycle of ThDP-dependent decarboxylases
can be subdivided into a number of microscopic steps
[14]. In the first step the thiazolium C2 atom is deprot-
onated, forming a highly nucleophilic ylide. Recent
studies have shown that the N4¢ group of the cofactor,
in its imino tautomeric form, functions as the general
base in this step [22,47,48]. Upon substrate entry the
ylide attacks the substrate carbonyl forming a first
covalent tetrahedral intermediate. With phenylpyruvate
as substrate this intermediate can be called 2-(3-phe-
nyl-lactyl)-ThDP. Decarboxylation of this intermediate
leads to a second carbanion which is resonance-stabil-
ized by its enamine form. Protonation of this interme-
diate gives a second tetrahedral intermediate called
2-(1-hydroxy-2-phenyl-ethyl)-ThDP. Deprotonation of
the a-hydroxyethyl group with concomitant cleavage
of the carbon–carbon bond between C2 of ThDP and
Ca of the intermediate finally releases the product phe-
nylacetaldehyde. Assignment of specific roles to partic-
ular amino acids in this catalytic cycle has proven to
be quite challenging despite some thorough site-direc-
ted mutagenesis studies combined with in-depth kinetic
measurements. The general architecture of the active
site, and hence probably the mechanism, of AbPPDC
generally corresponds to that of PDCs and IPDC,
although some remarkable differences exist.
PPDC binds two molecules of the cofactors ThDP
and Mg2+ per tight dimer (Fig. 3A). The ThDP mole-
cules are bound in the typical V-conformation in clefts
at the interface of the PYR and PP domains. This
V-conformation brings the C2 and N4¢ atoms of the
cofactor in close proximity, at 3.11 A of each other
(Fig. 5). Typically a large hydrophobic residue is
wedged underneath the two rings of the ThDP molecule
hence enforcing the V-conformation. In PPDC this resi-
due is a methionine (Met404) in contrast to an isoleu-
cine normally found in PDCs, but analogous to the
methionine found at this position in POX and benzalde-
hyde lyase (EC 4.1.2.38) [22,27]. The diphosphate moi-
ety of the cofactor is tightly anchored by bidentate
interaction with the octahedral coordinated Mg2+ and
numerous hydrogen bonds with amino acid residues
(Fig. 5). Although some small differences exist between
ScPDC and PPDC in the interaction pattern with the
diphosphate moiety of ThDP, the diphosphate-binding
sequence fingerprint G-D-G-X24-N-N (with Asp429,
Asn455 and Asn456) is conserved [49]. The Mg2+ ion is
coordinated by the side chains of Asp429 and Asn456,
by the main chain carbonyl of Ser458, by the phos-
phates of ThDP and by a water molecule; a pattern
completely conserved among different ThDP-dependent
enzymes [27]. Three conserved and catalytically import-
ant hydrogen bonds are also formed with the amino-
pyrimidine ring of ThDP: one between Glu48 and the
N1¢ atom, one between the main chain carbonyl of
Ala402 and the N4¢, and one looser hydrogen bond
between the main chain amino group of Met404
and the N3¢. Glu48 has been proposed to induce the
1¢4¢-imino tautomer of the aminopyrimidine ring, allow-
ing this N4¢ group to alternate between the amino and
imino form [39,50]. The hydrogen bond between the
main chain carbonyl of Ala402 and the N4¢-imino
group might be used subsequently to position the lat-
ter’s lone pair electrons in suitable orientation to trans-
fer a proton from C2 to the 4¢-imine.
W. Versees et al. Structure of Azospirillum brasilense PPDC
FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS 2369
The current picture of the AbPPDC substrate-bind-
ing pocket is incomplete due to the flexibility of loop
104–119 and the C-terminal residues. In analogy with
other ThDP-dependent decarboxylases (e.g., form B
ScPDC, ZmPDC and EcPPDC) it can be anticipated
that the active site loop will close over the active site
at some point during the catalytic cycle, thus provi-
ding two extra catalytically important histidines
(His112 and His113) [40]. Apart from these two histi-
dine residues, only two other polar residues, Asp25
and Thr71, line the active site pocket (Fig. 6). In
recent years a number of X-ray structures of ThDP-
dependent enzymes in complex with covalent reaction
intermediate analogues have been solved [51–53].
These structures allow modelling of the intermediates
in the active site of AbPPDC to gain further insight
in the possible catalytic and binding interactions,
as shown for the 2-a-lactyl-ThDP intermediate in
Fig. 6A. However, care should be taken in the inter-
pretation because protein conformational changes are
possible (and even likely) along the reaction trajectory.
The modelled structure shows a possible interaction
between the substrate’s carboxylate leaving group and
the main chain amine and side chain carboxyl of
Asp25. Recently Tittmann and coworkers used a com-
bined approach of site-directed mutagenesis and the
direct observation of reaction intermediates by rapid
quench ⁄NMR, to delineate the role of active site resi-
dues in EcIPDC. They found that the residue corres-
ponding to Asp25 is part of a Glu-Asp-His triad
Fig. 5. Cofactor binding site of A. brasilense
PPDC. (A) Schematic diagram of ThDP bind-
ing interactions in AbPPDC. Interactions and
distances (in A) are indicated in magenta.
The short distance between the C2 and N4¢atoms of the cofacter, imposed by the
V-conformation, is indicated in red.
(B) Cofactor binding site. Residues belong-
ing to different subunits are coloured grey
and yellow, respectively, while the carbon
atoms of the ThDP cofactor are shown in
green. The Mg2+ ion and an interacting
water molecule are represented as grey and
red spheres, respectively. Interactions are
indicated by magenta dotted lines. The elec-
tron density, contoured at 5 r, of an fo-fc
simulated annealed omit map calculated
without ThDP is also shown (blue mesh).
Structure of Azospirillum brasilense PPDC W. Versees et al.
2370 FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS
(Glu468-Asp29-His115 in EcIPDC numbering) that
forms a substrate independent carboxylate pocket also
conserved in PDCs [54]. This triad appeared to be
involved on the one hand in binding of substrates and
decarboxylation of the first tetrahedral intermediate,
and on the other hand in the protonation of the carb-
anion ⁄ enamine intermediate. Asp25 together with
His112 of the flexible loop probably fulfils the same
dual role in PPDC. A remarkable difference of PPDC
compared to most other ThDP-dependent decarboxy-
lases is the presence of a leucine at position 462, which
corresponds to the glutamate in the Glu-Asp-His triad
of IPDC and PDC (Glu477 in ScPDC and Glu468 in
EcIPDC, Fig. 6B). This leucine is located directly
underneath the carboxylate group of the lactyl-ThDP
model, on the opposite side of Asp25, and could con-
tribute to catalysis by increasing the hydrophobic
character (low dielectric constant) of the active site.
Indeed, a large fraction of the catalytic power of
ThDP-dependent enzymes is established by providing
an environment of low dielectric constant [55,56]. Such
an environment significantly stabilizes key zwitterionic
intermediates along the reaction trajectory such as the
ylide and the enamine ⁄ carbanion. Hence, Leu462 in
PPDC could compensate for the loss of one partner in
the Glu-Asp-His catalytic triad by an enhanced stabil-
ization of zwitterionic intermediates through solvent
effects. Apart from the interaction with Asp25 (and
the interactions with His112 and His113, which will
likely occur in the closed structure), no specific inter-
actions seem to occur between the enzyme and the
substrate’s carboxyl or C2a-hydroxyl group. This
agrees with the emerging view that the cofactor by
itself carries out the bulk of the catalysis with the
cofactor’s 4¢-amino ⁄ imino group probably being the
central general acid ⁄base in protonation of the sub-
strate’s carbonyl (associated with substrate binding)
and in ionization of the a-hydroxyl group during acet-
aldehyde release [50,54,57].
The aromatic moiety of the substrate can be accom-
modated in the pocket lined by the nonpolar residues
Met380, Phe385, Met461, Phe465 and Phe532 (Fig. 6A).
Although the substrate specificity of PPDC can prob-
ably not be ascribed to a single amino acid substitu-
tion, a key residue in this respect appears to be
Met380. This residue is replaced by a threonine in
ScPDC and ZmPDC (Thr388 in ScPDC) and by a glu-
tamine in EcIPDC (Gln383). A similar role in sub-
strate discrimination was already proposed for the
Gln383 in EcIPDC, because mutation of this residue
to threonine significantly increased the affinity for pyru-
vate compared to indolepyruvate [54]. The role of the
active site residues in catalysis and substrate specificity,
and in particular the implications of the amino acid
Fig. 6. Catalytic substrate binding site of
PPDC versus PDC. (A) Stereo picture of the
active site of AbPPDC with the modelled
reaction intermediate analogue 2-lactyl-
ThDP. The model of 2-lactyl-ThDP was
taken from PDB 2EZ8 [51]. The carbon
atoms of the 2-lactyl-ThDP model are col-
oured yellow, and Mg2+ is depicted as a
grey sphere. Residues from the A and B
subunit are indicated without and with an
asterisk , respectively. The phenyl or indole
ring of the substrate could be accommoda-
ted by the large hydrophobic pocket lined by
Met380, Phe385, Met461, Phe465 and
Phe532. (B) Stereo representation of the
superposition of the active sites of AbPPDC
(residues shown in green) and ScPDC (PDB
1QPB, residues shown in grey). Residues of
the A and B subunits are indicated as
above. Both cofactors are taken from the
AbPPDC structure. His114 and His115 of
ScPDC are provided by an active site loop
that is missing in the AbPPDC model due to
flexibility.
W. Versees et al. Structure of Azospirillum brasilense PPDC
FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS 2371
substitutions compared to PDCs and IPDC, are cur-
rently being further investigated.
Experimental procedures
Protein expression and purification
The ipdC open reading frame from A. brasilense was cloned
into a pET28a vector (Novagen, Darmstadt, Germany) as
will be described elsewhere in detail (S. Spaepen, W. Ver-
sees, D. Gocke, M. Pohl, J. Steyaert & J. Vanderleyden,
unpublished results). The wildtype AbPPDC was expressed
with an N-terminal hexahistidine-tag in Escherichia coli
Bl21-CodonPlus (DE3)-RP cells (Stratagene, La Jolla, CA,
USA). Cells were grown at 37 �C with shaking until the
cells reached an A600 between 0.6 and 0.8. The medium was
then cooled to 30 �C and isopropyl thio-b-d-galactosidewas added to initiate the expression of the recombinant
protein. After lysis of the cells the presence of the his-tag
allowed for a simple two-step purification protocol consist-
ing of a Ni-nitrilotriacetic acid affinity chromatography
step (Qiagen, Hilden, Germany) and gel filtration on a
Superdex-200 column (GE Healthcare, Uppsala, Sweden)
using 10 mm Mes ⁄KOH pH 6.5, 150 mm KCl, 0.1 mm
ThDP, 2.5 mm MgSO4 as buffer in the latter step. The
concentration of pure protein (expressed per monomer)
was determined spectrophotometrically using a e280 of
36840 m)1Æcm)1. SDS-PAGE was used to confirm enzyme
purity. During purification the enzyme activity of the pro-
tein was monitored with phenylpyruvate as substrate using
the established coupled optical test with horse liver alcohol
dehydrogenase and NADH as described previously [13].
Crystallization and data collection
The purified AbPPDC (in crystallization buffer: 10 mm
Mes ⁄KOH pH 6.5, 150 mm KCl, 0.1 mm ThDP, 2.5 mm
MgSO4) was adjusted to a final concentration of 8 mgÆmL)1
using a Vivaspin centrigugal concentrator (Sartorius Viva-
science, Goettingen, Germany, 30 kDa cut-off). PPDC was
crystallised by the hanging drop vapour diffusion method.
Equal volumes of protein solution and precipitant contain-
ing 15% PEG4000 (w ⁄ v), 10% glycerol (v ⁄ v) in 100 mm
Hepes buffer pH 7.5 were mixed and equilibrated at 20 �C.Crystals appeared after one to three weeks and grew to
maximal dimensions of 0.5 · 0.2 · 0.05 mm. The crystals
were transferred to a cryo solution containing 20%
PEG4000 (w ⁄ v), 25% glycerol (v ⁄ v) in 100 mm Hepes
pH 7.5 and transferred immediately to the cryo stream.
X-ray diffraction data were collected at 100 K to a resolu-
tion of 1.5 A on beamline X11 (EMBL, DESY, Hamburg,
Germany) using an X-ray wavelength of 0.8131 A.
The diffraction data were indexed and integrated using
denzo and scaled using scalepack [58]. Intensities were
converted to structure factor amplitudes using truncate
[59]. Table 1 summarizes the data collection and processing
statistics.
Structure determination and refinement
Initial phases were obtained by molecular replacement with
the program phaser [60] using a monomer of the E. cloacae
IPDC (PDB 1OVM) as a search model. A solution was
found, consisting of the expected two subunits forming a
relevant tight dimer in the asymmetric unit. However, the
relatively low sequence identity between AbPPDC and
EcIPDC (26%) did not allow a straightforward refinement
of this solution. Simulated annealing refinement of this
solution in cns (R ¼ 48.5%, Rfree ¼ 54% after refinement)
yielded very poor electron density maps [61,62]. The
molecular replacement solution was subsequently used as a
starting model for 20 rounds of automated model building
(WarpNTrace) in arp ⁄ warp [63]. After 20 rounds the
model had converged at a connectivity index of 0.98 with
1030 out of the 1130 residues docked in the electron den-
sity. The warp model was subjected to the simulated
annealing procedure as implemented in cns and manual
model building, checking and inspection of electron density
was performed in coot [64]. After several cycles of posi-
tional and temperature factor refinement combined with
manual corrections, solvent molecules, cofactors and alter-
native conformations were included in the model. Structure
refinement was considered complete after crystallographic
R-factor and free R-factor had converged, and the differ-
ence density was without interpretable features. The final
model was checked with the Molprobity web server [65].
Refinement statistics are summarized in Table 1.
The structural superpositions were performed using the
DALI server [66] and ⁄ or the program lsqman [67]. Inter-
face accessible surface areas were calculated with the
program provided by the protein–protein interaction
server (http://www.biochem.ucl.ac.uk/bsm/PP/server/). Fig-
ures were prepared with pymol [68] and molscript [69].
The coordinates and structure factors have been depos-
ited in the Protein Data Bank with accession code 2NXW.
Acknowledgements
The authors acknowledge the use of synchrotron beam
time on the EMBL X11 beamline (DESY, Hamburg,
Germany) and on the ID14-2 beamline at the ESRF
(Grenoble, France). This work was supported by a
research grant of the Fund for Scientific Research-
Flanders (FWO-Vlaanderen). Wim Versees is a recipient
of a postdoctoral grant from the FWO-Vlaanderen.
Stijn Spaepen is financed in part by the FWO-Vlaand-
eren (G.0085.03) and in part by the IAP (IUAP P5 ⁄ 03).
Structure of Azospirillum brasilense PPDC W. Versees et al.
2372 FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS
The authors also wish to thank Lieven Buts, Remy
Loris and Klaas Decanniere for helpful discussions.
References
1 Okon Y (1985) Azospirillum as a potential inoculant for
agriculture. Trends Biotechnol 3, 223–228.
2 Steenhoudt O & Vanderleyden J (2000) Azospirillum, a
free-living nitrogen-fixing bacterium closely associated
with grasses: genetic, biochemical and ecological aspects.
FEMS Microbiol Rev 24, 487–506.
3 Dobbelaere S, Croonenborghs A, Thys A, Vande Broek
A & Vanderleyden J (1999) Phytostimulatory effect of
Azospirillum brasilense wild-type and mutant strains
altered in IAA production on wheat. Plant Soil 212,
155–164.
4 Costacurta A, Keijers V & Vanderleyden J (1994) Mole-
cular cloning and sequence analysis of an Azospirillum
brasilense indole-3-pyruvate decarboxylase gene. Mol
Gen Genet 243, 463–472.
5 Prinsen E, Costacurta A, Michiels K, Vanderleyden J &
Vanonckelen H (1993) Azospirillum brasilense indole-
3-acetic acid biosynthesis: evidence for a non-trypto-
phan dependent pathway. Mol Plant Microbe Interact 6,
609–615.
6 Vande Broek A, Gysegom P, Ona O, Hendrickx N,
Prinsen E, Van Impe J & Vanderleyden J (2005) Tran-
scriptional analysis of the Azospirillum brasilense indole-
3-pyruvate decarboxylase gene and identification of a
cis-acting sequence involved in auxin responsive expres-
sion. Mol Plant Microbe Interact 18, 311–323.
7 Somers E, Ptacek D, Gysegom P, Srinivasan M & Van-
derleyden J (2005) Azospirillum brasilense produces the
auxin-like phenylacetic acid by using the key enzyme for
indole-3-acetic acid biosynthesis. Appl Environ Microbiol
71, 1803–1810.
8 Woodward AW & Bartel B (2005) Auxin: regulation,
action, and interaction. Ann Bot 95, 707–735.
9 Soto-Urzua L, Xochinua-Corona YG, Flores-Encarna-
cion M & Baca BE (1996) Purification and properties of
aromatic amino acid aminotransferases from Azospirillum
brasilense UAP 14 strain. Can J Microbiol 42, 294–298.
10 Vuralhan Z, Morais MA, Tai SL, Piper MD & Pronk
JT (2003) Identification and characterization of phenyl-
pyruvate decarboxylase genes in Saccharomyces cerevi-
siae. Appl Environ Microbiol 69, 4534–4541.
11 Guo Z, Goswami A, Nanduri VB & Patel RN (2001)
Asymmetric acyloin condensation catalysed by phenyl-
pyruvate decarboxylase. Part 2: Substrate specificity and
purification of the enzyme. Tetrahedron Asymmetry 12,
571–577.
12 Koga J, Adachi T & Hidaka H (1992) Purification and
characterization of indolepyruvate decarboxylase. A
novel enzyme for indole-3-acetic acid biosynthesis in
Enterobacter cloacae. J Biol Chem 267, 15823–15828.
13 Schutz A, Golbik R, Tittmann K, Svergun DI, Koch
MH, Hubner G & Konig S (2003) Studies on structure-
function relationships of indolepyruvate decarboxylase
from Enterobacter cloacae, a key enzyme of the indole
acetic acid pathway. Eur J Biochem 270, 2322–2332.
14 Schowen RL (1998) Thiamin-dependent enzymes. In
Comprehensive Biological Catalysis. A Mechanistic Ref-
erence. Reactions of Nucleophilic ⁄Carbanionoid Carbon.
Vol. II. (Sinnott M, ed), pp. 217–266. Academic Press,
San Diego, CA.
15 Hubner G, Weidhase R & Schellenberger A (1978) The
mechanism of substrate activation of pyruvate decar-
boxylase: a first approach. Eur J Biochem 92, 175–181.
16 Sergienko EA & Jordan F (2002) New model for activa-
tion of yeast pyruvate decarboxylase by substrate con-
sistent with the alternating sites mechanism:
demonstration of the existence of two active forms of
the enzyme. Biochemistry 41, 3952–3967.
17 Krieger F, Spinka M, Golbik R, Hubner G & Konig S
(2002) Pyruvate decarboxylase from Kluyveromyces lac-
tis. An enzyme with an extraordinary substrate activa-
tion behaviour. Eur J Biochem 269, 3256–3263.
18 Raj KC, Talarico LA, Ingram LO & Maupin-Furlow
JA (2002) Cloning and characterization of the Zymobac-
ter palmae pyruvate decarboxylase gene (PDC) and
comparison to bacterial homologues. Appl Environ
Microbiol 68, 2869–2876.
19 Duggleby RG (2006) Domain relationships in thiamine
diphosphate-dependent enzymes. Acc Chem Res 39,
550–557.
20 Furey W, Arjunan P, Chen L, Sax M, Guo F & Jordan
F (1998) Structure-function relationships and flexible
tetramer assembly in pyruvate decarboxylase revealed
by analysis of crystal structures. Biochim Biophys Acta
1385, 253–270.
21 Schutz A, Sandalova T, Ricagno S, Hubner G, Konig S
& Schneider G (2003) Crystal structure of thiamindi-
phosphate-dependent indolepyruvate decarboxylase
from Enterobacter cloacae, an enzyme involved in the
biosynthesis of the plant hormone indole-3-acetic acid.
Eur J Biochem 270, 2312–2321.
22 Muller YA & Schulz GE (1993) Structure of the thia-
mine- and flavin-dependent enzyme pyruvate oxidase.
Science 259, 965–967.
23 Arjunan P, Umland T, Dyda F, Swaminathan S, Furey
W, Sax M, Farrenkopf B, Gao Y, Zhang D & Jordan F
(1996) Crystal structure of the thiamin diphosphate-
dependent enzyme pyruvate decarboxylase from the
yeast Saccharomyces cerevisiae at 2.3 A resolution.
J Mol Biol 256, 590–600.
24 Dobritzsch D, Konig S, Schneider G & Lu G (1998)
High resolution crystal structure of pyruvate decarboxy-
lase from Zymomonas mobilis. Implications for substrate
activation in pyruvate decarboxylases. J Biol Chem 273,
20196–20204.
W. Versees et al. Structure of Azospirillum brasilense PPDC
FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS 2373
25 Hasson MS, Muscate A, McLeish MJ, Polovnikova LS,
Gerlt JA, Kenyon GL, Petsko GA & Ringe D (1998)
The crystal structure of benzoylformate decarboxylase
at 1.6 A resolution: diversity of catalytic residues in
thiamin diphosphate-dependent enzymes. Biochemistry
37, 9918–9930.
26 Berthold CL, Moussatche P, Richards NG & Lindqvist
Y (2005) Structural basis for activation of the thiamin
diphosphate-dependent enzyme oxalyl-CoA decarboxy-
lase by adenosine diphosphate. J Biol Chem 280,
41645–41654.
27 Mosbacher TG, Mueller M & Schulz GE (2005) Struc-
ture and mechanism of the ThDP-dependent benzalde-
hyde lyase from Pseudomonas fluorescens. FEBS J 272,
6067–6076.
28 Lu G, Dobritzsch D, Konig S & Schneider G (1997)
Novel tetramer assembly of pyruvate decarboxylase
from brewer’s yeast observed in a new crystal form.
FEBS Lett 403, 249–253.
29 Kutter S, Wille G, Relle S, Weiss MS, Hubner G &
Konig S (2006) The crystal structure of pyruvate decar-
boxylase from Kluyveromyces lactis. Implications for the
substrate activation mechanism of this enzyme. FEBS J
273, 4199–4209.
30 Bringer-Meyer S, Schimz KL & Sahm H (1986) Pyruvate
decarboxylase from Zymomonas mobilis. Isolation and
partial characterization. Arch Microbiol 146, 105–110.
31 Candy JM & Duggleby RG (1998) Structure and prop-
erties of pyruvate decarboxylase and site-directed muta-
genesis of the Zymomonas mobilis enzyme. Biochim
Biophys Acta 1385, 323–338.
32 Polovnikova ES, McLeish MJ, Sergienko EA, Burgner
JT, Anderson NL, Bera AK, Jordan F, Kenyon GL &
Hasson MS (2003) Structural and kinetic analysis of
catalysis by a thiamin diphosphate-dependent enzyme,
benzoylformate decarboxylase. Biochemistry 42,
1820–1830.
33 Davies DD (1967) Glyoxylate as a substrate for pyruvic
decarboxylase. Biochem J 104, 50P.
34 Baburina I, Gao Y, Hu Z, Jordan F, Hohmann S &
Furey W (1994) Substrate activation of brewers’ yeast
pyruvate decarboxylase is abolished by mutation of cys-
teine 221 to serine. Biochemistry 33, 5630–5635.
35 Baburina I, Li H, Bennion B, Furey W & Jordan F
(1998) Interdomain information transfer during sub-
strate activation of yeast pyruvate decarboxylase: the
interaction between cysteine 221 and histidine 92.
Biochemistry 37, 1235–1244.
36 Li H, Furey W & Jordan F (1999) Role of glutamate 91
in information transfer during substrate activation of
yeast pyruvate decarboxylase. Biochemistry 38, 9992–
10003.
37 Li H & Jordan F (1999) Effects of substitution of tryp-
tophan 412 in the substrate activation pathway of yeast
pyruvate decarboxylase. Biochemistry 38, 10004–10012.
38 Wang J, Golbik R, Seliger B, Spinka M, Tittmann K,
Hubner G & Jordan F (2001) Consequences of a modi-
fied putative substrate-activation site on catalysis by
yeast pyruvate decarboxylase. Biochemistry 40,
1755–1763.
39 Kern D, Kern G, Neef H, Tittmann K, Killenberg-Jabs
M, Wikner C, Schneider G & Hubner G (1997) How
thiamine diphosphate is activated in enzymes. Science
275, 67–70.
40 Lu G, Dobritzsch D, Baumann S, Schneider G & Konig
S (2000) The structural basis of substrate activation in
yeast pyruvate decarboxylase. A crystallographic and
kinetic study. Eur J Biochem 267, 861–868.
41 Joseph E, Wei W, Tittmann K & Jordan F (2006) Func-
tion of a conserved loop of the beta-domain, not
involved in thiamin diphosphate binding, in catalysis
and substrate activation in yeast pyruvate decarboxy-
lase. Biochemistry 45, 13517–13527.
42 Jordan F, Nemeria N, Guo FS, Baburina I, Gao YH,
Kahyaoglu A, Li HJ, Wang J, Yi JZ, Guest JR & Furey
W (1998) Regulation of thiamin diphosphate-dependent
2-oxo acid decarboxylases by substrate and thiamin
diphosphate. Mg(II) – evidence for tertiary and quater-
nary interactions. Biochim Biophys Acta 1385, 287–306.
43 Jordan F, Nemeria N & Sergienko EA (2005) Multiple
modes of active center communication in thiamin
diphosphate-dependent enzymes. Acc Chem Res 38,
755–763.
44 Seifert F, Golbik R, Brauer J, Lilie H, Schroder-Tittmann
K, Hinze E, Korotchkina LG, Patel MS & Tittmann K
(2006) Direct kinetic evidence for half-of-the-sites reactiv-
ity in the E1 component of the human pyruvate dehydro-
genase multienzyme complex through alternating sites
cofactor activation. Biochemistry 45, 12775–12785.
45 Khailova LS & Korochkina LG (1985) Half-of-The-Site
Reactivity of the Decarboxylating Component of the
Pyruvate-Dehydrogenase Complex from Pigeon Breast
Muscle with Respect to 2-Hydroxyethyl Thiamine Pyro-
phosphate. Biochem Int 11, 509–516.
46 Frank RAW, Titman CM, Pratap JV, Luisi BF &
Perham RN (2004) A molecular switch and proton wire
synchronize the active sites in thiamine enzymes. Science
306, 872–876.
47 Jordan F, Zhang Z & Sergienko EA (2002) Spectro-
scopic evidence for participation of the 1¢,4¢-imino tau-
tomer of thiamin diphosphate in catalysis by yeast
pyruvate decarboxylase. Bioorg Chem 30, 188–198.
48 Jordan F, Nemeria NS, Zhang S, Yan Y, Arjunan P &
Furey W (2003) Dual catalytic apparatus of the thiamin
diphosphate coenzyme: acid-base via the 1¢,4¢-iminopy-
rimidine tautomer along with its electrophilic role. J Am
Chem Soc 125, 12732–12738.
49 Hawkins CF, Borges A & Perham RN (1989) A Com-
mon Structural Motif in Thiamin Pyrophosphate-Bind-
ing Enzymes. FEBS Lett 255, 77–82.
Structure of Azospirillum brasilense PPDC W. Versees et al.
2374 FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS
50 Jordan F (2003) Current mechanistic understanding of
thiamin diphosphate-dependent enzymatic reactions.
Nat Prod Rep 20, 184–201.
51 Wille G, Meyer D, Steinmetz A, Hinze E, Golbik R &
Tittmann K (2006) The catalytic cycle of a thiamin
diphosphate enzyme examined by cryocrystallography.
Nat Chem Biol 2, 324–328.
52 Arjunan P, Sax M, Brunskill A, Chandrasekhar K,
Nemeria N, Zhang S, Jordan F & Furey W (2006) A
thiamin-bound, pre-decarboxylation reaction intermedi-
ate analogue in the pyruvate dehydrogenase E1 subunit
induces large scale disorder-to-order transformations
in the enzyme and reveals novel structural features in
the covalently bound adduct. J Biol Chem 281,
15296–15303.
53 Jordan F & Nemeria N (2005) Experimental observa-
tion of thiamin diphosphate-bound intermediates on
enzymes and mechanistic information derived from
these observations. Bioorg Chem 33, 190–215.
54 Schutz A, Golbik R, Konig S, Hubner G & Tittmann
K (2005) Intermediates and transition states in thiamin
diphosphate-dependent decarboxylases. A kinetic and
NMR study on wild-type indolepyruvate decarboxylase
and variants using indolepyruvate, benzoylformate, and
pyruvate as substrates. Biochemistry 44, 6164–6179.
55 Zhang S, Zhou L, Nemeria N, Yan Y, Zhang Z, Zou Y
& Jordan F (2005) Evidence for dramatic acceleration
of a C-H bond ionization rate in thiamin diphosphate
enzymes by the protein environment. Biochemistry 44,
2237–2243.
56 Jordan F, Li HJ & Brown A (1999) Remarkable stabili-
zation of zwitterionic intermediates may account for a
billion-fold rate acceleration by thiamin diphosphate-
dependent decarboxylases. Biochemistry 38, 6369–6373.
57 Nemeria N, Baykal A, Joseph E, Zhang S, Yan Y,
Furey W & Jordan F (2004) Tetrahedral intermediates
in thiamin diphosphate-dependent decarboxylations
exist as a 1¢,4¢-imino tautomeric form of the coenzyme,
unlike the michaelis complex or the free coenzyme.
Biochemistry 43, 6565–6575.
58 Otwinowski Z & Minor W (1997) Processing of X-Ray
Diffraction Data Collected in Oscillation Mode. In
Methods in Enzymology, Macromolecular Crystallogra-
phy Part A Vol. 276 (Carter CW & Sweet RM, eds),
pp. 307–326. Academic Press, New York, NY.
59 French S & Wilson K (1978) On the treatment of
negative intensity observations. Acta Crystallogr A34,
517–525.
60 McCoy AJ, Grosse-Kunstleve RW, Storoni LC & Read
RJ (2005) Likelihood-enhanced fast translation func-
tions. Acta Crystallogr D61, 458–464.
61 Brunger AT, Krukowski A & Erickson JW (1990)
Slow-cooling protocols for crystallographic refinement
by simulated annealing. Acta Crystallogr A46,
585–593.
62 Brunger AT, Adams PD, Clore GM, DeLano WL, Gros
P, Grose-Kunstleve RW, Jiang J-S, Kuszewski J, Nilges
N, Pannu NS, Read RJ, Rice LM, Simonson T &
Warren GL (1998) Crystallography and NMR system
(CNS): a new software system for macromolecular
structure determination. Acta Crystallogr D54, 905–921.
63 Perrakis A, Morris RM & Lamzin VS (1999) Auto-
mated protein building combined with iterative structure
refinement. Nat Struct Biol 6, 458–463.
64 Emsley P & Cowtan K (2004) Coot: model-building
tools for molecular graphics. Acta Crystallogr D60,
2126–2132.
65 Lovell SC, Davis IW, Adrendall WB, de Bakker PIW,
Word JM, Prisant MG, Richardson JS & Richardson
DC (2003) Structure validation by C alpha geometry:
phi,psi and C beta deviation. Proteins 50, 437–450.
66 Holm L & Sander C (1998) Touring protein fold space
with Dali ⁄FSSP. Nucleic Acids Res 26, 316–319.
67 Collaborative Computational Project Number 4 (1994)
The CCP4 suite: programs for protein crystallography.
Acta Crystallogr D50, 760–763.
68 DeLano WL (2002) The PyMOL Molecular Graphics
System. DeLano Scientific, San Carlos, CA.
69 Kraulis PJ (1991) MOLSCRIPT: a program to produce
both detailed and schematic plots of protein structures.
J Appl Crystallogr 24, 946–950.
70 Koga J, Adachi T & Hidaka H (1991) Molecular Clon-
ing of the Gene for Indolepyruvate Decarboxylase from
Enterobacter cloacae. Mol Gen Genet 226, 10–16.
W. Versees et al. Structure of Azospirillum brasilense PPDC
FEBS Journal 274 (2007) 2363–2375 ª 2007 The Authors Journal compilation ª 2007 FEBS 2375