Investigating terephthalate biodegradation: Structural characterization of a putative decarboxylating cis-dihydrodiol dehydrogenase

Article abstract of DOI:10.1016/j.jmb.2012.07.022

As a highly coveted precursor molecule, terephthalate (Tph) continues to be used extensively for the production of polyethylene Tph bottles, polyester films, and textile fibers worldwide. Based on its detrimental physiological effects, Tph is now recognized as a serious environmental pollutant. While amenable to biodegradation and, in fact, traditionally neutralized by aerobic microbiological processes, our current lack of understanding of the enzymatic degradation of Tph at the molecular level presents a major impediment in the development of robust bioremediation strategies. The biodegradation of Tph proceeds through a single metabolic intermediate (a cis-dihydrodiol), which is subsequently converted to the end product (protocatechuate) by a decarboxylating cis-dihydrodiol dehydrogenase (TphB). Using iodide single-wavelength anomalous dispersion, we report the first structural characterization of TphB to 1.85 A resolution. Contrary to prior speculations, a fluorescent scan unambiguously shows that TphB coordinates Zn^{2 +} and not Fe ^{2 +}. The molecular architecture of TphB provides a rationale to the primary-level divergence observed between TphB and other cis-dihydrodiol dehydrogenases while explaining its intriguingly close evolutionary clustering with non-dihydrodiol dehydrogenases belonging to the isocitrate/isopropylmalate family of enzymes. Sequence and structural analyses reveal a putative substrate-binding pocket proximal to the bound Zn^{2 +}. In silico substrate modeling in this putative binding pocket suggests a mechanistic sequence relying on H291, K295, and Zn^{2 +} as core mediators of catalytic turnover. Overall, this study reveals novel structural and mechanistic insights into a decarboxylating cis-dihydrodiol dehydrogenase that mediates one of the two catalytic steps in the biodegradation of the environmental pollutant Tph.

Full text of DOI:10.1016/j.jmb.2012.07.022

http://dx.doi.org/10.1016/j.jmb.2012.07.022
J. Mol. Biol. (2012) 423, 284–293
Contents lists available at www.sciencedirect.com
Journal of Molecular Biology
journal homepage: http://ees.elsevier.com.jmb
COMMUNICATION
Investigating Terephthalate Biodegradation:
Structural Characterization of a Putative
Decarboxylating cis-Dihydrodiol Dehydrogenase
1
2
1
Jasleen Bains †, Jeremy E. Wulff and Martin J. Boulanger ⁎
1
Department of Biochemistry and Microbiology, University of Victoria, PO Box 3055 STN CSC, Victoria,
BC, Canada V8W 3P6
2
Department of Chemistry, University of Victoria, PO Box 3055 STN CSC, Victoria, BC, Canada V8W 3P6
Received 29 May 2012;
received in revised form
As a highly coveted precursor molecule, terephthalate (Tph) continues to
be used extensively for the production of polyethylene Tph bottles,
polyester films, and textile fibers worldwide. Based on its detrimental
physiological effects, Tph is now recognized as a serious environmental
pollutant. While amenable to biodegradation and, in fact, traditionally
neutralized by aerobic microbiological processes, our current lack of
understanding of the enzymatic degradation of Tph at the molecular level
presents a major impediment in the development of robust bioremediation
strategies. The biodegradation of Tph proceeds through a single metabolic
intermediate (a cis-dihydrodiol), which is subsequently converted to the
end product (protocatechuate) by a decarboxylating cis-dihydrodiol
dehydrogenase (TphB). Using iodide single‐wavelength anomalous
dispersion, we report the first structural characterization of TphB to
1.85Å resolution. Contrary to prior speculations, a fluorescent scan
2
4 July 2012;
accepted 28 July 2012
Available online
11 August 2012
Edited by G. Schulz
Keywords:
Burkholderia xenovorans
LB400;
tph pathway;
cis-dihydrodiol;
TphB;
2
+
2+
unambiguously shows that TphB coordinates Zn
and not Fe . The
zinc
molecular architecture of TphB provides a rationale to the primary‐level
divergence observed between TphB and other cis-dihydrodiol dehydro-
genases while explaining its intriguingly close evolutionary clustering
with non-dihydrodiol dehydrogenases belonging to the isocitrate/iso-
propylmalate family of enzymes. Sequence and structural analyses reveal
2
+
a putative substrate-binding pocket proximal to the bound Zn . In silico
substrate modeling in this putative binding pocket suggests a mechanistic
2
+
sequence relying on H291, K295, and Zn as core mediators of catalytic
turnover. Overall, this study reveals novel structural and mechanistic
insights into a decarboxylating cis-dihydrodiol dehydrogenase that
mediates one of the two catalytic steps in the biodegradation of the
environmental pollutant Tph.
©
2012 Elsevier Ltd. All rights reserved.
*
Corresponding author. E-mail address: mboulang@uvic.ca.
†
Currently a Natural Sciences and Engineering Research Council Postdoctoral Fellow at ABM, Richmond, BC, Canada.
Abbreviations used: LB400, Burkholderia xenovorans LB400; Tph, terephthalate; TphB, terephthalate 1,2-cis-dihydrodiol
dehydrogenase; SEC, size‐exclusion chromatography; EC, Enzyme Commission; IDH, isocitrate dehydrogenase; PDB,
Protein Data Bank; EDTA, ethylenediaminetetraacetic acid.
0022-2836/$ - see front matter © 2012 Elsevier Ltd. All rights reserved.

X-ray Crystallography Studies of TphB
285
Introduction
dehydrogenase (TphB), and a putative reductase
(
TphA3).
Terephthalate (Tph), isophthalate, and phthalate
are nonreactive aromatic compounds (benzene
dicarboxylate isomers) that can enter the environ-
The gene products of the tph operon catalyze the
conversion of Tph into protocatechuate (Fig. 1a) via
the only pathway intermediate cis-dihydrodiol
[variously referred to in the literature as either cis-
4,5-dihydroxycyclohexa-1(6),2-diene-1,4-dicarbox-
ylate or (3R,4S)-3,4-dihydroxycyclohexa-1,5-diene-
1,4-dicarboxylic acid]. It is noteworthy that dihy-
drodiols are key, nonaromatic intermediates
formed during the metabolism of aromatic com-
pounds in both prokaryotes and eukaryotes and
represent an activated form of the parent benzenoid
nucleus. Typically, dihydrodiols with cis
1
,2
ment from industrial or natural sources.
The
production of purified Tph alone ranks among the
top 50 chemicals manufactured in the world, which
translates to high levels of Tph and other organic
contaminants (acetate, benzoate, and p-toluate) as
3
waste products. Based on their wide-ranging
adverse effects on animal cells,
4
–11
as well as
considering the decades of global industrial use as
plasticizers and polymers, these aromatic isomers
are now recognized as serious and ubiquitous
3
8
stereochemistry (as mentioned above with re-
spect to the tph pathway) are formed during
microbial aromatic metabolism, while trans-dihy-
drodiols are often generated during the metabolism
of aromatic compounds by the mammalian
4
,12
pollutants.
With microbial biodegradation as
the primary portal for mediating the bio-friendly
and cost‐effective neutralization of these pollutants,
it is imperative that such bioprocesses are investi-
gated in sufficient detail to enable design and
implementation of improved bioremediation strate-
gies. While some studies have indeed been con-
ducted relevant to these benzene dicarboxylate
3
9,40
liver.
A dedicated set of mostly stereo-selective
dehydrogenases [Enzyme Commission (EC)
1.3.1.X] are employed to catalyze the NAD(P) ‐
+
dependent oxidation of trans-dihydrodiols and, in
select cases, an oxidative decarboxylation of cis-
dihydrodiols. While trans-dihydrodiol dehydroge-
nases have been implicated in the metabolism of
1
,13,14
isomers,
dation processes involved, especially for Tph and
our understanding of the biodegra-
1
5–17
isophthalate, remains limited.
Furthermore, the
4
1
biodegradation of Tph (1,4-dicarboxybenzene) also
holds promise for an alternative bio-friendly appli-
cation that entails use of Tph as a starting compound
for the production of biodegradable 2-pyrone-4,6-
endogenous steroids, prostaglandins,
and
4
2
xenobiotics and suggested as prognostic markers
4
3
for certain cancers, the cis-dihydrodiol dehydro-
genases are fundamental to the microbial biodeg-
radation of both natural and xenobiotic aromatic
compounds.
1
8–20
dicarboxylate polyamides and polyesters.
Several microorganisms have demonstrated the
2
1–24
capability to degrade Tph.
Studies at the
The cis-dihydrodiol dehydrogenase from the tph
pathway (Q13RP4), variously identified as Tph 1,2-
cis-dihydrodiol dehydrogenase, 1,2-dihydroxy-3,5-
cyclohexadiene-1,4-dicarboxylate dehydrogenase,
or DCD dehydrogenase (TphB hereinafter), is one
such enzyme that catalyzes the oxidative decarbox-
ylation of cis-4,5-dihydroxycyclohexa-1(6),2-diene-
1,4-dicarboxylate to protocatechuate (Fig. 1a). Based
on the reaction catalyzed, TphB delineates its own
EC niche (EC 1.3.1.53) within the existing repertoire
of cis-dihydrodiol dehydrogenases, most of which
have been studied with respect to the biodegrada-
tion of benzene, toluene, ethylbenzene, naphthalene,
benzoate, p-chlorotoluene, and 5-amino-4-chloro-2-
genetic and biochemical level encompass members
from both the Gram-positive (Rhodococcus sp.
2
5
26
strain DK17,
Rhodococcus sp. strain RHA1 )
2
,27
29
and Gram-negative (Comamonas testosteroni T-2,
2
8
C. testosteroni YZW-D, Comamonas sp. strain E6,
3
0
and Delftia tsuruhatensis strain T7 ) bacteria. To
probe Tph degradation from a new microbe, and one
that can be effectively and holistically harnessed for
bioremediation applications, owing to an outstanding
repertoire of 11 “central aromatic” and 20 “peripheral
3
1
aromatic” pathways,
attention on Burkholderia xenovorans LB400 (LB400
hereinafter). In addition to a putative tph operon
we choose to focus our
4
4
(
functionally uncharacterized), LB400's uniquely rich
phenyl-3-(2H)-pyridazinone. Interestingly, TphB
represents one of the four cis-dihydrodiol dehydro-
genases that have, in addition, a decarboxylating
component to their reaction chemistry and the only
one that funnels into protocatechuate versus the
relatively more popular central metabolite catechol/
substituted catechol. Currently, representatives
metabolic inventory endows this bacterium with the
capability to degrade various environmental toxins,
3
2
including, but not limited to, nitro-aromatics,
3
3
polycyclic aromatic hydrocarbons, and chlorinat-
3
4–36
ed aromatic compounds.
A closer, compara-
tive analysis of both the organization and the
nucleotide sequences of the protein and enzymatic
components within the putative tph operon con-
firms its annotation in the database. The tph operon
in LB400 encodes a regulatory protein, a transport
protein, the large and small subunit of a putative
dioxygenase (TphA1A2), a putative cis-dihydrodiol
4
5
from only two sub‐subclasses (EC 1.3.1.28 and
4
6
EC 1.3.1.56 ) of cis-dihydrodiol dehydrogenases
stand structurally characterized, which warrants
further investigation of these physiologically impor-
tant group of enzymes. Therefore, with an aim to
investigate the molecular basis of Tph degradation

286
X-ray Crystallography Studies of TphB
Fig. 1. (a) Schematic for the tph pathway that begins with Tph and terminates with the formation of protocatechuate,
which is eventually metabolized into Krebs cycle intermediates. (b) The elution profile of TphB reveals a stable dimeric
conformation with an approximate molecular mass of 73kDa. The protein peaks in the standard curve refer to conalbumin
(
I—75kDa), ovalbumin (II—45kDa), and carbonic anhydrase (III—29kDa). (c) The TphB dimer shown in violet (chain A)
and lime green (chain B). (d) The subdomain division of the TphB monomer. Shown in orange is the N-terminal subdomain
D1 while shown in blue is the C-terminal subdomain D2. (e) Surface topology of the dimeric interface for TphB that is
largely mediated by α-helices. (f) Left: Chemical representation of an octahedral geometry alongside a ball‐and‐stick
2
+
representation of the metal site with Zn shown as a gray sphere, waters as red spheres, a molecule of β-mercaptoethanol,
and the three histidine residues contributed by both the monomers in TphB. Right: 2F −F electron density map contoured
at 1σ for Zn and its coordinating ligands. All the figures have been generated using PyMOL.
o
c
2
+
37
as well as explore cis-dihydrodiol dehydrogenases
in general, we report the 1.85‐Å‐resolution crystal
structure of TphB from LB400 solved using iodide
phasing. Our structure is the first reported for a
TphB and reveals mechanistic insight into this
important group of enzymes.
and size‐exclusion chromatography (SEC). The
elution profile from the SEC column revealed a
stable homodimeric quaternary structure with an
overall molecular weight of approximately 73kDa
(Fig. 1b) similar to that observed for the only other
comprehensively, biochemically characterized TphB
2
7
homolog from C. testosteroni T-2. Crystallization of
TphB revealed a dimer in the asymmetric unit of
C2221 with the native dimeric configuration gener-
ated by a 2-fold noncrystallographic axis of symme-
try (Fig. 1c). A complexation significance score of
Results and Discussion
Homodimerization of TphB creates the requisite
scaffold for metal binding
1
.00 indicates that the buried interface is energeti-
To investigate the structural basis of Tph degra-
dation, we recombinantly produced TphB in Escher-
ichia coli BL21 (DE3) and purified the protein to
homogeneity using nickel affinity chromatography
cally sufficient to promote assembly into a dimeric
quaternary conformation. The buried surface area
2
within the dimeric interface (~1200Å ) accounts for
approximately 10% of the total surface area and is

X-ray Crystallography Studies of TphB
287
4
7
48
comparable to other known dehydrogenases. The
final model was refined to a resolution of 1.85Å,
starts at residue 9 for chain A and at residue 8 for
chain B, and includes an unmodeled loop region
for chain A (residues 38–49) and chain B (residues
are catalytic. Furthermore, treatment with EDTA
did not compromise the structural integrity of TphB
as determined by the elution profile on a size‐
2
+
exclusion column. With the two Zn sites formed
at the dimer interface, it appears that dimerization is
necessary to generate a functional enzyme.
3
8–48) before terminating with residue 347 in both
chains. Data collection and model refinement
statistics are presented in Supplementary Table 1.
Overall, the TphB monomer adopts an alpha–beta
fold that comprises an inner core formed by an
extended, mixed β-sheet comprising 11 β-strands.
The β-sheet is twisted along its longitudinal axis and
sandwiched betweena total of 10α-helices. Despite no
clear demarcation for domain structure at the level of
primary sequence, the structure of the TphB monomer
reveals two subdomains, D1 (residues 1–167; 306–349)
and D2 (residues 168–305) (Fig. 1d). The D1 and D2
subdomains adopt an extended V-shaped structure
that pivots in the middle at the two antiparallel β-
strands, one from each subdomain. The N and C
termini are encompassed within D1 while the dimeric
interface is formed through the interactions between
the D2 subdomains of the TphB monomers.
Substrate‐specific reaction chemistry likely
necessitates the presence of the divergent
D2 domain
All known cis- and trans‐dihydrodiol dehydroge-
nases are categorized as one of 15 EC 1.3.1 sub-
subclasses. Other than the trans-dihydrodiol dehy-
drogenases belonging to EC 1.3.1.20, which have been
4
9
the most extensively structurally characterized,
structures for representative members (2,3-dihydro-
4
5
2,3-dihydroxybenzoate dehydrogenase and cis-2,3-
4
6,50
dihydro-2,3-dihydroxybiphenyl dehydrogenase
)
from only two other categories (EC 1.3.1.28 and EC
1.3.1.56, respectively) have been reported, under-
scoring our limited understanding of this important
group of enzymes. Structural comparison between
TphB and the previously characterized dihydrodiol
dehydrogenases (Fig. 2a–c) reveals a pattern whereby
the homology is confined to the D1 subdomain. While
low sequence identity can, in part, explain this
observation, the presence of a completely divergent
D2 subdomain relative to these dihydrodiol dehydro-
During refinement, 2F −F electron density maps
o
c
of native TphB revealed two symmetry‐related peaks
that were too large for water molecules and
suggestive of transition metal ions. Coincidentally,
a previous characterization of a TphB homolog
revealed that ethylenediaminetetraacetic acid
2
7
51
(
EDTA) inhibited enzyme activity. Based on the
genases is intriguing. However, a DALI structural
5
3
coordination geometry, type of interacting side‐chain
moieties, and associated topological features, zinc
appeared to be a likely candidate. A source of
ambiguity, however, arose from previous work on
homology search confirms our initial BLAST
analysis, indicating that TphB closely resembles
(spanning both the subdomains) dehydrogenases
that act on non-dihydrodiol substrates (Fig. 2d) and
are members of the 3-isopropylmalate and isocitrate
dehydrogenase (IDH) superfamily of enzymes.
Specifically, a 4-hydroxy threonine-4-phosphate
27
the same (aforementioned) TphB homolog where
2
+
Fe were shown to have a stimulatory effect on
enzyme activity alongside a restoration effect follow-
ing treatment with EDTA. To establish the nature of
the bound metal, we performed a K-edge fluores-
47
dehydrogenase from E. coli [Protein Data Bank
(PDB) code 1PS6; IDH member] displays the highest Z
score of 31.6 (DALI score) with an r.m.s.d. of 4.1Å over
2
+
cence scan on the TphB crystal for both Fe and
2
+
α
Zn . While no signal was detected at the K-edge of
328 C atoms with 27% identity. Currently, the PDB
2
+
Fe , a clear fluorescence peak was observed at the
lists five crystal structures for 4-hydroxy threonine-4-
phosphate dehydrogenases (2HI1, putative Salmonella
typhimurium; 1YXO, Pseudomonas aeruginosa; 3LXY,
Yersinia pestis; CO92 and 3TSN, Campylobacter jejuni;
and 1PS6/1PS7/1PTM, E. coli) of which only 1PS6/
2
+
2+
Zn K‐edge consistent with Zn bound to TphB.
2
+
The two Zn are positioned at the surface and bulk
solvent‐exposed facet of the dimer interface and
ε2
coordinated by the N imidazole nitrogens of H191
47
and H287 from one monomer and H235 from the
second monomer (Fig. 1f; note that residue number-
ing is offset by +3 with respect to native sequence of
TphB due to the presence of three additional residues
remaining from thrombin cleavage). Two water
molecules and a molecule of β-mercaptoethanol
complete the octahedral coordination sphere of each
1PS7/1PTM appear in a published journal. In order
to assess the phylogenetic categorization of TphB
(Q13RP4) with regard to both the dihydrodiol and
non-dihydrodiol dehydrogenases, a phylogenetic tree
was constructed (Fig. 2e). Based on the availability of
annotated amino acid sequences, our analysis includes
members from 8 of the 15 categories for the
dihydrodiol dehydrogenases ranging from 248 to
400 amino acids in length as well as one representative
member from 4-hydroxythreonine-4-phosphate dehy-
drogenase, 3-isopropylmalate dehydrogenase, and
isocitrate dehydrogenase. Among the representative
members displayed in Fig. 2e, the non-dihydrodiol
2
+
Zn for the native data set solved in C222 (Fig. 1f).
1
2
+
The distances between the Zn and the coordinating
nitrogen, oxygen, and sulfur range from 2.1 to 2.8Å. It
is noteworthy that pure structural metal sites
comprise four protein ligands with no bound water
2
+
molecule(s), suggesting that the Zn sites in TphB

288
X-ray Crystallography Studies of TphB
Fig. 2. Structural overlay of TphB monomer (violet) with (a) cis-2,3-dihydrobiphenyl-2,3-diol dehydrogenase
EC1.3.1.56), PDB code 1BDB; (b) trans-1,2-dihydrobenzene-1,2-diol dehydrogenase (EC1.3.1.20), PDB code 2O48; (c)
,3-dihydro-2,3-dihydroxybenzoate dehydrogenase (EC 1.3.1.28), PDB code 2FWM; and (d) 4-hydroxy threonine-4-
(
2
5
1
phosphate dehydrogenase (EC 1.1.1.262), PDB code 1PS6 on the DALI server. The statistics for these overlays are as
α
α
follows: Z score: 3.5, aligned C atoms: 126, r.m.s.d.: 3.8Å with 13% identity for 1BDB. Z score: 3.2, aligned C atoms: 99,
α
r.m.s.d.: 3.8Å with 14% identity for 2O48. Z score: 3.7, aligned C atoms: 116, r.m.s.d.: 3.7Å with 9% identity for 2FWM. Z
score: 31.6, aligned C atoms: 328, r.m.s.d.: 4.1Å with 27% identity for 1PS6. (e) A phylogenetic tree constructed using the
α
5
2
program MultAlin. Protein accession codes are listed.
and dihydrodiol dehydrogenases are well segregated
with the exception of TphB that clusters with the non-
dihydrodiol dehydrogenases.
and 1YXO, respectively) alongside the relatively
more divergent IDHs that are known to coordinate
2
+
2+
either a Mg or Mn as the metal cofactor. It is
unknown if the other three mechanistic homologs of
TphB (among the dihydrodiol dehydrogenases)
require a metal ion, although based on the length of
their primary sequence and the associated distant
phylogeny (Q1LDM9) with TphB, it is unlikely that
they possess a similar D2 subdomain. Thus, it
appears that the D2 subdomain in TphB is designed
to support a reaction mechanism with a requirement
for metal-based catalytic chemistry.
It is noteworthy that only four types of dihydrodiol
dehydrogenases (EC 1.3.1.25, 1.3.1.53, 1.3.1.67, and
1
.3.1.68) catalyze reactions that involve both a
dehydrogenation and a decarboxylation event of
which TphB (1.3.1.53) reported in this article is the
first to be structurally characterized. Like TphB, all
structurally related non-dihydrodiol dehydroge-
nases have an obligate requirement for a metal ion.
This could explain the evolutionary requirement for a
D2 subdomain‐like substructure, which in this case
provides the appropriate scaffold to accommodate
the metal ion. The latter explanation is supported by
the observation that the 4-hydroxythreonine-4-phos-
phate dehydrogenase, which shares the highest
A conserved, metal proximal site exhibits favorable
chemical topology to bind the native substrate
Structural comparisons with TphB homologs
2
+
similarity, has a nearly identical Zn
site with
combined with the identification of the catalytic
2
+
TphB. Furthermore, the three other 4-hydroxythreo-
nine-4-phosphate dehydrogenases (unpublished
Zn site enabled the localization and characteriza-
tion of the active‐site structure in TphB, which is
niched within the cradle‐like surface topology
2
+
2+
structures) coordinate a Ni or Mg (3TSN, 3LXY,

X-ray Crystallography Studies of TphB
289
formed at the subdomain interface (Fig. 3a; black
square boxes) and extending to the metal ion
coordinated at the dimeric interface. Sequence
alignment between TphB from LB400 and TphB
from other prokaryotes supports an important role
for this region with the clustering of highly
conserved residues (Fig. 3b and c). To further aid
in the demarcation of the active site, we generated
an electrostatic surface of TphB (Fig. 3d), which
revealed a large basic patch extending from the
involved in hydrogen‐bonding interactions with
δ1
δ1
the O of E238 and O of D288, respectively
(Fig. 4a). These interactions likely serve to ensure
correct orientation for these histidine residues as
also suggested for the analogous histidine–aspar-
tate interaction observed in 4-hydroxythreonine-4-
4
7
phosphate dehydrogenase. It is noteworthy that
equivalents of all these residues are, to the best of
our knowledge, absolutely conserved in Tph 1,2-
cis-dihydrodiol dehydrogenases (sequence align-
ment; Supplemental Data). Considering the stereo-
chemistry and redox potential of the native
substrate of TphB and the need for correct spatial
juxtaposition to enable metal-mediated reaction
chemistry, we predict that TphB will be highly
specific for its native substrate.
2
+
bound Zn capable of complementing the negative
charge of the TphB substrate.
Overall, the active‐site pocket is substantially
solvent exposed (Fig. 4a), which is not surprising
considering the highly polar nature of the native
substrate (Fig. 1a). However, three leucine resi-
dues (L176, L178, and L190; Fig. 4b) define one
face of this putative binding pocket forming a
hydrophobic region capable of accommodating the
nonpolar carbon backbone of the dihydrodiol
Proposed reaction mechanism
The current theory for the mechanism of TphB‐
mediated conversion of the cis-dihydrodiol pathway
intermediate into protocatechuate includes an initial
+
substrate and the lipophilic part of the NAD(P)
cofactor. The proximal GHGSG sequence in TphB
Fig. 4a and b), which closely resembles the
2,28
(
dehydrogenation event followed by decarboxylation.
phosphate binding consensus sequence GXGXXG,
likely participates in ensuring proper juxtaposition
Due to the lack of commercial availability of the
native substrate to support co-crystallization, we
constructed an in silico model of the enzyme–
substrate complex, taking into account all of the
residues of the putative active site (Fig. 4c). To assist
in the design of the model, we first examined several
small-molecule structures in the Cambridge Struc-
+
of the NAD(P) cofactor for hydride transfer. A
lysine (K295) that sits atop this shallow pocket is
the likely candidate for coordinating the carboxyl-
ate that remains unchanged between the substrate
δ1
and product (Fig. 4a and b). The N atoms of
2
+
2+
two out of the three Zn coordinating histidines,
H235 and H287 (one from each monomer), are
tural Database where Zn was bound to α-hydroxy
acid or α,β-dihydroxy acid functions. This search
Fig. 3. (a) An overview of the active‐site localization in TphB positioned at the inter-subdomain interface, which can be
distinguished by a groove‐like surface topology (indicated by black square boxes). (b and c) An illustration of sequence
conservation (yellow) mapped on to TphB structure, where (b) denotes the highly conserved (represented as yellow) dimer
face while (c) depicts the more divergent surface opposing the dimer interface. Note that the color scheme employed for
depicting the sequence conservation is purely arbitrary. (d) Electrostatic representation of the dimer interface.

290
X-ray Crystallography Studies of TphB
Fig. 4. (a) An overview of the location of the proposed active site highlighting the modeled native substrate that is in
close proximity to the zinc site. A close-up view of the active site showcasing key residues that form this active site as well
as those likely involved in catalysis; the black dashes indicate the GHGSG sequence tag in TphB, which bears semblance to
the consensus sequence for binding the phosphate of the coenzyme. (b) A depiction of a rotated view of (b) to highlight the
interaction between the modeled native substrate (with hydrogens shown in this case) and both H291 and K295. (c) A
2
+
mechanistic schematic highlighting the roles of Zn , H291, and K295 in catalysis.
revealed a marked preference for α-hydroxy acids to
bind in a bidentate fashion, using one oxygen of the
carboxylate together with the alcoholic oxygen. α,β-
Dihydroxy acids bound with a qualitatively similar
geometry; other possible binding modes (e.g.,
bidentate coordination through both oxygen atoms
of the carboxylate group or tridentate coordination
through the carboxylate and both alcohols) were not
represented in the database.
must be expelled as part of the dehydrogenation
event. In addition, the hydridic hydrogen points into
the open cavity of the putative active site, where it
would be well placed for delivery to bound
+
NAD(P) . Based on these calculations, we hypoth-
esize that the mechanism follows the process
outlined in Fig. 4c where dehydrogenation promot-
ed by H291 results in the transient formation of a β-
keto acid, which would rapidly lose CO2 to generate
protocatechuate and promote de-coordination of the
We therefore constructed our model by beginning
with a bidentate coordination of the α-hydroxy acid
function in the cis-dihydrodiol intermediate to the
tris-histidine–zinc complex (Fig. 4b and c). Geome-
try optimization by both molecular mechanics
2
+
active-site Zn , thereby allowing the enzyme to
return to its catalyst-ready state. Aside from orient-
ing the substrate, which seems to be the primary role
2
+
of Zn , the modeling data suggest that having one
2
+
(
MMFF) and semiempirical (PM3) methods placed
hydroxyl group bound to Zn
(Fig. 4b and c)
the vinylic carboxylate function in close proximity to
K295, as predicted from our preliminary examina-
tion of the active‐site geometry. It is also possible
that R300 plays a role in binding the distal
carboxylate, but this would require a substantial
rearrangement of the arginine side chain.
greatly improves hydrogen bonding to the neigh-
boring alcohol in the substrate. This effect is due to
optimized geometry (binding of the OH to the zinc
requires that the proton point back toward the
unbound alcohol) and electrochemistry (binding of
2
+
the OH to the Zn should also lower the pK of the
a
The geometry-optimized model placed the non-
coordinated alcohol in close proximity to H291.
Intriguingly, when this histidine was modeled in its
deprotonated state, geometry optimization (Fig. 4c)
bound functional group). The resulting increase in
intramolecular hydrogen bonding should enhance
the acidity of the unbound alcohol, which in turn
will facilitate the reaction.
resulted in a very close contact (1.8Å) between the
ε2
N
imidazole nitrogen of H291 and the proton of
Conclusions
the β-hydroxyl group. Moreover, this “protic”
hydrogen is in an almost perfectly antiperiplanar
orientation relative to the “hydridic” hydrogen that
Here, we report the three‐dimensional structure of
TphB, the first of a decarboxylating dehydrogenase

X-ray Crystallography Studies of TphB
291
to act on a cis-dihydrodiol substrate structural
characterization of the dehydrogenase that mediates
the second and final step in the biodegradation
pathway of the major environmental pollutant Tph.
A structural comparison of TphB with other cis-
dihydrodiol and non-dihydrodiol dehydrogenases
reveals a trend that is closely matched at the
phylogenetic level, whereby, despite catalyzing the
turnover of a cis-dihydrodiol substrate, TphB shares
closer homology with dehydrogenases that are
active on non-dihydrodiol substrates. Based on our
structural data that pertain to the associated metal
cofactor, we propose that this trend stems from an
obligate requirement for a metal-based catalytic
chemistry in case of TphB. In silico substrate
modeling corroborates our structural finding(s)
regarding the putative active site, while providing
a compelling mechanistic scheme that involves the
selective constraints (refer to the Supplemental Data for
detailed methodology regarding substrate modeling).
PDB accession number
Coordinates and structure factors have been deposited
in the PDB with accession numbers 4aty and r4atysf,
respectively.
Acknowledgements
This work was supported by research grants from
the Natural Sciences and Engineering Research
Council to M.J.B. and J.E.W. J.B. was supported by
a Natural Sciences and Engineering Research
Council CGSD fellowship. M.J.B. and J.E.W. are
Canada Research Chairs and Michael Smith Foun-
dation for Health Research Scholars. The authors
gratefully acknowledge the staff of beamline 9-2 at
the Stanford Synchrotron Radiation Lightsource.
2
+
Zn site, H291, and K295 as key components to
perform the required chemistry for catalysis.
Materials and Methods
The chromosomally encoded tph gene from B. xenovorans
LB400 was cloned into pET-28a(+) in frame with an N-
terminal hexahistidine tag and expressed in E. coli BL21
Supplementary Data
Supplementary data to this article can be found
online at http://dx.doi.org/10.1016/j.jmb.2012.07.022
(DE3) cells. Following purification using nickel affinity
chromatography and SEC, crystals of TphB were obtained
using the vapor-diffusion method for crystallization (for
details, refer to the Supplemental Data). Diffraction data on
TphB crystal(s) were collected on a Rigaku R-axis IV++
area detector (home source) while the fluorescent scan for
the K‐edge for iron and zinc was performed at Stanford
Synchrotron Radiation Lightsource, California. Diffraction
References
1. Keyser, P., Pujar, B. G., Eaton, R. W. & Ribbons, D. W.
(1976). Biodegradation of the phthalates and their
esters by bacteria. Environ. Health Perspect. 18, 159–166.
2. Schlafli, H. R., Weiss, M. A., Leisinger, T. & Cook,
A. M. (1994). Terephthalate 1,2-dioxygenase system
from Comamonas testosteroni T-2: purification and
some properties of the oxygenase component.
J. Bacteriol. 176, 6644–6652.
3. Macarie, H., Noyola, A. & Guyot, J. P. (1992).
Anaerobic treatment of a petrochemical wastewater
from a terephthalic acid plant. Water Sci. Technol. 25,
223–235.
4. Colborn, T., vom Saal, F. S. & Soto, A. M. (1993).
Developmental effects of endocrine-disrupting chemi-
cals in wildlife and humans. Environ. Health Perspect.
101, 378–384.
5. David, R. M., Moore, M. R., Cifone, M. A., Finney, D. C.
& Guest, D. (1999). Chronic peroxisome proliferation
and hepatomegaly associated with the hepatocellular
tumorigenesis of di(2-ethylhexyl)phthalate and the
effects of recovery. Toxicol. Sci. 50, 195–205.
5
4
data were processed using the CCP4 suite of programs.
A
total of 10 iodide sites were identified and refined using
autoSHARP, which resulted in high‐quality phases that
enabled building and registering of approximately 70% of
5
5
the backbone using ARP/wARP. The remaining struc-
ture was built manually and solvent atoms were selected
56
57
using Coot and refined with REFMAC to an Rcryst of
1.3% and an Rfree of 26.5%. In total, 187,478 reflections
2
were used in refinement with a sigma cutoff of 2.0. The
Ramachandran plot of the refined TphB structure showed
good stereochemistry with 94.5% of the residues in the
most favored conformations and no residues modeled in
disallowed orientations. Overall, 5% of the reflections were
set aside for calculation of Rfree. Data collection statistics
are presented in Supplementary Table 1. For in silico
substrate modeling, residues identified in the crystal
structure to project into the putative active site were
selected in PyMOL and imported into Spartan ‘06. Briefly,
the amide backbone was maintained for residues where
the backbone carbonyl group could conceivably interact
with a substrate; remaining residues were trimmed back to
the α-carbon for computational efficiency. All non-hydro-
gen atoms were frozen and the structure was subjected to a
semiempirical geometry optimization (PM3) to ensure the
correct placement of the hydrogen atoms. An α-hydroxy
acid function was built off of the zinc atom and the
structure was optimized for the final output after placing
6. Ganning, A. E., Brunk, U. & Dallner, G. (1984). Phthalate
esters and their effect on the liver. Hepatology, 4, 541–547.
7. Harris, C. A., Henttu, P., Parker, M. G. & Sumpter, J. P.
(1997). The estrogenic activity of phthalate esters in
vitro. Environ. Health Perspect. 105, 802–811.
8. Huff, J. E. & Kluwe, W. M. (1984). Phthalate esters
carcinogenicity in F344/N rats and B6C3F1 mice. Prog.
Clin. Biol. Res. 141, 137–154.

292
X-ray Crystallography Studies of TphB
9
. Jobling, S., Reynolds, T., White, R., Parker, M. G. &
degradation by Rhodococcus sp. strain DK17. FEMS
Microbiol. Lett. 252, 207–213.
26. Hara, H., Eltis, L. D., Davies, J. E. & Mohn, W. W.
(2007). Transcriptomic analysis reveals a bifurcated
terephthalate degradation pathway in Rhodococcus sp.
strain RHA1. J. Bacteriol. 189, 1641–1647.
Sumpter, J. P. (1995). A variety of environmentally
persistent chemicals, including some phthalate plasti-
cizers, are weakly estrogenic. Environ. Health Perspect.
103, 582–587.
1
1
0. Kluwe, W. M. (1982). Overview of phthalate ester
pharmacokinetics in mammalian species. Environ.
Health Perspect. 45, 3–9.
1. Nakai, M., Tabira, Y., Asai, D., Yakabe, Y., Shimyozu,
T., Noguchi, M. et al. (1999). Binding characteristics of
dialkyl phthalates for the estrogen receptor. Biochem.
Biophys. Res. Commun. 254, 311–314.
2. Staples, C. A., Parkerton, T. F. & Peterson, D. R. (2000).
A risk assessment of selected phthalate esters in North
American and Western European surface waters.
Chemosphere, 40, 885–891.
3. Ribbons, D. W., Keyser, P., Kunz, D. A., taylor, B. F.,
Eaton, R. W. & Anderson, B. N. (1984). Microbial
degradation of phthalates. In Microbial Degradation of
Organic Compounds (Gibson, D. T., ed.), pp. 371–397,
Dekker, New York.
27. Saller, E. L. H., Schlafioppenberg, H. R. & Cook, A. M.
(1995). Purification and some properties of (1R,2S)-
1,2-dihydroxy-3,5-cyclohexadiene-1,4-dicarboxylate
dehydrogenase from Comamonas testosteroni T-2.
FEMS Microbiol. Lett. 130, 97–102.
28. Wang, Y. Z., Zhou, Y. & Zylstra, G. J. (1995).
Molecular analysis of isophthalate and terephthalate
degradation by Comamonas testosteroni YZW-D. Envi-
ron. Health Perspect. 103, 9–12.
29. Sasoh, M., Masai, E., Ishibashi, S., Hara, H.,
Kamimura, N., Miyauchi, K. & Fukuda, M. (2006).
Characterization of the terephthalate degradation
genes of Comamonas sp. strain E6. Appl. Environ.
Microbiol. 72, 1825–1832.
30. Shigematsu, T., Yumihara, K., Ueda, Y., Numaguchi,
M., Morimura, S. & Kida, K. (2003). Delftia tsuruha-
tensis sp. nov., a terephthalate-assimilating bacterium
isolated from activated sludge. Int. J. Syst. Evol.
Microbiol. 53, 1479–1483.
31. Chain, P. S., Denef, V. J., Konstantinidis, K. T., Vergez,
L. M., Agullo, L., Reyes, V. L. et al. (2006). Burkholderia
xenovorans LB400 harbors a multi-replicon, 9.73-Mbp
genome shaped for versatility. Proc. Natl Acad. Sci.
USA, 103, 15280–15287.
32. Johnson, G. R., Jain, R. K. & Spain, J. C. (2000).
Properties of the trihydroxytoluene oxygenase from
Burkholderia cepacia R34: an extradiol dioxygenase
from the 2,4-dinitrotoluene pathway. Arch. Microbiol.
173, 86–90.
33. Mueller, J. G., Devereux, R., Santavy, D. L., Lantz, S. E.,
Willis, S. G. & Pritchard, P. H. (1997). Phylogenetic and
physiological comparisons of PAH-degrading bacteria
from geographically diverse soils. Antonie Van Leeu-
wenhoek, 71, 329–343.
34. Arnett, C. M., Parales, J. V. & Haddock, J. D. (2000).
Influence of chlorine substituents on rates of oxidation
of chlorinated biphenyls by the biphenyl dioxygenase
of Burkholderia sp. strain LB400. Appl. Environ. Micro-
biol. 66, 2928–2933.
1
1
1
4. Ribbons, D. W. (1988). More about phthalates. In
Microbial Catabolism and the Carbon Cycle (Hagedorn,
S. R., Hanson, R. S. & Kunz, D. A., eds), pp. 85–100,
Harwood, Chur.
1
5. Cartwright, C. D., Owen, S. A., Thompson, I. P. &
Burns, R. G. (2000). Biodegradation of diethyl phthal-
ate in soil by a novel pathway. FEMS Microbiol. Lett.
186, 27–34.
1
1
1
6. Chang, H. K. & Zylstra, G. J. (1998). Novel organiza-
tion of the genes for phthalate degradation from
Burkholderia cepacia DBO1. J. Bacteriol. 180, 6529–6537.
7. Gu, J. D., Li, J. & Wang, Y. (2005). Biochemical
pathway and degradation of phthalate ester isomers
by bacteria. Water Sci. Technol. 52, 241–248.
8. Otsuka, Y., Nakamura, M., Shigehara, K., Sugimura,
K., Masai, E., Ohara, S. & Katayama, Y. (2006).
Efficient production of 2-pyrone 4,6-dicarboxylic
acid as a novel polymer-based material from proto-
catechuate by microbial function. Appl. Microbiol.
Biotechnol. 71, 608–614.
9. Shigehara, K., Y. Katayama, S. Nishikawa, and Y.
Hotta. (2001). Polyester and process for producing the
same. U.S. patent 6,303,745.
1
2
2
0. Shigehara, K., Y. Katayama, S. Nishikawa, and Y.
Hotta. (2002). Polyamide and process for producing
the same. U.S. patent 6,340,739.
35. Goris, J., De Vos, P., Caballero-Mellado, J., Park, J.,
Falsen, E., Quensen, J. F., 3rd et al. (2004). Classifica-
tion of the biphenyl- and polychlorinated biphenyl-
degrading strain LB400T and relatives as Burkholderia
xenovorans sp. nov. Int. J. Syst. Evol. Microbiol. 54,
1677–1681.
1. Engelhardt, G., Rast, H. G. & Wallnofer, P. R.
(
1979). Degradation of aromatic carboxylic acids by
Nocardia spec DSM 43251. FEMS Microbiol. Lett. 5,
45–251.
2
2
2
2
2. Karegoudar, T. B. & Pujar, B. G. (1985). Degradation of
terephthalic acid by a Bacillus species. FEMS Microbiol.
Lett. 30, 217–220.
3. Kimura, T. & Ito, Y. (2001). Two bacterial mixed
culture systems suitable for degrading terephthalate
in wastewater. J. Biosci. Bioeng. 91, 416–418.
4. Sugimori, D., Dake, T. & Nakamura, S. (2000).
Microbial degradation of disodium terephthalate by
alkaliphilic Dietzia sp. strain GS-1. Biosci. Biotechnol.
Biochem. 64, 2709–2711.
36. Hubner, A., Danganan, C. E., Xun, L., Chakrabarty,
A. M. & Hendrickson, W. (1998). Genes for 2,4,5-
trichlorophenoxyacetic acid metabolism in Burkhol-
deria cepacia AC1100: characterization of the tftC and
tftD genes and locations of the tft operons on multiple
replicons. Appl. Environ. Microbiol. 64, 2086–2093.
37. Delano, W. L. (2002). The PyMOL Molecular Graphics
System.
38. Gibson, D. T. (1971). The microbial oxidation of
aromatic hydrocarbons. Crit. Rev. Microbiol. 1, 199–223.
39. Jerina, D. M., Daly, J. W., Witkop, B., Zaltzman-
Nirenberg, P. & Udenfriend, S. (1968). The role of
arene oxide-oxepin systems in the metabolism of
2
5. Choi, K. Y., Kim, D., Sul, W. J., Chae, J. C., Zylstra,
G. J., Kim, Y. M. & Kim, E. (2005). Molecular and
biochemical analysis of phthalate and terephthalate

X-ray Crystallography Studies of TphB
293
aromatic substrates. 3. Formation of 1,2-naphthalene
oxide from naphthalene by liver microsomes. J. Am.
Chem. Soc. 90, 6525–6527.
dehydrogenase from a PCB degrader at 2.0Å resolu-
tion. Protein Sci. 7, 1286–1293.
47. Sivaraman, J., Li, Y., Banks, J., Cane, D. E., Matte, A.
& Cygler, M. (2003). Crystal structure of Escherichia
coli PdxA, an enzyme involved in the pyridoxal
phosphate biosynthesis pathway. J. Biol. Chem. 278,
43682–43690.
48. Auld, D. S. (2001). Zinc coordination sphere in
biochemical zinc sites. Biometals, 14, 271–313.
49. Carbone, V., Hara, A. & El-Kabbani, O. (2008).
Structural and functional features of dimeric dihydro-
diol dehydrogenase. Cell. Mol. Life Sci. 65, 1464–1474.
50. Patil, D. N., Tomar, S., Sylvestre, M. & Kumar, P.
(2010). Expression, purification, crystallization and
preliminary crystallographic studies of cis-biphenyl-
2,3-dihydrodiol-2,3-dehydrogenase from Pandoraea
pnomenusa B-356. Acta Crystallogr., Sect. F: Struct.
Biol. Cryst. Commun. 66, 1517–1520.
51. Holm, L. & Park, J. (2000). DaliLite workbench for
protein structure comparison. Bioinformatics, 16, 566–567.
52. Corpet, F. (1988). Multiple sequence alignment with
hierarchical clustering. Nucleic Acids Res. 16,
10881–10890.
4
0. Oesch, F., Jerina, D. M., Daly, J. W., Lu, A. Y.,
Kuntzman, R. & Conney, A. H. (1972). A reconstituted
microsomal enzyme system that converts naphtha-
lene to trans-1,2-dihydroxy-1,2- dihydronaphthalene
via naphthalene-1,2-oxide: presence of epoxide hy-
drase in cytochrome P-450 and P-448 fractions. Arch.
Biochem. Biophys. 153, 62–67.
4
4
4
1. Matsuura, K., Hara, A., Deyashiki, Y., Iwasa, H.,
Kume, T., Ishikura, S. et al. (1998). Roles of the
C-terminal domains of human dihydrodiol dehydro-
genase isoforms in the binding of substrates and
modulators: probing with chimaeric enzymes. Bio-
chem. J. 336, 429–436.
2. Terada, T., Fujita, N., Sugihara, Y., Sato, R., Takagi, T.
&
Maeda, M. (2001). Site-directed mutagenesis studies
of bovine liver cytosolic dihydrodiol dehydrogenase:
the role of Asp-50, Tyr-55, Lys-84, His-117, Cys-145
and Cys-193 in enzymatic activity. Chem. Biol. Interact.
130–132, 833–845.
3. Hsu, N. Y., Ho, H. C., Chow, K. C., Lin, T. Y., Shih, C. S.,
Wang, L. S. & Tsai, C. M. (2001). Overexpression
of dihydrodiol dehydrogenase as a prognostic
marker of non-small cell lung cancer. Cancer Res.
53. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. &
Lipman, D. J. (1990). Basic local alignment search tool.
J. Mol. Biol. 215, 403–410.
61, 2727–2731.
54. Collaborative Computational Project Number 4.
(1994). The CCP4 Suite—programs for protein crys-
tallography. Acta Crystallogr., Sect. D, 50, 760–763.
55. Perrakis, A., Morris, R. & Lamzin, V. S. (1999).
Automated protein model building combined with
iterative structure refinement. Nat. Struct. Biol. 6, 458–463.
56. Emsley, P. & Cowtan, K. (2004). Coot: model-building
tools for molecular graphics. Acta Crystallogr., Sect. D,
60, 2126–2132.
57. Murshudov, G. N., Vagin, A. A. & Dodson, E. J.
(1997). Refinement of macromolecular structures by
the maximum-likelihood method. Acta Crystallogr.,
Sect. D, 53, 240–255.
4
4. Patel, T. R. & Gibson, D. T. (1974). Purification and
propeties of (plus)-cis-naphthalene dihydrodiol dehydro-
genase of Pseudomonas putida. J. Bacteriol. 119, 879–888.
5. Sundlov,J. A., Garringer, J. A., Carney, J. M., Reger, A. S.,
Drake, E. J., Duax, W. L. & Gulick, A. M. (2006).
Determination of the crystal structure of EntA, a 2,3-
dihydro-2,3-dihydroxybenzoic acid dehydrogenase
from Escherichia coli. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 62, 734–740.
4
4
6. Hulsmeyer, M., Hecht, H. J., Niefind, K., Hofer, B.,
Eltis, L. D., Timmis, K. N. & Schomburg, D. (1998).
Crystal structure of cis-biphenyl-2,3-dihydrodiol-2,3-