Mechanistic divergence of two closely related aldol-like enzyme-catalysed reactions

Article abstract of DOI:10.1039/b511661a

The mechanistic divergence of two closely related aldol-like enzyme-catalyzed reaction is studied. The interaction of threose 4-phosphate with 3-deoxyerythrose 4-phosphate with 3-deoxy -D-arabino-heptulosonate 7-phosphate synthase (DAH7PS) revealed previo

Full text of DOI:10.1039/b511661a

View Article Online / Journal Homepage / Table of Contents for this issue
C O M M U N I C A T I O N
Mechanistic divergence of two closely related aldol-like
enzyme-catalysed reactions†
Meekyung Ahn, Amy L. Pietersma, Linley R. Schofield and Emily J. Parker*
Institute of Fundamental Sciences, Massey University, Private Bag 11-222, Palmerston North,
New Zealand. E-mail: E.J.Parker@massey.ac.nz; Fax: +64 6 350 5682; Tel: +64 6 356 9099
Received 16th August 2005, Accepted 23rd September 2005
First published as an Advance Article on the web 6th October 2005
The analysis of the interaction of threose 4-phosphate
and 2-deoxyerythrose 4-phosphate with 3-deoxy-D-arabino-
heptulosonate 7-phosphate synthase (DAH7PS) reveals
previously unrecognised mechanistic differences between
the DAH7PS-catalysed reaction and that catalysed by the
closely related enzyme, 3-deoxy-D-manno-octulosonate 8-
phosphate synthase (KDO8PS).
on molecular mass and sequence.19 The type I enzymes are a
broad family of 3-deoxyald-2-ulosonate phosphate synthases
that includes the KDO8PSs. This group is also divided into two
subfamilies (Ia and Ib).16 Currently only DAH7PSs are known
in subfamily Ia, whereas subfamily Ib contains both DAH7PSs
(IbD) and KDO8PSs (IbK). It has been proposed that the ances-
tor of the type I enzymes was a DAH7PS of the IbD subfamily.16,17
While the similarities are clear, two key differences in these
enzyme-catalysed reactions have intrigued us. Firstly, whereas
all known DAH7PSs require a divalent metal for catalysis,
both metal-dependent and metal-independent KDO8PSs have
been characterised.20 Recent studies have shown that metal-
dependent KDO8PSs can be converted to metal-independent
enzymes following mutation of the metal-binding Cys to
Asn (the “natural” substitution found in metal-independent
enzymes).21,22 Secondly, the substrate specificity with respect
to the configuration at C2 of the aldose phosphate substrate
differs; and it has been reported that the C2 epimer of A5P (D-
ribose 5-phosphate, R5P) with E4P-like C2 configuration is not
a substrate for KDO8PSs from E. coli and A. aeolicus.2,23,24
In this study we have probed these two key differences by
mutation at the metal-binding site of the IbD enzyme from P. fu-
riosus, and by investigation of D-threose 4-phosphate (T4P), and
2-deoxy-D-erythrose 4-phosphate (2-deoxyE4P) as substrates
for DAH7PS, and R5P and 2-deoxy-D-ribose 5-phosphate (2-
deoxyR5P) as substrates for KDO8PS. Our analysis illuminates
significant and previously unrecognised differences in the cat-
alytic mechanisms of DAH7PS and KDO8PS.
3-Deoxy-D-arabino-heptulosonate
7-phosphate
synthase
(DAH7PS) and 3-deoxy-D-manno-octulosonate 8-phosphate
synthase (KDO8PS) are two functionally unrelated enzymes
that share many mechanistic and structural features. DAH7PS
(EC 2.5.1.54) catalyses the first committed step of the
shikimate pathway, responsible for the biosynthesis of aromatic
compounds.1 This enzyme catalyses an aldol-like condensation
reaction between phosphoenolpyruvate (PEP) and D-erythrose
4-phosphate (E4P) to generate DAH7P and inorganic
phosphate. KDO8PS (EC 2.5.1.55) catalyses an analogous
reaction using D-arabinose 5-phosphate (A5P) in place of E4P,
giving rise to the eight carbon phosphorylated sugar KDO8P
(Fig. 1). This reaction is a key step in the biosynthesis of cell
wall lipopolysaccharide in Gram-negative bacteria.2 As both
pathways are found in microorganisms but not in animals, the
enzymes of these pathways have attracted interest as targets for
the development of novel antibiotics.1,3
P. furiosus DAH7PS is the most closely related DAH7PS to
KDO8PS yet characterised.15 Unlike other DAH7PSs (and like
KDO8PSs) it is not subject to allosteric inhibition by shikimate
pathway end products, and its structure reveals a basic catalytic
(b/a)8-barrel with no significant extensions.15,25 It has also been
shown to be relatively ambiguous with respect to substrate
selection with an expanded ability to accept the five-carbon
phosphorylated sugars A5P and R5P.15 Therefore we chose this
enzyme to investigate the dispensability of metal-dependency
in DAH7PSs. The P. furiosus DAH7PS Cys31Asn mutant was
constructed and purified following established procedures.26
This enzyme appeared to be identical to wild-type by native
and SDS-PAGE. Comparison of the UV-visible spectrum of
the Cys31Asn mutant in the presence of Cu2+ to that obtained
with wild-type enzyme indicated that this mutant enzyme was
unable to bind metal ions.27 In contrast to the observations made
with the equivalent mutations in the metal dependent KDO8PSs
from both Aquifex pyrophilis and A. aeolicus,21,22 the Cys31Asn
mutant of P. furiosus DAH7PS showed no detectable enzymic
activity with or without EDTA or added Mn2+.
T4P was synthesised from D-diethyl tartrate in seven steps
with an overall yield of 25% (Scheme 1). D-Diethyl tartrate was
benzylated and reduced. The resulting diol was then monophos-
phorylated and oxidised using Dess–Martin periodinane, and
the aldehyde functionality was protected as the dimethyl acetal.
Hydrogenolysis and cleavage of the acetal gave rise to T4P. For
the preparation of 2-deoxyE4P (S)-b-hydroxy-c-butyrolactone
was benzylated and then reduced to the lactol (Scheme 2).
Following protection of the aldehyde functionality, the C4
Fig. 1 KDO8PS and DAH7PS reactions.
Despite low sequence similarity between DAH7PS and
KDO8PS many key mechanistic similarities have been shown.
Both enzymes catalyse the condensation of PEP with a phos-
phorylated aldose by a similar ordered-sequential mechanism
where PEP binds first and phosphate is released last.4,5 Both
reactions involve the cleavage of the C–O bond of PEP,6,7 and
are highly stereospecific with the si face of PEP coupling with
the re face of their respective sugar substrates.8,9 Additionally,
X-ray crystal structures of DAH7PSs (from Escherichia coli,
Thermotoga maritima, Pyrococcus furiosus and Saccharomyces
cerevisiae) have been found to be remarkably similar to those of
KDO8PSs (Aquifex aeolicus, E. coli).10–15
Based on these similarities the phylogenetic relation-
ship between DAH7PS and KDO8PS has recently received
attention.16–18 Two types of DAH7PS have been identified based
† Electronic supplementary information (ESI) available: Experimental
for the preparation of the aldose phosphate substrate analogues. See
DOI: 10.1039/b511661a
4 0 4 6
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 0 4 6 – 4 0 4 9
T h i s j o u r n a l i s
T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5
©

View Article Online
hydroxyl group was phosphorylated. Deprotection with H2
over platinum followed by dissolution in water gave 2-deoxyE4P.
T4P and 2-deoxyE4P were tested as an alternative substrate
for DAH7PSs from both E. coli (Ia) and P. furiosus (IbD).
Both DAH7PSs were able to accept T4P and 2-deoxyE4P as
alternative substrates to E4P (Table 1). Intriguingly, for P.
furiosus DAH7PS, 2-deoxyE4P was preferred as a substrate
to the “natural” substrate E4P, and T4P was utilised with
a comparable efficiency. For E. coli DAH7PS a significant
(but similar) increase in the Km value was recorded with
both alternative substrates. For both T4P and 2-deoxyE4P the
seven carbon phosphorylated sugar products of the enzymatic
reactions were isolated and characterised by both NMR and
HRMS and were identified as the expected 3-deoxy-D-lyxo-
heptulosonate 7-phosphate28 and 5-deoxyDAH7P29 respectively.
Commercially available R5P and 2-deoxyR5P were tested
as substrates for the KDO8PS from Neisseria meningitidis31
(Table 2). 2-DeoxyR5P was a poor substrate, and no evidence of
substrate activity was observed for R5P, even at high substrate
and enzyme concentrations. These observations parallel those
made for the KDO8PS from E. coli.23 Neither E4P, T4P or
2-deoxyE4P were able to act as alternative substrates for N.
meningitidis KDO8PS. However, DAH7PS from both E. coli and
P. furiosus were able to utilise the five carbon phosphorylated
sugars A5P, R5P and 2-deoxyR5P.15,32
Scheme 1 Synthesis of T4P.
What then can account for differences in metal requirement
and sugar substrate C2 configuration between DAH7PS and
KDO8PS as detailed in these studies? On one hand, DAH7PSs
appear to show an absolute requirement for metal for catalytic
function, whereas for KDO8PSs the metal-binding is dispens-
able. On the other hand, DAH7PSs show an ambivalence to
sugar substrate C2 configuration, yet KDO8PSs have a strict
requirement for the correct (and opposite) stereochemistry at
this position.
Careful comparison of the active sites of DAH7PS and
KDO8PS reveals three absolutely conserved substitutions to
residues that interact directly with either PEP or the aldose
phosphate substrate10–15,24 (Fig. 2): an Arg binds the PEP
carboxylate in DAH7PS, whereas Lys is found in KDO8PS;
secondly, an Arg in the PEP phosphate binding site in DAH7PS
is substituted by Phe in KDO8PS; and, thirdly, in the aldose
phosphate binding site, a Pro to AlaAsn substitution is found
in the absolutely conserved LysProArgThr motif of DAH7PS
(creating an equivalent conserved LysAlaAsnArgSer motif in
KDO8PS). We propose that these latter two differences account
for both the altered substrate specificity and metal requirement
and give rise to different catalytic mechanisms.
Scheme 2 Synthesis of 2-deoxyE4P.
Table 1 Kinetic parameters for E4P, T4P and 2-deoxyE4P with DAH7PS
Substrate
DAH7PS Sourcea
E4P
T4P
390 13
2.5 0.2
0.006 0.001
2-DeoxyE4P
410 40
E. coli
Km/lM
39
26
0.7 0.1
4
3
k
k
cat/s−1
25
3
cat/Km/s−1 lM−1
0.06 0.01
P. furiosus
Km/lM
9
115
21
1
6
115
k
k
cat/s−1
1.4 0.115
0.16 0.03
2.4 0.1
0.11 0.01
3.0 0.115
0.5 0.1
cat/Km/s−1 lM−1
a DAH7PS from both sources were assayed in accordance with previously reported procedures.15,30
Table 2 Kinetic parameters for A5P, R5P and 2-deoxyR5P with N. meningitidis KDO8PS
Substrate
A5P
12
2.7 0.6
0.23 0.05
R5P
2-DeoxyR5P
N. meningitidis KDO8PS31
Km/lM
kcat /s−1
1
Not a substrate
230 20
0.13 0.01
0.0006 0.0001
k
cat/Km/s−1 lM−1
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 0 4 6 – 4 0 4 9
4 0 4 7

View Article Online
the re face of PEP, observed in both DAH7PS and KDO8PS
structures, is the likely candidate, giving overall (favourable) anti
addition to PEP. Studies with enzyme mutants and substrate
analogues are currently underway to test further these proposed
mechanisms.
It should be noted that the structure of A. aeolicus KDO8PS
in complex with R5P has also been solved.24 Binding of R5P
appeared to disrupt the coordination of a water molecule
proposed at that time to act as the nucleophilic water in
the catalytic mechanism. However, a direct catalytic role for
the divalent metal ion in water activation has been largely
discounted as metal-independent KDO8PS enzymes arise due
to a single Cys to Asn mutation.21,22 What is clear from these
structures however, is that the carbonyl functionality in R5P
adopts a significantly different orientation, entirely consistent
with our analysis where we predict that appropriate interaction
with the C2 hydroxyl is vital for addition of C3 of PEP to
the carbonyl of A5P. The ability of 2-deoxyR5P to act as an
alternative, yet poor substrate for KDO8PS noted in this study
and by others23 is consistent with a greater likelihood of this
analogue accessing the reactive conformation. Likewise, this
analysis would also appear to account for the recently reported
disparity in behaviour with 3-fluoroPEP.33 In this study the
existence of significant differences between the PEP subsites of
KDO8PS and DAH7PS have been proposed in order to account
for the different abilities of the enzymes to process the two 3-
fluoroPEP diastereoisomers.
Identification of key mechanistic similarities, the discovery
of metal-dependent KDO8PSs, and phylogenetic analysis has
led to the assumption that a common mechanism applies to
these two related enzyme-catalysed reactions. This present study,
together with a reinterpretation of existing substrate specificity
and structural data, suggests that the evolutionary process that
led to altered substrate specificity also gave rise to different
mechanisms of catalysis. Consequently the divalent metal on
which some KDO8PSs rely for catalytic activity, plays an
altered and dispensable role in the enzyme-catalysed reaction.
Its presence, however, in some enzymes, albeit as an evolutionary
carry-over, provides further evidence for a common DAH7PS-
like ancestor for this enzyme family.
Fig. 2 Comparison of active sites and proposed (partial) reaction
mechanisms for (a) KDO8PS (A. aeolicus, PDB 1FWW12) (b) DAH7PS
(P. furiosus with E4P modelled, PDB 1ZCO15). E4P has been modelled
into this structure based on the observed binding of glycerol 3-phosphate
to S. cerevisiae DAH7PS13 and the proposed binding of E4P to
T. maritima DAH7PS.14 The key changes discussed in the text are
highlighted in green. Metal and metal ligands are in cyan and PEP
ligands are shown in blue. Substrates, PEP and A5P (or E4P) are shown
in black.
Substitution of positively charged Arg to hydrophobic Phe
in KDO8PS eliminates a salt bridge to the PEP phosphate that
is found in DAH7PS, and increases the hydrophobicity in the
vicinity of the PEP phosphate group in KDO8PS. Modelling
studies suggest that this allows the aldehyde functionality of
A5P in KDO8PS to be positioned differently, and PEP to be
bound to KDO8PS in its dianionic rather than the trianion
form. Therefore, in KDO8PS the phosphate moiety of PEP
may hydrogen-bond to the aldehyde oxygen of A5P. The second
key substitution (LysAlaAsnArgSer rather than LysProArgThr)
provides an additional binding contact for A5P ensuring correct
placement of the aldehyde moiety in KDO8PS close to the PEP
phosphate moiety.
Acknowledgements
This work was supported by the Royal Society of New Zealand
Marsden Fund (MAU008). The authors gratefully acknowledge
Geoffrey Jameson for helpful discussions and advice, and Fiona
Cochrane and Mark Patchett for preparation of N. meningitidis
KDO8PS.
The key chemical event in these condensation reactions is
attack by C3 of PEP on the aldehyde group of co-substrate
E4P or A5P. In DAH7PS, structural and modelling studies are
consistent with activation of the aldehyde by the metal13,14 (Lewis
acid catalysis), meaning that the divalent metal plays an essential
and indispensable role. In contrast, for KDO8PS, the metal
Notes and references
1 R. Bentley, Crit. Rev. Biochem. Mol. Biol., 1990, 25, 307.
2 P. H. Ray, J. Bacteriol., 1980, 141, 635.
3 C. R. Raetz, Annu. Rev. Biochem., 1990, 59, 129.
4 A. B. DeLeo, J. Dayan and D. B. Sprinson, J. Biol. Chem., 1973, 248,
2344.
5 T. Baasov, S. Sheffer-Dee-Noor, A. Kohen, A. Jakob and V. Belakhov,
Eur. J. Biochem., 1993, 217, 991.
6 A. B. DeLeo and D. B. Sprinson, Biochem. Biophys. Res. Commun.,
1968, 32, 873.
7 L. Hedstrom and R. Abeles, Biochem. Biophys. Res. Commun., 1988,
157, 816.
8 D. K. Onderka and H. G. Floss, J. Am. Chem. Soc., 1969, 91,
5894.
9 A. Kohen, R. Berkovich, V. Belakhov and T. Baasov, Bioorg. Med.
Chem. Lett., 1993, 3, 1577.
10 I. A. Shumilin, R. H. Kretsinger and R. H. Bauerle, Structure
(London), 1999, 7, 865.
11 S. Radaev, P. Dastidar, M. Patel, R. W. Woodard and D. L. Gatti,
J. Biol. Chem., 2000, 275, 9476.
12 H. S. Duewel, S. Radaev, J. Wang, R. W. Woodard and D. L. Gatti,
J. Biol. Chem., 2001, 276, 8393.
˚
(when present) is too far from the aldehyde functionality (∼6 A)
to be involved in electrophilic activation,12 and activation is by
protonation (Brønsted acid catalysis). In KDO8PS, activation
and positioning of the aldehyde moiety is more delicately
choreographed. For KDO8PS the C2 hydroxyl group plays a
critical role, as via coordination to metal or an Asn side chain
(most likely via an intermediate water), the dihedral angle about
the C1–C2 bond of A5P is controlled. Consequently, altering
the configuration of C2 of A5P is catalytically catastrophic.
DAH7PS on the other hand is tolerant to changes in the C2
position, with the P. furiosus accepting T4P or 2-deoxyE4P,
with similar or greater efficiency than the natural substrate. This
alternative placement of the aldehyde moiety means that despite
the different C2 configuration in A5P and E4P, attack by C3 of
PEP on its aldose co-substrate follows the Felkin-Anh model in
both condensation reactions.
For either enzymic reaction, formation of the oxycarbenium
ion intermediate (or transition state) will be followed by the
attack of water on C2 of PEP. A water molecule located on
4 0 4 8
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 0 4 6 – 4 0 4 9

View Article Online
13 V. Konig, A. Pfeil, G. H. Braus and T. R. Schneider, J. Mol. Biol.,
2004, 337, 675.
14 I. A. Shumilin, R. Bauerle, J. Wu, R. W. Woodard and R. H.
Kretsinger, J. Mol. Biol., 2004, 341, 455.
15 L. R. Schofield, B. F. Anderson, M. L. Patchett, G. E. Norris, G. B.
Jameson and E. J. Parker, Biochemistry, 2005, 44, 11950.
16 P. S. Subramaniam, G. Xie, T. Xia and R. A. Jensen, J. Bacteriol.,
1998, 180, 119.
17 R. A. Jensen, G. Xie, D. H. Calhoun and C. A. Bonner, J. Mol. Evol.,
2002, 54, 416.
18 M. R. Birck and R. W. Woodard, J. Mol. Evol., 2001, 52, 205.
19 G. E. Walker, B. Dunbar, I. S. Hunter, H. G. Nimmo and J. R.
Coggins, Microbiology (Reading, U. K.), 1996, 142, 1973.
20 H. S. Duewel and R. W. Woodard, J. Biol. Chem., 2000, 275, 22824.
21 S. Shulami, C. Furdui, N. Adir, Y. Shoham, K. S. Anderson and T.
Baasov, J. Biol. Chem., 2004, 279, 45110.
22 J. Li, J. Wu, A. S. Fleischhacker and R. W. Woodard, J. Am. Chem.
Soc., 2004, 126, 7448.
23 D. L. Howe, A. K. Sundaram, J. Wu, D. L. Gatti and R. W. Woodard,
Biochemistry, 2003, 42, 4843.
26 P. furiosus DAH7PS mutants were prepared utilizing a QuikChange
II Site-Directed Mutagenesis Kit (Stratagene). A conservative substi-
tution of a residue near the active site, yet not implicated in substrate
binding or catalysis (Ser64Thr), had activitycomparable to wild-type
enzyme.
27 P. A. Jordan, D. S. Bohle, C. A. Ramilo and J. N. S. Evans,
Biochemistry, 2001, 40, 8387.
28 The 1H NMR spectrum for this compound was very similar to that
reported for KDO8P where the configuration of C4, C5 and C6
is identical to that predicted for the product arising from reaction
between T4P and PEP.
29 R. M. Williamson, A. L. Pietersma, G. B. Jameson and E. J. Parker,
Bioorg. Med. Chem. Lett., 2005, 15, 2339.
30 E. J. Parker, E. M. M. Bulloch, G. B. Jameson and C. Abell,
Biochemistry, 2001, 40, 14821.
31 N. meningitidis KDO8PS was expressed, purified and assayed using
similar methods to those used for the cloning and characterisation
of the very closely related KDO8PS from Neisseria gonorrhoeae
(G. Y. Sheflyan, A. K. Sundaram, W. P. Taylor and R. W. Woodard,
J. Bacteriol., 2000, 182, 5005).
24 J. Wang, H. S. Duewel, R. W. Woodard and D. L. Gatti, Biochemistry,
2001, 40, 15676.
25 L. R. Schofield, M. L. Patchett and E. J. Parker, Protein Expression
Purif., 2004, 34, 17.
32 G. Y. Sheflyan, D. L. Howe, T. L. Wilson and R. W. Woodard, J. Am.
Chem. Soc., 1998, 120, 11027.
33 C. M. Furdui, A. K. Sau, O. Yaniv, V. Belakhov, R. W. Woodard, T.
Baasov and K. S. Anderson, Biochemistry, 2005, 44, 7326.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 0 4 6 – 4 0 4 9
4 0 4 9