Linear free energy relationships demonstrate a catalytic role for the flavin mononucleotide coenzyme of the type II isopentenyl diphosphate: Dimethylallyl diphosphate isomerase

Article abstract of DOI:10.1021/ja104090m

The type II isopentenyl diphosphate:dimethylallyl diphosphate isomerase (IDI-2) catalyzes the reversible isomerization of the two ubiquitous isoprene units, isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP), which are required to initiate the biosynthesis of all isoprenoid compounds found in nature. The overall chemical transformation catalyzed by IDI-2 involves a net 1,3-proton addition/elimination reaction. Surprisingly, IDI-2 requires a reduced flavin mononucleotide (FMN) coenzyme to carry out this redox neutral isomerization. The exact function of FMN in catalysis has not yet been clearly defined. To provide mechanistic insight into the role of the reduced flavin in IDI-2 catalysis, several FMN analogues with altered electronic properties were chemoenzymatically prepared, and their effects on the kinetic properties of the IDI-2 catalyzed reaction were investigated. Linear free energy relationships (LFERs) between the electronic properties of the flavin and the steady state kinetic parameters of the IDI-2 catalyzed reaction were observed. The LFER studies are complemented with kinetic isotope effect studies and kinetic characterization of an active site mutant enzyme (Q154N). Cumulatively, the data presented in this work (and in other studies) suggest that the reduced FMN coenzyme of IDI-2 functions as an acid/base catalyst, with the N5 atom of the flavin likely playing a critical role in the deprotonation of IPP en route to DMAPP formation. Several potential chemical mechanisms involving the reduced flavin as an acid/base catalyst are presented and discussed.

Full text of DOI:10.1021/ja104090m

Published on Web 07/01/2010
Linear Free Energy Relationships Demonstrate a Catalytic Role for the Flavin
Mononucleotide Coenzyme of the Type II Isopentenyl
Diphosphate:Dimethylallyl Diphosphate Isomerase
Christopher J. Thibodeaux, Wei-chen Chang, and Hung-wen Liu*
DiVision of Medicinal Chemistry, College of Pharmacy, Department of Chemistry and Biochemistry, and Institute of
Cellular and Molecular Biology, UniVersity of Texas, Austin, Texas 78712
Received May 12, 2010; E-mail: h.w.liu@mail.utexas.edu
Scheme 1. Potential IDI-2 Chemical Mechanisms
Abstract: The type II isopentenyl diphosphate:dimethylallyl diphos-
phate isomerase (IDI-2) catalyzes the reversible isomerization of the
two ubiquitous isoprene units, isopentenyl pyrophosphate (IPP) and
dimethylallyl pyrophosphate (DMAPP), which are required to initiate
the biosynthesis of all isoprenoid compounds found in nature. The
overall chemical transformation catalyzed by IDI-2 involves a net
1,3-proton addition/elimination reaction. Surprisingly, IDI-2 requires
a reduced flavin mononucleotide (FMN) coenzyme to carry out this
redox neutral isomerization. The exact function of FMN in catalysis
has not yet been clearly defined. To provide mechanistic insight into
the role of the reduced flavin in IDI-2 catalysis, several FMN
analogues with altered electronic properties were chemoenzymati-
cally prepared, and their effects on the kinetic properties of the IDI-2
catalyzed reaction were investigated. Linear free energy relationships
(LFERs) between the electronic properties of the flavin and the
steady state kinetic parameters of the IDI-2 catalyzed reaction were
observed. The LFER studies are complemented with kinetic isotope
effect studies and kinetic characterization of an active site mutant
enzyme (Q154N). Cumulatively, the data presented in this work (and
in other studies) suggest that the reduced FMN coenzyme of IDI-2
functions as an acid/base catalyst, with the N5 atom of the flavin likely
playing a critical role in the deprotonation of IPP en route to DMAPP
formation. Several potential chemical mechanisms involving the reduced
flavin as an acid/base catalyst are presented and discussed.
Isopentenyl diphosphate:dimethylallyl diphosphate isomerases (IDI)
catalyze the reversible interconversion of isopentenyl pyrophosphate
(IPP, 1) and dimethylallyl pyrophosphate (DMAPP, 2) used in the
biosynthesis of isoprenoids in all living organisms (see Scheme 1).1
Two structurally distinct IDIs exist in nature. The type I IDI (IDI-1)
requires only a divalent metal ion for catalysis, whereas the type II
enzyme (IDI-2) requires both a divalent metal ion and a reduced flavin
mononucleotide (FMN) coenzyme.2 Mechanistic studies of IDI-1
indicate that it uses acid-base chemistry provided by active site amino
acids to catalyze a proton addition-elimination reaction via an electron
deficient carbocation-like intermediate or transition state.3-6 The
mechanism of IDI-2 enzymes is less well understood, and in particular,
it is unclear what role the reduced flavin coenzyme plays in this
isomerization reaction with no net redox change.
Upon binding 1 or 2, the IDI-2 bound reduced FMN is rapidly
converted into a species whose UV-visible absorbance is most
consistent with either an anionic reduced FMNH- (3) with slightly
altered absorption properties relative to the resting state or a protonated
reduced flavin (FMNH2, either 4 or 5, Scheme 1).7,8 Following
formation of the enzyme ·IPP complex, several mechanisms in which
the reduced flavin plays a catalytic role can be envisioned for the redox-
neutral isomerization catalyzed by IDI-2 (see Scheme 1, paths
A-E).7-9 The detection of trace amounts of a magnetically isolated
neutral flavin semiquinone (6, Scheme 1, path A) in reaction mixtures
by EPR provided initial evidence for a radical mechanism. However,
no coupled substrate radical (such as 7) was observed.7,8 Subsequent
attempts to detect 6 in stopped-flow studies under single turnover
conditions failed,8 and several radical clock mechanistic probes also
failed to reveal the existence of radical substrate intermediates.10 These
results cast doubt on an isomerization mechanism involving single
electron transfer. Alternatively, the anionic, reduced FMN (3) could
help to electrostatically stabilize a substrate-derived carbocation (8,
Scheme 1, path B), generated through acid-base chemistry mediated
by active site amino acid residues. However, a recently solved crystal
structure of IDI-2 in complex with reduced FMN, Mg2+, and IPP (1)
showed no ionizable amino acid side chains in the vicinity of the C2
atom of the bound IPP substrate, raising questions as to the identity
9
9994 J. AM. CHEM. SOC. 2010, 132, 9994–9996
10.1021/ja104090m  2010 American Chemical Society

C O M M U N I C A T I O N S
of the putative proton donor and acceptor in this mechanism.11 Another
possibility is that the reduced FMN coenzyme participates directly in
acid-base catalysis, either with the assistance of an amino acid group
(Scheme 1, path C) or by serving as both an acid and a base via
FMNH2 or its zwitterionic tautomer (Scheme 1, paths D and E,
respectively).
flavin intermediate (Scheme 1, 3, 4, or 5) does not become significantly
rate-limiting as the flavin structure is altered, we analyzed the effects
of 7-Cl-FMN (12, the slowest of the flavin analogues employed in
this study) on the presteady state kinetics of flavin intermediate
formation and decay under single turnover conditions with IPP as the
substrate. As shown in Figure 2, the pattern of presteady state
absorption changes with FMN- and 7-Cl-FMN-reconstituted IDI-2 are
very similar. Namely, the flavin intermediate accumulates rapidly
relative to turnover and then decays to an equilibrium level in a
kinetically competent slow phase. As with kcat and kcat/Km, both of the
observed presteady state rate constants (kfast and kslow) were reduced
in the reaction with IDI-2•7-Cl-FMN relative to their values in the
IDI-2•FMN-catalyzed reaction. However, it is important to note that
the fast phase of flavin intermediate formation (kfast ) 12.1 ( 0.1 s-1
and 1.30 ( 0.01 s-1) remains ∼30- to 40-fold faster than steady state
turnover (kcat ) 0.47 ( 0.01 s-1 and 0.029 ( 0.001 s-1) for the FMN
and 7-Cl-FMN coenzymes, respectively. Thus, there do not appear to
be any major changes in the mechanism when the flavin coenzyme is
altered, and the 7-Cl-FMN intermediate still accumulates rapidly in
the presteady state, presumably due to a subsequent and partially rate-
determining chemistry step in the kinetic mechanism.
In our previous rapid-mix chemical-quench studies, it was found
that a step in the kinetic mechanism prior to or concomitant with
DMAPP formation limits turnover in the reaction catalyzed by IDI-2
from Staphylococcus aureus.8 In addition, a 1° substrate deuterium
kinetic isotope effect was measured on kcat when (R)-[2-2H]-IPP was
used as the substrate, suggesting that C2-H/D bond cleavage is at
least partially rate-limiting.8 Based on these observations, we reasoned
that if the reduced FMN plays a direct catalytic role in the isomerization
step, then alteration of the electronic properties of the reduced flavin
should affect the steady state kinetic parameters. To assess this
possibility, we chemoenzymatically synthesized a series of FMN
analogues (9-12, Figure 1A and Figure S1) substituted with various
electron-donating and -withdrawing groups at the 7- and 8- positions
(meta and para, respectively, to the N5 atom of FMN) and investigated
the effects of these analogues on the kinetics of the IDI-2-catalyzed
reaction. When the steady state kinetic parameters for the conversion
of IPP to DMAPP were determined using IDI-2 reconstituted with
the various reduced FMN analogues, a clear trend was noted. As shown
in Figures 1 and S3, a linear relationship was found between the
logarithm of the steady state kinetic constants (kcat and kcat/Km) versus
either the sum of the Hammett inductive substituent constants12 of
the 7- and 8-substituents (Figure 1B) or the estimated pKa of the N5
atom, taken to be equivalent to the pKa of the corresponding aniline
derivatives (Figure 1C).13 These results, combined with our previous
presteady state and kinetic isotope effect studies, strongly suggest that
the electronic properties of the reduced FMN coenzyme of IDI-2 play
a critical role in mediating the isomerization of IPP (1) to DMAPP
(2). The negative slopes in the Hammett plots (F ≈ -2.0) are consistent
with a mechanism where the flavin loses electron density in the step(s)
that limit steady state turnover. Similarly, the positive slopes in the
Brønsted plots (ꢀ ≈ 0.7) are consistent with a mechanism involving
proton transfer to the N5 atom of the reduced flavin. Together, these
results suggest that the flavin N5 atom may be serving as a general
base catalyst in the conversion of IPP (1) to DMAPP (2).
Figure 2. Formation and decay of the FMN intermediate (A) and the 7-Cl-
FMN intermediate (B) in the presteady state under single turnover
conditions. The amplitudes and rates for the two kinetic phases were
determined by nonlinear regression as described in the Supporting Informa-
tion. For the FMN-catalyzed reaction, the amplitudes and rates of the fast
and slow phases are as follows: Afast ) 0.179(4) AU, kfast ) 12.1(1) s-1
,
Aslow ) -0.090(4) AU, kslow ) 4.3(1) s-1, and offset, C ) 0.0164(1) AU.
For the 7-Cl-FMN-catalyzed reaction: Afast ) 0.0410(2) AU, kfast ) 1.30(1)
s-1, Aslow ) -0.0281(2) AU, kslow ) 0.096(1) s-1, and C ) 0.0197(1) AU.
The reduction in both presteady state rate constants (kfast and kslow
)
with 7-Cl-FMN suggests that, in addition to the chemical step involving
IPP C2-H bond cleavage, other steps in the kinetic mechanism (such
as the formation of the flavin intermediate upon IPP binding) are
sensitive to the electronic properties of the flavin. To assess whether
these other steps are more sensitive to the flavin substituents than the
C2-H bond cleavage step, we investigated the effects of the flavin
analogues on the substrate deuterium kinetic isotope effect (Dkcat) using
the stereospecifically deuterated analogue, (R)-[2-2H]-IPP (Figures 3
and S4). We anticipated that if other steps in the kinetic mechanism
are more (or less) sensitive to the flavin substiuents than the chemical
step involving C2-H/D bond cleavage, then the expression of the KIE
on kcat should change. The primary Dkcat values (ranging from 1.6-2.8)
suggest that C2-H/D bond cleavage is at least partially rate-limiting
with each flavin analogue (Figure 3A). Hammett plots of the reactions
with IPP and (R)-[2-2H]-IPP (Figure 3B) give F values within
experimental error of each other. Thus, if multiple steps in the kinetic
mechanism are sensitive to the flavin substituents, they appear to be
effected to similar extents, such that the overall expression of the KIE
on kcat is relatively unchanged. Interestingly, while the F values for
the two reactions are within experimental error of each other, the
magnitude of F for the (R)-[2-2H]-IPP catalyzed reaction appears to
Figure 1. Flavin substituent effects on the steady state kinetic parameters.
(A) Structures of the FMN analogues (in their oxidized forms) employed
in this study. Hammett (B) and Brønsted (C) plots of log kcat/Km and log
kcat (O and b, respectively). The data were fitted with linear regression
giving F values of -2.1(5) and -2.1(3) for kcat/Km and kcat, respectively, in
the Hammett plots, and ꢀ values of 0.7(2) and 0.7(1) for kcat/Km and kcat
,
respectively, in the Brønsted plots.
The linear relationship between the steady state kinetic constants
and the electronic properties of the flavins suggest that the overall
mechanism of the IDI-2 catalyzed reaction remains unchanged as the
flavin substituents are altered. In support of this, the absorbance changes
in the reduced, IDI-2 bound flavin analogues at 440 nm in the presence
of IPP are very similar (Figure S2), suggesting that a similar flavin
intermediate forms upon IPP binding. To ensure that formation of the
9
J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010 9995

C O M M U N I C A T I O N S
D
be slightly larger, leading to a slight increase in kcat as the flavin
becomes a less efficient catalyst. This trend would suggest that the
C2-H/D bond cleavage step becomes more rate-limiting as the flavin
becomes a poorer catalyst, but the sensitivity of our assay currently
precludes this conclusion.
base functional groups (Scheme 1, paths D and E). If the 1,5-dihydro-
FMNH2 tautomer (4) is utilized as the active species in the isomer-
ization reaction (Scheme 1, path D), the N1 atom may serve as the
proton source, with N5 serving as the base in the forward direction.
In this mechanism, the trends observed in the LFER studies could be
explained if the N5 atom “senses” the inductive effects of the
substituents more strongly than the N1 atom. If the zwitterionic FMNH2
tautomer (5) is employed (Scheme 1, path E), the substrate protonation
step may be faster than the deprotonation step in this obligate stepwise
mechanism, leading to the observed LFER trends. In these latter two
mechanisms, the Q154 residue may function to keep the substrate and
flavin oriented properly to ensure optimal proton transfer rates.
In summary, LFER studies using several flavin analogues have
provided evidence that the reduced flavin coenzyme of S. aureus
IDI-2 plays a direct role in the isomerization of IPP and DMAPP.
Our data are most consistent with a mechanism (Scheme 1C) where
the N5 atom of the flavin serves as a general acid-base catalyst,
perhaps in conjunction with Q154, to effect the isomerization of
the IPP double bond using proton addition-elimination chemistry.
These studies can now be added to the growing body of experi-
mental evidence7-11,14,15,17,19 suggesting that the flavin coenzyme
of IDI-2 serves a novel function as an acid-base catalyst. Studies
aimed at elucidating additional features of the IDI-2 catalyzed
reaction are currently in progress.
Additional support for a direct catalytic role for the FMN coenzyme
in the IDI-2-catalyzed reaction can be gleaned from previous studies.
For example, the reduced FMN can be covalently modified at the N5
atom by various electrophilic IPP and DMAPP analogues,10,14,15
suggesting that the N5 atom of the reduced flavin carries sufficient
electron density to serve as a nucleophile with these electrophilic
inhibitors. In addition, the nearly complete inactivity of IDI-2
reconstituted with reduced 5-deazaFMN16,17 suggests that the N5 atom
of reduced FMN is critical for isomerization chemistry and that the
role of the flavin in catalysis likely extends beyond the simple
electrostatic stabilization of a substrate carbocation intermediate
(Scheme 1, path B). Interestingly, the N5 atom of the reduced FMN
appears to be positioned appropriately for a role as an acid-base
catalyst, as it is only 3.2 Å away from C2 of IPP (1) in the crystal
structure.11 The substrate in this crystal structure is bound in a mode
that orients the pro-R C2 proton of IPP toward the N5 atom of FMN,
consistent with the known stereospecificity of C2-H abstraction.18
Furthermore, a hydrogen bond formed between the carbonyl group of a
conserved M66 residue and the N5 atom of FMN (separated by 2.9 Å)11
may help to modulate the pKa of the N5 atom for an acid/base catalyst
function during turnover. In conjunction with the linear free energy
relationships (LFER) studies presented here, these results are all consistent
with an IDI-2 chemical mechanism involving general base catalysis by
the N5 atom of the reduced flavin coenzyme (Scheme 1, paths C-E).
Acknowledgment. The authors gratefully acknowledge Steve
Sorey for his assistance in collecting the KIE data, Drs. Steven O.
Mansoorabadi and Mark Ruszczycky for their helpful discussions, and
the NIH (GM40541) and Welch Foundation (F-1511) for funding.
Supporting Information Available: Chemoenzymatic synthesis and
absorbance spectra of IDI-2 bound flavin analogues; kinetic data for
the LFER, KIE, and Q154A mutant enzyme studies; and descriptions
of all assays and data fitting. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Ramos-Valdiva, A. C.; van der Heijden, R.; Verpoorte, R. Nat. Prod. Rep.
1997, 14, 591–603.
(2) Kaneda, K.; Kuzuyama, T.; Takagi, M.; Hayakawa, Y.; Seto, H. Proc. Natl.
Acad. Sci. U.S.A. 2001, 98, 932–937.
(3) Reardon, J. E.; Abeles, R. H. Biochemistry 1986, 25, 5609–5616.
(4) Muehlbacher, M.; Poulter, C. D. Biochemistry 1988, 27, 7315–7328.
(5) Wouters, J.; Oudjama, Y.; Berkley, S. J.; Tricot, C.; Stalon, V.; Droogmans,
L.; Poulter, C. D. J. Biol. Chem. 2003, 278, 11903–11908.
(6) Lee, S.; Poulter, C. D. J. Am. Chem. Soc. 2006, 128, 11545–11550.
(7) Rothman, S.; Helm, T. R.; Poulter, C. D. Biochemistry 2007, 46, 5437–5445.
(8) Thibodeaux, C. J.; Mansoorabadi, S. O.; Kittleman, W.; Chang, W.-c.; Liu,
H.-w. Biochemistry 2008, 47, 2547–2558.
D
Figure 3. Flavin substituent effects on kcat. Reactions with IPP and (R)-
[2-2H]-IPP (O and b circles, respectively, in panel B) were carried out
under pseudo-first-order conditions as described in the Supporting Informa-
tion. (A) Calculated Dkcat from the data in panel (B). (B) Hammett plots: F
) -2.5(4) and -2.7(5) for IPP and (R)-[2-2H]-IPP, respectively.
Although the N5 atom of FMN is a reasonable candidate for one
of the acid/base groups required for turnover, the proton source for
the isomerization of IPP eludes us. If the mechanism of path C in
Scheme 1 is operative, then the proton required to complete the
isomerization of IPP may be supplied by a proton relay composed of
H10-E222-Q154 (S. aureus numbering) that connects the enzyme
surface to the active site. Interestingly, the absolutely conserved Q154
residue is in close proximity to the bound IPP substrate in the crystal
structure and is oriented on the opposite face of the IPP molecule
relative to the FMN N5 atom.11 As such, it appears to be positioned
appropriately for an anti 1,3-proton addition-elimination reaction.
Kinetic studies of the Q154N mutant enzyme indicate a substantial
reduction in kcat (170-fold) relative to the wild type enzyme with the
Km for IPP (1) being relatively unchanged, suggesting that this residue
may indeed play a crucial role in acid-base catalysis (Figure S5).
Alternatively, the flavin could carry out both the protonation and
deprotonation steps, precluding the need for amino acid derived acid/
(9) Mansoorabadi, S. O.; Thibodeaux, C. J.; Liu, H.-w. J. Org. Chem. 2007,
72, 6329–6342.
(10) Johnston, J. B.; Walker, J. R.; Rothman, S. C.; Poulter, C. D. J. Am. Chem.
Soc. 2007, 129, 7740–7741.
(11) Unno, H.; Yamashita, S.; Ikeda, Y.; Sekiguchi, S.; Yoshida, N.; Yoshimura,
T.; Kusunoki, M.; Nakayama, T.; Nishino, T.; Hemmi, H. J. Biol. Chem.
2009, 284, 9160–9167.
(12) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165–195.
(13) Dissociation constants of organic bases in aqueous solution; Perrin, D. D.,
Ed.; Butterworths: London, 1965.
(14) Hoshino, T.; Tamegai, H.; Kakinuma, K.; Eguchi, T. Bioorg. Med. Chem.
2006, 14, 6555–6559.
(15) Rothman, S. C.; Johnston, J. B.; Lee, S.; Walker, J. R.; Poulter, C. D.
J. Am. Chem. Soc. 2008, 130, 4906–4913.
(16) Hemmi, H.; Ikeda, Y.; Yamashita, S.; Nakayama, T.; Nishino, T. Biochem.
Biophys. Res. Commun. 2004, 322, 905–910.
(17) Kittleman, W.; Thibodeaux, C. J.; Zhang, H.; Liu, Y.-n.; Liu, H.-w.
Biochemistry 2007, 46, 8401–8413.
(18) Kao, C.-l.; Kittleman, W.; Zhang, H.; Seto, H.; Liu, H.-w. Org. Lett. 2005,
7, 5677–5680.
(19) Sharma, N. K.; Pan, J.-J.; Poulter, C. D. Biochemistry 2010, 49, (DOI:
10.1021/bi100844e).
JA104090M
9
9996 J. AM. CHEM. SOC. VOL. 132, NO. 29, 2010