Model Systems for Flavoenzyme Activity: Interplay of Hydrogen Bonding and Aromatic Stacking in Cofactor Redox Modulation

Article abstract of DOI:10.1021/ol036279g

(Equation presented) A model system has been developed to study the synergy between aromatic stacking and hydrogen bonding in the binding of a flavin derivative. The results show that the identity of both the hydrogen bonding and π-stacking units strongly determine the overall receptor affinity for flavin in both the oxidized and radical anion forms.

Full text of DOI:10.1021/ol036279g

ORGANIC
LETTERS
2004
Vol. 6, No. 3
385-388
Model Systems for Flavoenzyme
Activity: Interplay of Hydrogen Bonding
and Aromatic Stacking in Cofactor
Redox Modulation
Mark Gray,† Allan J. Goodman,† Joseph B. Carroll,† Kevin Bardon,†
Michael Markey,† Graeme Cooke,‡ and Vincent M. Rotello*
,†
Department of Chemistry, UniVersity of Massachusetts, Amherst, Massachusetts 01003,
and Centre for Biomimetic Design and Synthesis, Department of Chemistry, School of
Engineering & Physical Sciences, Heriot-Watt UniVersity, Edinburgh, UK EH14 4AS
rotello@chem.umass.edu
Received November 21, 2003
ABSTRACT
A model system has been developed to study the synergy between aromatic stacking and hydrogen bonding in the binding of a flavin
derivative. The results show that the identity of both the hydrogen bonding and π-stacking units strongly determine the overall receptor
affinity for flavin in both the oxidized and radical anion forms.
Flavoenzymes are a class of biological catalysts performing
a diverse range of functions in living systems. These func-
tions range from the oxidation of amines and thiols to imines
and disulfides, facilitation of electron-transfer reactions with
Fe-heme units and FeS-clusters, neurotransmitter regulation,
and xenobiotic detoxification.1 At the heart of a flavoenzyme
is the cofactor flavin, usually present as flavin mononucleo-
tide (FMN) or flavin adenine dinucloeotide (FAD), nonco-
valently bound into the active site of the protein through an
array of specific interactions. It is the identity of these
interactions that determines the redox properties of the
cofactor and thus the overall utility of the enzyme.
Intact flavoprotein systems have been successfully used
to probe qualitative cause-effect relationships between active
site structure and flavoprotein functionality.2 However, many
standard biochemical techniques such as site-directed mu-
tagenesis cause unavoidable changes to the subtle yet sig-
nificant secondary interactions that exist within the system
as well as affecting the primary interaction of interest. One
way to examine the weight that specific noncovalent interac-
tions have upon modulation of the physical properties of the
cofactor in a quantitative manner is through the development
of model systems.3
Within the reported active sites of the flavodoxin family
a common three-point hydrogen bonding pattern is displayed
between the imide functionality of the cofactor and the
† University of Massachusetts.
‡ Heriot-Watt University.
(1) For reviews of flavoprotein function, see: (a) Walsh, C. Acc. Chem.
Res. 1980, 13, 148. (b) Ghisla, S.; Massey, V. Eur. J. Biochem. 1989, 181,
1. (c) Bruice, T. C. Isr. J. Chem. 1984, 24, 54. (d) Mueller, F. In Topics in
Current Chemistry; Boschke, F. L., Ed.; Springer-Verlag: Berlin, 1983;
Vol. 108, p 71. (e) Ziegler, D. M. Trends Pharm. Sci. 1990, 321.
(2) Ghisla, S.; Massey, V. Biochem. J. 1986, 239, 1.
(3) (a) Niemz, A.; Rotello, V. M. Acc. Chem. Res. 1999, 32, 44. (b)
Kajiki, T.; Moriya, H.; Hoshino, K.; Kuori, T.; Kondo, S.; Nabeshima, T.;
Yano, Y. J. Org. Chem. 1999, 64, 9679.
10.1021/ol036279g CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/08/2004

Figure 2. Binding constants observed for (a) the DAT-based
receptors and (b) the DAP-based receptors with flavin in both its
oxidized and radical anion states. The colored areas depict
approximate zones of influence according to the maximum extent
of π-overlap.
Figure 1. (a) Example of hydrogen-bonded contacts to the imide
portion of flavin, taken from the structure of lipoamide dehydro-
genase.8 (b) Example of π-π interactions between protein and
cofactor, taken from the structure of DesulfoVibrio Vulgaris.9 (c)
Example of a donor atom-π interaction between a flavoprotein
residue and cofactor, taken from a red algae Chrondus-Crispus.10
(d) Adaptable general model of flavoprotein active sites.
More recently, we showed that the reduction potential of
1,8-naphthalimide, an electron-deficient aromatic imide simi-
lar to flavin, was much more strongly stabilized by a simple
DAP receptor than the DAT analogue.6 As a result, while
both receptors showed identical binding constants with the
oxidized imide (160 M-1) the increase in binding constant
upon reduction of the imide was far more pronounced for
the DAP system (46 000 M-1, a ∼300-fold increase) than
DAT (900 M-1, a ∼6-fold increase). Extensive use of
computational methods showed that these effects could be
attributed to a greater polarizability within the DAP-imide
complex than the DAT-imide system. In light of this work,
we have developed an improved class of flavin receptor that
combines the strong electronic modulating ability of a DAP-
based recognition unit with the general utility afforded by
the modular nature of a xanthene scaffold to give the most
powerful synthetic flavoprotein model yet reported. In this
paper, we compare and contrast energetic data associated
with this new DAP-based xanthene system with values
obtained for the older DAT-containing analogues. Three
different receptors were synthesized and evaluated for each
of the DAP and DAT systems, through variations of the
aromatic unit suspended below the flavin π-system via: (i)
an anthracyl unit to typify the strong π-π interactions
observed in DesulfoVibrio Vulgaris, (ii) a thiomethylphenyl
unit to illustrate a general donor atom-π effect as seen in
flavodoxins such as Chrondus-Crispus or flavocytochrome-C
sulfide dehydrogenase,7 and (iii) a simple phenyl unit which
does not extend significantly below the flavin as a control.
The association constants (Ka(ox)) between N(10)-isobu-
tylflavin and all six receptors were quantified through NMR
titration in CDCl3 by following the downfield changes in
apoprotein. The identity of these contacts can result in signifi-
cant stabilization of the radical anionic state (Figure 1a) and
thus acts to tune the reactivity of the flavin. In previous work,
we developed and evaluated a simple model based on the
dipropamide of 2,6-diaminopyridine (DAP) to replicate the
effects of the natural three-point hydrogen-bonding pattern
on the reduction potential of the cofactor.4 It was found that
the binding constant between DAP and the flavin radical
anion was enhanced by a factor 500 times larger than that
observed for flavin in the oxidized form. In addition, the
reduction potential of the complex was observed to take place
at a potential some 155 mV lower than that of the flavin
alone, effectively mimicking the tuning function of a natural
flavoprotein.
In other work, we showed that other noncovalent contacts
commonly observed in the active sites of flavodoxins such
as those observed between tryptophan and tyrosine residues
(Figure 2b) or donor atoms in the form of carbonyl groups
(Figure 2c) or disulfides, and the electron poor aromatic
surfaces of the cofactor could be replicated using a more
complex model. These systems employed a 2,6-diaminotri-
azine (DAT) unit appended onto a xanthene scaffold to help
orient and anchor the cofactor in place through a similar
three-point hydrogen-bonding interaction as used previously
between flavin and DAP. The receptor design is such that
any interaction of interest can be placed in contact with the
electron-deficient aromatic surface of the flavin, with the
general result being that strongly electron-donating substit-
uents cause dramatic increases in binding to the oxidized
flavin but reduced association constants with the same flavin
in the radical anionic form.5
(5) (a) Breinlinger, E. C.; Rotello, V. M. J. Am. Chem. Soc. 1997, 119,
1165. (b) Breinlinger, E. C.; Keenan, C. J.; Rotello, V. M. J. Am. Chem.
Soc. 1998, 120, 8606. (c) Goodman, A. J.; Breinlinger, E. C.; McIntosh, C.
M.; Grimaldi, L. N.; Rotello, V. M. Org. Lett. 2001, 3, 1531.
(6) Gray, M.; Cuello, A. O.; Cooke, G.; Rotello, V. M. J. Am. Chem.
Soc. 2003, 125, 7882.
(4) Niemz, A.; Imbriglio, J.; Rotello, V. M. J. Am. Chem. Soc. 1997,
119, 887.
(7) Chen, Z.-W.; Koh, M.; van Driessche, G.; van Beeumen, J.; Bartsch,
R.; Meyer, T.; Cusanovich, M.; Mathews, F. Science 1994, 266, 430.
386
Org. Lett., Vol. 6, No. 3, 2004

chemical shift that occurred to the N(3) imide proton of the
flavin as the amount of receptor present was increased. In
addition, the half-wave reduction potentials of the free flavin
(E1/2(unbound)) and the corresponding potentials for flavin in
the presence of each receptor (E1/2(bound)) were determined
in dichloromethane through cyclic voltammetry. Taken
together, these three pieces of experimental data were used
to obtain a binding constant for each receptor with flavin in
the radical anionic form through the following relationship.
1/2(bound)-E1/2(unbound))
Ka(red) ) Ka(ox)e[(nF/RT)(E
]
An examination of the measured binding constants be-
tween the fully oxidized flavin and the six receptors (Figure
2) shows that as expected the addition of either a donor
atom-π interaction or an enhanced π-π interaction causes
an increase in binding relative to the receptors containing
an unsubstituted phenyl unit below the flavin. Moreover, the
three receptors containing the DAT-imide recognition unit
displayed binding constants around 1.5 times greater than
the DAP-containing counterparts.
This effect may be explained from an examination of the
1H NMR spectra of the receptors themselves. The primary
difference lies in the chemical shift of the proton associated
with the amide group that attaches the aromatic units to the
xanthene skeleton. This resonance occurs at around 9.5 ppm
for the DAT series but significantly further upfield at 8.7
ppm for the DAP-based receptors. This suggests that in the
DAT series this proton is more strongly hydrogen bound to
the ether oxygen atom of the xanthene giving a more pre-
organized structure and would hence lead to stronger guest
binding from an entropic standpoint (Figure 3).
Figure 4. Cartoon showing the interplay of forces that occurs upon
reduction of flavin once bound to the xanthene based receptors.
between each host and guest. Within the DAT series of
receptors, binding to the unsubstituted phenyl receptor is
more than doubled compared to the value obtained for the
oxidized guest; however, addition of functionality through
either a donor atom or an extended π-surface both result in
lower binding constants for the reduced form in comparison
to the oxidized guest.
Reduction of the flavin to the corresponding radical anion
causes large changes in the strength of the interaction
In contrast, binding flavin with the DAP-based receptors
all show significant enhancement when the cofactor is in
the radical anion state. This behavior may be rationalized
through a dissection of the binding constant data into indi-
vidual energetic contributions from the hydrogen bonding
and π-stacking units.
For flavin in the neutral state, all of the intermolecular
forces promote binding, with the hydrogen-bonding interac-
tions between receptor and substrate reinforced by the
complementarity between the electron poor π-surface of the
cofactor and the aromatic units appended to the xanthene.
Upon reduction of the cofactor, however, the flavin becomes
(8) Mattevi, A. O. G.; Kalk, K. H.; van der Berkel, W. J. H.; Hol, W. G.
J. J. Mol. Biol. 1993, 230, 1200.
(9) (a) Muller, F. In Chemistry and Biochemistry of FlaVoenzymes;
Muller, F., Ed.; CRC: Ann Arbor, MI, 1992; Vol. III, p 557. (b) Muller,
F.; Vervoort, J.; Mierlo, P.; Mayhew, S.; van Berkel, W.; Bacher, A. In
FlaVins and FlaVoproteins (1990); Kurti, B., Ronchi., Zanetti, G., Ed.; De
Gruyer: Berlin, 1991 p 261.
Figure 3. Comparison of the NMR spectra of the p-thiometh-
ylphenyl (a) DAP- and (b) DAT-based receptors show the latter to
be more preorganized due to the presence of an intramolecular
hydrogen bond between the amide proton and the ether oxygen of
the xanthene.
(10) Fukuyama, K.; Matsubara, H. J. Mol. Biol. 1992, 225, 775.
Org. Lett., Vol. 6, No. 3, 2004
387

extremely electron rich; thus, interaction with the π-stacking
recognition units becomes actively disfavored. With the DAT
unit in place to bind the imide moiety of the flavin guest,
little enhancement in the strength of the hydrogen bonding
interactions takes place, even less upon contributions from
the stronger π-stacking units, leading to an overall diminution
in observed binding for the p-thiomethylphenyl and anthracyl
containing DAT receptors. However, as mentioned previ-
ously, the DAP recognition element is highly responsive to
changes in the oxidation state of guests with massive in-
creases in affinity being reported upon guest reduction. Thus,
for the three receptors bearing the DAP system, the binding
enhancement that occurs within the hydrogen-bonding units
strongly outweigh the change from attraction to repulsion
occurring simultaneously between the π-systems.
A second effect that may play a contributory role to the
overall observed effect lies in an extension of the role the
differences in receptor structure. The DAP systems have
appended π-stacking elements that are loosely preorganized
relative to the DAT systems where evidence of hydrogen
bonding between the amide proton and the xanthene oxygen
gives a more ordered class of receptor. The advantages this
difference lends to the DAT receptors for oxidized flavin
become disadvantages once the flavin is reduced. The loosely
preorganized DAP receptors may alleviate much of the
inter-π cloud repulsion through low energy bond rotations,
for the same rotations to occur within the DAT receptors
the amide-oxygen hydrogen bond must be broken giving
significant enthalpic cost to the process (Figure 4).
Table 1. Binding Energies and Energy Differences (kcal/mol)
Phena
Thiob
Anthc
DAT
DAP
∆G(ox)
∆G(red) -4.42
∆G(ox) -3.62
∆G(red) -5.21
0.22
-3.84
-4.97 -5.41
-4.87 -4.67
-4.58 -5.17
-5.42 -5.73
∆G(ox)DAP - ∆G(ox)DAT
∆G(red)DAP - ∆G(red)DAT
0.38
0.24
-0.79
-0.54 -1.06
a Phenyl-substituted receptors. b p-Thiomethylphenyl-substituted re-
ceptors. c Anthracyl-substituted receptors.
As a result, less negative charge occupies the region over-
lapping the electron-rich anthracene unit, thus lessening the
repulsion.
In summary, we have synthesized a flavoenzyme model
bearing a diamidopyridine recognition element and several
aromatic units to replicate π-π and donor atom-π interac-
tions commonly found in nature. In contrast to an earlier
series of models that incorporated a passive diaminotriazine
unit to bind imides, substantial increases in binding occurred
upon electrochemical reduction of the guest. It was shown
that a subtle interplay occurs between the hydrogen bonding
network and the aromatic units which allows for maximum
stabilization of the cofactor in the reduced state. Our further
efforts will involve investigating these systems using high
level computational methods and the application of the results
of this work to the rational design of redox active molecular
devices. The results of our subsequent investigations will
be disclosed in due course.
Finally, an examination of the differences in the free
energy changes associated with flavin binding in the pres-
ence of the two classes of receptor reveal a subtle syn-
ergy between the hydrogen bonding and π-stacking effects
(Table 1). If the two recognition events were electronically
independent then these free energy differences would be
constant. However, our results show that the anthracyl recep-
tors, which display the largest π-stacking effects, show larger
differences in stabilization than the phenyl and p-thiometh-
ylphenyl systems. This effect may be again attributed to
the strong mutual polarization between the DAP group and
the guest, which in this case shifts the electron density
away from the N(5) position toward the imide functionality.
Acknowledgment. This research was supported by the
National Institutes of Health (GM59249) and the National
Science Foundation (CHE-0213354).
Supporting Information Available: Synthesis of DAT
and DAP systems; 1H NMR, IR, HRMS, as well as 1H NMR
titrations and electrochemical data with flavin “guests”. This
material is available free of charge via the Internet at
http://pubs.acs.org.
OL036279G
388
Org. Lett., Vol. 6, No. 3, 2004