Decomposition of amino diazeniumdiolates (NONOates): Molecular mechanisms

Article abstract of DOI:10.1016/j.jinorgbio.2014.08.008

Although diazeniumdiolates (X[N(O)NO]^-) are extensively used in biochemical, physiological, and pharmacological studies due to their ability to release NOand/or its congeneric nitroxyl, the mechanisms of these processes remain obscure. In this

Full text of DOI:10.1016/j.jinorgbio.2014.08.008

Journal of Inorganic Biochemistry 141 (2014) 28–35
Contents lists available at ScienceDirect
Journal of Inorganic Biochemistry
journal homepage: www.elsevier.com/locate/jinorgbio
Decomposition of amino diazeniumdiolates (NONOates):
Molecular mechanisms
Nizamuddin Shaikh a, Marat Valiev b, Sergei V. Lymar a,
⁎
a
Chemistry Department, Brookhaven National Laboratory, Upton, NY 11973, United States
b
William R. Wiley Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, P.O. Box 999, Richland, WA 99352, United States
a r t i c l e i n f o
a b s t r a c t
Although diazeniumdiolates (X[N(O)NO]−) are extensively used in biochemical, physiological, and pharmaco-
logical studies due to their ability to release NO and/or its congeneric nitroxyl, the mechanisms of these processes
remain obscure. In this work, we used a combination of spectroscopic, kinetic, and computational techniques to
arrive at a quantitatively consistent molecular mechanism for decomposition of amino diazeniumdiolates (amino
NONOates: R2N[N(O)NO]−, where R = \N(C2H5)2 (1), \N(C3H4NH2)2 (2), or \N(C2H4NH2)2 (3)). Decomposi-
tion of these NONOates is triggered by protonation of their [NN(O)NO]− group with the apparent pKa and decom-
position rate constants of 4.6 and 1 s−1 for 1; 3.5 and 0.083 s−1 for 2; and 3.8 and 0.0033 s−1 for 3. Although
protonation occurs mainly on the O atoms of the functional group, only the minor R2N(H)N(O)NO tautomer
(population ~10−7, for 1) undergoes the N\N heterolytic bond cleavage (kd ~ 107 s−1 for 1) leading to amine
and NO. Decompositions of protonated amino NONOates are strongly temperature-dependent; activation en-
thalpies are 20.4 and 19.4 kcal/mol for 1 and 2, respectively, which includes contributions from both the
tautomerization and bond cleavage. The bond cleavage rates exhibit exceptional sensitivity to the nature of
R substituents which strongly modulate activation entropy. At pH b 2, decompositions of all three NONOates
that have been investigated are subject to additional acid catalysis that occurs through di-protonation of the
[NN(O)NO]− group.
Article history:
Received 9 February 2014
Received in revised form 12 August 2014
Accepted 13 August 2014
Available online 23 August 2014
Keywords:
Diazeniumdiolates
NONOates
NO donor
Mechanism
Tautomerization
Ab initio calculations
© 2014 Elsevier Inc. All rights reserved.
1. Introduction
diazeniumdiolates with X− being an electron and electron pair, respec-
tively, and the ON-[N(O)NO]− anion [8,10,12] should be recognized as a
A growing appreciation of the biological and pharmacological roles of
nitric oxide (NO) and nitroxyl (HNO/NO−) has naturally rekindled
interest in diazeniumdiolates, the compounds of the general structure
X-[N(O)NO]− shown in Scheme 1 [1–6]. Based on their known decompo-
sition pathways, these compounds can be viewed either as adducts of X-
NO with NO− or as complexes of Lewis bases X− with a Lewis acid N2O2,
an N\N dimer of NO. The latter view is somewhat more instructive be-
cause it captures the origin of the N2O2 stabilization in diazeniumdiolates.
An N\N bond in ONNO is extremely weak, but the addition of an elec-
tron dramatically stabilizes the resulting hyponitrite radical (ONNO−); ef-
fectively, the extra electron “glues” together the two NO fragments
[7–10]. Addition of another electron yields an even more stable
hyponitrite di-anion (ONNO2−) [11]. The ONNO stabilization in all other
diazeniumdiolates is achieved in the same manner by adding an electron
pair donor X−. As has been extensively reviewed [2], most frequently X−
is a carbanion or amide anion but diazeniumdiolates with X− = O2− (as
in Angeli's salt) or SO23− (as in Pelouze's salt) are also long-known. From
this perspective, both ONNO− and ONNO2− should be accepted as
bona fide nitroxy diazeniumdiolate with X− = NO−.
Decomposition of diazeniumdiolates can be triggered either by
photoexcitation [13,14] or by the addition of an electron pair acceptor
that counters the stabilization exerted by X−; commonly, a proton can
play this role. Depending on the nature of X−, this destabilization by
H+ can result in decomposition to various products as shown in
Scheme 1. However, the molecular mechanisms of these reactions
remain obscure; in most cases a rate determining step has not been
unambiguously established and quantitatively characterized. This lack
of mechanistic understanding hampers our ability to manipulate
diazeniumdiolates toward the desired properties.
Of all diazeniumdiolates, the amino variety with X− = R1R2N− is the
most extensively used in biochemical, physiological, and pharmacolog-
ical studies due to their ability to release NO in neutral media
R2NNðOÞNO− þ Hþ → R2NH þ 2NO:
ð1Þ
These compounds are often called amino NONOates, and we will use
this notation throughout this paper. Despite a wealth of phenomenolo-
gy consisting primarily of the decomposition rates and NO yields under
the physiological medium conditions and of comparative studies for a
⁎
Corresponding author. Tel.: +1 631 344 4333; fax: +1 631 344 5815.
E-mail address: Lymar@bnl.gov (S.V. Lymar).
http://dx.doi.org/10.1016/j.jinorgbio.2014.08.008
0162-0134/© 2014 Elsevier Inc. All rights reserved.

N. Shaikh et al. / Journal of Inorganic Biochemistry 141 (2014) 28–35
29
Cary 500 were adequate, but the faster reactions were measured using
SX20 stopped flow spectrometer (Applied Photophysics).
2.3. Computational techniques
Calculations were performed with the NWChem computational
chemistry package [27] and consisted of two stages. In the first stage,
quantum mechanical/molecular mechanics (QM/MM) approach was
used for structural characterization of the NONOate species in aqueous
environment. The QM/MM approach implemented in NWChem follows
standard electrostatic embedding scheme with no modifications of
the classical force parameters, and its performance has been validated
for various aqueous systems [28–32]. In the second stage, hydration
free energies for the configurations obtained in the first stage were
evaluated with the self-consistent reaction field theory of Klamt and
Schüürmann (COSMO) [33].
Scheme 1. General structure and known decomposition pathways of NONOates.
large number of amino NONOates [1,2,15–20], we are aware of only a
few credible attempts at understanding the underlying molecular
mechanisms [21–26]. Unfortunately, these attempts generally suffer
from the lack of complete pH dependencies for rates and from the
data interpretation inconsistencies that we will briefly address in the
Discussion section. In this work, we present a mechanistic investigation
of the amino NONOates shown in Scheme 2. Through a combination of
spectroscopic, kinetic, and computational techniques, we arrive at a
quantitatively consistent molecular mechanism for reaction 1 and dis-
cuss the main factors determining its rate.
The system for QM/MM calculations consisted of a NONOate mole-
cule placed in 30 Å cubic box of classical water molecules. The QM re-
gion containing only NONOate was treated at the density functional
level of theory (DFT/B3LYP [34]) with 6-31+G* basis set [35]. The rest
of the system (in this case, all the water molecules) was treated classi-
cally using SPC/E [36] water model and the periodic boundary condi-
tions based on the cutoff radius of 10 Å. Within this general scheme,
the following procedure for structural QM/MM calculations was
employed. First, an initial guess was generated either by hydrating a
gas-phase solute structure using NWChem prepare module or by mod-
ifications of previously optimized aqueous structure; e.g., the initial
guess for decomposition products was obtained by stretching the corre-
sponding N\N bond in the optimized aqueous reactant. With the initial
structure constructed, the QM/MM optimization of the entire system
was performed [31]. Then, with the NONOate structure fixed, the
surrounding solvent was equilibrated for 100 ps at 25 °C using classical
molecular dynamics calculations. This step was followed by another
round of QM/MM optimization of the entire system.
To determine the NONOate decomposition pathway, we used the
optimized structures of both hydrated reactant and product in conjunc-
tion with the previously developed QM/MM implementation [31] of the
nudged elastic band (NEB) method [37]. The transition state was ob-
tained as the highest energy structure on the 10-bead NEB path, and
the QM/MM calculations were performed to verify the presence of the
single imaginary vibrational mode in this structure, which is a transition
state indicator.
2. Materials and methods
2.1. Materials
Analytical grade chemicals and Milli-Q purified (ASTM type I) water
were used throughout. D2O (99.8%) was obtained from ICN Biomedicals.
Anionic diethylamine NONOate (1, diethylammonium salt) and zwit-
terionic dipropylenetriamine NONOate (2) and diethylenetriamine
NONOate (3) from Cayman Chemical were used as received. Their
stock solutions were prepared in 1–10 mM NaOH and stored on ice.
Concentrations of these solutions were determined spectrophotometri-
cally using molecular absorptivities ε248 = 8.89 × 103 for 1 [20], ε252
7.86 × 103 for 2, and ε252 = 7.64 × 103 M−1 cm−1 for 3 [16].
=
2.2. Experimental procedures
The solute aqueous free energy was obtained as G = E + Hc − TS +
ΔGh + 1.84 kcal/mol, where E is the electronic contribution to internal
energy computed in the gas phase at CCSD(T) level of theory [38] and
maug-cc-pVTZ basis set [39], Hc is the usual thermal enthalpy correc-
tion, TS accounts for the solute entropy contribution, ΔGh contains the
free energy of hydration from COSMO calculations, and the last term ac-
counts for the standard state change from 1 bar pressure to 1 M concen-
tration. Notably, both Hc and S were computed for the solute structure
and vibrational frequencies produced by the QM/MM model. For the
COSMO calculations of ΔGh, we used 78.0 for water dielectric constant
and solvation cavity defined by a set of intersecting atomic spheres
with radii suggested by Stefanovich and Truong (N 2.126, H 1.172,
O 1.576, and C 1.635 Å) [40]. The QM theory in COSMO calculations
was identical to that used for QM/MM calculations, namely DFT/B3LYP
with 6-31+G* basis set.
As a test of this method for the free energy computations, we per-
formed a tautomer analysis for nitramide that resembles NONOates in
terms of its structure and chemical composition, and for which the
free energy gap of 8.2 kcal/mol between its major nitro-tautomer
(H2N-NO2) and the higher energy aci-tautomer (HN_N(O)OH) had
been reliably estimated from experiment [41]. Our computations gave
8.3 kcal/mol (Table S1 and Fig. S3), which is nearly exact match with
the experimental value. Even if somewhat serendipitous, this match
Spectrophotometric pH titrations were carried out using a Cary 500
spectrophotometer (Varian). To avoid concentration changes due to the
NONOate decompositions during spectral scans that become significant
at low pH, a syringe-driven mixing system was employed in which an
alkaline NONOate solution was mixed with a desired buffer and the
spectra were recorded under continuous flow conditions. The buffer
used were: 0.01 M borate for pH 10–8, 0.1 M phosphate for pH 8–6,
0.01 M acetate for pH 5–3.7, 0.1 M phosphate for pH 3.5–2.6, and
HClO4 for pH 2–0.
The NONOate decomposition kinetics were measured under thermal
stabilization at 25 °C (except for the activation measurements) and
strictly anaerobic conditions in a head-space-free flow optical cell. For
the sufficiently slow reactions the above-described flow system and
O
O
O
O
O
O
N N
N N
N N
N
N
N
H3N
H3N
NH2
H2N
Scheme 2. Amino NONOates investigated in this work.

30
N. Shaikh et al. / Journal of Inorganic Biochemistry 141 (2014) 28–35
does give us a measure of confidence in the employed computational
technique.
A
0.6
0.3
0
3. Results
3.1. Acid–base equilibria
Complete and accurate quantitative description of protonation
states is essential for understanding the pH-dependent decomposition
of NONOates. Chemical intuition suggests that the terminal amines in
2 and 3 are the most basic sites in these NONOates. Because the two
amino groups are well-separated, we should expect that each amine is
protonated independently, and, for statistical reasons, the first apparent
pKa1 should be 0.3 units lower than the second pKa2. These expectations
are wholly confirmed by the titration data in Fig. 1. Upon dissolution of
solid 2 and 3 that are zwitterions with half-protonated terminal amines,
a buffered solution is obtained with pH = pKa2 = 9.2, which consumes
one acid equivalent on titration to pH 7 with the mid-point at pH =
pKa1 = 8.9. Thus at pH 7, 2 and 3 each carry a unit positive charge,
whereas 1 carries a unit negative charge.
In addition to charging, protonation of terminal amines slightly
perturbs the [N(O)NO]− chromophoric group as evidenced by a
~2 nm red shift in the spectrum of 3 on lowering pH from 12 to 7
(Fig. 1, inset). A very similar spectral change was observed for 2 but
not for 1 that lacks terminal amines. However, apart from this indirect
influence, protonation of terminal amines is not involved in the decom-
position of 2 and 3, as discussed below.
220
230
240
250
260
270
wavelength, nm
B
8
7
6
4
3
1
Much larger spectral changes are observed for all three NONOates at
pH b 7. These pH-dependent changes exhibit all the hallmarks of pro-
tonation of the chromophoric [N(O)NO]− group; that is, a significant
blue shift with decrease of absorption maximum upon protonation
and the isosbestic point as shown for 1 in Fig. 2A. The fits of the theoret-
ical titration equation
3
4
5
6
7
pH
Fig. 2. Panel A: changes of 1 absorption spectra upon lowering pH from 7.0 to 5.0, 4.8, 4.0,
3.3. Isosbestic point is observed at 233 nm. Panel B: pH dependencies of the apparent
molar absorptivities at 248 nm (red circles) and 226 nm (blue diamonds); the lines
show fits to Eq. (2) with pKmp = 4.6 (red) and 4.7 (blue). (For interpretation of the refer-
ences to color in this figure legend, the reader is referred to the web version of this article.).
ꢀ
ꢁ
εaKmp þ εmp Hþ
εapp
¼
ð2Þ
Kmp þ ½Hþꢀ
to the titration curves yielded pKmp = 4.6
0.1 (Fig. 2B). Here, εa
and εmp are the molar absorptivities of the anion and the conjugate
mono-protonated species and Kmp is the attendant acid dissociation
constant. Similar spectral and titration curves were observed for 2 and
3 (Figs. S1 and S2), and their corresponding pKa were determined as
3.8 0.1 and 3.5 0.1.
Because the spectrophotometric titration gives no information about
the protonation site, we have resorted to computational methods for
ranking the five possible tautomers of protonated 1 in terms of their
relative energies in aqueous environment. Fig. 3 shows the computed
optimized structures for 1 and the tautomers of 1-H, and Table 1 sum-
marizes their geometry and energetics. (See, Table S1 for the more de-
tails.) Recalling that the nominal bond lengths for N\N, N_N, N\O,
and N_O are 1.45, 1.25, 1.40, and 1.21 Å, respectively [43] we observe
that the O1\N2, N2\N3, and N3\O4 bonds in the anion can all be
best characterized as partially double, which corresponds to the Lewis
structures that we use in Scheme 2 and in Fig. 3. Notably, the N2\N3
double bond character gives rise to the E and Z isomers, and the calcu-
lations reveal that the lowest energy structures are all Z isomers. How-
ever for the 1-H(5) species, we have found that the E isomer is only
slightly (0.7 kcal/mol) above the corresponding Z isomer (Fig. S4), and
this energy difference is likely to be within the computational uncer-
tainty. The most consequential structural features of the mono-
protonated species in Table 1 are that their N3\N5 bond is essentially
single for all tautomers and, with the exception of very high energy
1-H(3), the N2\N3 bond remains partially double. Lewis structures
that best correspond to the computed ones are included in Fig. 3. As
can be seen from the bottom lines in Table 1, comparison between our
computed structure and the crystallographic data is very favorable,
which serves to validate our structural computations.
6
pH 12
4
pH 7
2
1.5
240
250
260
270
wavelength, nm
1.0
0.5
0
4
5
6
7
8
9
pH
Fig. 1. pH-metric titrations of 2 (green up triangles) and 3 (blue down triangles). Solid
NONOates were dissolved in 5 mL of pure water to give 5 mM solutions of pH 9.2 that
were titrated with 25 mM HCl. On the acidic side, these measurements were limited by
the rates of NONOate decompositions. Inset: spectral shift for 3 upon changing pH from
12 to 7. (For interpretation of the references to color in this figure legend, the reader is re-
ferred to the web version of this article.).
As expected and in agreement with previous computations [21,23],
protonation at oxygen atoms is preferential with the terminal O1

N. Shaikh et al. / Journal of Inorganic Biochemistry 141 (2014) 28–35
31
Fig. 3. Optimized structures of the hydrated 1 anion and its mono-protonated tautomers showing surrounding water molecules. Also shown are the corresponding Lewis structures based
on bond lengths in Table 1. Top to bottom and left to right: 1, 1-H(1), 1-H(2), 1-H(3), 1-H(4), and 1-H(5).
being the most basic site followed by the O4 site. Among the nitrogen
sites, protonation of N5 is most advantageous leading to overall third
lowest energy tautomer 1-H(5). The apparent Kmp derived from the
data in Fig. 2 relates to the Ka values of individual tautomers through
1-H(5) tautomer with pKa = −2.0 for N-cyanodiethylammonium [44]
suggests that, despite its overall negative charge, the \[N(O)NO]−
group is as strongly electron-withdrawing as the \CN group; the sub-
stitution of diethylammonium (pKa = 10.8 [45]) with either of these
groups increases its acidity by ~13 units. The \[N(O)NO]− group is, in
fact, the hyponitrite radical [7,9], and its considerable electron-
withdrawing strength is consistent with the high reduction potential
E0 = 0.96 V vs. NHE of this radical [7], which exerts large polarization
(inductive effect) on the N3\N5 σ-bond in 1-H(5).
i¼5
i¼5
X
X
1
1
1
¼
¼
Fi
ð3Þ
Kmp
Kað1‐HðiÞÞ Kað1‐Hð1ÞÞ
i¼1
i¼1
where Fi = Ka(1-H(1)) / Ka(1-H(i)) = exp(−ΔGi/RT) are the fractions
of individual tautomers. From this equation and ΔGi in Table 1, we esti-
mate: pKa(1-H(2)) ≈ −5.9, F2 ≈ 3 × 10−11; pKa(1-H(3)) ≈ −15,
F3 ≈ 6 × 10−20; pKa(1-H(4)) ≈ 1.7, F4 ≈ 10−3; and pKa(1-H(5)) ≈
−2.8, F5 ≈ 6 × 10−8. A comparison of pKa ≈ −2.8 for the
3.2. Decomposition kinetics
These are conveniently observed through the decay of the strong UV
absorption bands of the NONOates. In qualitative agreement with the
Table 1
Computed properties of the lowest energy structures for hydrated 1 and tautomers of 1-H shown in Fig. 3.
Species
Relative ΔGi,
Bond length, Å
Bond angles, degrees
kcal/mol
O1\N2
N2\N3
N3\O4
N3\N5
O1\N2\N3
N2\N3\O4
N5\N3\O4
1-H(1)
1-H(2)
1-H(3)
1-H(4)
1-H(5)
1
0.00
14.3
26.3
4.0
1.36
1.26
1.15
1.24
1.28
1.28
1.29
1.32
1.28
1.30
1.85
1.31
1.30
1.30
1.29
1.28
1.27
1.25
1.28
1.40
1.28
1.30
1.31
1.28
1.43
1.44
1.47
1.38
1.51
1.44
1.42
1.43
111
124
105
114
115
115
113
112
125
123
111
121
133
128
126
126
121
123
119
118
116
119
121
120
10.0
1, Na+
1, Et2NH+2
The bottom two lines show crystallographic data for the salts of 1 [42].

32
N. Shaikh et al. / Journal of Inorganic Biochemistry 141 (2014) 28–35
equation faithfully describes the entire pH dependencies in Fig. 4 for
both 1 and 2 and the low pH rate increase for 3 (the higher pH kinetics
for 3 were deemed rather uninformative and not studied for this very
stable NONOate). It is clear from the data that in all cases kdp ≫ kmp
and that pKdp is negative making the last denominator term in Eq. (4)
small even at pH ~0. Therefore, only the kdp/Kdp ratios, kmp, and pKm
could be determined from the data, and these are shown in Table 2.
Temperature dependencies of the decay rates were investigated for 1
and 2 at pH 2; that is, when the functional groups of these NONOates are
completely mono-protonated and yet the contribution to rate from the
kdp pathway is still negligible as it is obvious in Fig. 4. Under these con-
ditions, kobs is essentially identical with kmp (Eq. (4)). As can be seen
from Fig. 5 and Table 2, the decomposition of mono-protonated species
occurs with very significant activation energy that is practically identical
for 1 and 2 and the lower rate for 2 is entirely due to the much lower pre-
exponential factor. Our Ea value for 1 is in general agreement with a
crude estimate of 24 kcal/mol obtained by Ramamurthi and Lewis [17].
The hydrogen kinetic isotope effect (KIE) was measured for 2 by run-
ning its decomposition in acidic H2O and D2O under the otherwise iden-
tical conditions. The values for KIE = kmp(H2O) / kmp(D2O) = 0.82 and
0.85 were found in 10 and 20 mM HClO4, respectively. These numbers
indicate that the isotope effect is secondary [46] and inverse (KIE b 1);
that is, a hydrogen ion is not directly involved in the decomposition of
mono-protonated species and the measured reaction is likely to be not
elementary.
1.2
0.8
0.4
0
×5
×20
0
2
4
pH
6
8
Fig. 4. pH dependencies of the observed rate constants for 1 (blue circles), 2 (green
squares) and 3 (red diamonds). For 2 and 3 the kobs data are multiplied by 5 and 20, re-
spectively. The curves show fits to Eq. (4) with the parameters from Table 2. (For interpre-
tation of the references to color in this figure legend, the reader is referred to the web
version of this article.).
previous reports [17,22], the decays show a very good exponential be-
havior, and a sharp increase in the decay rate as the N(O)NO− groups
of the NONOates becomes protonated (Fig. 4). However, the rate
constants do not completely level off, and we observe a previously un-
reported further steep increase in rates at lower pH. A control experi-
ment has shown that the rate enhancement in the high acid
concentration region (0.001–1 M HClO4) is not due to the ionic strength
effect. The addition of 1 M NaClO4 at pH 1.0 (0.1 M HClO4) resulted only
in a 6% rate increase for 1; thus, high medium acidity opens an addition-
al decay pathway. This feature is consistent with the onset of the second
protonation that further decreases the NONOate stability.
4. Discussion
Apart from the pH region above 6 where the amino end-groups of 2
and 3 undergo the acid–base equilibration (Fig. 1), the protonation and
attendant decomposition patterns for all three NONOates are similar
(Scheme 3). It is clear from Table 2 that summarizes our kinetic data
that there is a good correspondence between spectrally and kinetically
determined pKmp values, as it should be if Scheme 3 is valid. This is in
variance with the previous report by Davies and co-workers who, as
shown in Table 2, found spectrally determined pKa values consistently
much lower than pKa obtained from kinetics. To explain this observa-
tion, they suggested two protonation steps at the [NN(O)NO]− group
between pH 8 and 2, both leading to rate increases but only that occur-
ring at lower pH reflected in the UV spectrum. Our data in Figs. 1, 2,
and 4 clearly show that this is not the case: in this pH range, there is
only one protonation step affecting both kinetics and spectrum.
We will now consider the molecular mechanism of decomposition
of amino NONOates, focusing primarily on the mono-protonated spe-
cies and compound 1 as a prototypical case. First, we note that the pla-
teaus for kobs around pH 2 in Fig. 4 and the very small inverse hydrogen
kinetic isotope effect that we have observed exclude from consideration
any mechanism with proton participation in the rate-determining step
including the tautomerization reaction and the mechanism proposed
in the theoretical work by Hall and co-workers [21]. Thus, the rate-
determining step must be the monomolecular decomposition of one
(or, less likely, more than one) of the 1-H tautomers shown in Fig. 3.
As for the equilibration between tautomers, its mechanism undoubtedly
involves deprotonation–protonation with intermediacy of the 1 anion
(see Scheme S1). Because the distribution of tautomer populations at
equilibrium is pH-independent, the experimentally observable values
kmp and Kmp (defined by Eqs. (3) and 4) are related to the monomolec-
ular decomposition rate constants, kd(i), of tautomers and their acidi-
ties, Ka(i), as
The most generic mechanism describing this behavior is shown in
Scheme 3 which corresponds to the following expression for the ob-
served rate constant
h
i
ꢀ
ꢁ
kdp
Kdp
1
2
kmp Hþ
þ
Hþ
kobs
¼
ð4Þ
h
i
2
Kmp þ ½Hþꢀ þ
Hþ
Kdp
where Kdp and Kmp are the acid dissociation constants and kdp and kmp
are the rate constants for decomposition of a NONOate with di-
protonated and mono-protonated functional groups, respectively. This
i¼5
X
Kmp
kmp
¼
kdðiÞ:
ð5Þ
KaðiÞ
i¼1
Recalling that kmp = 1 s−1 and pKmp = 4.6, we estimate that tauto-
mers with pKa more negative than −4.5 cannot appreciably contribute
Scheme 3. Proton-assisted decomposition of amino NONOates.

N. Shaikh et al. / Journal of Inorganic Biochemistry 141 (2014) 28–35
33
Table 2
Kinetic parameters for decompositions of amino NONOates.a
b s−1
t
½
,
c min at pH 7
kdp/Kdp
,
b M−1 s−1
A, 1013 s−1
ΔS#, cal/(mol K)
Ea (ΔH#), kcal/mol
pKa from ref. [22]
b
Species
pKmp
kmp
,
1
2
3
4.5 (4.6d)
3.6 (3.5d)
3.7 (3.8d)
1.0
0.083
0.0033
3.7
350
7000
0.57
0.13
0.019
330
4.0
nd
14.5
5.7
nd
21 (20.4)
20 (19.4)
nd
5.0 (4.5d)
nd
5.9 (3.1d)
a
Parameters are defined in Eq. (4).
At 25 °C.
Half-life estimated using Eq. (4) and parameters in the table.
Spectroscopically determined pKa.
b
c
d
to the decomposition with the observed rate (see Scheme S1 and its dis-
cussion), and we can safely exclude decomposition pathways through
the highest energy 1-H(2) and 1-H(3) tautomers.
and energetics of the transition states and the complex). The lowest en-
ergy decomposition pathway includes three steps: (a) equilibrium
tautomerization between 1-H(1) and 1-H(5) occurring through inter-
mediacy of their common conjugate base 1, (b) N3\N5 bond scission
in 1-H(5), and (c) decomposition of the nascent Et2NH⋯N2O2 complex
from step b. Applying the activated complex formalism to the observed
kmp temperature dependence, we obtain
From the bond lengths in Table 1 and the related discussion of their
structural implications we believe that the lowest energy pathways to
the amine and NO final products are through the rate determining
step that involves heterolytic scission of the N3\N5 single bond in
one of the three remaining tautomers as shown in Scheme 4. While
for 1-H(5) such a pathway gives relatively low-energy Et2NH and
N2O2 as the nascent products (rds 1), for 1-H(4) and 1-H(1) the emerg-
ing products are high-energy Et2N− and HN2O2+ (rds 2). With reference
to tautomerization energies in Table 1 and the notations in Scheme 4, it
is easy to see that the minimal free energy difference (in kcal/mol) be-
tween rds 2 and rds 1 is Δ2G − Δ2G = 1.36(pKa1 − pKa2) + Δt2G . Taking
into account that pKa1 = pKa(R2NH) ≈ 35 [46], pKa2 = pKa(HN2O+2 ) is
certainly negative, and Δt2G ≈ 4 kcal/mol, we estimate that the rds 1
pathway is at least by 52 kcal/mol more favorable than any of the rds
2 pathways. Because this number by far exceeds the activation energy
of 21 kcal/mol that we have observed, there is little, if any, doubt that
the amino NONOates decompose through the N5-protonated tautomer,
and Eq. (5) reduces to
!
!
#
#
kT
h
ΔGobs
kT
h
Δ
tauG þ ΔdG
kmp
¼
exp
−
¼
exp
−
ð7Þ
RT
RT
where ΔG#obs is the experimentally observed activation free energy and
the meaning of ΔtauG and ΔdG# is evident from Fig. 6. From data in
Table 2, we obtain ΔG#
= 16.1 kcal/mol; this number is in good
obs
agreement with the computed value for ΔtauG + ΔdG# = 17.8 kcal/mol.
Previously, Houk and co-workers [23] explored a mechanism quali-
tatively analogous to the one shown in Scheme 4 and derived kd(5) =
7.2 × 1011 s−1 and pKa(1-H(5)) = −6.9 by analyzing an incomplete
pH dependence of kobs available only in the pH 8–5.2 range [22] (that
is, without reaching the low pH “plateau”), and supported this very neg-
ative pKa by theoretical energy computations. Thus, 1-H(5) was de-
scribed as an extremely strong acid with virtually no stability toward
losing N2O2. In contrast, kd(5) ≈ 2 × 107 s−1 and pKa(1-H(5)) ≈ −2.8
obtained in the present work suggest that 1-H(5) is a much weaker
acid, whose decomposition with the N2O2 release is significantly
activated and thus relatively slow. This large disagreement merits
two brief comments. As has been mentioned above, it is impossible to
obtain both kd(5) and pKa(1-H(5)) even from a complete kobs vs pH
dependence; even the kd(5)/Ka(1-H(5)) ratio that can be obtained
from Eq. (6) is subject to a large uncertainty without the kinetic data
below pH 5, such as those in Fig. 4, that include a rate plateau. Second,
the values of kd(5) and pKa(1-H(5)) suggested by Houk and co-
workers are incompatible with the key premises behind Scheme 4,
which are equilibrium between tautomers and the N3\N5 bond het-
erolysis being the rate determining step, because it is inconceivable
Kmp
kmp ¼ kdð5Þ
ð6Þ
Kað1‐Hð5ÞÞ
so that kd(5) and Ka(1-H(5)) cannot be separately obtained from the ki-
netic data alone. However, from Eq. (6) and the data in Table 2, we de-
rive kd(5)/Ka(1-H(5)) = 4 × 104 M−1 s−1. Invoking our computational
data for pKa(1-H(5)) ≈ −2.8, we estimate kd(5) ≈ 2 × 107 s−1
.
We have also investigated the kd(5) decomposition pathway com-
putationally, and the result is shown in Fig. 6. We find that N3\N5
bond scission in 1-H(5) requires significant activation and results in a
metastable Et2NH⋯N2O2 complex (see Fig. S5 and Table S3 for structures
1
1
0
2
-1
3.0
3.2
3.4
3.6
1000/T, K-1
Fig. 5. Temperature dependencies in the Arrhenius coordinates of the observed decompo-
sition rate constants for 1 (squares) and 2 (circles) at pH 2 (HClO4). The activation energy
(Ea) and pre-exponential factor (A) derived from the linear fits are shown in Table 2.
Scheme 4. Free energy relationship between tautomerization and decomposition of 1-H.

34
N. Shaikh et al. / Journal of Inorganic Biochemistry 141 (2014) 28–35
Overall, the experimental and computational results and their inter-
Transition state
pretation presented in this work suggest that initiation by protonation
and involvement of several protonated tautomers are likely to be the
common features in decomposition of NONOates. Therefore, structural
and thermochemical characterization of these tautomers is a major
prerequisite for molecular-level understanding of the decomposition
process. Because this information is difficult, if not impossible, to obtain
from experimental measurements, computational modeling becomes of
significant value. Even in their uncharged, protonated state, NONOates
exhibit significant intramolecular charge separation, which places a
great demand on a computational model in both electronic structure
and solvation treatment. We believe that our approach, based on the
explicit treatment of water for the electronic structure characterization
followed by the hydration energy computation through the continuum
approximation, provides a reasonable practical solution to this problem.
However, as with any model, theoretical predictions are not infallible
and should be crosschecked with reality and the experimental data,
as has been done in present work. While essential, information on
the distribution of tautomers is only part of the picture because
the NONOate decomposition reactions are substantially activated
(Table 2). Thus, calculation of the free energy activation barriers
is another place where theoretical modeling can make a substantial
impact.
15
10
5
dG#
Et2NH...N2O2
1-H(5)
pKa -2.7
-
Et2N(N2O2)
tauG
pKa=4.6
1-H(1)
0
To illustrate these points, let us consider tautomers 1-H(2) and
1-H(5) from Fig. 3 and Table 1. The former can, in principle, decompose
to nitrosamine and nitroxyl
Et2NH+2NO
R2NNðOÞNðHÞO → R2NNO þ HNO
ð8Þ
Fig. 6. Computed energy profile for the proton-initiated decomposition of 1. Solid and
dashed red arrows show the pre-equilibrium and decomposition steps, respectively. The
transition state corresponds to the N3\N5 bond scission.
but this reaction is not a major channel with R = ethyl, apparently be-
cause 1-H(2) is higher in energy then 1-H(5). However, it is conceivable
that replacement of the ethyl by a less electron-donating group could
decrease the basicity of N5 to a greater degree than that of N2, thus
narrowing or even reversing the ~4 kcal/mol energy gap between
1-H(2) and 1-H(5); with a favorable free energy activation barrier,
this change in relative basicity can lead to the predominance of HNO-
producing pathway. The computational approach can certainly help
with fundamental understanding of how overall reaction process can
be influenced by the choice of functional groups in NONOates, and the
effort in this direction is under way [47].
that equilibration of the so strongly acidic 1-H(5) can compete with its
decomposition occurring at ~1012 s−1
.
It can be shown that
tautomerization 1-H(1)
→ 1-H(5) becomes rate limiting if
kd(5) exceeds ~109 s−1 and pKa(1-H(5)) is more negative than about
−4.5 (see Scheme S1 and its discussion); thus the numerical values
for kd(5) close to 1012 s−1 and pKa(1-H(5)) ≈ −7 that were obtained
in [23] invalidate the very mechanism that they were intended to quan-
titatively describe. In contrast, our values for kd(5) and pKa(1-H(5)) are
well within the required limits.
Acknowledgments
With the decomposition mechanism for di-protonated NONOates
that becomes increasingly important at pH b 2 (Fig. 4), the situation is
less clear-cut. There are 10 di-protonated tautomers and their detailed
exploration in aqueous environment is computationally expensive. It
is also doubtful that the second plateaus somewhere at pH b 0 in the
pH-dependencies for kobs in Fig. 4 can be reached and the values of kdp
and Kdp can be determined experimentally. Nevertheless, a reasonable
mechanistic proposition can be made. As with the mono-protonated
species, the decomposing di-protonated tautomer will probably need
a proton on the N5 atom to provide for the lowest energy pathway by
releasing amine, an excellent leaving group. Of the remaining sites,
the O1 will probably remain the most basic after N5 has been protonat-
ed. Thus, we tend to think that the di-protonated species decompose as
shown in Scheme 5. Our calculations have shown that the 1-H2(1,5)+
tautomer does exist in water as a bound structure (Fig. S6), which
makes this mechanism plausible.
We thank Dr. V. Shafirovich (NYU) for valuable discussions. This
research was supported by the U.S. Department of Energy, Office of
Basic Energy Sciences, Division of Chemical Sciences, Geosciences, and
Biosciences. Research at Brookhaven National Laboratory was carried
out under contract DE-AC02-98CH10886. Research at Pacific Northwest
National Laboratory was performed in part using the Molecular Science
Computing Facility in the William R. Wiley Environmental Molecular
Sciences Laboratory.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jinorgbio.2014.08.008. These data include MOL files
and InChiKeys of the most important compounds described in this
article.
References
OH
O
+
[1] L.K. Keefer, R.W. Nims, K.M. Davies, D.A. Wink, Methods Enzymol. 268 (1996)
281–293.
NH
+
N2O2
H
+
N N
[2] J.A. Hrabie, L.K. Keefer, Chem. Rev. 102 (2002) 1135–1154.
[3] J.E. Saavedra, L.K. Keefer, J. Chem. Educ. 79 (2002) 1427–1434.
[4] L.K. Keefer, Annu. Rev. Pharmacol. Toxicol. 43 (2003) 585–607.
[5] M.R. Miller, I.L. Megson, Br. J. Pharmacol. 151 (2007) 305–321.
[6] L.K. Keefer, ACS Chem. Biol. 6 (2011) 1147–1155.
NH
1-H2(1,5)+
Scheme 5. Proposed structure and decomposition pathway for the 1-H2(1,5)+ tautomer.

N. Shaikh et al. / Journal of Inorganic Biochemistry 141 (2014) 28–35
35
[7] G.A. Poskrebyshev, V. Shafirovich, S.V. Lymar, J. Am. Chem. Soc. 126 (2004) 891–899.
[8] S.V. Lymar, V. Shafirovich, G.A. Poskrebyshev, Inorg. Chem. 44 (2005) 5212–5221.
[9] G.A. Poskrebyshev, V. Shafirovich, S.V. Lymar, J. Phys. Chem. A 112 (2008)
8295–8302.
[27] M. Valiev, E.J. Bylaska, N. Govind, K. Kowalski, T.P. Straatsma, H.J.J. Van Dam, D.
Wang, J. Nieplocha, E. Apra, T.L. Windus, W. de Jong, Comput. Phys. Commun. 181
(2010) 1477–1489.
[28] J. Chen, H.Y. Yin, D.Y. Wang, M. Valiev, Chem. Phys. Lett. 559 (2013) 30–34.
[29] T.T. Wang, H.Y. Yin, D.Y. Wang, M. Valiev, J. Phys. Chem. A 116 (2012) 2371–2376.
[30] H.Y. Yin, D.Y. Wang, M. Valiev, J. Phys. Chem. A 115 (2011) 12047–12052.
[31] M. Valiev, B.C. Garrett, M.K. Tsai, K. Kowalski, S.M. Kathmann, G.K. Schenter, M.
Dupuis, J. Chem. Phys. 127 (2007) 051102.
[10] M. Valiev, S.V. Lymar, J. Phys. Chem. A 115 (2011) 12004–12010.
[11] F.T. Bonner, M.N. Hughes, Comments Inorg. Chem. 7 (1988) 215–234.
[12] V. Shafirovich, S.V. Lymar, Proc. Natl. Acad. Sci. U. S. A. 99 (2002) 7340–7345.
[13] C.M. Pavlos, H. Xu, J.P. Toscano, Free Radic. Biol. Med. 37 (2004) 745–752.
[14] S.V. Lymar, V. Shafirovich, J. Phys. Chem. B 111 (2007) 6861–6867.
[15] C.M. Maragos, D. Morley, D.A. Wink, T.M. Dunams, J.E. Saavedra, A. Hoffman, A.A.
Bove, L. Isaac, J.A. Hrabie, L.K. Keefer, J. Med. Chem. 34 (1991) 3242–3247.
[16] J.A. Hrabie, J.R. Klose, D.A. Wink, L.K. Keefer, J. Org. Chem. 58 (1993) 1472–1476.
[17] A. Ramamurthi, R.S. Lewis, Chem. Res. Toxicol. 10 (1997) 408–413.
[18] K.M. Miranda, T. Katori, C.L.T. de Holding, L. Thomas, L.A. Ridnour, W.J. MeLendon,
S.M. Cologna, A.S. Dutton, H.C. Champion, D. Mancardi, C.G. Tocchetti, J.E.
Saavedra, L.K. Keefer, K.N. Houk, J.M. Fukuto, D.A. Kass, N. Paolocci, D.A. Wink, J.
Med. Chem. 48 (2005) 8220–8228.
[19] S. Griveau, C. Dumezy, P. Goldner, F. Bedioui, Electrochem. Commun. 9 (2007)
2551–2556.
[20] Q. Li, J.R. Lancaster Jr., Nitric Oxide 21 (2009) 69–75.
[21] D.K. Taylor, I. Bytheway, D.H.R. Barton, C.A. Bayse, M.B. Hall, J. Org. Chem. 60 (1995)
435–444.
[22] K.M. Davies, D.A. Wink, J.E. Saavedra, L.K. Keefer, J. Am. Chem. Soc. 123 (2001)
5473–5481.
[32] M. Valiev, E.J. Bylaska, M. Dupuis, P.G. Tratnyek, J. Phys. Chem. A 112 (2008)
2713–2720.
[33] A. Klamt, G. Schuurmann, J. Chem. Soc. Perkin Trans. 2 (1993) (1993) 799–805.
[34] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785–789.
[35] R. Krishnan, J.S. Binkley, R. Seeger, J.A. Pople, J. Chem. Phys. 72 (1980) 650–654.
[36] H.J.C. Berendsen, J.R. Grigera, T.P. Straatsma, J. Phys. Chem. 91 (1987) 6269–6271.
[37] G. Henkelman, B.P. Uberuaga, H. Jonsson, J. Chem. Phys. (2000) 9901–9904.
[38] R.J. Bartlett, M. Musial, Rev. Mod. Phys. 79 (2007) 291–352.
[39] E. Papajak, H.R. Leverentz, J.J. Zheng, D.G. Truhlar, J. Chem. Theory Comput. 5 (2009)
3330.
[40] E.V. Stefanovich, T.N. Truong, Chem. Phys. Lett. 244 (1995) 65–74.
[41] C.H. Arrowsmith, A. Awwal, B.A. Euser, A.J. Kresge, P.P.T. Lau, D.P. Onwood, Y.C. Tang,
E.C. Young, J. Am. Chem. Soc. 113 (1991) 172–179.
[42] L.K. Keefer, J.L. Flippen-Anderson, C. George, A.P. Shanklin, T.A. Dunams, D.
Christodoulou, J.E. Saavedra, E.S. Sagan, D.S. Bohle, Nitric Oxide 5 (2001) 377–394.
[43] J.E. Huheey, Inorganic Chemistry, 3rd ed. Harper & Row, New York, 1983.
[44] G.W. Stevenson, D. Williamson, J. Am. Chem. Soc. 80 (1958) 5943–5947.
[45] W.M. Haynes (Ed.), CRC Handbook of Chemistry and Physics, CRC Press, Boca Raton,
FL, 2013, (Section 5, No 94).
[23] A.S. Dutton, J.M. Fukuto, K.N. Houk, Inorg. Chem. 43 (2004) 1039–1045.
[24] A.S. Dutton, C.P. Suhrada, K.M. Miranda, D.A. Wink, J.M. Fukuto, K.N. Houk, Inorg.
Chem. 45 (2006) 2448–2456.
[25] D. Andrei, D.J. Salmon, S. Donzelli, A. Wahab, J.R. Klose, M.L. Citro, J.E. Saavedra, D.A.
Wink, K.M. Miranda, L.K. Keefer, J. Am. Chem. Soc. 132 (2010) 16526–16532.
[26] D.J. Salmon, C.L.T. de Holding, L. Thomas, K.V. Peterson, G.P. Goodman, J.E. Saayedra,
A. Srinivasan, K.M. Davies, L.K. Keefer, K.M. Miranda, Inorg. Chem. 50 (2011)
3262–3270.
[46] R.P. Bell, The Proton in Chemistry, Cornell University Press, Itheca, New York, 1959.
[47] Y.-N. Wang, J. Collins, R.J. Holland, L.K. Keefer, J. Ivanic, J. Phys, Chem. A 117 (2013)
6671–6677.