Hydrolysis and DFT structural studies of dinuclear Zn(II) and Cu(II) macrocyclic complexes of m -12N3O-dimer and the effect of pH on their promoted HPNP hydrolysis rates

Article abstract of DOI:10.1080/00958972.2016.1157177

The synthesis of the ligand, m-12N₃O-dimer (1,3-bis(1-oxa-4,7,10-triazacyclododecan-7-yl)methyl)benzene, L), and the stability and hydrolysis constants of its dinuclear Zn(II) and Cu(II) complexes are reported, in addition to the effect of pH on HPNP (2-hydroxypropyl-4-nitrophenylphosphate) hydrolysis reaction rates promoted by these complexes. Various structural possibilities of the [Zn₂L] and [Cu₂L] hydrolytic species derived from solution equilibrium modeling are predicted from density functional theory (DFT) studies to correlate with the promoted HPNP hydrolysis reaction rates and to establish the structure-function-reactivity relationship. Upon deprotonation [Zn₂L(OH)]³⁺ tends to form a structure with a closed-form conformation where it is not possible for para-isomers. At pH >8, the formation of the closed-form [Zn₂L(OH)₂]²⁺ and [Zn₂L(μ-OH)(OH)₂]⁺ species led to faster promoted HPNP hydrolysis rates than the [Zn₂L(OH)]³⁺ species. On the other hand, the observed rates of the Cu₂L-promoted HPNP hydrolysis reaction were much slower than those of the [Zn₂L]-promoted ones due to formation of the inactive, di-μ-OH^- bridged closed-form [Cu₂L(μ-OH)₂]²⁺ structure at high pH. The effects of solvent molecules and the use of higher DFT computation levels, i.e., M06 and M06-2X, in conjunction with cc-pVDZ and cc-pVTZ basis sets on the DFT-predicted structures for both [Cu(12N₄)(H₂O)]²⁺ and [Zn(12N₃O)(H₂O)₂]²⁺ complexes were also evaluated and compared with those using the B3LYP/6-31G? method.

Full text of DOI:10.1080/00958972.2016.1157177

Journal of Coordination Chemistry
ISSN: 0095-8972 (Print) 1029-0389 (Online) Journal homepage: http://www.tandfonline.com/loi/gcoo20
Hydrolysis and DFT structural studies of dinuclear
Zn(II) and Cu(II) macrocyclic complexes of
m-12N3O-dimer and the effect of pH on their
promoted HPNP hydrolysis rates
C. Allen Chang, Tsung-Ta Wu & Hwa-Yu Lee
To cite this article: C. Allen Chang, Tsung-Ta Wu & Hwa-Yu Lee (2016): Hydrolysis and DFT
structural studies of dinuclear Zn(II) and Cu(II) macrocyclic complexes of m-12N3O-dimer and
the effect of pH on their promoted HPNP hydrolysis rates, Journal of Coordination Chemistry,
DOI: 10.1080/00958972.2016.1157177
To link to this article: http://dx.doi.org/10.1080/00958972.2016.1157177
View supplementary material
Accepted author version posted online: 25
Feb 2016.
Published online: 16 Mar 2016.
Submit your article to this journal
Article views: 2
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=gcoo20
Download by: [University Library Utrecht]
Date: 21 March 2016, At: 21:57

JOURNAL OF COORDINATION CHEMISTRY, 2016
http://dx.doi.org/10.1080/00958972.2016.1157177
Hydrolysis and DFT structural studies of dinuclear Zn(II) and Cu(II)
macrocyclic complexes of m-12N3O-dimer and the effect of pH on
their promoted HPNP hydrolysis rates
C. Allen Changa,b,c,d, Tsung-Ta Wub and Hwa-Yu Leea
aDepartment of Biomedical Imaging and Radiological Sciences, National Yang-Ming University, Taipei, Taiwan,
R.O.C; bDepartment of Biological Science and Technology, National Chiao Tung University, Hsinchu, Taiwan, R.O.C;
cBiophotonics & Molecular Imaging Research Center (BMIRC), National Yang-Ming University, Taipei, Taiwan, R.O.C;
dBiomedical Engineering Research and Development Center (BERDC), National Yang-Ming University, Taipei, Taiwan,
R.O.C
ABSTRACT
ARTICLE HISTORY
Received 6 October 2015
Accepted 8 February 2016
The synthesis of the ligand, m-12N3O-dimer (1,3-bis(1-oxa-4,7,10-
triazacyclododecan-7-yl)methyl)benzene, L), and the stability and hydrolysis
constants of its dinuclear Zn(II) and Cu(II) complexes are reported, in addition
to the effect of pH on HPNP (2-hydroxypropyl-4-nitrophenylphosphate)
hydrolysis reaction rates promoted by these complexes. Various structural
possibilities of the [Zn2L] and [Cu2L] hydrolytic species derived from solution
equilibrium modeling are predicted from density functional theory (DFT)
studies to correlate with the promoted HPNP hydrolysis reaction rates
and to establish the structure–function–reactivity relationship. Upon
deprotonation [Zn2L(OH)]3+ tends to form a structure with a “closed-form”
conformation where it is not possible for para-isomers. At pH >8, the
formation of the closed-form [Zn2L(OH)2]2+ and [Zn2L(μ-OH)(OH)2]+ species
led to faster promoted HPNP hydrolysis rates than the [Zn2L(OH)]3+ species.
On the other hand, the observed rates of the Cu2L-promoted HPNP hydrolysis
reaction were much slower than those of the [Zn2L]-promoted ones due
to formation of the inactive, di-μ-OH− bridged closed-form [Cu2L(μ-OH)2]2+
structure at high pH. The effects of solvent molecules and the use of higher
DFT computation levels, i.e., M06 and M06–2X, in conjunction with cc-pVDZ
and cc-pVTZ basis sets on the DFT-predicted structures for both [Cu(12N4)
(H2O)]2+ and [Zn(12N3O)(H2O)2]2+ complexes were also evaluated and
compared with those using the B3LYP/6–31G* method.
KEYWORDS
Macrocyclic ligands; Zn(II)
Ion; hydrolysis; dinuclear
species; DFT calculations
CONTACT C. Allen Chang
cachang@ym.edu.tw
Supplemental data for this article can be accessed at http://dx.doi.org/10.1080/00958972.2016.1157177.
© 2016 Taylor & Francis

2
C. A. CHANG eT AL.
1. Introduction
Zinc is an essential element of all forms of life. At the end of 2010, the Protein Data Bank (PDB) contained
6170 structures of zinc-binding proteins. Out of these, about 1925 proteins use zinc as a catalytic cofac-
tor and hydrolases, in particular, account for more than half of them [1]. Studies of zinc enzymes range
from their catalytic reaction mechanisms, structures, evolution, bioinformatics, and genomics [2–7] to
computational redesign of these enzymes [8]. A large number of studies using zinc complexes with
relatively low molecular weights have been reported to mimic selected biological functions of these
enzymes as anion (e.g. phosphate [9, 10]) receptors or artificial DNA/RNA-cleaving agents, including a
photoactive dinuclear zinc(II) complex [11]. Among these artificial DNA/RNA cleaving agents, dinuclear
complexes are more reactive than mononuclear ones [12, 13].
Recently, Morrow et al. reported the use of the dinuclear Zn(II) complex of 1,4-bis(1-oxa-4,7,10-tri-
azacyclododecan-7-yl)methyl)benzene (p-12N3O-dimer, scheme 1) for promotion of the cleavage of a
simple RNA analog lacking a nucleobase, HPNP (2-hydroxypropyl-4-nitrophenylphosphate, scheme 2)
[14]. In comparison with the mononuclear analog Zn(II) complex of 1-oxa-4,7,10-triazacyclododecane,
it was found that little cooperative effect was observed for the dinuclear Zn(II) complex on promoting
the hydrolysis of the RNA analog, HPNP. Prior to and after this, a number of dinuclear [15, 16], trinuclear
[17] and tetranuclear [18, 19] Zn(II) complexes of structural analogs based on cyclen (1,4,7,10-tetraaza-
cyclododecan, 12N4) were reported to bind phosphate, barbiturate, malonate, and DNA ions, separately.
Several reports on efficient phosphate diester hydrolysis by di-Zn(II) complexes with different macro-
cyclic ligand ring sizes or dinuclear ligand backbones have also appeared [12, 13, 20–23].
In this paper, we report the synthesis and characterization of dinuclear Zn(II) and Cu(II) macrocy-
clic complexes of 1,3-bis(1-oxa-4,7,10-triazacyclododecan-7-yl)methyl)benzene (m-12N3O-dimer, L,
scheme 1) as well as the effect of pH on the promoted HPNP hydrolysis rates by the two complexes. Of
particular interest is the use of density functional theory (DFT) calculations to evaluate various possible
structures of the hydrolytic species of these complexes [24–30] and to predict the most abundant one
which allows the elucidation of the structure–function–activity relationships of the dinuclear metal
complex catalyzed HPNP hydrolysis reactions. In contrast to the para-isomer reported by Morrow et al.,
our ligand, m-12N3O-dimer (L), is a meta-isomer which is capable of forming a “closed-form” structure
and is potentially capable of cooperatively promoting the cleavage of the RNA analog, HPNP.
Scheme 1. Structural formulas of m-12N3O-dimer (left) and p-12N3O-dimer (right). If the oxygen is changed to NH, the resulting
formulas would be m-cyclen-dimer and p-cyclen-dimer, respectively (vide infra).
Scheme 2. Hydrolysis of HPNP.

3
2. Results and discussion
2.1. Synthesis and purification of m-12N3O-dimer
The synthesis of the ligand m-12N3O-dimer has not been reported previously. However, it could be
done in a straightforward manner using previously published methods with minor modifications
[31, 32], as shown in scheme 3. Protection of the three secondary ring nitrogens from reaction with
DMF dimethyl acetal to form tricyclic 12N3O can be easily done in DMF solution. Reaction of 12N3O-
tricyclic with xylene dibromide produced the dipositive cationic bromide salt that precipitated in dry
acetonitrile solution and was easily filtered and purified for the final hydrolysis reaction to obtain the
m-12N3O-dimer ligand. A 13C NMR distortionless enhancement by polarization transfer (DePT) spectrum
of the m-12N3O-dimer for determining the presence of primary, secondary, and tertiary carbons is shown
in figure S1 (Supplemental data) with peak assignments. Note that the signal from the p-carbons at
δ = 138.8 ppm with no attached protons is absent (due to the lack of attached protons).
2.2. Standardization of ligand solutions
even in the crystalline state, the hydrochloride salt of the ligand may not have exact integral ratio of HCl
to the fully deprotonated ligand. Thus, a general standardization procedure of the ligand (L) solutions
has been established by pH titration using a standard tetramethylammonium hydroxide solution with
and without an excess of added Cu(II) salt (i.e. [Cu2+]>2.0[L] in the present case). The free ligand titration
gives an end-point after all the “weaker”-bound protons with relatively low pKa values were neutral-
ized. In the m-12N3O-dimer case, these protons are those with pKa values lower than 6 (equation (1),
figure 1). In the presence of excess of strongly binding Cu(II) ions, all the protons were expelled and
another end point was obtained. The difference between the two end points in mmole unit would be
the total amount of the“stronger”ligand bound protons (i.e. with pKa values > 6) which in the m-12N3O-
dimer case is four times of the ligand quantity (equation (2)). We used Cu(II) for the standardization of
the ligand because it forms stronger complexes than the Zn(II) by five orders of magnitude.
HnLn + +(n - 4)OH− ⇆ H4L4 + +(n - 4)H2O
(1)
2Cu2 + + H4L4 + + 4OH− ⇆ Cu2L + 4H2O
(2)
2.3. Ligand protonation constants and protonation sites
Figure 1 shows the respective titration curves of m-12N3O-dimer in the absence and presence of Zn(II)
and Cu(II). The logarithmic protonation constants of m-12N3O-dimer determined by the potentiometric
method are 9.83, 9.31, 7.98, 7.28 and 2.74. These values are mostly higher than those of the p-12N3O-di-
mer reported previously [14], i.e., 9.29, 9.15, 7.62 and 7.33, indicating that the m-12N3O-dimer is more
basic than the p-12N3O-dimer. Note that the first two logarithmic protonation constants are 10.11 and
Scheme 3. Synthesis of 1,3-bis(1-oxa-4,7,10-triazacyclododecan-7-yl)methyl)benzene (m-12N3O-dimer).

4
C. A. CHANG eT AL.
8.53 for the parent 12N3O monomeric macrocycle, and the first two protonations occur simultaneously
at the three nitrogens of the macrocycle but to a higher degree at the two nitrogens next to the ether
oxygens, as confirmed by NMR studies [33]. For the m-12N3O-dimer, because the basicity of the sec-
ondary nitrogen is greater than that of the tertiary nitrogen, the four protonations are all expected to
be at the secondary nitrogen. This is consistent with the results of DFT calculations (to be published).
2.4. Stabilities and hydrolysis of the [Zn2(m-12N3O-dimer)] and [Cu2(m-12N3O-dimer)]
complexes
The complex formation reactions of both Zn(II)/Cu(II)-m-12N3O-dimer (L) systems are relatively fast, as
evidenced by fast equilibrium pH establishment after each addition of standard base and the stability
constants could be conveniently determined by potentiometric pH titration techniques. The logarithmic
formation constants of the 1 : 1 and 2 : 1 metal–ligand complexes and their hydrolysis species (vide
infra) are calculated from the equilibrium titration data and are listed in table 1. Those of structurally
analogous complexes are also listed in table 1 for comparison.
Figure 1. Potentiometric titration curves of m-12N3O-dimer (L) and 1 : 1/2 : 1 ratios of Zn(II)/Cu(II)-L. [L] = 1.0 mM; 25 0.1 °C. Ionic
strength = 0.10 M (CH3)4NCl.
Table 1. Logarithmic formation constants of the Zn(II)/Cu(II)-m-12N3O-dimer (L) complex systems, 25 °C, ionic strength 0.1. The
standard deviations for the protonation constants are 0.03–0.05; for metal complex formation and hydrolysis constants, 0.04–0.20.
No.
equilibrium
[Zn2L]
[Cu2L]
[Zn2L1]a [Zn2L2]b [Zn2L3]c [ZnL4]d [CuL4]d
1
2
3
4
5
6
7
8
MHL3++ H+ ⇌ MH2L4+
7.03
7.97
11.53
−9.16
−10.05
6.45
−6.75
−8.36
−9.51
9.83
9.31
7.98
7.28
2.74
7.36
8.96
14.83
−9.48
n/a
12.36
−7.89
−10.07
n/a
ML2++ H+ ⇌ MHL3+
M2++ L ⇌ ML2+
10.43
−7.68
15.63
−8.73
ML(H2O)2+ ⇌ ML(OH)++ H+
ML(H2O)(OH)+ ⇌ ML(OH)2 + H+
ML2++ M2+ ⇌ M2L4+
(16.72)
−7.34
−7.82
−10.22
9.29
9.15
7.62
M2L(H2O)4+ ⇌ M2L(OH)3++ H+
M2L(H2O)(OH)3+ ⇌ M2L(OH)22++ H+
M2L(H2O)(OH)23+ ⇌ M2L(OH)3+ + H+
L + H+ ⇌ LH+
−6.7
−8.5
−7.23
−7.88
9
10
11
12
13
14
10.11
8.52
1.56
10.11
8.52
1.56
2+
LH+ + H+ ⇌ LH2
LH22++ H+ ⇌ LH3
3+
LH33++ H+ ⇌ LH45+
7.33
4+
LH44++ H+ ⇌ LH5
aL1 = p-12N3O-dimer (Ref. 7). The logβM2L value of Zn2L1 is shown in parenthesis (i.e. 16.72, Table 1, equlibrium no. 6) corresponding
to the equilibrium: 2 M2+ + L ⇌ M2L.
bL2 = m-cyclen-dimer (scheme 1, [34]).
cL3 = p-cyclen-dimer (scheme 1, [8a]).
dL4 = 12N3O [35, 36].

5
The stability constant logβM2L for the Zn(II)- m-12N3O-dimer complex (2 M2+ + L ⇌ M2L) can be
calculated from data in table 1 to be 11.53 + 6.45 = 17.98 (equilibria no. 3 and 6) which is higher than
that of the Zn(II)-p-12N3O–dimer complex, i.e., 16.72, consistent with the fact that the basicity of the
m-12N3O-dimer is greater than the p-12N3O-dimer. Because Zn(II) can be either five- or six-coordinate
and Cu(II) can only be five-coordinate when complexed by the 12N3O-monomeric ligand (L’) with four
donors, species such as [ML’(OH)2] and [M2L’(OH)3]+ can only be found for the Zn(II) complex, i.e., M=Zn.
Indeed, the addition of the Cu2LH−3 into the speciation model to fit our titration data always failed to
converge. Therefore, both [Zn2(m-12N3O-dimer)] and [Zn2(p-12N3O-dimer)] with potentially two coor-
dinated water molecules on each Zn(II) may be deprotonated to form the [M2L(OH)3]+ species (L = m- or
p-12N3O-dimer), consistent with previous reports that, depending on the coordination environment,
Zn(II) can be either five- or six-coordinate [35]. It was proposed that each Zn(II) in [Zn2(m-cyclen-dimer)]
[17] and [Zn2(p-cyclen-dimer)] [15] has only one coordinated water and each Zn(II) is five-coordinate.
Similar to the reference dinuclear complexes, [Zn2(p-12N3O-dimer)], [Zn2(m-cyclen-dimer)] and
[Zn2(p-cyclen-dimer)], the dinuclear [Zn2(m-12N3O-dimer)] can form various hydrolytic species with
distinct pKa values. The first pKa of the dinuclear [Zn2(m-12N3O-dimer)], i.e., pKa1 = 6.75 (table 1, equi-
librium no. 7), is similar to that of the [Zn2(m-cyclen-dimer)] complex, i.e., pKa1 = 6.7 and is among
the lowest pKa1 values reported [20, 21] (vide infra). The pKa1 value of the monomeric Zn(12N3O)
complex is either 8.18 [35] or 7.68 [36] and the pKa1 value of a trinuclear Zn(II)-bound H2O can be
as low as 6.08 [17]. The pKa1 values for the Zn(II) complexes of the m-isomers are lower than those
of the p-isomers. The respective differences of the first two pKa values (ΔpKa = pKa2 - pKa1) of the
[Zn2(p-12N3O-dimer)] and [Zn2(p-cyclen-dimer)] are only 0.48 and 0.65 (i.e. pKa2(7.82)–pKa1(7.34) and
pKa2(7.88)–pKa1(7.23), equilibria no. 7 and 8 in table 1) which are smaller than those of the [Zn2(m-
12N3O-dimer)] and [Zn2(m-cyclen-dimer)], i.e., ΔpKa = 1.61 and 1.80. The [Cu2(m-12N3O-dimer)]
complex has a ΔpKa value 2.18 which is also greater than those of the Zn2(p-12N3O/cyclen-dimer)
complexes.
For the [Zn2(m-12N3O-dimer)], after dissociation of the first proton from coordinated water on one
Zn(II), the resulting coordinated OH− could form an intramolecular H-bond with the water coordinated
to the other Zn(II) ion and form a “closed-form” structure as shown in chart 1 (left) [15, 20]. This is not
possible for the para-isomer because the intramolecular H-bond is not likely to form. Hence, the pKa1
value is lower for [Zn2(m-12N3O-dimer)]. Once the closed form structure is formed, dissociation of the
second proton from the complex becomes more difficult because the H-bond stabilizes the water O–H
bond which leads to higher pKa2. However, if the second proton is dissociated at higher pH, the result-
ing dinuclear complex could also form an “open-form” structure (chart 1, middle; vide infra) [15]. The
observation on the variations of the ΔpKa values for the [Zn2(m-12N3O-dimer)], [Zn2(m-cyclen-dimer)],
[Cu2(m-12N3O-dimer)], [Cu2(m-cyclen-dimer)] and [Zn2(p-12N3O-dimer)] complexes could, therefore,
be rationalized in the above-discussed manner.
The respective ΔpKa values for the pKa2 and pKa3 values (equilibria no. 8 and 9 in table 1) for [Zn2(m-
12N3O-dimer)] and [Zn2(p-12N3O-dimer)] are 1.15 and 2.40, indicating that the third proton is easier to
dissociate from the meta-isomer than from the para-isomer. This might indicate that after dissociation
Chart 1. Possible structures of the “closed-form” [Zn2(m-cyclen-dimer)(H2O)(OH)]+ (left), “open-form” [Zn2(m-cyclen-dimer)(OH)2]
(middle), and μ-OH-bridged “closed-form”species of the [Zn2(m-12N3O-dimer)(μ-OH)(OH)2]− complexes (right).

6
C. A. CHANG eT AL.
of the third proton, the resulting dinuclear [Zn2(m-12N3O-dimer)(OH)3]− could form a more stabilized
μ-OH-bridged species in the“closed-form”depicted also in chart 1 (right) [20] (vide infra). Note that the
bridging μ-OH− is considered to be a poor nucleophile as compared to the coordinated, nonbridged
OH− [21].
The formation constants for [CuL] and [Cu2L] (L = m-12N3O-dimer) complexes are higher than those
of the corresponding Zn(II) complexes, consistent with many reports published previously [36]. The fact
that the respective pKa values (equilibria no. 7 and 8 in table 1) of [Cu2L] (i.e. 7.89 and 10.07) are higher
than those of the [Zn2L] complexes (i.e. 6.75 and 8.36) indicates that after complexation the acidity of
Cu(II) is weaker than that of Zn(II). This could be one of the reasons why nature employs Zn(II) instead
of Cu(II) ions more often as catalytic cofactors for enzyme catalyzed hydrolysis reactions.
Figure 2 shows the speciation diagrams of [Zn2L] and [Cu2L], M = Zn or Cu. Note that M2L=[M2(m-
12N3O-dimer)]4+, M2LH−1=[M2(m-12N3O-dimer)(OH)]3+, M2LH−2=[M2(m-12N3O-dimer)(OH)2]2+ and
M2LH−3=[M2(m-12N3O-dimer)(OH)3]+. In general, more species are found for the [Zn2L] complex than
for [Cu2L]. Formation of M2LH−1 is at lower pH for [Zn2L] than for [Cu2L] because the former has lower
pKa. Note that the titration data only allow the accurate calculation of the formation constants up to
the Zn2LH−3 species because the uncertainty of pH measurements at high pH, i.e. beyond pH ~11, due
to the well-known alkaline error for pH measurements. It is not certain if the formation of the Zn2LH−4
species is possible at even higher pH.
2.5. DFT computations of Zn(II) and Cu(II) structures
Because various hydrolytic species were present simultaneously in solution, particularly at pH >7, single
crystals of the present hydrolytic complex species were difficult to be obtained. Thus, by using a similar
platform computation technology established previously [24, 25], we have performed DFT calculations
to predict the possible ligand protonation sites and structures of various hydrolytic species of the Zn2L
and Cu2L complexes, as well as the mononuclear [Zn(12N3O)(H2O)]2+ and [Cu(12N3O)(H2O)]2+ for compar-
ison (vide supra). This platform technology could be used as a complementary one to that of the X-ray
crystallography for structures in the solid state. The latter usually provides the most stable structure
and sometimes could be influenced by crystal lattice energy; the former can calculate the energies for
several possible structures and help predict the most abundant one in the absence of crystal structures.
Usually the most appropriate DFT parameters suitable for the complexes with known structures were
Zn2(m-12N3O-dimer)
Cu2(m-12N3O-dimer)
100
80
60
40
20
0
100
80
60
40
20
0
M2LH-1
M2L
M
M
M2LH-1
M2LH-2
M2L
M2LH-3
M2LH-2
MLH2
MLH2
MLH
6
4
8
10
2
4
6
8
10
pH
pH
Figure 2. Speciation diagrams of [Zn2(m-12N3O-dimer)] (left) and [Cu2(m-12N3O-dimer)] (right). The concentration of each
complex is 1.0 mM. M2L = M2(m-12N3O-dimer)4+, M2LH-1 = [M2(m-12N3O-dimer)(OH)]3+, M2LH−2 = [M2(m-12N3O-dimer)(OH)2]2+, and
M2LH−3 = [M2(m-12N3O-dimer)(OH)3]+; M = Zn, Cu.

7
determined first and then used to predict unknown structures. The following is the preliminary results
using the B3LYP/6–31G* in PCM/UFF aqueous solvent model and a more detailed report will be pub-
lished elsewhere. In general, both Zn(II) and Cu(II) complexes have similar structures when each metal
ion is five-coordinate, i.e., four from the 12N3O macrocyclic ligand and one from coordinated water.
Note that Zn(II) ion can also be six-coordinate depending on ligand environment.
Our initial studies on the monomeric complexes revealed that when complexed by 12N3O mon-
omer, Zn(II) can bind two more aqua ligands to form six-coordinate [Zn(12N3O)(H2O)2]2+ which
is more stable than the sum of the energies of five-coordinate [Zn(12N3O)(H2O)]2+ and H2O by
5.15 kcal mole−1. The average Zn–O(aqua) bond distance 2.156 Å of [Zn(12N3O)(H2O)2]2+ is longer than
those of [Zn(12N3O)(H2O)]2+ and [Cu(12N3O)(H2O)]2+ with only one coordinated water on each metal
ion, i.e., the respective estimated M–O(aqua) distances are 2.050 Å and 2.086 Å. When coordinated
to two aqua ligands, the Zn(II)–O(12N3O ether oxygen) bond distance (2.411 Å) is longer than
those of Zn(II)–N(12N3O) (average 2.141 Å). The Zn–12N3O plane distance of [Zn(12N3O)(H2O)2]2+
(0.94 Å) is also longer than those of the [Zn(12N3O)(H2O)]2+ and [Cu(12N3O)(H2O)]2+, i.e. 0.82 and
0.77 Å, respectively, making the Zn(II) in [Zn(12N3O)(H2O)2]2+ have higher coordination space to
adopt two aqua ligands (vide infra).
For the dinuclear complexes, selected DFT calculations predicted H2O-bound and OH−-bound struc-
tures in the open-forms and closed-forms are shown in figure 3.
Figure 3. The open-form and closed-form structures of [Zn2(m-12N3O-dimer)]. (A) Open-form [Zn2L(H2O)4]4+ (upper left); (B) Closed-
form [Zn2L(H2O)(OH)]3+ (upper right); (C) Closed-form [Zn2L(μ-OH)]3+ (middle left); (D) Open-form [Zn2L(OH)2]2+ (middle right); (E)
Closed-form [Zn2L(OH)2]2+ (lower left); (F) Closed-form [Zn2L(OH)3]+, i.e. [Zn2L(μ-OH)(OH)2]+ (lower right).

8
C. A. CHANG eT AL.
Among these structures, the“open-form”[Zn2L(H2O)4]4+ with six-coordination on each Zn(II) is more
stable than [Zn2L(H2O)2]4+ with five-coordination on each Zn(II) ion (figure 3, upper left). However, when
the coordinated waters dissociate to form various hydrolytic species, both open-form and closed-form
structures are possible and the closed-form structures were found to have lower energies and were
more stable than their corresponding open-form structures. For example, after dissociation of the
first proton, the resulting [Zn2L(H2O)(OH)]3+ structure (figure 3, upper right) with five-coordination
on each Zn(II) and (Zn–OH2⋯HO–Zn) H-bond is similar to the one proposed previously (chart 1, left)
and is more stable than the μ-OH bridged [Zn2L(μ-OH)]3+ (figure 3, middle left) and other species with
six-coordination on each Zn(II). This (Zn–OH2⋯HO–Zn) intramolecular H-bond is likely the reason why
the pKa1 value (i.e. 6.75) is among the lowest as compared to others reported previously (vide supra).
The open-form [Zn2L(OH)2]2+ structure (figure 3, middle right) is ~7.6 kcal mole−1 less stable than the
closed-form [Zn2L(OH)2]2+ structure (figure 3, lower left). The two coordinated OH− ions are linked
by a Zn–OH⋯HO–Zn H-bond in the closed-form [Zn2L(OH)2]2+ structure, which is ~6.3 kcal mole−1
more stable than the di-μ-OH− bridged, closed-form [Zn2L(μ-OH)2]2+ structure (figure 4, lower right).
Thus, favorable formation of the open-form [Zn2L(OH)2]2+ structure proposed previously (chart 1,
middle) is ruled out by DFT calculations. Lastly, the lowest energy closed-form [Zn2L(OH)3]+ structure
(figure 3, lower right) has one bridged μ-OH− and two mono-coordinated OH− ions, very similar to the
one depicted in chart 1 (right), i.e., [Zn2L(μ-OH)(OH)2]+. The respective calculated Zn⋯Zn distances for
the complexes in figure 3(A)-(F) are 10.39 Å, 5.30 Å, 3.80 Å, 10.10 Å, 4.73 Å, and 3.24 Å. Other selected
calculated bond distances and bond angles using the B3LYP/6–31G* in PCM/UFF aqueous solvent
model for these open-form and closed-form structures of the [Zn2L] complex system are listed in
table S1 (Supplemental data).
Figure 4. The open-form and closed-form structures of [Cu2(m-12N3O-dimer)]. (A) Open-form [Cu2L(H2O)2]4+ (upper); (B) Closed-form
[Cu2L(OH2)(OH)]3+ (lower left); (C) Closed-form [Cu2L(μ-OH)2]2+ (lower right).

9
The structures of the [Cu2L] system is similar to those of the [Zn2L] system, except that [Cu2L(H2O)2]4+
adopts an open-form anti-configuration and each Cu(II) ion is five-coordinate (figure 4, upper). This
dihydrate structure is similar to that of the published one of [Cu2(m-cyclen-dimer)(H2O)2]4+ with a
syn-configuration [30]. Note that the anti- and syn-configurations represent two examples of energet-
ically similar rotamers. After dissociation of one proton from [Cu2L(H2O)2]4+, the resulting [Cu2L(OH2)
(OH)]3+, i.e. Cu2LH−1, species takes a closed-form structure with a (Cu–OH2⋯HO–Cu) intramolecular
H-bond (figure 4, lower left). An interesting exception is that the dihydroxide species, Cu2LH−2, i.e.
[Cu2L(μ-OH)2]2+, adopts an unusual di-μ-OH− bridged closed-form configuration (figure 4, lower right)
instead of a structure similar to that of the closed-form [Zn2L(OH)2]2+ complex (figure 3, lower right).
The closed-form [Cu2L(OH)3]+, i.e. Cu2LH−3, structure does not exist based on DFT calculations, consist-
ent with the solution titration results. The respective calculated Cu⋯Cu distances for the complexes
in figure 4(A), (B) and (C) are 10.09 Å, 5.37 Å, and 3.03 Å. Other selected calculated bond distances and
bond angles using the B3LYP/6–31G* in PCM/UFF aqueous solvent model for these open-form and
closed-form structures of the [Cu2L] complex system are listed in table S2 (Supplemental data).
2.6. Effect of solvent molecules on the DFT predicted structures
To further verify whether the simplified DFT calculations predict the structures of these complexes, we
have studied the effects of solvent molecules. Using the approaches of previously published methods
[37–39], we have added one and up to four water molecules which form H-bond(s) with the [Cu(12N4)
(H2O)]2+ (i.e. [Cu(cyclen)(H2O)]2+) and [Zn(12N3O)(H2O)2]2+ complexes and examined their effects on the
complex structures and energies. Figure 5 shows the top and side views of [Cu(12N4)(H2O)]2+ with four
H-bonded water molecules to the amine H and the [Zn(12N3O)(H2O)2]2+ complex with three H-bonded
water molecules to the amine H. Cu(II) ion is closer to the 12N4 plane (i.e. 0.50 Å) and the Zn(II) ion is a
Figure 5. (A) Top (upper left) and side (upper right) views of [Cu(12N4)(H2O)]2+ with four H2O molecules H-bonded to the amine H; (B)
Top (lower left) and side (lower right) views of the [Zn(12N3O)(H2O)2]2+ complex with three H2O molecules H-bonded to the amine
H. The numbers refer to the H-bond distances.

10
C. A. CHANG eT AL.
Table 2. Experimental and DFT calculated average bond lengths (Å) and angles (°) and zero-point energies (E0) of [Cu(12N4)(H2O)]2+
with up to four H-bonded water molecules to the amine H calculated by B3LYP/6–31G* in PCM/UFF aqueous solvent model.
[Cu(12N4)(H2O)]2+
exp.a
2.147
2.035
105.03
86.08
0.53
+0H2O
+1H2O
+2H2O
+3H2O
+4H2O
Cu–Ow
Cu–N
Ow–Cu–N
N–Cu–N
Cu–N4 plane
∆E0
2.204
2.034
106.16
86.14
0.53
2.176
2.034
104.93
86.17
0.53
−10.62
2.230
2.028
104.46
86.41
0.51
−18.16
2.191
2.026
104.35
86.47
0.50
−26.69
2.226
2.026
104.32
86.47
0.50
−35.86
–
0
aX-ray crystallography experimental data from Ref. [35].
Table 3. DFT calculated bond lengths (Å) and angles (°) and zero-point energies (E0) of [Zn(12N3O)(H2O)2]2+ with up to three H-bond-
ed water molecules to the amine H calculated by B3LYP/6–31G* in PCM/UFF aqueous solvent model.
[Zn(12N3O)(H2O)2]2+
+0H2O
+1H2O
+2H2O
+3H2O
Zn–Ow1
Zn–Ow2
2.122
2.190
151.01
2.195
2.169
150.03
2.219
2.117
152.44
2.269
2.154
152.51
121.47
68.80
84.66
91.19
93.83
134.84
135.21
2.135
2.116
2.449
2.136
2.209
84.96
76.38
75.05
83.11
79.88
0.93
Ow1–Zn–N1
Ow1–Zn–N2
Ow1–Zn–O
Ow1–Zn–N4
Ow2–Zn–N1
Ow2–Zn–N2
Ow2–Zn–O
Ow2–Zn–N4
Zn–N1
Zn–N2
Zn–O
Zn–N4
Zn–N(O)
N1–Zn–N2
N2–Zn–O
O–Zn–N4
N4–Zn–N1
N(O)–Zn–N(O)
Zn–N(O) plane
∆E0
123.09
72.66
87.40
84.62
91.43
141.04
135.22
2.161
2.122
2.411
2.140
2.209
83.82
76.56
75.62
81.96
79.49
0.94
123.77
72.48
86.76
85.97
92.14
139.18
134.85
2.155
2.132
2.390
2.131
2.202
83.94
77.00
75.90
82.33
79.79
0.93
121.84
69.61
85.70
87.72
91.19
137.04
136.90
2.146
2.120
2.448
2.123
2.209
84.45
76.20
75.69
82.62
79.74
0.93
0
−6.755
−16.96
−22.63
bit farther away from the 12N3O plane (i.e. 0.93 Å). Tables 2 and 3 list the average bond lengths, bond
angles, and zero-point energies of the respective [Cu(12N4)(H2O)]2+ and [Zn(12N3O)(H2O)2]2+ complexes
with up to four or three H-bonded water molecules to the amine H calculated by B3LYP/6–31G* in PCM/
UFF aqueous solvent model, including the experimentally determined data by X-ray crystallography
[29], where available.
From the data shown in tables 2 and 3, the more solvent molecules added to the complex system,
the lower the stabilization energies. For the Cu(II) complex, the more solvent water molecules H-bonded
to the ligand amine H, the closer the Cu(II)-12N4 plane distance and the more deviation from the single
crystal data. For the Zn(II) complex, the effect of solvent water molecules on the overall structure is less
obvious as compared to that of the Cu(II) complex.
2.7. Effects of higher DFT computation levels on the complex structures and energies
Two higher level functionals, i.e., M06 and M06–2X, in conjunction with cc-pVDZ and cc-pVTZ basis
sets were used to perform the structure and energy computations and the results were compared with
those using the B3LYP/6–31G* method. The method using the higher level M06–2X/cc-pVTZ would
normally be expected to be better, however, with much longer computation time and higher cost
[40, 41]. The mononuclear [Cu(12N4)(H2O)]2+ and the dinuclear [Cu2(m-cyclen-dimer)(H2O)2]4+ have been

11
Table 4. Average calculated and experimental bond distances (Å) and angles (°) of [Cu(12N4)(H2O)]2+ by B3LYP/(6–31G*, cc-pVTZ),
M06/(6–31G*, cc-pVDZ, cc-pVTZ), and M06–2X/(6–31G*, cc-pVDZ, cc-pVTZ) methods in PCM/UFF aqueous solvent model.
B3LYP
M06
M06–2X
[Cu(12N4)(H2O)]2+
exp.a
6–31G* cc-pVTZ 6–31G* cc-pVDZ cc-pVTZ 6–31G* cc-pVDZ cc-pVTZ
Cu–Ow
Ow–Cu–N
Cu–N
N–Cu–N
Cu–12N4 plane
2.147
105.03
2.035
86.08
0.53
2.204
106.16
2.034
86.14
0.53
2.239
105.62
2.060
85.84
0.55
2.177
104.26
2.012
86.50
0.50
2.170
105.46
2.042
85.91
0.55
2.207
104.96
2.035
86.16
0.53
2.141
106.42
2.058
85.53
0.57
2.147
106.84
2.079
85.17
0.60
2.164
106.52
2.074
85.35
0.59
aRef. [29].
Table 5. Average calculated and experimental bond distances (Å) and angles (°) of [Cu2(m-cyclen-dimer)(H2O)2]4+ by B3LYP/(6–
31G*, cc-pVTZ) and M06/(6–31G*, cc-pVDZ, cc-pVTZ) methods in PCM/UFF aqueous solvent model.
B3LYP
M06
[Cu2(m-cyclen-dimer)(H2O)2]4+
exp.a
2.151
6–31G*
cc-pVTZ
6–31G*
cc-pVDZ
cc-pVTZ
Cu–Ow
2.205
104.70
104.67
2.042
2.042
86.33
86.35
0.52
2.236
105.39
105.31
2.071
2.070
85.98
86.03
0.55
2.174
104.29
104.15
2.019
2.018
86.53
86.59
0.49
2.166
105.43
105.30
2.051
2.050
85.95
86.02
0.54
2.200
105.09
104.87
2.045
2.043
86.12
86.23
0.53
Ow1–Cu1–N
Ow2–Cu2–N
Cu1–N
104.42
104.00
2.028
2.029
86.50
86.70
0.49
Cu2–N
N–Cu1–N
N–Cu2–N
Cu–12N4 plane
aRef. [30].
chosen for these computations because the results can be compared with the crystallographic data
determined by the X-ray diffraction method. Tables 4 and 5 summarize the results of the [Cu(12N4)
(H2O)]2+ and [Cu2(m-cyclen-dimer)(H2O)2]4+ complexes, respectively.
From the data in table 4 and by using the experimental X-ray data of [Cu(12N4)(H2O)]2+ for compar-
ison, the results obtained by both the B3LYP/6–31G* and the M06/cc-pVTZ methods are comparable
and acceptable. However, the M06/cc-pVTZ method would take longer time and, thus, be more costly
and is not desirable for routine studies. On the other hand, the results obtained by the M06–2X method
are not as good as those either by the M06/cc-pVTZ method or by the B3LYP/6–31G* method. More
examples are required to be tested and to obtain statistically significant conclusions.
From the data in table 5 and by using the experimental X-ray data of [Cu2(m-cyclen-dimer)(H2O)2]4+
as reference, the results obtained by the B3LYP/6–31G* method are also quite acceptable. However, the
method using M06/6–31G* is even better and more similar to the experimental data. Note that the use of
the M06/cc-pVTZ method does not generate better results. Moreover, the use of M06–2X method results
in imaginary frequency and cannot be used for the present computation. Overall, the use of higher level
DFT computation methods involving cc-pVDZ and cc-pVTZ basis sets do not generate better results.
Our current B3LYP/6–31G* method seems to be adequate with less computing time and lower cost.
The calculated bond distances and bond angles of [Zn(12N3O)(H2O)2]2+ by the B3LYP/(6–31G*,
cc-pVTZ), M06/(6–31G*, cc-pVDZ, cc-pVTZ), and M06–2X/(6–31G*, cc-pVDZ, cc-pVTZ) methods in PCM/
UFF aqueous solvent model are given in table S3 (Supplemental data).
From a theoretical point of view, instead of comparing the calculated structural results with the
experimental data obtained from X-ray crystallography, one could consider the result with the lowest
zero point energy to be the most desirable one. Table S4 (Supplemental data) lists the calculated zero-
point energies (E0), free energies (G0) and the energy differences (∆E and ∆G) of [Cu(12N4)(H2O)]2+
,
[Cu2(m-cyclen-dimer)(H2O)2]4+ and [Zn(12N3O)(H2O)2]2+ by B3LYP/(6–31G*, cc-pVTZ), M06/(6–31G*,
cc-pVDZ, cc-pVTZ), M06–2X/(6–31G*, cc-pVDZ, cc-pVTZ) method in PCM/UFF aqueous solvent model.

12
C. A. CHANG eT AL.
Figure 6. Left: [Zn2(m-12N3O-dimer)] (●) and [Cu2(m-12N3O-dimer)] (○) promoted HPNP hydrolysis reaction rate constants,
obs, at various pH. [Complex] = 1.0 mM. Right: Plot of kobsvs. [Zn2(m-12N3O-dimer)] at pH 7.0, slope = (1.10 0.02)×10−3 M−1s−1
intercept = 2.26 × 10−7 s−1, r2 = 0.999. [HPNP] = 0.0362 mM, [Buffer] = 0.12 M, ionic strength = 0.1 M, T = 25.0 0.1 °C.
k
,
If the result of the B3LYP/6–31G* method is arbitrarily defined to be zero in energy, and the relative
energies are calculated by ∆E (or ∆G)=E0 (or G0)(other method)–E0 (or G0) (B3LYP/6–31G*), it is clear
that the B3LYP/cc-pVTZ method gives the lowest relative energies for all three complexes. Next is the
M06–2X/cc-pVTZ method if the [Cu2(m-cyclen-dimer)(H2O)2]4+ dimer is not considered. For the [Cu2(m-
cyclen-dimer)(H2O)2]4+ dimer, the second best method is the B3LYP/6–31G* method which is the one
we choose for the present study.
2.8. Effects of pH on the HPNP hydrolysis reactions promoted by [Zn2(m-12N3O-dimer)] and
[Cu2(m-12N3O-dimer)]
Two mechanistic pathways exist for the hydrolysis of HPNP, metal-catalyzed cleavage which is opera-
tional starting at pH 7 and below, and OH−-catalyzed cleavage, which dominates at pH 10 and above.
The purpose of this study is not a comprehensive kinetic mechanistic study but to evaluate which
species based on the hydrolysis and DFT studies might be the most probable one to promote the
HPNP hydrolysis reaction at variable solution pH to establish a structure–function–reactivity relation-
ship. Figure 6 (left) shows plots of the HPNP hydrolysis reaction rate constants (kobs) promoted by
[Zn2(m-12N3O-dimer)] and [Cu2(m-12N3O-dimer)] at pH up to pH 11. The numerical data are listed in
table S5 (Supplemental data). Note that the OH− ion catalyzed HPNP hydrolysis reaction second-order
rate constant (kOH) (i.e. kobs = kOH[OH−]) was determined to be 6.3 × 10−2 M−1s−1 which was similar to
that (6.7 × 10−2 M−1s−1) published previously [14]. Therefore, at pH 10, 10.5, and 11, the contributions
by OH− catalyzed pathway to the HPNP hydrolysis kobs value would be 6.3 × 10−6 s−1, 2.0 × 10−5 s−1
and 6.3 × 10−5 s−1, respectively, which confirm that the kobs values beyond pH 10 were dominated by
OH− catalysis. Thus, we will discuss the rate results up to pH 10.
For [Zn2(m-12N3O-dimer)] promoted HPNP hydrolysis reactions, the rate constants increase with
increasing pH in the pH range 7–10. From the speciation diagram of [Zn2(m-12N3O-dimer)] (figure 2,
left) at pH 7 the most abundant species is Zn2LH−1; between pH 8.5–9.0, Zn2LH−2, and beyond pH 9.5,
Zn2LH−3. The kobs,corr. values obtained from the experimental kobs values after subtraction of the contri-
butions from OH− catalysis and the % speciation data for the Zn2LH−1, Zn2LH−2, and Zn2LH−3 species
are listed in table S6 (Supplemental data). Using the data in table S6 and equation 3 and multiple lin-
ear regression analysis [42], we could calculate the following values: kZn2LH−1=(0.6 2.0)×10−3 M−1s−1
Zn2LH−2=(2.1 0.2)×10−2 M−1s−1 and kZn2LH−3=(1.6 0.2)×10−2 M−1s−1, where kZn2LH−1, kZn2LH−2 and kZn2LH−3
are the respective individual second-order rate constants for each hydrolytic species.
,
k

13
[
]
[
]
[
]
kobs,corr. = kZn2LH - 1 Zn2LH−1 +kZn2LH - 2 Zn2LH−2 +kZn2LH - 3 Zn2LH−3
(3)
NotethatthekZn2LH-1 valuehadalargererrorwhichcouldbeduetotheun-weightedanalysis. Nevertheless,
it was clear that the HPNP hydrolysis activities were in the order Zn2LH−2 > Zn2LH−3 > Zn2LH−1. Although
the HPNP hydrolysis promotion activities of all the Zn2L species were inferior to that of OH−, they were
more effective at lower, e.g. physiological, pH.
To evaluate the effect of promoter concentration, the promoted kobs values for the HPNP hydrolysis
reaction at pH 7, 25.0 0.1 °C were measured and plotted against the total concentration of [Zn2(m-
12N3O-dimer)] from 1.0 to 5.0 mM (ionic strength 0.10 M), and a straight line with a slope (i.e. second-or-
der rate constant) of 0.0011 M−1s−1 was obtained (figure 6, right). This second-order rate constant was
lower than those found previously for the monomeric [Zn(12N3O)]2+ and dimeric [Zn2(p-12N3O-dimer)]
complexes, i.e., 1.9 × 10−3 and 3.8 × 10−3 M−1s−1, respectively [14]. At pH 7, the % speciation of Zn2LH-1
is ~59%, which translates to a value of kZn2LH-1 = 1.9 × 10−3 M−1s−1, which is within the error limit of the
value found from multiple linear regression analysis of the kobs data at various pH values discussed above.
Because [Zn2(m-12N3O-dimer)(OH)]3+ (i.e. Zn2LH−1) is likely in the closed-form conformation with the
intramolecular (Zn–OH2⋯HO–Zn) H-bond as predicted by DFT calculations, its low HPNP hydrolysis
promotion activity as compared to the open-form [Zn2(p-12N3O-dimer)] complex could be rationalized.
It is possible that the more stable but less reactive closed-form dimeric [Zn2(m-12N3O-dimer)(OH)]3+
was in equilibrium with a more reactive open-form dimeric species and resulted in an overall reactivity
similar to that of the monomeric [Zn(12N3O)]2+ at pH 7.
The relatively more reactive [Zn2(m-12N3O-dimer)(OH)2] (i.e. Zn2LH−2) and [Zn2(m-12N3O-dimer)
(OH)3]+ (i.e. [Zn2L(μ-OH)(OH)2]+ or Zn2LH−3) as compared with [Zn2(m-12N3O-dimer)(OH)]3+ (i.e. Zn2LH−1
)
are also predicted by DFT computations to be in closed-form conformations, the former with a Zn–
OH⋯HO–Zn H-bond (figure 3, lower right) and the latter with a mono-μ-OH−bridged structure depicted
in chart 1 (right). Besides the poor nucleophilic nature of the μ-OH−, both [Zn2(m-12N3O-dimer)(OH)2]
and [Zn2(m-12N3O-dimer)(μ-OH)(OH)2]+ have two mono-coordinated, nonbridged OH− ions. However,
the former has higher hydrolysis promotion reactivity. This might be attributed to the more compact
“pocket” environment of [Zn2(m-12N3O-dimer)(μ-OH)(OH)2]+ resulting from μ-OH− bridging with
reduced accessibility of the HPNP substrate. This is manifested by the shorter Zn⋯Zn distance of the
latter (3.24 Å) as compared with that of the former (4.73 Å). Unfortunately, no corresponding rate data
are available for [Zn2(p-12N3O-dimer)] at higher pH for comparison.
The low reactivities of [Cu2(m-12N3O-dimer)] as compared to those of the [Zn2(m-12N3O-dimer)]
are then easy to explain because Cu(II) can only be five coordinate and is less acidic when coordinated
by the 12N3O macrocycle. The kobs values of [Cu2(m-12N3O-dimer)] increased from pH 7 to 8 and then
decreased from pH 8 to 10, indicating that the active species is most likely Cu2LH−1 which has a more
stable closed-form [Cu2(m-12N3O-dimer)(H2O)(OH)]3+ structure. At pH 7, the kobs value of the [Cu2(m-
12N3O-dimer)] solution is roughly 0.4 times of that of the [Zn2(m-12N3O-dimer)] solution, presumably
because the effective concentration of the active [Cu2(m-12N3O-dimer)(H2O)(OH)]3+ complex is lower
than that of [Zn2(m-12N3O-dimer)(H2O)(OH)]3+. From the speciation diagram (figure 2, right), the con-
centration of the Cu2LH−1 (i.e. [Cu2(m-12N3O-dimer)(H2O)(OH)3+]) increases from pH 6 to a maximum at
pH ~9 and then decreases as pH is further increased. The kobs values follow the same trend. The unusual
di-μ-OH− bridged closed-form conformation adopted by Cu2LH−2 (i.e. [Cu2(m-12N3O-dimer)(μ-OH)2]3+
)
is inactive with respect to promotion of HPNP hydrolysis. The kobs value found at pH 10 was even lower
than that by OH− alone which might be attributed to the binding of the HPNP substrate by Cu2LH−1
leading to lower effective substrate concentration.
3. Conclusion
Studies of solution speciation and hydrolytic behaviors at various pH values coupled with DFT structural
computations turned out to be a very powerful platform which led to understanding of the most possi-
ble solution structures of various species of the dinuclear Zn(II) and Cu(II) complexes of the macrocyclic

14
C. A. CHANG eT AL.
ligand, m-12N3O-dimer. These results rationalize some physicochemical properties such as trends on
pKa values and correlate with the promoted HPNP hydrolysis reaction rates by the two complexes to
establish structure–function–reactivity relationships. Specifically, formation of the closed-form [Zn2(m-
12N3O-dimer)(OH)]3+ (i.e. Zn2LH-1), [Zn2(m-12N3O-dimer)(OH)2]+ (i.e. Zn2LH−2) and [Zn2(m-12N3O-dimer)
(μ-OH)(OH)2]+ (i.e. Zn2LH−3) allows rationalization of the respective smaller pKa1, larger pKa2 and smaller
pKa3 values of the meta-isomer as compared with those of the para-isomer, i.e. [Zn2(p-12N3O-dimer)].
The favorable formation of the open-form [Zn2(m-12N3O-dimer)(OH)2]+ (i.e. Zn2LH−2) proposed previ-
ously was also ruled out.
Concerning the promoted HPNP hydrolysis reactivity, at pH 7, the promoted reaction rate for Zn2LH−1
with an apparent second-order rate constant 1.9 × 10−3 M−1s−1 was smaller than that of the para-isomer.
At pH >8, the formation of the closed-form [Zn2(m-12N3O-dimer)(OH)2]+ (i.e. Zn2LH−2) and [Zn2(m-12N3O-
dimer)(μ-OH)(OH)2]+ (i.e. Zn2LH−3) led to faster catalyzed HPNP hydrolysis rates which was eventually
surpassed by that of OH− catalysis at pH >10. The individual second-order rate constants could be
obtained from multiple linear-regression analysis of the observed kobs data to show their activities in
the order Zn2LH−2 > Zn2LH−3 > Zn2LH−1. The observed rates of the Cu2L-promoted HPNP hydrolysis were
much slower than those of the Zn2L-promoted ones due to lower acidity of the Cu(II) ion after macr-
ocyclic ligand complexation and formation of the inactive [Cu2L(μ-OH)2]2+ structure with an unusual
di-μ-OH− bridged closed-form configuration at high pH.
The effect of solvent molecules on the DFT predicted structures were also evaluated and it was found
that for the Cu(II) complex, more solvent water molecules H-bonded to the ligand amine H, the closer the
Cu(II)-12N4 plane distance and the more deviation from the single crystal data. For the Zn(II) complex, the
effect of solvent water molecules on the overall structure is less obvious than for the Cu(II) complex. The use
of higher DFT computation levels, i.e. M06 and M06–2X, in conjunction with cc-pVDZ and cc-pVTZ basis sets
on the complex structures and energies were compared with those using the B3LYP/6–31G* method. From
the structural point of view, the results obtained by both the B3LYP/6–31G* and the M06/cc-pVTZ methods
are comparable and quite acceptable. However, the M06/cc-pVTZ method would take longer and cost more
and is not as desirable for routine studies. The results obtained by the M06–2X method are not as good as
those either by the M06/cc-pVTZ method or by the B3LYP/6–31G* method. From the zero point energy,
for [Cu(12N4)(H2O)]2+, [Cu2(m-12N4-dimer)(H2O)2]4+ and [Zn(12N3O)(H2O)2]2+, the B3LYP/cc-pVTZ method
gives the lowest relative energies for all three complexes. The next best is the M06–2X/cc-pVTZ method if
the [Cu2(m-12N4-dimer)(H2O)2]4+ dimer is not considered. If [Cu2(m-12N4-dimer)(H2O)2]4+ is considered, the
second best method is the B3LYP/6–31G* method which is the one we chose for the present study. This
study could be extended to other complex systems not only in solution but also in the solid state structures,
spectroscopic and other properties [43, 44].
4. Experimental
4.1. Materials and standard solutions
Analytical reagent-grade chemicals and buffers, unless otherwise stated, were purchased from Sigma (St.
Louis, MO), Aldrich (Milwaukee, Wl) or Merck (Darmstadt, Germany) and were used as received. Disodium
ethylenediaminetetraacetic acid (Na2H2eDTA) was purchased from Fisher. The ligand m-12N3O-dimer
was prepared and purified as described below. Carbonate-free deionized water was used for all solution
preparations.
The concentration of the m-12N3O-dimer (L) stock solution (ca. 0.01 M) was determined by pH
titration using a standard tetramethylammonium hydroxide solution (0.1 M), with and without an
excess of added Cu(II) salt (i.e. [Cu2+]>2.0[L]), and was also checked by complexometric back-titration.
The concentrations of the lanthanide nitrate stock solutions were ca. 0.01 M and were standardized
by eDTA titration using xylenol orange as indicator. The eDTA solution was standardized by titrating a
calcium carbonate primary standard solution (first dissolved in HCl solution) at pH 10 using calmagite
as the indicator.

15
The 0.1 M tetramethylammonium hydroxide solution was prepared by diluting a 20% (CH3)4NOH-
methanol solution obtained from Aldrich (carbonate-free). The aqueous (CH3)4NOH solution was
standardized by using reagent grade primary standard potassium hydrogen phthalate. A 0.1 M HCl
solution was prepared by diluting a reagent-grade HCl solution and standardized by using the standard
(CH3)4NOH solution. A 1.0 M stock solution of tetramethylammonium chloride (Aldrich) was prepared
and diluted to 0.1 M for each titration to maintain a constant ionic strength (0.1 M).
2-Hydroxypropyl-4-nitrophenyl phosphate, barium salt (HPNP, MW 344.82, barium salt form) was
prepared according to a published method [45].
4.2. Synthesis of m-12N3O-dimer (scheme 3)
Tricyclo[8,2,1,0[4, 13]]1-oxa-4,7,10-triazacyclotridecane (12N3O-tricyclic): To 0.03 mol (5.22 g) of
1-oxa-4,7,10-triazacyclododecane (12N3O) was added 90 mL toluene in a round bottom flask under
nitrogen at 60 °C with stirring. To the solution was then added 0.03 mol (3.57 g) dimethylformamide
dimethyl acetal and the mixture was allowed to react for 6 h, cooled and rotary evaporated to dryness
to obtain the yellow oily material (yield = 84%). NMR (CDCl3, 300 MHz), 1H NMR, δ (ppm): 2.6, 2.7, 2.8,
3.3, 3.4, 4.7; 13C NMR, δ (ppm): 51.7, 53.3, 53.8, 71.5, 100.8.
1,3-Bis(1-oxa-4,7,10-triazacyclododecan-7-yl)methyl)benzene (m-12N3O-dimer): To 0.031 mol (5.67 g)
12N3O-tricyclic in a round bottom flask was added 80 mL dry acetonitrile and the solution was stirred under
nitrogen.To the solution was then added 0.0125 mol (3.30 g) xylene dibromide in 15 mL dry acetonitrile and
the mixture was allowed to react at 50 °C for 6 h to obtain a white precipitate.The precipitate was filtered and
washed with 10 mL cold dry acetonitrile three times and redissolved in 80 mL 2 N NaOH solution.The mixture
was refluxed for two days, cooled to room temperature, and extracted three times with 30 mL chloroform.
The chloroform extracts were combined and treated with anhydrous MgSO4 and the solution was filtered and
rotary evaporated to dryness and the solids were recrystallized in ethanolic HCl aqueous solution to obtain
the white product (yield = 94%). Anal. Calcd for m-12N3O-dimer·HCl salt (C24H44N6O2·6HCl·5H2O): C, 38.06;
H, 7.97; N, 11.09. Found: C, 38.09; H, 8.05; N, 10.96. NMR (CDCl3, 300 MHz), 1H NMR, δ (ppm): 2.58, 2.63, 2.73,
3.53, 3.58, 7.19; 13C NMR, δ (ppm): 44.1, 46.2, 51.4, 59.2, 66.3, 128.1, 128.5, 130.1, 138.8. eSI(+)-MS, m/z: 449.5.
4.3. Potentiometric titrations
Acid-base and complex formation titrations were carried out at a constant ionic strength of 0.10 M
(CH3)4NCl using reported procedures [46]. A Model 720 Metrohm Titroprocessor in conjunction with
Metrohm Combination electrode was employed to monitor the pH. Generally, equal molar concen-
trations of the ligand and the metal ion were employed (~1.0 mM) for titration. The ionic strength of
the solution was adjusted to 0.1 M using 1 M (CH3)4NCl. The (CH3)4NOH solution was delivered from a
10 mL automatic Brinkmann Metrohm Model 665 Dosimat buret with a reading accuracy of 0.001 mL.
The pH-metric titration data were used to calculate the stepwise ligand protonation constants
defined in equation (4), where n = 1–6 for 12N3O-dimer (L):
[
] ([
][ ])
Kn= HnL ∕
H
n - 1L H+
(4)
All equilibrium calculations were performed using the program Hyperquad2008 Version 5.2.15. The
averaged values are presented with the standard deviations calculated from valid data points. Speciation
diagrams were generated using Hyss 2006 software.
4.4. DFT computations of Zn(II) and Cu(II) complex structures
The simulated annealing (SA) method [25] employing the Universal and Dreiding force field in the
Forcite Tools of Material Studio software [26] was employed with a temperature cycle of 298.0–800.0 K
to find the lowest-energy mononuclear, and the open-form and closed-form conformers of the Zn/Cu
(12N4/12N3O) mononuclear and dinuclear complexes. The results are then refined by DFT calculations.

16
C. A. CHANG eT AL.
All DFT calculations were performed by employing hybrid DFT with the B3LYP exchange correlation
functional of the Gaussian 09 package [27]. Full geometry optimizations of the mononuclear model
compounds Zn/Cu(12N4/12N3O) as well as the Zn2(m-12N3O-dimer) and Cu2(m-12N3O-dimer) systems
(i.e. with various hydrolytic species) were obtained by using the 6–31G* basis set for carbon, hydrogen,
nitrogen, and oxygen. The initial structures of various Zn/Cu(12N4/12N3O) complexes were obtained
from the X-ray crystallographic experimental results of the same or structurally similar complexes,
including [Zn(12N4)(etOH)]2+ [28] for [Zn(12N4)(H2O)]2+, [Cu(12N4)(Br)]+ [29] for [Cu(12N3O)(H2O)]2+, and
[Cu2(m-12N3O-dimer)(H2O)2]4+ [30]. The geometries of the complexes were then fully optimized by
B3LYP/6–31G* in aqueous solutions by using a polarizable continuum model (PCM/UFF). each molecular
structure has been fully optimized and characterized as minimum by frequency analysis.
The M06 and M06–2X functionals as well as the cc-pVDZ and cc-pVTZ basis sets were also used
to study the effect of higher level computational methods. All calculations were performed with the
software and facilities of the National Center for High-performance Computing in Hsinchu, Taiwan.
4.5. Kinetic measurements
The Zn(II) and Cu(II) m-12N3O–dimer complex solutions were freshly prepared by mixing 2.00:1.02
metal salt to ligand molar ratio solutions. The pH of each solution was adjusted to 6.0–6.5 by adding
appropriate amount of (CH3)4NOH solution. To each solution, the HPNP solution was added and the
final pH adjusted by adding an appropriate amount of buffer stock solution and used immediately
after preparation. MPS (3-[N-morpholino]propanesulfonic acid), TAPS (3-[[tris(hydroxylmethyl)methyl]
amino] propane sulfonic acid), CHeS (2-[cyclohexyl amino]ethanesulfonic acid) and CAPS (3-[cyclohex-
ylamino]-1-propanesulfonic acid) with the respective pKa values 7.2, 8.4, 9.3, and 10.4 were used to
prepare buffer solutions with the desired pH. The final HPNP concentration was kept at 0.0362 mM and
the Zn(II)/Cu(II) complex concentrations were 1.0 mM to fulfill pseudo-first-order reaction conditions.
The ionic strength was adjusted to 0.1 M with (CH3)4NCl except otherwise specified. A HP 8453 UV–vis
spectrophotometer was used to measure the absorption increase with time at 400 nm due to formation
of nitrophenolate after HPNP hydrolysis. Pseudo-first-order rate constants were calculated using the
initial rate method or the integral method. Sigma plot was use for curve fitting. Rate constants quoted
represent the average values of two to three kinetic runs and were reproducible within 10% [47].
4.6. Multiple linear-regression analysis
Multiple linear regression analysis of the rate data was conducted similarly to a previous reported paper with
minor modifications using the computer software-Statistical Package for the Social Sciences (SPSS) [42].With
a multiple-regression model, the relationship is described using a generalization of a straight-line equation (5):
Y = ꢀ0 + ꢀ1X1 + ꢀ2X2 + … + ꢀpXp + e
(5)
where Y denotes the response variable (i.e. kobs values), β0 is a constant (intercept), X1, X2, …, Xp is a vec-
tor of p predictor variables (i.e. concentration of each species at different pH), β1, β2, …, βp is the vector
of p regression coefficients (i.e. individual second order rate constant of each species). each predictor
variable has its own coefficient, and the outcome variable is predicted from a combination of all the
variables multiplied by their respective coefficients plus an error term. The relationship was determined
using the least squares fitting technique.
Acknowledgement
The authors wish to thank the National Science Council and the Atomic energy Council of the Republic of China (Taiwan)
for financial support (grant numbers NSC-98-2113-M-010-001-MY3, NSC 96-NU-7-009-003, and NSC-100-2811-M-010-003)
of this work.

17
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
This work was supported by the National Science Council and the Atomic energy Council of the Republic of China (Taiwan)
for financial support [grant number NSC-98-2113-M-010-001-MY3], [grant number NSC 96-NU-7-009-003], [grant number
NSC-100-2811-M-010-003] of this work.
References
[1] C. Andreini, I. Bertini. J. Inorg. Biochem., 111, 150 (2012).
[2] W.J. Marer. J. Inorg. Biochem., 111, 110 (2012).
[3] R.J.P. Williams. J. Inorg. Biochem., 111, 104 (2012).
[4] J.W. Yun, T. Tanase, S.J. Lippard. Inorg. Chem., 35, 7590 (1996).
[5] B. Bauer-Siebenlist, F. Meyer, e. Farkas, D. Vidovic, J.A. Cuesta-Seijo, R. Herbst-Irmer, H. Pritzkow. Inorg. Chem., 43,
4189 (2004).
[6] C. Andreini, I. Bertini, A. Rosato. Acc. Chem. Res., 42, 1471 (2009).
[7] J.-H. Shin, H.J. Jung, Y.J. An, Y.-B. Cho, S.-S. Cha, J.-H. Roe. Proc. Natl. Acad. Sci. USA, 108, 5045 (2011).
[8] S.D. Khare, Y. Kipnis, P. Greisen, R. Takeuchi, Y. Ashani, M. Goldsmith, Y. Song, J.L. Gallaher, I. Silman, H. Leader, J.L.
Sussman, B.L. Stoddard, D.S. Tawfik, D. Baker. Nature. Chem. Biol., 8, 294 (2012).
[9] S.H. Gellman, R. Petter, R. Breslow. J. Am. Chem. Soc., 108, 2388 (1986).
[10] e. Boseggia, M. Gatos, L. Lucatello, F. Mancin, S. Moro, M. Palumbo, C. Sissi, P. Tecilla, U. Tonellato, G. Zagotto. J. Am.
Chem. Soc., 126, 4543 (2004).
[11] A. Panja, T. Matsuo, S. Nagao, S. Hirota. Inorg. Chem., 50, 11437 (2011).
[12] M. Ruf, K. Weis, H. Vahrenkamp. J. Am. Chem. Soc., 118, 9288 (1996).
[13] Y. Gultneh, Allwar, B. Ahvazi, D. Blaise, R.J. Butcher, J. Jasinski, J. Jasinski. Inorg. Chim. Acta, 241, 31 (1996).
[14] C.S. Rossiter, R.A. Mathews, I.M.A. Delmundo, J.R. Morrow. J. Inorg. Biochem., 103, 64 (2009).
[15] T. Koike, M. Takashige, e. Kimura, H. Fujioka, M. Shiro. Chem. Eur. J., 2, 617 (1996).
[16] S. Develay, R. Tripier, M. La Baccon, V. Patinec, G. Serratrice, H. Handel. Dalton Trans., 3016, (2005).
[17] e. Kimura, S. Aoki, T. Koike, M. Shiro. J. Am. Chem. Soc., 119, 3068 (1997).
[18] M. Zulkefeli, T. Sogon, K. Takeda, e. Kimura, S. Aoki. Inorg. Chem., 48, 9567 (2009).
[19] H. Fujioka, S. Kishida, T. Ishizu, M. Shiro, e. Kinoshita, T. Koike. Bull. Chem. Soc. Jpn., 83, 267 (2010).
[20] C. Bazzicalupi, A. Bencini, e. Berni, A. Bianchi, P. Fornasari, C. Giorgi, B. Valtancoli. Inorg. Chem., 43, 6255 (2004).
[21] C. Bazzicalupi, A. Bencini, C. Bonaccini, C. Giorgi, P. Gratteri, S. Moro, M. Palumbo, A. Simionato, J. Sgrignani, C. Sissi,
B. Valtancoli. Inorg. Chem., 47, 5473 (2008).
[22] M. Arca, A. Bencini, e. Berni, C. Caltagirone, F.A. Devillanova, F. Isaia, A. Garau, C. Giorgi, V. Lippolis, A. Perra, L. Tei, B.
Valtancoli. Inorg. Chem., 42, 6929 (2003).
[23] M. Subat, K. Woinaroschy, C. Gerstl, B. Sarkar, W. Kaim, B. Konig. Inorg. Chem., 47, 4661 (2008).
[24] L.F. Lindoy, T. Rambusch, B.W. Skelton, A.H. White. J. Chem. Soc., Dalton Trans., 1857, (2001).
[25] C.A. Chang, H.Y. Lee, C.-L. Chen. Dalton Trans., 42, 6397 (2013).
[26] BIOVIA. Materials studio overview. Available online at: http://accelrys.com/products/materials-studio/index.html.
[27] Gaussian 09, Revision A.1, M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.e. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino,
G. Zheng, J.L. Sonnenberg, M. Hada, M. ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, T. Vreven, J.A.Montgomery Jr., J.e. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, e. Brothers, K.N. Kudin,
V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi,
N. Rega, N.J. Millam, M. Klene, J.e. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.e. Stratmann,
O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth,
P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox.
Gaussian, Inc., Wallingford, CT (2009).
[28] A. Schrodt, A. Neubrand, R. van eldik. Inorg. Chem., 36, 4579 (1997).
[29] Y.W. Ren, J. Li, S.M. Zhao, F.X. Zhang. Struct. Chem., 16, 439 (2005).
[30] S. el Ghachtouli, C. Cadiou, I. Déchamps-Olivier, F. Chuburu, M. Aplincourt, V. Turcry, M. Le Baccon, H. Handel. Eur. J.
Inorg. Chem., 2658, (2005).
[31] D.D. Dischino, e.J. Delaney, J.e. emswiler, G.T. Gaughan, J.S. Prasad, S.K. Srivastava, M.F. Tweedle. Inorg. Chem., 30,
1265 (1991).
[32] M. Le Baccon, F. Chuburu, L. Toupet, H. Handel, M. Soibinet, I. Dechamps-Olivier, J.-P. Barbier, M. Aplincourt. New J.
Chem., 25, 1168 (2001).

18
C. A. CHANG eT AL.
[33] M.T.S. Amorim, J.R. Ascenso, R. Delgado, J.J.R. Fraústo da Silva. J. Chem. Soc., Dalton Trans., 3449, (1990).
[34] H. Fujioka, T. Koike, N. Yamada, e. Kimura. Heterocycles, 42, 775 (1996).
[35] C.S. Rossiter, R.A. Mathews, J.R. Morrow. Inorg. Chem., 44, 9397 (2005).
[36] M.T.S. Amorim, S. Chaves, R. Delgado, J. Dasilva. J. Chem. Soc., Dalton Trans., 3065, (1991).
[37] X. Zhang, H. Gao, H. Xu, J. Xu, H. Chao, C. Zhao. J. Mol. Catal. A: Chem., 368, 53 (2013).
[38] A.L.O. de Noronha, L. Guimarães, H.A. Duarte. J. Chem. Theory Comput., 3, 930 (2007).
[39] V.S. Bryantsev, M.S. Diallo, W.A. Goddard III. J. Phys. Chem. A, 113, 9559 (2009).
[40] N.F. Brás, M.A.S. Perez, P.A. Fernandes, P.J. Silva, M.J. Ramos. J. Chem. Theory Comput., 7, 3898 (2011).
[41] T. Kupka, B. Ruscic, R.e. Botto. J. Phys. Chem., 106, 10396 (2002).
[42] T.C.T. Lo, Y.H. Jiang, A.L.J. Chao, C.A. Chang. Anal. Chim. Acta, 584, 50 (2007).
[43] P. Fuentealba, L. Serón, C. Sánchez, J. Manzur, V. Paredes-Garcia, N. Pizarro, M. Cepeda, D. Venegas-Yazigi, e. Spodine.
J. Coord. Chem., 67, 3894 (2014).
[44] H.P. ebrahimi, J.S. Hadi, Z.A. Abdulnabi, Z. Bolandnazar. Spectrochim. Acta, Part A, 117, 485 (2014).
[45] D.M. Brown, D.A. Usher. J. C. S. (Resumed), 6558, (1965).
[46] C.A. Chang, P.H.L. Chang, S. Qin. Inorg. Chem., 27, 944 (1988).
[47] C.A. Chang, B.H. Wu, B.Y. Kuan. Inorg. Chem., 44, 6646 (2005).