Enabling Mitochondrial Uptake of Lipophilic Dications Using Methylated Triphenylphosphonium Moieties

Article abstract of DOI:10.1021/acs.inorgchem.8b03380

Triphenylphosphonium (TPP⁺) species comprising multiple charges, i.e., bis-TPP⁺, are predicted to be superior mitochondrial-targeting vectors and are expected to have mitochondrial accumulations 1000-fold greater than TPP⁺, the current "gold standard". However, bis-TPP⁺ vectors linked by short hydrocarbon chains (n < 5) are unable to be taken up by the mitochondria, thus hindering their development as mitochondrial delivery vectors. Through the incorporation of methylated TPP⁺ moieties (T*PP⁺), we successfully enabled the accumulation of bis-TPP⁺ with a short linker chain in isolated mitochondria, as measured by high performance liquid chromatography. These experimental results are further supported by molecular dynamics and ab initio calculations, revealing the strong correlations between mitochondria uptake and molecular volume, surface area, and chemical hardness. Most notably, the molecular volume has been shown to be a strong predictor of accumulation for both mono- and bis-TPP⁺ salts. Our study underscores the potential of T*PP⁺ moieties as alternative mitochondrial vectors to overcome low permeation into the mitochondria.

Full text of DOI:10.1021/acs.inorgchem.8b03380

Article
Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/IC
Enabling Mitochondrial Uptake of Lipophilic Dications Using
Methylated Triphenylphosphonium Moieties
†
,#
†,#
‡
‡
§
How Chee Ong, Zhang Hu, Joao T. S. Coimbra, Maria J. Ramos, Oi Lian Kon,
̃
†
†
,‡
,†
Bengang Xing, Edwin K. L. Yeow, Pedro A. Fernandes,* and Felipe García*
†School of Physical and Mathematical Sciences, Division of Chemistry and Biological Chemistry, Nanyang Technological University,
1 Nanyang Link, 637371, Singapore
2
‡
UCIBIO, REQUIMTE, Departamento de Química e Bioquímica, Faculdade de Cien
Alegre s/n, 4169-007, Porto, Portugal
̂
cias, Universidade do Porto, Rua do Campo
§
Division of Medical Sciences, Laboratory of Applied Human Genetics, Humphrey Oei Institute of Cancer Research, National
Cancer Centre, 169610, Singapore
*
S Supporting Information
+
ABSTRACT: Triphenylphosphonium (TPP ) species comprising
+
multiple charges, i.e., bis-TPP , are predicted to be superior
mitochondrial-targeting vectors and are expected to have mitochon-
+
drial accumulations 1000-fold greater than TPP , the current “gold
+
standard”. However, bis-TPP vectors linked by short hydrocarbon
chains (n < 5) are unable to be taken up by the mitochondria, thus
hindering their development as mitochondrial delivery vectors.
+
+
Through the incorporation of methylated TPP moieties (T*PP ),
+
we successfully enabled the accumulation of bis-TPP with a short
linker chain in isolated mitochondria, as measured by high
performance liquid chromatography. These experimental results
are further supported by molecular dynamics and ab initio
calculations, revealing the strong correlations between mitochondria
uptake and molecular volume, surface area, and chemical hardness. Most notably, the molecular volume has been shown to be a
+
+
strong predictor of accumulation for both mono- and bis-TPP salts. Our study underscores the potential of T*PP moieties as
alternative mitochondrial vectors to overcome low permeation into the mitochondria.
+
INTRODUCTION
The uptake of TPP by the mitochondrion is modeled by
■
1
0,20
the Nernst equation,
which predicts the equilibrium
Mitochondria organelle are well-known for their pivotal role in
+
1
concentration of bis-TPP to be 1000-fold greater than
mono-TPP (−180 mV), thus suggesting enhanced mitochon-
drial targeting properties. In addition, although mono-TPP
aerobic respiration and other essential cellular functions, such
+
2
,3
4
as cell signaling
and apoptosis. Furthermore, aberrant
21
+
mitochondria has been linked to many neurodegenerative and
cardiovascular diseases such as Alzheimer’s disease, athero-
sclerosis, amyotrophic lateral sclerosis (ALS), and cancer.
Hence, as mitochondria is increasingly recognized to play a
critical role in human disease, it is vital to develop more
efficient vectors that selectively target mitochondria.
Typically, mitochondrial targeting vectors are positively
charged species that selectively accumulate within the organelle
due to the large negative membrane potential across the inner
22,23
can show selectivity between cancer and noncancer cells,
bis-TPP+ species are more sensitive to small changes in
5
−9
24
membrane potentials, which may confer enhanced selectivity.
For instance, cancerous cells display higher |Δψ| (ca. 60 mV)
than normal cells, which corresponds to a 10-fold increase in
+
selectivity (relative to mono-TPP ) according to the Nernst
25,26
equation.
However, an increased cationic charge elevates
+
the kinetic barrier for the transport of bis-TPP across
4
27,28
mitochondrial membrane (Δψ = −150−180 mV). Common
examples include rhodamines, cyanines, mitochondrion-target-
ing peptides, and triphenylphosphonium (TPP ) salts.
phospholipid bilayers,
a principle corroborated by the
work by Murphy et al., who demonstrated that incorporation
of methylene bridges comprising less than five carbon atoms
resulted in the unsuccessful mitochondrial uptake of bis-TPP
+
10−13
+
+
The most extensively studied, “gold standard” vector, TPP ,
boasts ease of synthesis, stability in biological systems, and
facile conjugation to molecular cargo, and has been used to
successfully deliver a wide range of molecular cargo into
21
(Figure 1).
Received: December 14, 2018
1
4−19
mitochondria.
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.8b03380
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
crystallization in a mixture of acetonitrile and diethyl ether
(
isolated crystalline product yields between 65 and 70%).
Nuclear magnetic resonance (NMR) spectroscopic analysis
reveals the successful formation of the phosphonium salts, as
shown by the characteristic downfield shifts relative to the free
3
1
1
phosphine in the P{ H} NMR spectra (−5.4, −7.9, −5.8
3
0,31
ppm
vs 25.2, 24.0, and 24.6 ppm for 1−3 respectively),
consistent with the chemical shifts reported in the
literature.
2
1,29
13
1
The C{ H} NMR spectra confirms the
successful linkage of two phosphonium moieties on the same
molecule, as seen from the ABX splitting patterns observed
(
Cortho, Cmeta, and P−Calkyl) as a result of second-order spin
+
21
Figure 1. Previously reported bis-TPP cations: (top) Murphy et al.
and (bottom) Lisanti et al.
coupling.
35
The single-crystal structural data show that the P−C bond
Ar
lengths ranges from 1.789−1.802 Å, 1.785−1.801 Å, and
+
In light of the apparent inability of bis-TPP salts linked by
1.790−1.806 Å, while the P−C ranges from 1.796−1.801 Å,
alkyl
short methylene linkers (n ≤ 5) to accumulate, we set out to
1.792−1.793 Å and 1.789−1.804 Å for compounds 1, 2, and 3
respectively. This is highly consistent with the bond lengths
expected of alkyltriarylphosphonium P−C bonds at 1.796
+
28
study the effect of T*PP moieties on the mitochondrial
+
uptake of short-chained bis-TPP dicationic species. In stark
32
contrast with their previously reported TPP counterparts, we
report the successful accumulation of dicationic bis-T*PP+
species containing short linker chains. This work demonstrates
(±0.017) Å for P−C and 1.798 (±0.031) Å for P−C
.
Ar
Alkyl
The bond angles for C−P−C bonds for all three compounds
ranges from 105.10 to 113.56°, which is consistent with a
+
33
that the use of T*PP moieties allows for the rational fine-
tetrahedral geometry expected of phosphonium moieties.
tuning of steric and electronic parameters toward their
accumulation in mitochondria and represents a straightforward
strategy to readily enhance mitochondrial uptake of dicationic
species.
Partition Coefficient and Mitochondrial Uptake of
Compounds 1−3. The octanol−water partition coefficients
(
P) for compounds 1−3 (i.e., ratio of the concentration of
dication in 1-octanol to the concentration of dication in water)
were determined to establish their relative lipophilicities
(Table 1), which has been shown to influence membrane
RESULTS AND DISCUSSION
■
34
Synthesis and Structure of Compounds 1−3. In our
study, we based the initial choice of phosphine starting
materials on previously reported work, where T*PP+
mitochondrial delivery vectors comprising p-tolyl or 3,5-
dimethylphenyl trisubstituted phosphonium moieties signifi-
cantly increase the accumulation of both the T*PP cations
and their conjugated cargo into the mitochondria organelle.
permeability. The observed trend indicates that the bis-
T*PP+ dication 3, containing dimethyl substituted phenyl
rings, has the greatest lipophilicity, followed by methyl
substituted dication 2 and last compound 1, which comprises
+
TPP moieties (Table 1 and Figure 3).
+
In order to evaluate if the hydrophobic variation observed
for 1−3 is translated to differences in half-maximal inhibitory
concentrations (IC ), their 72-h IC values in HeLa cells
29
Dications 1−3 were obtained in quantitative spectroscopic
yield upon reacting the respective phosphines with 1,3-
dibromopropane in acetonitrile under reflux (Scheme 1).
5
0
50
were determined. It was observed that with an increasing
degree of methylation, the IC50 decreased from 180 μM to 21
μM and 6 μM for 1 to 3, respectively, suggesting increasing
mitochondrial accumulation with increasing degree of
substitutionconsistent with their monocationic counter-
+
+
Scheme 1. Synthesis of Bis-TPP and Bis-T*PP salts 1−3
2
9
parts. Moreover, a reduction in IC values for mono- and
5
0
+
dicationic TPP species has been recently linked to reduced
ATP levels (i.e., suggestive of increased mitochondrial
+
accumulation) in studies proposing the use of TPP and bis-
TPP+ cations as a novel strategy toward the selective
35
eradication of cancer stem cells (CSCs).
To rule out any differences in intrinsic toxicity between the
dications 1−3and to further corroborate if the variation in
IC values was caused by enhanced mitochondrial uptake
5
0
In agreement with previous observations for monocationic
the accumulation in isolated mitochondria (C ) was
m
+
29
T*PP species, attempts to obtain dicationic species with
more heavily substituted phosphines, such as trimesitylphos-
phine, were unsuccessful. Compounds 1−3 were purified by
established (see Supporting Information).
The trend of Cm values obtained shows a significant
improvement in the rate of uptake with increasing methylation,
Table 1. Summary of Experimental and Computational Results for Compounds 1, 2, and 3
C /mmol g‑
1
volume/Å3
SASA/Å2
ΔGaq→oct /kcal mol‑1 ΔΔGaq→oct
exp
vdW
/kcal mol‑1 η/eV
compound
log P
IC /μM
5
0
m
2
1
3
−0.12 ± 0.03
−1.62 ± 0.13
0.70 ± 0.08
21.3 ± 1.6
181.3 ± 7.6
5.6 ± 0.3
5.2 ± 0.4
0.0
20.3 ± 1.2
1862.7 ± 1.5
1570.9 ± 1.1
2120.0 ± 1.6
1014.5 ± 1.0
860.3 ± 0.8
1108.9 ± 1.0
0.16 ± 0.04
2.21 ± 0.18
−0.95 ± 0.11
−5.56 ± 0.30
0.00 ± 0.35
−7.94 ± 0.45
3.65
3.79
3.52
B
DOI: 10.1021/acs.inorgchem.8b03380
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 2. Selected bond lengths (Å) and angles (deg) of compound 1 C21−P1 1.800(6) C22−P1 1.798(6) C21−P1−C22 110.8(3) C22−P1−
C28 108.9(3); Compound 2 C22−P1 1.789(10) C1−P1 1.799(10) C1−P1−C22 110.9(5) C1−P1−C8 107.0(4); Compound 3 C25−P1
1
.793(4) C1−P1 1.800(8) C1−P1−C25 113.6(5) C1−P1−C9 105.2(7). Thermal ellipsoids are shown at the 50% probability level. Hydrogen
atoms and bromide anions have been omitted for clarity.
found to be proportional to the accessible surface area of the
2
28,37
cation (r ).
Although the ionic radius is commonly used to calculate the
Born and hydrophobic free energy, compounds 1−3 are far
from spherical. Hence, using an average radius to calculate the
kinetic barrier of transport for a nonspherical ion might render
the results obtained inaccurate. Moreover, previous reports of
+
substituted TPP moieties as mitochondrial delivery agents
were mostly limited to biological activity quantification, while
the ones describing structural correlation with accumulation
did not consider nonspherical effects.
Figure 3. Experimental results obtained for log P, IC , and C . Error
bars (±) refer to the standard deviation (SD).
5
0
m
35,39−42
Solvent Accessible Surface Area (SASA) and Molec-
ular Volume (V). For this purpose, molecular dynamics (MD)
calculations were carried out (see Supporting Information) to
determine the impact of methylation on the volume (V) and
solvent accessible surface area (SASA) as an alternative
parameter to molecular radius. MD simulations revealed that
as the number of methyl groups increases, a larger (hydro-
phobic) volume (and SASA) is accessible to the solvent
decreasing the polar interactions with the molecules of water,
resulting in an increasingly poorer cation solvation. On the
other hand, the larger nonpolar volume enhances the level of
nonpolar interactions leading to a favorable partition into the
organic solvent (i.e., more negative ΔGaq→oct), as seen from the
from negligible uptake of 1 to 5.2 mmol g− and 20.3 mmol g−1
in 2 and 3, respectively. The negligible accumulation of
compound 1 is in agreement with reports by Murphy et al.
1
2
1
(
vide supra). The experimental trend suggests that the
increased IC50 values is not due to the intrinsic cytotoxicity of
the compounds but is rather due to its enhanced accumulation
within the HeLa cells.
Furthermore, the supernatant concentration of compound
3
the most accumulating compound presented in our
studywas measured using HPLC after treatment of the
mitochondria with carbonyl cyanide p-trifluoromethoxyphe-
nylhydrazone (FCCP), to abolish the membrane potential
across the mitochondrial membrane, resulting in the efflux of
compound 3 (see Supporting Information). This indicates that
2
strong linear correlation (r = 0.986 and 0.999) for V and
SASA, respectively (See Figure 4).
3
does not accumulate in the mitochondria when Δψ is
While the relationship between mitochondrial accumulation
and nonpolar molecular SASA has been experimentally
abolished, and the accumulation of 3 is thus driven by the
presence of the membrane potential.
+
29
established in T*PP compounds, and is also reflected in
the case of bis-T*PP+ molecules, the strong correlation
observed between molecular volume (whose surrogates are
surface and radius) and mitochondrial accumulation has never
been described. Furthermore, the importance of molecular
volume as a meaningful parameter to predict mitochondrial
accumulation for multiply charged TPP+ species will be
discussed further in the text (vide infra).
THEORETICAL STUDIES
■
To better understand the mechanism underlying the observed
increase in mitochondrial uptake, we considered the
corresponding membrane transport kinetics. The kinetic
barrier for the permeation of small charged molecules across
a phospholipid membrane in an aqueous environment can be
approximated by the free energy of transport from the aqueous
Neutral and Electrostatic Components for the Free
Energy of Transfer. Thermodynamic integration (TI)
calculations were used to quantify the nonpolar component
(i.e., vdW) of the relative free energy of transfer of compounds
36
phase into the center of the lipid membran, and is hence
closely linked to lipophilicity. This barrier mainly comprises
the Born and hydrophobic (or neutral) free energies,
which can be treated as two independent components.
21
28,37
34,38
vdw
1−3 from water to 1-octanol (ΔΔGaq→oct ) taking com-
2
The Born free energy, W = 339(Z /r), the most significant
pound 1 as the reference (Table 1). Similar to previous reports
on monocationic species, initial TI results showed that the
electrostatic component (i.e., Born component) of compounds
1−3 showed negligible variation (see Supporting Information
B
electronic component of this kinetic barrier, is directly
proportional to the square of the ionic charge (Z), and
inversely proportional to the ionic radius (r) for a spherical
ion, whereas the hydrophobic free energy was empirically
29
for details). Thus, the results presented were obtained by
C
DOI: 10.1021/acs.inorgchem.8b03380
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 2. Vertical Ionization Energy (IP), Electron Affinity
+
(
EA), and Hardness Values for the Three Bis-TPP Cations
in This Work (Cations 1, 2, and 3) and for the Previously
Reported Bis-TPP+ Cations by Murphy et al. (PC2P, PC4P,
a
PC5P, PC6P, and PC10P)
cation
IP/eV
EA/eV
hardness/eV
PC2P
1
PC4P
PC5P
PC6P
2
13.45
13.20
12.98
12.81
12.64
12.58
12.15
12.25
5.85
5.63
5.40
5.24
5.11
5.29
4.78
5.20
3.80
3.79
3.79
3.78
3.76
3.65
3.69
3.52
PC10P
3
a
+
For the bis-TPP cations proposed by Murphy et al., we have used
the same nomenclature as in the original work. Compounds are
ordered in increasing molecular weight.
expt
Figure 4. Correlation graphs between (a) ΔG
and
aq→oct
vdW
ΔΔG
against log P (b) IC and C against hardness, (c)
volume and SASA against log P, (d) log P and volume against
properties of 1−3 in relation to their mitochondrial uptake
trends.
aq→oct
50
m
hardness. Error bars (±) refers to the standard deviation (SD).
The ESP maps (see Figure 5) revealed that the electron
+
density located in the volume between the two P atoms
increases with increased substitution. The presence of electron-
donating methyl groups increasingly promotes the dissipation
of the positive charge across the molecule from 1 to 3 and
renders the methylated compounds more lipophilic in nature.
This agrees with the calculated chemical hardness values,
which decrease from 3.79, 3.65, to 3.52 eV in 1−3,
respectively. Chemical hardness is negatively correlated with
the mitochondrial uptake, indicating that softening the
phosphonium cation improves the membrane permeability
and hence the mitochondrial accumulation.
considering only the nonelectrostatic components. The results
vdw
show a strong negative correlation between ΔΔG
and
aq→oct
the mitochondrial uptake, with decreasing free energy of
transport from compounds 1 to 3 (0.00, −5.56, and −7.94 kcal
−
1
mol , respectively).
The thermodynamic integration calculations trends were
exp
consistent with the ΔG
calculated from log P (ΔG =
aq→oct
−
2.303 RT log P) (Figure 4 and Table 1). The calculated
decrease in kinetic barrier to permeate the membrane can be
attributed to the increased number of nonpolar substituents
The same phenomenon could also be observed from bis-
+
+
21
present on bis-T*PP species (i.e., 2 and 3) as compared to
TPP previously investigated by Murphy and co-workers.
+
+
bis-TPP . This increased affinity of the bis-T*PP for the
hydrophobic phase lowers the permeation energy barrier
resulting in greater mitochondrial uptake, which is in
agreement with our experimentally observed Cm and IC50
values. Taken together, these findings suggest that the use of
Further calculations revealed that the increase of the alkyl
chain length was also accompanied by a decrease in chemical
21
hardness (see Table 2) and lipophilicity. Furthermore, it can
be observed from the calculated value for all eight dications
that the chemical hardness decreased with increasing molecular
weight. However, dication 2 displays a lower chemical
hardness than PC P, which comprises one additional carbon
+
T*PP moietieswith varying levels of substitutionis a
highly effective approach to improving the effectiveness of bis-
1
0
+
TPP species as mitochondrial vectors.
atomsuggesting methyl substitution as a more effective
approach to dissipating the positive charge on this species than
the commonly used chain extension approach.
Electrostatic Potential and Chemical Hardness for
Compounds 1−3 and Previously Reported Dications.
To examine the changes in the electronic interactions due to
methylation, density functional theory (DFT) calculations
These results therefore suggest that lipophilicity may be
tunable and improved from structural modifications targeted to
reducing charge density as well as the HOMO−LUMO gap for
(
B3LYP/6-31+G(d)) were performed, determining the elec-
+
trostatic potential (ESP) maps and chemical hardness of 1−3.
Electrostatic potential maps depict the local electron density
distribution, serving as a measure of the available polar surface
T*PP compounds.
+
+
Unifying Mono-T*PP and Bis-T*PP Accumulation:
Volume Per Charge Unit (V ). Comparison between the
PC
14
+
29
+
area for polar interactions, and it has been observed from
previously reported mono-T*PP
and bis-T*PP 1−3
2
1
14
works of Murphy and Hartley that a reduction in the
electrostatic potential in cationic species can be correlated with
an increase in membrane permeation.
species revealed that the correlation between accumulation of
+
TPP salts and lipophilicity was not satisfactory. Although the
trends between accumulation and lipophilicity are internally
+
The chemical hardness serves as an indicator of electron
cloud polarizability, and the strength of hydrophobic
consistent within each type of TPP salts (mono- or bis-
+
T*PP ), the linear correlation cannot be generalized to both
4
3
interactions, and was obtained for 1−3 from the vertical
ionization potentials (IP) and electron affinities (EA),
calculated at the B3LYP/6-311+G(2df,p) level, with the
following expression: η = (IP-EA)/2 (see Table 2). We herein
utilize these parameters as intrinsic molecular descriptors to
correlate and further our basic understanding of the electronic
(Figure 6a). To bridge the gap between the differing
+
accumulations in mono and bis-T*PP , we investigated the
new structural parameters described above (V/SASA/η) and
their correlation with C . As η correlates well with logP
m
(Figure 4d), two separate trendlines (similar to the correlation
with logP) were observed. Thus, a satisfactory correlation
D
DOI: 10.1021/acs.inorgchem.8b03380
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 5. ESP maps (B3LYP/6-31G(d)) for compounds 1, 2, and 3. Regions with high electron density (less positive electrostatic potential) are
colored red, and regions with low electron density (more positive electrostatic potentials) are colored blue. Yellow and green correspond to3
intermediate electrostatic potentials. The same color scale was used: from 0.115 (red) to 0.238 au (dark blue). An isodensity value of 0.0020 e a0−
was used to define the outer surface of the molecules on which the ESP was calculated.
Table 3. Solvent Accessible Surface Area (SASA) and
Volume (V) Computational Results for Compounds 1−3
+
and the Previously Reported Bis-TPP Cations by Murphy
a
et al. (PC P, PC P, PC P, PC P, and PC P) in water
2
4
5
6
10
cation
SASA/Å2
V/Å3
mitochondrial accumulation
PC2P
1
PC4P
PC5P
PC6P
2
855.0 ± 0.8
860.3 ± 0.8
932.6 ± 1.1
968.4 ± 1.0
985.0 ± 1.3
1014.5 ± 1.0
1110.0 ± 1.3
1108.9 ± 1.0
1534.2 ± 1.1
1570.9 ± 1.1
1650.9 ± 1.2
1708.4 ± 1.2
1750.9 ± 1.5
1862.7 ± 1.5
1964.3 ± 1.6
2120.0 ± 1.0
NO
NO
NO
YES
YES
YES
YES
YES
Figure 6. Correlations plots between (a) C against logP and (b) C
m
m
+
against volume/charge for mono- and bis-T*PP . Error bars
PC P
+
10
correspond to the SD. Mono- and bis-T*PP are colored blue and
3
green, respectively.
a
+
For the bis-TPP cations proposed by Murphy et al., we have used
the same nomenclature as in the original work.
between mono- and bis-T*PP+ was not possible (see
Supporting Information, Figure S4).
Notably, the molecular volume, on the other hand, upon
normalization with respect to charge (V/Z)volume per
Quantitative Structure−Activity Relationship (QSAR)
Models. In order to confirm that the mitochondrial uptake is
consistent with the current QSAR models, we considered the
charge (V )provided a generalizable linear correlation
PC
(Figure 6b). Although the volume and SASA follows the
4
4−46
same general trend, we focused on volume due to the slightly
model developed by Horobin et al.,
which relates
2
higher r values. The experimentally obtained equationV
=
physiochemical parameters of compounds to the intracellular
localization. These parameters are used to determine the
subcellular localization of compounds and thus can be used to
predict whether the compounds will be mitochondria-
targeting. For mitochondrial localization, the model considers
the following parameters: amphiphilic index (AI), conjugated
PC
8
.06*C + 852underscores a strong relationship between
m
+
the mitochondrial accumulation of T*PP species and their
V . Furthermore, the data obtained suggests the existence of a
PC
min
3
minimum VPC (VPC ) threshold of 852 Å required for the
effective permeation of T*PP species into mitochondria
organelle.
+
bond number (CBN), log P, pK , and electric charge (Z). For
a
To further assess the validity of the obtained equation, the
volume and SASA of compounds studied by Murphy et al. (i.e.,
PC P, PC P, PC P, PC P, PC P) were calculated and
compounds 1−3, the pK and AI are not considered due to a
a
lack of acidic protons and amphiphilicity. As such, for selective
mitochondria accumulation, the relevant criteria for com-
pounds 1−3 to satisfy are CBN < 40, 5 > logP > 0, Z > 0.
From the structures of compounds 1−3, the key
determining factor for mitochondria targeting lies with the
logP value. All three compounds have an identical CBN and Z
value, and they satisfy the criteria for CBN and Z as previously
described (CBN = 18, Z = 2). For the lipophilicity, compound
1 falls far below the logP requirement with a logP of −1.62,
resulting in negligible mitochondrial accumulation, as
predicted by the QSAR model. Compound 3, on the other
hand, had a logP of 0.70, which fell within the range of 0 <
logP < 5, which resulted in a high mitochondrial accumulation.
Compound 2 had a logP of −0.12, which was slightly lower
than the lower limit, resulting in a poor mitochondrial uptake,
as reflected by its Cm value. The uptake observed for the
2
4
5
6
10
compared with the values obtained for compounds 1−3 (see
Table 3). From the experimentally obtained equation, the
min
) required for a bis-TPP+
minimum permeation volume (V
PC
3
is calculated to be 1704 Å . Thus, it is predicted that PC P (V
1650.9 Å ) would not accumulate in the mitochondria, while
PC P (V = 1708.4 Å ) would be able to do so. This is
4
3
=
3
5
consistent with their previously reported experimental
2
1
results, demonstrating that the herein reported methodology
is robust. However, it should be noted that this relationship
may not hold if the molecular structure varies significantly
+
from that of TPP salts presented in this work. Overall, these
findings suggest that the volume parameter could be useful for
the prediction of mitochondrial accumulation, especially for
+
multivalent TPP compounds.
E
DOI: 10.1021/acs.inorgchem.8b03380
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
compounds studied thus conforms well and is consistent with
current QSAR models.
This model shows that bis-TPP linked by short chains do
possess the necessary characteristics for mitochondrial
accumulation. However, these systems can be limited by the
lipophilicity, as corroborated by both the QSAR model, as well
ORCID
T. S. Coimbra: 0000-0001-9138-7498
Joao
̃
+
Maria J. Ramos: 0000-0002-7554-8324
Bengang Xing: 0000-0002-8391-1234
Edwin K. L. Yeow: 0000-0003-0290-4882
Felipe García: 0000-0002-9605-3611
Author Contributions
as the C values for compounds 1−3. This further supports the
m
strategy of phenyl methylation, as it can be used to overcome
this logP limitation by boosting the logP into the optimal
range, thus enabling more efficient accumulation within the
mitochondria.
#
H.C.O. and Z.H. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
F.G. would like to thank A*STAR AME IRG A1783c0003,
NTU for a start-up grant (M4080552) and MOE Tier 1 grants
RG 11/15 and RG 113/16) for financial support. P.A.F.,
M.J.R., and J.T.S.C. acknowledge financial support from the
European Union (FEDER funds POCI/01/0145/FEDER/
07728) and National Funds (FCT/MEC, Fundaca
Ciencia e Tecnologia and Ministerio da Educaca e Cien
under the Partnership Agreement PT2020 UID/MULTI/
4378/2013.
CONCLUSIONS
■
■
+
In summary, we report the use of T*PP moieties to address
+
the inability of TPP dications comprising short chain linkers
(
to be taken up by the mitochondria organelle. The
incorporation of nonpolar methyl groups increases the volume
+
and SASA of the resulting bis-T*PP saltswhile dispersing
0
̧
o
̃
para a
their high charge density and decreasing their chemical
hardnesswhich in turn enhances their lipophilicity and
mitochondrial accumulation. Furthermore, ESP maps and
chemical hardness calculations could serve as useful tools to
̂
́
̧
o
̃
̂
cia)
0
+
predict the efficacy of TPP mitochondrial vectors in silico.
REFERENCES
These results demonstrate that substituents on the phenyl
rings increase the volume and that mitochondrial accumulation
can be readily enhanced by fine-tuning these properties
through structural modifications. In particular, molecular
volume has been shown to be a key parameter for the+
prediction of compounds incorporating multiple TPP
moieties. From our observations across various mono- and
■
(
1) Kuhlbrandt, W. Structure and function of mitochondrial
membrane protein complexes. BMC Biol. 2015, 13, 89−99.
2) Rizzuto, R.; De Stefani, D.; Raffaello, A.; Mammucari, C.
Mitochondria as sensors and regulators of calcium signalling. Nat. Rev.
Mol. Cell Biol. 2012, 13 (9), 566−578.
3) Chandel, N. S. Mitochondria as signaling organelles. BMC Biol.
(
(
2
014, 12 (1), 34−40.
+
bis-TPP , we propose the minimum permeation volume per
(4) Murphy, M. P.; Smith, R. A. Drug delivery to mitochondria: the
key to mitochondrial medicine. Adv. Drug Delivery Rev. 2000, 41 (2),
235−250.
min
charge (VPC ), which can be used as a new structural
parameter for the prediction of mitochondrial accumulation for
+
(
5) Di Lisa, F.; Kaludercic, N.; Carpi, A.; Menabo, R.; Giorgio, M.
TPP molecules.
Mitochondria and vascular pathology. Pharmacol. Rep. 2009, 61 (1),
Our work not only highlights the use of readily tunable
molecular descriptors as key parameters to design future TPP+
mitochondrial delivery vectors with enhanced properties,
123−130.
(
6) Frantz, M. C.; Wipf, P. Mitochondria as a target in treatment.
Environ. Mol. Mutagen. 2010, 51 (5), 462−475.
7) Reddy, P. H. Mitochondrial medicine for aging and neuro-
degenerative diseases. NeuroMol. Med. 2008, 10 (4), 291−315.
8) Celsi, F.; Pizzo, P.; Brini, M.; Leo, S.; Fotino, C.; Pinton, P.;
+
thereby enhancing bioavailability of TPP -conjugated drugs,
(
but also illustrates the need for a broader family of T*PP+
species for future enhanced mitochondrial therapies.
(
Rizzuto, R. Mitochondria, calcium and cell death: a deadly triad in
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.8b03380.
neurodegeneration. Biochim. Biophys. Acta, Bioenerg. 2009, 1787 (5),
■
3
(
35−344.
*
S
9) Haelterman, N. A.; Yoon, W. H.; Sandoval, H.; Jaiswal, M.;
Shulman, J. M.; Bellen, H. J. A mitocentric view of Parkinson’s
disease. Annu. Rev. Neurosci. 2014, 37 (1), 137−159.
(10) Zielonka, J.; Joseph, J.; Sikora, A.; Hardy, M.; Ouari, O.;
Vasquez-Vivar, J.; Cheng, G.; Lopez, M.; Kalyanaraman, B.
Mitochondria-Targeted Triphenylphosphonium-Based Compounds:
Syntheses, Mechanisms of Action, and Therapeutic and Diagnostic
Applications. Chem. Rev. 2017, 117 (15), 10043−10120.
Experimental procedures, details for MD and DFT
calculations, NMR/HRMS spectra for compounds 1−3
(PDF)
(
11) Johnson, L. V.; Walsh, M. L.; Chen, L. B. Localization of
Accession Codes
mitochondria in living cells with rhodamine 123. Proc. Natl. Acad. Sci.
U. S. A. 1980, 77 (2), 990−994.
CCDC 1478287−1478289 contain the supplementary crys-
tallographic data for this paper. These data can be obtained
free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by
emailing data_request@ccdc.cam.ac.uk, or by contacting The
Cambridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033.
(12) Laris, P. C.; Bahr, D. P.; Chaffee, R. R. J. Membrane-Potentials
in Mitochondrial Preparations as Measured by Means of a Cyanine
Dye. Biochim. Biophys. Acta, Bioenerg. 1975, 376 (3), 415−425.
(13) Szeto, H. H. Cell-permeable, mitochondrial-targeted, peptide
antioxidants. AAPS J. 2006, 8 (2), E277−E283.
(14) Robb, E. L.; Gawel, J. M.; Aksentijevic, D.; Cocheme, H. M.;
Stewart, T. S.; Shchepinova, M. M.; Qiang, H.; Prime, T. A.; Bright,
T. P.; James, A. M.; Shattock, M. J.; Senn, H. M.; Hartley, R. C.;
Murphy, M. P. Selective superoxide generation within mitochondria
by the targeted redox cycler MitoParaquat. Free Radical Biol. Med.
2015, 89, 883−894.
AUTHOR INFORMATION
Corresponding Authors
(F.G.) E-mail: fgarcia@ntu.edu.sg.
(P.A.F.) E-mail: paferman@fc.up.pt.
■
*
*
F
DOI: 10.1021/acs.inorgchem.8b03380
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(
15) Jameson, V. J.; Cocheme, H. M.; Logan, A.; Hanton, L. R.;
the Cambridge Structural Database and visualizing crystal structures.
Acta Crystallogr., Sect. B: Struct. Sci. 2002, 58 (3−1), 389−397.
(33) Corbridge, D. E. Phosphorus: Chemistry, Biochemistry and
Technology; CRC Press, 2016.
Smith, R. A.; Murphy, M. P. Synthesis of triphenylphosphonium
vitamin E derivatives as mitochondria-targeted antioxidants. Tetrahe-
dron 2015, 71 (44), 8444−8453.
(
16) Millard, M.; Gallagher, J. D.; Olenyuk, B. Z.; Neamati, N. A
(34) Winiwarter, S.; Ridderstrom, M.; Ungell, A. L.; Andersson, T.
̈
selective mitochondrial-targeted chlorambucil with remarkable
cytotoxicity in breast and pancreatic cancers. J. Med. Chem. 2013,
B.; Zamora, I. In Comprehensive Medicinal Chemistry II; Triggle, D. J.,
Ed.; Elsevier: Oxford, 2007; pp 531−554.
5
(
6 (22), 9170−9179.
(35) Ozsvari, B.; Sotgia, F.; Lisanti, M. P. Exploiting mitochondrial
targeting signal(s), TPP and bis-TPP, for eradicating cancer stem cells
(CSCs). Aging 2018, 10 (2), 229−240.
17) Finichiu, P. G.; Larsen, D. S.; Evans, C.; Larsen, L.; Bright, T.
P.; Robb, E. L.; Trnka, J.; Prime, T. A.; James, A. M.; Smith, R. A.;
Murphy, M. P. A mitochondria-targeted derivative of ascorbate:
MitoC. Free Radical Biol. Med. 2015, 89, 668−678.
(36) Ketterer, B.; Neumcke, B.; Lauger, P. Transport Mechanism of
Hydrophobic Ions through Lipid Bilayer Membranes. J. Membr. Biol.
1971, 5 (3), 225−245.
(
18) Kelso, G. F.; Porteous, C. M.; Coulter, C. V.; Hughes, G.;
Porteous, W. K.; Ledgerwood, E. C.; Smith, R. A.; Murphy, M. P.
Selective targeting of a redox-active ubiquinone to mitochondria
within cells: antioxidant and antiapoptotic properties. J. Biol. Chem.
(37) Ross, M. F.; Kelso, G. F.; Blaikie, F. H.; James, A. M.;
Cocheme, H. M.; Filipovska, A.; Da Ros, T.; Hurd, T. R.; Smith, R. A.
J.; Murphy, M. P. Lipophilic triphenylphosphonium cations as tools in
mitochondrial bioenergetics and free radical biology. Biochemistry
2
(
001, 276 (7), 4588−4596.
2
(
005, 70 (2), 222−230.
19) Kaur, A.; Brigden, K. W. L.; Cashman, T. F.; Fraser, S. T.; New,
38) El Tayar, N.; Testa, B.; Carrupt, P. A. Polar intermolecular
E. J. Mitochondrially targeted redox probe reveals the variations in
oxidative capacity of the haematopoietic cells. Org. Biomol. Chem.
interactions encoded in partition coefficients: an indirect estimation
of hydrogen-bond parameters of polyfunctional solutes. J. Phys. Chem.
2
(
015, 13 (24), 6686−6689.
1
(
992, 96 (3), 1455−1459.
20) Murphy, M. P.; Smith, R. A. Targeting antioxidants to
39) Haslop, A.; Wells, L.; Gee, A.; Plisson, C.; Long, N. One-pot
mitochondria by conjugation to lipophilic cations. Annu. Rev.
multi-tracer synthesis of novel (18)F-labeled PET imaging agents.
Pharmacol. Toxicol. 2007, 47, 629−656.
Mol. Pharmaceutics 2014, 11 (11), 3818−3822.
(
21) Ross, M. F.; Da Ros, T.; Blaikie, F. H.; Prime, T. A.; Porteous,
(40) Zhang, Z.; Jenni, S.; Zhang, C.; Merkens, H.; Lau, J.; Liu, Z.;
C. M.; Severina, I. I.; Skulachev, V. P.; Kjaergaard, H. G.; Smith, R. A.;
Murphy, M. P. Accumulation of lipophilic dications by mitochondria
and cells. Biochem. J. 2006, 400 (1), 199−208.
Perrin, D. M.; Benard, F.; Lin, K. S. Synthesis and evaluation of
18)F-trifluoroborate derivatives of triphenylphosphonium for my-
ocardial perfusion imaging. Bioorg. Med. Chem. Lett. 2016, 26 (7),
675−1679.
41) Chen, S.; Zhao, Z.; Zhang, Y.; Fang, W.; Lu, J.; Zhang, X. Effect
(
(
22) Morrison, D. E.; Aitken, J. B.; de Jonge, M. D.; Issa, F.; Harris,
1
(
H. H.; Rendina, L. M. Synthesis and Biological Evaluation of a Class
of Mitochondrially-Targeted Gadolinium(III) Agents. Chem. - Eur. J.
of methoxy group position on biological properties of (18)F-labeled
2
(
014, 20 (50), 16602−16612.
benzyl triphenylphosphonium cations. Nucl. Med. Biol. 2017, 49, 16−
23) Ucar, E.; Seven, O.; Lee, D.; Kim, G.; Yoon, J.; Akkaya, E. U.
2
(
3.
Selectivity in Photodynamic Action: Higher Activity of Mitochondria
Targeting Photosensitizers in Cancer Cells. ChemPhotoChem 2019, 3
42) While increased molecular radius from additional methyl
groups can be approximated from the molecular volume (assuming a
hypothetical sphere of the same volume), this approach is incapable of
reproducing the effect of adding individual methyl groups on a
nonspherical molecular scaffold.
(
3), 129−132, DOI: 10.1002/cptc.201800231.
(
24) Logan, A.; Pell, V. R.; Shaffer, K. J.; Evans, C.; Stanley, N. J.;
Robb, E. L.; Prime, T. A.; Chouchani, E. T.; Cocheme, H. M.;
Fearnley, I. M.; Vidoni, S.; James, A. M.; Porteous, C. M.; Partridge,
L.; Krieg, T.; Smith, R. A.; Murphy, M. P. Assessing the
Mitochondrial Membrane Potential in Cells and In Vivo using
Targeted Click Chemistry and Mass Spectrometry. Cell Metab. 2016,
(43) Ghanty, T. K.; Ghosh, S. K. Correlation between Hardness,
Polarizability, and Size of Atoms, Molecules, and Clusters. J. Phys.
Chem. 1993, 97 (19), 4951−4953.
(44) Horobin, R. W. In Mitochondrial Medicine: Vol. II, Manipulating
2
(
3 (2), 379−385.
Mitochondrial Function; Weissig, V., Edeas, M., Ed.; Springer: New
York, NY, 2015; pp 13−23.
25) Hu, Q.; Gao, M.; Feng, G.; Liu, B. Mitochondria-targeted
cancer therapy using a light-up probe with aggregation-induced-
emission characteristics. Angew. Chem., Int. Ed. 2014, 53 (51),
(45) Horobin, R. W.; Rashid-Doubell, F.; Pediani, J. D.; Milligan, G.
Predicting small molecule fluorescent probe localization in living cells
using QSAR modeling. 1. Overview and models for probes of
structure, properties and function in single cells. Biotech. Histochem.
1
4225−14229.
26) Chen, L. B. Mitochondrial membrane potential in living cells.
Annu. Rev. Cell Biol. 1988, 4 (1), 155−181.
27) Flewelling, R. F.; Hubbell, W. L. The membrane dipole
(
2
(
013, 88 (8), 440−460.
(
46) Horobin, R. W.; Rashid-Doubell, F. Predicting small molecule
potential in a total membrane potential model. Applications to
hydrophobic ion interactions with membranes. Biophys. J. 1986, 49
fluorescent probe localization in living cells using QSAR modeling. 2.
Specifying probe, protocol and cell factors; selecting QSAR models;
predicting entry and localization. Biotech. Histochem. 2013, 88 (8),
(
2), 541−52.
(
28) Honig, B. H.; Hubbell, W. L.; Flewelling, R. F. Electrostatic
461−476.
interactions in membranes and proteins. Annu. Rev. Biophys. Biophys.
Chem. 1986, 15, 163−193.
(
29) Hu, Z.; Sim, Y.; Kon, O. L.; Ng, W. H.; Ribeiro, A. J.; Ramos,
M. J.; Fernandes, P. A.; Ganguly, R.; Xing, B.; Garcia, F.; Yeow, E. K.
Unique Triphenylphosphonium Derivatives for Enhanced Mitochon-
drial Uptake and Photodynamic Therapy. Bioconjugate Chem. 2017,
2
(
8 (2), 590−599.
30) Li, P.; Wischert, R.; Met
Oxides with Phosphites To Access Phosphines. Angew. Chem., Int. Ed.
017, 56 (50), 15989−15992.
31) Culcasi, M.; Berchadsky, Y.; Gronchi, G.; Tordo, P. Anodic
́
ivier, P. Mild Reduction of Phosphine
2
(
behavior of crowded triarylphosphines. ESR study of triarylphospho-
niumyl radicals, Ar3P+. J. Org. Chem. 1991, 56 (11), 3537−3542.
(
32) Bruno, I. J.; Cole, J. C.; Edgington, P. R.; Kessler, M.; Macrae,
C. F.; McCabe, P.; Pearson, J.; Taylor, R. New software for searching
G
DOI: 10.1021/acs.inorgchem.8b03380
Inorg. Chem. XXXX, XXX, XXX−XXX