Tuning the Cell Uptake and Subcellular Distribution in BODIPY–Carboranyl Dyads: An Experimental and Theoretical Study

Article abstract of DOI:10.1002/chem.202002600

A set of BODIPY-carboranyl dyads synthesized by a Sonogashira cross-coupling reaction, where different C-substituted ortho- and meta-carboranyl fragments have been linked to a BODIPY fluorophore is described. Chemical, photophysical and physicochemical an

Full text of DOI:10.1002/chem.202002600

Accepted Article
Accepted Article
Title: Tuning the cell uptake and subcellular distribution in BODIPY-
carboranyl dyads: An experimental and theoretical study.
Authors: Pablo Labra-Vázquez, Ricardo Flores-Cruz, Aylin Galindo-
Hernández, Justo Cabrera-González, Cristian Guzmán-
Cedillo, Arturo Jiménez-Sánchez, Pascal G. Lacroix, Rosa
Santillan, Norberto Farfán, and Rosario Núñez
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202002600
Link to VoR: https://doi.org/10.1002/chem.202002600
01/2020

10.1002/chem.202002600
Chemistry - A European Journal
Tuning the cell uptake and subcellular distribution in BODIPY-
carboranyl dyads: An experimental and theoretical study.
Pablo Labra-Vázquez,a,d Ricardo Flores-Cruz,b Aylin Galindo-Hernández,a Justo
Cabrera-González,c Cristian Guzmán-Cedillo,a Arturo Jiménez-Sánchez,b Pascal G.
Lacroix,d Rosa Santillan,e Norberto Farfán,*,a and Rosario Núñez*,c
a
Facultad de Química, Departamento de Química Orgánica, Universidad Nacional
Autónoma de México, 04510 Ciudad de México, México.
b
Instituto de Química, Universidad Nacional Autónoma de México, 04510 Ciudad de
México, México.
c Institut de Ciència de Materials de Barcelona (ICMAB-CSIC), Campus de la UAB, E-08193
Bellaterra, Barcelona, Spain.
d
Laboratoire de Chimie de Coordination du CNRS, 205 route de Narbonne, F-31077,
Toulouse, France.
e Departamento de Química, Centro de Investigación y de Estudios Avanzados del IPN, Apdo.
Postal 14-740, 07000, Ciudad de México, México.
1
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
For Table-of-Contents use only.
Round 1! Fluorophores bearing carborane clusters often display remarkably different cell
incomes, with recurrent higher permeabilities for meta-carboranes compared to ortho-
isomers. Here, we study a set of BODIPY-carborane dyads to provide an explanation for
these differences, which originate from important variations in their partition coefficients and
static dipole moments.
2
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
ABSTRACT
A set of BODIPY-carboranyl dyads synthesized by a Sonogashira cross-coupling reaction,
where different C-substituted ortho- and meta-carboranyl fragments have been linked to a
BODIPY fluorophore is described. Chemical, photophysical and physicochemical analyses
are presented, including NMR and SXRD experiments, optical absorption/emission studies
and log P measurements. These studies, supported by DFT computations (M06-2X/6-31G**)
provide an explanation to the largely divergent cell income that these fluorescent carboranyl-
based fluorophores display, for which a structural or physicochemical explanation remains
elusive. By studying the cell uptake efficiency and subcellular localization for our set of
dyads on living HeLa cells, we tracked the origins of these differences to significant
variations in their static dipole moments and partition coefficients, which tune their ability to
interact with lipophilic microenvironments in cells. Remarkably, m-carboranyl-BODIPY
derivatives with a higher lipophilicity are much better internalised by cells than their
homologous with o-carborane, suggesting that m-isomers are potentially better theranostic
agents for in vitro bioimaging and boron carriers for BNCT.
3
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
1. INTRODUCTION
Dicarba-closo-dodecaboranes (icosahedral 12-vertex species usually named carboranes,
C2B10H12) are chemically intriguing boron clusters bearing two endocyclic carbon atoms.[1]
They possess remarkable chemical, thermal and photochemical stabilities,[2] and spherical
aromaticity through -delocalized electron densities, owed to the ubiquitous occurrence of
multicentered bonds within their skeletons.[3] Consequently, these boranes display some
electronic properties and sizes similar to those of a rotating benzene ring[4] and, accordingly,
isomers 1,2- 1,7- and 1,12-closo-C2B10H12 are typically named ortho-, meta- and para-
carboranes, respectively. In this sense, it has been demonstrated that the different electronic
properties of carborane isomers can notably change the photophysical behaviour of
fluorophores linked through a carbon cluster atom (Cc).[2a, 5] Their unique characteristics have
given carborane derivatives increasing research interest during the last decade in diverse
fields, including catalysis,[6] optoelectronics[7] nonlinear optics,[8] and materials science,[9]
particularly in the preparation of highly thermally stable polymers.[2c] Moreover, carboranes
show low toxicity and high stability in biological systems,[10] and their hydrophobic nature
has proved to enhance the diffusion across the cellular membranes of carborane-bearing
drugs.[11] Owing to these features, one of the largest field of study covers the potential of
carborane-based molecules in biomedical applications, in especial as candidates to be used
as agents for boron neutron capture therapy (BNCT).[4b, 12] This anti-cancer therapy relies on
10
the controlled cell death due to ionizing radiation released from B nuclei after irradiation
with thermal neutrons.[13]
Currently, one of the most outstanding areas of development is the design of new
fluorophores as efficient in vitro imaging probes to be applied in optical imaging. The
4
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
integration of one imaging agent with a target drug in a single system leads to the
development of a theranostic agent, allowing simultaneously diagnosis and treatment of a
specific disease. Nevertheless, relevant structure-related properties such as subcellular
biodistribution and dynamic of fluorescent-derived carboranes are still an unmet need. To
move towards this goal, several chemical architectures containing carborane fragments have
been synthetized by our group,[14] however their molecular structures cannot be easily related
to their subcellular distribution and function, even though some correlation of general
physicochemical observables has been reported to be helpful on achieving specific biological
properties for a great variety of compounds.[1b, 10d, 15]
Among the wide family of fluorophores addressed to bioimaging and chemosensing,
BODIPY (4,4-difluoro-4-bora-3a,4a-diaza-s-indacene) dyes have found an increased interest
due to their unique spectroscopic properties,[16] synthetic versatility,[17] low toxicity and good
stability under physiological conditions.[18] Furthermore, synthesis of BODIPYs bearing
carborane[19] and metallacarborane clusters[20] have been reported and their photophysical
properties assessed. Although some of the studied BODIPY-carboranyl dyads have displayed
exceptional cellular uptake toward HeLa cells, making these clusters good candidates for in
vitro cell tracking, it is often encountered that small changes in the molecular structure deeply
modify their cellular internalization ability.[19f, g]
There exists a consistent evidence suggesting that cell permeability towards carborane-
based compounds is directly influenced by the lipophilic character of the compounds,[21]
which depends on the partition coefficient that in turn is modulated by the dipole moment
and the van der Waals volume of the molecule.[22] According to this, we decided to
investigate a new set of BODIPY-carboranyl dyads specifically intended to correlate these
5
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
structural parameters, not only with their cellular uptake efficiency but also with their
subcellular distribution features.
The target compounds are shown in Chart 1. BODIPY-o-Me-CB and BODIPY-o-Ph-CB
have disubstituted ortho-carborane cages, where the difference substituents linked to the
adjacent Cc (methyl and phenyl) is expected to yield a significant change in the van der Waals
volume, while the dipole moment should be similar. Conversely, when comparing BODIPY-
o-Ph-CB with BODIPY-m-Ph-CB, the van der Waals volume is expected to be relatively
unaltered, whereas there should be a significant change in the static dipole moments between
both isomers, in agreement with previous reports on the chemistry of carboranes.[1b] Thus,
these comparisons would allow us to stablish separately the influence of these structural
features on the cellular uptake efficiency. Finally, BODIPY-Ref, having no carborane
fragment, was included as a reference compound to evaluate how the insertion of different
C-substituted o- and m-carboranyl units into the BODIPY fluorophore influence the cellular
uptake and biodistribution.
Chart 1. BODIPY derivatives under investigation.
6
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
2. RESULTS AND DISCUSSION
Synthesis and Characterization
The synthesis of the target compounds (Scheme 1) was carried out employing a Sonogashira
cross-coupling reaction as the key synthetic step from BODIPY 1, which was coupled to 4-
iodotoluene and the carboranyl derivatives o-Me-CB, o-Ph-CB and m-Ph-CB (further
synthetic details are gathered in the SI),[5] to produce the fluorophores in moderate to good
yields. The identity of these compounds was stablished using multinuclear 1H, 13C{1H}, 11B
19
and F-NMR and HRMS and SXRD experiments (Details are gathered in the Supporting
Information).
The unambiguous assignment of the 1H and 13C{1H} signals for the target compounds was
carried out with the help of homonuclear 1H-1H (COSY, NOESY) and heteronuclear 1H-13C
(HSQC/HMBC) 2D-NMR experiments. Signals from the B-H protons of the clusters
1
appeared as broad multiplets, from  = 1.74 to 2.96 ppm, due to the ubiquitous H-11B and
1H-10B scalar couplings. Aliphatic protons from the methylene (H-16) in the carboranyl
fluorophores appeared in the  = 3.12 - 3.49 ppm interval; resonating at higher frequencies
when compared to the protons of the methyl group in BODIPY-Ref (H-16) with  = 2.39
ppm. This observation is in good agreement with the broadly recognized inductive electron-
withdrawing nature of boranes,[23] here providing a rationale to the iterative deshielding of
these nuclei. Although this influence further echoes strongly on the resonance of C-16, with
δ = 21.55 ppm for BODIPY-Ref and  = 40.78 - 43.01 ppm for the carboranyl derivatives,
the chemical shifts for the aromatic nuclei of the 4 fluorophores were seemingly unaffected,
suggesting a localized withdrawing effect and a rather low electronic coupling between the
carboranyl and BODIPY cores. This result matches well with the optical spectra and the DFT
7
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
computations for these compounds (vide infra), which confirmed the lack of electronic
communication between the two sub-molecular fragments. 11B NMR spectra also confirmed
the formation of the expected compounds showing for all of them similar patterns, with one
resonance at -0.60 ppm assigned to the BF2 group and broad resonances in the range from 
-6 to -15 ppm characteristic from closo-carborane clusters.
Scheme 1. Synthesis and numbering for the target BODIPY derivatives. Reagents and
general conditions: (i) Pd(PPh3)2Cl2, CuI, THF, Et3N, reflux.
8
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
X-Ray Diffraction Studies
Crystals of m-Ph-CB and all of the target BODIPY derivatives suitable for single-crystal X-
ray diffraction analysis were grown by room temperature evaporation of their saturated
solutions in acetone/n-hexane mixtures. Excepting BODIPY-Ref, which crystallized with a
single fluorophore molecule per asymmetric unit (Z’ = 1), all the other BODIPY derivatives
crystallized with Z’ = 2. In the BODIPY-o-Me-CB, the asymmetric unit is composed by two
fluorophore molecules crystallographically independent yet geometrically close. BODIPY-
o-Ph-CB and BODIPY-m-Ph-CB crystallized with one molecule of the fluorophore and half
of a disordered n-hexane coming from the crystallization process. Figures 1-2 show the
molecular structures for these compounds including relevant distances and angles (further
crystallographic details are gathered in Table 1). No significant differences in bond lengths
and angles were found for m-Ph-CB when compared to the previously reported crystal
structure of its ortho- isomer, o-Ph-CB.[5] The introduction of the BODIPY fluorophore did
not alter significantly the bond lengths related to the dicarbaborane cluster as compared with
m-Ph-CB. Indeed, bond lengths among the annular Cc atoms (C8-C9 for m-Ph-CB,
otherwise C25-C26) and the C atoms directly attached to them remained unchanged, even
though the BODIPY core is a strong withdrawing group owed to the delocalized positive
charge within its tricyclic framework. This observation is in good agreement with the lack of
electronic coupling between the carboranyl and BODIPY sub-molecular fragments as
evidenced by the NMR and DFT studies (vide infra). Furthermore, the delocalization within
the BODIPY core is further evidenced by the almost equal N1-B1/N2-B1 distances and N1-
B1-F1/N2-B1-F2 angles, depicting similar bond strengths among these atoms.
9
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
Table 1. Crystal Structure and Refinement Data
m-Ph-CB
BODIPY-o-Me-CB BODIPY-o-Ph-CB BODIPY-m-Ph-CB BODIPY-Ref
empirical formula
C15H21B10I
C27H29B11F2N2
C32H31B11F2N2,
0.5(C6H14)
293(2)
Triclinic
P-1
8.2770(4)
11.8101(8)
18.2600(10)
86.936(5)
80.300(4)
79.934(5)
1731.82(18)
2
C32H31B11F2N2,
0.5(C6H14)
297(2)
Triclinic
P-1
10.6616(7)
12.0496(8)
13.4696(8)
89.052(2)
89.018(2)
88.341(2)
1729.21(19)
2
C24H17BF2N2
temperature (K)
crystal system
space group
a (Å)
b (Å)
c (Å)
130(2)
Orthorhombic
P212121
7.7896(3)
11.6611(5)
20.9589(9)
90
299(2)
Monoclinic
P21/n
14.8970(7)
15.5300(7)
26.5342(11)
90
294(2)
Triclinic
P-1
7.6541(5)
7.8995(5)
16.6764(11)
100.375(2)
97.336(2)
103.098(2)
950.97(11)
2
 (deg)
 (deg)
90
103.5900(10)
90
90
 (deg)
volume (Å)3
Z
1903.81(14)
4
5966.8(5)
8
density
1.522
1.199
1.218
1.236
1.335
(gcm-3)
Crystal
0.41 x 0.36 x 0.32 0.40 x 0.32 x 0.19
0.45 x 0.26 x 0.12
0.25 x 0.25 x 0.12
0.52 x 0.39 x 0.26
size (mm)
3.89 - 29.94
2.20 - 26.45
190178/11516
R1 = 0.0591
wR2 = 0.1273
R1 = 0.1370
3.51 - 30.00
17155/6305
R1 = 0.0534
wR2 = 0.1339
R1 = 0.0685
2.25 - 23.27
32185/4577
R1 = 0.0679
wR2 = 0.1686
R1 = 0.0943
2.52 - 30.13
78656/5606
R1 = 0.0551
wR2 = 0.1298
R1 = 0.1284
 range (deg)
reflns collected/unique 7955/4458
final R indices
[I >2σ(I)]
R indices
R1 = 0.0246
wR2 = 0.0574
R1 = 0.0273
(all data)
wR2 = 0.0588
wR2 = 0.1613
wR2 = 0.1467
wR2 = 0.1891
wR2 = 0.1606
10
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
Figure 1. Solid-state structure for m-Ph-CB. Thermal ellipsoids are drawn at 50 %
probability for every atom other than hydrogen. Selected bond lengths [Å]: C7-C8 1.530(5),
C8-B1 1.710(5), C9-B1 1.708(5), C9-C10 1.511(4), I1-C1 2.086(3).
Figure 2. Solid-state structure for the target BODIPY derivatives with the thermal ellipsoids
drawn at 50 % probability for every atom other than hydrogen. Selected bond lengths [Å]
and angles [°] for BODIPY-o-Me-CB: B1-F1 1.375(3), B1-F2 1.382(3), B1-N1 1.532(4),
B1-N2 1.537(4), C24-C25 1.532(3), C25-C26 1.674(3), C26-C27 1.517(4); F1-B1-F2
108.6(2), N1-B1-N2 106.4(2). For BODIPY-o-Ph-CB: B1-F1 1.3888(17), B1-F2
1.3779(18), B1-N1 1.5489(19), B1-N2 1.545(2), C24-C25 1.5299(18), C25-C26 1.7158(17),
C26-C27 1.5011(18); F1-B1-F2 109.91(11), N1-B1-N2 105.62(10). For BODIPY-m-Ph-
CB: B1-F1 1.386(3), B1-N1 1.527(4), B1-N2 1.536(4), B1-F2 1.372(3), C24-C25 1.530(3),
C26-C27 1.508(3); F1-B1-F2 108.5(2), N1-B1-N2 106.7(2). For BODIPY-Ref: B1-F1
1.380(2), B1-F2 1.385(2), B1-N1 1.531(3), B1-N2 1.542(3); F1-B1-F2 108.29(15), N1-B1-
N2 106.18(14). For clarity, only one of the crystallographically independent molecules is
presented for BODIPY-o-Me-CB.
11
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
At this point, two crystal structures deserve further attention. As mentioned earlier, both
BODIPY-o-Ph-CB and BODIPY-m-Ph-CB differing only on the position of the annular Cc
atoms within the dicarbaborane cluster, crystallized with disordered n-hexane molecules in a
BODIPY/n-hexane 2/1 ratio. Nonetheless, their self-assemblies are considerably divergent,
as markedly different crystallographic environments for the n-hexane molecules can be
observed. As depicted in Figure 3, for BODIPY-o-Ph-CB the solvent molecules are confined
within supramolecular pores formed by the antiparallel head-to-tail arrangement of pairs of
fluorophores, spreading through the crystallographic a axis. Contrastingly, for BODIPY-m-
Ph-CB, the n-hexane molecules are very closely surrounded by fluorophore molecules.
Importantly, in the ortho-isomer no close contacts are observed for the solvent molecules,
whereas in the meta- one, 12 solvent-fluorophore short contacts (less than the sum of van der
Waals radii) were found. This suggests that in the first case the n-hexane molecules move
freely throughout the supramolecular channels, whilst in the latter one the solvent molecules
are densely trapped within the crystal lattice, resulting in these supramolecular interactions.
12
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
Figure 3. Crystallographic projections for (a) BODIPY-o-Ph-CB and (b) BODIPY-m-Ph-
CB along the crystallographic a and b axes, respectively. Disordered n-hexane molecules are
shown in blue.
In order to provide theoretical insights to support these observations, the difference in the
solvation of the different species may tentatively be related to the DFT computation of the
ΔG° (Gibbs free energies), that is associated to the following reaction:
BODIPY + hexane → BODIPY : hexane
(1)
It is important to point out that this approach is based on a single hexane-BODIPY
interaction (one explicit hexane along with a polarizable continuum model simulation for
hexane). Therefore, it cannot be fully relevant to establish the complete behavior of the
different isomers in the presence of solvent. Therefore, it must be used with care. The
optimized geometries for both BODIPY-o-Ph-CB – hexane and BODIPY-m-Ph-CB –
13
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
hexane interacting species are shown in Figure 4. Both starting models were built assuming
that the molecules of solvent closest to the BODIPY-carboranyl dyads, in the solid state
(present X-ray data), can account for most of the solvation interaction. Under such
assumption ΔG° is equal to +6.74 kcal mol-1 and +2.42 kcal mol-1, for BODIPY-o-Ph-CB
and BODIPY-m-Ph-CB, respectively. Therefore, the affinity for hexane appears clearly
reduced in the case of the ortho isomer, in agreement with the X-ray structures.
An additional computational insight on the relative polarity of BODIPY-o-Ph-CB and
BODIPY-m-Ph-CB, and hence their affinity to hexane is available by DFT. The computed
dipole moments are equal to 8.22 D and 4.86 D for BODIPY-o-Ph-CB and BODIPY-m-
Ph-CB, respectively. These values confirm that the solvation in non-polar medium (hexane)
is less favorable for the more polar BODIPY-o-Ph-CB species.
As discussed in detail later, these solid-state observations correlate well with the higher
partition coefficients (log P) values observed for BODIPY-m-Ph-CB compared to
BODIPY-o-Ph-CB. Although formally, a direct comparison is not possible as the SXRD
experiments and log P measurements are performed in different phases, it is not surprising
to find out that the more lipophilic isomer BODIPY-m-Ph-CB interacts with non-polar n-
hexane molecules, while for BODIPY-o-Ph-CB the n-hexane molecules remain isolated
from the clusters.
14
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
Figure 4. Computed geometries for BODIPY-o-Ph-CB (left) and BODIPY-m-Ph-CB
(right) with one molecule of hexane.
Photophysical Properties and TD-DFT Computations
The photophysical properties of the target compounds were analysed by means of UV-Vis
and fluorescence spectroscopies. Figure 5 shows the absorption and emission spectra for
these fluorophores and Table 2 contains a summary of their experimental and computed
photophysical properties. From Figure 5 it might be observed that the shape and energies
associated to the electronic transitions remain practically unaltered regardless the substituent
at the Cc (Me or Ph) or the cluster isomer, when compared with the BODIPY-Ref. Only
some differences in the extinction coefficients and fluorescence intensities were appreciated
(Table 2). Both computation and experimental data match well with the observation of three
local absorption maxima located around 500 nm, 380 nm, and 280 nm, which are readily
attributed to transitions 1 → 2, 1 → 3, and 1 → 7, respectively. These closely related
15
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
transitions involve the dominant contribution of HOMO → LUMO, HOMO-1 → LUMO,
and HOMO-1 → LUMO+1 orbitals for transitions 1 → 2, 1 → 3, and 1 → 7, respectively.
These sets of orbitals are shown in Figure 6, with their related energies. Regarding the
electronic spectra, the contribution of the carboranyl cluster fragments appears negligible in
all of them, leading to the expectations that the four BODIPY derivatives exhibit similar UV-
Vis spectra. The wavelength of the emission maxima for all the compounds is centred at λem
= 519 nm (Figure 5) (λexc= 450 nm), regardless of the type of carboranyl group introduced,
suggesting that there is not an important electronic communication between the BODIPY
and the carborane cage. Moreover, the fluorescence quantum yields (f) of o-carboranyl-
BODIPY dyads are two-fold the f of the non-substituted BODIPY-Ref, whereas a similar
value is obtained for the m-carborane derivative, indicating that linking the cluster to the
BODIPY produces an enhancement of the fluorescence efficiency of the fluorophores (Table
2). Due to the solution-state photophysical behaviour described above, these compounds
were explored as fluorophores for in vitro imaging by confocal microscopy.
Figure 5. Normalized absorption (solid lines) and emission (dashed lines) spectra in ethanol
for the target BODIPY derivatives.
16
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
Table 2. Experimental and computed UV-Vis data and relative fluorescence quantum yields (f
for the target BODIPY derivatives. Both computed and experimental values were
measured/computed in ethanol.
UV-Vis
TD-DFT
λmax (nm) [ f ]
374 425
λem
a
f
λmax (nm) [ (M-1 cm-1)]
(nm)
BODIPY-
Ref
280
381
501
275
519 0.0068
520 0.0108
520 0.0120
521 0.0060
[94309]
280
[35036] [137908] [0.770] [0.792] [0.510]
381 501 275 367 426
BODIPY-
o-Me-CB
BODIPY-
o-Ph-CB
BODIPY-
m-Ph-CB
[127502] [59536] [178213] [0.870] [0.886] [0.505]
280
[53213]
280
381
[79293]
381
501
[69151]
501
275
[0.829] [0.884] [0.505]
275 368 426
367
426
[114067] [71390] [163988] [0.839] [0.876] [0.502]
a Obtained by irradiation at λ = 450 nm, using Rhodamine B as standard with f = 0.5 in ethanol.[24]
Cellular Uptake and Intracellular Localization
The fluorescence properties of all the dyads allow us to assess their subcellular distribution
in live HeLa cells cultures, at different concentrations, by confocal microscopy at the green
(λexc = 486 nm, λem = 500 nm) and red (λexc = 535 nm, λem = 610 nm) channel setups (Fig. 7,
S17 and S18); this is because in confocal microscopy-spectrally resolved studies, the
fluorescence spectral bands of many fluorescence localizers can be strongly widespread in
vitro conditions. In fact, we observed fluorescence signal at higher wavelengths (up to 550
nm), which spatially correlate with the stronger green channel. As shown in Figure 7a-b,
BODIPY-Ref and BODIPY-m-Ph-CB showed the best permeability and localization
characteristics in this cell line. On the contrary, the ortho-derivatives BODIPY-o-Me-CB
and BODIPY-o-Ph-CB poorly translocate the cell membrane probably because their less
lipophilic character strongly favoured aggregation in the cell surface. Nonetheless, by using
a low-dose stimulation with the ortho-derivatives, a better intake was observed (Figure 7c-
17
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
d). While BODIPY-o-Me-CB exhibited a cytosolic distribution with still subtle particle
agglomeration (having no membrane affinity), BODIPY-o-Ph-CB shows a perinuclear
membrane localization, although non-specific. Importantly, the meta-derivative BODIPY-
m-Ph-CB showed excellent fluorescence distribution features, where the plasma membrane
was mostly localized in a wide range of fluorophore concentrations, from 10 to 40 µM,
having almost no variation in such subcellular distribution. Bright-field images as well as
comparison with commercial plasma membrane localizer using CellMask® is shown in the
Supporting Information (Fig. S19). It is worth mentioning that BODIPY-Ref also exhibited
interesting membrane permeability with a non-specific localization in the perinuclear zone.
Further, this fluorophore was able to specifically localize apoptotic vesicles and cellular blebs
during cell oxidative stress.[25] The low-dose stimulus with H2O2 was used in order for the
blebs to be well-contrasted under the green channel illumination, allowing a bright in-plane
bleb visualization with clean green-emission on these membranes.
18
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
Figure 6. Dominant frontier orbitals involved in the three intense 1 → 2, 1 → 3, and 1 → 7 electronic transitions for BODIPY-Ref,
BODIPY-m-Ph-CB, BODIPY-o-Ph-CB and BODIPY-o-Me-CB.
19
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
Figure 7. Confocal images of 10 μM fluorophores in live HeLa cells after 30 min incubation.
Panels from left to right show the confocal green channel (λexc = 486 nm, λem = 500 nm), red
channel (λexc = 535 nm, λem = 610 nm) and merged channels, respectively.
20
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
At this point, we were motivated to investigate the origin of the divergent cell uptake
efficiency and subcellular localization of these molecules, with the most striking differences
arising from BODIPY-o-Ph-CB and BODIPY-m-Ph-CB. The molecular structure of these
two compounds differ exclusively on the position of the endocyclic Cc atoms, however they
displayed completely different self-assemblies and fluorophore-cells interactions. Having in
mind that the theoretical computations discussed earlier (vide supra) suggested a higher
affinity of BODIPY-m-Ph-CB for non-polar molecules as n-hexane than its homologous
ortho-isomer, we wondered if a difference in lipophilicity would provide an explanation for
the divergent confocal microscopy results. Thus, as the lipophilic partition of some organic
fluorophores is determined by their partition coefficients (log P),[26] we measured log P for
these fluorophores (Table 3, Figure S20). Those molecules showing the best cell
internalization also possessed the higher log P values, which correlate with a higher
lipophilicity of BODIPY-Ref, BODIPY-m-Ph-CB and consequently a better interaction
with the cell membranes. To gain further insight on these observations, we determined the
van der Waals volume of these molecules and computed by DFT their static molecular
dipoles. These two parameters could help us to rationalize the differences in log P and
consequently in the internalization of the fluorophores into the cells. Both the molecular
volume and the polarity of a molecule have long been recognized as critical factors
determining its permeability.[27] Although a correlation between log P and the size/volume
of the molecules is absent, a negative proportional relationship of log P and the dipole
moment appears to be present, correlating low polarities with higher log P values. Although
this correlation provides a satisfactory explanation for the marked differences observed in
the cell intake of fluorophores, further studies would be needed to correlate a larger interval
of log P values and dipolar moments with the cell income of carboranyl derivatives. In this
21
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
manner, the straightforward DFT computation of static dipole moments could eventually
guide the design of fluorescent carboranyl derivatives with potential theragnostic
applications.
Table 3. Partition coefficients (log P), dipolar moments () and van der Waals volumes (Vvdw).
log P a
 (D)b
Vvdw (Å3/molecule)c
BODIPY-Ref
0.324
-0.166
-0.211
0.677
8.29
10.07
8.22
950.97
1491.71
1595.85
1589.56
BODIPY-o-Me-CB
BODIPY-o-Ph-CB
BODIPY-m-Ph-CB
4.86
a log P values were calculated using a octanol:water partitioning protocol.[28] b Computed at the M06-
2X/6-31G(d,p) level of theory. c Computed from the crystal structures using CSD-Materials.[29]
As a final remark, we would like to stress the significance of these results, as they provide
an explanation for the specific and recurrent observation that diverse fluorophores bearing
meta-carboranyl derivatives show better cell incomes than those of their ortho-
counterparts.[19g] Furthermore, the high lipophilicity of meta-isomers correlates well with
their subcellular distributions, as these species generally localize in lipophilic membranous
organelles. The understanding of the reasons behind this specific behaviour paves the way
for designing new and better fluorescent carboranes for in vitro cell imaging.
3. CONCLUSIONS
The cell uptake efficiency and subcellular localization of a set of boron-dipyrromethene
difluoride (BODIPY) dyes decorated with Cc-substituted carboranyl derivatives were
investigated by confocal microscopy on living HeLa cells. They exhibited markedly
22
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
divergent behaviours depending on the cluster isomer linked to the BODIPY core, whereas
not great alterations with the substituents at Cc was observed. A rationale for these differences
was provided from a detailed structural and physicochemical characterization of these
clusters, supported by a combination of experimental measurements and theoretical DFT
computations. Altogether, our results show that the differences in cell income and subcellular
distribution of the fluorophores originate from an important variation in their static dipole
moments and partition coefficients, which modulates the ability of these molecules to interact
with the lipophilic microenvironments in cells. It can be highlighted that these properties are
indeed related to the type of cluster isomer linked to the BODIPY, being the m-carborane
derivatives better internalized than their ortho- analogues. This evidence provides a
molecular design strategy for improving the prospective applications of BODIPY-carboranyl
dyads as potential fluorescence imaging agents and boron carriers for BNCT.
4. EXPERIMENTAL
General Experimental Considerations
Materials. All starting materials were purchased from Sigma-Aldrich and used without
further purification. Solvents were purified by distillation over appropriate drying agents.
Starting materials, BODIPY 1, o-Me-CB, o-Ph-CB and m-Ph-CB were obtained following
modified literature procedures;[5, 30] spectroscopic data is in good agreement.
Instrumentation. NMR spectra were recorded on Jeol ECA 500, Varian MR-400 and Bruker
ARX 300 spectrometers using deuterated solvents; chemical shifts for 1H and 13C NMR data
are relative to the residual nondeuterated solvent signal, fixed at  = 7.26 ppm for 1H-NMR
23
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
and  = 77.00 ppm for 13C-NMR. Infrared spectra were registered on a FT-IR Varian ATR
spectrometer. HRMS data was acquired using an Agilent G1969A MS TOF spectrometer.
Synthesis of the Target BODIPY Derivatives.
General Sonogashira Cross-Coupling procedure. To a round-bottom flask filled with
BODIPY 1 (1 mmol), Pd(PPh3)2Cl2 (0.05 mmol), CuI (0.1 mmol) and a carboranyl benzyl
iodide (o-Me-CB, o-Ph-CB, m-Ph-CB) or 4-iodotoluene (1 mmol) as coupling partner, were
added THF (30 mL) and triethylamine (30 mL). The mixture was magnetically stirred
overnight at room temperature, heated to reflux for 2 hours and evaporated under reduced
pressure, the resulting residue was dissolved in methylene chloride (100 mL), washed with
saturated ammonium chloride (3x40 mL), dried over anhydrous sodium sulfate and purified
through a chromatographic column packed with silica gel and eluted with hexane/acetone
mixtures.
BODIPY-o-Me-CB. Red solid. Yield: 56 %. 1H-NMR (CDCl3, 400 MHz)  7.96 (s, 2H, H-
1), 7.68 (d, J = 8.2 Hz, 2H, H-8), 7.61 – 7.48 (m, 4H, H-7, H-13), 7.22 (d, J = 8.1 Hz, 2H,
H-14), 6.95 (d, J = 4.1 Hz, 2H, H-3), 6.57 (d, J = 2.9 Hz, 2H, H-2), 3.49 (s, 2H, H-16), 2.18
(s, 3H, H-19), 2.64 - 1.89 (m, 10H, B-H). 13C-NMR (CDCl3, 100 MHz)  146.34 (C-5),
144.35 (C-1), 135.59 (C-15), 134.69 (C-4), 133.60 (C-9), 131.91 (C-13), 131.59 (C-8),
131.39 (C-3), 130.57 (C-7), 130.47 (C-14), 125.90 (C-6), 122.56 (C-12), 118.74 (C-2), 91.48
11
(C-11), 89.15 (C-10), 77.20 (C-18), 74.79 (C-17), 41.04 (C-16), 23.71 (C-19). B-NMR
(CDCl3, 128 MHz)  -0.69 (t, J = 28.8 Hz, 1B, BF2), -5.61 (d, J = 270.5 Hz, 4B, BH), -9.13
– -12.58 (m, 6B, BH). 19F-NMR (CDCl3, 376 MHz)  -144.89 (q, J = 28.7 Hz). HRMS (ESI-
24
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
TOF+) m/z: [M+-F]+ Observed: 521.3356. Anal. calcd. for C27H29B11N2F: 521.3338. Error:
3.45 ppm.
BODIPY-o-Ph-CB. Red solid. Yield: 81 %. 1H-NMR (CDCl3, 400 MHz)  7.96 (s, 2H, H-
1), 7.75 – 7.70 (m, 2H, H-20), 7.68 – 7.65 (m, 2H, H-8), 7.57 (dt, J = 8.2, 1.5 Hz, 2H, H-7),
7.55 – 7.52 (m, 1H, H-22), 7.50 – 7.47 (m, 2H, H-21), 7.44 (dt, J = 8.5, 2.0 Hz, 2H, H-13),
6.95 (d, J = 4.2 Hz, 2H, H-3), 6.85 (d, J = 8.3 Hz, 2H, H-14), 6.58 – 6.55 (m, 2H, H-2), 3.12
(s, 2H, H-16), 2.96 – 1.95 (m, 10H, BH). 13C-NMR (CDCl3, 100 MHz)  146.35 (C-5),
144.34 (C-1), 135.84 (C-15), 134.71 (C-4), 133.56 (C-9), 131.67 (C-13), 131.56 (C-8),
131.48 (C-20), 131.37 (C-3), 130.95 (C-22), 130.70 (C-19), 130.55 (C-7), 130.19 (C-14),
129.11 (C-21), 125.95 (C-6), 122.30 (C-12), 118.70 (C-2), 91.59 (C-11), 89.01 (C-10), 83.70
11
(C-18), 81.45 (C-17), 40.78 (C-16). B-NMR (CDCl3, 128 MHz)  -0.69 (t, J = 28.8 Hz,
1B, BF2), -4.29 (br s, 2B, BH), -8.99 – -15.94 (m, 8B, BH).19F-NMR (CDCl3, 376 MHz)  -
144.90 (q, J = 28.9 Hz). HRMS (ESI-TOF+) m/z: [M+-F]+ Observed: 583.3519. Anal. calcd.
for C32H31B11N2F: 583.3495. Error: 4.11 ppm.
BODIPY-m-Ph-CB. Orange solid. Yield: 61 %. 1H-NMR (CDCl3, 400 MHz)  7.96 (s, 2H,
H-1), 7.68 (d, J = 8.4 Hz, 2H, H-8), 7.55 (dd, J = 15.4, 8.3 Hz, 4H, H-7, H-13), 7.36 (dd, J =
8.2, 1.4 Hz, 2H, H-20), 7.29 – 7.23 (m, 3H, H-21, H-22), 7.16 (d, J = 8.2 Hz, 2H, H-14), 6.95
(d, J = 4.1 Hz, 2H, H-3), 6.59 – 6.55 (m, 2H, H-2), 3.30 (s, 2H, H-16), 2.68 – 1.74 (m, 10H,
B-H). 13C-NMR (CDCl3, 100 MHz)  146.40 (C-5), 144.34 (C-1), 137.59 (C-15), 135.14 (C-
19), 134.73 (C-4), 133.53 (C-9), 131.85 (C-13), 131.56 (C-8), 131.38 (C-3), 130.56 (C-7),
130.06 (C-14), 128.64 (C-22), 128.31 (C-21), 127.76 (C-20), 126.05 (C-6), 122.01 (C-12),
11
118.72 (C-2), 91.79 (C-11), 88.86 (C-10), 77.20 (C-18), 75.63 (C-17), 43.01 (C-16). B-
NMR (CDCl3, 128 MHz)  -0.69 (t, J = 28.8 Hz, 1B, BF2), -6.22 – -15.41 (m, 10B, BH).
25
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
19F-NMR (CDCl3, 376 MHz)  -144.90 (q, J = 28.8 Hz). HRMS (ESI-TOF+) m/z: [M+-F]+
Observed: 583.3502. Anal. calcd. for C32H31B11N2F: 583.3495. Error: 1.19 ppm.
BODIPY-Ref. Red solid. Yield: 45 %. 1H-NMR (CDCl3, 400 MHz)  7.95 (br s, 2H, H-1),
7.67 (d, J = 8.1 Hz, 2H, H-7), 7.56 (d, J = 8.1 Hz, 2H, H-8), 7.47 (d, J = 8.0 Hz, 2H, H-13),
7.20 (d, J = 8.0 Hz, 2H, H-14), 6.95 (d, J = 4.0 Hz, 2H, H-3), 6.58 – 6.55 (m, 2H, H-2), 2.39
(s, 3H, H-16). 13C-NMR (CDCl3, 100 MHz)  146.51 (C-5), 144.24 (C-1), 139.10 (C-15),
134.70 (C-4), 133.21 (C-9), 131.63 (C-13), 131.47 (C-8), 131.39 (C-3), 130.53 (C-7), 129.24
(C-14), 126.42 (C-6), 119.54 (C-12), 118.67 (C-2), 92.56 (C-11), 87.73 (C-10), 21.55 (C-
16). 19F-NMR (CDCl3, 376 MHz)  -144.86 (q, J = 28.8 Hz). 11B-NMR (CDCl3, 128 MHz)
 -0.68 (t, J = 28.8 Hz). HRMS (ESI-TOF+) m/z: [M++H]+ Observed: 383.1544. Anal. calcd.
for C24H18BF2N2: 383.1531. Error: 3.39 ppm.
Single X-Ray Diffraction studies. The intensity data were collected on Oxford Diffraction
Gemini "A" (m-Ph-CB) and Bruker D8 Venture CMOS (BODIPY-Ref, BODIPY-o-Me-
CB, BODIPY-o-Ph-CB, BODIPY-m-Ph-CB) diffractometers equipped with CCD area
detectors (MoK = 0.71073 A˚, monochromator: graphite) at 130 K and room temperature,
respectively. The crystals were mounted on conventional MicroLoops.™ All heavier atoms
were found by Fourier map difference and refined anisotropically. All reflection data set were
corrected for Lorentz and polarization effects. The first structure solution was obtained using
the SHELXS-2018 program and the SHELXL-2018 was applied for refinement and output
data.[31] All software manipulations were performed through the ShelXle program.[32]
Mercury 3.7 and ORTEP-3 were used to prepare artwork representations.[33]
26
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
CCDC 1975442-1975446 contain the supplementary crystallographic data for this paper.
These data can be obtained free of charge via https://summary.ccdc.cam.ac.uk/structure-
summary-form (or from the Cambridge Crystallographic Data Centre, 12, Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033).
Steady State Spectroscopy. Absorption spectra were acquired in a 10 mm path-length
quartz-cell Cary-50 (Varian) spectrophotometer with dual beam, Czerny-Turner
monochromator, 190–1100 nm wavelength range, approximately 1.5 nm fixed spectral
bandwidth, full spectrum Xe pulse lamp, dual Si diode detectors, quartz overcoated optics,
scan rates up to 24000 nm/min, at room temperature (20 ± 1 °C) under aerated conditions.
Fluorescence experiments were measured on a FluoroMax spectrofluorometer from
HORIBA Scientific with Vertically mounted, CW, 150 W Ozone-free xenon arc
lamp. Emission: R928P photon counting PMT (185-850 nm) and reference photodiode for
monitoring lamp output at room temperature (20 ± 1 °C) under aerated conditions. All the
photophysical measurements were performed on 1 mM solutions of the analytes in ethanol.
Optical Microscopy. The microscopy images were taken using an inverted Zeiss LSM
880 microscope maintaining 5% CO2 and 37°C during the experiments. HeLa cells were
seeded in 8 well μ–slides (iBidi, Germany) at a density of 20000 cells per well one day prior
to experiments in MEM alpha with 10% FBS. On treatment day, cells were washed once in
MEM alpha with no FBS, and incubated with compounds at different concentrations, from
10 to 40 µM, for 30 minutes. Cells were then washed twice in MEM alpha with no FBS and
imaged. No signs of apoptotic or morphologically altered cells were observed during these
experiments.
27
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
Computational Methods. BODIPY-Ref, BODIPY-m-Ph-CB, BODIPY-o-Ph-CB and
BODIPY-o-Me-CB were fully optimized at the M06-2X/6-31G* level, using the Gaussian-
09 program package[34] within the framework of the Density Functional Theory (DFT). The
computation were performed assuming the presence of ethanol as the solvent, which was
modeled by the Polarizable Continuum Model (SCRF=PCM method).[35] For the
computations devoted to the interaction with hexane (Equation 1), no other solvent effect
was taken into account than the presence of one molecule of hexane in the vicinity of the
cluster. We selected the M06-2X hybrid functional following the previous report by
Jacquemin,[36] who recommends M06-2X as the most reasonable DFT method for molecules
based on the BODIPY units. The optimized coordinates are provided in Supplementary
Materials. The UV-vis spectra were computed at the M06-2X/6-311+G** level.[36]
AUTHOR INFORMATION
* E-mail: rosario@icmab.es
* E-mail: norberto.farfan@gmail.com
Conflict of interest
The authors declare no conflict of interest.
ACKNOWLEDGEMENTS
The work has been performed within the framework of the French-Mexican International
Laboratory (LIA-LCMMC). Authors acknowledge support from PAPIIT (IN222819) and
CONACyT. The work was supported by Spanish Ministerio de Economía y Competitividad,
MINECO (CTQ2016-75150-R) and Generalitat de Catalunya (2017/SGR/1720). Doctoral
28
This article is protected by copyright. All rights reserved.

10.1002/chem.202002600
Chemistry - A European Journal
scholarships were granted by CONACyT to P.L.-V (337958) and R. F.-C. (576635). We
thank María E. Ochoa (CINVESTAV-IPN), Marco A. Leyva (CINVESTAV-IPN) and Flor
R. Vega (FQ-UNAM) for NMR/HRMS experiments, SXRD experiments and technical
support, correspondingly.
Keywords
Carborane, BODIPY, imaging, BNCT, theranostic
ASSOCIATED CONTENT
1
Supporting Information Available: Synthesis of the starting materials, H, 13C NMR and
HRMS spectra, atomic coordinates for the four DFT-optimized BODIPY derivative, confocal
images and plots of log P.
REFERENCES
[1] a) R. N. Grimmes, Carboranes, Academic Press, Boston, 2016 b) M. Scholz and E. Hey-
Hawkins, Chem. Rev. 2011, 111, 7035-7062; c) N. S. Hosmane, Boron Science: New
Technologies and Applicatons, Taylor & Francis, Bosa Roca, 2012.
[2] a) J. Cabrera-González, A. Ferrer-Ugalde, S. Bhattacharyya, M. Chaari, F. Teixidor, J.
Gierschner and R. Núñez, J. Mater. Chem. C 2017, 5, 10211-10219; b) A. Ferrer-Ugalde, E.
J. Juárez-Pérez, F. Teixidor, C. Viñas and R. Núñez, Chem. Eur. J. 2013, 19, 17021-17030;
c) R. Núñez, I. Romero, F. Teixidor and C. Viñas, Chem. Soc. Rev. 2016, 45, 5147-5173.
[3] a) Z. Chen and R. B. King, Chem. Rev. 2005, 105, 3613-3642; b) J. Poater, M. Solà, C.
Viñas and F. Teixidor, Chem. Eur. J. 2013, 19, 4372-4372.
[4] a) J. Poater, M. Solà, C. Viñas and F. Teixidor, Angew. Chem. Int. Ed. 2014, 53, 12191-
12195; b) P. Stockmann, M. Gozzi, R. Kuhnert, M. B. Sárosi and E. Hey-Hawkins, Chem.
Soc. Rev. 2019, 48, 3497-3512.
29
This article is protected by copyright. All rights reserved.