Near-infrared light activated azo-BF2 switches

Article abstract of DOI:10.1021/ja508125n

Increasing the electron density in BF₂-coodinated azo compounds through para-substitution leads to a bathochromic shift in their activation wavelength. When the substituent is dimethyl amine, or the like, the trans/cis isomerization process can be efficiently modulated using near infrared light. The electron donating capability of the substituent also controls the hydrolysis half-life of the switch in aqueous solution, which is drastically longer for the cis isomer, while the BF2-coodination prevents reduction by glutathione.

Full text of DOI:10.1021/ja508125n

Communication
pubs.acs.org/JACS
Near-Infrared Light Activated Azo-BF2 Switches
Yin Yang, Russell P. Hughes, and Ivan Aprahamian*
Department of Chemistry, Dartmouth College, Hanover, New Hampshire 03755, United States
S
* Supporting Information
property enabled us to switch the azo-compounds using visible
(i.e., blue and green) light. Density functional theory (DFT)
calculations predicted that increasing the electron-density in the
system could further red-shift the absorption bands and hence
activation wavelength of the system. To test this hypothesis, we
designed a series of azo-BF2 complexes having electron donating
para- and ortho-substituents (Scheme 1).17 Here we report how
ABSTRACT: Increasing the electron density in BF2-
coodinated azo compounds through para-substitution
leads to a bathochromic shift in their activation wave-
length. When the substituent is dimethyl amine, or the
like, the trans/cis isomerization process can be efficiently
modulated using near infrared light. The electron donating
capability of the substituent also controls the hydrolysis
half-life of the switch in aqueous solution, which is
drastically longer for the cis isomer, while the BF2-
coodination prevents reduction by glutathione.
Scheme 1. Synthesis of Ortho- and Para-substituted BF2-
Coordinated Azo Complexes
ight penetration through tissue is primarily regulated by the
L
absorptions of hemoglobin and water,1 which limit its
“therapeutic window” to the 600−1200 nm range. In principle,
the more red-shifted the wavelength, the deeper the penetration;
hence, near-infrared (NIR) is far better than red light in this
aspect.2 One way of using this property of light is to couple it with
photochromic3 compounds that are capable of reversibly
modulating biological processes. This is the main objective of
the fields of photopharmacology4 and opto(chemical)genetics.5
Azo compounds6 are the most commonly used light activated
switches7 in these research areas because of their efficient trans/
cis photoisomerization. However, this process generally relies on
UV light, which might limit the biocompability of such switches.8
Consequently, there has been intense activity in the field in trying
to shift the activation wavelength of these photochromic
compounds to the visible region9 and beyond. The burgeoning
activity has led to the development of a number of visible light-
activated azo compounds, through appropriate derivatization,10
or the use of metal to ligand charge transfer,11 among other
approaches.12 This activity has recently paid off with the seminal
work by Woolley et al. describing the in vivo activation of an azo
compound using red light (635 nm).10c Despite these recent
advances there is still a need to develop more efficient systems
and push the activation wavelength of the azo compounds
beyond the red region, in order to gain access to deeper tissues.13
In this context, Qian et al. have recently showed14 how NIR
activated upconversion nanoparticles15 can be used in
manipulating azo compounds; however, this was accomplished
using the NIR light indirectly, as the azo switch was still
modulated using the UV light emitting from the excited
nanoparticles. Furthermore, the incorporation of nanpoarticles
complicates the system and leads to low isomerization
efficiencies, reducing the practicality of this approach in
photopharmacology and opto(chemical)genetics.
such substituents shift the azo-BF2 absorption bands to the red
and even NIR region, thus enabling us (for the first time to the
best of our knowledge) to directly and efficiently activate the
isomerization of an azo-compound using NIR light.18 We also
show how these systems are stable to glutathione reduction and
have relatively long-lived half-lives in aqueous solutions.
As predicted by DFT calculations, the introduction of a
methoxy group para to the azo linkage (Scheme 2) leads to a
Scheme 2. Visible Light-Induced Trans/Cis Isomerization of 2
bathochromic shift in the π−π* band of 2 (λmax = 594 nm; ε =
15,998 M−1 cm−1). The λmax of the cis photostationary state (PSS;
irradiation at 630 nm) of compound 2 is red-shifted by 40 nm
(Figure 1a), while the trans PSS (irradiation at 490 nm) is red-
shifted by 55 nm compared with the parent complex 1.16 The
switching process of 2 was accompanied by a strong color change
between cobalt blue and poppy red. High photoconversion ratios
(PSS490 = 92% trans; PSS630 = 96% cis) and a 93% trans isomer
ratio under dark were determined for 2 using 1H NMR
spectroscopic analysis (Figures S4 and S3 in the Supporting
Information). There are several spectral features of 2 that set it
We have recently discovered16 that the coordination of BF2
with an azo group’s nitrogen lone-pair leads to a reversal of the
positions of n-π* and π−π* transition energy levels. This
Received: August 7, 2014
Published: September 15, 2014
© 2014 American Chemical Society
13190
dx.doi.org/10.1021/ja508125n | J. Am. Chem. Soc. 2014, 136, 13190−13193

Journal of the American Chemical Society
Communication
in the trans isomer can be inferred by its blue-shifted absorption
band, and its destabilization is evident from the lower trans
isomer ratio (58%) under dark.
Encouraged by the results from compound 2, we shifted our
focus to the BF2-azo complexes 4−8, which were synthesized
with the intention of further pushing the activation wavelength of
the systems to even lower energy levels. The attachment of a
para-dimethylamino group shifted the absorption peak of
complex 4 to 680 nm (ε = 36 255 M−1 cm−1) with a tail
extending out to 760 nm (Figure 2). This property allowed us to
Figure 1. (a) UV/vis spectral changes upon the photoisomerization of 2
in CH2Cl2 (0.2 mM). The black trace is of 2 equilibrated under dark
(mainly trans), which upon irradiation at λ = 630 nm gives the cis PSS
(blue trace). Irradiating the latter at λ = 490 nm gives the trans PSS (red
trace). (b) UV/vis spectral changes upon the photoisomerization of 3 in
CH2Cl2 (0.1 mM). The black trace is of 3 equilibrated under dark, which
upon irradiation at λ = 640 nm gives the cis PSS (blue trace). Irradiating
the latter at λ = 490 nm gives the trans PSS (red trace).
apart from the parent compound; instead of a sharp band, the
π−π* band of the trans-dominant state exhibits a well-resolved
vibrational fine structure, with the highest-intensity peak
observed at λmax = 594 nm. The appearance of the sub-bands
can be attributed to the intensified vibrational transitions caused
by the electron-donating para-substituent.19 In addition, the
isomerization process in 2 is more efficient than 1 based on its
quantum yields (Φtrans→cis = 71% and Φcis→trans = 95%). This
efficiency enhancement results from a better separation of the cis
and trans states’ π−π* bands, which leads to less overlap of their
irradiation windows compared to the parent azo complex 1. The
cis isomer of 2 has a half-life of 10.4 h at 294 K in deoxygenated
methylene chloride (Figure S25 in the Supporting Information),
compared to 25 min in regular solvent (not deoxygenated).16,20
Hecht et al. recently identified10b that ortho-tetrafluoroazo-
benzene has a slightly blue-shifted π−π* absorption when
compared to its parent azobenzene. In 2011, Woolley’s group
reported10a that the substitution of the ortho positions with
methoxy groups leads to the separation of the n−π* orbital of the
cis and trans isomers and a red-shift of the π−π* band in the
latter. In order to study the ortho-substitution effect on our
system, we prepared complex 3 (Scheme 3). The extra ortho-
Figure 2. (a) NIR light-induced trans/cis isomerization of 4. (b) UV/vis
spectral changes upon the photoisomerization of 4 in CH2Cl2 (0.2 mM).
The black trace is of 4 equilibrated under dark (mainly trans), which
upon irradiation at λ = 710 nm gives the cis populated state (red trace).
activate the trans to cis isomerization process using NIR light!
The UV/vis spectrum of the trans isomer of 4 exhibits better-
resolved vibrational fine structure compared to 2. The half-life of
the cis isomer of 4 was determined to be 250 s (Figure S26 in the
Supporting Information), with no obvious difference in rate
observed by deoxygenating the methylene chloride solution.
Such an enhancement in the thermal relaxation rate compared to
the parent compound 1 (t1/2 = 12.5 h)16 is not surprising. The
effective bond order of NN bond is strongly influenced by the
substituents on the phenyl rings (vide infra). Electron-donating
groups would cause an increase in the electron density of the π*
(antibonding) orbital and thus a decrease in the effective bond
order of the NN bond, which leads to a lower thermal
isomerization barrier.6,21 Under dark, complex 4 is almost
quantitatively (97%) composed of the thermodynamically more
stable trans form. Upon irradiation with 710 nm light,
isomerization to the cis isomer occurs quickly. Because of the
fast cis to trans isomerization rate, we were unable to record the
PSS at 710 nm. The lowest estimation of the amount of cis isomer
at PSS710 is 63% based on 1H NMR spectroscopy (Figure S12 in
the Supporting Information).22
Scheme 3. Visible Light-Induced Trans/Cis Isomerization of 3
methoxy group in 3 causes a 20 nm blue shift of its trans
absorption band relative to 2 (Figure 1b), whereas its cis isomer’s
π−π* band is red-shifted by 15 nm. The combined effect is the
generation of a pronounced overlap between the absorption
bands of the two configurations. This overlap resulted in a drastic
decrease in the efficiency of the switch, including its PSS ratios
(PSS490 = 56% and PSS640 = 79%) and photoswitching
quantum yields (Φcis→trans = 51% and Φtrans→cis = 42%). We
hypothesize that the steric hindrance of the ortho-OMe group
destabilizes the trans isomer and prevents the phenyl ring from
lying coplanar with the rest of the molecule. The loss of planarity
Next we studied the isomerization and photochromic
properties of the pyrrolidinyl (5), piperidinyl (6), methylpiper-
izinyl (7), and morpholinyl (8) BF2-azo derivatives (Figure 3,
and Figure S31 in Supporting Information). As can be seen by
their UV/vis spectra, switches 5 and 6−8 undergo isomerization
13191
dx.doi.org/10.1021/ja508125n | J. Am. Chem. Soc. 2014, 136, 13190−13193

Journal of the American Chemical Society
Communication
Figure 4. (a) Switching cycles of 4 in acetonitrile/PBS buffer (1:1)
monitored by following the absorbance at λ = 681 nm (black trace) upon
irradiation at λ = 710 nm then leaving in the dark. (b) UV/vis spectra
following an acetonitrile/PBS buffer (1:1) solution of the azo-BF2
complex 4 at 25 °C. The interval between each scan is 20 min.
existence of only two species in solution, which in this case are
the azo-BF2 complex and the hydrazone starting material. These
results, along with similar stability tests conducted for
compounds 5−8 (Figures S34−S41 in the Supporting
Information), demonstrate the robustness of this class of NIR
switches in high concentrations of GSH. Moreover, they indicate
that coordination with BF2 might be a viable strategy to
preventing glutathione reduction of azo compounds.
By comparing the degradation half-life of the complexes in
PBS buffer (2.6, 2.3, 2.0, 1.2, and 0.9 h for 5, 4, 6, 7, and 8,
respectively), we can conclude that the higher electron-donating
ability of the para-substituent, the more stable the switch is to
hydrolysis. According to the crystallographic analysis (Figure 5),
Figure 3. UV/vis spectral changes upon photoisomerization of (a) 5,
(b) 6, and (c) 7. The black trace is the equilibrated complex in the dark
(mainly trans), which upon irradiation at λ = 710 or 730 nm gives the cis
populated state (red trace). (d) Overlay of the π−π* absorption bands
of the trans isomers of 4−8, represented by green, magenta, black, blue,
and red lines, respectively.
upon irradiation with 730 and 710 nm light, respectively, and
maintain a high ratio of trans isomer under dark (98%, 98%, 97%
and 94% trans ratio, with molar extinction coefficient ε = 41 839,
30 458, 29 536, and 28 025 M−1 cm−1, for 5, 6, 7, and 8,
respectively). The small variation in the electronic characteristics
of the amino groups greatly impacts their photophysical
properties, including their absorption bands and half-lives. The
overlay of the absorption bands of compounds 4−8 (Figure 3d)
reveals that the resolution of their vibrational fine structure
decreases in the order of 5 > 4 > 6 > 7 > 8, which is in accordance
to their para-substituent’s electron-donating ability.23 A similar
trend is also observed in their absorption maxima (λmax = 690,
681, 661, and 652 nm for 5, 6, 7, and 8, respectively). In addition,
the half-life relaxation rate from cis to trans is substantially
enhanced in switch 8 (t1/2 = 900 s) as compared to 5 (t1/2 = 120
s), 6 (t1/2 = 150 s), and 7 (t1/2 = 400 s), which is again in
accordance with their electron donation capability. Such a
dramatic change in the half-life of these switches was also
reflected by their PSS values. The highest PSS ratio for 5 that we
could measure was 23%, while for 6, 7, and 8 it was 28%, 49%,
and 83%, respectively. The latter value is comparable to that
obtained for the parent azo-BF2 complex 1.
To test the stability of the switches in an aqueous environment,
we conducted multiple switching cycles (10 cycles shown in
Figure 4a) of 4 in an acetonitrile/PBS buffer (1:1) mixed solvent
system. These experiments were conducted by irradiating the
sample using 710 nm light, followed by thermal relaxation under
dark. Although switch 4 is stable for short periods of time under
these conditions, it gradually undergoes hydrolysis (Figure S32
in the Supporting Information), reverting back to the starting
hydrazone24 compound (Figure 4b) with a half-life of 2.3 h. We
also tested the stability of switch 4 to reduction by glutathione
(GSH). The switch was incubated in 10 mM reduced glutathione
in acetonitrile/PBS buffer (1:1) solution. No obvious difference
was observed in its degradation half-life (t1/2 = 2.5 h)25 compared
to that measured in the acetonitrile/PBS buffer (Figure S33 in
the Supporting Information). The multiple isosbestic points
observed in both cases (λ = 286, 413, and 537 nm) suggest the
Figure 5. ORTEP drawing (50% probability ellipsoids) of the crystal
structures of (a) 2 and (b) 4. The hydrogen atoms were removed for
clarity.
the stronger the electron-donating capability of the para-
substituent, the longer the N(1)N(2) bond (1.294(1) and
1.298(2) Å in 2 and 4, respectively) and shorter the B(1)−N(3)
bond and B(1)−N(2) bond lengths become (1.550(2) and
1.624(2) Å in 2, and 1.542(2) and 1.615(2) Å in 4, respectively).
We speculate that the latter trend (i.e., the strengthening of the
BN bonds) makes the compounds less susceptible to hydrolysis.
Remarkably, we found out that switch 4 is relatively stable
toward hydrolysis if it is kept in its cis form through constant
irradiation with 710 nm light. We monitored aqueous solutions
of 4 (with and without glutathione) using UV/vis spectroscopy
(Figure 6) and observed at most 6% hydrolysis over a period of 8
h. To have a better understanding of this phenomenon we
calculated the cis and trans isomers of 4 (Figure S47 and S48 in
the Supporting Information) and found out that the B(1)−N(2)
bond (Figure 5) in the cis isomer is shorter by 0.02 Å compared
to the trans form. The same trend was found for compounds 1, 5,
13192
dx.doi.org/10.1021/ja508125n | J. Am. Chem. Soc. 2014, 136, 13190−13193

Journal of the American Chemical Society
Communication
2011, 50, 12156−12182. (b) Brieke, C.; Rohrbach, F.; Gottschalk, A.;
Mayer, G.; Heckel, A. Angew. Chem., Int. Ed. 2012, 51, 8446−8476.
(6) Bandara, H. M. D.; Burdette, S. C. Chem. Soc. Rev. 2012, 41, 1809−
1825.
(7) (a) Kay, E. R.; Leigh, D. A.; Zerbetto, F. Angew. Chem., Int. Ed.
2007, 46, 72−191. (b) Balzani, V.; Credi, A.; Venturi, M. Molecular
Devices and Machines: Concepts and Perspectives for the Nanoworld; Wiley-
VCH: Weinheim, Germany, 2008. (b) Stoddart, J. F. Chem. Soc. Rev.
2009, 38, 1802−1820. (d) Feringa, B. L.; Browne, W. R., Eds. Molecular
Switches; Wiley-VCH: Weinheim, Germany, 2011.
(8) (a) Brash, D.; Rudolph, J.; Simon, J.; Lin, A.; Mckenna, G.; Baden,
H.; Halperin, A.; Ponten, J. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 1012−
10128. (b) Tamai, T. K.; Vardhanabhuti, V.; Foulkes, N. S.; Whitmore,
D. Curr. Biol. 2004, 14, R104−R105. (c) Banerjee, G.; Gupta, N.;
Kapoor, A.; Raman, G. Cancer Lett. 2005, 223, 275−284.
Figure 6. UV/vis spectra following (a) an acetonitrile/PBS buffer (1:1)
solution and (b) an acetonitrile/PBS buffer (1:1) with 10 mM reduced
glutathione solution of the azo-BF2 complex 4 kept under continuous
irradiation with 710 nm light at 25 °C. The interval between each scan is
4 h.
(9) (a) Sadovski, O.; Beharry, A. A.; Zhang, F.; Woolley, G. A. Angew.
Chem., Int. Ed. 2009, 48, 1484−1486. (b) Wegner, H. A. Angew. Chem.,
Int. Ed. 2012, 51, 4787−4788.
and 6 (Figures S46−S52 in the Supporting Information)
indicating that this is a general trend. This finding corroborates
our speculation that strengthening the BN bonds (in this case the
B(1)−N(2) bond, which happens to be the longest/weakest of
the two bonds) contributes to the stability of the system toward
hydrolysis.
In conclusion, we have shown how the para-substitution of
azo-BF2 compounds with electron donating groups leads to
photochromic compounds that can be activated with NIR light.
Structure−property analysis showed that the hydrolysis process
in these systems can be modulated and slowed down using strong
electron donating para-substituents. Moreover, we showed that
the cis isomer is drastically more stable toward hydrolysis than
the trans isomer and that these switches are not susceptible to
reduction by glutathione, most probably because of the
coordination with BF2. These results are very promising and
pave the way for using these BF2-coordinated azo compounds in
photopharmacolgical4 and opto(chemo)genetical5 applications.
(10) (a) Beharry, A. A.; Sadovski, O.; Woolley, G. A. J. Am. Chem. Soc.
2011, 133, 19684−19687. (b) Bleger, D.; Schwarz, J.; Brouwer, A.;
Hecht, S. J. Am. Chem. Soc. 2012, 134, 20597−20600. (c) Samanta, S.;
Beharry, A. A.; Sadovski, O.; McCormick, T. M.; Babalhavaeji, A.;
Tropepe, V.; Woolley, G. A. J. Am. Chem. Soc. 2013, 135, 9777−9784.
(d) Samanta, S.; McCormick, T. M.; Schmidt, S. K.; Seferos, D. S.;
Woolley, G. A. Chem. Commun. 2013, 49, 10314−10316.
(11) (a) Kurihara, M.; Hirooka, A.; Kume, S.; Sugimoto, M.; Nishihara,
H. J. Am. Chem. Soc. 2002, 124, 8800−8801. (b) Venkataramani, S.; Jana,
U.; Dammaschk, M.; Sonnichsen, F. D.; Tuczek, F.; Herges, R. Science
̈
2011, 331, 445−448.
(12) Siewertsen, R.; Neumann, H.; Buchheim-Stehn, B.; Herges, R.;
Nather, C.; Renth, F.; Temps, F. J. Am. Chem. Soc. 2009, 131, 15594−
15595.
(13) Light at 630 nm penetrates only the first 5 mm of tissue, while
light between 700 and 800 nm has a penetration of 1−2 cm. See ref 2.
(14) Wang, L.; Dong, H.; Li, Y.; Xue, C.; Sun, L.-D.; Yan, C.-H.; Li, Q. J.
Am. Chem. Soc. 2014, 136, 4480−4483.
(15) Sun, L.-D.; Wang, Y.-F.; Yan, C.-H. Acc. Chem. Res. 2014, 47,
1001−1009.
(16) Yang, Y.; Hughes, R. P.; Aprahamian, I. J. Am. Chem. Soc. 2012,
134, 15221−15224.
ASSOCIATED CONTENT
* Supporting Information
■
S
(17) The para- and/or ortho-substituted derivatives were synthesized
following the synthetic protocol used in the making of the parent
complex 1. See ref 16.
Experimental procedures, NMR spectra of key compounds,
photoisomerization studies, kinetic measurements, X-ray
crystallography, and computational data. This material is
available free of charge via the Internet at http://pubs.acs.org.
(18) For a NIR absorbing BF2-azo compound that does not switch,
please see: Li, Y.; Patrick, B. O.; Dolphin, D. J. Org. Chem. 2009, 74,
5237−5243.
(19) Mustroph, H. Dyes Pigm. 1991, 15, 129−137.
(20) The nature of the effect of oxygen on the isomerization rate is
under investigation.
(21) Blevins, A. A.; Blanchard, G. J. J. Phys. Chem. B 2004, 108, 4962−
4968.
(22) We are unable to measure the quantum yield of the process
because of the competition with the fast thermal isomerization process
and instrumentation limitations.
(23) Mustroph, H. Dyes Pigm. 1991, 16, 223−230.
(24) (a) Ray, D.; Foy, J. T.; Hughes, R. P.; Aprahamian, I. Nat. Chem.
2012, 4, 757−762. (b) Su, X.; Voskian, S.; Hughes, R. P.; Aprahamian, I.
Angew. Chem., Int. Ed. 2013, 52, 10734−10739. (c) Tatum, L.; Su, X.;
Aprahamian, I. Acc. Chem. Res. 2014, 47, 2141−2149. (d) Su, X.;
Aprahamian, I. Chem. Soc. Rev. 2014, 43, 1963−1981.
(25) The hydrolysis half-life in the glutathione containing solutions is
in general slightly longer than in the regular acetonitrile/PBS buffer
(1:1) mixtures. We currently have no hypothesis to explain this effect.
AUTHOR INFORMATION
Corresponding Author
ivan.aprahamian@dartmouth.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This work was supported by Dartmouth College and the Burke
Research Initiation Award.
■
REFERENCES
■
(1) Kalka, K.; Merk, H.; Mukhtar, H. J. Am. Acad. Dermatol. 2000, 42,
389−413.
(2) Smith, K. C., Ed. The Science of Photobiology; Plenum Press: New
York, 1977.
(3) (a) Irie, M. Chem. Rev. 2000, 100, 1683−1684. (b) Durr, H.; Bouas-
Laurent, H., Eds. Photochromism: Molecules and Systems; Elsevier:
Amsterdam, The Netherlands, 2003.
(4) Velema, W. A.; Szymanski, W.; Feringa, B. L. J. Am. Chem. Soc.
2014, 136, 2178−2191.
(5) Sjulson, L.; Miesenbcç k, G. Chem. Rev. 2008, 108, 1588−1602.
(b) Fehrentz, T.; Schonberger, M.; Trauner, D. Angew. Chem., Int. Ed.
̈
13193
dx.doi.org/10.1021/ja508125n | J. Am. Chem. Soc. 2014, 136, 13190−13193