Photoswitching Azo compounds in vivo with red light

Article abstract of DOI:10.1021/ja402220t

The photoisomerization of azobenzenes provides a general means for the photocontrol of molecular structure and function. For applications in vivo, however, the wavelength of irradiation required for trans-to-cis isomerization of azobenzenes is critical si

Full text of DOI:10.1021/ja402220t

Subscriber access provided by Columbia Univ Libraries
Article
Photoswitching azo compounds in vivo with red light.
Subhas Samanta, Andrew A. Beharry, Oleg Sadovski, Theresa M.
McCormick, Amir Babalhavaeji, Vince Tropepe, and G. Andrew Woolley
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/ja402220t • Publication Date (Web): 10 Jun 2013
Downloaded from http://pubs.acs.org on June 16, 2013
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Photoswitching azo compounds in vivo with red light.
Subhas Samanta‡, Andrew A. Beharry‡, Oleg Sadovski, Theresa M. McCormick, Amirhossein
Babalhavaeji, Vince Tropepe# & G. Andrew Woolley*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Department of Chemistry, University of Toronto, 80 St. George St., Toronto, ON, M5S 3H6, Canada and #Deꢀ
partment of Cell & Systems Biology and Centre for the Analysis of Genome Evolution and Function, University of
Toronto, 25 Harbord St., Toronto, ON, M5S 3G5, Canada.
KEYWORDS azobenzene, photoisomerization, photo-control, optical control
ABSTRACT: The photoisomerization of azobenzenes provides a general means for the photoꢀcontrol of molecular
structure and function. For applications in vivo, however, the wavelength of irradiation required for transꢀtoꢀcis isomꢀ
erization of azobenzenes is critical since UV and most visible wavelengths are strongly scattered by cells and tissues. We
report here that azobenzene compounds in which all four positions ortho to the azo group are substituted with bulky
electronꢀrich substituents can be effectively isomerized with red light (630ꢀ660 nm), a wavelength range that is orders
of magnitude more penetrating through tissue than other parts of the visible spectrum. When the ortho substituent is
chloro, the compounds also exhibit stability to reduction by glutathione enabling their use in intracellular environments
in vivo.
Previously we found that the tetraꢀorthoꢀmethoxy
substituted azobenzene derivative (1-rf) has nꢀπ* transiꢀ
tions in the visible (blue/green) region that are well sepꢀ
arated for the cis and trans isomers making photochemiꢀ
cal switching possible without using UV light.17 Unlike
derivatives in which the πꢀπ* transitions are shifted to
longer wavelengths,18,19 these orthoꢀsubstituted species
showed slow (hours) thermal back reactions.
INTRODUCTION
Photochemical switches are broadly useful tools;
they permit one to use a highly versatile reagent – light –
to direct molecular responses in a reversible (on/off)
manner. A number of molecular classes are known to
undergo photoisomerization.1 The majority of these,
however, require UV light for at least one of the phoꢀ
toswitching directions. Except in certain model organꢀ
isms such as C. elegans and the zebrafish,2 and in speꢀ
cialized organs such as the eye, in vivo single photon
optical applications require red (or near infrared) light to
achieve significant penetration through tissue.3 Naturally
occurring photoswitches such as the phytochromes and
the rhodopsins have evolved to undergo redꢀlight photoꢀ
chemistry and several of these are being engineered as
tools for manipulating molecular processes in vivo.4,5
These photoswitches are limited to protein photoꢀ
control, however, and their requirement for large protein
coꢀfactors limits versatility for nanotechnology applicaꢀ
tions.
Cl
NH
O
NH
O
Br
Cl
X
O
X
O
N
O
N
O
O
N
O
O
N
O
N
N
X
Cl
X
O
O
NH
Cl
NH
Br
4: X = Cl
4-rf: X = H
1: X = Cl
1-rf: X = H
2
NH
O
Br
NH
O
O
Br
X
N
N
N
O
O
N
O
X
Br
O
O
NH
Br
5-rf
NH
3: X = Cl
3-rf: X = H
CHART 1. Tetraꢀortho substituted azobenzene derivaꢀ
tives studied here.
C2ꢀbridged azobenzene species have also been reꢀ
ported that have well separated nꢀπ* transitions and can
be isomerized with visible light.20ꢀ22 In these cases, howꢀ
ever, the trans isomer, which has the longer wavelength
absorption band, is thermodynamically less stable than
the cis isomer. Control of molecular processes in settings
where only long wavelength light can penetrate may be
easier to achieve if light drives production of the therꢀ
modynamically less stable isomer and the system relaxes
thermally in the dark.
The photoisomerization of azobenzene, discovered
in 19376, is a wellꢀdefined and extensively applied proꢀ
cess. The parent molecule is approximately planar in its
thermally stable trans conformation. Irradiation with UV
light produces the cis isomer, which has a bent conforꢀ
mation and a much larger dipole moment. These moꢀ
lecular changes have been used to drive a myriad of funcꢀ
tional outcomes in various settings.7,8 While application
of azo compounds in living cells and tissues has been
achieved in specific cases,2,9ꢀ14 the limitations of using UV
light in vivo have driven efforts to prepare azo comꢀ
pound that operate with longer wavelengths of light.15,16
Recently Aprahamian and colleagues23 and Hecht and
colleagues24 have reported that BF2 coordinated azobenꢀ
zenes and orthoꢀfluoro substituted azobenzenes, respecꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 9
tively, can also exhibit visible light switching properties. highly polar side product was formed, which we could
1
2
3
4
5
6
7
8
Since visible light switching is appealing for use in bioꢀ
logical systems that are sensitive to UV light, we decided
to explore more fully the behavior of this type of phoꢀ
toswitch. We report the discovery that a variety of tetraꢀ
ortho substituted azobenzene compounds can be effecꢀ
tively isomerized with red light. In addition, ortho chloro
substitution confers stability to reduction by glutathione
enabling use of these switches in intracellular environꢀ
ments in vivo.
not characterize. The yields of Suzuki crossꢀcoupling
reactions were also comparatively low. Even though we
used nearly 3ꢀ4 equivalent of arylboronic acids long reacꢀ
tion times, significant amounts of mono coupled prodꢀ
ucts were isolated. Yields of the bisꢀcoupled products
might be improved in by systematic testing of different
solvents and reaction conditions.
O
O
O
NO2
O
NH2
O
g
9
h
i
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Br
O
Br
Br
Br
MATERIALS AND METHODS
11
12
Synthetic methods: The compound 1-rf was prepared
previously.17 The route employed in that case proved difꢀ
ficult to adapt to the synthesis of a chloroacetamidoꢀ
substituted version (1) suitable for crossꢀlinking to pepꢀ
tides. Instead we adopted the synthetic route shown beꢀ
low:
NHR
O
N
O
N
O
O
j
N
O
N
O
O
Br
O
RHN
2
13; R = H
O
O
O
NH2
O
NHTs
O
NHTs
O
a
b
l
k
O2N
O
O
O
O
3; R = COCH2Cl
6
7
N
HN
NH
N
O
O
O
O
NH2
O
3-rf
NH2
O
c
d (i, ii)
N
O
N
SCHEME 2. Synthesis of 3 and 3-rf. (g) HNO3, Ac2O,
O2N
O
O2N
0oC to rt, 4h, yield 19%; (h) Fe, NH4Cl, MeOH, reflux 4 h,
yield 57%; (i) KMnO4ꢀCuSO4.5H2O, dichloromethane,
reflux, 48 h, yield 20%; (j) 4ꢀaminophenylboronic acid
8
9
hydrochloride,
Pd(PPh3)4,
NaHCO3,
anh.
1,2ꢀ
O
N
NHR
dimethoxyethane, 90oC, 24 h, yield 54%; (k) chloroacetyl
chloride, K2CO3, anh. DMF, 0oCꢀrt, 3 h, yield 62%; (l) 4ꢀ
acetamidophenylboronic acid, Pd(PPh3)4, K2CO3, anh.
1,4ꢀdioxane, 110oC, 24 h, yield 33%.
O
N
O
e
1
; R = COCH2Cl
O
f
RHN
10
; R = H
Syntheses of the tetraꢀorthoꢀhalo derivatives (4, 4-rf,
and 5-rf) are outlined in Scheme 3. AgO oxidation of
commercially available 4ꢀacetamidoꢀ2,6ꢀdibromoaniline
led to 5-rf in only 10% yield and several highly polar to
nonꢀpolar side products were produced, which we could
not characterize. One could prepare 5-rf via a diazonium
coupling reaction sequence of four steps as used for 4-rf,
but the overall yield would appear to be less than 10%. In
general, we obtain low yields for diazo coupling reactions
of diꢀorthoꢀsubstituted donors and diꢀorthoꢀsubstituted
acceptors, which we believe is due to steric hindrance
caused by the ortho substituents. Higher yielding synꢀ
thetic approaches to these types of compounds would be
very useful.
SCHEME 1. Synthesis of 1. (a) pꢀToluenesulfonyl
chloride, pyridine, 100oC, 1 h, yield 80 %; (b) NaNO2,
HNO3/AcOH, yield 60%; (c) Conc. H2SO4, 16 h, yield
70%; (d) i) NaNO2, dil. HCl, 0ꢀ5oC, 20 min ii) 3, 5ꢀ
dimethoxyaniline, H2O, sat. NaHCO3 (pH 8ꢀ9), 5oC to rt,
12 h, yield 31%; (e) SnCl2, anh. DMSO, 48 h, yield 7%; (f)
chloroacetyl chloride, triethylamine, CHCl3, 5oC to rt, 12
h, yield 62%.
This approach makes use of the paraꢀdirecting properꢀ
ty of the tosyl group under the specific reaction condiꢀ
tions to cleanly give 7 as originally noted by Wepster.25
Reduction of 9 to give 10 gave low yields likely for the
same reasons that the final compound is sensitive to reꢀ
duction by glutathione (vide infra).
Peptide synthesis and cross-linking: The peptides
listed in Table I were prepared using standard Fmocꢀ
based solidꢀphase peptide synthesis methods and puriꢀ
fied by reverseꢀphase HPLC on a semiꢀpreparative RXꢀ
C8 column (Zorbax, 9.4 mm ID x 255 mm) using a linear
gradient of 10ꢀ65% acetonitrile/H2O (containing 0.1%
trifluoroacetic acid) over a course of 25 min. The molecuꢀ
lar compositions of peptides were confirmed by ESIꢀMS
[M+].
The syntheses of 3, and 3-rf, were achieved as shown
on Scheme 2. In this case a reductive step was avoided;
the azo unit was formed using potassium permanganate
and copper sulfate,26 and elaborated using a Suzuki reacꢀ
tion to give 13. Initial nitration of 1ꢀbromoꢀ3,5ꢀ
dimethoxybenzene led to the production of two isoꢀ
mers. The major product was 1ꢀbromoꢀ3,5ꢀdimethoxyꢀ2ꢀ
nitrobenzene. The production of 12 using KMnO4ꢀ
CuSO4.5H2O was successful but the yield was low; a
ACS Paragon Plus Environment

Page 3 of 9
Journal of the American Chemical Society
(Figure S6). Fluorescence imaging and time courses were
1
2
3
4
5
6
7
8
carried out on an Olympus IX71 microscope with a UPꢀ
lanFL 10x/0.30 NA objective lense. The illumination
source was a 103W/2 short arc mercury lamp (3000 luꢀ
mens). Photoswitching was carried out at room temperaꢀ
ture (~23oC). A fluorescein filter set (Omega Optical Inc.
Alpha Vivid XF100ꢀ3, ex. 450–490 nm, em. 505ꢀ580
nm) was used for imaging the embryos and as the blue
light source to trigger cisꢀtoꢀtrans isomerization. A cusꢀ
tomized Cy5 filter set with excitation filter
(630AF50/25R (606ꢀ650nm)) was used as the red light
source to cause transꢀtoꢀcis isomerization. Exposure to
~5 seconds of red light (nominally (out of focus) 60
mW/cm2) was routinely used to produce the maximum
amount of the cis isomer.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
Red light photoswitching of (1). To explore the
suitability of tetraꢀorthoꢀmethoxy substituted azobenꢀ
zene derivatives for biomolecule photoꢀcontrol, we synꢀ
thesized the thiolꢀreactive derivative 1 suitable for linkꢀ
ing to peptide and protein targets. We used 1 to crossꢀ
link the helixꢀforming peptide FK11, which has Cys resiꢀ
dues spaced i, i+11 in the sequence (Table I). This pepꢀ
tide is wellꢀestablished as a sequence where helixꢀcoil
transitions can be driven by cisꢀtrans isomerization of
azobenzene switches.28
SCHEME 3. Synthesis of 4, 4-rf, and 5. (m) HCHO,
NaHSO3, EtOH/H2O, 70oC, 24 h, yield 84%; (n) i)
NaNO2, H2SO4, AcOH/DMF, 0o C, 2 h, ii) 14, DMF, 0oC
to rt, 72 h, iii) 20% NaOH, 70oC, 2 h, overall yield 26%;
(o) Na2S, 1,4ꢀdioxane/EtOH/H2O, 90oC, 24 h, yield 42%;
(p) chloroacetyl chloride, pyridine/diethyl ether, 0oC to
rt, 1 h, yield 67%; (q) acetic anhydride, pyridine, 0oC to
rt, 12 h, yield 70%; (r) AgO, acetone, rt, 72h, yield 10%.
Surprisingly, attachment to the peptide caused a red
shift in the absorbance of 1 relative to the reference
compound 1-rf in water (λmax 490 vs. 470 nm) (Fig. 1a).
While this λmax is still a long way from the red region of
the spectrum, the nꢀπ* band has a tail past 600 nm (Fig.
1a). Moreover, if one plots the ratio of the absorbance of
the trans and cis isomers of 1 linked to the FK11 peptide,
it is seen that the trans isomer has a larger absorbance
than the cis throughout the wavelength range 500ꢀ700
nm (Fig. 1b). This behavior is not observed with common
azobenzene photoswitches that do not have the tetraꢀ
ortho substitution pattern. We therefore decided to irraꢀ
diate the sample with red (635 nm, 90 mW/cm2) light
and found nearly complete (~98%, based on comparison
with the isolated cis isomer of 1-rf)17 conversion to the
cis isomer (Fig. 1a). As expected, transꢀtoꢀcis isomerizaꢀ
tion of the photoswitch also drove helical folding of the
attached peptide (Fig. 1c,d). We found that irradiation
with 660 nm light of similar intensity (80 mW/cm2) also
produced transꢀtoꢀcis isomerization but at a slower rate
consistent with a weaker absorbance of the photoswitch
at longer wavelengths (Fig. 1a). Because the thermal back
reaction is slow (τ½~ 6h, Table I) even a relatively slow
rate of photochemical transꢀtoꢀcis isomerization producꢀ
es the cis isomer at a competitive rate. Near IR waveꢀ
lengths (760 nm, 1W/cm2) however, proved ineffective in
causing isomerization. The trans isomer can be regenerꢀ
ated (to ~85%) rapidly with blue light (Fig. S1).
The crossꢀlinking reactions of peptides were generally
carried out in 50% DMSO/ water. A solution of 0.5 mM
peptide (freshly purified by HPLC) and 2 mM crossꢀ
linker in 50 mM Tris buffer at pH 8 was stirred at 40oC
under a nitrogen gas atmosphere for 20 h. The progress
of the reaction was monitored using MALDI mass specꢀ
trometry. The reaction was dried under high vacuum,
and the crossꢀlinked peptide was purified by HPLC (SBꢀ
C8 column, as described above) using a linear gradient of
10ꢀ70% acetonitrile/water (containing 0.1% trifluoroaceꢀ
tic acid) over the course of 25 min. The compositions of
crossꢀlinked peptides were confirmed by ESIꢀMS.
Microinjection into zebrafish embryos and fluores-
cence imaging in vivo:
Zebrafish embryos (Danio rerio, AB strain) were
maintained at 28oC and staged as described in Kimmel et
al.27 3 ꢁL of each peptide: Flꢀ(Pro)9ꢀFK11) linked to 4 (in
PBS, 1.0 mM), and the unlinked control peptide Flꢀ
(Pro)9ꢀFK11) (in PBS, 1.3 mM) were backfilled into glass
needles and ~ 3 nL of each were injected into the yolk of
1ꢀ2 cell stage embryos (n > 25 for each group). Embryos
were maintained in the dark. Images were captured with
a Lumar V12 fluorescent stereomicroscope (Zeiss) with
an Axiocam MRc digital camera and software. Developꢀ
ment appeared normal as determined by comparing inꢀ
jected embryos to unꢀinjected embryos at each stage
From a biological applications standpoint, the imꢀ
portance of being able to drive isomerization with red
rather than green light can be appreciated simply by
shining a red vs. a green laser pointer at one’s hand; the
absorption coefficients of oxygenated mammalian tissue
differ by a factor of ~100 between 635 nm (red) and 535
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 9
nm (green) wavelengths.3 Red light driven isomerization
thus makes optical manipulation of targets in higher orꢀ
ganisms possible.
gree of coꢀplanarity of the aromatic rings. This hypotheꢀ
1
2
3
sis is consistent with the finding that the position of the
nꢀπ* transition of the trans isomer of 1-rf is temperature
sensitive in water.17
Red light photoswitching occurs with a variety
of peptides. Concerned that this result was due to a
peculiarity of the interaction of 1 with the specific pepꢀ
tide sequence FK11, we made a series of analogues of the
peptide that varied greatly in helical propensity and/or
amino acid composition near the photoswitch attachꢀ
ment site (Table 1). The λmax of the nꢀπ* transition of the
trans isomer varied significantly among these peptides
(Table 1). However, all of these, even those with extremeꢀ
ly weak absorbance in the red region of the spectrum,
could be isomerized effectively with red light (Fig. S1).
Red light driven transꢀtoꢀcis isomerization resulted in a
decrease in the αꢀhelix content of the attached peptide in
each case (Fig. S2). As expected, the rate of transꢀtoꢀcis
conversion with red light of a given intensity and a given
wavelength depended on the molar extinction coefficient
of the switch/target combination.
4
5
6
7
8
9
Table 1. Sequences and properties of crossꢀlinked
peptides
Peptide Sequence
λmax
(πꢀπ*)
λmax
(nꢀπ*)
Cis
τ½
#
FK11
pAib
EdblꢀA
R10ꢀA
R5ꢀA
R5R10A
Eꢀrich
A8ꢀG
WGEACAREAAAREAACRQ
363 nm
367 nm
350 nm
372 nm
362 nm
373 nm
366 nm
360 nm
359 nm
359 nm
358 nm
357 nm
327 nm
369 nm
489 nm
474 nm
482 nm
490 nm
483 nm
493 nm
491 nm
477 nm
471 nm
470 nm
480 nm
479 nm
480 nm
471 nm
6 h
13 h
12 h
2.2 h
4.3 h
2.0 h
7.2 h
17 h
27 h
ꢀ
36 h
45 h
3.5h
ꢀ
WGEACAABAABAABACRQ
WGEACARAAAARAAACRQ
WGEACAREAAAAEAACRQ
WGEACAAEAAAREAACRQ
WGEACAAEAAAAEAACRQ
WGEACAREAEAREAECRQ
WGEACAREAGAREAACRQ
WGEACAREGGGREAACRQ
EACARVBAACEAAARQ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
pG
JRKꢀ7
SSꢀ17*
FK11*
FK11**
GSH
WGACEAAAREAAAREAAARECAAQ
WGEACAREAAAREAACRQ
WGEACAREAAAREAACRQ
γECG (X2)
“B” denotes an Aib residue; all peptides are Nꢀacetylated and Cꢀ
amidated *crossꢀlinked with 3, ** crossꢀlinked with 4. #measured
at 37°C
FIGURE 2. Xꢀray crystal structures of 2 without solvent
(a) and with coꢀcrystallized chloroform molecules (b). In
the solventꢀfree form, the rings are highly twisted at an anꢀ
gle of 65 degrees. In the second crystal form, the H atoms of
each of two chloroform molecules interact with an azo niꢀ
trogen atom and oxygen atoms from two methoxy groups.
These interactions result in a much more planar ring sysꢀ
tem. (c) Calculated (TDꢀDFT, B3LYP/6ꢀ311++G**) spectra
of 2 based on the Xꢀray structures before and after optimiꢀ
zation (opt). Only the nonꢀplanar structures show long
wavelength nꢀπ* absorption bands.
FIGURE 1. (a) UVꢀVis spectra of the peptide FK11
(WGEACAREAAAREAACRQ) crossꢀlinked with 1. Darkꢀ
adapted (trans) spectra and redꢀlight irradiated (635 nm, 90
mW/cm2 for 5 min or 660 nm, 80 mW/cm2 for 25 min)
spectra in 5 mM NaPi buffer pH 7.0, 20oC. (b) Ratio of the
absorbance of the darkꢀadapted crossꢀlinked peptide to that
of the redꢀirradiated sample (trans/cis) (c) Model of the
FK11 peptide crossꢀlinked with 1 in trans and cis states. (d)
Circular dichroism (CD) spectra of 1ꢀcrossꢀlinked FK11 indiꢀ
cate a red light driven decrease in helix content.
A specific example of the effects of Hꢀbonding on the
absorption properties of a tetraꢀorthoꢀmethoxy azobenꢀ
zene compound was encountered during the preparation
of derivatives of 1. Two distinct crystal forms of paraꢀ
bromo tetraꢀorthoꢀmethoxy azobenzene (2) were isolatꢀ
ed from a chloroform/methanol solution. In one form,
the aromatic rings are highly twisted at an angle of 65
degrees (Fig. 2a). In the second form, the H atoms of
each of two chloroform molecules Hꢀbond with an azo
nitrogen atom and oxygen atoms from two methoxy
groups (Fig. 2b). These interactions result in a nearly
planar ring system. Timeꢀdependent density functional
theory calculations (B3LYP/6ꢀ311++G**) on each of theꢀ
se structures confirm large changes in the position of the
nꢀπ* (and πꢀπ*) bands (Fig. 2c) and indeed the two crysꢀ
tal forms have distinct colors (orange, yellow).
Factors affecting the position of the absorption
bands. Time dependent density functional theory calcuꢀ
lations suggest that the red shift of the nꢀπ* band of the
trans isomer arises due to a destabilization of the HOMO
centered on the azo nitrogen lone pairs because of the
close proximity of the methoxy oxygen lone pairs.17 We
propose that the variability of the position of the trans nꢀ
π* band among 1ꢀcrossꢀlinked peptides is due to variaꢀ
tion in solvent Hꢀbonding to the azo nitrogens and
methoxy groups and concomitant alteration in the deꢀ
ACS Paragon Plus Environment

Page 5 of 9
Journal of the American Chemical Society
the cis form with red light (Fig. S3,S4). Crossꢀlinking of
1
2
3
the peptide FK11 with 4 again led to photocontrol of heꢀ
lix content upon red light illumination (Fig. 4). The reꢀ
sult is therefore quite general.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
FIGURE 4. (a) UVꢀVis spectra and (b) CD spectra of comꢀ
pound 4 crossꢀlinked to peptide FK11 (see Table 1 for seꢀ
quence) under the indicated irradiation conditions in 10
mM sodium phosphate buffer, pH 7, 20oC.
We did not synthesize the corresponding fluoro derivaꢀ
tive since it was not predicted (B3LYP, 6ꢀ31G*) to show a
large separation of nꢀπ* bands. However, Hecht and colꢀ
leagues have reported that, in fact, this derivative does
show separation of the nꢀπ* bands, particularly when an
electron withdrawing para substituent is present.24 It
would be of significant interest to test if this compound
too undergoes red light switching. The trans isomer of
the BF2 substituted azo compound reported by Apraꢀ
hamian et al has a higher extinction coefficient than the
cis isomer at wavelengths greater than 600 nm, and so
may also be expected to undergo red light driven isomerꢀ
ization. The surprising enhancement of the thermal reꢀ
laxation rate reported for this compound in the presence
of oxygen may complicate biological applications, howꢀ
ever.23
FIGURE 3. (a) UVꢀVis spectra of FK11 crossꢀlinked with 3.
Darkꢀadapted (trans) spectra and redꢀlight irradiated (635
nm, 90 mW/cm2 for 5 min) spectra in 5 mM NaPi buffer pH
7.0, 20oC. (b) Xꢀray structure of trans 3-rf. (c) Model of
FK11 crossꢀlinked with 3 in trans and cis states. (d) CD
spectra of 3ꢀcrossꢀlinked FK11 showing a red light driven
increase in helix content.
Resistance to reduction by glutathione. A second
concern, besides the wavelength, for photoꢀcontrol in
vivo is the chemical and metabolic stability of the phoꢀ
toswitch. Azo compounds may undergo photobleaching,
reduction by intracellular glutathione or may be enzyꢀ
matically metabolized. These processes depend critically
on the specific structure of the azo compound and its
redox potential.13,31 The tetraꢀortho methoxy compounds
1, and 3 were found to be stable to photobleaching but
underwent slow reduction (τ1/2~1h) in the presence of 10
mM reduced glutathione in their trans isomeric states
(Fig. S5). This sensitivity to reduction means that only
extracellular use and short term intracellular use are
possible in vivo.
The generality of red-light switching with tetra-
ortho substituted azo compounds. To further test
the generality of the tetramethoxy scaffold as a redꢀlight
switch, we synthesized the extended derivatives 3 and 3ꢀ
rf (Scheme 2). The trans isomer of compound 3 has an
endꢀtoꢀend distance of ~20 Å and this changes to ~12 Å
upon isomerization to cis.29 Photoꢀcontrol of the strucꢀ
ture of larger peptides and proteins is likely to be faciliꢀ
tated longer rigid photoswitchable crossꢀlinkers of this
type.30 The Xꢀray structure of the trans isomer of 3-rf
showed the compound was also nonꢀplanar (Fig. 3b). A
redꢀshifted nꢀπ* band and effective red light driven
transꢀtoꢀcis isomerization (93% cis) was observed with
3-rf in solution (Fig. 3a, Fig. S3, S4). Attachment of phoꢀ
toswitch 3 to peptides with Cys residues spaced i, i+17
(trans) and i, i+11 (cis) was found to enable redꢀlight
driven helixꢀcoil transitions, either increasing or deꢀ
creasing helix content depending on the Cys spacing
(Fig. 3c,d, Fig. S2).
The tetraꢀortho chloroꢀsubstituted azobenzene comꢀ
pound 4 however, proved resistant to reduction by glutaꢀ
thione in both isomeric forms (Fig. 5) and was also stable
to photobleaching (Fig. 5). This result was unexpected
since chloro substituents are more electronꢀwithdrawing
than methoxy substituents and so might be expected to
increase susceptibility to reduction. The nonꢀplanar conꢀ
formation of the trans isomer decreases the electron doꢀ
nating ability of the methoxy groups via resonance but
the methoxy group is still less electron withdrawing than
the chloro group via induction.32
We then tested whether oxygen substituents were
strictly required by synthesizing ortho chloroꢀ and broꢀ
moꢀ derivatives (4-rf, 5-rf)(Scheme 3). The chloroꢀand
bromoꢀ species also gave trans isomers with red shifted
nꢀπ* transitions that could be effectively isomerized to
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 9
The mechanism of reduction of azo compounds by thiꢀ as shown in Figure 6(e). This may occur prior to, or in
1
2
3
4
ols has been studied by Kosower and colleagues who deꢀ
termined that, for diazenedicarboxylic acid derivatives,
the thiol attacks one azo nitrogen atom and the second
azo nitrogen atom becomes protonated to give a sulfenyl
hydrazine derivative.33
concert with, nucleophilic attack.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
FIGURE 6. Protonation of the azo group of 1-crossꢀlinked
FK11 (a) becomes significant below pH 4 whereas no protoꢀ
nation is evident for 4ꢀcrossꢀlinked FK11 even at pH 1.3. The
pHs are indicated in the figure. (b). Calculated minimum
energy structures (B3LYP/6ꢀ311++G**) of 1-rf (c) and 4-rf
(d) indicate that protonation leads to a planar structure for
1-rf. The protonated 1ꢀrf may be represented as shown in
(e) and is expected to be attacked by thiols more easily than
the unprotonated form.
FIGURE 5. Stability to photobleaching and to reduction by
glutathione of 4ꢀcrossꢀlinked FK11 (a) Multiple cycles of
red/blue light switching in 10 mM phosphate buffer, pH 7.
(b) Photoswitch 4, in the trans state, crossꢀlinked to FK11
was incubated in 10 mM reduced glutathione for 16 h at
25oC. The first scan is shown as a solid black line. Intermeꢀ
diate scans are shown as red lines. The final scan is shown
as a black dotted line. No change is evident. Normal
red/blue light photoswitching (as in Fig. 4) was observed
after this incubation. (c) Photoswitch 4, in the cis state,
crossꢀlinked to FK11 was incubated in 10 mM reduced glutaꢀ
thione for 16 h at 25oC. Thermal cisꢀtoꢀtrans isomerization
also occurs over this time period. Normal red/blue light
photoswitching was observed after this incubation.
Red-light photoswitching in vivo. To examine the
possibility of photoswitching with red light with phoꢀ
toswitch 4 directly in vivo, a fluorescent reporter seꢀ
quence (Fluoresceinꢀ[D]ꢀPPPPPPPPPEACAREAAAREAꢀ
ACRQ; Flꢀ(Pro)9ꢀFK11) was designed. Dꢀamino acids
were used to avoid degradation of the reporter peptide
by proteases in vivo.13 Fluorescein was chosen since its
emission spectrum overlaps to different extents with the
nꢀπ* bands of the trans and cis isomers of 4 so fluoresꢀ
cence should be quenched to different extents by the
trans and cis isomers via fluorescence resonance energy
transfer. Blue light was expected to excite fluorescein
and simultaneously drive cisꢀtoꢀtrans isomerization of
the photoswitch. Also, intense redꢀlight was expected to
drive transꢀtoꢀcis isomerization without exciting (or
bleaching) fluorescein.
The azo compounds studied here are more electron
rich than the compounds studied by Kosower. We hyꢀ
pothesize that thiol attack in the present case is facilitatꢀ
ed by prior, or concomitant protonation of the azo group
to enhance its electrophilicity. If this is the case, the enꢀ
hanced rate of reduction of the tetraꢀortho methoxy
compound by glutathione could be due a higher effective
pKa of the azo group. We determined the ease of protoꢀ
nation of the tetraꢀortho methoxy (1) and tetraꢀortho
chloro (4) substituted FK11 peptides by measuring UVꢀ
Vis spectra as a function of pH (Fig. 6, panels a,b). These
data confirm that the tetraꢀortho methoxy species is
more easily protonated than the tetraꢀortho chloro (4)
substituted FK11 peptide. The conformations of the proꢀ
tonated crossꢀlinkers were determined using computaꢀ
tional methods (B3LYP/6ꢀ311++G**) and are shown in
Figure 6 (panels c,d). The tetraꢀortho methoxy comꢀ
pound is planar while the tetraꢀortho chloro species reꢀ
mains twisted. The planar conformation of the protonatꢀ
ed tetraꢀortho methoxy compound again implicates Hꢀ
bonding to the methoxy groups as playing an important
role in the behavior of these compounds. Protonation of
an azo nitrogen is expected to facilitate attack by thiols
Figure 7 shows models of the fluorescent reporter pepꢀ
tide crossꢀlinked with 4. A solution of the peptide was
first irradiated with red light to convert the photoswitch
to the cis isomeric state. The solution was then exposed
to blue light and the time course of fluorescence emisꢀ
sion was monitored. As expected, blue light produced a
time dependent fluorescence decrease as it converted cis
species into trans (Fig. 7b). The rate of this process deꢀ
pended on the intensity of the blue light source (which
was varied by varying the excitation slit width in the fluꢀ
orescence spectrophotometer in Fig. 7b).
ACS Paragon Plus Environment

Page 7 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
FIGURE 7. (a) Models of the fluorescent reporter peptide
fluoresceinꢀ[ ](Pro)9ꢀFKꢀ11 crossꢀlinked with 4 in the trans
D
(blue irradiated and cis (red irradiated) states. The trans
state results if greater quenching of fluorescein emission.
(b) In vitro photoswitching in 20% DMSO in phosphate
buffered saline. Samples were preꢀirradiated with red light
to convert them to the cis state. Samples were then irradiatꢀ
ed with blue light (460 nm) and fluorescence emission at
520 nm was monitored as a function of time. Blue light exꢀ
cites fluorescein and simultaneously switches the azo group
from cis to trans. As the blue light intensity was increased
by increasing the excitation monochromator slit width (inꢀ
dicated), the rate of cisꢀtoꢀtrans isomerization increased. No
fluorescence change was seen if the sample was not first
converted to the cis isomer with red light (black trace).
FIGURE 8. (a) Fluorescence image of a zebrafish embryo
(early pharyngula stage/~30 hpf) containing the fluorescent
reporter peptide fluoresceinꢀ[D](Pro)9ꢀFKꢀ11 crossꢀlinked
with 4. (b) Measurements of fluorescence intensity at the
zones indicated by black boxes in (a), as a function of time
during blue light irradiation. As was observed in vitro (Fig.
6), a change in fluorescence emission was seen only if the
peptide was first switched with red light (indicated by red
bars). No preꢀirradiation with red light before blue light
produces no exponential (switching) transient, only a backꢀ
ground slope due to fluorescein reporter photobleaching.
Note the blue light intensity is ~1000 fold higher in here
than in Figure 6. (c) Fluorescence image of a zebrafish emꢀ
bryo (~30 hpf) containing the fluorescent reporter peptide
To test photoisomerization in vivo, the reporter pepꢀ
tide was microinjected into zebrafish embryos. The
zebrafish provides a convenient model system for in vivo
testing of photoswitchable peptides and proteins.13 First,
microinjection into the yolk at the yolkꢀcell interface
during embryogenesis ensures the photoswitchable bioꢀ
molecule is carried throughout the organism to all types
of cells and tissues. Second, the optical transparency of
the fish embryos enables convenient fluorescence moniꢀ
toring of the photoswitch using a blueꢀlight absorbꢀ
ing/greenꢀlight emitting dye, and finally, blue light can
be used to reset the switch in order to rapidly test many
cycles of photoswitching. Zebrafish embryos (n>25 for
each group) were injected at the 1ꢀ2 cell stage with the
Flꢀ(Pro)9ꢀFK11 peptide crossꢀlinked with 4 or an unꢀ
crossꢀlinked control peptide and subsequently imaged at
5.5 hours post fertilization (hpf) or 24 hpf. In all cases,
there was no evidence of gross morphological changes or
delays in the rate of development indicating no apparent
peptide toxicity (Fig. S6). Robust fluorescence changes
confirming red light/blue light switching were seen at
least until the early pharyngula period of development27
(~30 h)(Fig. 8a,b) The fluorescently labelled peptide
fluoresceinꢀ(Pro)9ꢀFK11 lacking a photoswitch did not
show any time dependent emission, only a linear bleachꢀ
ing (Fig. 8c,d).
fluoresceinꢀ[D](Pro)9ꢀFK11 without a photoswitch. (d)
Measurement of fluorescence intensity of the nonꢀ
switchable peptide as a function of time during blue light
irradiation after preꢀirradiation with red light as in (b).
CONCLUSIONS
As a class, the tetraꢀortho substituted azobenzene
compounds described here extend the versatility of azoꢀ
benzene based switches for in vivo use by providing the
capability of conformational control using red light. Red
light can effectively penetrate tissue in most organisms
and red light absorbing compounds can permit studies
on whole living animals.34,35 These red light switchable
azo compounds can be readily synthesized and derivatꢀ
ized for attachment to a wide variety of targets. Their
small size combined with their robust and predictable
conformational changes can be used to rationally design
a wide range of light driven molecular responses.
ASSOCIATED CONTENT
Supporting Information. Synthetic procedures, compuꢀ
tational details, NMR data and further UV/Vis and CD data.
This material is available free of charge via the Internet at
http://pubs.acs.org
AUTHOR INFORMATION
Corresponding Author
* G. Andrew Woolley (awoolley@chem.utoronto.ca)
Author Contributions
‡These authors contributed equally
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 9
(28) Ihalainen, J. A.; Paoli, B.; Muff, S.; Backus, E. H.;
Funding Sources
1
2
3
4
5
6
7
8
Bredenbeck, J.; Woolley, G. A.; Caflisch, A.; Hamm, P. Proc.
Natl. Acad. Sci. U. S. A. 2008, 105, 9588-93.
(29) Standaert, R. F.; Park, S. B. J. Org. Chem. 2006, 71, 7952-
66.
(30) Beharry, A. A.; Chen, T.; Al-Abdul-Wahid, M. S.; Samanta,
S.; Davidov, K.; Sadovski, O.; Ali, A. M.; Chen, S. B.; Prosser, R.
S.; Chan, H. S.; Woolley, G. A. Biochemistry 2012, 51, 6421-31.
(31) Levine, W. G. Drug. Metab. Rev. 1991, 23, 253-309.
(32) Ritchie, C. D.; Sager, W. F. Prog. Phys. Org. Chem. 1964, 2,
323-400.
This research has been supported by the Natural Sciences
and Engineering Research Council of Canada and the Naꢀ
tional Institutes of Health (USA)(R01 MH086379)
ABBREVIATIONS
CDꢀcircular dichroism; IR –infrared; UVꢀultraviolet.
REFERENCES
(1) Molecular switches; 2nd Ed. ed.; Feringa, B.; Browne, W. R.,
Eds.; Wiley-VCH, 2011.
(2) Wyart, C.; Del Bene, F.; Warp, E.; Scott, E. K.; Trauner, D.;
Baier, H.; Isacoff, E. Y. Nature 2009, 461, 407-10.
(3) Wang, L. V.; Wu, H.-i. Biomedical optics: Principles and
imaging; Wiley, 2007.
(4) Fenno, L.; Yizhar, O.; Deisseroth, K. Annu. Rev. Neurosci.
2011, 34, 389-412.
(5) Levskaya, A.; Weiner, O. D.; Lim, W. A.; Voigt, C. A. Nature
2009, 461, 997-1001.
(6) Hartley, G. S. Nature 1937, 140, 281-281.
(7) Beharry, A. A.; Woolley, G. A. Chem. Soc. Rev. 2011, 40,
4422-37.
(8) Yager, K. G.; Barrett, C. J. J. Photochem. Photobiol. A. Chem.
2006, 182, 250-261.
(9) Hoppmann, C.; Schmieder, P.; Domaing, P.; Vogelreiter, G.;
Eichhorst, J.; Wiesner, B.; Morano, I.; Ruck-Braun, K.; Bey-
ermann, M. Angew. Chem. Int. Ed. Engl. 2011, 50, 7699-702.
(10) Volgraf, M.; Gorostiza, P.; Numano, R.; Kramer, R. H.;
Isacoff, E. Y.; Trauner, D. Nat. Chem. Biol. 2006, 2, 47-52.
(11) Polosukhina, A.; Litt, J.; Tochitsky, I.; Nemargut, J.; Sychev,
Y.; De Kouchkovsky, I.; Huang, T.; Borges, K.; Trauner, D.; Van
Gelder, R. N.; Kramer, R. H. Neuron 2012, 75, 271-82.
(12) Fehrentz, T.; Schonberger, M.; Trauner, D. Angew. Chem.
Int. Ed. 2011, 50, 12156-12182.
(13) Beharry, A. A.; Wong, L.; Tropepe, V.; Woolley, G. A. An-
gew. Chem. Int. Ed. Engl. 2011, 50, 1325-7.
(14) Zhang, F.; Timm, K. A.; Arndt, K. M.; Woolley, G. A. An-
gew. Chem. Int. Ed. Engl. 2010, 49, 3943-6.
(15) Mourot, A.; Kienzler, M. A.; Banghart, M. R.; Fehrentz, T.;
Huber, F. M.; Stein, M.; Kramer, R. H.; Trauner, D. ACS Chem.
Neurosci. 2011, 2, 536-43.
(16) Wegner, H. A. Angew. Chem. Int. Ed. 2012, 51, 4787-4788.
(17) Beharry, A. A.; Sadovski, O.; Woolley, G. A. J. Am. Chem.
Soc. 2011, 133, 19684-19687.
(18) Sadovski, O.; Beharry, A. A.; Zhang, F.; Woolley, G. A.
Angew. Chem. Int. Ed. Engl. 2009, 48, 1484-6.
(19) Li, Y.; Patrick, B. O.; Dolphin, D. J. Org. Chem. 2009, 74,
5237-5243.
(20) Samanta, S.; Qin, C. G.; Lough, A. J.; Woolley, G. A. An-
gew. Chem. Int. Ed. 2012, 51, 6452-6455.
(21) Siewertsen, R.; Neumann, H.; Buchheim-Stehn, B.; Herges,
R.; Nather, C.; Renth, F.; Temps, F. J. Am. Chem. Soc. 2009, 131,
15594-5.
(22) Sell, H.; Nather, C.; Herges, R. Beilstein J. Org. Chem. 2013,
9, 1-7.
(23) Yang, Y.; Hughes, R. P.; Aprahamian, I. J. Am. Chem. Soc.
2012, 134, 15221-15224.
(24) Bleger, D.; Schwarz, J.; Brouwer, A. M.; Hecht, S. J. Am.
Chem. Soc. 2012, 134, 20597-600.
(25) Wepster, B. M. Rec. Trav. Chim. Pays-Bas 1954, 73, 809-
818.
(26) Merino, E. Chem. Soc. Rev. 2010, 40, 3835-53.
(27) Kimmel, C. B.; Ballard, W. W.; Kimmel, S. R.; Ullmann, B.;
Schilling, T. F. Dev. Dyn. 1995, 203, 253-310.
9
(33) Kosower, E. M.; Kanety-Londner, H. J. Am. Chem. Soc.
1976, 98, 3001-7.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(34) Ntziachristos, V. Annu. Rev. Biomed. Eng. 2006, 8, 1-33.
(35) Shcherbo, D.; Shemiakina, II; Ryabova, A. V.; Luker, K. E.;
Schmidt, B. T.; Souslova, E. A.; Gorodnicheva, T. V.; Strukova,
L.; Shidlovskiy, K. M.; Britanova, O. V.; Zaraisky, A. G.; Luky-
anov, K. A.; Loschenov, V. B.; Luker, G. D.; Chudakov, D. M.
Nat. Methods 2010, 7, 827-9.
ACS Paragon Plus Environment

Page 9 of 9
Journal of the American Chemical Society
TOC artwork
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
ACS Paragon Plus Environment