Induction of intracellular reductive stress with a photoactivatable phosphine probe

Article abstract of DOI:10.2533/chimia.2018.241

Reductive stress is a condition present in cells that have an increased concentration of reducing species, and it has been associated with a number of pathologies, such as neurodegenerative diseases and cancer. The tools available to study reductive stres

Full text of DOI:10.2533/chimia.2018.241

Laureates: Junior Prizes of the sCs faLL Meeting 2017
CHIMIA 2018, 72, No. 4 241
doi:10.2533/chimia.2018.241
Chimia 72 (2018) 241–244 © Swiss Chemical Society
Induction of Intracellular Reductive Stress
with a Photoactivatable Phosphine Probe
§
Alina Tirla * and Pablo Rivera-Fuentes*
§
SCS-DSM Award for best poster presentation in Chemical Biology
Abstract: Reductive stress is a condition present in cells that have an increased concentration of reducing spe-
cies, and it has been associated with a number of pathologies, such as neurodegenerative diseases and cancer.
The tools available to study reductive stress lack both in selectivity and specific targeting and some of these
shortcomings can be addressed by using photoactivatable compounds. We developed a photoactivatable phos-
phonium probe, which upon irradiation releases a fluorescent molecule and a trialkyphosphine. The probes can
permeate through the plasma membrane and the photoreleased phosphine can induce intracellular reductive
stress as proven by the detection of protein aggregates.
Keywords: Photoactivation · Protein aggregation · Reductive stress · Tributylphosphine
intensively,[6] there is still a need for bio-
logical tools to investigate reductive stress.
Pharmacological agents, such as N-acetyl-
1
1
. Introduction
.1 Reductive Stress
Redox homeostasis is essential for ꢀ-cysteine[ or dithiothreitol (DTT)
2]
[7]
can
regulating fundamental processes such as be used to effectively induce reductive
cellular signaling pathways, transcription- stress, but these compounds lack selectiv-
al and post-transcriptional activities, and ity and spatial resolution.
protein folding.[ Redox homeostasis can
1]
be defined as an intracellular equilibrium 1.2 Photoactivatable Compounds
between oxidative and reducing species.
Photoactivatable molecules (often
A balanced redox environment is main- called photoremovable or ‘photocaged’)
Alina Tirla obtained her MSc in me- tained by the reduced to oxidized ratios are an important class of compounds as
dicinal chemistry from the University of of redox molecular duos, such as nicotin- they provide spatial and temporal control
Glasgow in 2015. During her studies, she amide adenine dinucleotide (phosphate) for the release of a variety of chemicals.[
8,9]
+
did a one-year internship at Roche, Basel, (NAD(P)H/NAD (P)) and glutathione A photoactivatable probe is composed of
where she developed small molecules for (GSH/GSSG).[ Th e reduced (GSH) and
2,3]
a photocleavable protecting group (PPG)
the treatment of neurodegenerative dis- oxidized (GSSG) glutathione couple is the and an inactive molecule of interest, which
eases. In 2015, she started her PhD in the major redox buffer inside the cell and it is is either directly covalently attached to the
laboratory of Prof. Pablo Rivera-Fuentes at also responsible for maintaining individual PPG or through a functional group.[ By
10]
E TH Zurich. Her research interests include
redox environments needed in some organ- irradiating at a specific wavelength, the
[
2]
the synthesis and implementation of chem- elles.
ical probes to study redox biology.
protecting group is removed and the mol-
For example, the cytosol needs, in ecule becomes active. In 1977, Engels and
general, a more reducing environment for Schlaeger were presumably the first to
the reversible oxidation and reduction of employ this strategy in a biological sys-
the cytosolic protein thiols and it has a tem by photoreleasing cyclic adenosine
range of GSH/GSSG of 30:1 to 100:1.[
2]
[11]
monophosphate (cAMP).
Soon after,
Mitochondria also need a more reductive Hoffman and co-workers described the
potential and the reported range of GSH/ photolytic release of adenosine 5'-triphos-
+
+
GSSG is from 20:1 to 40:1. In contrast, the phate (A TP ) for use in studying the Na /K
endoplasmic reticulum (ER) has a more pump.[ Th ese reports paved the way for
oxidative environment necessary for the a new research field and, over the years,
12]
folding of secretory proteins through the there have been numerous reports of pho-
[
2]
[13]
formationofdisulfidebridges. Th erefore,
toactivatable bioagents such as ions,
[14]
[15]
in the ER, the ratio of GSH/GSSG varies neurotransmitters, lipids, nucleic ac-
from 1:1 to 3:1 with millimolar concentra- ids,[ peptides, and proteins.
16]
[17]
[18]
tions of both components present.
In general, the properties of the photo-
When there is a shift in this balance, removable protecting group (PPG) can be
oxidative or reductive stress can occur, adjusted based on the desired application,
and these processes are linked to many but a good PPG should comply with the
*
Correspondence: A. Tirla, Prof. Dr. P. Rivera-Fuentes
Laboratorium für Organische Chemie, ETH Zurich,
Vladimir-Prelog-Weg 3, CH-8093 Zurich
E-mail: alina.tirla@org.chem.ethz.ch; pablo.rivera-
fuentes@org.chem.ethz.ch
[8,9]
pathologies, such as diabetes, inflamma- following criteria:
A) Th e photoreac-
tion, and neurodegenerative diseases.[
4,5]
tion should be clean and occur with a high
www.rivera-fuentes.chem.ethz.ch
Whereas oxidative stress has been studied quantum yield of release, Φ ; B) the chro-
rel

242 CHIMIA 2018, 72, No. 4
Laureates: Junior Prizes of the sCs faLL Meeting 2017
mophore should have strong absorption at (Scheme 1).[23] Th e complex contained a
a trialkylphosphine, however, was pub-
wavelengths higher than 300 nm, where photoswitchable o-tosylamide azobenzene lished in 2012 by Zhuang and co-workers
there is weaker absorption of light by the ligand and, following irradiation with 406 (Scheme 4).[
25]
biological environment and less phototox- nm light, changed its conformation from
icity; C) the PPGs should be stable and sol- the Z to the E isomer. Th is isomerization
Th e authors described that tris(2-car-
boxyethyl)phosphine ( TC EP), a water-sol-
uble in the target media (most often aque- released quantitative amounts of triphen- uble phosphine, quenched the fluorescence
ous, in case of biological applications); lyphosphine, which was used as catalyst of the carbocyanine fluorophore Cy5, fol-
D) the photochemical byproducts should for the initiation of an aza-Morita–Baylis– lowing the 1,4-addition of the phosphine
not be toxic and should not absorb at the Hillman reaction.
wavelength of irradiation; and E) to facili- Photoactivatable
to the polymethylene bridge of the dye.
triarylphosphines Th e fluorescence of the dye could be re-
tate monitoring the photoreaction, one of were described in the context of light-ac- covered after irradiation with UV light.
the photoproducts should be fluorescent.
tivated Staudinger–Bertozzi ligations.[
10,24]
Even though the Cy5– TC EP adduct is not
Using different PPGs, a multitude of As seen in Scheme 2, Lam and co-workers thermally stable, the TC EP-induced fluo-
functional groups can be photoactivated, reported the cleavage of a triarylphosphine rescence quenching was successfully ap-
such as carboxylates, phosphates, sulfo- from 9-anthracenylmethyl diphenlyphos- plied in super-resolution microscopy and
[
8,19]
[24]
nates, alcohols, and thiols.
phonium chloride. Th e released phos- cell internalization assays.
Photoactivatable phosphines were first phinothioester was subsequently used as
described by Weiss and co-workers in the reagent in a Staudinger–Bertozzi ligation,
1
970s. Th ey reported the release of a free which upon reaction with an azide formed 2. Results and Discussion
triphenylphosphine following a photoly- an amide and triarylphosphine oxide as
sis reaction, at 300 nm, of either a phos- products.
TC EP and other trialkylphosphines
[
20]
[21]
phorane or a triphenylphosphonium
Shah et al. independently reported the are known to reduce biological disulfide
compound. Peranovich et al. then reported photorelease of a triarylphosphine and bonds.[ In an intracellular environment,
26]
that irradiation of 9-anthracenenylmethyl- its use as Staudinger–Bertozzi reagent breaking disulfide bonds can lead to an ac-
triphenylphosphonium chloride, at 300 nm (Scheme 3).[ Th e liberated phosphine
cumulation of both reducing agents as well
10]
in degassed isopropanol, produced triphe- was used to label an azide-tagged glyco- as misfolded proteins and the accumula-
[
22]
nylphosphine in 25% yield.
protein present on the cell surface, either in tion of unfolded proteins in the ER lumen
vitro, in fixed mammalian cells, or in vivo has been known to induce ER reductive
To the best of our knowledge, the
[5,27]
first application of a photoactivatable tri- employing zebrafish embryos.
arylphosphine was reported in 2016 by Deo All these reports focused on the photo-
et al. when they released triphenylphos- activation of triarylphosphines. Th e only
stress.
Th e aim of our work was to develop
a photoactivatable compound that upon
phine from a ruthenium-arene complex literature precedence for the release of photoactivation would release a trialkyl-
phosphine, for inducing intracellular re-
ductive stress, and a fluorescent reporter,
+
+
for ease of monitoring, as the only photo-
[
28]
products.
O
NTs
Th e probe design was based on the ther-
mally unstable Cy5– TC EP adduct and, us-
ing density functional theory (DF T) calcu-
lations, various substituents on the cyanine
backbone were screened to increase the
thermal barrier of the C–P bond cleavage.
From this screening, we identified that a
N,N-dimethylcoumarin–indolenine hybrid
had the best predicted stability and was
chosen as model compound for synthesis
+
Ph
irr 406 nm
S, S = solvent
PPh3
catalyst
+
Ts
Ts
Ru
N
S Ru
N
N
TsHN
Ph
O
Ph3P
N
N
N
Z
E
Scheme 1. Photorelease of triphenylphosphine from a ruthenium complex and its subsequent use
as a catalyst for an aza-Morita–Baylis–Hillman reaction.
(Fig. 1A). We tuned the thermal stability
of the probes further by varying the sub-
stituents on the indolenine core (Fig. 1B).
Probes 1a–d were synthesized by
Knoevenagel condensation of the N,N-
diethylamino coumarin aldehyde 2 with
Fischer’s bases 3a–d to result in the cou-
marin–carbocyanine dyes 4a–d (Scheme
O
O
UV light
O
R-N3
Ph
P
Ph
R
+
P
Ph
S
N
H
Ph
Staudinger–Bertozzi
ligation
SH
PPh2
S
O
5
). Th ese dyes were subsequently treated
with tributylphosphine to afford probes
a–d.
Absorbance and fluorescence spec-
Scheme 2. Photorelease of phosphinothioester and its use in a Staudinger–Bertozzi ligation.
1
O
OMe
HO
tra of dyes 4a–d and probes 1a–d were
measured in phosphate buffered-saline
(PBS). Dyes 4a–d displayed an absor-
bance maxima (λmax) at 570 nm, whereas
OH
HO
OH
OH
R1
PPh
X
2
OH
OMe
OMe
X =
O
COO-
O
COO-
O
H
N
H
N
N
H
O
O2N
N3
O
UV light
HO
R
PPh2
O
HO
probes 1a–d exhibited λ at 415 nm be-
max
cause addition of the phosphine interrupts
the conjugation of the molecules. Cuvette
experiments were performed to determine
the thermal stability of the probes, in the
Scheme 3. Photorelease of a triarylphosphine and its use as a Staudinger–Bertozzi reagent for
labeling azido-tagged glycoproteins present on the cell surface.

Laureates: Junior Prizes of the sCs faLL Meeting 2017
CHIMIA 2018, 72, No. 4 243
compound is membrane permeable and the
photoactivation is still functional in a cel-
lular environment.
fluorescent
non-fluorescent
-O3S
SO3-
-O S
3
SO3
-
TCEP
UV
+PR3
Next, we tested the ability of the pho-
toreleased phosphine to increase the con-
centration of free thiols inside the cell. For
this experiment, cells were pretreated with
monobromobimane (mBB), a thiol sensi-
tive fluorescent probe.[ Following incu-
bation of 1c and irradiation, we detected
a significant increase in the mBB signal,
which indicated the generation of intra-
N+
N
N
N
R=(CH2)2COO-
Cy5
Cy5-TCEP adduct
29]
Scheme 4. Conjugate addition of TCEP to Cy5 forms a non-fluorescent Cy5–TCEP adduct which
can be reversed with UV light. TCEP = tris(2-carboxyethyl)phosphine.
dark, in PBS that was supplemented with of 1c in presence of GSSG was monitored cellular reducing species which could be
GSSG. Th e role of GSSG was to trap employing high-performance liquid chro- either the generated thiols or the released
[30]
the released phosphine, thus avoiding re- matography (HPLC, Fig. 2). As indicated trialkylphosphine (Fig. 3D–E).
attack on the dye. Monitoring the increase by the chromatogram, the photoreaction
As mentioned before, common reduc-
in absorbance at 570 nm, we determined produced GSH from the reduction of tive stress indicators are the accumula-
the half-lives of the probes in absence of GSSG by the released trialkylphosphine.
light. Th e results of this experiment ( Ta ble Because compound 1c had the best following disulfide bond cleavage. We
, column 2) confirmed our hypothesis that thermal stability and good photoactiva- employed Th ioflavin T ( Th T) , a common
an electron-donating group, like methoxyl, tion efficiency, it was chosen for live-cell fluorescent stain for amyloid protein ag-
tion of unfolded proteins and aggregation
[
27]
1
[
31]
would decrease the thermal stability of the experiments. Human cervical cancer cells gregates, to test whether protein aggre-
probes, whereas an electron-withdrawing (HeLa) were incubated with probe 1c for gates are formed following irradiation of
group, like trifluoromethyl, would increase 10 minutes in imaging medium, washed, 1c. HeLa cells that were irradiated with
their stability.
and the fluorescence was recorded in the 1c and then treated with Th T
signal than
ried out in a fluorimeter ( Ta ble 1, column BP) before and after irradiation with a 405 cells that were stained with Th T
and ir-
). Solutions of probes 1a–d in PBS with nm laser (Fig. 3A–C). Th e increase in fluo-
radiated, but had not been treated with 1c
exhibited
Photoactivation experiments were car- red channel (λ = 561 nm; λ = 605/52 a significantly higher Th T
ex
em
3
GSSG were irradiated with 415 nm light rescence upon irradiation proved that the (Fig. 3F–I).
and the increase in fluorescence was moni-
tored at 630 nm, where the dyes 4a–d have
1
31
Fig. 1. New probe
design based on
their emission maxima. H and P NMR
analyses confirmed the identity of the
photoproducts as the corresponding dye
and tributylphosphine. Whereas probes
-
SO3-
A
O S
3
computational model-
ing. A) Evolution of
structures from the
Cy5 core structure
to the stable N,N-
dimethylcoumarin–in-
dolenine hybrid. B)
Further tuning of the
stability of the phos-
phonium adduct by
variation of substitu-
ents on the indolenine
core.
Me P+
3
N
Et
N
Et
1
a–c underwent photoactivation, no reac-
tion was observed for 1d. A combination
of theoretical calculations and additional
photophysical experiments indicated that
the phosphine was released following a
Enamine donor
Indolenine
Me N
Me N
O
O
2
2
photoinduced electron transfer (Pe T) from
the indolenine core to the coumarin moi-
ety. Th e trifluoromethyl group is too elec-
tron withdrawing, which lowers the energy
of the HOMO level of the indolenine, dis-
N
N
+
+
Me P
Me P
3
3
N,N-dimethylanilino
N,N-dimethylamino
Indolenine
Indolenine
donor
coumarin donor
favoring Pe T, and thus probe 1d was not
photoactivatable.
With these probes in hand, we also
B
Et N
O
O
2
1
1
1
a R = OMe
proved that photoactivatable phosphines
are capable of reducing GSSG to increase
the concentration of GSH. Photoirradiation
b R = H
N
c R = Cl
Bu P+
3
1d R = CF
3
R
Table 1. Half-lives of phosphine release. (dark
Scheme 5. Synthesis
of probes 1a–d.
=
thermal hydrolysis in absence of light; PA =
3
a-d
Et2N
O
O
hydrolysis during photoactivation with 405 nm
light; n.d. = not determined).
Et2N
O
O
EtOH, 75 °C, 5 h
N+
O
3
0–85%
O
2
4a-d
probe
t1/2 (dark) (min)
9.2(2)
t1/2 (PA) (min)
4(1)
R
1
a
Et2N
O
PBu3, CH2Cl2
20 °C, 5 min
N+ I–
1
b
27.3(3)
85(2)
11(2)
N
R: a = OMe
b = H
4a-d
5
0–90%
Bu3P+
c = Cl
1
c
12(2)
R
d = CF3
1
a-d
R
3a-d
1
d
>120
n.d

244 CHIMIA 2018, 72, No. 4
Laureates: Junior Prizes of the sCs faLL Meeting 2017
Fig. 2. In vitro reduc-
tion of GSSG upon
photoactivation of 1c.
[
1] M. Narasimhan, N. S. Rajasekaran, Biochim.
Biophys. Acta, Mol. Basis Dis. 2015, 1852, 53.
Irradiation of 1c in
3
2
1
00
00
00
0
presence of GSSG
[2] H. Zhang, P. Limphong, J. Pieper, Q. Liu, C. K.
Rodesch, E. Christians, I. J. Benjamin, FASEB
J. 2012, 26, 1442.
GSSG
GSH
[
3] N. Kaludercic, S. Deshwal, F. Di Lisa, Front.
Physiol. 2014, 5, 285.
[4] M. Wang, R. J. Kaufman, Nat. Rev. Cancer
014, 14, 581.
5] M. Schröder, R. J. Kaufman, Mutat. Res. 2005,
69, 29.
2
[
[
[
[
5
5
6
7
8
6] M. Schieber, N. S. Chandel, Curr. Biol. 2014,
R (min)
t
24, R453.
7] A. Tr ivedi, P. S. Mavi, D. Bhatt, A. Kumar, Nat.
Commun. 2016, 7, 11392.
8] P. Klán, T. Šolomek, C. G. Bochet, A. Blanc,
R. Givens, M. Rubina, V. Popik, A. Kostikov, J.
Wirz, Chem. Rev. 2013, 113, 119.
9] C. Brieke, F. Rohrbach, A. Gottschalk, G.
Mayer, A. Heckel, Angew. Chem. Int. Ed. 2012,
[
5
1, 8446.
[
[
[
[
[
[
10] L. Shah, S. T. Laughlin, I. S. Carrico, J. Am.
Chem. Soc. 2016, 138, 5186.
11] J. Engels, E.-J. Schlaeger, J. Med. Chem. 1977,
2
0, 907.
12] J. H. Kaplan, B. Forbush III, J. F. Hoffman,
Biochemistry 1978, 17, 1929.
13] C. Gwizdala, S. C. Burdette, Curr. Op. Chem.
Biol. 2013, 17, 137.
14] R. H. Kramer, D. L. Fortin, D. Tr auner, Curr.
Op. Neurobiol. 2009, 19, 544.
15] D. Höglinger, A. Nadler, C. Schultz, Biochim.
Biophys. Acta Mol. Cell Biol. Lipids 2014,
1
841, 1085.
16] Q. Liu, A. Deiters, Acc. Chem. Res. 2014, 47,
5.
17] D. S. Lawrence, Curr. Op. Chem. Biol. 2005, 9,
70.
18] A. S. Baker, A. Deiters, ACS Chem. Biol. 2014,
, 1398.
[
[
[
[
[
[
[
4
5
9
19] R. S. Givens, M. Rubina, J. Wirz, Photochem.
Photobiol. Sci. 2012, 11, 472.
20] R. R. da Silva, V. G. To scano, R. G. Weiss, J.
Chem. Soc., Chem. Commun. 1973, 567.
21] R. R. da Silva, F. H. Qulna, V. G. To scano, R. G.
Weiss, J. Phys. Chem 1979, 83, 1213.
22] T. M. S. Peranovich, M. E. R. Marcondes, V.
G. To scano, Phosphorus, Sulfur Silicon Relat.
Elem. 1990, 51, 314.
23] C. Deo, N. Bogliotti, P. Retailleau, J. Xie,
[
[
Organometallics 2016, 35, 2694.
24] P. Hu, T. Feng, C.-C. Yeung, C.-K. Koo, K.-C.
Lau, M. H. W. Lam, Chem. Eur. J. 2016, 22,
1
1537.
[
25] J. C. Vaughan, G. T. Dempsey, E. Sun, X.
Fig. 3. Photoactivation of compound 1c in live HeLa cells. A) Cells after 10 min incubation with
c (10 µM) before photoactivation. B) Same cells as in (A) after photoactivation (405 nm, 25 s). C)
Zhuang, J. Am. Chem. Soc. 2013, 135, 1197.
1
[26] J. A. Burns, J. C. Butler, J. Moran, G. M.
Whitesides, J. Org. Chem. 1991, 56, 2648.
[27] E. S. Christians, I. J. Benjamin, Am. J. Physiol.
Heart Circ. Physiol. 2012, 302, H24.
Quantification of the fluorescence intensity of 1c in panels (A) and (B). D) Cells treated only with
mBB (50 µM) and photoactivated. E) Generation of intracellular reducing agents in cells treated
with mBB (50 µM) and 1c (10 µM) after photoactivation. F) Quantification of the fluorescence
intensity of mBB in panels (D) and (E). G) Cells treated with ThT (5 µM) and photoactivated. H)
Generation of intracellular protein aggregates in cells treated with 1c (10 µM) for 10 min, then
photoactivated, washed, and stained with ThT (5 µM). I) Quantification of fluorescence intensity of
ThT in panels (G) and (H). mBB=monobromobimane, ThT=Thioflavin; Scale bar = 10 µm.
[
28] A. Ti rla, P. Rivera-Fuentes, Angew. Chem. Int.
Ed. 2016, 55, 14709.
[
29] A. Cossarizza, R. Ferraresi, L. Tr oiano, E. Roat,
L. Gibellini, L. Bertoncelli, M. Nasi, M. Pinti,
Nat. Protoc. 2009, 4, 1790.
30] D. E. Graham, K. C. Harich, R. H. White, Anal.
Biochem. 2003, 318, 325–328.
[
In conclusion, a series of photoactivat- Acknowledgements
[31] D. R. Beriault, G. H. Werstuck, Biochim.
Biophys. Acta Mol. Cell Res. 2013, 1833, 2293.
able phosphonium probes were synthe-
sized and their properties characterized.
Th is work was supported by E TH Zurich
and the Swiss National Science Foundation
(grant 200021_165551). Confocal microscopy
Th e probes are membrane permeable and
was carried out at the Scientific Center for
Optical and Electron Microscopy (ScopeM)
can be activated with 405 nm light. Live
cell photoactivation experiments resulted
at E TH Zurich and computational work was
in an increase in intracellular reducing
performed on the Euler cluster of the Swiss
species and protein aggregates, which are
telltale signs of reductive stress. Current
work in our lab is focusing on targeting
the probes to specific redox sensitive or-
ganelles, such as mitochondria and ER.
National Supercomputing Center.
Received: January 24, 2018