Improved PeT molecules for optically sensing voltage in neurons

Article abstract of DOI:10.1021/ja510602z

VoltageFluor (VF) dyes have the potential to measure voltage optically in excitable membranes with a combination of high spatial and temporal resolution essential to better characterize the voltage dynamics of large groups of excitable cells. VF dyes sens

Full text of DOI:10.1021/ja510602z

Subscriber access provided by SELCUK UNIV
Article
Improved PeT molecules for optically sensing voltage in neurons
Clifford R Woodford, E. Paxon Frady, Richard S Smith, Benjamin Morey, Gabriele Canzi, Sakina F
Palida, Ricardo C. Araneda, William B. Kristan, Clifford P. Kubiak, Evan W. Miller, and Roger Y. Tsien
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/ja510602z • Publication Date (Web): 13 Jan 2015
Downloaded from http://pubs.acs.org on January 18, 2015
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
Improved PeT molecules for optically sensing voltage in neurons
Clifford R. Woodford,1,‡ E. Paxon Frady,3 Richard S. Smith,4 Benjamin Morey,2 Gabriele Canzi,1 Sakina F. Palida,6 Ricardo
C. Araneda,4 William B. Kristan, Jr.,5 Clifford P. Kubiak,1 Evan W. Miller,2,†,‡,* and Roger Y. Tsien.1,2,7,*
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
1Departments of Chemistry and Biochemistry, 2Pharmacology, 3Neurosciences Graduate Group, 5Division of Biological Sci-
ences, 6Biomedical Sciences, and 7Howard Hughes Medical Institute, University of California, San Diego, La Jolla, Califor-
nia, 92093, United States and 4Department of Biology, University of Maryland, College Park, Maryland, 20742, United
States
ABSTRACT: VoltageFluor (VF) dyes have the potential to optically measure voltage in excitable membranes with the combination
of high spatial and temporal resolution essential to better characterize the voltage dynamics of large groups of excitable cells. VF
dyes sense voltage with high speed and sensitivity using photoinduced electron transfer (PeT) through a conjugated molecular wire.
We show that tuning the driving force for PeT (ꢀGPeT + w) through systematic chemical substitution modulates voltage sensitivity,
estimate (ꢀGPeT + w) values from experimentally measured redox potentials, and validate the voltage sensitivities in patch-clamped
HEK cells for 10 new VF dyes. VF2.1(OMe).H, with a 48% ꢀF/F per 100 mV, shows approximately 2-fold improvement over
previous dyes in HEK cells, dissociated rat cortical neurons, and medicinal leech ganglia. Additionally, VF2.1(OMe).H faithfully
reports pharmacological effects and circuit activity in mouse olfactory bulb slices, thus opening a wide range of previously inacces-
sible applications for voltage sensitive dyes.
The ability to interrogate the dynamics of cellular events with
precise spatial and temporal resolution is limited, to a large de-
gree, by the development of fluorescent indicators with suffi-
cient sensitivity and selectivity for events of interest. Changes
in electric potential across biological membranes are cellular
events of key interest in a number of biological systems such as
the heart, muscles, and brain. Understanding the spatial and
temporal dynamics of electrical changes in neural activity that
underlie basic human behaviors, emotions, perceptions, and
thought remains a key challenge to neurobiology. Monitoring
transmembrane potential of excitable cells forms the basis for
modern neurobiology, and yet methods to do so have been lim-
ited primarily to electrophysiological recordings with elec-
trodes.1 This places sharp constraints on the spatial information
gleaned from this approach. In general, the size of electrodes
and pipettes used for traditional electrophysiology restricts their
usage to a single neuronal soma, making recording from either
multiple cells simultaneously or from sub-cellular regions like
dendritic spines and axonal terminals extremely difficult or im-
possible.1 In a complementary fashion, imaging of Ca2+ ions has
taken center stage for measuring neuronal activity, where the
transient increase in [Ca2+]i acts as an indirect readout of neu-
ronal activity. However, while Ca2+ indicators are highly sensi-
tive and can in some cases be genetically encoded, they provide
very little information regarding hyperpolarization, sub-thresh-
old events, or the nature of the electrical changes that generated
the Ca2+ increase.
indicators include combinations of low sensitivity, slow re-
sponse kinetics, high capacitive load, low brightness, and poor
membrane localization.
In an effort to help meet the need for indicators that faithfully
report on voltage changes with high spatial and temporal reso-
lution, we recently disclosed the initial design and characteriza-
tion of VoltageFluor 2.1.Cl (VF2.1.Cl) for imaging voltage
changes in neurons with high spatial and temporal fidelity.23
VF2.1.Cl uses photo-induced electron transfer24 (PeT) through
a molecular wire as a platform to achieve fast, wavelength-in-
dependent voltage imaging in neurons. VoltageFluor dyes lo-
calize to the plasma membrane where the free energy for PeT is
affected by the local electric field. At hyperpolarized potentials
PeT is more favorable, and at depolarized potentials PeT is less
favorable. Considering PeT and fluorescence are competing
processes, the inverse is true for fluorescence, which can be
monitored via traditional fluorescence imaging (Fig. 1a). In this
paper, we show that the VoltageFluor platform offers a general
chemical strategy for voltage imaging and that voltage sensitiv-
ity can be rationally increased through modulation of donor and
acceptor electron affinities. We present the design and synthesis
of a series of 10 new structurally-related VoltageFluors, esti-
mate the driving force for PeT (ꢀGPeT), and establish their util-
ity for imaging transmembrane potential in cultured cells, dis-
sociated mammalian neurons, and ex vivo leech ganglia. Fi-
nally, we demonstrate that VF2.1(OMe).H can report on both
fast and slow voltage changes in acutely prepared rodent olfac-
tory bulb slices.
Imaging voltage dynamics offers an attractive solution to this
problem, and several types of voltage sensitive indicators have
been described. These include small molecule fluorescent ap-
proaches like merocyanines2, oxonols, and rhodamines,3
charge-shift electrochromic dyes,4-6 lipophilic anions,7-10 sec-
ond-harmonic generation11,12 and nanoparticles.13,14 Genetically
encoded voltage indicators are also known and make use of flu-
orescent protein fusions to endogenous voltage-sensing do-
mains15-20 or microbial opsins21,22 to transduce voltage changes
into photons. Limitations of these and other voltage-sensitive
Results
Design, Synthesis, and Characterization of VoltageFluors
Our strategy for voltage sensing relies on proper orientation
of a fluorophore-wire-donor construct into the plasma mem-
brane (Fig 1a). Sulfofluoresceins were an initial choice because
the persistently ionized sulfonic acid (pKa < -2) helps prevent
internalization of the sensor through the cellular membrane.
Sulfofluoresceins also have demonstrated utility in two-photon
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 9
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 1. Voltage sensing mechanism and synthesis of VoltageFluor dyes. a) Hyperpolarized (left) membrane potentials (negative inside
cell) promote PeT and quench fluorescence. Depolarization (positive inside cell) decreases PeT and increases fluorescence (right). Thus, the
quantum yield of VF dyes are related to the local membrane potential. b) General synthesis of VF dyes.
fluorescence imaging, leaving open possibilities for in vivo
applications.25 The low attenuation values and ease of chemical
synthesis26 of phenylenevinylene (PPV) molecular wires made
them an ideal choice for spacers between the donor and accep-
tor. Our previous study showed that 2 generations of PPV
spacer provided excellent voltage sensitivity while maintaining
sufficient loading and water solubility. Finally, nitrogenous do-
nors are frequently used in PeT sensors27-30 and in this case offer
the opportunity to tune the relative energetics of PeT by modu-
lation of the electron-richness of the aniline. Previous studies23
show that, while N,N-dibutylaniline-derived VF2.4.Cl had sim-
ilar voltage sensitivities as the N,N-dimethyl substituted
VF2.1.Cl, the dimethyl analog had better signal-to-noise, on ac-
count of increased uptake into cellular membranes—therefore,
dimethyl analogs were used throughout this study.
staining localized to the plasma membrane (Fig 2a and 3a, all
images depict cells not under voltage-clamp, where Vm is close
to -60 mV). Importantly, loading conducted at 37 °C did not
result in significant internalization of the dye molecule. Further-
more, throughout the course of evaluation in HEK cells (45 to
60 minutes), we observe negligible dye internalization or loss
of voltage sensitivity, as measured by the change in voltage sen-
sitivity recorded from cells patched at the beginning of a trial
(approximately 5-15 min post-loading) and at the end of a trial
(approximately 30-60 min post-loading).
The deviation in sensitivity is low (<5%), and this is reflected
in the small error bars in Figure 3b. This suggests that, under
these conditions, the initial orientation of the dye within the
membrane is established early during loading and very little
We sought to explore the voltage sensitivity of VF2.1.Cl
through substituent changes on both the fluorophore/acceptor
and donor. The modular nature of the VF dye synthesis (Fig 1b)
enabled rapid construction of several new derivatives listed in
Table 1. Scheme 1 outlines the synthesis of the VF family of
dyes (full details available in Supporting Information). All of
the VF sensors were characterized spectroscopically in aqueous
media (pH 9, 0.1% Triton X-100 to ensure complete solubility
and deprotonated form of the xanthene phenol) and had absorp-
tion maxima ranging from 504 nm (VF2.(1OMe).H) to 522 nm
(VF2.1Cl compounds) and emission profiles from between 522
nm to 538 nm (Table 1). The substituent pattern on the xan-
thene chromophore determined the absorbance band, with un-
substituted sulfofluorescein displaying ꢁmax values centered
around 506 nm, fluoro-substituted sulfofluorescein centered at
508 nm, methyl at 514 nm, and chloro at 522 nm. Additionally,
each VF compound exhibited a secondary absorbance band
characteristic of a conjugated phenylenevinylene molecular
wire, with ꢁmax values ranging from 392 to 410 nm (Fig 2b and
Supporting Information Figure 1).VF2.0.Cl had a secondary
ꢁmax substantially blue-shifted from the other VF dyes (363 nm),
owing to the lack of conjugation through the dimethylaniline.
The quantum yields for the dyes ranged from 0.04 for
VF2.1(OMe).Me to 0.50 for VF2.0.Cl (Table 1).
Figure 2. Characterization of VoltageFluor 2.1(OMe).H in HEK
cells. a) HEK cells stained with 200 nM VF2.1(OMe).H for 15
minutes at 37 °C. Scale bar is 20 ꢂm. b) Absorbance and emission
profile of VF2.1(OMe).H at pH 9, 0.01% Triton X-100. c) Fluores-
cence response of VF2.1(OMe).H in voltage-clamped cells from a),
plotted against time, during 50 ms steps from -60 mV to +100 mV
followed by steps decreasing in potential by 20 mV increments to
-100 mV. d) Linearity of VF2.1(OMe).H response (ꢀF/F) vs. final
membrane potential in the physiologically relevant range (±100
mV). Each data point represents three separate measurements. Er-
ror bars are ±SEM for ≥ 3 separate experiments.
Fluorescence Response to Potential Changes
The voltage response of the dyes was established in whole-
cell patched voltage-clamped HEK cells. Bath application of
200 nM VF dyes for 15 minutes at 37 °C resulted in cellular
ACS Paragon Plus Environment

Page 3 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 3. Staining and voltage sensitivity of VoltageFluor dyes in HEK cells. a) Epifluorescence images of HEK cells incubated in HBSS
buffer containing 200 nM of the indicated VoltageFluor dyes for 15 min at 37°C. All acquisition and analysis parameters are identical, to
enable an estimation of the relative brightness of the dyes in a cellular context. Scale bar is 20 ꢂm. Cells are not under voltage-clamp
conditions. b) Fluorescence response of representative VoltageFluor dyes vs. membrane potential. Voltage-clamped HEK cells were held at
-60 mV and then stepped to the indicated potential. The relative change in fluorescence (ꢀF/F) is plotted against the final membrane potential
for VoltageFluor dyes loaded in HEK cells at a concentration of 200 nM. Error bars are ±SEM for n ≥ 3 for each dye.
inversion of the VoltageFluor occurs. We corroborated this
finding by examining the cellular staining of VF dyes as a func-
tion of dye incubation time and found little change in the local-
ization after loading VF2.1(OMe).H for either 10 minutes or 85
minutes (Supporting Information Figure 2). Uniform orien-
tation of the dyes is critical for obtaining voltage sensitivity, be-
cause an equal distribution of “chromophore in” and “chromo-
phore out” orientations would result in net zero voltage sensi-
tivity (Fig 1a). Whole-cell, voltage-clamped HEK cells were
held at -60 mV and then stepped to hyper- and depolarizing po-
tentials 20 mV steps (range ±100 mV, Fig 2c and d, SI Movie
1). After background subtraction, voltage sensitivities ranging
from 4% to 49% ꢀF/F per 100 mV were measured for the VF
dyes, with VF2.1.Me least and VF2.1(OMe).Cl most sensitive,
with VF2.1(OMe).H registering 48% ꢀF/F per 100 mV (Sup-
porting Information Figure 3 indicates an example of region
of interest, ROI, selection for determining voltage sensitivity).
Importantly, control compound VF2.0.Cl, which lacks an elec-
tron-rich aniline donor, shows no voltage sensitivity (Support-
ing Information Figure 4).
To further explore the linearity of the VF series we applied a
more extreme hyper- and de-polarizing stepping paradigm to
HEK cells loaded with VF dyes, assaying a range spanning
±300 mV. Still larger steps proved too much of a strain on the
cells and could not be reliably measured. At extremely hyperpo-
larizing potentials (large negative potential inside the cell), we
hypothesized PeT would be maximally activated and the VF
dye would be at its dimmest state. Conversely, at larger depo-
larizing potential (large positive potential inside), PeT quench-
ing would be entirely relieved, and the dye would be at its max-
imal brightness. Fluorescence responses were measured as be-
fore, and steps were provided in 30 mV increments to cover the
range from ±300 mV. The first generation dye VF2.1.Cl, as well
as VF2.1(diOMe).H, VF2.1(OMe).H and VF2.1.F, fit within a
range that encompassed, or nearly encompassed, the minimum
and maximum, meaning that the 600 mV range of potentials
spanned the energetics required to switch PeT completely on or
off (Fig 3b). VF dyes bearing more electron rich aniline donors
(OMe, diOMe) and/or electron-poor fluorophores, such as
VF2.1(diOMe).Cl, VF2.1(OMe).Cl, and VF2.1(OMe).F,
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 9
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 4. Application of VF2.1(OMe).H to sensing action potentials in cultured neurons. a) Epifluorescence image of cultured rat hippo-
campal/cortical neurons stained with 200 nM VF2.1(OMe).H for 15 minutes at 37 °C. Scale bar is 20 ꢂm. b) Dual optical and electrophysi-
ological recordings of evoked action potentials in cultured neurons. c) Expanded time scale of the first stimulation in b).
showed fluorescence responses that generally comprised the
lower bound of the sigmoidal curve at hyperpolarizing poten-
tials, indicating that PeT processes dominate in this regime.
Only at extremely depolarizing potentials (>100 mV) did sig-
nificant voltage sensitivity manifest. A final grouping of dyes,
including VF2.1.H, VF2.1(OMe).Me, and VF2.1.Me, displayed
fluorescence responses confined to the asymptote achieved at
extreme depolarizing potentials. In this regime, radiative path-
ways predominate, giving bright signals that can only be
quenched upon extreme hyperpolarization of the cell mem-
brane.
ing the energy required to separate two charges. Due to the mi-
nor structural differences between the VF compounds, w re-
mains relatively constant across the VF series, though its abso-
lute magnitude remains difficult to estimate with precision. The
(ꢀGPeT + w) values are summarized in Table 1.
VF dyes bearing electron-withdrawing groups (Cl, F) on the
fluorophore and electron-donating groups (OMe, diOMe) on
the donor had the most negative (ꢀGPeT + w) values, while VF
dyes with electron-donating (H, Me) fluorophores and unsub-
stituted dimethylanilines had the highest (ꢀGPeT + w) values.
The experimentally estimated (ꢀGPeT + w) values correlates
well with the fluorescence response of the VF dyes at extreme
potentials. For the three dyes with (ꢀGPeT + w) values lower
than -0.23 eV, VF2.1(diOMe).Cl, VF2.1(OMe).Cl, and
VF2.1(OMe).F, voltage sensitivity was moderate to high (20%
to 49% ꢀF/F per 100 mV) in the physiologically relevant win-
dow of ±100 mV (Table 1 and Fig 3b). However, the response
was not linear, and significant quenching is observed at hy-
perpolarizing/neutral potentials. Only upon achieving more ex-
treme depolarizing potentials do these dyes unquench to give
significant fluorescence responses to changing membrane po-
tentials (Fig 3b). Conversely, for the three VF dyes with (ꢀGPeT
+ w) greater than -0.08 eV, sufficient driving force is not avail-
able to quench fluorescence, and so radiative pathways domi-
nate VF2.1.H, VF2.1(OMe).Me, and VF2.1.Me (Fig 3b).
Estimation of (ꢀGPeT + w)
To gain a more quantitative understanding of the factors that
contribute to voltage sensitivity, we measured the redox poten-
tials of both the donors and acceptors within the VF framework.
The oxidation and reduction potentials of dimethyl anilines and
substituted sulfofluoresceins were measured via cyclic voltam-
metry. As expected, Cl-substituted sulfofluorescein was most
readily reduced (Eo (A/A-) -2.02 V vs ferrocene, Fc), followed
by F-substituted sulfofluorescein (-2.10 V vs Fc). Unsubstituted
sulfofluorescein was 130 mV more difficult to reduce (-2.24 V
vs Fc), and Me-substituted sulfofluorescein was the most diffi-
cult to reduce (-2.32 V vs Fc) (Supporting Information Figure
5). Measuring oxidation potentials of the donors proved more
challenging, as the resulting radical cations undergo further ox-
idation and reaction to form benzidine species,31 limiting our
ability to report oxidation potentials of the pure dimethylaniline
species. We instead measured the oxidation potential of syn-
thetic intermediates 4, which, on account of substitution para-
to the dimethylamino group, restricted formation of confound-
ing dimeric species. Oxidation of the phenylenevinylene dime-
thylanilines were still fairly reactive, giving irreversible volt-
ammograms, but provided more reliable initial oxidation meas-
urements (Supporting Information Figure 5). The unsubsti-
tuted donor measured at the initial oxidation peak was the least
readily oxidized (Eo (D/D+) +0.129 V vs Fc). A single methoxy
substitution ortho- to the dimethylamino group gave an oxida-
tion potential of 0.090 V vs Fc, while 2,5-dimethoxy aniline was
the most readily oxidized at 0.033 V vs Fc. With these values
in hand, we could estimate the driving force for PeT through the
use of the Rehm-Weller equation,32 ꢀGꢁꢂꢃ = ꢀEꢄ° ꢅ ꢆ ꢀEꢇ°ꢂꢈ ꢆ
ꢀEꢉ,ꢉ ꢆ ꢊ, where ꢀE°ox is the oxidation potential of the donor,
ꢀE°red is the reduction potential of the acceptor/fluorophore,
ꢀE0,0 is the energy required to excite the chromophore into the
first electronically excited state, and w is a work term represent-
This results in low voltage sensitivity (5-16% ꢀF/F per 100
mV) in the physiological range that becomes larger only upon
increased hyperpolarization. Finally, the four VF dyes with in-
termediate (ꢀGPeT + w) values, ranging from -0.13 to -0.22 eV
show the most linear responses to membrane potential changes
within the physiologically relevant range of ±100 mV (Fig 3b),
with moderate (27%) to high (48%) voltage sensitivity (Table
1). These dyes include VF2.1.Cl, VF2.1.F, VF2.1(diOMe).H,
and VF2.1(OMe).H. These data suggest a threshold of (ꢀGPeT
+
w) values less than -0.08 eV, but no lower than -0.27 eV, are
required to achieve voltage sensitivity greater than 20% ꢀF/F
per 100 mV. Two dyes that meet this criteria, VF2.1(OMe).Cl
and VF2.1(OMe).H show high voltage sensitivity, at 49 and
48% ꢀF/F per 100 mV, a >80% improvement over the initial
VF2.1.Cl compound. Although the absolute sensitivity of
VF2.1(OMe).Cl is marginally greater than VF2.1(OMe).H, the
more negative (ꢀGPeT + w), which results in more quenching at
physiological membrane potential, makes VF2.1(OMe).H more
suitable for investigations of membrane potential fluc tuations
at or near -60 mV (nominal resting membrane potential for
many neurons). Additionally, the brighter staining of HEK cells
ACS Paragon Plus Environment

Page 5 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 5. Applications of VF2.1(OMe).H a) Epifluorescence image of H. medicinalis mid-body ganglia stained with VF2.1(OMe).H. Scale
bar is 50 ꢂm. b) Optical (red) and electrophysiological (black) recording of spontaneous activity in the leech ganglia from regions of interest
(ROI) indicated in a). The calibration bar is 0.3% ∆F/F, 4 mV, and 1 s. c) Epifluorescence image of a sagittal mouse olfactory bulb (OB)
slice loaded with VF2.1(OMe).H (500 nM). The dotted square represents the ROI used for quantification. The white arrow points to the
plasma membrane of a mitral cell (MC) loaded with VF dye. The scale bar is 10 ꢂm. d) Optical recordings (0.5 Hz acquisition) showing the
fluorescence response of VF2.1(OMe).H in the cell from (c) to glutamate (blue trace, +Glu, 100 ꢂM, 1.5 min.) and nicotine (red trace, +Nic,
30 ꢂM, 1.5 min.). The calibration bar is 2% ꢀF/F and 2 min. e) Epifluorescence wide field view of another olfactory bulb slice loaded with
VF2.1(OMe).H. A patch electrode was used to electrically stimulate in the vicinity of the MC layer (MCL). The top calibration bar is 50 ꢂm.
Orange and red circles indicate regions of interest used for analysis in f). f) Optical recordings (200 Hz acquisition) in the external plexiform
layer (EPL, orange) or granule cell layer (GCL, red). The timing of the stimulus pulse is indicated by black dots (10 ms, 10 Hz). The stimulus
produces a voltage change in the GLC (upper red trace) but not in the external plexiform layer (EPL, middle orange trace). The responses in
the GCL are abolished in the presence of ionotropic glutamate receptors (iGluR) blockers (APV, 100ꢂM; CNQX, 10ꢂM; bottom red trace).
The calibration bar is 2% ꢀF/F and 100 ms).
with VF2.1(OMe).H over VF2.1(OMe).Cl, made it the opti-
mal probe for subsequent biological applications.
important model system for understanding circuit dynamics, be-
cause it not only contains several hundred functionally con-
nected neurons, but also provides ready access to both electro-
physiological and voltage-sensitive optical recording. Small
molecule indicators have a distinct advantage over protein-
based indicators because of the difficulty of transfecting adult
leech neurons.34 We prepared desheathed mid-body ganglia
from H. medicinalis stained with VF2.1(OMe).H (200 nM, 20
min.). Following treatment with VF2.1(OMe).H, cellular mem-
branes in the ganglia showed strong fluorescence characteristic
of VF dyes (Fig 5a). Using paired electrode recording (20-40
Mꢄ) in conjunction with fluorescence imaging, we monitored
fluorescence as a function of membrane potential. Spontane-
ously-firing action potentials from the Retzius cell were clearly
distinguished optically (Fig 5b, red trace), and the time course
matched that of the intracellular electrode (Fig 5b, black trace).
VF2.1(OMe).H detects not only supra-threshold spikes, but also
slower depolarization and re polarizations following action po-
tential. These repolarizations were not observed in the rat opti-
cal or electrophysiological traces due to a stimulus artifact that
bled slightly into the repolarization phase of the action poten-
tial. In both cases (rat and leech), the optical trace clearly fol
lows the electrophysiology, establishing the fidelity of the VF
Characterization of VF2.1(OMe).H in Neurons
First, we applied VF2.1(OMe).H to cultured cortical neurons
(Fig 4). Bath application of the dye extensively stained neuronal
cell membranes, as determined by epifluorescence microscopy
(Fig 4a). Confocal imaging of VF2.1(OMe).H-stained neurons
established localization to the cellular membrane with very little
visible internalization (Supporting Information Figure 6).
These experiments also show little internalization even with an
extended loading time of 2 hours (Supporting Information
Figure 6). VF2.1(OMe).H readily detected action potentials in
single trials (Fig 4b and c, SI Movie 2) with a signal to noise
ratio of 28:1, near two-fold improvement over VF2.1.Cl.23 The
inherently fast nature by which VF dyes sense changes in mem-
brane potential (ꢃon/off < 150 ꢂs)23 offers superior fidelity in re-
producing electrical signals compared to protein based sensors
with longer response times, for example ArcLight Q239 (ꢃon/off
= 10-50 ms).18
One of the promises of voltage imaging is the ability to spa-
tially reconstruct patterns of activity within a functional net-
work. For this purpose, the medicinal leech has proven to be an
ACS Paragon Plus Environment

Journal of the American Chemical Society
Table 1. Properties of VoltageFluor Dyes
Page 6 of 9
1
2
3
4
% ꢀF/F
ꢀG°PeT
w
(eV)
+
E°(D+/D) E°(A/A−)
ꢁabs
ꢁem
ꢀG00
(eV)
b
Compound
R1=
R2=
R3=
R4=
ꢅFl
per 100
mVc
(V)a
(V)a
(nm)b (nm)b
5
6
7
8
VF2.1(diOMe).Cl N(Me)2 OMe OMe Cl
0.033
0.090
0.090
0.129
0.129
0.033
0.090
0.129
0.090
0.129
1.080d
-2.02
-2.02
-2.11
-2.02
-2.11
-2.24
-2.24
-2.24
-2.32
-2.32
-2.02
521
522
509
522
508
504
504
507
515
513
521
535
536
528
536
524
522
524
528
536
532
538
2.38
2.38
2.44
2.38
2.44
2.46
2.46
2.45
2.41
2.42
2.38
-0.325
-0.263
-0.243
-0.224
-0.209
-0.186
-0.130
-0.076
0.003
20
49
44
27
30
30
48
16
13
5
0.26
0.13
0.05
0.05
0.10
0.24
0.04
0.11
0.04
0.38
0.50
VF2.1(OMe).Cl
VF2.1(OMe).F
VF2.1.Cl
N(Me)2 OMe
N(Me)2 OMe
H
H
H
H
Cl
F
9
N(Me)2
N(Me)2
H
H
Cl
F
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
VF2.1.F
VF2.1(diOMe).H N(Me)2 OMe OMe
H
VF2.1(OMe).H
VF2.1.H
N(Me)2 OMe
N(Me)2
N(Me)2 OMe
H
H
H
H
H
H
H
H
VF2.1(OMe).Me
VF2.1.Me
Me
Me
Cl
N(Me)2
H
H
H
0.033
VF2.0.Cl
0.722
0
a vs. Ferrocene (Fc). b 0.01% Triton X-100, 5 mM sodium phosphate, pH 9. c Measured in voltage-clamped HEK cells. d Oxidation potential of
stilbene taken from Ref 33.
dyes for tracking real membrane potential changes. The ac-
tion potentials in the leech optical recording appear truncated
due to undersampling in the fluorescence channel (acquisition
speed is 50 Hz). The fractional change in fluorescence is ap-
proximately 3x larger than with VF2.1.Cl,23 consistent with the
improvement (approximately 2x) observed in HEK cells.
µM, 2 min), also depolarized MCs (ꢀF/F, 3.02 ± 0.27%, n=6
p<0.01). Next, we conducted electrical stimulations (10 ms
pulses, 10 Hz) in the vicinity of the MC layer (MCL), to synap-
tically activate GCs while optically recording (200 Hz frame
rate) from this cell layer (GCL). Focal electrical stimulation re-
vealed a fast, robust increase in fluorescence in the GCL (Fig
5e,f, red circle and red upper trace, ꢀF/F 1.91 ± 0.23%, n=5,
p<0.01), owing to activation of GCs by glutamate released from
MCs. Importantly, this excitation did not elicit a response in the
external plexiform layer (EPL, Fig 5e,f, orange circle and
trace), demonstrating that VF2.1(OMe).H efficiently tracks
synaptic transmission from the MCL to the GCL, and that this
response does not result from a non-selective depolarization of
the stimulated region. Accordingly, the excitation was sensitive
to ionotropic glutamate receptor (iGluR) blockade by APV and
CNQX (Fig 5f, red lower trace, 100 and 10 µM; ꢀF/F, 0.07 ±
0.15%, n=5, p<0.01). When the same experiments were con-
ducted with VF2.1.Cl in order to compare its performance vs.
VF2.1(OMe).H, the optical response to identical stimulation
was substantially larger for VF2.1(OMe).H, consistent with re-
sults in HEK cells, rat neurons, and leech ganglia. The response
from VF2.1(OMe).H was 2x larger than VF2.1.Cl for electrical
stimulation and 3x to 4x larger under stimulation with gluta-
mate or nicotine, respectively (p < 0.0015 in all cases, T-test, n
= 6; Supporting Information Figure 7). These experiments
and direct comparisons demonstrate the utility of the VF dyes—
and especially VF2.1(OMe).H—for conducting both fast and
slow in vitro slice network physiology.
Next, we assessed the ability of the dye to detect voltage
changes in brain slices—a more challenging endeavor consid-
ering the complexity, scale, and diversity of the preparation. Ef-
forts to visualize voltage changes in slice are hampered by small
fractional changes in fluorescence. State of the art indicators all
have significant shortcomings in this context. For example, the
low brightness of QuasAr requires intense laser power to
achieve sensing, limiting the observable field to ~ 200 ꢂm2 be-
fore tissue heating becomes a concern.22 Small molecule volt-
age-sensitive dyes have inherent limitations as well, particu-
larly with indiscriminate staining of all plasma membranes. In-
creasing the voltage sensitivity of indicators can partially ame-
liorate this obstacle, and so we were hopeful that the improved
sensitivity of VF2.1(OMe).H would allow us to measure mem-
brane potential changes optically in a brain slice.
We turned to the olfactory bulb (OB) of mice because it con-
tains a well-characterized synaptic connectivity which has
largely been explored through traditional electrophysiological
means35 or by optical imaging through absorbance techniques.36
Within the OB, sensory input excites principal neurons—the
mitral and tufted cells (MCs). Activation of MCs excites, via
ionotropic glutamate receptors, surrounding local inhibitory
neurons, or granule cells (GCs).37 Staining of OB slices with
VF2.1(OMe).H (500 nM, 20 min incubation), resulted in strong
fluorescence signals from the cell membranes of OB neurons,
in particular the large MCs (Fig 5c, white arrow). Activation
of MCs by bath application of glutamate gives a large increase
in fluorescence intensity, lasting several minutes (optical re-
cording rate = 0.5 Hz) and consistent with activation of the MCs
(Fig 5d, “G”, 100 ꢂM, 2 min; 4.13 ± 0.40%, n=6 p<0.01). Ad-
ditionally, in agreement with previous findings,35 application of
the acetylcholine receptor agonist, nicotine (Fig 5d, “N”, 30
Discussion
Voltage imaging with small molecules has been limited to
staining of single cells in a brain slice38 or to measurement of
optical “field potentials” that report population changes and re-
quire spike-timed averaging.39 One problem has been the lack
of fast, sensitive, non-disruptive probes to report on membrane
potential changes. To address this concern, we here now re-
ported the design, synthesis, and application of a new family of
voltage-sensitive dyes—VoltageFluors (VFs)—that make use
of photoinduced electron transfer from an electron-rich aniline
ACS Paragon Plus Environment

Page 7 of 9
Journal of the American Chemical Society
through a conjugated phenylenevinylene molecular wire to a
xanthene chromophore for fast, sensitive voltage sensing in a
Table 2. VF2.1(OMe).H Performance in Biological Samples
1
2
3
4
5
6
7
8
Sample
ꢀF/F
SNR
variety of neuronal contexts. The VF family of dyes displays
visible excitation and emission profiles that enable optical volt-
age recording at peak excitation and emission wavelengths,
large linear responses to changes in membrane potential, negli-
gible capacitance loading of membranes, and fast, turn-on opti-
cal responses capable of resolving action potentials and sub-
threshold membrane potential dynamics in neurons.
HEK cells
Rat Neurons
48%a
10%a
2-4%a,b
2%c
38:1
28:1
17:1
17:1
Leech Neurons
Olfactory Bulb Slice
a per 100 mV change, calibrated by simultaneous electro-
As a chemical platform, the VF dyes offer a tunable approach
to voltage sensing, in which alteration of the electron affinities
results in modulation of the voltage sensitivity of the dye. This
study demonstrates, for the first time, that modulation of ꢀGPeT
alters the voltage-sensitivity of VF dyes and establishes une-
quivocally the requirement of an electron-rich donor for voltage
sensing, as indicated by the lack of voltage response from
VF2.0.Cl. Analysis of the relative driving force for electron
transfer derived from experimentally determined oxidation/re-
duction potentials suggests that the range of -0.08 V to -0.27 V
yields sensors with high voltage sensitivity and linearity within
a physiologically relevant window of ±100 mV. The most sen-
sitive of the VF dyes—VF2.1(OMe).Cl and VF2.1(OMe).H—
have voltage sensitivities of 49 and 48% ꢀF/F per 100 mV, re-
spectively, and compare favorably to other “fast” voltage dyes
that function via an electrochromic sensing mechanism and
have typical voltage sensitivities ranging from 10 to 28% ꢀF/F
per 100 mV.38,40 Although high voltage sensitivities can be
achieved with electrochromic dyes by excitation at the far red
edge of the excitation spectrum and collection with a similarly
narrow emission filter, this requires off-peak excitation, the use
of potentially phototoxic high-intensity illumination, and sam-
pling of only a small fraction of emitted, voltage-sensitive pho-
tons.41 For voltage imaging, which is inherently photon-limited
due to the fast sampling nature (0.5-2 kHz range) of the exper-
iment, sacrificing a majority of excitation and emission photons
can be problematic, making approaches like VF dyes that use
all excitation and emission photons for sensing ideally suited
for voltage imaging. Due to its improved linear response to
membrane potential change, we used VF2.1(OMe).H to moni-
tor action potentials in cultured neurons with enhanced SNR
over previous VF dyes (28:1 vs 16:1). We also demonstrated
the enhanced utility of VF2.1(OMe).H in leech ganglia for
monitoring spontaneous activity with a 3-fold improvement of
SNR compared to VF2.1.Cl. Finally, applications in brain slices
show that VF2.1(OMe).H can report on both slow and fast net-
work physiology in a complex neuronal environment. The per-
formance of VF2.1(OMe).H in these preparations are summa-
rized in Table 2.
9
b
physiological measurement. No background subtraction.
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
c
Images sampled at 50 Hz. Uncalibrated; no background
subtraction.
here shows a greater than 100% ꢀF/F per 100 mV at ex-
tremely depolarized potentials, indicating greater sensitivities
can be achieved. Thirdly, longer wavelengths would be benefi-
cial for thick samples and multiplex imaging with currently
available probes. The generality of the VoltageFluor PeT plat-
form predicts the chemical tractability of these efforts, which
are ongoing in our labs.
ASSOCIATED CONTENT
Supporting Information.
Experimental details including synthesis and characterization of all
VoltageFluor dyes and precursors, cyclic voltammetry, electro-
physiology and cell imaging data. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Correspondence should be addressed to EWM (evanwmil-
ler@berkeley.edu) or RYT (rtsien@ucsd.edu).
Present Addresses
† Departments of Chemistry and Molecular & Cell Biology, UC
Berkeley, Berkeley, CA, 94720, United States.
Author Contributions
‡These authors contributed equally.
ACKNOWLEDGMENT
This work was supported by the Howard Hughes Medical Institute,
the NIH, R37NS027177 to R.Y.T., a Pathways to Independence
Award to E.W.M. (K99/R00NS078561), NIH MH43396 and NSF
IOB-0523959 to W.B.K., NIH Training Grant EB009380 and
MH020002 to E.P.F., NSF CHE-1145893 and AFOSR FA 9550-
13-1-0020 to C.P.K. and G.C., and NIH DC009817 to R.C.A.
The optimal voltage probe remains elusive. The best voltage-
sensitive proteins, QuasAr1,22 ASAP1,20 and ArcLight Q23918
have issues with brightness, sensitivity, and speed respectively.
In addition the turn-off response of ASAP1 and ArcLight Q239
to depolarization decreases sensitivity. These issues have been
addressed by VF dyes. Current VF dyes represent a significan-
tadvance in sensitivity over previous VF dye incarnations, alt-
hough much work remains. First, because VF dyes have no ge-
netically encoded component yet, promiscuous staining of non-
excitable cells in heterogeneous samples severely decreases the
apparent voltage sensitivity by raising the overall background
fluorescence. This trend is apparent in the decreased voltage
sensitivity and SNR upon going from HEK cells to brains slices
(Table 2). Genetic targeting of VF dyes in a two-component
system would enable analysis of genetically defined cells and
improve signal-to-noise. Secondly, VF2.1(OMe).H, described
REFERENCES
(1)
(2)
Scanziani, M.; Hausser, M. Nature 2009, 461, 930.
Salzberg, B. M.; Davila, H. V.; Cohen, L. B. Nature
1973, 246, 508.
(3)
Johnson, L. V.; Walsh, M. L.; Chen, L. B. Proc Natl
Acad Sci U S A 1980, 77, 990.
(4)
Grinvald, A.; Fine, A.; Farber, I. C.; Hildesheim, R.
Biophys J 1983, 42, 195.
(5)
Fluhler, E.; Burnham, V. G.; Loew, L. M. Biochemistry
1985, 24, 5749.
(6)
Fromherz, P.; Hubener, G.; Kuhn, B.; Hinner, M. J. Eur
Biophys J 2008, 37, 509.
(7)
69, 1272.
(8)
Gonzalez, J. E.; Tsien, R. Y. Biophysical Journal 1995,
Chanda, B.; Blunck, R.; Faria, L. C.; Schweizer, F. E.;
Mody, I.; Bezanilla, F. Nat Neurosci 2005, 8, 1619.
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 9
(9)
Bradley, J.; Luo, R.; Otis, T. S.; DiGregorio, D. A. J
(28) Kojima, H.; Nakatsubo, N.; Kikuchi, K.; Kawahara, S.;
1
2
3
4
5
6
7
8
Neurosci 2009, 29, 9197.
Kirino, Y.; Nagoshi, H.; Hirata, Y.; Nagano, T. Anal Chem 1998,
70, 2446.
(29) He, Q.; Miller, E. W.; Wong, A. P.; Chang, C. J. J Am
Chem Soc 2006, 128, 9316.
(30) Zheng, S.; Lynch, P. L. M.; Rice, T. E.; Moody, T. S.;
Gunaratne, H. Q. N.; de Silva, A. P. Photoch Photobio Sci 2012,
11, 1675.
(31) Larumbe, D.; Gallardo, I.; Andrieux, C. P. Journal of
Electroanalytical Chemistry and Interfacial Electrochemistry
1991, 304, 241.
(10) Sjulson, L.; Miesenbock, G. J Neurosci 2008, 28, 5582.
(11) Millard, A. C.; Jin, L.; Wei, M. D.; Wuskell, J. P.; Lewis,
A.; Loew, L. M. Biophys J 2004, 86, 1169.
(12) Reeve, J. E.; Corbett, A. D.; Boczarow, I.; Kaluza, W.;
Barford, W.; Bayley, H.; Wilson, T.; Anderson, H. L. Angew Chem
Int Ed Engl 2013, 52, 9044.
(13) Park, K.; Deutsch, Z.; Li, J. J.; Oron, D.; Weiss, S. ACS
Nano 2012, 6, 10013.
(14) Marshall, J. D.; Schnitzer, M. J. ACS Nano 2013, 7, 4601.
(15) Siegel, M. S.; Isacoff, E. Y. Neuron 1997, 19, 735.
(16) Tsutsui, H.; Karasawa, S.; Okamura, Y.; Miyawaki, A.
Nat Methods 2008, 5, 683.
9
(32) Rehm, D.; Weller, A. Israel J Chem 1970, 8, 259.
(33) Adams, B. K.; Cherry, W. R. Journal of the American
Chemical Society 1981, 103, 6904.
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
(17) Barnett, L.; Platisa, J.; Popovic, M.; Pieribone, V. A.;
Hughes, T. PLoS One 2012, 7, e43454.
(18) Jin, L.; Han, Z.; Platisa, J.; Wooltorton, J. R.; Cohen, L.
B.; Pieribone, V. A. Neuron 2012, 75, 779.
(34) Baker, M. W.; Macagno, E. R. J Neurosci Methods 2006,
156, 145.
(35) Smith, R. S.; Araneda, R. C. J Neurophysiol 2010, 104,
2963.
(36) Keller, A.; Yagodin, S.; Aroniadou-Anderjaska, V.;
Zimmer, L. A.; Ennis, M.; Sheppard, N. F., Jr.; Shipley, M. T. J
Neurosci 1998, 18, 2602.
(37) Shipley, M. T.; Ennis, M. Journal of Neurobiology 1996,
30, 123.
(38) Yan, P.; Acker, C. D.; Zhou, W. L.; Lee, P.; Bollensdorff,
C.; Negrean, A.; Lotti, J.; Sacconi, L.; Antic, S. D.; Kohl, P.;
Mansvelder, H. D.; Pavone, F. S.; Loew, L. M. Proc Natl Acad Sci
U S A 2012, 109, 20443.
(39) Refojo, D.; Schweizer, M.; Kuehne, C.; Ehrenberg, S.;
Thoeringer, C.; Vogl, A. M.; Dedic, N.; Schumacher, M.; von
Wolff, G.; Avrabos, C.; Touma, C.; Engblom, D.; Schutz, G.;
Nave, K. A.; Eder, M.; Wotjak, C. T.; Sillaber, I.; Holsboer, F.;
Wurst, W.; Deussing, J. M. Science 2011, 333, 1903.
(40) Kuhn, B.; Fromherz, P. J Phys Chem B 2003, 107, 7903.
(41) Kuhn, B.; Fromherz, P.; Denk, W. Biophys J 2004, 87,
631.
(19) Akemann, W.; Sasaki, M.; Mutoh, H.; Imamura, T.;
Honkura, N.; Knopfel, T. Sci Rep 2013, 3, 2231.
(20) St-Pierre, F.; Marshall, J. D.; Yang, Y.; Gong, Y.;
Schnitzer, M. J.; Lin, M. Z. Nat Neurosci 2014, 17, 884.
(21) Kralj, J. M.; Douglass, A. D.; Hochbaum, D. R.;
Maclaurin, D.; Cohen, A. E. Nat Methods 2012, 9, 90.
(22) Hochbaum, D. R.; Zhao, Y.; Farhi, S. L.; Klapoetke, N.;
Werley, C. A.; Kapoor, V.; Zou, P.; Kralj, J. M.; Maclaurin, D.;
Smedemark-Margulies, N.; Saulnier, J. L.; Boulting, G. L.; Straub,
C.; Cho, Y. K.; Melkonian, M.; Wong, G. K.; Harrison, D. J.;
Murthy, V. N.; Sabatini, B. L.; Boyden, E. S.; Campbell, R. E.;
Cohen, A. E. Nat Methods 2014.
(23) Miller, E. W.; Lin, J. Y.; Frady, E. P.; Steinbach, P. A.;
Kristan, W. B.; Tsien, R. Y. P Natl Acad Sci USA 2012, 109, 2114.
(24) Li, L.-s. Nano Letters 2007, 7, 2981.
(25) Tanner, G. A.; Sandoval, R. M.; Dunn, K. W. Am J
Physiol Renal Physiol 2004, 286, F152.
(26) Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski,
M. R. Nature 1998, 396, 60.
(27) Minta, A.; Kao, J. P.; Tsien, R. Y. J Biol Chem 1989,
264, 8171.
8
ACS Paragon Plus Environment

Page 9 of 9
Journal of the American Chemical Society
Table of Contents Graphic
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
VoltageFluor (VF) dyes have the potential to optically measure voltage in excitable membranes with the combination of high spatial and
temporal resolution essential to better characterize the voltage dynamics of large groups of excitable cells. VF dyes sense voltage with high
speed and sensitivity using photoinduced electron transfer (PeT) through a conjugated molecular wire. We show that tuning the driving force
for PeT (ꢀGPeT + w) through systematic chemical substitution modulates voltage sensitivity, estimate (ꢀGPeT + w) values from experimentally
measured redox potentials, and validate the voltage sensitivities in patch-clamped HEK cells for 10 new VF dyes. VF2.1(OMe).H, with a
48% ꢀF/F per 100 mV, shows approximately 2-fold improvement over previous dyes in HEK cells, dissociated rat cortical neurons, and
medicinal leech ganglia. Additionally, VF2.1(OMe).H faithfully reports pharmacological effects and circuit activity in mouse olfactory bulb
slices, thus opening a wide range of previously inaccessible applications for voltage sensitive dyes.
9
ACS Paragon Plus Environment