Cell Chemical Biology, Volume
24
Supplemental Information
Directed Evolution of a Bright
Near-Infrared Fluorescent Rhodopsin
Using a Synthetic Chromophore
Lukas Herwig, Austin J. Rice, Claire N. Bedbrook, Ruijie K. Zhang, Antti Lignell, Jackson
K.B. Cahn, Hans Renata, Sheel C. Dodani, Inha Cho, Long Cai, Viviana
Gradinaru, and Frances H. Arnold
1
Supplemental Information
Supplemental Figures
Figure S1. Synthesis of merocyanine retinal and subsequent synthesis of Schiff base.
Related to Figure 1.
(A)
Merocyanine
retinal was synthesized as shown above. Full characterization of the final compounds can be found at
the end of this document. The final aldehyde product yielded a 3:1
trans
:
cis
ratio at C16 (wavy line). This final
aldehyde product is used throughout this work, and concentration values represent the total merocyanine retinal
concentration (both trans and 16
-
cis
). (B) The aldehyde form of merocyanine retinal was converted to a fre
e Schiff
base as described in the supplemental experimental procedures.
2
Figure S2. Base denaturation of merocyanine
-
bound Arch variants and free merocyanine retinal controls.
Related to Figure 1.
(A) Normalized absorbance spectra of the aldehyde form of
merocyanine retinal (Mero) as
purified (
red
), in alkaline conditions (
light blue
), and in acidic conditions (green). The acid peak at 435 nm likely
represents protonated indolylidene ring. (B) Normalized absorbance spectra of the Schiff base (SB) f
orm of
merocyanine retinal prepared in crude reaction (
red
), in alkaline conditions (
light blue
), and in acidic conditions
(green). Synthesis and characterization of the SB product are described at the end of this document. Over the course
of minutes, the protonated Schiff base peak diminished and a protonated indolylidene ring peak formed at 435 nm
;
for clarity, only an early time point (within 1 min of adding acid) is shown, with its distinct yet incompletely formed
protonated Schiff base peak. (C) Normalized absorbance spectra of purified wild
-
type Arch bound to merocyanine
(
red
). Base denatur
ation yields a deprotonated Schiff base peak (
light blue
). Subsequent addition of acid yields a
broad protonated Schiff base peak (green). Similar results were observed for Mero2 and Mero6 with the protonated
Schiff base spectra shown in
orange
a
nd
dark blue
, respectively. All measurements were taken in buffer [25 mM
Tris; pH 6.5; 200 mM NaCl; 0.15% DDM] with serial reads taken after base denaturation (20
μ
L 5.6M NaOH added
3
to 600
μ
L sample), then acidification (40
μ
L 5M HCl added to 620
μ
L sample). Less base (5
μ
L) and acid (10
μ
L)
were added to the free
-
Schiff base control due to a tendency to form protonated indolylidene in the acidic spectrum.
Spectra were normalized to a maximal peak value of 1.
Figure S3. Merocyanine retinal spec
ific opsin fluorescence in
E. coli
and the effect of merocyanine retinal on
expression of Arch variants.
Related to Figure 2.
(A) Blue curve: Emission spectrum of Arch
-
WT (wild
-
type
Arch); red curve: ScADH6 (
Saccharomyces cerevisiae
cinnamyl alcohol dehydr
ogenase); both spectra were
recorded upon 720 nm excitation using a Tecan Infinite® M200 plate
-
reader; proteins were expressed in presence of
1 μM merocyanine retinal and 500 μM IPTG. ATR
-
bound Arch
-
WT excites and emits at 556/687 nm
(McIsaac et
al., 2014)
. (B)
Expression is quantified by the fluorescence intensity of a C
-
terminally fused CFP tag (ex/em at
425/475 nm). CFP florescence measurements are taken on IPTG
-
induced
E. coli
suspensions in the presence (blue
bars) and the absence (grey bars) of 10 μM
merocyanine retinal (Mero). For all tested variants, merocyanine retinal
has an insignificant effect on CFP fluorescence. A 10
-
fold greater concentration of merocyanine retinal was used in
this test, but all other factors are identical to screening conditi
ons (e.g. 500 μM IPTG). Error bars represent the
standard deviation of six (with merocyanine) or two measurements (without merocyanine). The control protein
(ScADH6 (
Saccharomyces cerevisiae
cinnamyl alcohol dehydrogenase)) does not encode CFP and represen
ts the
nonspecific cellular background signal.
4
Figure S4. NIR photo
-
bleaching of merocyanine
-
bound Arch variants in
E. coli
. Related to Figure 4.
Representative decay curves of CFP
-
tagged Arch
-
WT (A) and Mero
-
6 (B). Each
E. coli
image contained multipl
e
E.
coli
cells within the field of view. For each
E. coli
image, CFP and Arch fluorescence were respectively bleached
with 405 nm light (500 ms exposures;
teal
dots) then near
-
infrared (NIR) 727 nm light (250 ms exposures;
red
dots).
Fluorescence
was normalized for protein absorption profile and laser power. Each decay curve was fit with a single
exponential curve (black) between 3 and 60 seconds. The decay rates for the representative curves are shown in the
legends (units = s
-
1
). (C) Decay rates
determined for CFP
-
tagged Arch
-
WT (n = 5 samples) and Mero
-
6 (n = 4
samples), in both the CFP and NIR channels. (D) For each sample of Arch
-
WT (blue dots) and Mero
-
6 (green dots),
the decay rate of merocyanine
-
bound Arch (determined in the NIR channel; 72
7 nm) was normalized by the decay
rate of CFP (determined at 405
nm). The black horizontal line marks the mean normalized decay rate (Table 1) and
the black vertical bars represent the 95% confidence interval determined from 1000 bootstraps.
A
B
C
D
5
Figure S5.
Expression and NIR fluorescence in live eukaryotic (HEK) cells grown in D10 medium.
Related to
Figure 4.
(A) Eukaryotic construct for expression of a fusion of GFP and Arch mutant (Arch Mut) driven by a
CaMKIIa
promoter. TS: trafficking sequence. ES: Export signal. WPRE: woodchuck hepatitis virus post
-
transcriptional enhancer. (B) Representative NIR (ex/em 727/766
-
854 nm) and GFP (ex/em 473/>498 nm)
fluorescence images of HEK cells expressing wild
-
type Arch or M
ero
-
6. Low and high contrast images of the GFP
channel are shown due to significant differences in expression between Arch variants; such treatment was not
required in the NIR channel due to enhanced fluorescence intensity of Mero
-
6. Indicated HEK cells ar
e expressed
with merocyanine retinal (1 μM final concentration) for 24 hrs. Scale bar: 10 μm. (C) Mean GFP fluorescence
intensity of HEK cells expressing Arch
-
WT (n = 29) and Mero
-
6 (n = 17). (D) Mean NIR fluorescence of WT,
Mero
-
6, and non
-
transfected (NT
) cells with merocyanine (n = 29, 17, and 28, respectively) and without
merocyanine (n = 14, 15, and 21 respectively). (E) NIR fluorescence normalized by GFP fluorescence for HEK cells
with or without merocyanine added. All error bars represent one standar
d error of the mean. (F) Patch
-
clamp
electrophysiology of transfected HEK cells was used to test merocyanine retinal
-
bound, wild
-
type Arch (
n
= 7 cells)
and Mero
-
6 (
n
= 10 cells) for voltage sensitivity. Both variants were expressed in the presence of 1 μM
merocyanine
retinal for 24
hrs. With excitation at 650 nm and detection >780 nm, wild
-
type Arch showed approximately 2%
change in fluorescence intensity during a 100 mV step in the clamped voltage (voltage step from
-
70 mV to +30
mV). However, precise meas
urement of the intensity change is difficult due to high noise levels in the wild type
readings. Mero
-
6 readings were less noisy and showed a 4% change in fluorescence intensity with an identical
voltage step.
Δ
F/F = change in fluorescence from baseline, n
ormalized by baseline fluorescence.
6
Figure S6. Light
-
driven ion pumping in HEK cells. Related to Figure S5.
Patch
-
clamp electrophysiology of
transfected HEK cells was used to measure photocurrents of wild
-
type Arch and Mero
-
6 with short light pulses (0.
5
s) of varying wavelength. Recordings were done of cells incubated with media containing (A) 1% FBS with 1 μM
merocyanine retinal added or (B) 10% FBS without any added merocyanine retinal. The amount of FBS used
dictates the concentration of ATR in the m
edium. Photocurrents were measured for four different wavelengths of
light. Currents were only observed for wild
-
type Arch at 485 nm, 590 nm, and 650 nm but not 730 nm light with (
n
= 6 cells) or without (
n
= 10 cells) 1 uM merocyanine retinal. No currents
were observed for Mero
-
6 at any
wavelength with (
n
= 10 cells) or without (
n
= 7 cells) 1 μM merocyanine retinal. Example traces with 590 nm light
for wild
-
type Arch and Mero
-
6 are shown to the left for both conditions (A) 1% FBS with 1 μM merocyanine ret
inal
added and (B) 10% FBS. These results suggest that Mero
-
6 is incapable of pumping with either ATR or
merocyanine retinal.
7
Figure S7. Spectral scans of purified merocyanine retinal
-
bound Arch variants.
Related to Table 1.
(A)
Fluorescent emission scan (730
-
850 nm) with excitation set at 700 nm. (B) Excitation scan (610
-
790 nm) with
emission detected at 810 nm. Spectra were collected from 200
μ
L samples (diluted with DDM desalt buffer to an
absorbance (A
680
) below 0.07) usi
ng a SpectraMax: M5 plate reader (4 nm bandwidth; 1 nm steps). Spectra were
normalized to peak values.
8
Figure S8.
Effects of merocyanine retinal on HEK cell health. Related to Figure S5.
Representative images of
non
-
transfected HEK cells after a 48
hrs incubation with varying concentrations of merocyanine retinal (0
-
10 μM)
in medium (DMEM with 1% FBS).
9
Figure S9. Representative data used in quantum yield and extinction coefficient calculations of merocyanine
-
bound Arch variants.
Related to Table
1.
(A)
Representative fluorescent emission spectra (excitation at 680 nm)
from one of four dilutions made for each Arch variant and standard dye (Alexa Fluor 750). Baseline fluorescence
was measured for DDM desalt buffer (proteins) or PBS (dye). Spectra w
ere integrated as described in methods.
Absorbance values (A
680
) were determined for each sample and dilution, which were corrected by buffer controls.
(B) Plot of integrated fluorescence vs. absorbance. Equations are shown for linear regressions to each s
et of data. (C)
10
Time course of absorbance spectra (350
-
850 nm), collected during the hydroxylamine bleaching of merocyanine
-
bound Arch
-
WT every 12 min (C) and Mero
-
6 every 3 min (E). Peak values of pigmented opsin (758
-
761 nm
depending on mutant) and the f
ree oxime form of retinal (441 nm) were monitored. The change in absorbance over
time for pigmented opsin protein (
Δ
Opsin) is plotted against the concurrent change in oxime retinal for Arch
-
WT
(D) and Mero
-
6 (F). The linear region of the plot (blue spectra
in C and E contributing blue points in D and F) is
fitted with a linear regression (equation and R
2
, shown). Data points with minimal change in oxime retinal
absorption (red points from red spectra) were excluded from the regression. The slope of the line
ar regression
(
Δ
Opsin /
Δ
Retinal) was used in determining the extinction coefficient for each variant (see Supplemental
Experimental Procedures).
Supplemental
Data
Data S1.
NMR characterization o
f merocyanine retinal synthesis.
Related to experimental methods
and Figure S1.
NMR spectra are shown with the structure of corresponding intermediate and target
compounds. To help interpret the
1
H NOESY experiments, circled peaks are paired with indicated
interactions via color
-
coding.
Supplemental Experimental Procedures
Arch
-
His
6x
and Arch
-
CFP Nucleotide Sequences
>
Arch
-
His
6
ATGGACCCGATTGCTCTGCAGGCAGGTTACGACCTGCTGGGTGATGGCCGTCCGGAAACCCTGTGGCT
GGGTATTGGTACCCTGCTGATGCTGATCGGCACCTTCTACTTCCTGGTGCGTGGCTGGGGCGTCACTGA
TAAGGATGCTCGTGAATACTACGCTGTTACCATCCTGGTGCCAGGCATTGCGAGCGCAGCTTACCTGT
CTATGTTCTTCGGCATCGGTCTGACCGAAGTAACGGTAGGTGGTGAAATGC
TGGACATTTACTACGCA
CGTTACGCTGACTGGCTGTTCACTACCCCACTGCTGCTGCTGGATCTGGCTCTGCTGGCCAAAGTTGAC
CGTGTTACCATCGGCACTCTGGTAGGTGTGGACGCACTGATGATCGTGACCGGTCTGATTGGTGCTCT
GTCTCACACTGCTATCGCTCGTTATTCCTGGTGGCTGTTTAGCACCATCTGCATGATTGTGGTGCTGTA
CTTCCTGGCGACCTCTCTGCGCTCCGCTGCTAA
AGAGCGTGGTCCGGAAGTTGCTTCTACTTTTAACAC
CCTGACCGCTCTGGTGCTGGTGCTGTGGACCGCTTATCCTATCCTGTGGATCATCGGCACCGAAGGTG
CGGGTGTTGTTGGTCTGGGCATTGAAACCCTGCTGTTCATGGTTCTGGATGTTACTGCCAAAGTTGGCT
TTGGTTTTATTCTGCTGCGTTCCCGTGCGATCCTGGGTGACACCGAAGCCCCAGAACCGAGCGCTGGT
GCAGATGTTTCCGCT
GCGGAC
gcagctgcactcgag
CACCACCACCACCACCAC
TGA
>
Arch
-
e
CFP
ATGGACCCGATAGCACTACAGGCGGGATACGACCTACTCGGGGACGGTCGCCCCGAGACGTTGTGGT
TGGGTATCGGAACGTTACTAATGCTCATCGGGACCTTCTACTTCCTCGTCAGAGGGTGGGGGGTCACC
GACAAGGACGCCCGCGAGTACTACGCGGTCACGATCCTCGTGCCGGGGATCGCGTCGGCGGCGT
ACC
TGTCGATGTTCTTCGGCATCGGCCTGACGGAAGTCACGGTCGGTGGCGAAATGCTCGACATCTACTAC
GCGCGGTACGCGGACTGGCTGTTCACCACGCCGCTGCTGCTGCTCGACCTCGCGCTGCTCGCAAAGGT
CGACCGCGTCACCATCGGGACGCTCGTCGGCGTCGACGCGCTGATGATCGTCACCGGCCTCATCGGCG
CGCTCTCGCACACGGCGATCGCGCGGTACTCCTGGTGGCTGTTCAGCAC
GATTTGCATGATCGTCGTG
CTGTACTTCCTCGCCACGAGCCTCCGGAGCGCGGCGAAGGAGCGCGGACCTGAAGTCGCGAGCACCT
TCAACACGTTGACCGCGCTGGTCCTGGTGCTCTGGACGGCCTACCCGATCCTGTGGATCATCGGAACC
GAGGGCGCCGGCGTCGTCGGCCTCGGCATCGAGACCCTCCTGTTCATGGTTCTCGACGTGACGGCCAA
GGTCGGCTTCGGCTTCATCCTGCTCCGCAGCCGC
GCCATCCTCGGCGACACCGAGGCGCCGGAGCCCT
CCGCGGGCGCCGACGTCTCCGCCGCGGAC
gcggccgcactcgag
ATGGTGAGCAAGGGCGAGGAGCTGTTCA
CCGGGGTGGTGCCCATCCTGGTCGAGCTGGACGGCGACGTAAACGGCCACAAGTTCAGCGTGTCCGG
CGAGGGCGAGGGCGATGCCACCTACGGCAAGCTGACCCTGAAGTTCATCTGCACCACCGGCAAGCTG
11
CCCGTGCCCTGGCCCA
CCCTCGTGACCACCCTGACCTGGGGCGTGCAGTGCTTCAGCCGCTACCCCGA
CCACATGAAGCAGCACGACTTCTTCAAGTCCGCCATGCCCGAAGGCTACGTCCAGGAGCGCACCATCT
TCTTCAAGGACGACGGCAACTACAAGACCCGCGCCGAGGTGAAGTTCGAGGGCGACACCCTGGTGAA
CCGCATCGAGCTGAAGGGCATCGACTTCAAGGAGGACGGCAACATCCTGGGGCACAAGCTGGAGTAC
AA
CTACATCAGCCACAACGTCTATATCACCGCCGACAAGCAGAAGAACGGCATCAAGGCCAACTTCA
AGATCCGCCACAACATCGAGGACGGCAGCGTGCAGCTCGCCGACCACTACCAGCAGAACACCCCCAT
CGGCGACGGCCCCGTGCTGCTGCCCGACAACCACTACCTGAGCACCCAGTCCGCCCTGAGCAAAGAC
CCCAACGAGAAGCGCGATCACATGGTCCTGCTGGAGTTCGTGACCGCCGCCTAA
Purification of Arch variants
Arch variants of interest were transformed into NiCo
-
pLEMO cells and plated (LB; 100 μg/mL ampicillin; 34
μg/mL chloramphenicol) for lawn growth. The next day, cells were scraped from the plate, re
-
suspended in media
(2xYT; 10
0 μg/mL ampicillin; 34 μg/mL chloramphenicol), and 1 mL of re
-
suspended cells inoculated into 1 L
medium. For typical variants, at least 2 L of culture was prepared. Cultures were grown at 37 °C until an optical
density of 0.6
-
0.8 was reached. At this poin
t merocyanine retinal (at 10 mM in 100 % ethanol) and IPTG were
added to a final concentration of 1 μM and 500 μM, respectively. Cells were then grown for 3 hours at 30 °C in the
dark, harvested via centrifugation and stored at
-
80 °C.
Holo
-
Arch variants
were purified as reported
(McIsaac et al., 2014)
. Frozen cells were resuspended in
buffer components such that the final solution consisted of [20 mM Tris
-
HCl; pH 7.5; 200 mM NaCl; 20 mM
imidazole; 1% dodecylmaltopyranoside (DDM; Anatrace); 10 mg/mL lysozy
me (Sigma); ~1 mg/mL DNase
(Sigma)] with the final volume being ~5x the volume of the cells. Resuspended cells were incubated with shaking at
room temperature in the dark for 1.5 hours, then centrifuged in a microcentrifuge for 30 min 4 °C. From this point
on, the protein solution was kept on ice or at 4 °C and shielded from light. Variants were purified via nickel affinity
chromatography (ÄKTA Express) wherein a 1 mL HisTrap HP column (GE healthcare) was equilibrated with [20
mM Tris
-
HCl; pH 7.5; 200 mM Na
Cl; 20mM imidazole; 0.02% DDM] then over
-
loaded with lysate. The loaded
column was washed with 10 column volumes of equilibration buffer then eluted in a 20
-
500 mM imidazole gradient
over 20 column volumes. Absorbance at 280 and 760 nm was monitored, and p
igmented fractions with highest
purity were pooled and immediately desalted into ‘DDM desalt buffer’ [20 mM Tris
-
HCl; pH 6.5; 200 mM NaCl;
0.15% DDM] via PD
-
10 desalting columns (GE Healthcare).
To prepare
apo
protein for retinal binding assays,
Arch vari
ants were expressed as described with retinal
not added upon induction. To minimize temperature shock of what we expected to be a less stable
apo
protein, cells
were lysed with a microfluidizer in the absence of detergent. Lysate was centrifuged at low spe
ed (20,000 x g) to
remove cell debris, then at high speed (185,500 x g) to collect membranes. Membranes were resuspended in [20
mM Tris
-
HCl; pH
7.5; 200 mM NaCl] and stored at
-
80 ̊C. Membranes were thawed on ice and dissolved with
1.5% lauryl maltose
neopentyl glycol (LMNG) detergent. LMNG was found empirically to promote retinal binding
to
apo
-
Arch (wild
-
type
apo
-
Arch could not bind ATR or merocyanine retinal if purified in DDM; data not shown).
Dissolved
apo
-
Arch membranes were purified via Ni affini
ty chromatography as described above except that
0.0
15
% LMNG replaced DDM in all buffers,
4.5
mL of Ni
-
NTA
superflow
resin (Qiagen) was used, and protein
was eluted in one step. SDS
-
PAGE confirmed the presence and purity of
apo
proteins before PD
-
10 desalt
ing into
[20 mM Tris
-
HCl; pH 7.5; 200 mM NaCl; 0.0
15
% LMNG] (‘LMNG desalt buffer’). For the binding assay, purified
proteins were concentrated no more than 2x
via spin filtration (Millipore).
Quantum
y
ield and extinction coefficient determination
As described in Wall
et al
. (2015), the quantum yields of merocyanine
-
bound Arch variants were calculated by
comparison of each variant to the Alexa Fluor® 750 NHS Ester (succinimidyl ester) dye with known quantum yield
of 0.12 (
catalog number:
A20011, Life Technologies Corporation)
(Wurth et al., 2012)
. With peak
excitation/emission at 749/775 nm, this reference dye has similar spectral characteristics to the merocyanine
-
bound
Arch variants. For each Arch variant and the reference dye, four seri
al dilutions were made with corrected
absorbance values (A
680
) below 0.04. For each dilution, a Tecan Infinite® M200 plate
-
reader was used to measure
the A
680
, then a fluorescence emission scan between 708
-
850 nm when excited at 680 nm (Figure S9
A
). The
in
tegrated fluorescence from each baseline
-
corrected emission peak was calculated via trapezoidal integration.
Integrated fluorescence was plotted vs A
680
; for each Arch variant and the reference dye, a slope
(fluorescence/absorbance) was calculated (Figure
S9
B
). These slopes (m) were used in the following equation to
calculate the quantum yield of each protein (QY
P
): QY
P
= QY
ref
* (m
P
/ m
ref
) * (
η
P
2
/
η
ref
2
). Where
η
P
and
η
ref
are the
respective refractive indices of the protein buffer [20 mM Tris
-
HCl; pH 6.
5; 200 mM NaCl; 0.15% DDM] (
η
P
=