Discovery and Characterization of a Peptide That Enhances Endosomal Escape of Delivered Proteins in Vitro and in Vivo

Article abstract of DOI:10.1021/jacs.5b05694

The inefficient delivery of proteins into mammalian cells remains a major barrier to realizing the therapeutic potential of many proteins. We and others have previously shown that superpositively charged proteins are efficiently endocytosed and can bring associated proteins and nucleic acids into cells. The vast majority of cargo delivered in this manner, however, remains in endosomes and does not reach the cytosol. In this study we designed and implemented a screen to discover peptides that enhance the endosomal escape of proteins fused to superpositively charged GFP (+36 GFP). From a screen of peptides previously reported to disrupt microbial membranes without known mammalian cell toxicity, we discovered a 13-residue peptide, aurein 1.2, that substantially increases cytosolic protein delivery by up to ～5-fold in a cytosolic fractionation assay in cultured cells. Four additional independent assays for nonendosomal protein delivery collectively suggest that aurein 1.2 enhances endosomal escape of associated endocytosed protein cargo. Structure-function studies clarified peptide sequence and protein conjugation requirements for endosomal escape activity. When applied to the in vivo delivery of +36 GFP-Cre recombinase fusions into the inner ear of live mice, fusion with aurein 1.2 dramatically increased nonendosomal Cre recombinase delivery potency, resulting in up to 100% recombined inner hair cells and 96% recombined outer hair cells, compared to 0-4% recombined hair cells from +36-GFP-Cre without aurein 1.2. Collectively, these findings describe a genetically encodable, endosome escape-enhancing peptide that can substantially increase the cytoplasmic delivery of cationic proteins in vitro and in vivo.

Full text of DOI:10.1021/jacs.5b05694

Subscriber access provided by UNIV OF LETHBRIDGE
Article
Discovery and characterization of a peptide that enhances
endosomal escape of delivered proteins in vitro and in vivo
Margie Li, Yong Tao, Yilai Shu, Jonathan R LaRochelle, Angela Steinauer,
David Thompson, Alanna Schepartz, Zheng-Yi Chen, and David R Liu
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.5b05694 • Publication Date (Web): 14 Oct 2015
Downloaded from http://pubs.acs.org on October 17, 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 12
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Discovery and characterization of a peptide that enhances
endosomal escape of delivered proteins in vitro and in vivo
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
Margie Li†, Yong Tao‡, Yilai Shu‡°¦, Jonathan R. LaRochelle#, Angela Steinauer§, David Thompson†, Alanna
ǁ
Schepartz§,#, Zheng-Yi Chen‡, and David R. Liu*†
†Department of Chemistry & Chemical Biology, Harvard University, 12 Oxford St, Cambridge, MA 02138 USA
‡Department of Otolaryngology and Program in Neuroscience, Harvard Medical School and Eaton Peabody Laboratory, Massachu-
setts Eye and Ear Infirmary, Boston, MA 02114 USA
°Department of Otology and Skull Base Survery, Eye, Ear, Nose and Throat Hospital, Shanghai Medical College, Fudan University,
Shanghai 200031, China
¦Key Laboratory of Hearing Medicine, Ministry of Health, Shanghai, 200031, China
#Department of Molecular, Cellular and Developmental Biology, Yale University, New Haven, Connecticut 06520-8107, United States
§Department of Chemistry, Yale University, New Haven, Connecticut 06520-8107, United States
ǁHoward Hughes Medical Institute, Harvard University, 12 Oxford St, Cambridge, MA 02138 USA
ABSTRACT: The inefficient delivery of proteins into mammalian cells remains a major barrier to realizing the therapeutic potential of many
proteins. We and others have previously shown that superpositively charged proteins are efficiently endocytosed and can bring associated
proteins and nucleic acids into cells. The vast majority of cargo delivered in this manner, however, remains in endosomes and does not reach
the cytosol. In this study we designed and implemented a screen to discover peptides that enhance the endosomal escape of proteins fused to
superpositively charged GFP (+36 GFP). From a screen of peptides previously reported to disrupt microbial membranes without known
mammalian cell toxicity, we discovered a 13-residue peptide, aurein 1.2, that substantially increases cytosolic protein delivery by up to ~5-fold
in a cytosolic fractionation assay in cultured cells. Four additional independent assays for non-endosomal protein delivery collectively suggest
that aurein 1.2 enhances endosomal escape of associated endocytosed protein cargo. Structure-function studies clarified peptide sequence
and protein conjugation requirements for endosomal escape activity. When applied to the in vivo delivery of +36 GFP–Cre recombinase fu-
sions into the inner ear of live mice, fusion with aurein 1.2 dramatically increased non-endosomal Cre recombinase delivery potency, resulting
in up to 100% recombined inner hair cells and 96% recombined outer hair cells, compared to 0-4% recombined hair cells from +36-GFP-Cre
without aurein 1.2. Collectively, these findings describe a genetically encodable, endosome escape-enhancing peptide that can substantially
increase the cytoplasmic delivery of cationic proteins in vitro and in vivo.
INTRODUCTION
We previously reported the discovery of superpositively-
charged proteins, a class of engineered and naturally occurring
proteins that have abnormally high net positive charge, and their
ability to potently deliver proteins and nucleic acids into mammali-
an cells14-17. While superpositive proteins are very efficiently endo-
cytosed18,19 and can be more effective for protein delivery than
CPPs15, the vast majority of endocytosed proteins remain seques-
tered in endosomes18 that either mature into lysosomes, resulting in
protein degradation, or are recycled to the surface of the cell20,
resulting in extracellular protein release (Figure 1A). As a result,
relatively high concentrations (µM) of exogenous protein are typi-
cally needed for modest cytosolic or nuclear delivery. Although
superpositively charged proteins can slow endosomal maturation18,
the inefficiency of endosomal escape enables only a small fraction
of delivered protein to reach the cytosol21.
Proteins that bind extracellular targets, including monoclonal
antibodies1, Fc fusions2, and cytokines3, have served as important
therapeutics4. Fully realizing the therapeutic potential of proteins,
however, requires methods to enable exogenous proteins to access
intracellular targets. Because the vast majority of proteins cannot
spontaneously cross cell membranes, the development of intracel-
lular protein delivery methods could facilitate applications includ-
ing enzyme replacement therapies for metabolic diseases5, tran-
scription factor-driven changes in cell fate6, and genome editing7.
Several methods for protein delivery have been explored in the
past decade, including cell-penetrating peptides (CPPs)8, penta-arg
proteins9, receptor ligands10, and lipid nanoparticles11. While these
and other methods have advanced the field of protein delivery,
challenges including cytotoxicity, lack of generality, low potency, or
poor in vivo activity continue to limit their therapeutic rele-
To address this protein delivery bottleneck, we sought to dis-
cover peptides that facilitate endosomal escape when fused to en-
docytosed superpositively charged proteins such as +36 GFP.
vance12,13
.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 12
Membrane-active peptides such as influenza-derived HA2 have A major challenge to developing agents that enhance endoso-
been reported to be endosomolytic22. However, many of these
peptides, including HA2, are cytotoxic at concentrations required
for protein delivery23,24. Antimicrobial peptides (AMPs) are a class
of membrane-active peptides that penetrate microbial membranes
to provide defense against bacteria, fungi, and viruses, often with
high selectivity25. Given that many AMPs exhibit minimal toxicity
to mammalian cells26, we hypothesized that the altered endosomal
environment or endosomal membrane curvature could induce
some AMPs to be endosomolytic without exhibiting significant
mammalian cell toxicity at useful concentrations. To test this hy-
pothesis, we performed a screen of AMPs for their ability to in-
crease protein delivery into the cytosol.
mal escape is the lack of well-established assays that can distinguish
proteins trapped in the endosomes from proteins released into the
cytosol. Commonly used enzyme delivery assays involve substrates
and products that can freely diffuse through membranes and can-
not differentiate between endosomal and cytosolic proteins. To
overcome this challenge, we used multiple independent assays that
reflect the interaction of a variety of cargo with a variety of cytosolic
targets to evaluate endosomal escape of AMP–protein fusions.
1
2
3
4
5
6
7
8
9
In this study, we discovered aurein 1.2 (GLFDIIKKIAESF) as a
peptide that enhances the endosomal escape of a variety of cargo
fused to +36 GFP. We elucidated structure-function relationships
within aurein 1.2 using alanine scanning and mutational analysis.
Results from three independent delivery assays confirmed that
treatment of mammalian cells with cargo proteins fused to aurein
1.2–+36 GFP result in more efficient cytosolic delivery than the
same proteins fused to +36 GFP alone. Finally, we describe the
ability of aurein 1.2 to enhance non-endosomal protein delivery in
vivo. Cre recombinase enzyme was delivered into hair cells in the
cochlea (inner ear) of live mice with much greater (> 20-fold) po-
tency when fused with aurein 1.2 than in the absence of the pep-
tide. These results together provide a simple molecular strategy for
enhancing the cytosolic delivery of proteins in cell culture and in
vivo that is genetically encoded, localized to cargo molecules, and
does not require global treatment with cytotoxic small molecules.
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
A
Electrostatic
interaction with
proteoglycans
Endosomal
escape
(+)
(-)
(+)
(+)
Mature
endosomes
(+)
(+)
Lysosomal
Endocytosis
degradation
Early
endosomes
(+)
Membrane recycling
Decreasing pH
RESULTS
Preparation of antimicrobial peptide conjugates of super-
charged GFP–Cre fusion proteins. We sought AMPs from the
Antimicrobial Peptide Database27 that are ≤ 25 amino acids long
(to facilitate their preparation and conjugation to +36 GFP), lack
post-translational modifications, and are not known to be toxic to
mammalian cells. Based on these criteria, we identified 36 AMPs
ranging from 9 to 25 amino acids in length (Table 1). Each of the
peptides was synthesized on solid phase with an LPETGG se-
quence appended to their C-terminus to enable sortase-catalyzed
conjugation28 (Figure 1B). Assembly of proteins using sortase
proved more amenable to rapid screening than the construction
and expression of the corresponding fusions, especially since sever-
al AMP fusions do not express efficiently in E. coli.
B
GGG
GGG-+36GFP-Cre
Peptide stock
solutions
Sortase
GGG
LPETGG
Sortase-mediated peptide
conjugation
The peptides were conjugated to purified GGG–(+36 GFP)–
Cre using our previously described evolved sortase A enzyme
(eSrtA)28. Sortase catalyzes the transpeptidation between a sub-
strate containing the C-terminal LPETGG and a substrate contain-
ing an N-terminal glycine to form a native peptide bond linkage
and a protein identical to the product of translational fusion.
Figure 1. (A) Overview of protein delivery in mammalian cells. Cati-
onic macromolecules such as +36 GFP interact with anionic sulfated
proteoglycans on the cell surface and are endocytosed and sequestered
in early endosomes. The early endosomes can acidify into late endo-
somes or lysosomes. Alternatively, early endosomes may be trafficked
back to the cell surface as part of the membrane-recycling pathway. To
access the cytoplasm, an exogenous cationic protein must escape endo-
somes before it is degraded or exported. (B) Sortase-mediated conju-
gation of peptides with +36 GFP–Cre recombinase prior to screening.
Sortase was used to conjugate synthetic peptides containing a C-
terminal LPETGG with expressed +36 GFP–Cre containing an N-
terminal GGG. The resulting peptide–LPETGGG–+36 GFP–Cre
fusion proteins have the same chemical composition as expressed re-
combinant proteins, but are more easily assembled.
Table 1. List of peptides chosen from the Antimicrobial Pep-
tide Database (APD)
Conjugation
Label APD number Sequence
efficiency
A
B
C
D
E
AP00408
AP00405-11 FISAIASMLGKFL
FLFPLITSFLSKVL
55%
70%
-
AP00327
AP01434
AP00013
AP00025
AP00094
GWFDVVKHIASAV
FFGSVLKLIPKIL
GLFDIIKKIAESF
HGVSGHGQHGVHG
FLPLIGRVLSGIL
-
77%
20%
-
F
G
2
ACS Paragon Plus Environment

Page 3 of 12
Journal of the American Chemical Society
70
H
AP00012
AP00014
AP00015
AP00016
AP00017
AP00018
AP00019
AP00020
AP00021
AP00022
AP00101
AP00351
AP00352
AP00353
AP00567
AP00597
AP00818
AP00866
AP00870
AP00875
AP00898
AP01211
AP01249
GLFDIIKKIAESI
28%
60
1
2
3
4
5
6
7
8
I
GLLDIVKKVVGAFGSL
GLFDIVKKVVGALGSL
GLFDIVKKVVGAIGSL
GLFDIVKKVVGTLAGL
GLFDIVKKVVGAFGSL
GLFDIAKKVIGVIGSL
GLFDIVKKIAGHIAGSI
GLFDIVKKIAGHIASSI
GLFDIVKKIAGHIVSSI
FVQWFSKFLGRIL
GLFDVIKKVASVIGGL
GLFDIIKKVASVVGGL
GLFDIIKKVASVIGGL
VWPLGLVICKALKIC
NFLGTLVNLAKKIL
FLPLIGKILGTIL
-
50
J
13%
40
K
-
30
L
18%
20
M
N
-
10
-
0
(NP)
A
B
E
F
H
J
L
R
S
U
V
W
X
Y
AA AB AC AD AE AI (+)
O
-
P
-
9
Figure 2. Primary screen for cytosolic delivery of Cre recombinase in
BSR.LNL.tdTomato cells. Initial screen of 20 peptide–(+36 GFP)–
Cre conjugated proteins. Cytosolic Cre delivery results in recombina-
tion and tdTomato expression. The percentage of tdTomato positive
cells was determined by fluorescence image analysis. 250 nM+36
GFP–Cre was used as the no-peptide control (NP), and addition of
100 µM chloroquine was used as the positive control (+). Cells were
treated with 250 nM protein for 4 h in serum-free DMEM. Cells were
washed and supplanted with full DMEM and incubated for 48 h. Error
bars represent the standard deviation of three independent biological
replicates.
Q
-
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
R
51%
11%
-
S
T
U
4%
4%
34%
14%
86%
-
V
W
X
Y
FLPIIAKVLSGLL
Z
FLPIVGKLLSGLL
AA
AB
AC
AD
AE
AF
AG
AH
AI
AJ
FLSSIGKILGNLL
88%
70%
81%
20%
63%
-
The reporter BSR.LNL.tdTomato cells were incubated with
250 nM of each peptide–(+36 GFP)–Cre protein in serum-free
media. In the absence of any conjugated peptide, treatment of re-
porter cells with 250 nM +36 GFP–Cre protein resulted in 4.5% of
the cells expressing tdTomato, consistent with previous reports18.
The same concentration of protein incubated with 100 µM chloro-
quine as a positive control resulted in an average of 48% recom-
bined cells (Figure 2). The results of chloroquine treatment varied
substantially between independent replicates. As chloroquine is
known to be toxic to cells above 100 µM, we speculate that this
variability arises from the small differences between chloroquine's
efficacious and toxic dosages.
FLSGIVGMLGKLF
TPFKLSLHL
GILDAIKAIAKAAG
AP00013-G LFDIIKKIAESF
AP00013-2x LFDIIKKIAESGFLFDIIKKIAESF
AP00722-75 GLLNGLALRLGKRALKKIIKRLCR
-
His13
GHHHHHHHHHHHHH
FKCRRWQWRM
-
AP00512
AP00553
42%
-
KTCENLADTY
Peptides were synthesized with a C-terminal LPETGG tag to enable
conjugation with an evolved sortase (eSrtA). Conjugation efficiencies
were calculated based on LC-MS results using peak abundance as de-
termined through MaxEnt protein deconvolution.
Ten of the screened peptide conjugates resulted in recombina-
tion efficiencies that were significantly above that of +36 GFP–Cre
(Figure 2). The most potent functional delivery of Cre was ob-
served with aurein 1.2–+36 GFP–Cre (Table 1, entry E). Treat-
ment with aurein 1.2–+36 GFP–Cre resulted in an average of 40%
recombined cells, comparable to that of the chloroquine positive
control (Figure 2). To investigate the impact of differential conju-
gation efficiency on peptide performance, we compared citropin
1.3 (Table 1, entry U), which displayed a moderate level of recom-
bination and the lowest level of conjugation (4%), to aurein 1.2,
which has the highest level of recombination and also a high level of
conjugation (77%).
The efficiency of eSrtA-mediated conjugation varied widely
among the peptides (Supporting Information Figure S1). Of the 36
peptides chosen for screening, 20 showed detectable (4% to 88%)
sortase-mediated conjugation to +36 GFP–Cre, as observed by
LC-MS, to generate desired peptide–LPETGGG–(+36 GFP)–Cre
fusion proteins (Table 1). Unreacted peptide was removed by ul-
trafiltration with a 30-kD molecular weight cut off membrane.
Primary screen for endosomal escape. We assayed the ability
of each peptide–(+36 GFP)–Cre recombinase fusion when added
to culture media to effect recombination in BSR.LNL.tdTomato
cells15, a hamster kidney cell line derived from BHK-21 (Support-
ing Information Figure S2). Because Cre recombinase must enter
the cell, escape endosomes, enter the nucleus, and catalyze recom-
bination to generate tdTomato fluorescence, this assay reflects the
availability of active, non-endosomal recombinase enzyme that
reaches the nucleus. As a positive control, we treated cells with +36
GFP–Cre and chloroquine, a known endosome-disrupting small
molecule 29.
Both aurein 1.2–+36 GFP–Cre and citropin 1.3–+36 GFP–Cre
were cloned, expressed, and purified as fusion proteins. The re-
combination signal from treatment with 250 nM of expressed au-
rein 1.2–+36 GFP–Cre was 10.4-fold above that of +36 GFP–Cre.
In contrast, treatment with 250 nM expressed citropin 1.3–+36
GFP–Cre did not induce any enhanced Cre delivery. When the
treatment concentration was increased to 1 µM, aurein 1.2–+36
GFP–Cre and citropin 1.3–+36 GFP–Cre resulted in 3.8-fold and
3.0-fold higher recombination levels, respectively, than that of +36
GFP–Cre alone (Figure 3A). These results suggest that while au-
rein 1.2 and citropin 1.3 both enhance the delivery of functional,
non-endosomal +36 GFP–Cre protein at high concentrations,
aurein 1.2 has greater efficacy than citropin 1.3 at lower concentra-
tions.
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 12
concentration, the peptides are thought to alter the curvature of the
membrane enough to break apart the compartment.
A
1
2
3
4
5
6
7
8
90
80
70
60
50
40
30
20
10
0
Table 2. Site-directed mutagenesis of aurein 1.2
125 nM
250 nM
500 nM
1 µM
Label
Aurein 1.2
G1A
L2A
F3A
D4A
I5A
I6A
K7A
K8A
I9A
E11A
S12A
F13A
K7H
K8H
E11H
K7R
K8R
E11R
E11D
Sequence
GLFDIIKKIAESF
ALFDIIKKIAESF
GAFDIIKKIAESF
GLADIIKKIAESF
GLFAIIKKIAESF
GLFDAIKKIAESF
GLFDIAKKIAESF
GLFDIIAKIAESF
GLFDIIKAIAESF
GLFDIIKKAAESF
GLFDIIKKIAASF
GLFDIIKKIAEAF
GLFDIIKKIAESA
GLFDIIHKIAESF
GLFDIIKHIAESF
GLFDIIKKIAHSF
GLFDIIRKIAESF
GLFDIIKRIAESF
GLFDIIKKIARSF
GLFDIIKKIADSF
(NP)
aurein 1.2 (E) citropin 1.3 (U)
(+)
9
B
120
100
80
60
40
20
0
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
125 nM
250 nM
500 nM
1 µM
(NP)
aurein 1.2 (E) citropin 1.3 (U)
(+)
Figure 3. Efficacy and toxicity of recombinant expression fusions of
aurein 1.2 ("E") and citropin 1.3 ("U"). (A) Cytosolic Cre delivery
results in recombination and tdTomato expression. The percentage of
tdTomato positive cells was determined by flow cytometry. Protein
fusions were delivered at 125 nM, 250 nM, 500 nM, and 1 µM. (B)
Toxicity of aurein 1.2 and citropin 1.3 as determined by CellTiterGlo
(Promega) assay. Protein fusions were delivered at 125 nM, 250 nM,
500 nM, and 1 µM. The labeled concentration of +36 GFP–Cre was
used as the no peptide control (NP), and addition of 100 µM chloro-
quine was used as the positive control (+). Cells were treated with 250
nM protein for 4 h in serum-free media. Cells were washed and sup-
planted with full DMEM and incubated for 48 h. Error bars represent
the standard deviation of three independent biological replicates.
An alanine scan was performed on aurein 1.2 to determine positions
that tolerate mutation. Charged amino acids at tolerant positions were
then replaced with histidines or other charged amino acids in an at-
tempt to increase endosomal escape efficiency. All constructs were
expressed as recombinant fusion proteins with +36 GFP–Cre.
To identify the residues involved in enhancing non-endosomal
protein delivery, we performed an alanine scan of the 13 amino acid
positions of aurein 1.2 by cloning, expressing, and purifying each
alanine mutant of aurein 1.2–+36 GFP–Cre. The resulting fusion
proteins were assayed in BSR.LNL.tdTomato reporter cells as de-
scribed above (Table 2). Seven positions were moderately to highly
intolerant of alanine substitution. Six positions retained >70% of
the activity of unmutated aurein 1.2–+36 GFP–Cre (Figure 4A).
At each of these tolerant positions, which included three positions
with charged residues (K7, K8, and E11 from Table 2), we generat-
ed additional mutations in an effort to improve activity. In total, 19
mutants of aurein 1.2 were generated and tested using the Cre re-
combination assay. Two of the aurein variants, K8R and S12A,
exhibited potentially improved overall recombination efficiency
but also increased toxicity at 250 nM (Figure 4B). Given this in-
crease in toxicity, we decided to focus on the original peptide, au-
rein 1.2, and proceeded to characterize its potency through a series
of complementary secondary assays.
Next, we evaluated the toxicity of each fusion protein at a range
of concentrations (125 nM to 1 µM) using an ATP-dependent cell
viability assay at 48 h after treatment. For +36 GFP–Cre, we ob-
served no cellular toxicity up to 1 µM treatment, which resulted in
85% viable cells. Cells treated with 250 nM recombinant aurein
1.2–+36 GFP–Cre and citropin 1.3–+36 GFP–Cre displayed 87%
and 84% viability, respectively. Applying 1 µM treatments de-
creased cell viability to 70% and 66%, respectively (Figure 3B). In
light of its activity and low cytotoxicity at 250 nM, we characterized
in depth the ability of aurein 1.2 to enhance cytosolic protein deliv-
ery.
Site-directed mutagenesis of aurein 1.2. Aurein 1.2
(GLFDIIKKIAESF) is a potent AMP excreted from the Australian
tree frog, Litoria aurea30. Interestingly, citropin 1.3
(GLFDIIKKVASVIGGL) is a closely related peptide and is excret-
ed from a different Australian tree frog, Litoria citropa31. While the
properties of aurein 1.2 have been investigated for its anti-bacterial
and anti-tumorogenic abilities30, its ability to enhance endosomal
escape or macromolecule delivery has not been previously report-
ed. The free peptide is thought to adopt an amphipathic alpha heli-
cal structure in solution, but the length of the helix is too short to
span a lipid bilayer32. Therefore it has been theorized that aurein
1.2 disrupts membranes through a "carpet mechanism” in which
peptides bind to the membrane surface in a manner that allows
hydrophobic residues to interact with lipid tails and hydrophilic
residues to interact with polar lipid head groups33. Above a critical
4
ACS Paragon Plus Environment

Page 5 of 12
Journal of the American Chemical Society
A
A
1
mCherry/DAPI
90
80
70
60
50
40
30
20
10
0
2
3
4
5
6
7
8
9
D
B
C
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
B
100
90
80
70
60
50
40
30
20
10
0
Figure 5. Investigating the ability of +36 GFP and aurein 1.2–+36
GFP dexamethasone-conjugates to reach the cytosol and activate GR
translocation. (A) Images of HeLa cells expressing GR-mCherry treat-
ed in the presence and absence of 1 µM dexamethasone (Dex)-protein
conjugates for 30 min at 37 °C. (B) Nuclear-to-cytosol GR-mCherry
fluorescence ratios (translocation ratios) of respective Dex-protein
conjugates determined using CellProfiler®. (C) GR-mCherry translo-
cation ratios resulting from cells treated in the presence and absence of
+36 GFPDex and endocytic inhibitors. (D) GR-mCherry translocation
ratios resulting from cells treated in the presence and absence of aurein
1.2–+36 GFPDex and endocytic inhibitors. Statistical significance is
measured by P-value. ns= P > 0.05, * = P ≤ 0.05, ** = P ≤ 0.01, *** = P ≤
0.001.
Figure 4. Activity and cytotoxicity of aurein 1.2 variants fused to +36
GFP–Cre(A) The percentage of tdTomato positive cells was deter-
mined by flow cytometry. (B) Toxicity as determined by CellTiterGlo
(Promega) assay. For (b) and (c), 250 nM+36 GFP–Cre was used as
the no peptide control (NP), and addition of 100 µM chloroquine was
used as the positive control (+). Cells were treated with 250 nM pro-
tein for 4 h in serum-free DMEM. Cells were washed and supplanted
with full DMEM and incubated for 48 h.
We generated dexamethasone conjugates of +36 GFP (+36
GFPDex) and aurein 1.2–+36 GFP (aurein 1.2–+36 GFPDex) via
sortase-mediated conjugation (Supporting Information Figure S4).
Conjugated to these proteins, SDex is no longer cell permeable and
cannot activate the GR for nuclear translocation unless the protein-
SDex conjugate can access the cytosol. Treatment of HeLa cells
expressing GR-mCherry with 1 µM aurein 1.2–+36 GFPDex for 30
min resulted in a TR of 2.62, which was significantly greater (p <
0.05) than that of +36 GFPDex (TR = 2.23). As positive controls, we
treated these cells with canonical cell permeable peptides (TatDex
and Arg8Dex) and miniature proteins containing a penta-Arg motif
that reach the cytosol intact, with efficiencies exceeding 50% (5.3Dex
and ZF 5.3Dex)37. Aurein 1.2-+36 GFPDex (TR=2.62), activated sig-
nificantly greater levels of GR-mCherry translocation (p < 0.001)
than TatDex (TR = 1.87) and Arg8Dex (TR = 1.63) and similar levels
evoked by miniature proteins 5.3Dex (TR = 2.62) and ZF 5.3Dex (TR
= 2.38) (Figure 5A and 5B). Taken together, these results suggest
that aurein 1.2–+36 GFPDex exhibits an improved ability to access
the cytoplasm over +36 GFPDex and canonical cell permeable pep-
tides.
Independent assays of endosomal escape. Although endo-
somal escape is widely considered to be the major bottleneck of
cationic protein delivery34, few assays quantify the ability of pro-
teins to escape endosomes on a single-cell basis. To quantify cyto-
solic delivery of supercharged proteins in individual cells, we ap-
plied a glucocorticoid receptor (GR) translocationassay35 de-
scribed by Schepartz and colleagues9,36. In untreated HeLa cells
expressing mCherry-labeled GR (GR-mCherry), the GR distrib-
utes nearly uniformly throughout the cell interior, resulting in a
nuclear-to-cytoplasm translocation ratio (TR) of 1.17 (Figure 5A
and 5B). Upon treatment with the cell-permeable glucocorticoid
dexamethasone-21-thiopropionic acid (SDex) at a concentration of
1 µM for 30 min, GR-mCherry relocates almost exclusively to the
nucleus, resulting in a TR of 3.77 (Figure 5A and 5B).
As an additional, independent assay of non-endosomal protein
delivery, we tested the ability of aurein 1.2 to enhance the non-
endosomal delivery of an evolved biotin ligase (BirA) enzyme de-
veloped by Ting and coworkers38. BirA catalyzes the biotinylation
of a 15-amino acid acceptor peptide (AP). We transfected a
mCherry-AP fusion plasmid into HeLa cells. Biotinylation of
mCherry can only occur in the presence of cytosolic BirA. To as-
sess the non-endosomal delivery of +36 GFP–BirA protein,
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 12
mCherry-AP biotinylation was quantified by Western blot using Aurein 1.2 can greatly increase protein delivery efficiency
1
2
3
4
5
6
7
8
fluorophore-labeled streptavidin and normalized to the mCherry
signal (Supporting Information Figure S5A). Treatment with 250
nM aurein 1.2–+36 GFP–BirA resulted in a 50% increase in bioti-
nylation signal compared with 250 nM of +36 GFP–BirA alone
(Supporting Information Figure S5B). We also observed a dose-
dependent increase in AP-biotinylation across treatment concen-
trations (250 nM, 500 nM, and 1 µM) for both aurein 1.2–(+36
GFP)–BirA and unfused +36 GFP–BirA constructs. These results
are consistent with the results of the GR translocation assay, and
further suggest that aurein 1.2 enhances the endosomal escape of
superpositively charged proteins.
in vivo. To evaluate the ability of aurein 1.2 to increase the efficacy
of cationic protein delivery in vivo, we delivered proteins to the
inner ear of Cre reporter transgenic mice that express tdTomato
upon Cre-mediated recombination. This animal model was chosen
due to its confined injection volume, the presence of well-
characterized cell types, and the existence of genetic deafness mod-
els that would facilitate future studies of protein delivery to treat
hearing loss. We previously showed that +36 GFP–Cre alone can
be delivered to mouse retina,15 albeit resulting in only modest levels
of recombination consistent with inefficient endosomal escape.
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
Anesthetized postnatal day 2 (P2) mice were injected with 0.4
µL of 50 µM +36 GFP–Cre or aurein 1.2–+36 GFP–Cre solutions
in the scala media to access the cochlear cells. Five days after injec-
tion, the cochleas were harvested for immunolabeling of inner ear
cell markers and imaging for tdTomato florescence (Figure 6A).
We evaluated both the hair cells (Myo7a+) and supporting cells
(Myo7a-) for td Tomato signal. The total number of hair cells and
supporting cells (by DAPI labeling) in the sensory epithelium (SE)
was used to determine the relative toxicity of aurein 1.2–+36 GFP–
Cre to the baseline treatment of +36 GFP–Cre (Figure 6A). Over-
all, an average of 96%, 92% and 66% of cochlear cells survived au-
rein 1.2–+36 GFP–Cre treatment as compared to +36 GFP–Cre
treatment in the apex, middle, and base tissue samples, respectively
(Figure 6A). +36 GFP–Cre treatment resulted in low levels of re-
combination only in inner hair cells (IHC) of the apex of the coch-
lea (4.4%) but not in the middle or base of the cochlear hair cells or
any cochlear supporting cells. In contrast, treatment with aurein
1.2–+36 GFP–Cre resulted in very high Cre-mediated recombina-
tion levels throughout the apex, middle, and base samples of outer
hair cells (OHC) (96%, 91%, and 69%, respectively), inner hair
cells (100%, 94%, and 70%, respectively), as well as supporting cells
(arrows) (Figure 6A and 6C).
In order to directly quantify the increase in non-endosomal de-
livery resulting from aurein 1.2, we performed a cytosolic fractiona-
tion experiment to calculate the cytosolic concentrations of deliv-
ered protein. HeLa cells were treated with +36 GFP or aurein 1.2–
+36 GFP at 250 nM, 500 nM, and 1 µM. After 30 min of treatment,
cells were washed, homogenized, and fractionated by ultracentrifu-
gation. The cytosolic concentration of delivered protein was calcu-
lated from the GFP fluorescence of the cytosolic fraction together
with a standard curve relating fluorescence to known concentra-
tions of +36 GFP and aurein 1.2–+36 GFP added to cytosolic ex-
tract (Supporting Information Figures S6B and 6C). At 250 nM,
treatment with aurein 1.2–+36 GFP resulted in ~5-fold more deliv-
ered cytosolic protein than treatment with +36 GFP alone (Sup-
porting Information Figure S6A). This difference decreased with
increasing protein concentration, likely due to the influence of
alternate uptake pathways or delivery bottlenecks at high protein
concentrations. In contrast, the total amount of aurein 1.2–+36
GFP versus +36 GFP uptaken by cells was similar, with aurein 1.2–
+36 GFP showing 1.3-fold higher total cellular uptake at 250 nM.
These results directly demonstrate that aurein 1.2 increases the
cytosolic concentration of cationic proteins that enter cells pre-
dominantly through endosomes,14,18 and are consistent with the
above findings that aurein 1.2 has the greatest effect on enhancing
non-endosomal delivery at ~250 nM (Figure 3A).
Effect of endocytic inhibitors on +36 GFP and aurein 1.2–
+36 GFP delivery. We previously reported that endocytosis plays
a key role in the cytosolic delivery of superpositively charged pro-
teins18. To probe the role of endocytosis in the delivery of super-
charged proteins with or without aurein 1.2, we treated cells ex-
pressing GR-mCherry with either +36 GFPDex or aurein 1.2–+36
GFPDex in the presence of known endocytic inhibitors. The cortical
actin remodeling inhibitor N-ethyl-isopropyl amiloride (EIPA), the
cholesterol-sequestering agent methyl-β-cyclodextrin (MBCD),
and the endosomal vesicular ATPase inhibitor bafilomycin (Baf) all
strongly reduced the ability of both proteins to stimulate GR-
mCherry translocation. Blocking maturation of Rab5+ vesicles by
treatment with the phosphatidylinositol 3-kinase inhibitor wort-
mannin (Wort) did not influence reporter translocation of either
+36 GFPDex or aurein 1.2–+36 GFPDex (Figure 5C and 5D). In con-
trast, treatment with the small-molecule dynamin II inhibitor
Dynasore (Dyna) significantly suppressed the ability of +36 GFPDex
to stimulate GR-mCherry translocation (TR = 1.64) (Figure 5C)
but had little influence on the cytosolic delivery of aurein 1.2-+36
GFPDex (TR = 2.30) (Figure 5D). Taken together, these results
suggest that active endocytosis is required for uptake of +36 GFP
and aurein 1.2–+36GFP into the cell interior, and that the two
proteins may traffic differently into the cell interior.
6
ACS Paragon Plus Environment

Page 7 of 12
Journal of the American Chemical Society
some peptides can selectively disrupt the endosomal membrane
Apex
Middle
Base
Myo7a/tdTomato/DAPI
A
+36'GFP–Cre''
1
2
3
4
5
6
7
8
without disrupting the mammalian cell membrane. The effective-
ness of aurein 1.2 is highly dependent on its sequence, as several
other closely related peptides did not enhance protein delivery
(Tables 1 and 2). Subtle differences in amino acid composition led
to dramatic changes in membrane activity among peptides tested,
highlighting the difficulty of rationally designing peptides that en-
hance non-endosomal delivery. Moreover, the lack of correspond-
ence between peptide cationic charge and non-endosomal delivery
efficiency suggests that aurein 1.2 does not enhance non-
endosomal delivery simply by promoting endocytosis. While none
of the tested variants of aurein 1.2 substantially outperformed the
original peptide, we identified several amino acids that could be
altered without loss of activity. These findings also provide a start-
ing point for further optimization to discover next-generation en-
dosomolytic peptides with improved efficacy and reduced toxicity.
OHC'
IHC'
SE'
aurein'1.2–+36'GFP–Cre''
B
9
70
60
50
40
30
20
10
0
25
20
15
10
5
OHC
IHC
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
0
Apex
Middle
Base
Apex
Middle
Base
aurein 1.2–+36 GFP–Cre
+36 GFP–Cre
aurein 1.2–+36 GFP–Cre
+36 GFP–Cre
Four independent assays for non-endosomal protein delivery
(Cre recombination, GR translocation, BirA activity on a cyto-
plasmic peptide, and cytosolic fractionation), together with the
peptide mutational studies described above, collectively suggest
that aurein 1.2-fusion enhances endosomal escape of superpositive-
ly charged proteins. Moreover, these assays collectively demon-
strated the ability of aurein 1.2 to mediate the non-endosomal de-
livery of +36 GFP fused to different proteins (or small molecules),
suggesting that aurein 1.2 facilities endosomal escape in a manner
that is at least somewhat cargo-independent.
C
120
100
80
60
40
20
0
120
100
80
60
40
20
0
OHC
IHC
Apex
Middle
Base
Apex
Middle
Base
aurein 1.2–+36 GFP–Cre
+36 GFP–Cre
aurein 1.2–+36 GFP–Cre
+36 GFP–Cre
The in vivo protein delivery experiments described above re-
vealed dramatic increases in non-endosomal functional Cre re-
combinase delivery into the diverse inner ear cell types including
hair cells and supporting cells of live mice upon fusion with aurein
1.2. Indeed, aurein 1.2-fused +36 GFP–Cre construct resulted in
highly efficient recombination levels across the main cochlear sen-
sory epithelial cell classes studied in this work, all but one of which
were unaffected by +36 GFP–Cre treatment. Taken together, these
results suggest that aurein 1.2 is a 13-residue, potent, genetically
encodable, endosome escape-enhancing peptide that can greatly
increase the efficiency of non-endosomal cationic protein delivery
in vitro and in vivo without requiring the use of additional compo-
nents beyond the protein of interest.
Figure 6. In vivo protein delivery of Cre recombinase into mouse neo-
natal cochleas. 0.4 µL of 50 µM +36 GFP–Cre or aurein 1.2–+36 GFP–
Cre were injected into the scala media. (A) Five days after injection,
cochlea were harvested. Inner hair cells (IHC), outer hair cells (OHC)
and supporting cells in the sensory epithelium (SE) were imaged for
the presence of tdTomato, which is only expressed following Cre-
mediated recombination. Hair cells were labeled with antibodies
against the hair-cell marker Myo7a. Gray/white = Myo7a, Red =
tdTomato, Blue = DAPI. (B) To evaluate cytotoxicity, the number of
outer hair cells and inner hair cells were measured by counting DAPI-
stained cells (C) The percentage of tdTomato positive cells, reflecting
successful delivery of functional Cre recombinase, was determined by
fluorescence imaging.
The observed levels of recombination in the inner hair cells
from aurein 1.2–+36 GFP–Cre are comparable to that of adeno-
associated virus type 1 (AAV1) gene transfection39. For outer hair
cells, we have previously shown similar levels of recombination
using liposome-mediated delivery of supernegatively-charged
GFP–Cre40. The aurein 1.2–+36 GFP–Cre delivery system is the
only method that showed significant recombination levels in both
inner and outer hair cells39,41, and does not require any virus or
other molecules beyond a single polypeptide. Significantly, aurein
1.2–+36 GFP–Cre also extended delivered recombinase activity to
additional cochlear supporting cells. These results suggest aurein
1.2-+36 GFP-Cre delivery system to be a promising method for in
vivo protein delivery into both hair cells and supporting cells of the
inner ear42.
EXPERIMENTAL SECTION
Construction of expression plasmids. Sequences of all con-
structs used in this paper are listed in the Supporting Information.
All protein constructs were generated from previously reported
plasmids for protein of interest cloned into a pET29a expression
plasmid43. All plasmid constructs generated in this work will be
deposited with Addgene.
Expression and purification of proteins. E. coli BL21 STAR
(DE3) competent cells (Life Technologies) were transformed with
pET29a expression plasmids. Colonies from the resulting expres-
sion strain was directly inoculated in 1 L of Luria-Bertani (LB)
broth containing 100 μg/mL of ampicillin at 37 °C to OD600 = ~1.0.
Isopropyl β-D-1- thiogalactopyranoside (IPTG) was added at 0.5
mM to induce expression and the culture was moved to 20 °C.
After ~16 h, the cells were collected by centrifugation at 6,000 g
and resuspended in lysis buffer (Phosphate buffered saline (PBS)
with 1 M NaCl). The cells were lysed by sonication (1 sec pulse-on,
1 sec pulse-off for 6 min, twice, at 6 W output) and the soluble
lysate was obtained by centrifugation at 10,000 g for 30 min.
CONCLUSION
We discovered a 13-residue peptide, aurein 1.2, that can in-
crease the efficiency of non-endosomal protein delivery by screen-
ing a panel of known membrane-active peptides. The results from a
small screen of 22 peptides are consistent with our hypothesis that
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
The cell lysate was incubated with His-Pur nickel-nitriloacetic added to 800 mL of ice water and Dexamethasone-21-Mesylate
Page 8 of 12
1
2
3
4
5
6
7
8
acid (Ni-NTA) resin (Thermo Scientific) at 4 °C for 45 min to
capture His-tagged protein. The resin was transferred to a 20-mL
column and washed with 20 column volumes of lysis buffer plus 50
mM imidazole. Protein was eluted in lysis buffer with 500 mM
imidazole, and concentrated by Amicon ultra centrifugal filter (Mil-
lipore, 30-kDa molecular weight cut-off) to ~50 mg/mL. The elu-
ent was injected into a 1 mL HiTrap SP HP column (GE
Healthcare) after dilution into PBS (5-fold). Protein was eluted
with PBS containing a linear NaCl gradient from 0.1 M to 1 M over
five column volumes. The eluted fractions containing protein were
concentrated to 50 µM as quantified by absorbance at 488 nm as-
suming an extinction coefficient of 8.33 x 104 M-1cm-1 as previously
determined 14, snap-frozen in liquid nitrogen, and stored in aliquots
at -80 °C.
(Dex-21-OMs) formed a white precipitate. The slurry was filtered
and the precipitate washed with 800 mL of ice water, dried under
high vacuum overnight and quantified by LC-MS (m/z 471.19 Da,
83% yield).
Dexamethasone-21-thiopopionic acid (SDex) was prepared as
previously described.46 2.05 g of Dex-21-OMs was added to 2 eq.
thiopropionic acid and 4 eq. triethylamine stirring in anhydrous
acetone at room temperature overnight. The following morning,
the reaction was added to 800 mL of ice water and acidified with 1
N HCl until SDex, visible as an off-white solid, precipitation was
complete. The mixture was filtered, washed with 800 mL ice cold
water acidified to pH 1 with HCl, dried under high vacuum over-
night and analyzed by LC-MS (m/z 481.21 Da, 63% yield) (Sup-
porting Figure 7).
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
Synthesis and Purification of GGGKDex. GGGKDex was syn-
thesized on Fmoc-Lys (Mtt)-Wang resin (BACHEM, D-2565)
using microwave acceleratin (MARS, CEM). Coupling reactions
were performed using 5 equivalents of Fmoc-Gly-OH (Novabio-
chem, 29022-11-5), 5 equivalents of PyClock (Novabiochem,
893413-42-8) and 10 equivalents of diisopropylethylamine
(DIEA) in N-methylpyrrolidone (NMP). Fmoc groups were re-
moved using 25% piperidine in NMP (efficiency quantified;
ε299=6234 M-1cm-1 in acetonitrile) and Mtt groups were removed by
incubating the Fmoc-GGGK(Mtt)-resin with 2% trifluoroacetic
acid (TFA) in dichloromethane (DCM) for 20 min, after which the
resin was washed with 2% TFA in DCM until the characteristic
yellow color emitting from the Mttcation subsided. After Mtt re-
moval, SDex-COOH (Dex-21-thiopropinonic acid35) was coupled
to the Nε of the lysine side-chain by incubating the Fmoc-GGGK-
resin with 2.5 eq. SDex-COOH, 2.5 eq. HATU, 2.5 eq. HOAt, 5 eq.
DIEA and 5 eq. 2,6-lutidine in 2.5 mL NMP overnight, at room
temperature, on an orbital shaker. After SDex-labeling, Fmoc-
GGGKDex-resin was washed thoroughly with NMP and DCM, the
N-terminal Fmoc was removed using 25% piperidine in NMP, and
crude peptides were dissociated from the resin by incubating the
GGGKDex-resin in a cleavage cocktail composed of 81.5% trifluoro-
acetic acid (TFA), 5% thioanisole, 5% phenol, 5% water, 2.5%
ethanedithiol and 1% triisopropylsilane for 30 min at 38 °C. Crude
peptides were precipitated in 40 mL cold diethyl ether, resuspend-
ed in water, lyophilized and purified via reverse phase high-pressure
liquid chromatography (HPLC) using a linear gradient of acetoni-
trile and water with 0.1% TFA across a C18 (VYDAC, 250mm x 10
mm ID) column. Purified peptides were lyophilized and stored at
4 °C. Polypeptide identity was confirmed by mass spectrometry on
a Waters QToF LC-MS, and purity was measured by analytical
reverse-phase HPLC (Shimadzu Instruments) using a C18 column
(Poroshell 120 SB-C18, 2.7 µm, 100 mm x 3 mm ID, Agilent).
Cell Culture. All cells were cultured in Dulbecco's modifica-
tion of Eagle's medium (DMEM w/glutamine, Gibco) with 10%
fetal bovine serum (FBS, Gibco), 5 I.U. penicillin, and 5 g/mL
streptamycin. All cells were cultured at 37 °C with 5% CO2.
Peptide synthesis. Peptides were ordered from ChinaPeptides
Co., LTD, each 4 mg, purity > 90%. HPLC and MALDI data were
provided with lyophilized peptides. Peptides were resuspeneded in
DMSO to a final concentration of 10 mM.
Sortase conjugation. All reactions were performed in 100mM
Tris buffer (pH 7.5) with 5mM CaCl2 and 1M NaCl. For peptide
conjugation to the N-terminus of GGG–+36–GFP , 20 µM of pro-
tein with N-terminal Gly-Gly-Gly was incubated with 400 µM of
peptide with C-terminal LPETGG and 1 µM eSrtA for 2 h at room
temperature in a 50 µL reaction. The unreacted peptides were re-
moved through spin filtration with an Amicon Ultra-0.5 Centrifu-
gal Filter Unit (Millipore, 30-kDa molecular weight cut-off). The
reaction mixture was washed twice with 500 µL of buffer each time
to a final concentration of 50 µL. Conjugation efficiency was de-
termined through LC-MS (Agilent 6220 ESI-TOF) using protein
deconvolution through MaxEnt (Waters) by comparing relative
peak intensities.
For conjugation of GGGKDex to +36–GFP–LPETG–His6, 10
µM of protein was incubated with 400 µM of peptide and 2 µM
eSrtA at room temperature. The reaction was quenched with 10
mM ethylenediaminetetraacetic acid (EDTA) after 2 h. For aurein
1.2–+36–GFP–LPETG–His6, a N-terminal His6–ENLYFQ was
added to prevent sortase reaction with the N-terminal glycine of
aurein 1.2. The N-terminal tag was removed with 200µM TEV
protease at 4 °C for 16 h to release the native N-terminal sequence
of aurein 1.2–+36–GFP. Successful conjugation of GGGKDex re-
moves the C-terminal His6 tag and allows for purification through
reverse Ni-NTA column. Unreacted protein binds to the Ni-NTA,
and the unbound protein was collected and concentrated as de-
scribed above.
Image processing for primary screen. BSR.LNL.tdTomato
cells were plated at 10,000 cells per well in black 384-well plates
(Aurora Biotechnologies). Cells were treated with Cre fusion pro-
teins diluted in serum-free DMEM 24 h after plating and incubated
for 4 h at 37 °C. Following incubation, the cells were washed three
times with PBS + 20 U/mL heparin. The cells were incubated a
further 48 h in serum-containing media. Cells were fixed in 3%
paraformaldehyde and stained with Hoescht 33342 nuclear dye.
Images were acquired on an ImageXpress Micro automated micro-
scope (Molecular Devices) using a 4× objective (binning 2, gain 2),
with laser- and image-based focusing (offset −130 µm, range 50
Plasmid transfection. Plasmid DNA was transfected using
Lipofectamine 2000 (Life Technologies) according the manufac-
turer’s protocol.
Synthesis of Dexamethasone-21 Thiopropionic Acid
(SDex). Synthesis of dexamethasone-21-mesylate was performed
as previously described.44,45 2 g of dexamethasone stirring in 38 mL
anhydrous pyridine under nitrogen was reacted with 467.2 m
gmethanesulfonyl chloride (1.2 eq.) on ice for 1h, after which an-
other 311 methanesulfonmethanesulfonyl chloride was added and
allowed to react overnight (16h) on ice. Next, the reaction was
8
ACS Paragon Plus Environment

Page 9 of 12
Journal of the American Chemical Society
µm, step 25 µm). Images were exposed for 10 ms in the DAPI 250 nM, 500 nM and 1 µM concentrations. 250 nM +36 GFP–BirA
1
2
3
4
5
6
7
8
channel (Hoechst) and 500 ms in the dsRed channel (tdTomato).
Image analysis was performed using the cell-scoring module of
MetaXpress software (Molecular Devices). All nuclei were detect-
ed with a minimum width of 1 pixel, maximum width of 3 pixels,
and an intensity of 200 gray levels above background. Positive cells
were evaluated for uniform signal in the dsRed channel (minimum
width of 5 pixels, maximum width of 30 pixels, intensity > 200 gray
levels above background, 10 µm minimum stained area). In total,
nine images were captured and analyzed per well, and 16 wells were
treated with the same fusion protein. The primary screen was com-
pleted in biological triplicate.
+ 100 µM chloroquine was also used as a positive control for endo-
somal escape. The cells were washed three times with PBS contain-
ing heparin to remove excess supercharged proteins that were not
internalized. The cells were then treated with 100 µL of 10 µM
biotin and 1 mM ATP in PBS for 10 min. The reaction was
quenched with excess (10 µL of 8 mM) synthesized AP before cells
were trypsinized and lysed. To verify that extracellular BirA was not
generating signal during lysis, 1 µM +36 GFP–BirA or aurein 1.2–
+36 GFP–BirA was added during the quench step to untreated
wells. Cells were lysed with 100 µL of trypsin and lysed with QI-
Ashredder columns (Qiagen). 30 µL of lysate was loaded onto 4-
12% Bis-Tris Bolt gels in Bolt-MES buffer (Life Technologies) and
ran for 20 min at 200 volts. Gels were transferred to PVDF mem-
brane using iBlot2 transfer system (Life Technologies). Biotinyla-
tion was measured through western blotting using the LI-COR
quantitative infrared fluorescent antibodies and the Odyssey Im-
ager detection system. To normalize for transfection and gel load-
ing variables, the ratio of biotin signal to mCherry signal was used
for comparison.
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
Cre delivery assay. Uptake and delivery assays for Cre fusion
proteins were performed as previously described 15. Briefly, proteins
were diluted in serum-free DMEM and incubated on the cells in
48-well plates for 4 h at 37 °C. Following incubation, the cells were
washed three times with PBS + 20 U/mL heparin. The cells were
incubated a further 48 h in serum-containing media prior to tryp-
sinization and analysis by flow cytometry. All flow cytometry were
carried out on a BD Fortessa flow cytometer (Becton-Dickinson)
using 530/30 nm and 610/20 nm filter sets. Toxicity for aurein 1.2
and citropin 1.3 validation assays was determined using CellTiter-
Glo assay (Promega) in 96-well plates following manufacturer
protocol. Toxicity for alanine scan mutational analysis was deter-
mined with LIVE/DEAD fixable far-red dead cell stain (Life Tech-
nologies) with 635 nm laser and 670/30 nm filter.
Cytosolic fractionation assay. One day prior to fractionation,
4x106 HeLa cells were plated in 20 mL of DMEM (10% FBS, 1x
PenStrep, no phenol red) in 175-cm2 culture flasks and allowed to
adhere for 15 hours. The following day, the media was removed
from each flask and the cells were washed twice with clear DMEM
(no FBS, no PenStrep, no phenol red). The media was replaced
with 7 mL of clear DMEM containing +36 GFP or aurein 1.2–+36
GFP at a concentration of 250 nM, 500 nM, or 1 μM. Several flasks
were treated with clear DMEM to be used as negative controls and
to generate calibration curves with the cytosolic extracts. The cells
were incubated for 30 min at 37 °C, 5% CO2 after which they were
washed three times with PBS. Using a cell-scraper, the cells were
suspended in 5 mL of PBS, transferred into a 15 mL Falcon tube,
and pelleted at 500 g for 3 min. The cells were resuspended in 1 mL
PBS, counted using an automated cell counter (Auto T4, Cellome-
ter®), and pelleted again at 500 g for 3 min. The cell pellet was re-
suspended in ice-cold isotonic sucrose (290 mM sucrose, 10 mM
imidazole, pH 7.0 with 1 mM DTT, and cOmplete™, EDTA-free
protease inhibitor cocktail) and transferred to a glass test tube on
ice. The cells were homogenized with an Omni TH homogenizer
outfitted with a stainless steel 5 mm probe for three 30 s pulses on
ice with 30 s pauses between the pulses. The homogenized cell
lysate was sedimented at 350 Kg in an ultracentrifuge (TL-100;
Beckman Coulter) for 30 min at 4 °C using a TLA 120.2 rotor. The
supernatant (cytosolic fraction) was analyzed in a 96-well plate on
a fluorescence plate reader (Synergy 2, BioTek, excitation = 485
+/- 10 nm, emission = 528 +/- 10 nm). The concentration of the
protein conjugate in the cytosol was determined using a standard
curve relating fluorescence to known protein concentrations. To
generate the standard curve, known concentrations of +36 GFP
and aurein 1.2–+36 GFP between 0.2 nM and 1 μM were added to
cytosolic extracts of the untreated negative controls. For back-
ground subtraction, several wells containing cytosolic extracts from
untreated cells were averaged, and this average was subtracted from
each well.
GR-mCherry translocation assay. One day prior to transfec-
tion 10,000 HeLa cells in 200 µL of DMEM (10% FBS, 1x Pen-
Strep) were plated into single wells of a 96-well MatriCal glass
bottom microplate (MGB096-1-2-LG-L) and allowed to adhere
overnight. The following day, cells were transfected with GR-
mCherry9 using Lipofectamine® 2000 technologies. Following
transfection, cells were allowed to recover overnight in DMEM (+
10% FBS). The following day, cells were treated with dexame-
thasone (Dex) or 1 µM Dex-protein conjugate in the presence or
absence of inhibitor diluted into DMEM (without phenol red,
+300 nM Hoescht33342). After 30 min, cells were washed twice
with 200 µL of HEPES−Krebs−Ringer’s (HKR) buffer (140 mM
NaCl, 2 mM KCl, 1 mM CaCl2, 1 mM MgCl2, and 10 mM HEPES
at pH 7.4), after which 100 µL of HKR buffer was overlaid onto the
cells and images were acquired on a Zeiss Axiovert 200M epifluo-
rescence microscope outfitted with Ziess AxiocammRM camera
and an EXFO-Excite series 120 Hg arc lamp. The translocation
ratio (the ratio of median GFP intensity in the nuclear and sur-
rounding regions) for individual cells was measured using CellPro-
filer® as described36. To examine the effect of endocytosis inhibi-
tors, HeLa cells were pretreated for 30 min with DMEM (without
phenol red) containing inhibitors (80 µM Dynasore, 5 mM
MBCD, 50 µM EIPA, 200 nM bafilomycin or 200 nM wortman-
nin) at 37 °C for 30 min before incubation with Dex or Dex-protein
conjugates.
BirA translocation assay. One day prior to transfection,
100,000 HeLa cells in 1 mL of DMEM (10% FBS, 1x PenStrep)
were plated into single wells of a 12-well tissue culture plate and
allowed to adhere overnight. Cells were transfected with mCherry-
AP fusion protein using Lipofectamine® 2000 technologies accord-
ing to manufacture guidelines24 h before protein treatment. Next
day, transfected cells were treated for 1 h at 37 °C with +36 GFP–
BirA or aurein 1.2–+36 GFP–BirA diluted in serum-free DMEM at
Total protein delivery assay. One day prior to the experi-
ment, 100,000 HeLa cells/well were plated in DMEM (10% FBS,
1x PenStrep, no phenol red) in 6-well plates and allowed to adhere
for 15 hours. The following day, the media was removed from each
well and the cells were washed twice with clear DMEM (no FBS,
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 12
(1)
Nelson, A. L.; Dhimolea, E.; Reichert, J. M. Nat Rev
no PenStrep, no phenol red). The media was replaced with 1 mL of
clear DMEM containing +36 GFP or aurein 1.2–+36 GFP at con-
centrations of 250 nM, 500 nM, or 1 μM. The cells were incubated
for 30 min at 37 °C, 5% CO2 after which they were washed three
times with PBS containing 20 U/mL heparin (Sigma) to remove
surface-bound cationic protein. The cells were trypsinized for 5
min, pelleted in serum-containing DMEM for 3 min at 500 g,
washed with 1 mL PBS, and pelleted again for 3 min at 500 g. The
cell pellet was resuspended in 100 μL PBS. Flow cytometry was
performed on a BD Accuri C6 Flow Cytometer at 25 °C. Cells were
analyzed in PBS (excitation laser = 488 nm, emission filter = 533
+/- 30 nm). At least 10,000 cells were analyzed for each sample.
For background subtraction, wells were treated with clear DMEM
only. The average of three untreated wells was subtracted from
each +36 GFP conjugate-containing well.
Drug Discov 2010, 9, 767.
1
2
3
4
5
6
7
8
(2)
Huang, C. Current opinion in biotechnology 2009, 20,
692.
(3)
(4)
Hafler, D. A. Nat Rev Immunol 2007, 7, 423.
Leader, B.; Baca, Q. J.; Golan, D. E. Nat Rev Drug Discov
2008, 7, 21.
(5)
Schiffmann, R.; Kopp, J. B.; Austin, H. A.; Sabins, S.;
Moore, D. R.; Weibel, T.; Balow, J. E.; Brady, R. O. JAMA 2001, 285, 2743.
(6)
(7)
Spiegelman, B. M. Diabetes 1998, 47, 507.
Mali, P.; Esvelt, K. M.; Church, G. M. Nat Meth 2013,
9
10, 957.
(8)
Mueller, J.; Kretzschmar, I.; Volkmer, R.; Boisguerin, P.
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
Bioconjugate Chemistry 2008, 19, 2363.
(9)
Appelbaum, Jacob S.; LaRochelle, Jonathan R.; Smith,
Betsy A.; Balkin, Daniel M.; Holub, Justin M.; Schepartz, A. Chemistry & Biology
2012, 19, 819.
(10)
Rizk, S. S.; Luchniak, A.; Uysal, S.; Brawley, C. M.;
Rock, R. S.; Kossiakoff, A. A. Proceedings of the National Academy of Sciences
2009, 106, 11011.
Microinjection of proteins to mouse inner ear. P1-2
Gt(ROSA)26Sortm14(CAG-tdTomato)Hze mice were used for aurein 1.2–
+36–GFP–Cre and +36–GFP–Cre injection. The Rosa26-
tdTomato mice were from the Jackson Laboratory. Animals were
used under protocols approved by the Massachusetts Eye &Ear
Infirmary IACUC committee. Mice were anesthetized by hypo-
thermia on ice. Cochleostomies were performed by making an
incision behind the ear to expose the cochlea. Glass micropipettes
held by a micromanipulator were used to deliver the complex into
the scala media, which allows access to inner ear hair cells. The
total delivery volume for every injection was 0.4 µL per cochlea and
the release was controlled by a micromanipulator at the speed of 69
nL/min.
(11)
Hasadsri, L.; Kreuter, J.; Hattori, H.; Iwasaki, T.;
George, J. M. Journal of Biological Chemistry 2009, 284, 6972.
(12)
Fu, A.; Tang, R.; Hardie, J.; Farkas, M. E.; Rotello, V. M.
Bioconjugate Chemistry 2014, 25, 1602.
(13)
Pisal, D. S.; Kosloski, M. P.; Balu-Iyer, S. V. Journal of
pharmaceutical sciences 2010, 99, 2557.
(14)
McNaughton, B. R.; Cronican, J. J.; Thompson, D. B.;
Liu, D. R. Proceedings of the National Academy of Sciences 2009, 106, 6111.
(15)
Cronican, J. J.; Thompson, D. B.; Beier, K. T.;
McNaughton, B. R.; Cepko, C. L.; Liu, D. R. ACS Chemical Biology 2010, 5,
747.
(16)
Cronican, James J.; Beier, Kevin T.; Davis, Tina N.;
Tseng, J.-C.; Li, W.; Thompson, David B.; Shih, Allen F.; May, Erin M.; Cepko,
Constance L.; Kung, Andrew L.; Zhou, Q.; Liu, David R. Chemistry & Biology
2011, 18, 833.
Immunohistochemistry and quantification. 5 days after in-
jection, the mice were sacrificed and cochlea were harvested by
standard protocols47. For immunohistochemistry, antibodies
against hair-cell markers (Myo7a) and supporting cells (Sox2)
were used following a previously described protocol47. To quantify
the number of tdTomato positive cells after aurein 1.2–+36–GFP–
Cre and +36–GFP–Cre, we counted the total number of inner and
outer hair cells in a region spanning 100 µm in the apex, middle,
and base turn of the cochlea.
(17)
Lawrence, M. S.; Phillips, K. J.; Liu, D. R. Journal of the
American Chemical Society 2007, 129, 10110.
(18)
Thompson, David B.; Villaseñor, R.; Dorr, Brent M.;
Zerial, M.; Liu, David R. Chemistry & Biology 2012, 19, 831.
(19)
(20)
Fuchs, S. M.; Raines, R. T. ACS Chemical Biology 2007,
2, 167.
Pirie, C. M.; Hackel, B. J.; Rosenblum, M. G.; Wittrup,
K. D. Journal of Biological Chemistry 2011, 286, 4165.
(21)
Varkouhi, A. K.; Scholte, M.; Storm, G.; Haisma, H. J.
Journal of Controlled Release 2011, 151, 220.
(22)
Wadia, J. S.; Stan, R. V.; Dowdy, S. F. Nat Med 2004,
ASSOCIATED CONTENT
10, 310.
(23)
Neundorf, I.; Rennert, R.; Hoyer, J.; Schramm, F.;
Supporting Information. Additional information and experimental
methods. List of all protein constructs included. This material is availa-
ble free of charge via the Internet at http://pubs.acs.org.
Löbner, K.; Kitanovic, I.; Wölfl, S. Pharmaceuticals 2009, 2, 49.
(24)
Sugita, T.; Yoshikawa, T.; Mukai, Y.; Yamanada, N.;
Imai, S.; Nagano, K.; Yoshida, Y.; Shibata, H.; Yoshioka, Y.; Nakagawa, S. British
journal of pharmacology 2008, 153, 1143.
AUTHOR INFORMATION
(25)
(26)
Zasloff, M. Nature 2002, 415, 389.
Lohner, K.; Blondelle, S. Combinatorial chemistry & high
Corresponding Author
throughput screening 2005, 8, 241.
(27)
Wang, Z.; Wang, G. Nucleic acids research 2004, 32,
*drliu@fas.harvard.edu
D590.
(28)
Chen, I.; Dorr, B. M.; Liu, D. R. Proceedings of the
AKNOWLEDGMENTS
National Academy of Sciences 2011, 108, 11399.
This work was supported by the Howard Hughes Medical Institute and
the NIH/NIGMS (R01 GM095501) the NIH/NIDCD (R01
DC006908). M.L. was supported by fellowships from Harvard CCB.
Y.T. and Y.S were supported by the Frederick and Ines Yeatts Hair Cell
Regeneration grant. Y.S. was supported by The National Nature Sci-
ence Foundation of China (NSFC81300824). Angela Steinauer is a
Howard Hughes Medical Institute International Student Research
fellow. We are grateful to Thomas Hasaka and Nicola Tolliday of the
Broad Institute for assistance with the primary screen.
(29)
Dijkstra, J.; Van Galen, M.; Scherphof, G. L. Biochimica
et Biophysica Acta (BBA)-Molecular Cell Research 1984, 804, 58.
(30)
Rozek, T.; Bowie, J. H.; Wallace, J. C.; Tyler, M. J.
Rapid Communications in Mass Spectrometry 2000, 14, 2002.
(31)
Wegener, K. L.; Wabnitz, P. A.; Carver, J. A.; Bowie, J.
H.; Chia, B. C.; Wallace, J. C.; Tyler, M. J. European journal of biochemistry /
FEBS 1999, 265, 627.
(32)
Balla, M.; Bowie, J. H.; Separovic, F. European
Biophysics Journal 2004, 33, 109.
(33)
Fernandez, D. I.; Le Brun, A. P.; Whitwell, T. C.; Sani,
M.-A.; James, M.; Separovic, F. Physical Chemistry Chemical Physics 2012, 14,
15739.
REFERENCES
10
ACS Paragon Plus Environment

Page 11 of 12
Journal of the American Chemical Society
(34)
Sahay, G.; Querbes, W.; Alabi, C.; Eltoukhy, A.; Sarkar,
(42)
Izumikawa, M.; Minoda, R.; Kawamoto, K.; Abrashkin,
S.; Zurenko, C.; Karagiannis, E.; Love, K.; Chen, D.; Zoncu, R. Nature
biotechnology 2013, 31, 653.
K. A.; Swiderski, D. L.; Dolan, D. F.; Brough, D. E.; Raphael, Y. Nature Medicine
2005, 11, 271.
1
2
3
4
(35)
(36)
Yu, P.; Liu, B.; Kodadek, T. Nat Biotech 2005, 23, 746.
Holub, J. M.; LaRochelle, J. R.; Appelbaum, J. S.;
(43)
Thompson, D. B.; Cronican, J. J.; Liu, D. R. In Methods
in Enzymology; Wittrup, K. D., Gregory, L. V., Eds.; Academic Press: 2012; Vol.
Volume 503, p 293.
Schepartz, A. Biochemistry 2013, 52, 9036.
(37)
LaRochelle, J. R.; Cobb, G. B.; Steinauer, A.; Rhoades,
(44)
1980, 45, 3084.
(45)
1982, 39, 1.
(46)
Simons, S. S.; Pons, M.; Johnson, D. F. J Org Chem
Dunkerton, L. V.; Markland, F. S.; Li, M. P. Steroids
Kwon, Y. U.; Kodadek, T. J Am Chem Soc 2007, 129,
Sage, C.; Huang, M.; Karimi, K.; Gutierrez, G.;
5
6
7
8
E.; Schepartz, A. Journal of the American Chemical Society 2015, 137, 2536.
(38)
(39)
Howarth, M.; Ting, A. Y. Nature protocols 2008, 3, 534.
Akil, O.; Seal, Rebecca P.; Burke, K.; Wang, C.; Alemi,
A.; During, M.; Edwards, Robert H.; Lustig, Lawrence R. Neuron 2012, 75, 283.
(40) Zuris, J. A.; Thompson, D. B.; Shu, Y.; Guilinger, J. P.;
Bessen, J. L.; Hu, J. H.; Maeder, M. L.; Joung, J. K.; Chen, Z.-Y.; Liu, D. R. Nat
Biotech 2015, 33, 73.
9
1508.
(47)
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
Vollrath, M. A.; Zhang, D.-S.; García-Añoveros, J.; Hinds, P. W.; Corwin, J. T.;
(41)
Taura, A.; Taura, K.; Choung, Y. H.; Masuda, M.; Pak,
Corey, D. P.; Chen, Z.-Y. Science 2005, 307, 1114.
K.; Chavez, E.; Ryan, A. F. Neuroscience 2010, 166, 1185.
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 12
Table of Contents graphic:
1
2
3
4
5
Conjugation
to synthetic
peptide library
Endocytosis
6
7
Screening for
non-endosomal
protein delivery
+36 GFP-Cre
8
9
Discovery of endosomal
escape-promoting peptides
Superpositively
charged protein
Endocytosis
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
Sequestration
within endosomes
ACS Paragon Plus Environment
12