2 H-Azirine-Based Reagents for Chemoselective Bioconjugation at Carboxyl Residues Inside Live Cells

Article abstract of DOI:10.1021/jacs.9b12116

Protein modification by chemical reagents has played an essential role in the treatment of human diseases. However, the reagents currently used are limited to the covalent modification of cysteine and lysine residues. It is thus desirable to develop novel methods that can covalently modify other residues. Despite the fact that the carboxyl residues are crucial for maintaining the protein function, few selective labeling reactions are currently available. Here, we describe a novel reactive probe, 3-phenyl-2H-azirine, that enables chemoselective modification of carboxyl groups in proteins under both in vitro and in situ conditions with excellent efficiency. Furthermore, proteome-wide profiling of reactive carboxyl residues was performed with a quantitative chemoproteomic platform.

Full text of DOI:10.1021/jacs.9b12116

Subscriber access provided by - Access paid by the | UCSB Libraries
Article
2H-Azirine-based Reagents for Chemoselective
Bioconjugation at Carboxyl Residues inside Live Cells
Nan Ma, Jun Hu, Wenyan Liu, Minhao Huang, Youlong Fan, Xingfeng
Yin, Jigang Wang, Ke Ding, Wencai Ye, Zhengqiu Li, and Zhi-Min Zhang
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b12116 • Publication Date (Web): 11 Mar 2020
Downloaded from pubs.acs.org on March 11, 2020
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 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 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.
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 8
Journal of the American Chemical Society
1
2
3
4
5
2H-Azirine-based Reagents for Chemoselective Bioconjugation at Carboxyl
Residues inside Live Cells
6
7
8
9
Nan Ma,† Jun Hu,† Zhi-Min Zhang,† Wenyan Liu, † Minhao Huang,† Youlong Fan, † Xingfeng Yin, ‡ Jigang Wang,*
Ke Ding,*† Wencai Ye, *† Zhengqiu Li*†
§
† School of Pharmacy, Jinan University, International Cooperative Laboratory of Traditional Chinese Medicine Modernization and
Innovative Drug Development, Ministryof Education (MOE) of China, 601 Huangpu Avenue West, Guangzhou, China 510632
‡ Institute of Life and Health Engineering, College of Life Science and Technology, Jinan University
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
§
Artemisinin Research Center, and Institute of Chinese Materia Medica, China Academy of Chinese Medical Sciences, Beijing, China,
100700
Abstract: Protein modification by chemical reagents has played an
essential role in the treatment of human diseases. However, the
reagents currently used are limited to the covalent modification of
cysteine and lysine residues. It is thus desirable to develop novel
methods that can covalently modify other residues. Despite the fact
that the carboxyl residues are crucial for maintaining the protein
function, few selective labeling reactions are currently available.
Here, we describe a novel reactive probe, 3-phenyl-2H-azirine, that
enables chemoselective modification of carboxyl groups in proteins
under both in vitro and in situ conditions with excellent efficiency.
Furthermore, proteome-wide profiling of reactive carboxyl residues
was performed with a quantitative chemoproteomic platform.
form an unstable ester product in the bioconjugation, but a more
stable N-phenacylacetamide was generated in the reaction with 3-
phenyl-2H-azirine. Therefore, we reasoned that 3-phenyl-2H-azirine
could be an ideal reagent for the selective modification of carboxylic
acid groups in a protein. The proposed bioconjugation mechanism is
shown in Figures 1A and 1B. The carboxylic acid reacts readily with
3-phenyl-2H-azirine, and this is followed by intramolecular
nucleophilic addition and ring expansion to form the N-
phenacylacetamide adduct.[9] Owing to this unique reaction
mechanism and the chemical properties of 3-phenyl-2H-azirine, we
envisioned that this probe could confer the following advantages: 1)
good stability of the probe in physiological conditions; 2) high
reactivity with the carboxylic acid group due to the strong
electrophilicity of the azirine; and 3) high stability of the ligation
product, N-phenacylacetamide.
Introduction
Chemoselective modification of amino acids offers a powerful
way for the study of protein function and structure.[1] When coupled
with activity-based protein profiling (ABPP) technology, it can
enable the discovery of therapeutic inhibitors and druggable
targets.[2] The methods for protein bioconjugation reported to date are
mainly based on selective reactions at cysteine, lysine, serine,
tyrosine, histidine and methionine residues.[3] The carboxylic acid
group is a residue in protein that acts as both a hydrogen-bond donor
and acceptor, and possesses various functions.[4,5] Notwithstanding,
very few selective reactions are available for the bioconjugation at
carboxylic acid residues, especially in live cells. It is thus important
to develop novel methods for the covalent modification of acidic
residues. This will not only expand the protein functionalization
toolbox, but also facilitate the discovery of druggable hotspots and
new types of covalent inhibitors. Recently, Raines et al. and Lei et al.
reported using diazo compounds to esterify the carboxylic acid
groups in proteins[6] and Waldmann et al. developed Woodward’s
reagent K for covalent protein labeling at glutamic acids.[7] These
approaches however have not been applied in live cells. We recently
discovered that tetrazole-based probes can selectively crosslink with
Glu and Asp residues in live cells,[8] but UV irradiation is required to
achieve covalent binding. In this study, we further developed a novel
reactive probe that can be directly used in live cells for selective
modification of carboxyl residues at protein binding sites.
(A) Bioconjugation at carboxylic acid residues
O
O
Asp/Glu > 12%
N
COOH
H
O
Reactive group
N
Live cells
O
AZ-9
Labeled protein
(B) Reaction mechanism
N
O
O
O
H
NHBoc
O
N
HO
O
NHBoc
O
O
O
O
AZ-11
S20
NHBoc
O
O
NHBoc
O
O
O
O
NH
O
NH
O
O
O
Figure 1. (A) Proposed bioconjugation of the probe AZ-9 with Glu/Asp
in proteins. (B) Reaction mechanism of the azirine with N-Boc-
protected glutamic acid.
Results and discussion
Due to the intrinsically weak nucleophilicity of the carboxylic acid
group, a functional group needs to possess high reactivity and at the
same time avoid unspecific side reactions to achieve selective
labeling.[6,7] Further, the covalent adducts formed should be
sufficiently stable for subsequent analysis of the protein function.
Based on these considerations, various chemical reactions that have a
carboxylic acid involved as a reactant were examined. It was found
that the 3-phenyl-2H-azirine, a stable compound under normal
conditions, can react with carboxylic acids under ambient
conditions.[9] Previously reported modifiers of carboxyl residues
To test this idea, a 3-phenyl-2H-azirine based probe, AZ-9, was
readily assembled based on typical synthetic methods (Figure 1A and
Scheme S9).[10] The synthesis starts with the reaction of styrene with
NaN3 and ICl to give the 2H-azirine. An alkyne handle was
embedded into the probe for the subsequent target analysis, and an
alkyne handle-free analogue (AZ-11) was also prepared as a
competitor. In addition, a series of aliphatic 2H-azirines (AZ-1/2),
2,3-diphenyl-2H-azirine (AZ3-8) and 2,2-dialkyl-3-phenyl-2H-
azirine (AZ-10) probes were synthesized for a comparison with 3-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 8
1
2
3
(A)
3-phenyl-2H-azirines
Aliphatic-2H-azirines
2,3-diphenyl-2H-azirines
Control probes
4
5
6
7
8
O
N
N
N
N
I
ButO2C
N
H
AZ-1
R
O
O
O
IAA-alkyne
H
AZ-9
9
Me
N
AZ-6
AZ-11
R=
AZ-3
AZ-4
R=
R=
R=
O
O
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
N
CO2But
CO2Me
OMe
AZ-7
AZ-8
R=
R=
ButO2C
N
O
O
ButO2C
AZ-10
AZ-2
NO2
AZ-5
O
NHS-alkyne
(B)
(C)
O
NHBoc
H
N
O
AZ-11
O
O
O
S20
Figure 2. (A) Chemical strcutures of the probes AZ1-11 and the control probes IAA-alkyne, NHS-alkyne. (B) HPLC analysis of the reaction
betwen AZ-11 and N-Boc-protected glutamic acid. (C) Binding site analysis of AZ-9 with unprotected peptide of SYWCDFENTQHRMK.
phenyl-2H-azirine (AZ-9, Figure 2A). Two cysteine- and lysine-
specific probes, IAA-alkyne and NHS-alkyne, were used as controls
in the following biological experiments.
Subsequently, we assessed the performance of the probes in
labeling pure enzymes, commercially available or purified proteins,
including bovine serum albumin (BSA), ribonuclease A (RNaseA),
myosin, β-amylase, human serum albumins (HSA), YopJ and
catalase were selected as model systems. YopJ is a ubiquitin-like
protein protease that exerts a pathogenic effect.[11] First, labeling
profiles of the probes with BSA were carried out. Different from
previous studies that high probe concentrations (millimolar) were
With these probes in hand, we first used NMR to test the stability
of the representative probes (AZ-3/6/7/11), which showed that they
can be stored in organic solvent, aqueous solution or the solid state
for three days without degradation, demonstrating the good stability
(Figure S5). To determine if the probes could react with protected
carboxylic acids in aqueous buffer, we examined the reaction
between AZ-11 and N-Boc-protected glutamic acid in a mixture of
THF/H2O (1:1). Although the reaction appeared to be slow (Figure
2B), the N-phenacylacetamide product (S20) was successfully
isolated and characterized, which has been proven to be stable in
aqueous solution by NMR analysis (Scheme S12, Figure S5). We
checked the reaction with the side chain of other protected amino
acids under the same conditions, which showed that AZ-11 can also
react with cysteine, histidine, lysine and tyrosine residues after 12 h,
but not with serine and methionine residues (Figure S6). This
supports that the probe modifies a small number of residues (e.g.
lysine, tyrosine and cysteine) other than the carboxylic acid groups in
the following labeling of proteins and live cells. Subsequently, the
required in the labeling of carboxylic acid residues,
a low
micromolar concentration was used in the following experiments
with the aim of reducing nonspecific labeling. After incubation of
AZ1-10 (2.5 μM) with bovine serum albumin (BSA) for 1.5 h in PBS,
the mixture was clicked with TAMRA-N3, and the labeled proteins
were separated by SDS-PAGE and analyzed by in-gel fluorescence
scanning. As shown in Figure 3A, strong fluorescent bands were
observed from the samples treated with AZ-1/3/4/5/8/9/10, which is
stronger than IAA-alkyne-treated samples, indicating that these
probes can crosslink with more amino acid residues than cysteine.
Pretreatment of excess NHS and IAA failed to decrease the
fluorescence intensity, indicating that the labeling does not
predominantly occur at cysteine and lysine residues (Figures 3B/S10).
Time- and concentration-dependent labeling experiments showed
that an incubation time of 1 h and 0.5 μM probe concentration is
sufficient to produce a visible band, signifying excellent labeling
efficiency (Figures 3C/S11). Under similar conditions, AZ-9 can
successfully label myosin, β-amylase, catalase, HSA, ribonuclease A
and YopJ proteins (Figures 3D/S12). Interestingly, the fluorescence
intensity becomes much weaker upon pretreatment with as little as
one equivalent of the competitor AZ-11, demonstrating that the
labeling is generated from the probe, not from nonspecific labeling.
Finally, the binding sites of AZ-9 with BSA, YopJ and β-amylase
were characterized by LC-MS/MS. The AZ-9 labeled proteins were
reactivity
of
AZ-9
with
an
unprotected
peptide,
SYWCDFENTQHRMK was evaluated. After incubation of the
peptide with AZ-9 in phosphate buffered saline (PBS, pH = 7.3) for 2
h, the resulting mixture was directly analyzed by HPLC and LC-
MS/MS. Interestingly, all glutamic acid (E) and aspartic acid (D)
residues were efficiently modified by AZ-9 and no other
modifications were observed (Figures 2C and S7). HPLC analysis
confirmed the completeness of the reaction (Figure S8), indicating
that the physiological condition and the microenvironment of peptide
could promote the bioconjugation of the probe with carboxylic acid
residues.
ACS Paragon Plus Environment

Page 3 of 8
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
(B)
AZ-9 + BSA
AZ-9 + HSA
(A)
(C)
0
0.05 0.5 2.5
5
0
0.05 0.5
2.5
5
IAA (μM)
IAA (μM)
69 kDa
69 kDa
FL
FL
FL
CBB
CBB
CBB
NHS (μM)
NHS (μM)
FL
0
0.05 0.5 2.5
5
0
0.05 0.5 2.5
5
0.1
1
2
5
10
20
30
40
50
60 120
Time/min
FL
FL
9
CBB
CBB
CBB
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
(D)
(E)
(F)
(G)
myosin
AZ-9
RNase A
AZ-9
YopJ
AZ-9
Modification sites
Modification sites
Modification sites
AZ-11
+
+
+
Glu: 26
Asp: 15
Lys: 1
Glu: 6 Asp: 1
Glu: 5
Asp: 1
Tyr : 1
*
*
*
FL
CBB
catalase
AZ-9
β-amylase
HSA
AZ-9
AZ-9
AZ-11
FL
+
+
+
*
*
*
CBB
BSA
β-amylase
YopJ
(H)
(I)
C172
E378
D53
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. (A) Labeling of BSA (1 μg/μL) with the probes AZ1-10 (final probe concentration is 2.5 μM). FL = in-gel fluorescence scanning. CBB
= Coomassie gel. (B) Labeling of BSA and HSA in the presence of IAA (iodacetamide) and NHS (N-hydroxysuccinimide) with AZ-9 (2.5 μM).
(C) Time-dependent labeling of BSA with AZ-9 (2.5 μM). (D) Labeling of myosin, RNaseA, YopJ, catalase, β-amylase and HSA in the presence
or abesence of competitor (AZ-11, 1 equivalent). Modification site analysis of AZ-9 (2.5 μM probe) with BSA (43 μM) (E), YopJ (92 μM) (F)
and β-amylase (50 μM) (G). (H) Homology modeling structure of modification sites in YopJ (template PDB code: 6BE0) (I) Crystal structure of
active site E378 from β-amylase (PDB code: 2XGB).
clicked with acid-cleavable biotin-azide (DADPS, Table S1), after
affinity purification and trypsin digestion, the probe-labeled peptides
were released and analyzed by LC-MS/MS. Modified peptides that
were identified from at least two of triplicate runs were recorded. As
shown in Figures 3E/3F/3G and S15/16/17, 41, 7, and 6 carboxyl
residues were covalently modified from BSA, YopJ and β-amylase,
respectively. Only one side product was generated from BSA and β-
amylase at a lysine and a tyrosine residue, respectively, which
confirmed the amino acid selectivity of this method. The number of
modified Glu residues is much more than those of Asp, showing
marked preference of glutamic acid over aspartic acid, which could
be accounted for by the less steric hindrance of glutamic acid residue
than aspartic acid residue. Amongst the identified residues, D53 from
YopJ is located in the active site pocket, and the distance between
D53 and catalytic site C172 is 6Å (Figure 3H), indicating that this
site might be covalently modified to mediate the protein function.
The E378 from β-amylase that was identified is located in the active
site and serves as a proton acceptor (Figure 3I). Crystal structure
analysis revealed that most of the modified carboxyl residues are
located in protein binding pockets or buried sites, rather than on the
protein surface (Table S2/3/4). These data demonstrate that AZ-9 is
an efficient probe for protein functionalization at carboxyl residues in
protein binding sites.
To determine if AZ-9 can selectively label the functional glutamic
acid residue under in vitro and in situ conditions, the labeling profile
of aldehyde dehydrogenase (ALDH1),[12] an enzyme containing a
catalytic glutamic acid residue was carried out. As showed in Figure
4A, most probes can successfully label ALDH1 protein at a low
probe concentration. Similarly, the labeling profiles were not affected
by the treatment of excess IAA and NHS, but become weaker in the
presence of competitor AZ-11 (Figures 4B/4C). To characterise the
binding sites of AZ-9 with ALDH1, pull-down experiments with
recombinant ALDH1 were carried out, and the probe labeled
peptides were analysed by LC-MS/MS. As shown in Figure 4D, only
the functional E269 residue was modified by AZ-9, which confirmed
the excellent labeling specificity and efficiency. To verify the
selective labeling capability of AZ-9 under in situ conditions, the
catalytic residue (E269) was mutated to an alanine and then
transfected into MCF-7 cells, and wild-type (WT) enzyme was tested
simultaneously. After incubation of the probes with transfected cells
for 3 h, the cells were lysed and conjugated with TAMRA-N3, and
the resulting samples were then separated by SDS-PAGE and imaged
by fluorescence scanning. As shown in Figure 4E, AZ-
9 preferentially labels WT ALDH1 over the E269A mutant. Docking
experiments demonstrate that AZ-9 is accommodated well with the
binding pocket of ALDH1 (Figure 4F). These data demonstrate that
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 8
1
2
3
4
5
6
7
8
9
Inspired by these results, we
(E)
(A)
(B)
further evaluated the proteome
reactivity profiles of the probes in
live mammalian cells. First, the
labeling profile of AZ1-10 with a
breast cancer cell line (MCF-7) was
carred out. The procedures were
similar to those described above, as
expected, the 3-phenyl-2H-azirine
probe (AZ-9) appeared to be
superior to the other probes,
including IAA-alkyne, in labeling
endogenous proteins (Figure 5A).
Strong labeling bands throughout
the whole lane were observed from
AZ-9 treated samples. In contrast,
55 kDa
FL
FL
55 kDa
55 kDa
55 kDa
WB
CBB
(Anti-flag)
(C)
ALDH1 Labeling
AZ-9 + ALDH1
AZ-9 + ALDH1
NHS (μm)0 0.05 0.5 2.5
AZ-9 + ALDH1
+
20
15
10
5
IAA (μm) 0 0.05 0.5 2.5
5
5
AZ-11
FL
CBB
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
(D)
0
T
K
A
9
C
W
6
O
2
E
M
(F)
AZ-5-treated sample showed
a
selective labeling profile toward a
~60 kDa protein (Figure 5A, *
marked band), and the other
Figure 4. (A) Labeling of ALDH1 (1 μg/μL) with the probes AZ1-10 (final probe concentration is 2.5 μM).
FL = in-gel fluorescence scanning. CBB = Coomassie gel. Labeling of ALDH1 in the presence of IAA
(iodacetamide), NHS (N-hydroxysuccinimide) (B) and AZ-11 (C) with AZ-9 (2.5 μM). (D) Modification
site analysis of AZ-9 (2.5 μM probe) with ALDH1 at E269 by LC-MS/MS. (E) Labeling of WT and E269A
ALDH1 with AZ-9, respectively (top gel). The western blots (bottom gel) confirmed amount of WT and
E269A enzymes using flag antibodies. (F) Docking experiments to predict the binding model of AZ-9 with
ALDH1 (PDB code: 4WB9).
azirine-based probes produced only
faint or negligible bands. This
phenomenon could be attributed to
the substituents on the 2 position of
azirine drastically affecting the
probe binding with proteins in live
the probe possesses high amino acid-selectivity and bioconjugation
cells. All probes (AZ1-10) showed weak labeling profiles with MCF-
7 cell lysates (Figure S13), which underscored proteome reactivity
efficiency under both in vitro and in situ conditions.
(A)
(B)
(C)
AZ-11
(D)
AZ-9
AZ-9
AZ-9
2.5
AZ-9
0.1 0.5
AZ-9
2.5
Conc
Time
/h
IAA
NHS
(μM)
0
0.5
1
2
3
4
0
5
10
0
1
2.5
5
0
5
10
+
/μM
(μM)
kDa
75
FL
45
*
25
CBB
AZ-9
Probe/NU
IAA-alkyne
Probe/NU
(E)
(F)
(G)
Probe
Probe
Modification sites from MCF-7 cells
Glu Asp Others
isoTOP ABPP analysis of
AZ-9-treated MCF7 cells
Probe
20
10
>
BASP1
Probe+
IAA
NUDC
TPD54
5
DMSO
DMSO/NU
Probe+
AZ-11
0
0
200
400
600
800
69 binding sites
peptides
Figure 5. (A) Proteome reactivity profiles of AZ1-10 with MCF-7 cells ( final probe concentration is 2.5 μM), FL = in-gel fluorescence scanning.
CBB = Coomassie gel. (B) Time- and concentration-dependent labeling profiles of AZ-9 with MCF-7 cells. (C) Competitive labeling profiles of AZ-
9 (2.5 μM) with MCF-7 cells in the presence or absence of competitor (1×AZ-11). (D) Competitive labeling profiles of AZ-9 (2.5 μM) with MCF-7
cells in the presence of IAA (iodacetamide) and NHS (N-hydroxysuccinimide). (E) Cellular imaging of AZ-9 in the presence or absence of the
competitors (2.5 μM probe concentration, 1×AZ-11). Scale bar = 10 μm. (F) Quantitative mass spectrometry-based profiling of AZ-9 binding
proteins, DMSO treated sample as a negative control. (G) Analysis of binding sites of AZ-9 from live MCF-7 cells.
under in situ conditions. The labeling profiles with other cancers cell
lines such as THP1, HeLa and HEK293 cells also demonstrated that
ACS Paragon Plus Environment

Page 5 of 8
1
Journal of the American Chemical Society
2
3
4
5
6
7
8
9
AZ-9 was more efficient than other probes in labeling endogenous
unlike the IAA-alkyne probe where the fluorescence signals were
spread across the whole cell, AZ-9 was localized mainly in the cell
nucleus and partially in the cell cytosol, indicating that the probe is
able to label nuclear proteins. Consistent with the trend of labeling
profiles, the fluorescence intensity of AZ-9 was obviously stronger
than that of the IAA-alkyne probe, which was abolished by the
treatment with competitors AZ-11 but not by IAA (Figure 5E).
Control imaging experiments with DMSO under similar conditions
gave only minimal background fluorescence compared to labeled
cells (Figure 5E). Taken together, results from these competitive
labeling and imaging experiments demonstrate the excellent
proteome reactivity of AZ-9 in live-cell settings.
proteins (Figure S13). Thus, we focused on AZ-9 in the subsequent
experiments with live MCF7 cells. Time- and concentration-
dependent labeling profiles showed that AZ-9 can produce visible
bands at as low as 2.5 μM probe concentration and with as little as
0.5 h of incubation time, which again proved excellent labeling
efficiency under in situ conditions (Figure 5B). Consistent with the
results from the competitive labeling profiles with pure enzymes
(Figures 3D/4C), the fluorescence intensity decreased sharply upon
the treatment of competitors (AZ-11), but not by IAA and NHS
(Figures 5C/D/S14).
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
To track the subcellular location of the probes in live cells,
cellular imaging experiments were carried out. Live MCF-7 cells
were incubated with AZ-9 for 3 h, and were then fixed and
permeabilized. Subsequently, the samples were clicked with
TAMRA-N3 and imaged under previously optimized conditions,[13]
and IAA-alkyne was used as a control concurrently. Interestingly,
To identify the labeled proteins and binding sites in live cells,
tandem orthogonal proteolysis (TOP)-ABPP experiments were
carried out. Heavy- and light-labeled (SILAC) MCF-7 cells were
incubated with AZ-9 and DMSO, respectively, and this was followed
by cell lysis and conjugation with acid-cleavable biotin-N3.
(B)
(A)
(C)
2.0 < R10:1 < 5.0
R100:10 < 2.0
100
80
60
40
20
0
AZ-9 10:1
AZ-9 1:1
Glu: 1128 sites
>10
Asp: 549 sites
Others: 72 sites
5
0
Annotated actived site+calcium
binding regions
0
500
Peptide no.
1000
Others annotated Asp/Glu residues
No functional annotation
Glu
Asp
others
(D)
Proteins
Peptide
Labeling
Function
R100:10
sites
Superoxide dismutase [Cu-Zn]
Superoxide dismutase [Cu-Zn]
Protein S100-A4
GDGPVQGIINFE(1)QK
GD(1)GPVQGIINFEQK
LMSNLD(1)SNR
E22
D12
D63
D10
E303
Natural variant-in ALS1
Unannotated
Calcium binding
Unannotated
Zinc finger-C4-type
1.73
3.13
1.49
3.00
1.99
Protein S100-A4
Eukaryotic translation initiation factor 2
subunit 2
ALD(1)VMVSTFHK
LYFLQCE(1)TCHSR
Eukaryotic translation initiation factor 2
subunit 2
SPD(1)TILQK
D288
D92
Zinc finger-C4-type
Unannotated
2.70
4.25
1.67
Eukaryotic translation initiation factor 2
subunit 2
IFDID(0.979)EAEEGVK
Site-Contributes to redox
potential value
Thioredoxin domain-containing protein 17
Fatty aldehyde dehydrogenase
SWCPD(1)CVQAEPVVR
HLTPVTLE(1)LGGK
IVE(1)MSTSK
D45
E207
E42
Active site
Nature variant
1.45
1.28
Eukaryotic translation initiation factor 5A-2
NAD(P)H dehydrogenase [quinone] 1
Poly(U)-binding-splicing factor PUF60
Unannotated
Unannotated
1.01
1.27
IQILE(1)GWK
E206
E482
DIDDDLEGE(0.975)VTE
ECGK
(E)
(F)
(G)
Non-Drugbank proteins
49%
E22
E236
AZ-9
51%
AZ-9
SODC
LDHA
Drugbank proteins
102 proteins
Figure 6. (A) Identified isoTOP-ABPP ratios for peptides from MCF7 cells labeled with two pairwise AZ-9 probe concentrations (10:10 μM,
100 : 10 μM). (B) The percentage of functionally annotated residues searched from the UniProt database. (C) The percentage of functionally
annotated residues searched from the UniProt database. (D) Subset of residues selectively identified by AZ-9 with low isoTOP-ABPP ratios. (E)
The percentage of labeled proteins possessing unique peptides with highly reactive carboxylate residues (R100:10 < 2) and found in Drugbank. (F)
Crystal structure of the modification sites with isoTOP ABPP ratios (R100:10) < 2 and the known active-site ligand (SODC, PDB code: 4A7T) (left
figure), and the docking experiments to predict the binding of AZ-9 with SODC (right figure). (G) Crystal structure of the modification sites
E236 and the known active-site ligand in LDHA (left figure , PDB code: 6BAD), and the docking experiments to predict the binding of AZ-9
ACS Paragon Plus Environment
with LDHA (right figure).

Journal of the American Chemical Society
Page 6 of 8
1
2
Subsequently, the probe-labeled proteins were pulled-down, after
protein, such as E22/D12 of superoxide dismutase [Cu-Zn] (SODC),
3
4
5
6
7
8
9
trypsin digestion and acid cleavage, the probe labeled proteins and
modified sites were tested by LC-MS/MS. The obtained protein hits
were analyzed by corresponding plots of SILAC ratios. As shown in
Figure 5F and Table S6, a series of protein hits including BASP1
(brain acid-soluble protein 1), TPD54 (tumor protein D54), PTMA
(prothymosin alpha), NUDC (nuclear migration protein nudC),
NACA (nascent polypeptide-associated complex), TYB10 (thymosin
β10) and NASP (nuclear autoantigenic sperm protein) with high
ratios were identified. Consistent with the imaging results (Figure
5E), most protein hits such as BASP1, PTMA, NUDC, NACA,
NASP and LASP1 were located in the nucleus, indicating the
reliability of these protein hits. Two proteins in Figure 5F and Table
S6, BASP1 and PTMA, are biomarkers for pancreatic cancer and
esophageal squamous cell carcinoma (ESCC), respectively.[15]
TYB10 is associated with unfavorable prognosis in hepatocellular
carcinoma.[16] TPD54, a biomarker for tumors that is sensitive to
metformin treatment, has important biological functions in breast
cancer.[17] Further analysis of the binding sites revealed that up to
98.5% of the probe-modified sites were Glu and Asp, which confirms
the excellent carboxylic acid selectivity (Figure 5G, Table S7).
Significantly, most binding sites such as D7 of PTMA, E34 of
BASP1, E52 of NUDC and E36 of TYB10 were located in the
protein binding pockets (Table S7), demonstrating that this probe
provides an efficient and unique way to access these important
proteins. Competitive proteomics with AZ-9 in the presence of
different concentrations of AZ-11 was carried out, a series of
interesting targets, such as DDX5/17, KCNKA, TPM4 were
identified (Figure S19 and Table S8).
E52/D63/51/10 of protein S100-A4 (S10A4), E303/D92/288 of
eukaryotic translation initiation factor 2 subunit 2 (IF2B). As shown
in Figure 6D, the annotated active sites yield low R100:10 values (E22
of SODC, R100:10=1.73; D63 of S10A4, R100:10=1.49, D45 of TXD17,
R100:10=1.67), indicating that the isoTOP ABPP ratios can be used to
predict the activity and function of the identified sites. In addition,
the sites identified in the TOP-ABPP experiments described above
(Figure 5G, Table S7), such as D7 of PTMA, E34 of BASP1and E36
of TYB10 were also detected in this experiment, which demonstrates
the reproducibility of these data. D7 of PTMA and E36 of TYB10
showed low ratios (R100:10 ≈ 2), indicating that they are reactive and
functional sites. There are 145 other highly reactive carboxyl
residues (R100:10 < 2) that were identified in 102 proteins of
uncharacterized function. 51% of these proteins were found in the
Drugbank database and about 49% of the proteins still lack small-
molecule probes (Figure 6E). The crystal structure analysis of these
sites such as E236 of LDHA, E84 of PPIA, E286 of VIME, E133 of
6PGD, E210 of AL3A1 reveals that they are located in the active
binding pockets (Figures 6F/6G and S21), and the distances between
the modified sites and the known binding ligands are within 5 Å,
indicating that these might be potential new ways to mediate the
protein functions.
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
Conclusions
In conclusion, a chemoselective probe, AZ-9, has been created as
a protein bioconjugation reagent at the carboxyl residues. The
reagent has displayed excellent labeling efficiency and specificity
under both in vitro and in situ conditions, and can be complementary
to existing methods. During the proteome reactivity profiling
experiments, a series of highly reactive carboxyl groups with
annotated or unannotated functions were identified by the iso-TOP
ABPP approach. These carboxyl groups might be ligandable residues
to perturb the activities of a diverse range of proteins. Meanwhile,
this new reactive warhead, 3-phenyl-2H-azirine, could be used in the
development of new types of covalent inhibitors.
To quantitatively assess the intrinsic reactivity and functionality of
the probe-modified carboxylic acid residues, the well-established
isotopic tandem orthogonal proteolysis–activity-based protein
profiling (isoTOP-ABPP) was performed on
a proteome-wide
scale.[18] Two parallel experiments with the live MCF-7 cells using
pair-wise AZ-9 concentrations of 10:10 μM, 100:10 μM (light :
heavy) were carried out. The purpose of using higher probe
concentrations in this experiment is to profile more reactive residues.
As a result, a total of 1749 binding sites from 583 proteins were
identified and 95.9% of these were proved to be Asp and Glu
residues (Figures 6A, 6B and Table S9), which again proved the
excellent selectivity of this method. Amongst the identified
carboxylic acid residues, three residues are in the C-terminus, and the
other modification sites include a small number of histidine, lysine
and cysteine etc. (Table S9). In contrast, no C-terminus was detected
in the labeling of synthetic peptides and recombinant proteins
(Figures 2C, 3E/F/G), which could be accounted for by the different
ASSOCIATED CONTENT
Supporting Information:
This material is available free of charge via the internet at http://pubs.
acs.org.
Keywords:
Protein modification ·2H-azirine · carboxyl residues· activity-based
protein profiling · imaging
[19]
biological conditions between in vitro and in situ environment.
AUTHOR INFORMATION
Consistent with previous studies,[2] a large fraction of peptides that
carry the AZ-9 modification exhibited isoTOP ABPP ratios (R100:10
)
Corresponding Author
Pharmlzq@jnu.edu.cn,
chywc@aliyun.com
of 5 or greater (Figure 6A). In contrast, a small subset of highly
reactive carboxyl residues (8.3% of total peptides) has ratios below 2
(Figure 6C). Amongst the reactive sites, many carboxyl residues such
as E207 of fatty aldehyde dehydrogenase (AL3A2), E174 of
tropomyosin alpha-3 chain (TPM3), D288/E303 of eukaryotic
dingke@jnu.edu.cn
jgwang@icmm.ac.cn
translation initiation factor
2 subunit 2 (IF2B) and D45 of
Notes
thioredoxin domain-containing protein 17 (TXD17) are annotated
active sites by searching from the UniProt database. They can display
a wide range of biological activities such as zinc/calcium binding,
natural variant, contribution to redox (Figure 6D). Meanwhile, we
found that multiple binding sites were identified from the same
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
ACS Paragon Plus Environment

Page 7 of 8
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
[8] a) Cheng, K.; Lee, J. S., Hao, P.; Yao, S. Q.; Ding, K.; Li, Z.
Funding was provided by National Key R&D Program of China
Tetrazole ‐ Based Probes for Integrated Phenotypic Screening,
Affinity‐Based Proteome Profiling, and Sensitive Detection of a
Cancer Biomarker. Angew. Chem., Int. Ed. 2017, 56, 15044–
15048. b) Li, Z.; Qian, L.; Li, L.; Bernhammer, J. C.; Huynh, H.
V.; Lee, J. S.; Yao, S. Q. Tetrazole photoclick chemistry:
reinvestigating its suitability as a bioorthogonal reaction and
potential applications. Angew. Chem. Int. Ed. 2016, 55, 2002–
2006.
(2019YFC1711000), National Natural Science Foundation of China
(81803389, 21877050), Science and Technology Program of
Guangdong Province (2017A050506028, 2019B151502025), Science
and Technology Program of Guangzhou (201704030060,
201805010007),
China
Postdoctoral
Science
Foundation
(No.2017M622923), the PhD Start-up Fund of Natural Science
Foundation of Guangdong Province, China (NO. 2018A030310651) .
We thank S. Q. Yao (NUS, Singapore) for the invaluable suggestions
and support on this work.
[9] a) David, S. C. B.; James, E. D. Direct conversion of α-halo acids
into α-halo amides by reaction with 2-phenylazirine. Aust. J.
Chem., 1978, 31, 2313–2315. b) Palacios, F.; Aparicio, D.;
Ochoa De Retana, A. M.; de los Santos, J. M.; Gil, J. I.; Alonso, J.
M. Asymmetric synthesis of 2 H-azirines derived from phosphine
oxides using solid-supported amines. Ring opening of azirines
with carboxylic acids. J. Org. Chem. 2002, 67, 7283–7288.
[10] Wang, Y.; Lei, X.; Tang, Y. Rh (ii)-catalyzed cycloadditions of
1-tosyl 1, 2, 3-triazoles with 2 H-azirines: switchable reactivity of
Rh-azavinylcarbene as [2C]-or aza-[3C]-synthon. Chem.
Commun., 2015, 51, 4507–4510.
[11] Orth, K.; Xu, Z.; Mudgett, M. B.; Bao, Z. Q.; Palmer,L. E.;
Bliska, J. B.; Mangel, W. F.; Staskawicz, B.; Dixon, J. E.
Disruption of signaling by Yersinia effector YopJ, a ubiquitin-
like protein protease. Science. 2000, 290, 1594–1597.
[12] Weerapana, E.; Simon, G. M.; Cravatt, B. F. Disparate proteome
reactivity profiles of carbon electrophiles. Nat. Chem. Biol. 2008,
4, 405–407.
[13] a) Zhu, D.; Guo, H.; Chang, Y.; Ni, Y.; Li, L.; Zhang, Z.-M.;
Hao, P.; Xu, Y.; Ding, K.; Li, Z. Cell ‐ and Tissue ‐ Based
Proteome Profiling and Dual Imaging of Apoptosis Markers with
Probes Derived from Venetoclax and Idasanutlin. Angew. Chem.,
Int. Ed. 2018, 57, 9284–9289. b) Li, Z.; Qian, L.; Li, L.;
Bernhammer, J. C.; Huynh, H. V.; Lee, J. S.; Yao, S. Q.
Tetrazole photoclick chemistry: reinvestigating its suitability as a
bioorthogonal reaction and potential applications. Angew. Chem.
Int. Ed. 2016, 55, 2002–2006.
[14] Liu, W.; Zhang, Z.; Zhang, Z.-M. Hao, P.; Ding, K.; Li, Z.
Integrated phenotypic screening and activity-based protein
profiling to reveal potential therapy targets of pancreatic cancer.
Chem. Commun. 2019, 55, 1596–1599.
[15] a) Zhou, Q.; Andersson, R.; Hu, D.; Bauden, M.; Kristl, T.;
Sasor, A.; Pawłowski, K.; Pla, I.; Hilmersson, K. S.; Zhou, M.
Lu, F.; Marko-Varga, G.; Ansari. D. Quantitative proteomics
identifies brain acid soluble protein 1 (BASP1) as a
prognostic biomarker candidate in pancreatic cancer tissue
EBioMedicine. 2019, 19, 30242–30247. b) Zhu, Y.; Qi, X.;
Yu, C.; Yu, S.; Zhang, C.; Zhang, Y.; Liu, X.; Xu, Y.; Yang,
C.; Jiang, W.; Tian, G. ； Li, X.; Bergquist, J.; Zhang, J.;
Wang, L. ； Mi, J. Identification of prothymosin alpha
(PTMA) as a biomarker for esophageal squamous cell
carcinoma (ESCC) by label-free quantitative proteomics and
Quantitative Dot Blot (QDB). Clin. Proteomics. 2019, 16, 12.
[16] Song, C.; Su, Z.; Guo, J. Thymosin β 10 is overexpressed and
associated with unfavorable prognosis in hepatocellular
carcinoma. Biosci. Rep. 2019, 39, BSR20182355.
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
REFERENCES
[1] a) Counihan, J. L.; Ford, B.; Nomura, D. K. Mapping proteome-
wide interactions of reactive chemicals using chemoproteomic
platforms. Curr. Opin. Chem. Biol. 2016, 30, 68–76. b) Krall, N.;
Da Cruz, F. P.; Boutureira, O.; Bernardes, G. J. Site-selective
protein-modification chemistry for basic biology and drug
development. Nat. Chem. 2016, 8, 103–13.
[2] a) Roberts, A. M.; Ward, C. C.; Nomura, D. K. Activity-based
protein profiling for mapping and pharmacologically
interrogating proteome-wide ligandable hotspots. Curr. Opin.
Biotechnol. 2017, 43, 25−33. b) Weerapana, E.; Wang, C.; Simon,
G. M.; Richter, F.; Khare, S.; Dillon, M. B.D.; Bachovchin, D.
A.; Mowen, K.; Baker, D.; Cravatt, B. F. Quantitative reactivity
profiling predicts functional cysteines in proteomes. Nature 2010,
468, 790. c) Backus K, M.; Correia, B. E.; Lum, K. M.; Forli, S.;
Horning, B. D.; González-Páez, G. E.; Chatterjee, S.; Lanning, B.
R.; Teijaro, J. R.; Olson, A. J.; Wolan, D. W.; Cravatt, B. F.
Proteome-wide covalent ligand discovery in native biological
systems. Nature 2016, 534, 570.
[3] a) Tamura, T.; Hamachi, I. Chemistry for covalent modification
of endogenous/native proteins: from test tubes to complex
biological systems. J. Am. Chem. Soc. 2019, 141, 2782–2799. b)
Ward, C. C.; Kleinman, J. I.; Nomura, D. K. NHS-esters as
versatile reactivity-based probes for mapping proteome-wide
ligandable hotspots. ACS Chem. Biol. 2017, 12, 1478–1483. c)
Lin, S.; Yang, X.; Jia, S.; Weeks, A. M.; Hornsby, M.; Lee, P. S.;
Nichiporuk, R. V.; Iavarone, A. T.; Wells, J. A.; Toste, F. D.;
Chang, C. J. Redox-based reagents for chemoselective
methionine bioconjugation. Science 2017, 355, 597. d) Taylor, M.
T.; Nelson, J. E.; Suero, M. G.; Gaunt, M. J. A protein
functionalization platform based on selective reactions at
methionine residues. Nature 2018, 562, 563.
[4] a) Brosnan, J. T.; Brosnan, M. E. Glutamate: a truly functional
amino acid. Amino Acids. 2013, 45, 413–418. b) Lin, J.;
Pozharski, E.; Wilson, M. A. Short carboxylic acid–carboxylate
hydrogen bonds can have fully localized protons. Biochemistry.
2017, 56, 391–402. c) Kudryavtseva, A. A.; Osetrova, M. S.;
Livinyuk, V. Y.; Manukhov, I. V.; Zavilgelsky, G. B. The
importance of C-terminal aspartic acid residue (D141) to the
antirestriction activity of the ArdB (R64) protein. Mol Biol
(Mosk). 2017, 51, 831–835. d) Yoshikawa, S.; Tanimura, T.;
Miyawaki, A.; Nakamura, M.; Yuzaki, M.; Furuichi, T.;
Mikoshiba, K. Molecular cloning and characterization of the
inositol 1, 4, 5-trisphosphate receptor in Drosophila melanogaster.
J. Biol. Chem. 1992, 267, 16613–16619.
[17] Zhuang, Y.; Ly, R. C.; Frazier, C. V.; Yu, J.; Qin, S.; Fan, X. Y.;
Goetz, M. P.; Boughey, J. C.; Weinshilboum, R.; Wang, L. The
novel function of tumor protein D54 in regulating pyruvate
dehydrogenase and metformin cytotoxicity in breast cancer.
Cancer Metab. 2019, 7, 1.
[18] Weerapana, E.; Speers, A. E.; Cravatt, B. F. Tandem orthogonal
proteolysis-activity-based protein profiling (TOP-ABPP)—a
general method for mapping sites of probe modification in
proteomes. Nat. Protoc. 2007, 1414–1425.
[5] Chen, K.; Kurgan, L. Investigation of atomic level patterns in
protein—small ligand interactions. PLoS One. 2009, 4, e4473.
[6] a) McGrath, N. A.; Andersen, K. A.; Davis, A. K.; Lomax, J. E.;
Raines, R. T. Diazo compounds for the bioreversible
esterification of proteins. Chem. Sci. 2015, 6, 752–755. b) Zhang,
X.; Wang, J. H.; Tan, D.; Li, Q.; Li, M.; Gong, Z.; Tang, C.; Liu,
Z.; Dong, M. Q.; Lei, X. Carboxylate-selective chemical cross-
linkers for mass spectrometric analysis of protein structures. Anal.
Chem. 2018, 90, 1195–1201.
[7] Martín-Gago, P.; Fansa, E. K.; Winzker, M.; Murarka, S.;
Janning, P.; Schultz-Fademrecht, C.; Baumann, M.; Wittinghofer,
A.; Waldmann, H. Covalent protein labeling at glutamic acids.
Cell Chem. Biol. 2017, 24, 589–597.e5.
[19] Bloom, S.; Liu, C.; Kölmel, D. K.; Qiao, J. X.; Zhang, Y.; Poss,
M. A.; Ewing, W. R.; MacMillan, D. W. C. Nat Chem. 2018, 10,
205–211.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 8
1
2
3
Entry for the Table of Contents
4
5
6
7
8
9
A novel reactive probe, 3-phenyl-2H-azirine, was developed for
chemoselective modification of carboxyl groups in proteins under
both in vitro and in situ conditions with excellent efficiency. It can
facilitate the discovery of druggable hotspots and new types of
covalent inhibitors.
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
ACS Paragon Plus Environment