Iminoboronates as Dual-Purpose Linkers in Chemical Probe Development

Article abstract of DOI:10.1002/chem.202005115

Chemical probes that covalently modify proteins of interest are powerful tools for the research of biological processes. Important in the design of a probe is the choice of reactive group that forms the covalent bond, as it decides the success of a probe. However, choosing the right reactive group is not a simple feat and methodologies for expedient screening of different groups are needed. We herein report a modular approach that allows easy coupling of a reactive group to a ligand. α-Nucleophile ligands are combined with 2-formylphenylboronic acid derived reactive groups to form iminoboronate probes that selectively label their target proteins. A transimination reaction on the labeled proteins with an α-amino hydrazide provides further modification, for example to introduce a fluorophore.

Full text of DOI:10.1002/chem.202005115

Accepted Article
Accepted Article
Title: Iminoboronates as dual purpose linkers in chemical probe
development
Authors: Antonie J. (Niek) van der Zouwen, Aike Jeucken, Roy
Steneker, Katharina F. Hohmann, Jonas Lohse, Dirk J.
Slotboom, and Martin Witte
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202005115
Link to VoR: https://doi.org/10.1002/chem.202005115
01/2020

10.1002/chem.202005115
Chemistry - A European Journal
COMMUNICATION
Iminoboronates as dual-purpose linkers in chemical probe
development
A.J. van der Zouwen, [a] Dr. A. Jeucken,[b] R. Steneker,[a] K.F. Hohmann,[a] Dr. J. Lohse,[a] Prof. Dr. D.J.
Slotboom,[b] and Prof. Dr. M.D. Witte*[a]
[a]
A.J. van der Zouwen, R. Steneker, K.F. Hohmann, Dr. J. Lohse, Prof. Dr. M.D.Witte
Chemical Biology II,
Stratingh Institute for Chemistry
Nijenborgh 7, 9747 AG, Groningen, The Netherlands
E-mail:
[b]
Dr. A. Jeucken, Prof. Dr. D.J. Slotboom
Membrane Enzymology,
Groningen Biomolecular Sciences and Biotechnology Institute,
9747 AG Groningen, The Netherlands
Supporting information for this article is given via a link at the end of the document.
Abstract: Chemical probes that covalently modify proteins of interest
are powerful tools for the research of biological processes. Important
in the design of a probe is the choice of reactive group that forms the
covalent bond, as it decides the success of a probe. However,
(CuAAC) reaction, which necessitated the incorporation of a click
handle in one of the probe fragments.
The hydrazone linkage, which is introduced during the synthesis
of the probe, could serve as an alternative bio-orthogonal handle.
Exchange of the ligand for a fluorophore by transimination of the
choosing the right reactive group is not
a simple feat and
methodologies for expedient screening of different groups are needed. hydrazone would overcome the need for an additional bio-
We here report a modular approach that allows easy coupling of a
reactive group to a ligand. α-Nucleophile ligands are combined with
2-formylphenylboronic acid-derived reactive groups to form
iminoboronate probes that selectively label their target proteins. A
transimination reaction on the labeled proteins with an α-amino
hydrazide provides further modification, for example to introduce a
fluorophore.
orthogonal handle. However, attempts to visualize the labeled
proteins in this fashion were unsuccessful in our hands due to
inefficient exchange. Recently, chemistries have been reported
that undergo exchange more readily. In particular, the
iminoboronate chemistry attracted our attention. 2-Formylphenyl
boronic acids (2-FPBA) and pinacol boronate esters thereof have
been shown to form iminoboronates with hydrazides and
alkoxyamines rapidly even at low concentrations, and have been
Protein profiling with chemical probes is commonly employed to
study biological processes in both fundamental and applied
settings.[1–3] The probes used in these profiling studies generally
consist of a ligand that binds to the protein of interest (POI), a
reporter group (e.g. bio-orthogonal handle, fluorophore, affinity
tag) that allows visualization and/or enrichment of the modified
proteins, and a reactive group that covalently links the probe to
the POI.
Over the past decade, many reactive groups have been
developed that can be employed for probe development.[4–7]
Selecting the appropriate reactive group is essential to obtain
highly selective and potent chemical probes. The reactive group
determines which residues are being targeted and it can even
alter the binding profile of the probe.[8] However, determining the
optimal reactive group is, in many cases, not trivial. Often multiple
probes have to be synthesized, for example by multi-component
reactions, and these probes then have to be screened against the
POI. [9,10]
To further simplify the identification of an appropriate reactive
group, we started to explore means in which probe synthesis and
screening could be performed in one simple operation.
Capitalizing on the pioneering work on hydrazone-containing
probes by Hamachi and co-workers,[6] we established a method
to prepare probes by linking different reactive groups to ligands
via hydrazone and oxime chemistry.[11] The resulting chemical
probes could be used without further purification in protein
labeling experiments. The labeled proteins could only be detected
Figure 1. A. Schematic representation of our approach. α-Nucleophile ligands
are coupled with 2-FPBA derived reactive groups to form iminoboronate
chemical probes. After labeling the POI with the probe, the ligand is exchanged
with an α-amino hydrazide to introduce a reporter group. B, C. Structure of R1
(B) and FITC am-zide (C).
reliably by
a copper-catalyzed azide alkyne cycloaddition
1
This article is protected by copyright. All rights reserved.

10.1002/chem.202005115
Chemistry - A European Journal
COMMUNICATION
utilized in bioconjugation reactions.[12–16] The resulting reagents
can be transiminated under physiological conditions,[15,17] and gel-
stable tricyclic diazaborines (DAB) can be obtained by reacting 2-
FPBAs with α-amino hydrazides.[18] We hypothesized that
iminoboronates could serve as a dual-purpose linker (Figure 1).
Probes might be formed by reacting 2-FPBA reactive groups,
such as sulfonyl fluoride R1, with α-nucleophile ligands. The
probe-modified proteins might be detected by transimination of
the iminoboronate linkage with α-amino hydrazide fluorescein
(FITC am-zide) using the stability differences between
iminoboronates as the driving force. We here demonstrate the
feasibility of this approach.
In order to determine the suitability of iminoboronate chemistry for
probe synthesis and visualization, we synthesized sulfonyl
fluoride R1 and reacted it overnight with an equimolar amount of
biotin-hydrazide ligand L1 or non-targeted control hydrazide C1
(Figure 2A). This gave the crude probes, which we used to label
streptavidin (Strp) against a background of two control proteins,
bovine carbonic anhydrase II (CA-II) and ovalbumin (OVA). To
detect the modified proteins, we subjected the samples either to
CuAAC with fluorescein alkyne (Lumiprobe B41B0) or to
transimination with five equivalents of FITC am-zide. Read-out by
exchange of the iminoboronate linker gave a result that is very
similar to CuAAC functionalization of the azido group in ligands
L1 and C1 (Figure 2A). Pronounced labeling of the target protein
Strp was solely observed in samples containing the biotin-based
probe L1R1. Labeling was negligible in samples treated with the
control C1R1. Small differences between the methods were
noticeable. While the fluorescence signal for CA-II and OVA in the
CuAAC samples were comparable to the DMSO control, a small
amount of ligand-independent labeling was visible when
transimination was used as read-out. Nonetheless, these results
clearly indicate that tethering of the ligand to the reactive group
via an iminoboronate linkage and exchange of the ligand by FITC
am-zide are feasible. Labeling experiments on cell lysate of BirA-
overexpressing E. coli with ligands L1-L3, and on CA-II with
sulfonamide ligand L4 further confirmed this (Figure 2B&C).
The background signals in some of the gels prompted us to
optimize transimination protocol. First, we treated CA-II labeled
with L4R1 with varying amounts of reporter group. Approximately
two to three equivalents of FITC am-zide proved to be optimal
when performing the exchange at pH 8.2 (Figure 3A). At this
concentration, the fluorescent signal was still acceptable for the
target protein, while the background signal was minimal. Next, we
investigated how the pH affected the transimination reaction.
Lowering the pH of the samples during the exchange step
increased, as expected,[14] the efficiency (Figure 3B). Compared
to exchange at pH 8.2, the signal intensity for samples treated at
pH 5-6 with FITC am-zide was increased ~1.5 times. Finally, we
assessed the effect of the reaction time on transimination at pH
5.3. Exchange for 15 minutes resulted in a fluorescent signal that
is only 1.15 fold less intense than that obtained after two hours
exchange (Figure 3C). Moreover, extending the incubation to
overnight had a marginal effect on the signal intensity. Similar
trends for the number of equivalents, pH and time-dependency
were observed for BirA (Figure S2). The rapid exchange is a
significant improvement with respect to transimination of non-
iminoboronate linkers, which required twenty-four hours.[6]
Figure 2. A. Proof of concept labeling studies on Strp using biotin ligand L1, control C1 and reactive group R1. A mixture of Strp (25 µM), CA-II (5 µM) and OVA
(25 µM) in HEPES (pH 8.2) was incubated with 20 µM probe for 2 h. Read-out with CuAAC (fluorescein alkyne) or transimination (FITC am-zide). B. Labeling of
BirA in lysate of BirA-overexpressing E. coli (2.5 mg/mL, HEPES pH 8.2) with probes based on ligands L1-L3. Incubation with 20 µM probe for 2 h. Read-out with
FITC am-zide. C. Labeling of CA-II (5 µM) in the presence of OVA (25 µM) and avidin (25 µM) in HEPES (pH 8.2) with probes based on ligand L4. Incubation with
20 µM probe for 2 h. Read-out with FITC am-zide.
2
This article is protected by copyright. All rights reserved.

10.1002/chem.202005115
Chemistry - A European Journal
COMMUNICATION
observed when the probe is formed in the labeling mixture.
Experiments with L3 and R1 and BirA in lysate revealed that the
probes can even be formed in the context of a lysate, although in
this case the labeling was not as efficient as with the preformed
probe.
To demonstrate the suitability of the iminoboronate chemistry in
probe development, we expanded the panel of ligands with
sulfonamides L4-L9 and the panel of reactive groups with epoxide
R2 and arylazide photocrosslinker R3 (Figure 4). The
iminoboronate chemistry allowed straightforward screening of
different combinations of the reactive groups and ligands (Figure
4C, S8-S10) and it could be used to assess the influence of the
linker length on labeling efficiency (Figure S11).
The labeling properties of several of the new probe leads were
studied in greater detail (Figure S12-S14). Time-dependency
studies revealed that maximal labeling was reached within two to
four hours for R1 probes and after overnight incubation for R2
probes. Furthermore, irradiation with 312 nm light for 1 minute
was sufficient to label the target proteins with R3-based
photocrosslinker probes. For L6R1, the most promising CA-II
probe, we also assessed the selectivity and the sensitivity. We
spiked CA-II in E. coli lysate and subjected this to labeling with
L6R1 (Figure 4D). Some background labeling was observed
when we used 20 µM of L6R1. However, lowering the probe
concentration to 100 nM (Figure S15) allowed us to detect as little
as 1−2 ng CA-II in 10 µg lysate, while we did not observe any off-
target labeling in the lysate.
Figure 3. Optimization of the transimination conditions using labeled CA-II. A-
C. The effect of the amount of FITC am-zide (A), the pH (B) and the exchange
time (C) on the intensity of the fluorescent signal in the transimination reaction.
Relative fluorescent intensities were determined by normalizing the signal to 5
equivalents FITC am-zide (A), to pH 8.2 (B), and to 2 hours (C). All experiments
were performed in quadruplicate. D. Comparison of read-out by CuAAC with the
stable acyl hydrazone of R4 and transimination of probes made from R1. E.
Structure of R4. General labeling conditions: 5 µM CA-II or 2.5 mg/mL BirA-
overexpressing E. coli lysate; incubation with 20 µM probe for 2 h.
A head-to-head comparison of transimination and CuAAC
visualization on BirA demonstrated that the signal was similar for
both visualization methods (Figure 3D). Importantly, the
transimination method identified L3R1 as an efficient probe for
BirA, something which could not have been achieved using
CuAAC. Similar results were obtained for Strp, albeit that these
results are less straightforward to interpret due to the monomeric
and tetrameric species present on the gel (Figure S3).
Next, we investigated the effect of the iminoboronate chemistry
on the other steps of the protocol. Binding of the probe to glycans
during the labeling step, which could be induced by the boronic
acid moiety in the probe,[12,19] proved to be negligible, since an
excess of sucrose neither affected labeling of the target nor of the
glycosylated control proteins avidin and OVA (Figure S4). The
small amounts of off-target labeling, which are observed in all
experiments, could be suppressed partly by increasing the ligand-
to-R1 ratio (Figure S5). This observation suggests that hydrolysis
of the iminoboronate linkage may liberate a small amount of R1,
which reacts in a ligand-independent fashion.
The iminoboronate chemistry proved to be beneficial for probe
formation. Comparison of the labeling intensity of L4R1 formed
overnight with that of L4R1 prepared by mixing L4 and R1 prior
to the experiment revealed that the probes were formed within 15
minutes (Figure S6). Excitingly, the iminoboronate chemistry
allows the formation of probes in the presence of proteins. Already
30 minutes after adding ligand L4 and reactive group R1 to a
protein mixture, we observed fluorescent labeling of CA-II, albeit
not as high as the signal for the preformed L4R1 probe (Figure
S7). Importantly, the differences in signal intensities of the POI
disappeared when we increased the labeling time to 5 h. Only a
minimal increase in labeling by the non-targeted C1R1 is
Figure 4. A. Structures of sulfonamide ligands L4-L9, reactive groups R2 and
R3. Prior to probe formation, the trifluoroborate salt of R3 is converted into the
boronic acid by incubation in aqueous solution. B. Labeling of CA-II (5 µM) with
the new probes (20 µM). R3 is activated by irradiation at 312 nm for 15 minutes.
C. The indicated amount of CA-II was spiked in E. coli lysate (2 mg/mL) and
labeled with L6R1 (100 nM) in order to determine the sensitivity of the
iminoboronate probe.
3
This article is protected by copyright. All rights reserved.

10.1002/chem.202005115
Chemistry - A European Journal
COMMUNICATION
Finally, we investigated if the iminoboronate probes could be
applied on cells using BioY[20] as a model protein. Our group
In conclusion, we have shown that α-nucleophile ligands can be
coupled successfully to 2-FPBA reactive groups and the resulting
iminoboronate probes can be used without further purification to
label proteins in complex mixtures. Transimination with an α-
amino hydrazide reporter group enables reliable read-out of the
labeled proteins. Due to the fast reaction kinetics and the bio-
orthogonality of iminoboronate formation, the probes can even be
prepared in the context of a protein mixture. Due to the modular
nature of the iminoboronate linker, this methodology should be
widely applicable with any ligand that can be converted into an α-
nucleophile. Therefore, we believe that using iminoboronate
probes can greatly expedite the screening of reactive groups
towards the identification of new chemical probes.
previously employed
a His-tagged version of this biotin
transporter and the mutants BioY-N79K and BioY-R93K, which
carry an additional lysine near the binding site, to study on-cell
labeling of targeted diazotransfer reagents.[21] To perform the on-
cell labeling studies, we first had to identify an iminoboronate
probe for BioY. Therefore, we prepared a library of probes based
on ligands L1-L3 and reactive groups R1-R3. To increase the
chances further, we included two additional ligands (L10 and L11,
Figure 5A). We screened the library on membrane fractions of
Lactococcus lactis cells that either overexpressed BioY-wt or the
mutants BioY-N79K and BioY-R93K. As a control, we took the
membrane fraction of non-induced BioY-wt cells. For several of
the probes, a fluorescent signal at ~20 kDa, which corresponds
with the molecular weight of BioY, was observed. This signal was
absent in the non-induced samples and was less pronounced for
the non-targeted probes (Figure S16). Incubation of BioY-R93K
lysate with the probe L10R1 resulted in the most pronounced
labeling. Labeling of this protein by L10R1 was blocked in the
presence of an excess of biotin (Figure 5B), confirming that
labeling is ligand dependent and that BioY-R93K is being labeled.
We therefore used L10R1 for the labeling of BioY on L. lactis cells.
For the cell labeling, we added L10R1 to a suspension of either L.
lactis overexpressing BioY-R93K or, as a control, non-induced
BioY-wt in HEPES (pH 8.0) and left it to react for two hours. We
removed the excess of probe by washing the cells with buffer prior
to on-cell transimination with FITC am-zide at pH 5.3. To analyze
the labeled proteins, we lysed the cells, isolated the membrane
fraction and subjected it to SDS-PAGE. Analysis of the in-gel
fluorescence revealed that BioY-R93K had been successfully
labeled with L10R1, while limited or no labeling of BioY was
observed for the control C1R1 and for the cells where expression
had not been induced (Figure 5C). We also observed a
fluorescent band at high molecular weight in both samples,
indicating off-target labeling. Promisingly, the labeling of BioY-
R93K indicates that chemical probes containing an iminoboronate
linker can be used in cell labeling studies.
Acknowledgements
The Netherlands Organisation for Scientific Research (NWO-
VIDI) grant number 723.015.004 (M.D.W., A.J.V.D.Z.) is
acknowledged for funding.
Keywords: chemical probes • iminoboronate • bio-orthogonal
chemistry • protein labeling • affinity-based probes
[1]
[2]
[3]
[4]
B. F. Cravatt, A. T. Wright, J. W. Kozarich, Annu. Rev. Biochem. 2008,
77, 383–414.
Y. Takaoka, A. Ojida, I. Hamachi, Angew. Chem. Int. Ed. 2013, 52, 4088–
4106.
L. I. Willems, H. S. Overkleeft, S. I. Van Kasteren, Bioconjug. Chem.
2014, 25, 1181–1191.
Y. Liu, M. P. Patricelli, B. F. Cravatt, Proc. Natl. Acad. Sci. U. S. A. 1999,
96, 14694–14699.
[5]
[6]
A. Narayanan, L. H. Jones, Chem. Sci. 2015, 6, 2650–2659.
Y. Takaoka, H. Tsutsumi, N. Kasagi, E. Nakata, I. Hamachi, J. Am. Chem.
Soc. 2006, 128, 3273–3280.
[7]
[8]
[9]
Z. Li, P. Hao, L. Li, C. Y. J. Tan, X. Cheng, G. Y. J. Chen, S. K. Sze, H.-
M. Shen, S. Q. Yao, Angew. Chemie Int. Ed. 2013, 52, 8551–8556.
E. Weerapana, G. M. Simon, B. F. Cravatt, Nat. Chem. Biol. 2008, 4,
405–407.
J. Ge, X. Cheng, L. P. Tan, S. Q. Yao, Chem. Commun. 2012, 48, 4453.
[10] J. T. Bush, L. J. Walport, J. F. McGouran, I. K. H. Leung, G. Berridge, S.
S. van Berkel, A. Basak, B. M. Kessler, C. J. Schofield, Chem. Sci. 2013,
4, 4115.
[11] A. J. van der Zouwen, J. Lohse, K. F. Hohmann, L. H. E. Wieske, R. van
der Vlag, M. D. Witte, Chem. Commun. 2019, 55, 2050–2053.
[12] P. M. S. D. Cal, J. B. Vicente, E. Pires, A. V. Coelho, L. F. Veiros, C.
Cordeiro, P. M. P. Gois, J. Am. Chem. Soc. 2012, 134, 10299–10305.
[13] N. J. Gutiérrez-Moreno, F. Medrano, A. K. Yatsimirsky, Org. Biomol.
Chem. 2012, 10, 6960.
[14] P. Schmidt, C. Stress, D. Gillingham, Chem. Sci. 2015, 6, 3329–3333.
[15] A. Bandyopadhyay, J. Gao, Chem. Eur. J. 2015, 21, 14748–14752.
[16] K. Li, W. Wang, J. Gao, Angew. Chemie Int. Ed. 2020, 59, 14246–14250.
[17] H. Gu, S. Ghosh, R. J. Staples, S. L. Bane, Bioconjug. Chem. 2019, 30,
2604–2613.
[18] O. Dilek, Z. Lei, K. Mukherjee, S. Bane, Chem. Commun. 2015, 51,
16992–16995.
[19] E. Nakata, T. Nagase, S. Shinkai, I. Hamachi, J. Am. Chem. Soc. 2004,
126, 490–495.
Figure 5. A. Biotin-ligands L10 and L11. B. Labeling of membrane extracts (2
mg/mL) overexpressing BioY-R93K with L10R1 in the presence and absence
of biotin. Incubation with 20 µM probe for 2 h; Read-out: FITC am-zide. C. In-
gel fluorescence of L. lactis cells that were not induced or overexpressing
mutant BioY-R93K. Cells were labeled for 2 h with the indicated probe and the
ligand was subsequently exchanged on the cell surface by FITC am-zide.
[20] P. Hebbeln, D. A. Rodionov, A. Alíandega, T. Eitinger, Proc. Natl. Acad.
Sci. U. S. A. 2007, 104, 2909–2914.
[21] J. Lohse, L. J. Y. M. Swier, R. C. Oudshoorn, G. Médard, B. Kuster, D.
J. Slotboom, M. D. Witte, Bioconjug. Chem. 2017, 28, 913–917.
4
This article is protected by copyright. All rights reserved.

10.1002/chem.202005115
Chemistry - A European Journal
COMMUNICATION
Entry for the Table of Contents
Insert graphic for Table of Contents here.
Iminoboronate chemistry enables the synthesis of chemical probes that selectively label proteins of interest from 2-
formylphenylboronic acid reactive groups and hydrazide-functionalized ligands. The reversibility of the iminoboronate linkage is
exploited to detect the labeled proteins by ligand-fluorophore exchange. Remarkably, the fast iminoboronate chemistry even allows
probe formation within protein mixtures.
Institute and/or researcher Twitter usernames:@NiekZouwen @StratinghInst
5
This article is protected by copyright. All rights reserved.