New Methods for the Site-Selective Placement of Peptides on a Microelectrode Array: Probing VEGF-v107 Binding as Proof of Concept

Article abstract of DOI:10.1021/acschembio.6b00685

Cu(I)-catalyzed click reactions cannot be performed on a borate ester derived polymer coating on a microelectrode array because the Cu(II) precursor for the catalyst triggers background reactions between both acetylene and azide groups with the polymer surface. Fortunately, the Cu(II)-background reaction can itself be used to site-selectively add the acetylene and azide nucleophiles to the surface of the array. In this way, molecules previously functionalized for use in click reactions can be added directly to the array. In a similar fashion, activated esters can be added site-selectively to a borate ester coated array. The new chemistry can be used to explore new biological interactions on the arrays. Specifically, the binding of a v107 derived peptide with both human and murine VEGF was probed using a functionalized microelectrode array.

Full text of DOI:10.1021/acschembio.6b00685

Articles
pubs.acs.org/acschemicalbiology
New Methods for the Site-Selective Placement of Peptides on a
Microelectrode Array: Probing VEGF−v107 Binding as Proof of
Concept
Matthew D. Graaf, Bernadette V. Marquez, Nai-Hua Yeh, Suzanne E. Lapi,* and Kevin D. Moeller*,†
†
‡
†
,†,‡
†Department of Chemistry, Washington University in St. Louis, St. Louis, Missouri 63130, United States
‡Department of Radiology, Washington University in St. Louis, St. Louis, Missouri 63110, United States
*
S Supporting Information
ABSTRACT: Cu(I)-catalyzed “click” reactions cannot be per-
formed on a borate ester derived polymer coating on a
microelectrode array because the Cu(II) precursor for the catalyst
triggers background reactions between both acetylene and azide
groups with the polymer surface. Fortunately, the Cu(II)-back-
ground reaction can itself be used to site-selectively add the
acetylene and azide nucleophiles to the surface of the array. In this
way, molecules previously functionalized for use in “click” reactions
can be added directly to the array. In a similar fashion, activated
esters can be added site-selectively to a borate ester coated array. The new chemistry can be used to explore new biological
interactions on the arrays. Specifically, the binding of a v107 derived peptide with both human and murine VEGF was probed
using a functionalized microelectrode array.
icroelectrode arrays offer a novel platform for analyzing
the biological activity of a wide variety of molecules.
arrays offer an opportunity to simplify the analysis of medium
to larger molecular libraries by enabling their synthesis directly
on the device used for their analysis, to then analyze the
libraries over extended periods of time while new members are
added or built on the surface, and then characterize the
molecules made in order to make sure their structures are
correct both immediately after their synthesis and over time.
Three key components underlie every site-selective reaction
run on an array. The first is a porous polymer surface that coats
the array and provides attachment points for fixing molecules to
1
−12
M
They are particularly intriguing because binding events between
molecules fixed on the surface of an array and a biological target
can be monitored with the use of a benign redox mediator
13,14
(
usually and Fe(II)/Fe(III) couple).
The couple is oxidized
at the Pt electrodes in the array and reduced again at a remote
Pt counter-electrode, and the resulting current monitored at
each electrode in the array. When a binding event happens on
the surface of the array, the conductivity of the surface is
altered, leading to a change in current at the associated
electrode. This change in current is detected and its magnitude
used to judge the extent of the binding interaction. In this way,
the interaction is monitored in “real-time” without any need to
label the receptor, label the peptide, or develop an
immunological assay with labeled antibodies.
18
the surface of the electrodes. This polymer must be stable
over time, inert to a wide variety of chemical reagents, and
compatible with the electrochemical signaling studies. It was
this need for chemical inertness and long-term stability that led
us to choose polymer platforms for coating the surface and in
so doing avoid the less-stable self-assembled monolayer
However, what makes the use of microelectrode arrays truly
unique relative to alternative approaches is the synthetic
capability of the arrays. A wide variety of site-selective reactions
can be conducted at individual electrodes in an array, even
when the array has thousands of electrodes per square
19
platform used so successfully with alternative methods. The
second is a redox stable chemical reagent that can exist in an
inert form over the arrays and then be activated by the
electrodes in the array. Because of the use of the more stable
polymer platform for building a surface on the arrays, direct
electrochemical methods for modifying the array are not
1
3,15−17
centimeter.
Electrodes can be used to synthesize
oxidants, reductants, acids, bases, Lewis acids, nucleophiles,
electrophiles, and transition metal reagents and catalysts. In this
way, much of modern organic synthesis can be used to place,
construct, or manipulate molecules on the surface of an array.
These reactions can also be used to site-selectively recover
molecules built on the arrays so that they can be characterized
20
possible. While this initially appears problematic, the use of
indirect methods in which the electrode generates a chemical
reagent that in turn conducts reactions on the surface in
practice offers tremendous advantages. As mentioned earlier,
13
following a biological analysis. In this way, the synthetic
capabilities of the arrays allow for quality control analysis of an
array based library. Because of these combined capabilities, the
Received: August 9, 2016
Accepted: August 24, 2016
©
XXXX American Chemical Society
A
DOI: 10.1021/acschembio.6b00685
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
electrodes can be used to make a wide variety of chemical
reagents and catalysts. To date, Pd(0), Pd(II), Cu(I), Cu(II),
Scheme 1. Placement of Thiol Nucleophiles on a Borate
Ester Surface Using a Chan-Lam Coupling Procedure
Ce(IV), DDQ, Os(VIII), Ru(VII), triarylamine radical cations,
+
H , methoxide, H , and Sc(III) have all been used to conduct
2
syntheses on an array. Of course for these reactions to be
conducted at selected electrodes in the array, a third key
component is needed for each reaction. That component is a
confining agent that consumes the reagent generated at the
electrodes and returns it to its inactive solution phase state so
that the reaction cannot happen on the array at sites remote
13
Chan-Lam reaction only occurred at the desired electrodes.
The chemistry was very successful and allowed for peptides
functionalized for use in thiol-Michael approaches to be utilized
directly on the arrays.
With our first complementary method in place, we turned
our attention back to the development of an array-based variant
of the Huisgen-type “click reaction.” Unfortunately, initial
efforts to place molecules functionalized for use in the [3 + 2]-
cycloaddition onto arrays coated with the new, more desirable
diblock copolymer surface were not successful. The initial plan
was to simply move the previous array-based reaction from the
agarose surface (Scheme 2) to the diblock copolymer surface.
from the selected electrodes.
With the capability of running almost any chemical reaction
site-selectively on an array now in place, we turned our
attention toward the development of array-based variants of
transformations that are commonly employed in the con-
struction of bioconjugates. The goal of the work is to ensure
that the arrays can be used to complement other analytical
methods without any need to resynthesize the molecules being
evaluated. For example, peptides already functionalized with
21,22
acetylenes for use in Huisgen-type “click reactions”
can
potentially be site-selectively placed onto a microelectrode array
that is coated with the type of stable, chemically inert surface
that would allow it to be monitored over extended periods of
time. Peptides functionalized for use in Diels−Alder reaction
based routes to forming bioconjugates can be site-selectively
Scheme 2. Site-Selective Click Reaction on an Agarose
Coated Array
23
placed on an array with the same Diels−Alder strategy.
Of these methods, the use of a “click reaction” is particularly
attractive because of the reaction broad applicability in chemical
biology. For this reason, we developed a site-selective variant of
the reaction that coupled molecules with an acetylene to an
24,25
azide on an agarose coated array.
In this case, the agarose
surface served as a “practice” coating on the arrays for
developing the confinement strategy needed for the site-
selective reactions. Agarose is not stable and hence can be
removed very easily from an array. This is wonderful for reusing
the arrays, but impractical if one wants to build a stable surface
that enables signaling studies. For this reason, a more stable and
very versatile diblock copolymer surface has been developed for
18
use on the arrays (Figure 1).
In this previous study, an acetylene precursor was placed onto
24
the agarose surface by each of the electrodes in the array. The
3 + 2] cycloaddition between the surface bound acetylene and
[
an azide in solution was then accomplished by using selected
electrodes in the array to reduce a Cu(II) precursor to the
Cu(I) catalyst needed for the reaction. Air was used as the
“
confining agent” to oxidize any Cu(I) catalyst that migrated
away from the electrodes employed for the reaction. A
fluorescent tag was placed on the azide so that the quality of
the reaction could be assessed with the use of fluorescence
microscopy. For the reaction, a checkerboard pattern of
electrodes was used in order to probe the success of the
reaction over the entire surface of the array. As can be seen in
the image provided, the reaction and the confinement strategy
both worked well. The cycloaddition also proceeded nicely
when an azide was placed on the array and an acetylene used in
solution.
Figure 1. Diblock copolymer used to coat the microelectrode arrays.
This surface is ideal because it is stable for extended periods
of time, compatible with a wide variety of both synthetic and
analytical experiments, and tunable so that nonspecific binding
interactions with biological molecules can be minimized.
Our first strategy for placing peptides onto the borate ester
surface took advantage of a thiol nucleophile and the well-
26,27
RESULTS AND DISCUSSION
known Cu(II)-mediated Chan-Lam reaction (Scheme 1).
■
The Cu(II) needed for the reaction was generated by the
oxidation of a Cu(I) precursor at the electrodes selected for the
experiment. Excess thiol in solution was used as the confining
agent for the reaction by reducing any Cu(II) that migrated
away from the selected electrodes in the array. Hence, the
Surprisingly, the same reaction run on an array coated with the
borate-ester based surface led to the images shown in Scheme
3. For this experiment, a thiol substituted azide was added to
the surface of each electrode in the array using the Chan-Lam
coupling reaction illustrated in Scheme 1. An acetylene labeled
B
DOI: 10.1021/acschembio.6b00685
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Scheme 3. Initial Attempt at a Cu(I)-Catalyzed Click
Reaction with a Borate Ester-Base Polymer Surface on the
Array
nucleophile (Nu-H) adds to the Cu(II) reagent, which then
undergoes transmetalation with the arylborane prior to
oxidation and a reductive elimination reaction. It was certainly
possible that an acetylene might serve as the nucleophile for
this reaction.
This suggestion was easy to test by taking advantage of the
Cu(II) confinement strategy already developed for the thiol-
based Chan-Lam coupling reaction. To this end, an array
coated with the borate ester polymer was treated with a DMF
solution containing Cu(I) (generated from the reduction of
28,29
CuSO by the acetylene in solution),
an excess of the
4
pyrene labeled acetylene, and Bu NBF as an electrolyte. Blocks
4
4
of 12 electrodes (again, the pattern needed for subsequent
signaling studies) each were then used as anodes in order to
site-selectively generate Cu(II) from the Cu(I) in solution. The
resulting array was examined by fluorescence microscopy. As
shown in Scheme 5, the experiment did result in the site-
with a pyrene group was used as the solution phase substrate.
As can be seen, the reaction did not show the expected
fluorescent spot by each electrode used for the “click reaction.”
In this case, the pattern used on the arrays would have placed
the molecules down in blocks of 12 electrodes each (the
pattern needed for subsequent signaling studies). Instead, the
surface of the array surrounding the electrode was fluorescent
and the electrodes dark. Examination of the array using a green
filter for the fluorescent light (a wavelength of light that is not
ideal for the pyrene exciplex and hence only shows up with very
low concentrations of the dimer) showed that the reaction did
occur at a faster rate closer to the electrodes used for the
generation of Cu(I), but it did not occur on the electrode. So
what makes the “click reaction” on the borate ester surface so
different, and how do we solve this problem so that we can best
take advantage of peptide ligands already modified for “click
reaction” based applications?
One problem that can occur during an array-based reaction is
a very rapid background reaction between the fluorescent
substrate in solution and the unfunctionalized polymer coating
the array. Such reactions occur everywhere on the array except
where the solid phase substrate has taken the place of the
borate esters on the polymer. The result is a fluorescent surface
with “holes” over the electrodes just like the images shown in
Scheme 3. Three questions immediately arise about this
suggestion. What is the background reaction? Why did it
occur at a faster rate closer to the electrodes being used as
cathodes? And why did the desired Cu(I)-catalyzed [3 + 2]-
cycloaddition not occur on the electrodes?
Scheme 5. Addition of an Acetylene Nucleophile to the
Surface of the Array
selective placement of the acetylene nucleophile proximal to the
electrodes used for the oxidation. Since no reaction was
observed on the array remote from the electrodes used, the
reaction is not catalyzed by the Cu(I) precursor in solution, nor
is it the result of a “non-specific” interaction between the
solution phase substrate and the polymer surface. In this way,
the array reaction serves as its own control. Clearly, the reaction
only occurs in the presence of the Cu(II) mediator.
Support for the reaction shown in Scheme 5 was gained by
examining a similar solution phase reaction. In this case, the
deprotected boric acid (1) was used in place of the borate ester
because the boric acid moiety will undergo a Cu(II) catalyzed
reaction more quickly. This is important for the solution phase
reaction because array reactions are typically greatly accel-
Alternative Copper-Based Peptide Placement Reac-
tions. One possible answer to the first question is that the
Cu(II) added as a precursor for the reductive generation of
Cu(I) catalyzed a Chan-Lam type coupling between an
acetylene nucleophile and the borate ester surface. In the
mechanism for the Chan-Lam coupling reaction (Scheme 4), a
13
erated. In an array reaction, the “catalyst” is generated at the
surface of the electrode in large excess relative to the substrate
present. Suzuki and Heck reactions that take hours at reflux in
solution run in minutes on an array. Many of the previous array
reactions using the borate ester surface only proceed in solution
when the borate ester is deprotected. In fact, this is the case for
the reaction shown in Scheme 5. While a solution phase
reaction with a borate ester substrate led to no reaction, the
faster solution phase reaction with the deprotected boric acid
substrate proceeded nicely (Scheme 6). Two things about the
reaction are particularly relevant to the array reaction shown in
Scheme 5. First, the reaction did lead to the desired coupling
product (2, NMR yield based on the acetylene limiting reagent)
with the Cu(II) catalyst mediating the coupling of the acetylene
to the arylboronate. Second, the major product from the
reaction was an expected homocoupling product from the
acetylene (3, NMR yield based on the acetylene limiting
Scheme 4. Chan-Lam Reaction
C
DOI: 10.1021/acschembio.6b00685
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Scheme 6. A Solution-Phase Acetylene Coupling Reaction
Scheme 7. Addition of an Azide Nucleophile to the Surface
of the Array
2
8,29
reagent).
This reaction not only leads to the initial
generation of Cu(I) in solution (for oxidation at the anodes
in the array) but also is the reaction behind why the array
reaction in Scheme 5 was so well confined to the selected
electrodes. For confinement on an array, the reagent generated
at the electrodes must be quickly consumed in solution before
it can migrate to remote sites on the array. The oxidative
homocoupling reaction of the acetylene substrate accomplishes
this task perfectly because it reduces the copper(II) reagent
generated at the anodes in the array. Hence, the solution phase
reaction demonstrates both of the reactions necessary for a
successful array reaction. As a sidenote, a third product from
water addition to the excess aryl boronate used in the reaction
was also observed (4, NMR yield based on the aryl boronate).
This product is the result of the starting boronic acid being wet.
The success of the array reaction shown in Scheme 5 was
intriguing because it indicated that an acetylene substituted
peptide can be placed directly on the surface of a micro-
electrode array without any need to prefunctionalize the array
with the azide. In this way, the failure of the reaction shown in
Scheme 3 led to an even easier method for the placement of an
acetylene functionalized peptide onto an array (Scheme 5). As a
sidenote, the new method is also advantageous over the
previously developed Diels−Alder methods that also require
oligomers are oxidized to Pd(II) at the anode and then the
Pd(II) species generated reduced back to new, monomeric
Pd(0) catalyst at the cathode. The result is constant
regeneration of the active Pd(0) catalyst. If the reduction of
Cu(II) to Cu(I) at the cathodes selected for a “click reaction” is
followed by rapid reoxidation of the Cu(I) species by oxygen to
generate a more reactive Cu(II) reagent, then the Cu(II)
background reaction would occur more rapidly close to the
electrodes. Of course, more work needs to be done to
demonstrate if this is indeed the case.
Revisiting the Click Reaction. The third question asked
about the images shown in Scheme 3 wondered why there was
no fluorescence on the electrodes themselves. In spite of the
background reaction, the electrodes were used as cathodes to
generate Cu(I), and it was surprising that the expected [3 + 2]
cycloaddition reaction did not occur at those locations. With
the success of the reaction illustrated in Scheme 7, it was fair to
ask if the azide competed with the thiol nucleophile during the
Chan-Lam procedure used to make the starting array. If so,
then there would be no azide present on the surface of the
electrode and no subsequent cycloaddition.
If the cysteine derivative had been placed on the array with
the azide instead of the thiol, then there should be free thiol on
the surface of the electrodes. This possibility was tested. Thiols
can be detected on the surface of an array by treatment of the
array with Cu(II) and a solution phase thiol that is labeled with
a fluorescent group. The Cu(II) reaction catalyzes the
formation of a dithiane that fluorescently labels the surface of
the electrode. Two examples are shown in Scheme 8. In the
first (a), a bis-thiol compound is placed on the array (in an S-
pattern for sulfur) using the chemistry shown in Scheme 1, and
then the free thiol of the surface labeled as a control to show
that the reaction works. In the second (b), the experiment is
repeated with the cysteine derivative containing the azide. No
fluorescence was observed when the cysteine derivative was
used, an observation that indicated that the thiol was indeed the
nucleophile in the Chan-Lam reaction in spite of the presence
of the azide. So why was there no click reaction on the
electrodes in Scheme 3?
20,23
prefunctionalization of the array with a dienophile or diene.
The success of the reaction with the acetylene led us to
wonder if the azide partner for the “click reaction” might also
serve as a nucleophile for the Chan-Lam coupling reaction.
Similar solution phase reactions have been shown to afford aryl
3
0
amines. To this end, Cu(II) was site-selectively generated by
an N-pattern of electrodes in a microelectrode array (the N was
used for “nitrogen” so that the array could be distinguished
from the one shown in Scheme 5) coated with the borate ester
polymer. The reaction conditions were identical to those used
in Scheme 5 except for the use of a pyrene labeled azide as the
nucleophile (Scheme 7). Once again, the reaction worked
nicely and led to the selective placement of the pyrene by each
electrode used for the reaction.
The second question asked about the images shown in
Scheme 3 is not easy to answer. Since the background reaction
is clearly catalyzed by Cu(II) and not Cu(I), why did it occur
more rapidly close to the electrodes used to reduce Cu(II) to
Cu(I)? We can only speculate on a reason for this observation.
It is known that passing current through an organometallic
A second possibility was that the Cu(I) precursor for Cu(II)
in the array-based Chan-Lam reaction reduced the azide to an
amine following addition of the thiol to the array. The result
would be a free amine on the surface of the array and no azide
available for the click reaction. To test this possibility, an azide
substituted cysteine was placed onto the array using the Chan-
Lam coupling conditions using a “block pattern” of electrodes,
and then the resulting array incubated with a fluorescently
31
reaction can dramatically increase the rate of the reaction.
This is thought to occur by the continual regeneration of fresh
catalyst by the reduction and oxidation reactions that take place
at the electrodes. For example, Pd(0) catalyzed reactions often
suffer from oligomerization of the catalyst. When current is
passed through the reactions, it is thought that the Pd(0)
32
labeled acid fluoride which readily reacts with amines. The
reaction did label the electrodes used for the Chan-Lam
D
DOI: 10.1021/acschembio.6b00685
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Scheme 8. (a) Control Experiment Showing the Detection of
a Thiol on the Surface and (b) Chan-Lam Coupling with the
Azide Functionalized Cysteine Leading to No Free Thiol
reaction used to place the substrate on the array and not
necessarily a problem with the reaction itself. For this reason, a
site-selective Suzuki reaction was used to place an azide on the
33
array (Scheme 11). The entire array was then treated with a
Scheme 11. A “Click-Reaction” Conducted on a Borate Ester
Coated Array
reaction, indicating the presence of a reactive amine at those
sites (Scheme 9).
Scheme 9. Detection of an Amine
DMF/CH Cl2 solution that contained CuSO , a ligand
2
4
developed for optimization of the [3 + 2]-cycloaddition
34
reaction, electrolyte, and a fluorescently labeled acetylene.
Electrodes in a “C-pattern” (C- for “click” to distinguish the
reaction from the other azide and acetylene placement
reactions) were used as cathodes to reduce the Cu(II) to
Cu(I) and trigger the “click reaction.” Air was used as the
confining agent for the Cu(I) generated. As can be seen in the
image provided, the reaction worked very well, and the product
was generated exclusively at the electrodes used. Clearly, the
The presence of the amine readily explained the failure of the
click reactions” in Scheme 3. Reduction of the azide meant
that there was no surface bound substrate for the reaction. It
also suggested that the Chan-Lam coupling reaction using the
azide substituted cysteine proceeded through a pathway like the
one illustrated in Scheme 10. In this mechanism, the Cu(I)
“
“
click reaction” can be used to add molecules to the borate
ester surface as long as the Chan-Lam reaction is avoided as a
means to place the azide substrate onto the array.
Direct Esterification of the Borate Ester Surface. Since
pinacol-protected borate ester surfaces undergo reactions with
water to form boric acid groups and alcohols, we wondered if
these groups could be directly coupled to the C-terminus of a
peptide through an esterification reaction. This would provide
an alternative method for placing an existing peptide on the
arrays. For such reactions, the C-terminus of the peptide is
typically converted to an activated ester and then coupled to an
alcohol in the presence of a base catalyst. In past array
reactions, N-succinimide esters have been used to site-
selectively place peptides onto agarose and sucrose coated
Scheme 10. Mechanism of Azide Reduction and Cysteine
Placement on the Array
14
arrays with the use of an electrogenerated base.
The possibility of the same chemistry being used on the new
borate ester surface was tested by treating a borate ester coated
array with a pyrene labeled N-hydroxysuccinimide ester
(Scheme 12). As in earlier array esterification reactions, the
coupling of the NHS-ester to the surface of selected electrodes
in the array was accomplished by using the electrodes to reduce
vitamin B12 and in so doing generating a base. Tetramethy-
lammonium nitrate was used as the electrolyte for the reaction,
and blocks of 12 electrodes each were used for the reduction.
As image a shows in Scheme 12, the coupling reaction did
occur at the selected electrodes. However, the esterification lost
confinement, and fluorescence can be also be observed at
would undergo an oxidative addition to the azide followed by a
ligand exchange to place the thiol onto the resulting Cu(II)
intermediate. Transfer of the thiol to the borated ester surface
would lead to the free amine on the surface of the electrode.
We have found that Chan-Lam reactions with thiol
nucleophiles proceed much faster than similar reactions with
2
7
amines.
At this point, it became clear that the failure of the “click
reaction” in Scheme 3 was due to problems with the Chan-Lam
E
DOI: 10.1021/acschembio.6b00685
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Scheme 12. Esterification Reactions: (a) No Added Acid and
these peptides in mouse models of angiogenesis, especially
when both VEGF homologues are present in vivo.
(b) with Toluene Sulfonic Acid
In principle, the chemistry developed above makes it easy to
probe the binding of v107 analogs with both VEGF receptors.
The acetylene modified v107 peptide shown in Scheme 13 is
Scheme 13. Placement of a v107−Peptide on a
Microelectrode Array
neighboring regions of the array. The loss of confinement
typically occurs when generation of the reagent at the
electrodes overwhelms neutralization of the reagent in the
solution above the array. The problem can be resolved by
increasing the concentration of the confining agents. In the
current experiment, the reagent generated at the electrodes is a
base. Therefore, acid is needed as the confining agent. When
the concentration of acid above the array was increased with
the addition of toluene sulfonic acid, the esterification was
nicely confined to the electrodes used for the generation of base
desirable for its compatibility with making bioconjugates in
connection with the development of PET imaging agents where
“click reactions” are used to add radiometal binding ligands or
radiolabeled prosthetic groups to the peptide targeting group.
Using the chemistry highlighted in Scheme 5, this same
molecule can be placed by specific electrodes in an addressable
array, and then the electrodes used to monitor the binding
interactions in “real-time” without any need to utilize a labeled
receptor or develop an immunological assay or develop a
separate coupling strategy. The surfaces used on the arrays are
stable, so new peptide analogs of v107 can be added to the
array for comparison with the original derivatives, and in so
doing an addressable library of v107 analogs assembled.
(
Scheme 12b). In this way, the NHS esters developed for use
on sugar based surfaces could be employed directly on the
borate ester coated array.
Probing VEGF−v107 Binding. The reactions developed
above are important because they expand the utility of the
arrays as platforms for evaluating the biological activity of
molecules. For example, one problem that is of interest to us is
the interaction between the v107-peptide analog, v107 L19K
However, while such studies are easy to propose, the proof-
of-principle analytical studies conducted on the arrays to date
have all utilized binding interactions between molecules and
13−15
receptors with nanomolar affinities.
So, how useful are the
arrays for detecting weaker interactions? The v107−VEGF
interaction provides a perfect opportunity to answer this
question.
(
GGNECDIARMWEWECFERK-NH /cysteines are bridged
2
with a disulfide bond) and its vascular endothelial growth
35−37
factor (VEGF) target.
VEGF is important in tumor
To this end, an acetylene functionalized v107 peptide
(custom synthesized by CPC Scientific) was placed proximal
to 10 blocks of 12 electrodes each in an array that contained
angiogenesis and other disease states, and in many cases
overexpression of VEGF can be correlated to various stages of
disease with higher levels of VEGF being correlated to later
2
12 544 electrodes/cm (Scheme 13). A second set of 10 blocks
38
stages of the disease. The v107-peptide and its L19A analog
where the leucine at position 19 is replaced with an alanine are
of 12 electrodes each was functionalized with a linear
hexapeptide (DRDGSP) that served as a negative control.
This peptide was placed on the array using the serine side-chain
alcohol group using the Chan-Lam coupling reaction described
above. Alcohol and thiol nucleophiles have been shown to work
35,36
known weak ligands (K ca. 1 μM) for human VEGF.
d
Previous work has improved this binding interaction via
covalent binding to a specific lysine residue on VEGF, which
serves as a potential lead compound for the development of
Positron Emission Tomography imaging agents that will
27
equally well in this chemistry. In addition, 10 blocks of
electrodes that were coated only with the borate ester surface
were selected for comparison with the functionalized surfaces.
It should be noted that the acetylene functionalized v107
peptide used for the chemistry shown in Scheme 13 is a much
more complex substrate than the substrate used in Scheme 5,
and it is possible that the coupling reaction did not exclusively
involve the acetylene and the borate on the surface. However,
the binding properties of the array placed peptide in the studies
below are consistent with coupling of the molecule to the
surface at a location in the molecule remote from the v107
peptide sequence needed to bind the VEGF active site, and
therefore the biological studies themselves provide evidence
that the coupling reaction proceeded as planned. In addition, it
should be noted that the binding studies also provide evidence
37,39
selectively target VEGF expression.
This technique requires
substituting v107 L19 with a lysine (L19K) followed by
conjugating with amine-reactive cross-linkers for covalent
binding to a specific lysine residue on VEGF. To date, the
binding affinity of v107 L19K without the covalent cross-linker
has not yet been reported. It would be nice to know this affinity
so that v107 L19K could serve as a baseline peptide for
developing higher affinity analogs that did not possess the
covalent cross-linker. Additionally, the binding epitope to which
the peptide binds appears conserved in human and murine
VEGF-A homologues (please see the Supporting Information).
Determining binding affinities of v107 analogs to both VEGF
homologues could be useful in evaluating the biodistribution of
F
DOI: 10.1021/acschembio.6b00685
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
that the disulfide bridge in the v107 is intact. This is not a
surprise. While the disulfide linkage may be reduced by the
Cu(I)-precursor employed for the Chan-Lam reaction (as was
the azide in Scheme 9), such a reduction is reversible. So both
the Cu(II) generated at the anode during the Chan-Lam
reaction and the net oxidative environment of the signaling
experiment where the array is used as an anode would reform
the disulfide bridge and ensure its presence during the
analytical experiment.
Following placement of the peptide on the array, the array
was treated with a 1×PBS buffer solution containing 8 mM
potassium ferricyanide/potassium ferrocyanide in a 1:1 ratio. A
CV was run at the electrodes in the array and then the solution
replaced. This process was repeated three times until the
polymer surface on the array was saturated with the redox
mediator and the current stabilized. At this point, the array was
treated with the 1×PBS buffer solution containing 8 mM9
potassium ferricyanide/potassium ferrocyanide and 5.0 × 10−
M of the murine VEGF. A CV was recorded at three of the
blocks of electrodes functionalized with v107 (chosen at
different sites of the array), three functionalized with the
control linear hexapeptide, and three of the blocks function-
alized with only the polymer surface for each concentration of
VEGF (Gold Biotechnology) used. The solution above the
array was then replaced with a new solution having a higher
concentration of VEGF and the CV experiments repeated. This
was done until a CV was recorded at each set of electrodes used
for each concentration of the target protein. The cyclic
voltammetry data obtained for one of the blocks of electrodes
functionalized with the v107-peptide is shown in Figure 2.
Figure 3. Summary of murine VEGF−v107 peptide binding study.
Blue = v107; orange = DRDGSP; gray = unfunctionalized polymer.
Figure 3 summarizes the data obtained for the linear
hexapeptide control. No significant binding occurred between
this peptide and murine-VEGF. The gray line in Figure 3
showed that VEGF did not bind to the unfunctionalized
polymer. From the data, a 50% value for the change in current
would correspond to a K value of approximately 30 μM, a
d
35,36
value consistent with the literature.
Clearly, the arrays can
be used to detect the binding event in spite of the weaker
affinity of the peptide for murine VEGF.
A similar conclusion was reached when the v107 peptide was
tested for its binding to human VEGF (Figure 4). The
Figure 4. Summary of the human VEGF−v107-peptide binding study.
Blue = v107; gray = unfunctionalized polymer.
experiment was conducted in a fashion identical to the one
shown in Figure 3 other than the change in receptor. The gray
line again showed that the human VEGF did not bind to the
unfunctionalized surface of the array. The blue curve showed
that the binding event between the v107 peptide and the
human VEGF was very similar to the interaction between the
Figure 2. CVs taken for one block of electrodes functionalized with
v107 with varying concentrations of VEGF.
peptide and murine VEGF (K ca.1 μM). As with the murine
d
homologue, the array-based analytical experiment had no
trouble detecting the micromolar binding event.
The largest CV wave recorded was obtained for the lowest
concentration of VEGF (most iron reaching the electrode
located beneath the v107-L19K peptide ligand), and the
smallest CV wave recorded was obtained for the highest
concentration of VEGF used, indicating that the array could be
used to detect the interaction between the peptide and its
Conclusions. We have found that molecules having thiol
27
27
groups, alcohols, acetylenes, azides, and NHS esters can all
be added directly to a borate ester containing diblock
copolymer surface on a microelectrode array. The reactions
can be conducted in a manner that places the molecule by any
electrode or set of electrodes in the array. In this way,
molecules functionalized for use in Michael reactions, “click
reactions,” and esterification reactions can all be placed on the
arrays without any need to resynthesize the molecules or
prefunctionalize the surface of the array. This development
should allow the microelectrode arrays to be used in
combination with other analytical methods without the
additional barrier of having to further functionalize or modify
the molecules being studied. The utility of the method for
VEGF target even with its low K value.
d
The data obtained from the experiment is summarized in
Figure 3. For this figure, the current measured for each CV
wave (peak current for the oxidation wave−peak current for the
reduction wave) shown in Figure 2 was plotted relative to the
concentration of VEGF used. For the electrodes functionalized
with the v107-peptide, this gave rise to the blue curve. The
error bars provided show the spread in data for three blocks of
the v107-peptide functionalized electrodes. The orange line in
G
DOI: 10.1021/acschembio.6b00685
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
functionalizing an array and setting the stage for a bioanalytical
study was shown in connection with the binding of v107 to
both human- and murine-VEGF targets.
Electrochemical and Spectrochemical Aqueous Ion Testing. Sens.
Actuators, B 136, 177−185.
̈
(5) Beyer, M., Nesterov, A., Block, I., Konig, K., Felgenhauer, T.,
Fernandez, S., Leibe, K., Torralba, G., Hausmann, M., Trunk, U.,
Lindenstruth, V., Bischoff, F. R., Stadler, V., and Breitling, F. (2007)
Combinatorial Synthesis of Peptide Arrays onto a Microchip. Science
METHODS
■
Materials were purchased from Sigma-Aldrich and used without
further purification unless otherwise indicated. Amino acids were
purchased from Advanced Chemtech and used without further
purification. 4-Bromobenzylazide was purchased from Sigma. The
acetylene tagged v107-peptide was purchased from CPC scientific, and
the VEGF-A proteins were purchased from Gold Bio in St. Louis.
Fluorescence microscopy was carried out with a Nikon Eclipse E200
microscope connected to a Boyce Scientific M-100 burner and a
Nikon D5000 camera. Optical filters used: CFW-BP01-Clinical-000
3
(
18, 1888.
6) Devaraj, N. K., Dinolfo, P. H., Chidsey, C. E. D., and Collman, J.
P. (2006) Selective Functionalization of Independently Addressed
Microelectrodes by Electrochemical Activation and Deactivation of a
Coupling Catalyst. J. Am. Chem. Soc. 128, 1794−1795.
(
7) Wassum, K. M., Tolosa, V. M., Wang, J., Walker, E.,
Monbouquette, H. G., and Maidment, N. T. (2008) Silicon Wafer-
Based Platinum Microelectrode Array biosensor for Near Real-Time
Measurement of Glutamate in vivo. Sensors 8, 5023−5036.
(
Semrock) filter cube excitation 380−395 nm/emission 420−470 nm,
(8) Kerkhoff, H. G., Zhang, X., Mailly, F., Nouet, P., Liu, H., and
ET−GFP (FITC/Cy2) (Chroma) filter cube excitation 450−490 nm,
emission 500−550 nm, and TxRed-A-Basic-000 (Semrock) filter cube
excitation 540−580 nm, emission 590−670.
Richardson, A. (2008) A Dependable Micro-Electronic Peptide
Synthesizer using Electrode Data. VLSI Design 2008, 1.
(9) Zhang, Y., Wang, H., Nie, J., Zhang, Y., Shen, G., and Yu, R.
Sample Procedure for Spin-Coating Arrays with the Block
Copolymer. The microelectrode arrays were coated with a spin-
coater MODEL WS-400B-6NPP/LITE. The chip was inserted into a
socket in the spinner and adjusted to be horizontal, then three drops of
(2009) Individually Addressable Microelectrode Arrays Fabricated
with Gold-Coated Pencil Graphite Particles for Multiplexed and High
Sensitive Impedance Immunoassays. Biosens. Bioelectron. 25, 34−40.
(10) Maurer, K., Yazvenko, N., Wilmoth, J., Cooper, J., Lyon, W., and
0
.03 g/mL PCEMA-b-pBSt solution (4:1.5 DMF/THF) were added
Danley, D. (2010) Use of a Multiplexed CMOS Microarray to
Optimize and Compare Oligonucleotide Binding to DNA Probes
Synthesized or Immobilized on Individual Electrodes. Sensors 10,
onto the chip in order to cover the entire electrode area. The chip was
then spun 1000 rpm for 40 s. The coating was allowed to dry for 15
min and subjected to irradiation using a 100 W Hg lamp for 20 min
before use.
7
(
371−7385.
11) Li, X., Tian, Y., Xia, P., Luo, Y., and Rui, Q. (2009) Fabrication
Array-Based Synthetic Procedures. Site-selective reactions on
13
of TiO and Metal Nanoparticle-Microelectrode Arrays by Photo-
the arrays were performed using the published procedures, including
2
2
7
24
lithography and Site-Selective Photocatalytic Deposition. Anal. Chem.
81, 8249−8255.
the Cu(II)-mediated and Cu(I)-catalyzed methods.
Microelectrode array binding studies were also performed using the
13,14,33
already published procedures.
(12) Chan, E. W. L., and Yousaf, M. N. (2007) Site-Selective
Immobilization of Ligands with Control of Density on Electroactive
Microelectrode Arrays. ChemPhysChem 8, 1469−1472.
ASSOCIATED CONTENT
Supporting Information
■
(13) For a recent review, see: Graaf, M. D., and Moeller, K. D.
*
S
(2015) An Introduction to Microelectrode Arrays, The Site-Selective
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acschem-
bio.6b00685.
Functionalization of Electrode Surfaces, and the Real-Time Detection
of Binding-Events. Langmuir 31, 7697−7706.
(14) Stuart, M., Maurer, K., and Moeller, K. D. (2008) Moving
Known Libraries to an Addressable Array: A Site-Selective Michael
Reaction. Bioconjugate Chem. 19, 1514.
Synthetic procedures and spectral data are included for
all new compounds along homology modeling for the
human and murine VEGF binding epitopes (PDF)
(15) For a summary of early efforts to place peptides onto arrays:
Stuart-Fellet, M., Bartels, J. L., Bi, B., and Moeller, K. D. (2012) Site-
Selective Chemistry and the Attachment of Peptides to the Surface of a
Microelectrode Array. J. Am. Chem. Soc. 134, 16891−16898.
AUTHOR INFORMATION
Corresponding Authors
E-mail: lapi@uab.edu.
E-mail: moeller@wustl.edu.
■
(16) For a description of the chips used here, see: Dill, K.,
Montgomery, D. D., Wang, W., and Tsai, J. C. (2001) Antigen
Detection Using Microelectrode Array Microchips. Anal. Chim. Acta
*
*
4
44, 69−78. 1K chips: electrode diameter = 92 μm; distance between
Notes
the Pt-electrodes (rectangular cells) = 245.3 and 337.3 μm. 12K slide:
diameter = 44 μm; distance between the Pt-electrodes (square cells) =
The authors declare no competing financial interest.
3
3 μm.
ACKNOWLEDGMENTS
We thank the National Science Foundation (CBET 1262176)
for their generous support of our work.
(17) Microelectrode arrays and the power supply for addressing them
■
can be purchased from CustomArray, Inc., 18916 North Creek
Parkway, Suite 115, Bothell, WA 98011 (www.CustomArrayInc.com).
For a detailed discussion of how the array reactions are run using this
equipment, see the Supporting Information for: Bartels, J., Lu, P.,
Maurer, K., Walker, A. V., and Moeller, K. D. (2011) Site-Selectively
Functionalizing Microelectrode Arrays: The Use of Cu(I)-Catalysts.
Langmuir 27, 11199−11205.
(18) Hu, L., Graaf, M. D., and Moeller, K. D. (2013) The Use of UV-
Cross-Linkable Di-Block Copolymers as Functional Reaction Surfaces
for Microelectrode Arrays. J. Electrochem. Soc. 160, G3020−G3029.
(19) For the stability of SAM coated electrodes, see: Strulson, J. K.,
Johnson, D. W., and Maurer, J. A. (2012) Increased Stability of Glycol-
Terminated Self-Assembled Monolayers for Long-Term Patterned
Cell Culture. Langmuir 28, 4318.
REFERENCES
■
(
1) Sullivan, M. G., Utomo, H., Fagan, P. J., and Ward, M. D. (1999)
Automated Electrochemical Analysis with Combinatorial Electrode
Arrays. Anal. Chem. 71, 4369−4375.
2) Zhang, S., Zhao, H., and John, R. (2000) Development of a
Generic Microelectrode Array Biosensing System. Anal. Chim. Acta
21, 175−187.
3) Hintsche, R., Albers, J., Bernt, H., and Eder, A. (2000)
Multiplexing of Microelectrode Arrays in Voltammetric Measure-
ments. Electroanalysis 12, 660−665.
4) Gardner, R. D., Zhou, A., and Zufelt, N. A. (2009) Development
(
4
(
(
(20) For selected examples of the direct modification of an SAM
of a Microelectrode Array Sensing Platform for Combination
coated electrode, see: (a) Kwon, Y., and Mrksich, M. (2001)
H
DOI: 10.1021/acschembio.6b00685
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Dependence of the Rate of an Interfacial Diels-Alder Reaction on the
Steric Environment of the Immobilized Dienophile: an Example of
Enthalpy-Entropy Compensation. J. Am. Chem. Soc. 124, 806.
Irreversible Peptide Ligand for PET Imaging of Vascular Endothelial
Growth Factor. J. Nucl. Med. 55, 1030−1034.
(
b) Yousaf, M. N., and Mrksich, M. (1999) Diels-Alder Reaction for
the Selective Immobilization of Protein to Electroactive Self-
Assembled Monolayers. J. Am. Chem. Soc. 121, 4286.
(
21) For a lead reference, see: Kolb, H. C., Finn, M. G., and
Sharpless, K. B. (2001) Click Chemistry: Diverse Chemical Function
from a Few Good Reactions. Angew. Chem., Int. Ed. 40, 2004−2021.
(
22) For a review, see: Totobenazara, J., and Burke, A. J. (2015) New
Click-Chemistry Methods for 1,2,3-Triazoles Synthesis: Recent
Advances and Applications. Tetrahedron Lett. 56, 2853.
(
23) Bi, B., Maurer, K., and Moeller, K. D. (2009) Building
Addressable Libraries: Site-Selective Lewis-Acid (Sc(III)) Catalyzed
Reactions. Angew. Chem., Int. Ed. 48, 5872.
(
24) Bartels, J. L., Lu, P., Walker, A., Maurer, K., and Moeller, K. D.
(
2009) Building Addressable Libraries: A Site-Selective Click-Reaction
Strategy for Rapidly Assembling Mass Spec Cleavable Linkers. Chem.
Commun., 5573.
(
25) For related reactions using band electrodes, see: Devaraj, N. K.,
and Collman, J. P. (2007) Copper Catalyzed Azide-Alkyne Cyclo-
additions on Solid Surfaces: Applications and Future Directions. QSAR
Comb. Sci. 26, 1253−1260. as well as ref 6 above.
(
26) Uppal, S., Graaf, M. D., and Moeller, K. D. (2014)
Microelectrode Arrays and the Use of PEG-Functionalized Diblock
Copolymer Coatings. Biosensors 4, 318−328. and references therein.
(
27) Graaf, M. D., and Moeller, K. D. (2016) Chemoselectivity and
the Chan-Lam Coupling Reaction: Adding Amino Acid acids to
Polymer-Coated Microelectrode Arrays. J. Org. Chem. 81, 1527−1534.
(
28) Balaraman, K., and Kesavan, V. (2010) Efficient Cu(II) Acetate
Catalyzed Homo- and Heterocoupling of Terminal Alkynes at
Ambient Conditions. Synthesis 20, 3461.
(
29) Jia, X., Yin, K., Li, C., Li, J., and Bian, H. (2011) Copper-
catalyzed Oxidative Alkyne Homocoupling without Palladium,
Ligands, and Bases. Green Chem. 13, 2175.
(
30) Coeffard, B., Moreau, X., Thomassigny, C., and Greck, C.
2013) Transition-Metal-Free Amination of Aryl-boronic Acids and
Their Derivatives. Angew. Chem., Int. Ed. 52, 5684.
31) Tian, J., and Moeller, K. D. (2005) Electrochemically Assisted
Heck Reactions. Org. Lett. 7, 5381.
32) Valeur, E., and Bradley, M. (2009) Amide Bond Formation:
Bond the Myth of Coupling Reagents. Chem. Soc. Rev. 38, 606.
33) Hu, L., Stuart, M., Tian, J., Maurer, K., and Moeller, K. D.
2010) Building Addressable Libraries: The Site-Selective Use of
Pd(0)-Catalysts on Microelectrode Arrays. J. Am. Chem. Soc. 132,
6610.
34) Lewis, W. G., Magallon, F. G., Fokin, V. V., and Finn, M. G.
2004) Discovery and Characterization of Catalysts for Azide-Alkyne
Cycloaddition by Fluorescence Quenching. J. Am. Chem. Soc. 126,
152.
35) Fairbrother, W. J., Christinger, H. W., Cochran, A. G., Fuh, G.,
(
(
(
(
(
1
(
(
9
(
Keenan, C. J., Quan, C., Shriver, S. K., Tom, J. Y. D., Wells, J. A., and
Cunningham, B. C. (1998) Novel peptides selected to bind vascular
endothelial growth factor target the receptor-binding site. Biochemistry
3
(
7, 17754−64.
36) Pan, B., Li, B., Russell, S. J., Tom, J. Y. K., Cochran, A. G., and
Fairbrother, W. J. (2002) Solution Structure of a Phage-derived
Peptide Antagonist in Complex with Vascular endothelial Growth
Factor. J. Mol. Biol. 316, 769−787.
(
37) Marquez, B. V., Beck, H. E., Aweda, T. A., Phinney, B.,
Holsclaw, C., Jewell, W., Tran, D., Day, J. J., Peiris, M. N., Nwosu, C.,
Lebrilla, C., and Meares, C. F. (2012) Enhancing Peptide Ligand
Binding to Vascular Endothelial Growth Factor by Covalent Bond
Formation. Bioconjugate Chem. 23, 1080−1089.
(
38) Chekhonin, V. P., Shein, S. A., Korchagina, A. A., and Gurina, O.
I. (2013) VEGF in tumor progression and targeted therapy. Curr.
Cancer Drug Targets 13, 423−443.
(
39) Marquez, B. V., Ikotun, O. F., Parry, J. J., Rogers, B. E., Meares,
C. F., and Lapi, S. E. (2014) Development of a Radiolabeled
I
DOI: 10.1021/acschembio.6b00685
ACS Chem. Biol. XXXX, XXX, XXX−XXX