Backscattering Interferometry: An alternative approach for the study of hydrogen bonding interactions in organic solvents

Article abstract of DOI:10.1021/ol200757a

Intermolecular interactions involving hydrogen bonds are responsible for catalysis and recognition. Traditional methods used to study hydrogen-bonding interactions are generally limited to relatively large volumes and high substrate concentrations. Backscattering Interferometry (BSI) provides a microfluidic platform to study these interactions in nonaqueous media at micromolar to nanomolar concentrations in picoliter volumes by monitoring changes in the refractive index.

Full text of DOI:10.1021/ol200757a

ORGANIC
LETTERS
2
011
Vol. 13, No. 10
654–2657
Backscattering Interferometry: An
Alternative Approach for the Study
of Hydrogen Bonding Interactions
in Organic Solvents
2
†
Esther N. Pesciotta, Darryl J. Bornhop,* and Robert A. Flowers II*
,‡
,†
Department of Chemistry, Lehigh University, Bethlehem, Pennsylvania 18015, United
States, and Department of Chemistry and Vanderbilt Institute for Chemical Biology,
Vanderbilt University, Nashville, Tennessee 37235, United States
rof2@lehigh.edu; darryl.bornhop@vanderbilt.edu
Received March 21, 2011
ABSTRACT
Intermolecular interactions involving hydrogen bonds are responsible for catalysis and recognition. Traditional methods used to study hydrogen-
bonding interactions are generally limited to relatively large volumes and high substrate concentrations. Backscattering Interferometry (BSI)
provides a microfluidic platform to study these interactions in nonaqueous media at micromolar to nanomolar concentrations in picoliter volumes
by monitoring changes in the refractive index.
3
Intermolecular interactions, especially those involving
hydrogen bonds, drive the structure and function of
macromolecules that catalyze reactions responsible for
maintaining living systems. In organic chemistry, hydro-
gen bonding partners have been utilized as the backbone of
approaches used to drive crystal formation, molecular
2
recognition and catalysis. Traditionally hydrogen bonding
interactions are studied using isotope incorporation,
4
5
6
NMR, UVÀvis spectroscopy, fluorescence, and
7
calorimetry. While these tools are widely used to char-
1
acterize molecular interactions that arise through the for-
mation of hydrogen bonds, they often require high
substrate concentrations and volumes. Given the synthetic
effort required to produce the components for the study of
systems that utilize hydrogen bonds for recognition and
catalysis, it is desirable to be able to study their interactions
using small quantities of substrates. Recently backscatter-
ing interferometry (BSI) has been shown to effectively
determine the binding affinity of several biomolecular
†
Lehigh University.
Vanderbilt University.
‡
(
1) Taylor, M. S.; Jacobsen, E. N. Angew. Chem., Int. Ed. 2006, 45,
520–1543.
2) (a) Hunter, C. A. Angew. Chem., Int. Ed. 2004, 43, 5310–5324. (b)
1
(
Kim, S.-G.; Kim, K.-H.; Kim, Y. K.; Shin, S. K.; Ahn, K. H. J. Am.
Chem. Soc. 2003, 125, 13819–13824.
(7) (a) Jadhav, V. D.; Schmidtchen, F. P. J. Org. Chem. 2008, 73,
1077–1087. (b) Solomonov, B. N.; Novikov, V. B. J. Phys. Org. Chem.
2007, 21, 2–13.
(8) (a) Bornhop, D. J.; Latham, J. C.; Kussrow, A.; Markov, D. A.;
Jones, R. D.; Sorensen, H. S. Science 2007, 317, 1732–1736. (b) Latham,
J. C.; Stein, R. A.; Bornhop, D. J.; McHaourab, H. S. Anal. Chem. 2009,
81, 1865–1871. (c) Kussrow, A.; Kaltgrad, E.; Wolfenden, M. L.;
Cloninger, M. J.; Finn, M. G.; Bornhop, D. J. Anal. Chem. 2009, 81,
4889–4897.
(
3) (a) Brown, A. R.; Kuo, W. H.; Jacobsen, E. N. J. Am. Chem. Soc.
010, 132, 9286–9288. (b) Zotova, N.; Broadbelt, L. J.; Armstrong, A.;
Blackmond, D. G. Bioorg. Med. Chem. Lett. 2009, 19, 3934–3937.
4) Tarkanyi, G.; Kiraly, P.; Varga, S.; Vakulya, B.; Soos, T. Chem.;
Eur. J. 2008, 14, 6078–6086.
2
(
(
(
5) Noziere, B.; Cordova, A. J. Phys. Chem. A 2008, 112, 2827–2837.
6) Hoang, L.; Bahmanyar, S.; Houk, K. N.; List, B. J. Am. Chem.
Soc. 2003, 125, 16–17.
1
0.1021/ol200757a r 2011 American Chemical Society
Published on Web 04/21/2011

8
interactions ranging from interactions with proteins to
9
nucleic acid hybridization. BSI measures the change in
In addition, these catalysts are very useful in diastereo-
3a
and enantioselective reactions. As a result, the strength of
refractive index of an aqueous solution resulting from
intermolecular association of two components. The key
features of BSI that make it appealing for the study of
intermolecular interactions are the small sample sizes (pL
to μL range) and concentrations (nM to μM) and the
ability to carry out experiments without functionalizing or
surface immobilizing one of the binding partners. To
date, BSI has been employed in aqueous systems where one
or more of the components are macromolecules. Given the
importance of hydrogen bonding in nonaqueous media,
we posed the following questions: (1) Can BSI be used to
study hydrogen bonding in nonaqueous media? (2) What
are the limits of detection in a nonaqueous environment
when the binding partners are small molecules? (3) Can
BSI be used to distinguish between similar hydrogen
bonding partners? Herein we show that BSI can be used
to study the interaction of diphenyl ureas and thioureas
with benzoate in acetonitrile (MeCN) with quantities
several orders of magnitude lower than other commonly
utilized techniques.
the interaction between substrate and catalyst is important
information for a synthetic chemist hoping to utilize such
reactions. The results discussed below exemplify how BSI
can be used to determine the affinity of commonly employed
urea and thiourea derivatives in a miniaturized format.
Throughout the course of this study, we found BSI to be a
very interaction-efficient method for the study of small
molecule interactions.
Initially end-point BSI experiments of TMAB interac-
tion with DPU and DPTU in MeCN were performed.
Since the BSI signal is dependent on the strength of the
interaction, experiments were carried out in the low to mid
micromolar range. For these experiments, DPU and
DPTU were held constant at 10 μM while TMAB was
varied from 5 to 60 μM. The BSI experiments were carried
out in a steady-state manner, in which samples were mixed
and allowed to equilibrate for several hours before exam-
ining at 25 °C. Prior to the experiment, the laser and
temperature controller were allowed to equilibrate for an
hour and the instrument was aligned with respect to the
microfluidic channel and the detector to obtain a single
frequency Fourier transform. The samples were analyzed
by pipetting 1 μL of each concentration directly into
8
a
BSI utilizes a low power HeÀNe laser focused perpendi-
cularly onto a microfluidic channel in a glass chip to
8
a
generate a backscattered interference fringe pattern. The
introduction of two binding partners into the channel
creates a change in refractive index, causing a spatial shift
in the fringe pattern. The magnitude of this shift depends on
the precise fringes interrogated, the concentration of the
binding pairs, conformational changes initiated upon bind-
8
ing, changes in water of hydration, and binding affinity. To
date, BSI has only been used to study interactions in
aqueous solvents. To determine if this technique can be
extended to studies in organic solvents, we examined the
complexation of tetramethylammonium benzoate (TMAB)
with 1,3-diphenyl urea (DPU), 1,3-diphenylthiourea
(
(
DPTU), 1,3-bis(p-nitrophenyl)urea (DNPU), and 1,3-bis-
p-nitrophenyl)thiourea (DNPTU) in MeCN. Addition-
ally, the limits of BSI detection was studied by carrying out
experiments with DPU interacting with either TMAB or
tetramethylammonium p-toluene sulfonate (TMAS).
Ureaandthioureahavebeenwidelystudied inmolecular
recognition because of their ability to form strong hydro-
Figure 1. Steady-state BSI data of DPU complexed to TMAB
(
9) and TMAS (2). The large signal shift for TMAB compared
to TMAS shows that BSI detects concentration-dependent
binding.
10
gen bonds. Hydrogen bonding through urea and thiourea
derivatives are used to recognize carboxylic acids, sulfonic
1
1
acids, and nitrates. Ureas and thioureas also act as acid
catalysts in a variety of organic reactions including the
the channel well and recording the signal for 45 s. The
zero point of DPU/DPTU only and the TMAB-only
calibration curve were subtracted from the binding data
to obtain the final binding curve (Supporting Information
Figure S1). Figure 1 shows the representative BSI plot and
curve-fit of DPU complexation with TMAB. The BSI
signal levels out at the high concentration of TMAB,
showing a saturation binding curve that can be fit to a
1
2
DielsÀAlder reaction and Claisen rearrangement.
(
9) Pesciotta, E. N.; Bornhop, D. J.; Flowers, R. A. Chem.ÀAsian J.
011, 6, 70–73.
10) (a) Kelly, T. R.; Kim, M. K. J. Am. Chem. Soc. 1994, 116, 7072–
2
(
7080. (b) Schmidtchen, F. P.; Berger, M. Chem. Rev. 1997, 97, 1609–
1646.
(
11) (a) Takemoto, Y. Org. Biomol. Chem. 2005, 3, 4299–4306. (b)
Werner, F.; Schneider, H. Helv. Chim. Acta 2000, 83, 465–478. (c) Smith,
P. J.; Reddington, M. V.; Wilcox, C. S. Tetrahedron Lett. 1992, 33, 6085–
one-sitebinding hyperbolatoobtainK values. BothDPU
D
and DPTU (Supporting Information Figure S2) have
similar affinity for TMAB with KD values of 18.56 and
6088.
(
12) (a) Kelly, T. R.; Meghani, P.; Ekkundim, V. S. Tetrahedron Lett.
1990, 31, 3381–3384. (b) Etter, M. C. Acc. Chem. Res. 1990, 23, 120–126.
23.20 μM, respectively (Table 1). Next, experiments were
(c) Etter, M. C.; Urbancyzk-Lipowska, Z.; Zia-Ebrahimi, M.; Panunto,
T. W. J. Am. Chem. Soc. 1990, 112, 8415–8426.
carried out on DNPU and DNPTU as these substrates are
Org. Lett., Vol. 13, No. 10, 2011
2655

1
2c,13
expected to have a higher affinity for benzoate.
Ex-
MeCN at 25 °C. As with the BSI studies, the DNPU/
DNPTU ITC experiments were run at slightly lower
concentrations, with 2 mM TMAB titrated into 0.2 mM
of either DNPU or DNPTU. Additionally, ITC experi-
ments were carried out with 10 mM TMAS titrated into 1
mM DPU. The results from these experiments are shown
in Table 1, with DPU having a slightly higher affinity for
TMAB than DPTU, with respective dissociative binding
constants of 21.75 and 27.43 μM. The opposite trend was
observed for TMAB interaction with DNPU and
DNPTU, with DNPTU having a slightly lower KD of
amination of a higher affinity interaction is important to
test the limits of BSI an analytical tool for small molecule
interactions and determine the degree in which partners
a
Table 1. K
D
Values and Standard Error from BSI and ITC
BSI ITC
TMAB complexation
K
D
(μM)
D
K (μM)
DPU
18.56 ( 4.46
23.20 ( 4.55
0.54 ( 0.08
0.42 ( 0.08
146 ( 40
21.75 ( 6.04
27.43 ( 3.72
1.26 ( 0.06
1.18 ( 0.17
356 ( 27
1
.18 μM and DNPU with a K of 1.26 μM. The BSI K
DPTU
D D
DNPU
valuesareall withinexperimental error of the ITC data and
exhibit the same trends. The only exception is the interac-
tion between DPU-TMAS, where the K values between
DNPTU
TMAS-DPU
D
a
BSI and ITC are slightly different. This interaction is most
likely near the upper detection limit of BSI and ITC, given
that smaller signal changes are observed in both methods.
The data described above clearly show that BSI is
capable of measuring the affinity of small molecules in
nonaqueous media. Although BSI is useful for examining a
range of intermolecular interactions, the basis for the
signal obtained in these experiments is not fully under-
stood. Itis our supposition, that changes inrefractiveindex
arise from the release or capture of solvent upon associa-
tion. To examine this hypothesis, ellipsometry experiments
were carried out for the DPU-TMAB interaction to de-
termine the relationship between refractive index and a
binding event. These experiments were run using a 60°
liquid prism cell and calculating the refractive index at 633
nm, the same wavelength used in the BSI experiments. To
obtain measurable changes in the refractive index, the
concentrations of the binding pairs were increased, with
DPU held constant at 25 μM and TMAB varied from 25 to
Thermodynamic data is available in the Supporting Information.
with different affinities can be distinguished from each
other. BSI experiments were carried out at lower concen-
trations than the previous system with DNPU/DNPTU
held constant at 1 μM and TMAB varied from 0.5 to 10
μM. Again, final binding curves were obtained after sub-
tracting out the TMAB calibration data. The binding
affinity for both DNPU and DNPTU were similar
(Table 1) with K values of 0.54 and 0.42 μM, respectively.
D
As expected, the data obtained shows the enhanced bind-
ing to TMAB due tothe presence of the p-NO groups. The
2
data in Table 1 and subsequent binding curves (Supporting
Information Figures S3 and S4) show similar radian shifts
2
and binding saturation with R values of 0.97 or higher.
BSI detects specific binding interactions in a concentration
dependent manner. To illustrate this point, experimental
results were compared for DPU-TMAB and DPU-TMAS
interactions. TMAS is not expected to have a strong affinity
for DPU because the sulfonate oxygens are much less
electronegative in comparison to the benzoate oxygens, as
200 μM. The final binding data was obtained by subtract-
ing out the zeropoint of25μM DPU from the binding data
and then subtracting out the subsequent refractive index
demonstrated by the lower pK for benzoate (10) vs benze-
b
10a
nesulfonate (20). Given the low affinity for DPU-TMAS,
performing a BSI experiment under the same conditions as
the DPU-TMAB experiment should produce very different
results. As expected, Figure 1 shows the negligible signal shift
due to DPU-TMAS complexation whereas DPU-TMAB
generates a large signal shift that saturates at much lower
concentrations. However, when the DPU-TMAS BSI
experiment was run at higher concentrations with 50 μM
DPU and 0À600 μM TMAS, a saturation binding curve was
produced (Supporting Information Figure S5) and fit to
obtain a K value of 0.146 (0.04 mM. This value agrees with
D
previously reported KD values between sulfonate and
(thio)urea derivatives ranging from 0.006 to 22.2 mM, with
the lower values attributed to the precense of electron with-
10a,13
drawing groups on the (thio)urea derivatives.
To benchmark the BSI studies, isothermal titration
calorimetry (ITC) experiments were performed. Initial
experiments were carried out with 5 mM TMAB titrated
into the sample cell containing 0.5 mM DPU or DPTU in
Figure 2. Plot of ellipsometry refractive index measurements at
633 nm. The relative refractive index of equilibrated samples
containing 25 μM DPU and 0À200 μM TMAB show a satura-
tion binding curve. Curve fitting to a one-site binding hyperbola
2
yielded a K
D
of 19.40 ( 6.23 μM with an R of 0.97.
(
13) Wilcox, C. S.; Kim, E.; Romano, D.; Kuo, L. H.; Burt, A. L.;
Curran, D. P. Tetrahedron 1995, 51, 621–634.
2
656
Org. Lett., Vol. 13, No. 10, 2011

calibration curve of the varying concentration of TMAB.
Analysis of the ellipsometry data was analogous to BSI
since the experiments were performed at steady-state and
also measured the change in refractive index. Figure 2
shows the ellipsometry average plot with the curve fit
between ca. 325.2 million molecules. When compared to
the concentrations and volumes used for the other techniques
in this study, BSI is 6 orders of magnitude more sensitive than
ITC and 8 orders of magnitude more than ellipsometry. This
makes BSI interaction-efficient, with the ability to detect a
relatively small number of discrete interactions in comparison
to other free-solution techniques. Furthermore, the straight-
forward, user-friendly design of BSI provides an elegant
technique for screening intermolecular interactions of small
molecules in organic solvents making it potentially useful for
application in high-throughput screening.
obtaining a K value of 19.40 ( 6.23 μM. The ellipsometry
D
data correlate with the BSI K values, providing evidence
that the BSI signal shift is related to variations in refractive
index during complex formation.
D
In conclusion, the present study demonstrates the ap-
plicability of BSI as a tool for studying small molecule
interactions in nonaqueous solvents. Not only is BSI able
to measure the formation of just two hydrogen bonds, but
it can also be used to distinguish between TMAB com-
plexation with DPU/DPTU and DNPU/DNPTU with an
affinity difference of more than 1 order of magnitude. BSI
experiments have the advantage of using smaller volumes
and lower concentrations than ITC and ellipsometry. The
microfluidic channel used in these BSI experiments has a
Acknowledgment. R.A.F. and D.J.B. are grateful to the
National Science Foundation for supporting this research
(CHE 0848788). R.A.F. also thanks the National Science
Foundation for funding the purchase of the spectroscopic
ellipsometer (CHE 0923370) at Lehigh University.
2
cross-sectional area of 3600 μm that, when interrogated
Supporting Information Available. General methods,
experimental protocols, and supporting data. This mate-
rial is available free of charge via the Internet at http://
pubs.acs.org.
by a 100 μm diameter laser, provides an optical probe
volume of ca. 360 pL. Therefore, at the lowest concentrations
used in these experiments, BSI can detect an interaction
Org. Lett., Vol. 13, No. 10, 2011
2657