Molecularly imprinted nanoparticles as tailor-made sensors for small fluorescent molecules

Article abstract of DOI:10.1039/c4cc01516a

Water-soluble nanoparticles molecularly imprinted against naphthyl derivatives could bind the templates with high affinity and excellent selectivity among structural analogues in aqueous solution. Fluorescent dansyl groups installed during template polyme

Full text of DOI:10.1039/c4cc01516a

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Molecularly imprinted nanoparticles as tailor-
made sensors for small fluorescent molecules†
Cite this: Chem. Commun., 2014,
50, 5752
Joseph K. Awino and Yan Zhao*
Received 27th February 2014,
Accepted 7th April 2014
DOI: 10.1039/c4cc01516a
www.rsc.org/chemcomm
Water-soluble nanoparticles molecularly imprinted against naphthyl micelles in water.7 The nanoparticles imprinted against a bile salt
derivatives could bind the templates with high affinity and excellent derivative were found to bind the template among its structural
selectivity among structural analogues in aqueous solution. Fluorescent analogues with excellent selectivity and affinity. Because the radius
dansyl groups installed during template polymerization allowed these of the MINP (ca. 1.5 nm for the hydrophobic core and 2.5 nm
¨
nanoparticles to detect the presence of the target analytes by Forster including the surface ligands) is within the Forster distance (R0) of
resonance energy transfer.
¨
many fluorophore pairs,8 we reasoned that a MINP functionalized
with an appropriate fluorophore should be able to detect analytes
¨
Chemical sensors are important for a wide range of applications through Forster resonance energy transfer (FRET). As pointed out in
including clinical diagnostics, environmental remediation, drug a recent review, ‘‘ability to spectroscopically characterize binding
analysis, and chemical detection. Sensors ideally should detect sites’’ is a highly desirable feature for MIPs, especially if the materials
specific chemicals of interest with minimal interference from can be made ‘‘either soluble or insoluble’’ and ‘‘readily processable’’.9
other chemicals present in the same sample. What is vital to
To demonstrate the concept, we first prepared an aqueous
the sensing specificity is typically a molecular-recognition unit in solution of cross-linkable surfactant 1, template 2 (or 3), and fluores-
the sensor that binds the analyte with high affinity and selectivity. cent dansyl derivative 4 that has two polymerizable methacrylate
Molecular imprinting is a technique to create guest- groups (Scheme 1). Surfactant 1 has a critical micelle concentration
complementary binding sites, most often in a cross-linked polymer (CMC) of 0.55 mM and an aggregation number of 50 in water.6 With
matrix.1 It usually involves polymerization of a mixture of imprint [1] = 10 mM and [1]/[2 or 3] = 50/1, the resulting MINP was expected
molecules (i.e., the templates), functional monomers, and cross- to contain on average one binding site per particle. To enable the
linkers into a highly cross-linked material. Template-complementary resulting MINP to detect the template (the target analyte) by FRET,
binding sites are created upon the removal of the templates from the we chose to employ a naphthalene-containing template (2 or 3) that
polymer matrix. Because molecularly imprinted polymers (MIPs) could serve as a FRET donor for the dansyl acceptor to be incorpo-
potentially can be prepared for any molecule that can form a suitable rated into the MINP through co-polymerization of 4.
template-functional monomer complex, molecular imprinting is a
powerful technique for preparing synthetic receptors.
The details of the MINP synthesis and characterization are
reported in the ESI† (Fig. S1–S6). As shown in Scheme 1, the micelles
A key benefit of MIP is its predetermined binding selectivity of 1 were first cross-linked via click chemistry on the surface by
(for the template or its mimics). This feature is enormously diazide 5 using Cu(I) catalysts. At this point, the organic additives
useful to molecular sensing in which the molecules of interest including 4, DVB (divinylbenzene), and DMPA (2,2-dimethoxy-2-
are typically known.1c Indeed, when coupled with optical,2 mass,3 phenylacetophenone, a photoinitiator) should simply be trapped
refractive index,4 or other signal-transducing mechanisms, MIPs within the SCM. Immediately after the surface-cross-linking, a sugar-
have been used as sensors for a variety of analytes.5
derived azide (6) was added to the mixture to functionalize the
We recently reported a method to prepare molecularly imprinted surface of the alkynyl-SCM. The alkynyl-SCM had extra alkynes on
nanoparticles (MINPs)6 by surface-core cross-linking of surfactant the surface because the ratio of [1]/[5] was 1.2 in the reaction mixture
while surfactant 1 had 3 alkynyl groups and cross-linker 5 only
2 azides. After surface-functionalization, UV irradiation triggered free
radical polymerization of the methacrylate groups of 1 and 4, as well
as DVB solubilized within the SCM core. The micelles were able to
Department of Chemistry, Iowa State University, Ames, Iowa 50011-3111, USA.
E-mail: zhaoy@iastate.edu; Fax: +1-515-294-0105; Tel: +1-515-294-5845
† Electronic supplementary information (ESI) available: Syntheses and characteri-
solubilize one DVB per surfactant and this high level of DVB was
found to enhance the rigidity of the core and the binding selectivity.6
zation of materials, experimental procedures for the binding studies, additional
data and Figures. See DOI: 10.1039/c4cc01516a
5752 | Chem. Commun., 2014, 50, 5752--5755
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
Fig. 1 (a) Normalized excitation spectra of fluoro-MINP (2) in the presence
of different concentrations of 2 and (b) the excitation spectra with the
contribution of fluoro-MINP (2) subtracted. (c) Normalized excitation spectra
of fluoro-MINP (3) in the presence of different concentrations of 3 and
(d) the excitation spectra with the contribution of fluoro-MINP (3) subtracted.
The emission for the dansyl acceptor at 500 nm (lem) was monitored as the
excitation wavelength (lex) was scanned from 250 to 450 nm. [MINP] = 0.25 mM
in 50 mM Tris buffer (pH 7.4).
Scheme 1 Preparation of fluoro-MINP.
At the end of the core-cross-linking, the fluoro-MINPs were recovered
by precipitation from acetone, followed by methanol washing
(to remove the imprint molecules).
The entire cross-linking process could be monitored easily by When the contribution of the acceptor (i.e., the fluoro-MINP) was
1H NMR spectroscopy. As shown in Fig. S1 and S2 (ESI†), upon subtracted, a distinctive peak near 300–310 nm from the donor
1
surface-cross-linking, the sharp H NMR signals of the surfactant (2 or 3) appeared (Fig. 1b and d), indicative of increasing FRET
were replaced by broad peaks and the protons near the ammonium with higher concentrations of the template added to the solution.
headgroup (e.g., propargylic protons) disappeared. The meth- When the ‘‘wrong’’ template was added, e.g., 3 to fluoro-MINP (2)
acrylate and the DVB protons, although visible after surface- or 2 to fluoro-MINP (3), as shown in Fig. S7 and S8 (ESI†), the
cross-linking, disappeared completely after core-cross-linking.
FRET signal was either absent or much weaker. Similar observations
Fig. 1a and c shows normalized excitation spectra of fluoro- were made when the MINPs were titrated with other structural
MINP (2) and fluoro-MINP (3) in the presence of different analogues (7–10), including 7 that only differed from 2 by the
concentrations of 2 and 3 in Tris buffer (pH 7.4), when the dansyl position of a carboxylate (Fig. S9–S16, ESI†).
emission at 500 nm was monitored. In the absence of binding, the
The above results indicate that non-specific binding (from
donor fluorophore (2 or 3) would stay largely in solution, far from generic hydrophobic and electrostatic interactions) between the
the dansyl acceptors embedded within the fluoro-MINPs. Titra- MINPs and the negatively charged template analogues could not
tion of the fluoro-MINP with 2 or 3 would not affect the excitation trigger FRET, even those with very similar structures.11 FRET was
spectrum (of the dansyl) since all the emission would be caused apparently a result of strong and specific binding, which was
by direct excitation of the dansyl in this scenario. In the event of a confirmed by isothermal titration calorimetry (ITC) shown in
binding, the donor molecule bound by the MINP would absorb Fig. 2a and b. With ITC, we could obtain the binding data even
light, undergo excitation, and transfer the excited energy to the for those structural analogues that caused no change in the
dansyl acceptor in the same nanoparticle. In the latter case, the fluorescence excitation spectra. Additionally, the technique allowed
donor would contribute to the acceptor emission and thus peaks us to determine the number of binding sites (N) on the MINP.
corresponding to the donor absorption would appear in the
excitation spectrum of the dansyl acceptor.
The ITC binding data in Table 1 shows that the MINPs were
highly selective in their binding. For MINP (2), the template itself
We chose to have a 2 : 1 ratio of dansyl derivative 4 to the gave a binding constant of Ka = 0.43 Â 106 MÀ1, which translates to
template (2 or 3) so that each MINP had 2 dansyl groups on a binding free energy of ÀDG = 7.7 kcal molÀ1. The affinity was
average and a good chance existed for the naphthyl template to be quite remarkable for a small molecule like 2 and should have
within the Forster distance (R0 = 2.2 nm)10 of the dansyl acceptor resulted from the combination of hydrophobic interactions and
¨
during rebinding. Indeed, the characteristic contribution from the electrostatic interactions between the oppositely charged MINP and
donor absorption (@ 290–310 nm) appeared when the template the guest. None of the other anionic analogues, whether larger or
molecule was added to the ‘‘correct’’ fluoro-MINP (Fig. 1a and c). smaller than 2, showed any comparable binding; all the Ka values
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 5752--5755 | 5753

View Article Online
ChemComm
Communication
Fig. 3 (a) Excitation spectra of MINP (3) with 2 mM of compound 3, titrated
with 0–12 mM of compound 2. (b) Excitation spectra of MINP (3) with 2 mM of
compound 3, titrated with 0–12 mM of compound 8. The dotted spectrum in
black was obtained by subtracting the MINP spectrum from that of the MINP
plus compound 3. The emission for the dansyl acceptor at 520 nm (lem) was
monitored as the excitation wavelength (lex) was scanned from 250 to
450 nm. [MINP] = 0.50 mM in 50 mM Tris buffer (pH 7.4).
Fig. 2 ITC titration curve obtained at 298 K for the bindings between (a) fluoro-
MINP (2) and 2 and (b) between fluoro-MINP (3) and 3 in 50 mM Tris buffer
(pH 7.4). Additional ITC curves (Fig. S17–S18) are reported in the ESI.†
Table 1 Binding data for MINPs obtained by ITCa
Ka
ÀDG
Entry
MINP
Guest
(Â 106 MÀ1
)
(kcal molÀ1
)
N
the donor to the MINP acceptor was clearly visible in the excitation
spectrum (Fig. 3a, compare the MINP spectra before and after the
addition of compound 3; the dotted spectrum in black was
obtained by subtracting the MINP spectrum from that of MINP
plus 3, showing lmax = 310 nm from the donor).14 Significantly,
when 2–12 mM of compound 2 (Fig. 3a), 7 (Fig. S19, ESI†), or 9
(Fig. S20, ESI†), was added,15 the excitation spectra showed essentially
no change. Compound 8 did show some interference (Fig. 3b). Since
8 and 9 were bound by MINP (3) similarly, the interference from 8
should derive from its spectroscopic instead of binding properties.
We also examined the interference of two additional analogues of 3,
with the methyl ester hydrolyzed (in 11) and replaced with a longer,
hexyl group (in 12), respectively. As shown in Fig. S21 and S22 (ESI†),
these analogues did not affect the FRET signal at all, despite their
similarity to 3.
In summary, we have demonstrated that fluorescently-labelled
MINPs can be generated against hydrophobic guests for highly
specific binding among their structural analogues. The combination
of predetermined binding properties from molecular imprinting
and easy-to-perform FRET-based detection make these MINPs
potentially very useful as sensors for small fluorescent mole-
cules in water. Since the fluorophore was introduced indepen-
dently from the molecular recognition-aspect of the imprinting,
the FRET-detection and molecular imprinting in principle are
orthogonal to each other.
1
2
3
4
5
6
7
8
MINP (2)
MINP (2)
MINP (2)
MINP (2)
MINP (2)
MINP (2)
MINP (3)
MINP (3)
MINP (3)
MINP (3)
MINP (3)
MINP (3)
2
3
7
8
9
10
2
3
7
0.43 Æ 0.01
7.7
—
1.1
—
b
b
b
—
0.0023 Æ 0.0004
0.0015 Æ 0.0001
0.0033 Æ 0.0003
0.0011 Æ 0.0001
0.0070 Æ 0.0002
1.00 Æ 0.04
4.6
4.3
4.8
4.1
5.2
8.2
4.3
5.4
5.3
5.4
0.8
1.1
1.0
0.8
1.2
1.2
1.0
0.7
1.1
0.5
9
0.0015 Æ 0.0002
0.0095 Æ 0.0002
0.0082 Æ 0.0010
0.0095 Æ 0.0003
10
11
12
8
9
10
a
The titrations were generally performed in duplicates and the errors
in Ka between the runs were generally o20%. Binding was measured in
b
50 mM Tris buffer (pH = 7.4). Binding was not detectable by ITC.
were at least two orders of magnitude lower than that for the
template itself (Table 1, entries 2–6).
It is significant to note that stereoisomer 7 was bound by
MINP (2) nearly 200 times weaker than 2. The result highlighted
the importance of hydrophilic anchoring during the imprinting:
template 2 had to place its carboxylate group on the surface of the
micelle; the ionic anchor must have oriented the hydrophobic
group so that the resulting binding pocket could not accommodate
the naphthyl and a misplaced carboxylate.
MINP (3) was also very selective. It bound its own template 3
with a Ka value of 1.00 Â 106
M
À1, more than twice as that
We thank NSF for supporting the research.
between MINP (2) and 2. The higher affinity12 was reasonable
given the larger hydrophobic size of the guest, as hydrophobic
interactions are known to be proportional in strength to the
hydrophobic surface area buried upon binding.13 None of the
other guests, despite their similarities, was bound by MINP (3)
in comparable affinity (Table 1, entries 9–12).
Notes and references
1 (a) G. Wulff, Angew. Chem., Int. Ed. Engl., 1995, 34, 1812; (b) G. Wulff,
Chem. Rev., 2001, 102, 1; (c) K. Haupt and K. Mosbach, Chem. Rev.,
2000, 100, 2495; (d) K. J. Shea, Trends Polym. Sci., 1994, 2, 166;
(e) B. Sellergren, Molecularly Imprinted Polymers: Man-Made Mimics of
Antibodies and Their Applications in Analytical Chemistry, Elsevier,
Amsterdam, 2001; ( f ) B. Sellergren, Angew. Chem., Int. Ed., 2000,
39, 1031; (g) M. Komiyama, Molecular Imprinting: From Fundamentals
to Applications, Wiley-VCH, Weinheim, 2003; (h) M. Yan and
Convinced of the highly selective binding, we examined the
FRET signals in the presence of potentially interfering structural
analogues. Because of the stronger FRET of MINP (3) with its
template, we examined the FRET detection of 3 in the presence of
various structural analogues as potential interfering species. When
2 mM of 3 was added to a solution of 0.50 mM MINP (3), FRET from
¨
O. Ramstrom, Molecularly Imprinted Materials: Science and Technology,
Marcel Dekker, New York, 2005; (i) C. Alexander, H. S. Andersson,
L. I. Andersson, R. J. Ansell, N. Kirsch, I. A. Nicholls, J. O’Mahony and
M. J. Whitcombe, J. Mol. Recognit., 2006, 19, 106.
5754 | Chem. Commun., 2014, 50, 5752--5755
This journal is ©The Royal Society of Chemistry 2014

View Article Online
Communication
ChemComm
2 (a) A. L. Jenkins, O. Manuel Uy and G. M. Murray, Anal. Commun.,
1997, 34, 221; (b) P. Turkewitsch, B. Wandelt, G. D. Darling and
W. S. Powell, Anal. Chem., 1998, 70, 2025; (c) S. A. Piletsky,
E. Terpetschnig, H. S. Andersson, I. A. Nicholls and O. S. Wolfbeis,
Fresenius’ J. Anal. Chem., 1999, 364, 512; (d) J. Matsui, M. Higashi and
T. Takeuchi, J. Am. Chem. Soc., 2000, 122, 5218; (e) T. Takeuchi,
T. Mukawa, J. Matsui, M. Higashi and K. D. Shimizu, Anal. Chem.,
7 (a) S. Zhang and Y. Zhao, Macromolecules, 2010, 43, 4020; (b) S. Zhang and
Y. Zhao, J. Am. Chem. Soc., 2010, 132, 10642; (c) X. Li and Y. Zhao,
Langmuir, 2012, 28, 4152; (d) S. Zhang and Y. Zhao, Chem. Commun., 2012,
48, 9998; (e) G. Chadha and Y. Zhao, Chem. Commun., 2014, 50, 2718.
8 B. W. v. d. Meer, G. Coker and S. Y. S. Chen, Resonance Energy
Transfer: Theory and Data, VCH, New York, 1994.
9 S. C. Zimmerman and N. G. Lemcoff, Chem. Commun., 2004, 5.
2001, 73, 3869; ( f ) A. Tong, H. Dong and L. Li, Anal. Chim. Acta, 2002, 10 L. Stryer, Annu. Rev. Biochem., 1978, 47, 819.
466, 31; (g) E. Mertz and S. C. Zimmerman, J. Am. Chem. Soc., 2003, 11 We also tried 1-, and 2-naphthaleneamine as the guests and observed
125, 3424; (h) N. T. Greene and K. D. Shimizu, J. Am. Chem. Soc., 2005,
127, 5695.
no FRET either. They were not studied further because their water-
insolubility did not allow us to determine their binding by ITC.
3 (a) F. L. Dickert, O. Hayden and K. P. Halikias, Analyst, 2001, 12 Although compound 3 was significantly larger than 2, the former had
126, 766; (b) M. Zayats, M. Lahav, A. B. Kharitonov and I. Willner,
Tetrahedron, 2002, 58, 815; (c) K. Das, J. Penelle and V. M. Rotello,
Langmuir, 2003, 19, 3921.
quite a number of oxygen and nitrogen atoms that might be hydrated
when the molecule was in solution. The substantial desolvation during
binding might be the reason why the binding for 3 was only marginally
larger than that for 2 by their corresponding MINPs.
4 A. Kugimiya and T. Takeuchi, Biosens. Bioelectron., 2001, 16, 1059.
5 (a) Molecularly Imprinted Materials–Sensors and Other Devices, 13 C. Tanford, The Hydrophobic Effect: Formation of Micelles and
ed. K. J. Shea, M. Yan and M. J. Roberts, Materials Research Society, Biological Membranes, Krieger, Malabar, Fla, 2nd edn, 1991.
Warrendale, PA, 2002; (b) L. Ye and K. Haupt, Anal. Bioanal. Chem., 14 The lmax of the donor (310 nm) was quite close to that of the acceptor
2004, 378, 1887; (c) M. J. Whitcombe, I. Chianella, L. Larcombe,
S. A. Piletsky, J. Noble, R. Porter and A. Horgan, Chem. Soc. Rev.,
2011, 40, 1547.
(340 nm). An ideal FRET sensor should have very little overlap between
the donor and the acceptor absorptions.
15 Compound 10 contained the same dansyl as the MINP (3) itself and
thus was not compared in this study.
6 J. K. Awino and Y. Zhao, J. Am. Chem. Soc., 2013, 135, 12552.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 5752--5755 | 5755