Tetrahedron 73 (2017) 4950e4954

Contents lists available at ScienceDirect

Tetrahedron
journal homepage: www.elsevier.com/locate/tet

A modular self-assembled sensing system for heavy metal ions with

tunable sensitivity and selectivity
Subhabrata Maiti, Leonard J. Prins*
Department of Chemical Sciences, University of Padova, Via Marzolo 1, 35131, Padova, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Here we describe a self-assembled sensing system composed of three separate modules: gold nano-
Received 14 March 2017 particles, a reporter element, and a recognition element. The gold nanoparticles serve as a multivalent
Received in revised form platform for the interaction with both the reporter and recognition element and the gold nucleus serves
2 May 2017
to affect the fluorescent properties of the reporter. The reporter element serves for generation of the
Accepted 8 May 2017
output signal. The recognition element serves to make the assay selective. The working principle is that
Available online 10 May 2017
the interaction of the analyte with the recognition element leads to an increased affinity for the gold
nanoparticle, which causes a displacement of the reporter and a turn-ON of fluorescence. It is shown that
Keywords:
Gold nanoparticles
the modular nature of the system permits straightforward tuning of the dynamic detection range, the
Chemosensor sensitivity, and the selectivity, simply by changing the recognition module. The system can detect Hg2þ
Self-selection and Agþ metal ions at nanomolar concentrations in aqueous buffer.
Self-assembly © 2017 Elsevier Ltd. All rights reserved.
Displacement assay
Heavy metal ions

1. Introduction emphasized that the main focus of this work is on demonstrating the
novel features offered by a self-assembly approach to chemosensor
The development of innovative sensing systems for the detection development. It will be shown that the different modules of the
of analytes plays a crucial role in, amongst others, molecular di- sensing system can be independently changed to alter the selectivity,
agnostics, the detection of environmental pollution, food contami- sensitivity and dynamic detection range of the system,27 while
nation, and counter-terrorism measures.1e6 Intensive research over maintaining the same fluorescence output signal. In addition, this
the past decades has led to the development of a wide variety of work provides a relevant implementation of our recently developed
chemosensors able to convert the presence of low concentrations of protocol for dynamic combinatorial chemistry (DCC) on a multivalent
analytes into a detectable and easily readable output signal (optical, nanoparticle surface.28
electrical, thermal etc.).7e13 To favour practical applications, these
systems have often been designed based on criteria of robustness and
simplicity. However, there is a current interest in developing sensing 2. Results and discussion
systems of higher complexity, because of features that are difficult to
obtain using simple molecules.14e17 In particular, attention is being The use of Au NP 1 for the detection of peptides, nucleotides, and
paid to the use of self-assembly as a design principle because it en- small molecules has been reported by our group in recent
ables a modular approach through which the selectivity of the sys- years.11,29e31 Au NP 1 is composed of a gold nucleus with a diameter
tems towards different analytes can be tuned in a straightforward of around 1.6 (±0.3) nm covered with a monolayer of alkyl thiols
manner simply by changing the building blocks.18e22 terminating with a 1,4,7-triazacyclononane (TACN)$Zn2þ complex
Here we report a self-assembled fluorescence turn-ON sensing (Fig. 1). Fluorescent displacement assays have been developed
system that is able to detect the heavy metal ions Hg2þ and Agþ at based on the displacement of a negatively charged fluorescent in-
nanomolar concentrations in water. Although it is worth mentioning dicator from the Au NP 1 surface by analytes able to compete with
that the detection of toxic heavy metals is extremely important from binding to the surface. Recently, we have reported on a new signal
an environmental as well as a health perspective,23e26 it is transduction pathway relying on the formation of a ternary com-
plex between two thymidine nucleotides and Hg2þ metal ions with
* Corresponding author. an increased affinity for Au NP 1 compared to the separate nucle-
E-mail address: [email protected] (L.J. Prins). otides.21,25 An interesting feature of that study was that the

http://dx.doi.org/10.1016/j.tet.2017.05.028
0040-4020/© 2017 Elsevier Ltd. All rights reserved.
 S. Maiti, L.J. Prins / Tetrahedron 73 (2017) 4950e4954 4951

Fig. 1. Schematic representation of the signal generation process upon the addition of Hg2þ to the system composed of Au NP 1, probe A, and both TDP and cTMP. Low concen-
trations of Hg2þ result in formation of the ternary complex TDP$Hg2þ$TDP. As soon as TDP is depleted, further amounts of Hg2þ are complexed with cTMP. This way the dynamic
sensing regime of the system is enlarged covering a Hg2þ concentration range low nanomolar to high micromolar.

multivalent NP surface drives the equilibrium towards the more concentration of the cTMP$Hg2þ$cTMP complex is required to elicit
stable thermodynamic complex and self-selects the Hg2þ complex a displacement of probe A from Au NP 1. For TDP the fluorescence
with the highest affinity. Indeed, it was observed that the addition intensity started to increase already after the addition of 20 nM of
of Hg2þ to a mixture of TMP and cTMP complex in aqueous medium Hg2þ and continued until a maximum intensity was reached at
in the absence of Au NP 1 led to the formation of all possible ternary around 7 mM of Hg2þ, which corresponds roughly to the concen-
complexes (TMP$Hg2þ$TMP, TMP$Hg2þ$cTMP, cTMP$Hg2þ$cTMP), tration at which all free TDP has been depleted (Fig. 2a). On the other
but with an equilibrium composition shifted towards the more
stable cTMP$Hg2þ$cTMP complex. However, the presence of Au NP
1 shifted the equilibrium entirely to the other side and exclusive
formation of the TMP$Hg2þ$TMP complex was observed. This shift
is driven by the high density of negative charges in this complex
which cause a stronger interaction with the multivalent cationic
surface of the Au NP 1. This study showed that the affinity can be
modulated simply by changing the signal transduction unit (the
phosphates) without altering the recognition unit (the nucleobase).
This stimulated us to investigate the possibility of exploiting the
simultaneous use of different nucleotides to increase the dynamic
detection range, which would emphasize the unique possibility
offered by the signal transduction pathway that is operative in this
self-assembled sensing system.

2.1. Dynamic detection range

Previous studies had shown that the sensitivity of the system is

much higher when TDP is used (compared to TMP or cTMP),
because of the increased number of negative charges.21 We pre-
sumed that the combination of TDP and cTMP in the same system
would enlarge the dynamic detection range with TDP operating in
the nanomolar regime and cTMP in the micromolar regime. The
anionic probe A was selected because of the following reasons: i)
the high quantum yield of coumarin343 (lex ¼ 445 nm;
lem ¼ 493 nm) is advantageous for creating a response even at low
concentrations, ii) the carboxylate-probe is readily displaced by
phosphate competitors, and iii) the fluorescence properties of the
probe are not affected by the analytes.28
The detection range for TDP and cTMP separately was deter- Fig. 2. (a) Normalized change in fluorescence intensity (a.u.) at 493 nm as a function of
mined by measuring the fluorescence intensity after the addition of the concentration of Hg2þ added to a solution containing Au NP 1, probe A, and TDP
increasing amounts of Hg2þ to a buffered aqueous solution con- (orange), cTMP (green), or a mixture of TDP and cTMP (blue). Experimental conditions:
[TACN$Zn2þ] ¼ 20 ± 1 mM; [A] ¼ 7.3 mM, [TDP] ¼ 16 mM, [cTMP] ¼ 800 mM,
taining Au NP 1 ([TACN$Zn2þ] ¼ 20 ± 1 mM), A (7.3 mM) and either [HEPES] ¼ 10 mM, pH 7.0, T ¼ 37  C, fluorescence slit width ¼ (2.5/5) nm. (b) Hg2þ-
TDP (16 mM) or cTMP (800 mM). A much higher concentration of detection range in the presence of TDP (orange), cTMP (green), or a mixture of TDP and
cTMP was used compared to TDP, because a much higher cTMP (blue).
4952 S. Maiti, L.J. Prins / Tetrahedron 73 (2017) 4950e4954

hand, when cTMP was present the increase in fluorescence intensity that are able to complex the added analyte. Thus, the addition of
started at 2 mM Hg2þ and continued until 50 mM. These results Hg2þ to a mixture of 4 nucleotides (dAMP, dGMP, TMP, and dCMP)
clearly show that the negative charges present in the probes dictate resulted in the self-selection of TMP as the optimal recognition
the detection range. Next, the response of the system was measured element for Hg2þ metal ions. On the other hand, it was found that
in the presence of both TDP and cTMP at the same concentrations as the addition of Agþ to the same mixture of nucleotides resulted in
before. We were very pleased to observe that also in this case already the capturing of dGMP on Au NP 1. As such, this was a proof-of-
at low nanomolar concentrations of Hg2þ a fluorescence signal was principle demonstration that self-selection procedures can indeed
generated following the response curve of only TDP. Yet, at around be used to affect the composition of dynamic responsive multiva-
5 mM of Hg2þ the new response curve deviated and continued at lent surfaces. Here, we show a logical follow-up of that study in
lower intensity reaching a final plateau level at around 50 mM of which we exploit the information obtained from these self-
Hg2þ similar to that observed for cTMP alone. The change in selection experiments for the development of a sensing system
response curve at around 5 mM of Hg2þ originates from the fact that able to selectively detect either Agþ or Hg2þ metal ions (or both) at
the depletion of TDP starts to statistically favour the formation of low nanomolar concentrations in water.
complexes TDP$Hg2þ$cTMP first and cTMP$Hg2þ$cTMP later. The The results described in the previous section have already
result is a significant increase in the dynamic range of the sensing shown that the use of thymidine nucleotides as recognition module
system which ranges now from 20 nM to 50 mM with linearity in the provides a system that produces a fluorescent signal upon the
0e5 mM concentration range (Fig. 2b). addition of Hg2þ. Next, we were interested whether the replace-
ment of the TMP for GMP, while keeping all other components the
2.2. Sensitivity same, would make the system responsive to Agþ instead. Thus,
titration experiments were performed by adding increasing
Previously, we exploited a dynamic combinatorial approach for amounts of Agþ metal ions to a buffered aqueous solution con-
the target driven self-selection of recognition units on the surface taining Au NP 1 ([TACN$Zn2þ] ¼ 20 ± 1 mM), A (7.3 mM) and dGMP
of Au NP 1.28 In line with the approach described above, the ability (6 mM) followed by measurement of the fluorescence intensity after
of target metal ions to induce clustering of receptor units led to an each addition (Fig. 3a). We were pleased to observe an increase in
accumulation of only those receptor units on the monolayer surface fluorescence as a function of Agþ concentration, but regrettably

Fig. 3. (a) Change in fluorescence intensity (a.u.) at 493 nm as a function of the concentration of Agþ added to a solution containing Au NP 1, probe A (7.3 mM) and GDP (circles) or
dGMP(squares). (b) Fluorescence intensity (a.u.) at 493 nm as a function of the concentration of Agþ (50e2000 nM). Experimental conditions: [TACN$Zn2þ] ¼ 20 ± 1 mM;
[A] ¼ 7.3 mM, [HEPES] ¼ 10 mM, pH 7.0, T ¼ 37  C, [dGMP] ¼ [GDP] ¼ 6 mM (for Fig. 3a); [GDP] ¼ 6 mM (for Fig. 3b); fluorescence slit width ¼ (2.5/5) nm (for Fig. 3a) and (5/5) nm (for
Fig. 3b).
 S. Maiti, L.J. Prins / Tetrahedron 73 (2017) 4950e4954 4953

with a low signal strength in the low micromolar region (DFIdGMP/D Previously we had shown that the use of the TDP recognition unit
[Agþ] ¼ 1.0), indicating that the complex formed between Agþ and gave a highly selective sensing system for Hg2þ. For comparison,
dGMP has a poor capacity to displace probe A from Au NP 1. the already published data from a comparative study between a
However, the modular nature of the sensing systems permits a series of 12 metal ions (amongst which include Pb2þ, Cd2þ, Pd2þ,
straightforward solution to this problem by increasing the number Agþ, As5þ) at a constant concentration of 1 mM are reproduced in
of negative charges in the nucleotide. Since the transduction Fig. 4a. A repetition of the same experiment using GDP instead of
element (the phosphates) is independent from the recognition TDP completely changed the selectivity of the system for Agþ. Only
element (the nucleobase), this alteration does not (or to a very for Agþ a significant increase in fluorescence intensity was
minimal extent) affect the interaction with the analyte. Indeed, observed, indicating the inability of the other metal ions to form
when the experiment was repeated using GDP instead of dGMP (at complexes with GDP able to compete with probe A for binding to
the same concentration) a 40-fold stronger response (DFIGDP/D Au NP 1 (Fig. 4b). Similar as observed for Hg2þ, the increase in
[Agþ] ¼ 44.7) was measured (Fig. 3a). The use of GDP and optimized fluorescence intensity induced by Agþ (1 mM) was hardly affected
instrument settings permitted the obtainment of a linear response by the presence or not of a mixture of all other metal ions (1 mM
curve for Agþ in the 100e2000 nM concentration range (Fig. 3b). each; final column, Fig. 4b).
A peculiar feature of the sensing system presented here is that
the selectivity of the system for Hg2þ or Agþ is determined by the
2.3. Selectivity type of nucleobase present in the nucleotide (thymine or guanine,
respectively). However, after this selective recognition event the
Finally, we were interested in the selectivity of the sensing signal transduction pathway is identical for both analytes, because
system as a function of the added nucleotide (TDP or GDP).

Fig. 4. (a) Changes in the fluorescence intensity at 493 nm upon the addition of 1 mM of a series of metal ions to a solution containing Au NP 1, probe A, and TDP. Reproduced with
permission from reference.21 (b) Changes in the fluorescence intensity at 493 nm upon the addition of 1 mM of a series of metal ions to a solution containing Au NP 1, probe A, and
GDP. (c) Changes in the fluorescence intensity at 493 nm upon the addition of 1 mM of a series of metal ions to a solution containing Au NP 1, probe A, TDP and GDP. Experimental
conditions 4aec: [TACN$Zn2þ] ¼ 20 ± 1 mM; [CGDD] ¼ 7.3 mM, [TDP] ¼ 16 mM (for 4a) and 10 mM (for 4c), [GDP] ¼ 6 mM (for both 4b and c), [HEPES] ¼ 10 mM, pH 7.0, T ¼ 37  C,
fluorescence slit width ¼ 5/5 nm (for 4a) and 2.5/5 nm (for 4b and c).
4954 S. Maiti, L.J. Prins / Tetrahedron 73 (2017) 4950e4954

the affinity of the formed complexes for Au NP 1 originates in both were purchased from Sigma-Aldrich and their stock solutions were
cases from the same phosphate groups. This gives the possibility to prepared by weight.
develop an assay that reports on the presence of either one (or both) UVeVisible spectra were measured on a Varian Cary50 spec-
of the analytes Hg2þ or Agþ in a complex mixture simply by using trophotometer equipped with thermostatted multiple cell holders.
the two recognition elements TDP and GDP contemporarily. This Fluorescence measurements were performed on a Varian Cary
was demonstrated by an experiment in which both TDP (10 mM) and Eclipse fluorescence spectrophotometer also equipped with a
GDP (6 mM) were present. Their concentrations were chosen such thermostatted cell holder.
that 1 mM of Hg2þ and Agþ would give a comparable increase of Protocols for the displacement experiments and selectivity
fluorescence intensity (DFIHg2þ z 41a.u.; DFIAgþ z 43a.u.). Indeed, a studies have been described before.21,28
screening of the set of metal ions now gave positive signals when The normalization in Fig. 2a has been done by following the
either Hg2þ or Agþ was added, but not for any of the other metals equation mentioned below:
(Fig. 4c). As a side note, it is of interest to note that the simultaneous Normalized value ¼ (F e Fmin)/(Fmax e Fmin); where
addition of both Hg2þ and Agþ resulted in an increase of fluores- F ¼ fluorescence value in presence of any concentration of added
cence intensity practically equal to the sum of the individual values Hg2þ ion, Fmin ¼ fluorescence value when no Hg2þ ion was added,
(Fig. 4c). Response systems like this are attracting interest as logic Fmax ¼ fluorescence value in presence of maximum amount of
gates for chemical computing.32,33 added Hg2þ ion. Fmin and Fmax values for TDP were 176 and 427; for
TDP þ cTMP were 181 and 419; for cTMP were 3.5 and 233,
3. Conclusions respectively under the experimental condition described in the
main text.
In conclusion, we have developed a self-assembled sensing
system for the selective detection of Hg2þ and Agþ at nanomolar Acknowledgements
concentrations in water. Apart from the strong analytical perfor-
mance, the novelty of the system is that it is composed of modules This work was financially supported by the European Commis-
that can be independently changed to adapt the selectivity, sensi- sion (grant MSCA 657486) through the H2020-program.
tivity, and, in principle, also the output signal. This is a unique
feature when compared to numerous other nanoparticle based References
sensing systems reported in the literature, that typically rely on the
presence of recognition units that are covalently attached to the 1. Wu J, Kwon B, Liu W, Anslyn EV, Wang P, Kim JS. Chem Rev. 2015;115:7893.
2. Patel PD. Trends Anal Chem. 2002;21:96.
monolayer.12,13 It is emphasized that all components of the system 3. Johnson KJ, Rose-Pehrsson SL. Annu Rev Anal Chem. 2015;8:287.
are either commercially available or synthesized in a few steps 4. Chinen AB, Guan CM, Ferrer JR, Barnaby SN, Merkel TJ, Mirkin CA. Chem Rev.
(including Au NP 1). In addition, contrary to most other thymine 2015;115:10530.
5. Zhang S-W, Swager TM. J Am Chem Soc. 2003;125:3420.
and guanosine-based sensing systems for the detection of Hg2þ and 6. Pezzato C, Maiti S, Chen JL-Y, Cazzolaro A, Gobbo C, Prins LJ. Chem Commun.
Agþ, respectively, unmodified nucleotides are used, which permits 2015;51:9922.
a straightforward modulation of the systems response. Currently, 7. Hagiya M, Konagaya A, Kobayashi S, Saito H, Murata S. Acc Chem Res. 2014;47:
1681.
we are working at expanding the scope of the assay by incorpo-
8. Albert KJ, Lewis NS, Schauer CL, et al. Chem Rev. 2000;100:2595.
rating appropriate recognition modules for the recognition of small 9. De Silva AP, Fox DB, Huxley AJM, Moody TS. Coord Chem Rev. 2000;205:41.
(bio)molecules. 10. Duke RM, Veale EB, Pfeffer FM, Kruger PE, Gunnlaugsson T. Chem Soc Rev.
2010;39:3936.
11. Prins LJ. Acc Chem Res. 2015;48:1920.
4. Experimental section 12. Saha K, Agasti SS, Kim C, Li X, Rotello VM. Chem Rev. 2012;112:2739.
13. Jain PK, Huang X, El-Sayed IH, El-Sayed MA. Acc Chem Res. 2008;41:1578.
The synthesis and characterization of Au NP 1 has been 14. Lavigne JJ, Anslyn EV. Angew Chem Int Ed. 2001;40:3118.
15. Rana S, LeNgoc DB, Mout R, et al. Nat Nanotechnol. 2015;10:65.
described elsewhere.31 Stock solutions of Au NP 1 were preserved 16. Lee B, Chen S, Heinis C, Scopelliti R, Severin K. Org Lett. 2013;15:3456.
at 4  C in mQ water. The concentration of TACN-head groups in Au 17. He L, Dong B, Liu Y, Lin W. Chem Soc Rev. 2016;45:6449.
NP 1 was determined from kinetic titrations as reported previ- 18. Neri S, Pinalli R, Dalcanale E, Prins LJ. ChemNanoMat. 2016;2:489.
19. Liu C-L, Zhou L-P, Tripathy D, Sun Q-F. Chem Commun. 2017;53:2459.
ously.30,31 Zn(NO3)2 stock solution were standardized using EDTA 20. Salvia M-V, Salassa G, Rastrelli F, Mancin F. J Am Chem Soc. 2015;137:11399.
following standard procedures. The synthesis and characterization 21. Maiti S, Pezzato C, Garcia Martin S, Prins LJ. J Am Chem Soc. 2014;136:11288.
of the fluorescence probe (A) has been described before.30 22. Broman SL, Andersen CL, Jevric M, Tortzen CG, Hammerich O, Nielsen MB.
Tetrahedron. 2016;72:5831.
The buffer, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid 23. Zhang Y, Jiang H, Wang X. Anal Chim Acta. 2015;870:1.
(HEPES) was obtained from Sigma-Aldrich and used without 24. Lee J-S, Mirkin CA. Anal Chem. 2008;80:6805.
further purification. All the nucleotides (TDP, cTMP and GDP, GMP) 25. Ono A, Togashi H. Angew Chem Int Ed. 2004;43:4300.
26. Gruelich C, Braun D, Peetsch A, et al. RSC Adv. 2012;2:6981.
were obtained from Sigma-Aldrich and their stock solutions were
27. Chen Q, Wu X, Wang D, Tang W, Li N, Liu F. Analyst. 2011;136:2572.
prepared both by weight and UVevis spectroscopy using the 28. Maiti S, Prins LJ. Chem Commun. 2015;51:5714.
following molar extinction co-efficients: 3268 (TDP, 29. Garcia Martin S, Prins LJ. Chem Commun. 2016;52:9387.
cTMP) ¼ 9600 M1cm1; ε272 (GDP, GMP) ¼ 13700 M1cm1.30 All 30. Pezzato C, Lee B, Severin K, Prins LJ. Chem Commun. 2013;49:469.
31. Pieters G, Cazzolaro A, Bonomi R, Prins LJ. Chem Commun. 2012;48:1916.
metal salts [Hg(NO3)2, AgNO3, Cu(NO3)2, Fe(NO3)3, Pb(NO3)2, 32. De Silva AP, Uchiyama S. Nat Nanotechnol. 2007;2:399.
Ni(NO3)2, Co(NO3)2, NaAsO4, KNO3, Cd(NO3)2, Ca(NO3)2, PdCl2] 33. Willner I, Shlyahovsky B, Zayats M, Willner B. Chem Soc Rev. 2008;37:1153.