Establishment of Logic Gates Based on Conformational Changes in a

May 13, 2019 - DNA-based logic gates stimulate the development of molecular scale ..... logic gate, and reversibility evaluation of the proposed syste...
0 downloads 0 Views 1MB Size
Download PDF
Letter pubs.acs.org/ac

Cite This: Anal. Chem. 2019, 91, 6971−6975

Establishment of Logic Gates Based on Conformational Changes in a Multiple-Factor Biomolecule Interaction Process by Dual Polarization Interferometry Shuang Wang,†,‡ Jiahui Zhao,†,§ Shasha Lu,†,‡ Jianshe Huang,*,† and Xiurong Yang*,†,‡,§ †

State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry, Changchun, Jilin 130022, China University of Science and Technology of China, Hefei, Anhui 230026, China § University of Chinese Academy of Sciences, Beijing 100039, China Downloaded via 5.101.220.51 on July 19, 2019 at 08:58:44 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

‡

S Supporting Information *

ABSTRACT: DNA-based logic gates stimulate the development of molecular scale computers and show enormous potential in nanotechnology, biotechnology, and medicine. However, the reported detectors to date usually require one to label appropriate signal probes, resulting in not only a high cost but also potentially tedious manipulation. For the first time, we established a label-free logic gate by regarding the structure-related signal as output. Dual polarization interferometry (DPI) was employed to reveal the detailed conformational transitions occurring in the multiple-factor biomolecule interactions and then was utilized as a detection tool of logic gate. As a vital merit of this system, the dependence of the density output signal on the interaction with multiple-factor input can mimic the function of signal communication in OR, INHIBIT, and IDENTITY logic gates and the INHIBIT−OR cascade circuit. Additionally, the DPI signal with logic stringency can unambiguously distinguish conformational polymorphisms and compare structural stability. This study provides a new way for the construction of a label-free logic gate, supplements information deficiency of reaction details, and extends the application of DPI in logic operation.

S

readout and a long waveguide, the DPI technique possesses higher sensitivity and accuracy than the usual structuremeasured technology, such as nuclear magnetic resonance (NMR), X-ray crystallography, and high resolution mass spectrometry (HRMS). Thus, it is of special interest to explore the structural changes of protein,19 DNA,20,21 lipid,22 biomembrane,23 polymer,24 and the functional surface25 by DPI. The above properties manifest that DPI undoubtedly is an ideal detection tool for structure-based logic operation and the biomolecule interaction process. The exploration of the biomolecule interaction at the molecular level is beneficial for understanding the roles of the primary factors including ions and energetic contents in functions of the human body and interconnections with diseases.26,27 As a powerful tool, DPI was widely used to investigate the interaction processes, such as simulating the pathogenesis of Alzheimer’s disease by monitoring binding events of the amyloid-β peptide in different morphologies and its receptor28 and deeply investigating the interaction mechanism of the antimicrobial peptide and biomimic

ince Adleman solved a complex mathematical problem by only performing logical operations based on DNA oligonucleotides in 1994,1 molecular logic gates have stimulated the development of molecular-scale computers and shown enormous potential in nanotechnology, biotechnology, and medicine. DNA has attracted a great deal of attention as a potential material for designing logic gates due to its wellordered structure, highly specific recognition, and predictable merits. DNA-based logic gates are usually detected by fluorescent,2−7 colorimetric,8−10 and other electronic methods.11−14 However, these reported detectors to date usually require one to label an appropriate fluorophore/electron donor and quencher/acceptor or fluorescent/electrochemical probes, resulting in not only a high cost but also potentially tedious manipulation. Furthermore, various reported DNA logic gates15−17 utilized different output signals between the initial and final states of structure changes to execute logic operation, but there is no demonstration of direct use of structure-related signal as output to implement a logic gate. Dual polarization interferometry (DPI) described as a molecular ruler is one of the most powerful and precise techniques to label-free and real-time record the dynamic conformational information, whose resolving dimension reaches picometer scale.18 Due to exploiting interferometric © 2019 American Chemical Society

Received: March 14, 2019 Accepted: May 13, 2019 Published: May 13, 2019 6971

DOI: 10.1021/acs.analchem.9b01319 Anal. Chem. 2019, 91, 6971−6975

Letter

Analytical Chemistry membrane to develop effective antibacterial drugs.29 However, the above studies focus on the effect of one factor on target, which is not in line with the complex human body environment. Thus, establishing a model to simulate a complex reaction in vivo is highly significant and emerging but still has a great challenge. Further incorporating multiple-factor interaction into the design of logic operation would not only accomplish the sophisticated logic circuit but also endow the complex system with logic stringency. In this study, a successive conformational interconversion among a single stranded, double helix, and G-quadruplex (G4) structure of nucleic acid modulated by multiple factors is presented as a model. We chose the AS1411 aptamer to realize the transition of DNA structures, because it consists of guanine-rich bases that guarantee the formation of G4 and thymine-rich bases that could form a duplex with the help of different ions. As depicted in Scheme 1, we used positively

two strong positive peaks at 210 and 265 nm appeared, representing the formation of a parallel G4 structure, but the CD peaks of Pb2+-stabilized G4 were stronger; also there was an additional negative peak at 300 nm,36 which suggested the polymorphisms of the same parallel G4 was modulated by different ions. With the addition of Hg2+, the decreased negative band at 241 nm and positive band at 262 nm verified the formation of a stem-loop structure originating from specific binding between Hg2+ and the T−T mismatch.37 The control DNA had no evident changes, no matter which ion was added (Figure S1B). In the presence of K+ and Pb2+, the spectrum was basically in line with that in the presence of solely Pb2+. Additionally, as long as Hg2+ existed, the CD peaks were located at 241 and 262 nm (Figure S1C,D), which is the same as that of AS1411-Hg2+. We also employed two fluorescent dyes N-methylmorpholine (NMM) and SYBR Green II to prove the conformational transitions. NMM as a typical G4 binder shows distinctly enhanced fluorescence when binding with G4.38 The fluorescent spectra in Figure S1E demonstrate that K+-stabilized G4 exhibits enhanced fluorescence, while Pb2+-stabilized G4 shows faint fluorescence, because Pb2+stabilized G4 had almost no ability to bind porphyrins.33 When Hg2+ was added in the whole system, the fluorescence was decreased obviously, because the binding site of NMM was destroyed in the T−Hg2+−T duplex structure. SYBR Green II is one of most sensitive dyes known for detecting RNA or single stranded DNA.39 The fluorescence of AS1411 incubated by K+, Pb2+, or both of them was weaker than free AS1411, while as long as Hg2+ existed, the fluorescence decreased further (Figure S1F). The reason might be that G4 was closer to a free single strand, which was beneficial for the binding of SYBR Green II, while the compact T−Hg2+−T duplex prohibited the binding. The phenomenon was also in accordance with the CD spectra. All of these results verify the stability of different conformations: T−Hg2+−T duplex > Pb2+-stabilized G4 > K+-stabilized G4 > single strand, which meant the conformational transition is feasible. The whole multiple-factor interaction process among oligonucleotides, ions, and small molecules was monitored using DPI. When PEI/AS1411/Pb2+/Hg2+/Cys (Figure 1) was taken as an example, the abrupt changes in mass, thickness, and density at the beginning of the addition of PEI, AS1411, Pb2+, Hg2+, and Cys were observed; then, all of them gradually

Scheme 1. Mechanism of the Multifactor Biomolecule Interactiona

a

Construction of the biosensing interface and structure transition due to the multifactor interaction by the DPI technique.

charged PEI as a connector to immobilize negatively charged AS1411 on chip according to a previous study in our group.30,31 Then, multiple factors (K+, Pb2+, Hg2+, and cysteine) were introduced into the sensing surface in sequence. K+ and Pb2+ were introduced to realize the conversion from a single strand to G4.32,33 Furthermore, Pb2+-stabilized G4 was more stable than K+-stabilized G4 due to the formation of more compact DNA folds.34 On the basis of this, the unidirectional conformation transition between two kinds of G4 was implemented. Furthermore, a Hg2+-mediated T−T base pair could modulate the proper folding of G4 DNAs,35 inducing the formation of a duplex structure. Finally, the robust affinity between Hg2+ and cysteine (Cys) prompted the detachment of Hg2+ from T−Hg2+−T complexes, followed by transforming the conformation to a single strand again. Making use of the density change generated from the above structural transition, we successfully establish typical OR, INHIBIT, and IDENTITY logic gates and realize the cascade of INHIBIT and OR. From the structural signal with logic stringency, we could easily distinguish the unintelligible polymorphic conformations of G4 and the stability of different conformations as well as understand the multiple-factor interaction process.

■

Figure 1. Whole structural transition process. DPI-based real-time measurements of mass (red color), thickness (blue color), and density (black color) for the whole immobilization process of PEI (a) and AS1411 (b) and the interaction process with Pb2+ (c), Hg2+ (d), and Cys (e) on the bare chip surface.

RESULTS AND DISCUSSION Circular dichroism (CD) spectra were used to preliminarily prove the feasibility of our design. As depicted in Figure S1A, upon addition of K+ or Pb2+, a negative peak at 240 nm and 6972

DOI: 10.1021/acs.analchem.9b01319 Anal. Chem. 2019, 91, 6971−6975

Letter

Analytical Chemistry reached a stable state after finishing the sample injection. From the height of the platforms, we could roughly compare the mass, thickness, and density of each stabilized layer. Combined with the three key parameters (thickness, mass, and density) of every stable layer in Table S1, the general variation trend could be achieved. For instance, AS1411 layers had a lower density level (changing from 0.794 to 0.757) because they had more water-swollen properties than the PEI layers, which was in accordance with the results of previous reports.31,40 The evident changes of Pb2+ (increased thickness, mass, and density) were ascribed to the formation of compact G4. Upon the further introduction of Hg2+, the increased density (from 0.763 to 0.846 in Table S1) at a rapid rate implied that mismatched T−T could bind Hg2+ to form a more compact state. Cys captured Hg2+ from T−Hg2+−T complexes, resulting in the change from rigid hairpin-like to random coil conformation (the thickness, density, and mass decreased from 3.681 to 3.625, 0.846 to 0.764, and 3.113 to 2.766, respectively), which agreed well with those earlier observations that double stranded DNA formed a thicker layer than single stranded DNA.31 (A similar PEI/AS1411/K+/Hg2+/Cys interaction process was shown in Figure S2.) The results powerfully proved DPI could achieve structural transition information from beginning to end that can be regarded as output signal to accomplish logic operations. In order to display the PEI/AS1411/Pb2+/Hg2+/Cys process more explicitly, the thickness and density of every stable layer were plotted as a function of respective mass loading (Figure S3). The adsorption rate could be inferred from the intensive degree of data points because of the data collected at the same time interval.40 Just as the distribution of the data points was scattered at the start of sample adsorption, suggesting the adsorption rate was fast. Subsequently, the data points were dense, suggesting the association and disassociation gradually reached equilibrium, accompanied by conformational rearrangement to a slight degree. Along the direction of the arrows, we could distinguish the conformational change at each step in the interaction process. For AS1411, the density decreased and the thickness increased in a linear scale as the mass increased, suggesting the capture probes were adsorbed on the sensing platform continually; however, at last, the thickness decreased and the density increased as the mass decreased to some extent according to the original route, demonstrating the layer underwent a slight disassociation and conformational rearrangement.31 When Pb2+ was involved, the density and thickness increased as the mass increased, which meant Pb2+stabilized G4 was stable and in a compact state. Upon substitution of Pb2+ by Hg2+, both the thickness and density increased along with the increase of mass, suggesting that this whole dynamic process might be Hg2+ continually entering the G4 and aiming at the position of T−T bases; eventually, the conformation changed from Pb2+-stabilized G4 to a T−Hg2+− T stem loop structure. The changes of layer structure after addition of Cys were divided into three sections. At the start of the injection, the thickness and density decreased with the mass decreasing, which might be ascribed to desorption induced by flow. Next, when more Cys was introduced into the channel, more Cys captured the Hg2+ from the T−Hg2+−T complexes, but there was not enough time to wash them away in the restricted space, which corresponded to the increased thickness and the slightly decreased density as the mass increased. At the end of the process, the density increased and thickness decreased continuously as the mass decreased until

all of Hg2+ bonded in the AS1411 was captured by Cys. In the PEI/AS1411/K + /Hg 2+ /Cys interaction process, similar changes trends were observed except that the density displayed a completely opposite trend upon adding K+, which was ascribed to the formation of incompact G-quartets stabilized by H-bonding and cation binding. This feature could be used to develop a discriminator of G4 modulated by different ions. According to the detailed structural changes in the process of the multiple-factor interaction, the logic gates were developed by assigning the density signal as the output and the multiple-factors as inputs. The density of 0 was set as threshold value. The increased density (higher than 0) was defined as the output of 1, and the decreased density (lower than 0) was defined as the output of 0 after introducing diversified inputs. In this logic system, the AS1411/PEI layer was considered as the initial (0/0) state, showing decreased density and giving an output of 0. When Pb2+ and Hg2+ were regarded as two inputs, either of them (1/0, 0/1) and both of them (1/1) gave outputs of 1 because the compact Pb2+stabilized G4 and T−Hg2+−T duplex structures led to the increased density, in which a typical OR logic gate was achieved (Figure 2).

Figure 2. “OR” logic gate. (A) Schematic illustration of the operational results of the “OR” gate. (B) The density changes recorded at 150 s upon adding different inputs. (C) The corresponding density changes measured by DPI when introducing different inputs. (D) The truth table of the “OR” logic gate. (E) Electronic equivalent circuitry.

While using the two input pairs of Hg2+ and Cys could establish INHIBIT logic gate (Figure 3), only in the presence of the input of Hg2+ (1/0) did the density increase, leading to a 1 output. Otherwise, any one input including Cys by itself (0/ 1) or both Hg2+ and Cys (1/1) decreased the density, which yielded 0 output. The reason is that Cys could not bind with AS1411 tightly and disturb the conformational changes by itself, but its strong affinity for Hg2+ would take the Hg2+ out of T−Hg2+−T duplex; additionally, the conformation changed from compact duplex to random single strand. As expected, the cascade of INHIBIT and OR logic gates was implemented by introducing three inputs Hg2+/Cys/Pb2+. The DPI signal in Figure 4 suggested that the existence of Pb2+ and Hg2+ on their own, both of them, Pb2+ and Cys, and the mixture of the above three samples could produce increased density, corresponding to 1/0/0, 0/0/1, 1/0/1, 0/1/1, and 1/ 1/1, respectively. While introducing nothing, Cys and both Hg2+ and Cys (0/0/0, 0/1/0, 1/1/0) displayed decreased density, giving an output of 0. This is because the T−Hg2+−T duplex and Pb2+-stabilized G4 were more compact structures 6973

DOI: 10.1021/acs.analchem.9b01319 Anal. Chem. 2019, 91, 6971−6975

Letter

Analytical Chemistry

output value, we could distinguish the different configurations of G4 without any label and probes. To investigate the reversibility of our proposed system, we have taken Pb2+ input in OR logic gate as an example. AS1411 changed from a single strand to Pb2+-stabilized G4 when adding Pb2+, displaying an output of 1. Then, Hg2+ was introduced to turn Pb2+-stabilized G4 into the T−Hg2+−T duplex. Finally, Cys was used to capture the Hg2+, and the T− Hg2+−T duplex changed to single strand again. As shown in Figure S6, the AS1411 could reversibly execute the logic operation for 5 times though the density signal had a distinct decrease. The decreased density signal might ascribe to the desorption of AS1441 from the PEI layer. The conformations caused by other inputs in different logic gates could be returned to the original single strand state according to the above method. These results show the acceptable reversibility of the proposed logic gates.

Figure 3. “INHIBIT” logic gate. (A) Schematic illustration of the operational results of the “INHIBIT” gate. (B) The density changes recorded at 150 s upon adding different inputs. (C) The corresponding density changes measured by DPI when introducing different inputs. (D) The truth table of the “INHIBIT” logic gate. (E) Electronic equivalent circuitry.

■

CONCLUSIONS In conclusion, the detailed conformational transitions occurring in the multiple-factor biomolecule interactions were visualized by DPI and then were utilized subtly to operate the logic gate. As a vital merit of this system, the dependence of the density output on the interaction with multiple-factor inputs can mimic the function of signal communication in OR, INHIBIT, and IDENTITY logic gates and would be regarded as a DPI-detected logic gate system for the first time. The DPI signal with logic stringency can unambiguously distinguish conformational polymorphisms and compare structural stability. This study supplements the information deficiency of reaction details and extends the application of DPI in logic operation. This new method of establishing a logic gate also lays the foundation of constructing a practical logic sensing system.

■

ASSOCIATED CONTENT

S Supporting Information *

Figure 4. “INHIBIT−OR” logic gate combination. (A) Schematic illustration of the operational results of the “INHIBIT−OR” gate. (B) The corresponding density changes measured by DPI when introducing different inputs. (C) The density changes histogram of different inputs. (The outputs corresponded to the density values recorded at 150 s in (B). (D) The truth table of the “INHIBIT−OR”. (E) Electronic equivalent circuitry.

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.9b01319. Experimental section, CD and fluorescent characterization, structural parameters, the whole structural transition process of PEI/AS1411/K+/Hg2+/Cys, density and thickness changes as the mass changes, IDENTITY logic gate, and reversibility evaluation of the proposed system (PDF)

than the single stranded state; and only Hg2+ existed in the system, Cys could play its role in suppressing the formation of the T−Hg2+−T duplex. We also achieved IDENTITY gate based on the competitive interaction of K+ and Pb2+ and K+ and Hg2+ with AS1411. Integrating the density change induced by K+ and Pb2+ into one system (Figure S4), we could find that the density of K+stabilized G4 decreased distinctly (1,0), accompanied with an output of 0, while that of Pb2+-stabilized G4 increased (0,1), outputting 1, because Pb2+ could form shorter Pb−O and O− O bonds than K+-stabilized G4.33 Therefore, even though K+ and Pb2+ coexisted in one system (1,1), AS1411 kept the Pb2+stabilized G4 conformation, giving an output of 1, which is in accord with the results of CD. While regarding K+ and Hg2+ as inputs, the system outputs 1 only when Hg2+ is present, also fitting the feature of an IDENTITY gate.41 The detailed logic operation results were displayed in Figure S5. Simply from the

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Tel.: +86 431 85262056. *E-mail: [email protected]. Tel.: +86 431 85262964. ORCID

Shasha Lu: 0000-0002-1289-1248 Xiurong Yang: 0000-0003-0021-5135 Author Contributions

The manuscript was written through contributions of all authors. Notes

The authors declare no competing financial interest. 6974

DOI: 10.1021/acs.analchem.9b01319 Anal. Chem. 2019, 91, 6971−6975

Letter

Analytical Chemistry

■

(30) Wang, J.; Coffey, P. D.; Swann, M. J.; Yang, F.; Lu, J. R.; Yang, X. Anal. Chem. 2010, 82, 5455−5462. (31) Wang, Y.; Zheng, Y.; Yang, F.; Yang, X. Chem. Commun. 2012, 48, 2873−2875. (32) Sen, D.; Gilbert, W. Nature 1990, 344, 410−414. (33) Li, T.; Wang, E.; Dong, S. J. Am. Chem. Soc. 2009, 131, 15082− 15083. (34) Smirnov, I. V.; Kotch, F. W.; Pickering, I. J.; Davis, J. T.; Shafer, R. H. Biochemistry 2002, 41, 12133−12139. (35) Williamson, J. R. Annu. Rev. Biophys. Biomol. Struct. 1994, 23, 703−730. (36) Bagheri, Z.; Ranjbar, B.; Latifi, H.; Zibaii, M. I.; Moghadam, T. T.; Azizi, A. Int. J. Biol. Macromol. 2015, 72, 806−811. (37) Chen, H.; Zhang, X.; Sun, H.; Sun, X.; Shi, Y.; Xu, S.; Tang, Y. Analyst 2015, 140, 7170−7174. (38) Oh, S. S.; Plakos, K.; Lou, X. H.; Xiao, Y.; Soh, H. T. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 14053−14058. (39) Luo, Q.; Chen, B.; Xu, J.; Ma, W.; Lao, C.; Li, Y.; Tan, J.; Tang, Y.; Huang, C.; Liu, W.; Chen, Z. Intervirology 2019, 61, 230−236. (40) Zheng, Y.; Yang, C.; Yang, F.; Yang, X. Anal. Chem. 2014, 86, 3849−3855. (41) Zhou, C.; Wang, K.; Fan, D.; Wu, C.; Liu, D.; Liu, Y.; Wang, E. Chem. Commun. 2015, 51, 10284−10286.

ACKNOWLEDGMENTS We are thankful for the financial support from the National Key Research and Development Program of China (2016YFA0201301), the National Natural Science Foundation of China (Nos. 21435005, 21627808, and 21605139), and Key Research Program of Frontier Sciences, CAS (QYZDY-SSWSLH019)

■

REFERENCES

(1) Adleman, L. M. Science 1994, 266, 1021−1024. (2) Elbaz, J.; Wang, F.; Remacle, F.; Willner, I. Nano Lett. 2012, 12, 6049−6054. (3) Freeman, R.; Finder, T.; Willner, I. Angew. Chem., Int. Ed. 2009, 48, 7818−7821. (4) Zadegan, R. M.; Jepsen, M. D.; Hildebrandt, L. L.; Birkedal, V.; Kjems, J. Small 2015, 11, 1811−1817. (5) Tang, W.; Wang, H.; Wang, D.; Zhao, Y.; Li, N.; Liu, F. J. Am. Chem. Soc. 2013, 135, 13628−13631. (6) Pei, H.; Liang, L.; Yao, G.; Li, J.; Huang, Q.; Fan, C. Angew. Chem., Int. Ed. 2012, 51, 9020−9024. (7) He, K.; Li, Y.; Xiang, B.; Zhao, P.; Hu, Y.; Huang, Y.; Li, W.; Nie, Z.; Yao, S. Chem. Sci. 2015, 6, 3556−3564. (8) Bi, S.; Yan, Y.; Hao, S.; Zhang, S. Angew. Chem., Int. Ed. 2010, 49, 4438−4442. (9) Xu, X.; Zhang, J.; Yang, F.; Yang, X. Chem. Commun. 2011, 47, 9435−9437. (10) Freeman, R.; Sharon, E.; Tel-Vered, R.; Willner, I. J. Am. Chem. Soc. 2009, 131, 5028−5029. (11) Zhu, X.; Xu, H.; Gao, X.; Li, X.; Liu, Q.; Lin, Z.; Qiu, B.; Chen, G. Chem. Commun. 2011, 47, 9080−9082. (12) Li, X.; Ding, T.; Sun, L.; Mao, C. Biosens. Bioelectron. 2011, 30, 241−248. (13) Pérez-Inestrosa, E.; Montenegro, J.-M.; Collado, D.; Suau, R.; Casado, J. J. Phys. Chem. C 2007, 111, 6904−6909. (14) Xiang, Y.; Qian, X.; Chen, Y.; Zhang, Y.; Chai, Y.; Yuan, R. Chem. Commun. 2011, 47, 2080−2082. (15) Seelig, G.; Soloveichik, D.; Zhang, D. Y.; Winfree, E. Science 2006, 314, 1585−1588. (16) Lake, A.; Shang, S.; Kolpashchikov, D. M. Angew. Chem., Int. Ed. 2010, 49, 4459−4462. (17) Wang, S.; Sun, J.; Zhao, J.; Lu, S.; Yang, X. Anal. Chem. 2018, 90, 3437−3442. (18) Escorihuela, J.; Gonzalez-Martinez, M. A.; Lopez-Paz, J. L.; Puchades, R.; Maquieira, A.; Gimenez-Romero, D. Chem. Rev. 2015, 115, 265−294. (19) Freeman, N. Proteins at Solid-Liquid Interfaces 2006, 75−104. (20) Xu, P.; Huang, F.; Liang, H. Biosens. Bioelectron. 2013, 41, 505− 510. (21) Hu, T.; Yang, F.; Yang, X. Chin. J. Anal. Chem. 2017, 45, 944− 950. (22) Mashaghi, A.; Swann, M.; Popplewell, J.; Textor, M.; Reimhult, E. Anal. Chem. 2008, 80, 3666−3676. (23) Terry, C. J.; Popplewell, J. F.; Swann, M. J.; Freeman, N. J.; Fernig, D. G. Biosens. Bioelectron. 2006, 22, 627−632. (24) Bijelic, G.; Shovsky, A.; Varga, I.; Makuska, R.; Claesson, P. M. J. Colloid Interface Sci. 2010, 348, 189−197. (25) Volcke, C.; Gandhiraman, R. P.; Gubala, V.; Doyle, C.; Fonder, G.; Thiry, P. A.; Cafolla, A. A.; James, B.; Williams, D. E. J. Colloid Interface Sci. 2010, 348, 322−328. (26) Monopoli, M. P.; Aberg, C.; Salvati, A.; Dawson, K. A. Nat. Nanotechnol. 2012, 7, 779−786. (27) Nel, A. E.; Madler, L.; Velegol, D.; Xia, T.; Hoek, E. M. V.; Somasundaran, P.; Klaessig, F.; Castranova, V.; Thompson, M. Nat. Mater. 2009, 8, 543−557. (28) Hu, T.; Wang, S.; Chen, C.; Sun, J.; Yang, X. Anal. Chem. 2017, 89, 2606−2612. (29) Lee, T. H.; Hall, K. N.; Aguilar, M. I. Curr. Top. Med. Chem. 2016, 16, 25−39. 6975

DOI: 10.1021/acs.analchem.9b01319 Anal. Chem. 2019, 91, 6971−6975