Polyelectrolyte Ensembles as Artificial Tongue: Design

Jan 10, 2017 - Boronlectin/Polyelectrolyte Ensembles as Artificial Tongue: Design, Construction, and Application for Discriminative Sensing of Complex...
1 downloads 11 Views 2MB Size
Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

Boronlectin/polyelectrolyte Ensembles as Artificial Tongue: Design, Construction and Application for Discriminative Sensing of Complex Glycoconjugates from Panax ginseng Xiao-Tai Zhang, Shu Wang, and Guo-wen Xing ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.6b13363 • Publication Date (Web): 10 Jan 2017 Downloaded from http://pubs.acs.org on January 12, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Boronlectin/polyelectrolyte Ensembles as Artificial Tongue: Design, Construction and Application for Discriminative Sensing of Complex Glycoconjugates from Panax ginseng Xiao-tai Zhang, † Shu Wang‡ and Guo-wen Xing*,† † ‡

Department of Chemistry, Beijing Normal University, Beijing, 100875, China Key Laboratory of Organic Solids, Institute of Chemistry, Chinese Academy of Sciences, Beijing, 100190, China

KEYWORDS: ginsenosides, boronic acid, polyelectrolyte, sensor array, pattern recognition ABSTRACT: Ginsenoside is a large family of triterpenoid saponins from panax ginseng, which possesses various important biological functions. Due to the very similar structures of these complex glycoconjugates, it is crucial to develop powerful analytic method to identify ginsenosides qualitatively or quantitatively. We herein report an eight-channel fluorescent sensor array as artificial tongue to achieve the discriminative sensing of ginsenosides. The fluorescent cross-responsive array was constructed by four boronlectins bearing flexible boronic acid moieties (FBAs) with multiple reactive sites and two linear poly(phenylene-ethynylene) (PPEs). An “ON-OFF-ON” response pattern was afforded on the basis of superquenching of fluorescent indicator PPEs and analyte-induced allosteric indicator displacement (AID) process. Most importantly, it was found that the canonical distribution of ginsenosides data points analyzed by linear discriminant analysis (LDA) was highly correlated with the inherent molecular structures of the analytes, and the absence of overlaps among the five point groups reflected the effectiveness of the sensor array in the discrimination process. Almost all of the unknown ginsenoside samples at different concentrations were correctly identified based on the established mathematical model. Our current work provided a general and constructive method to improve the quality assessment and control of ginseng and its extracts, which are useful and helpful for further discriminating other complex glycoconjugate families.

INTRODUCTION Ginsenosides, a class of complex triterpenoid saponin derived glycoconjugates, act as the uppermost active and officinal ingredients in traditional herbs panax ginseng, which has been widely used as a tonic in eastern Asian for thousands of years.1,2 Due to the different substitution positions of glycosyl on the steroid nucleus, ginsenosides can be classified into protopanaxadiol and protopanaxatriol derivatives.3 The physiological activities of ginsenosides vary along with the glycosylation patterns including the degree of substitution, sequence of oligosaccharide chain as well as the aforementioned position of the glycosyl group, reflecting an explicit structure-function relationship.1 Nowadays, the application of ginseng plant material and corresponding extracts is transiting from traditional healthcare to elaborative clinical medicine. However, the ginsenosides composition in the commercially available ginseng products is susceptible to the cultivation conditions and processing technics.4,5 As a conquence, it is quite necessary to establish a set of efficient methods to achieve the quality assessment and control. Although various tools such as high performance liquid chromatography (HPLC), mass spectrometry (MS), enzyme immunoassay (EIA) and a series of hyphenated techniques have been developed during the past decades to analyze and identify ginsenosides qualitatively or quantitatively,3,6-8 an inexpensive, easy-to-perform and time-saving analysis strategy is still highly required. In the current work, an efficient and sensitive fluorescent sensor array with several cross-reactive receptors was constructed, and the resulting “artificial tongue” was applied to achieve the discriminative sensing of ginsenosides with similar chemical structures (Scheme 1). The concept of sensor

array was biomimetic and similar with the organization mode of olfactory system, which has been implanted into chemical sensing toward various analytes.9-13 The collected multidimensional signals guarantee the reliability of measurement, and

Scheme 1. Chemical structures of some ginsenosides (Glc = β-Dglucose, Rha = α-L-rhamnose, Xyl =β-D-xylose).

Scheme 2. Receptor and indicator elements in the current fluorescent sensor array.

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 9

Scheme 3. Representation of the fluorescent array for ginsenosides sensing.

with the aid of appropriate mathematical pattern recognition method, the imperceptible distinctions among chemical structures of the analytes can be amplified and distinguished. In consideration of the existence of glycosyls within ginsenoside molecules, boronic acids were chosen to be the receptors for their inherent property to covalently react with 1,2- or 1,3-diols reversibly and quickly.14,15 Herein, we designed and synthesized four novel boronlectins bearing flexible boronic acid moieties (FBAs) with different initial nucleus and high spatial freedom degree (Scheme 2). Among them, FBA1 has been used to capture simple saccharides in our previous work,16 FBA2, FBA3 and FBA4 adopt different arrangement of reactive sites. Introduction of the pyridinium salt moieties ensured the hydrophilicity of FBAs, and the existence of multiple boronic acid sites should facilitate the possible multivalent interaction between supramolecular host and guest. Integrated the boronlectins with two different linear conjugated polyelectrolytes poly(phenylene-ethynylene)s (PPEs) as the fluorescent indicators (IA and IB, Scheme 2) respectively, an eight-channel (4 FBAs×2 PPEs) fluorescent sensor array was obtained. The pairing of cationic FBAs containing pyridinium salt moieties and anionic PPEs indicators was inspired by the reported superquenching effect resulted from effective electron transfer (ET).17-19 Upon addition of analytes bearing diol moieties, the allosteric indicator displacement (AID) process occurred and accounted for the fluorescence recovery.20 Compared with the widely-investigated single-molecular boronic acid sensors for saccharides and their derivatives, the twocomponent systems circumvent elaborate design and complicated synthesis, and facilitate the modular construction of sensors.21-23 As shown in Scheme 3, eight 96-well plates which were loaded by different sensing ensembles were set for one ginsenoside. With the aid of microtiter plate assay, the “OnOff-On” fluorescence signals were collected for the further pattern recognition analysis. Up to now, there are quite few boronic acid-functionalized sensors reported for ginsenosides sensing, which employed single-molecular fluorescent sensor with a phenothiazine24/porphyrin moiety25, or UV-Vis absorption-based optical array.26 Our current work is the first time to develop an efficient fluorescent array for high throughput sensing of complex ginsenosides.

RESULTS AND DISCUSSION To begin with, the interactions between anionic PPEs and synthesized cationic FBAs were investigated systematically. As expected, the flexible pyridinium salt-bearing boronic acids (FBAs) dramatically quenched the fluorescence of the PPE indicator molecules (Figure S1, S2), since the existence of

superquenching, in which low concentration of an electrondeficient quencher induced efficient fluorescence quenching of the conjugated polyelectrolytes due to the fast migration of singlet exciton along the polymer chain and effective ET. In some cases, the Stern-Volmer quenching plot of F/F0 exhibited “S”-shape with the increase of FBAs concentrations (Figure S3), and the whole titration curves can be divided into linear, superlinear and gradually-saturated quenching ranges (Table 1). Quenching constants (Ksv) were calculated within the initial linear ranges, which were corresponded to the predominant static quenching. Compared with our previous study in which the interaction between the organic fluorescent indicator HPTS (8-hydroxypyrene-1,3,6-trisulfonic acid trisodium) and FBA1 was investigated,16 the Ksv of PPEs/FBA1 system increased by 30-100 times. Therefore, addition of few amount of FBAs brought about satisfying fluorescence quenching, and further improved the sensitivity of the sensing ensembles. The upward superlinear ranges were observed if increasing the amount of FBAs (Figure S3), since the polyvalent quenchers led to the formation of a compact aggregates, which was responsible for the interchain delocalization of the exciton, and consequently increased the quenching efficiency.27-30 To quench the fluorescence of PPEs to great extent, the required amount of boronic acids followed the order: FBA4 < FBA3 < FBA2 < FBA1 (Figure S1, S2), demonstrating that both the number and arrangement of positive charges of FBAs resulted in the variation of quenching capacity. Some other evidences enabled a deeper insight into the formation of PPE/FBA ensembles. Obvious decrease and bathochromic shift of UV-Vis absorption were observed upon addition of FBA4 into IA solution, and a typical isosbestic point arose at 450 nm (Figure S4), indicating the elongation of effective conjugated structure resulted from the on-going aggregation process.18,28,31,32 Actually, PPE indicators also aggregated to some extent in the absence of quenchers for their hydrophobic backbones, introduction of FBAs at the concentration of 5 µM generated the swelling or shrinking aggregates, which depended on linear quenching and superlinear quenching process, respectively (Figure S5, S6). Additionally, in contrast with the amorphous dried samples of IA or FBA4, the TEM images of IA/FBA4 ensembles with obvious meshy morphology verified the existence of ionic bridging induced by polyvalent quencher (Figure S7). With the PPE/FBA two-component sensing ensembles in hand, we investigated the AID process using IA/FBA4 as the model system. When titrated with ginsenoside Rb1, the generated anionic boronate esters neutralized the positive

ACS Paragon Plus Environment

Page 3 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Table 1. Summary of the quenching plots induced by FBAs in phosphate buffer (10 mM, pH 7.4; [IA] = 10 µM, λex = 421 nm, λem = 466 nm; [IB] = 10 µM, λex = 408 nm, λem = 526 nm). Indicator

Quencher

Ksv (M-1)a

Linear quenching range (µM)

Superlinear quenching range (µM)

FBA1 0 ~ 10 (9.77±0.07)×104 —b 5 FBA2 0~6 >6 (2.02±0.04)×10 IA FBA3 0~2 >2 (3.53±0.07)×105 FBA4 0 ~ 0.5 > 0.5 (1.25±0.05)×105 FBA1 0~4 4 ~ 10 (3.07±0.05)×105 5 0~2 >2 FBA2 (4.36±0.12)×10 IB FBA3 0 ~ 0.8 0.8 ~ 4 (6.49±0.12)×105 FBA4 0 ~ 0.8 0.8 ~ 3 (5.67±0.10)×105 a Ksv values were calculated via fitting F0/F as a function of FBAs concentrations within the linear ranges. b The ranges were not observed in the whole quenching plot.

charges of pyridinium salt, and the electrostatic interaction between IA and FBA4 was weakened. As a result, the fluorescence of solution recovered (Figure 1), and accompanied with an increasing UV-Vis absorption (Figure S8). Both of the spectra variations reflected a converse process of the quenching and the release of indicator from IA/FBA4 complex, which confirmed the covalent interactions between the boronic acid sites and the diol moieties of Rb1. In addition, it is noted that there were no obvious fluorescence responses for sucrose, which is concerned as the most important interference in the real ginsenoside-contained samples (Figure S9). Next, we performed the titration experiments toward five kinds of ginsenosides which belong to protopanaxadiols (Rb1, Rb3, Rd) and protopanaxatriols (Rg1, Re) under all eight sensing conditions (Figure S10, S11). Compared with IA, IB was easier to be quenched (Fig. S1, S2), therefore, a lower concentration of FBAs was used in the sensing systems indicated by IB to obtain the comparative quenching degrees. The detection limits of ginsenosides were calculated to be µM level, indicated the sensitivity of the sensing system (Table S1). In order to estimate the binding strength, a series of apparent binding constants were calculated via non-linear fitting of the titration data (Figure 2, Table S2).33,34 It was found that the thermodynamic binding capacities of a given PPE/FBA ensemble toward different ginsenosides exhibited obvious cross-reactivity. Generally, the binding affinities of protopanaxatriols were remarkably lower than those of protopanaxadiols when employing the same sensing ensemble, which can be ascribed to the geometrical complementation between FBAs and ginsenosides differed in the substitution positions of the glycosyl groups on the steroid nucleus. The results are completely different with the case involved in a porphyrin-based boronic acid receptor, which preferred to bind with protopanaxatriols rather than protopanaxadiols.25 Among the three protopanaxadiols, although Rd possesses fewer sugar rings than Rb1 or Rb3, its Kb values were calculated to be higher than the two counterparts. We speculated that the smaller molecular volume of Rd facilitated its insertion into the PPE/FBA aggregates, and then favored the AID process. Also, it is reasonable to deduce that the advantage of the terminal saccharide residue in covalent binding was thoroughly offset by the negative effect brought by its steric hindrance. The modest discrepancies between Rb1 and Rb3 in

Gradually-saturated quenching range (µM) > 10 —b —b —b > 10 —b >4 >3

Figure 1. Fluorescence recovery assay of IA/FBA4 ensemble upon addition of ginsenoside Rb1 (0–2 mM) in 10 mM pH 7.4 phosphate buffer containing 2% (v/v) DMSO. [IA] = 10 µM, [FBA4] = 5 µM, λex = 421 nm. Top dashed line: IA only, bottom dashed line: IA/FBA4, other solid lines: IA/FBA4/Rb1. Inset: photograph of the samples contained IA, IA/FBA4 or IA/FBA4/Rb1 taken under the illumination of a UV lamp (λex = 365 nm).

Figure 2. Apparent binding constants (Kb, M-1) of the five ginsenoside analytes using different PPE/FBA sensing ensembles.

Kb were discovered, possibly in virtue of the distinguishing binding preference of boronic acid sites toward the glucosyl and xylosyl moieties within ginsenoside molecules. It is noted that the IB-indicated system exhibited an extraordinarily great

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 9

substitution patterns of glycosyl group on the steroid nucleus, while the factor 2 acted as a “yardstick” for the number of sugar rings. Among the eight types of ensembles, three of them with relatively higher discriminative contributions were screened out through two continuous principle component analysis (PCA) processes: IB/FBA1, IB/FBA3 and IB/FBA4 (Figure S12).35-37 It was found that all of the chosen ensembles employed IB as the fluorescent indicator, and the corresponding low concentrations of the boronic acid receptors presumably induced the better discriminative capacity. Classification accuracy of the simplified sensor array with only three sensor elements was calculated to be up to 88.3% after performing a simple LDA process, and the majority of the misclassified samples belonged to ginsenosides Rb1 and Re (Figure S13). Improved discriminations were achieved via advanced processes such as quadratic discriminant analysis (QDA)37 and multi-layered LDA38,39, in which the misclassification of Re was eliminated while the classification accuracies of the other species almost remained (Table S4, S5).

Figure 3. a) 2D canonical score plot of the five ginsenosides at a concentration range (0.1–2 mM) analyzed by LDA (7 replicates for each concentration level). b) Classification accuracy of 99.1% was obtained via cross-validation.

Kb compared with the case indicated by IA when employing the same FBA to bind a certain protopanaxadiol, since the addition of fewer amount of FBAs enabled the early saturation at low concentrations of the analytes (Figure S11). Moreover, for the cases indicated by the same fluorescent PPE, the quenching capacities of different FBAs and their inherent binding affinities to ginsenosides had a synergistic effect on the significant differences of the Kb values. Therefore, a set of calculated Kb for a given ginsenoside were not comparable due to the ineffectiveness in evaluating the actual binding affinities between various boronic acids and the analyte. However, from a practical perspective, a “fingerprint” response of the analyte was still afforded. To achieve the pattern recognition of different ginsenosides, linear discriminant analysis (LDA) was used to process the relative fluorescence recovery (F/F0) data, which were obtained from the eight-channel sensor array. As shown in Figure 3a, 350 data points, which belong to five analytes at 10 concentration levels, were well-clustered and divided into five groups in the 2D canonical score plot, indicating an excellent capacity of the sensor array in discriminative sensing of ginsenosides. The first two canonical factors carried about 85.4% and 7.9% of the total variance, respectively (Table S3). The validity of the sensor array was further confirmed via “leave-one-out” cross-validation (LOOCV), only three samples of Rb1 and Rd were misclassified and gave rise to a 99.1% total classification accuracy (Figure 3b). It was interesting to find that the structural features of ginsenoside molecules were clearly reflected in the 2D canonical score plot: the factor 1 discriminated ginsenosides based on the

Figure 4. Distribution of the data points of unknown samples in the 2D canonical score plot (3 replicates for each case).

Afterwards, additional 15 ginsenoside samples (5 species× 3 concentration levels) were prepared to further verify the practicability of the eight-channel sensor array in predicting unknown ginsenosides. Fortunately, 13 out of 15 samples were correctly determined by the mathematical model established via LDA, with 86.7% accuracy (Table S6–S8, i.e. matrix A– C). Two Rb1 samples were misclassified to Rb3 and Rd, respectively, which was in correlation with the mutuallyadjacent occupied regions of the three protopanaxadiols in the 2D canonical score plot (Figure 3a). Additionally, with the aid of the non-standardized coefficient matrix of the discriminant functions (Table S9, i.e. matrix D), the canonical geometric coordinates of the unknown samples were obtained (Table S10, i.e. matrix E), and the visible discrimination was achieved consequently. As shown in Figure 4, the data points of Rd, Rg1 and Re were all correctly assigned to the respective two-dimensional regions which were divided by the linear boundaries. For the Rb1 and Rb3 samples, it was achievable to separate the two point groups from each other when introducing the third discriminant function (Z-axis), which undertook 5.4% of the total variance in LDA (Table S3).

ACS Paragon Plus Environment

Page 5 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

CONCLUSION In summary, four cationic flexible boronic acids (FBAs) with multiple binding sites were designed and synthesized, and integrated with two anionic PPE polymers to form an eightchannel sensor array for the pattern recognition of ginsenosides. The fluorescence of the PPE indicators can be efficiently quenched by FBAs via superquenching process resulted from the delocalization of singlet exciton, and both the number and arrangement of positive charges within FBAs were proved to affect the quenching capacity significantly. Addition of ginsenosides triggered the AID process, accompanied with obvious fluorescence recovery. The FBAsinduced quenching and the binding affinities of FBAs toward analytes were considered to account for the cross-reactivity of the sensing array. Information of the ginsenoside structures was explicitly reflected in the canonical score plot analyzed by LDA. Moreover, the established mathematical model was successfully applied to the identification of the unknown ginsenoside samples at different concentrations. The current work provided a unique and constructive method to improve the quality assessment and control of ginseng and its extracts, and should be useful and helpful for further discriminating other complex glycoconjugate families. EXPERIMENTAL SECTION General. All chemicals were purchased as reagent grade and used directly without further purification. The reactions were monitored by analytical thin-layer chromatography (TLC) on silica gel F254 glass plates and visualized under UV light (254 and 365 nm). Flash column chromatography was performed on silica gel (200-300 mesh). Reversed-phase column chromatography was performed using SiliaSphere C18 as packing. 1H NMR spectra were recorded with a Bruker Avance III 400 MHz NMR spectrometer at 20°C, and the chemical shift (in ppm) were referenced to the peaks of deuteriated solvent. Coupling constants in Hz were calculated from the one-dimensional spectra. 13C NMRspectra were recorded with the 400 MHz NMR spectrometer (100 MHz). High-resolution mass spectra were recorded with Waters LCT Premier XEmass spectrometer. Low-resolution mass spectra were determined by Agilent 7890A GC-5975C MS. Buffer solutions and the stock solutions of FBAs as well as PPEs were prepared with distilled H2O purified by a quartz sub-boiling distiller, while the stock solutions of ginsenoside analytes were prepared with DMSO because of their hydrophobicity resulted from the steroid nucleus. Fluorescence emission spectra and UV-Vis spectra were recorded on an Edinburgh Instruments FS5 Fluorescence Spectrometer and a Shimadzu UV-2600 spectrophotometer, respectively, in a quartz cell with 1 cm path length. The absolute quantum yields were measured on a HAMAMATSU Quantaurus-QY. Fluorescence responses upon addition of certain concentration of analytes that were used for LDA/QDA/PCA (performed by SPSS v20.0) were obtained on a microplate reader (Thermo Scientific, Varioskan Flash) using 96-well plates (Thermo Scientific, lighttight, flat bottom, non-sterile). Particle size data were obtained through dynamic laser scattering (ZetaPLUS, Brookhaven Instruments Corporation). Samples for transmission electron microscopy (TEM) were prepared by placing a drop of solution on a copper grid and then dried in

air, the TEM images were obtained on a Hitachi HT7700 microscope. Synthesis. N,N-bis[(2-boronic acid)benzyl pyridinium-2ylmethyl]amine dibromide (FBA1). Synthesis of FBA1 has been reported in the previous literature.7 N,N,N-tris[(2-boronic acid)benzyl pyridinium-2ylmethyl]amine tribromide (FBA2). Tris(2pyridylmethyl)amine (113 mg, 0.39 mmol) and 2(bromomethyl)phenylboronic acid (376 mg, 1.75 mmol, 4.5 equiv.) were dissolved in 6 mL THF, then the mixture was refluxed at 65°C for 48 h and a faint yellow solid was obtained. After cooling to room temperature, the supernatant was removed, and the resulting solid was washed with THF, followed by purification through reversed-phase column chromatography (SiliaSphere C18) eluting with a MeOH/H2O (1:9) system. MeOH and H2O in the eluate were respectively removed via rotary evaporation and lyophilization to afford FBA2 as a white flocculent solid (109 mg, 30.1%). 1H NMR (400 MHz, D2O): δ = 4.03 (s, 6H), 5.75 (s, 6H), 6.49 (d, J = 7.4 Hz, 3H), 7.16-7.27 (m, 6H), 7.56 (dd, J = 7.2, 1.3 Hz, 3H), 7.84 (t, J = 6.9 Hz, 3H), 7.97 (d, J = 7.9 Hz, 3H), 8.33 (t, J = 7.7 Hz, 3H), 8.64 (d, J = 5.9 Hz, 3H) ppm; 13C NMR (100 MHz, D2O): δ = 54.5, 61.3, 126.6, 127.3, 127.8, 128.8, 131.2, 135.2, 135.6, 146.9, 147.6, 152.6 ppm. ESI-MS for C39H42B3Br3N4O6: m/z 408.1 [M-2Br+3CH3OH-3H2O]2+. N,N,N’-tris[(2-boronic acid)benzyl pyridinium-2-ylmethyl]ethane-1,2-diamine tribromide (FBA3). The synthetic route of FBA3 was shown in Scheme S1. Compound 1 was synthesized following the reported protocols.40,41 Compound 1 (1.15 g, 3.46 mmol) was dissolved in 7 mL EtOH, then mixed with Boc-anhydride (0.91 g, 4.16 mmol, 1.2 equiv.) dissolved in 3 mL EtOH at 0°C. The starting material had been completely consumed as detected by TLC (DCM/MeOH =10/1) after stirring for 1 h at room temperature. EtOH was evaporated under vacuum and the residue was diluted with DCM (150 mL), washed with distilled water twice, and dried with K2CO3. After removal of the solvent, the residue was purified by column chromatography (DCM/MeOH = 50/1) to give compound 2 as white solid (1.44 g, 96.0%). 1H NMR (400 MHz, CDCl3): δ = 1.35 (s, 9H), 2.52-3.01 (m, 2H), 3.43 (t, J = 6.2 Hz, 1H), 3.58 (bs, 1H), 3.83 (bs, 4H), 4.43 (s, 1H), 4.56 (s, 1H), 7.08-7.22 (m, 4H), 7.48-7.70 (m, 5H), 8.42-8.55 (m, 3H) ppm; 13C NMR (100 MHz, CDCl3): δ = 28.3, 45.2, 52.3, 52.7, 52.8, 53.4, 60.3, 60.5, 79.7, 79.9, 120.4, 121.3, 122.0, 122.6, 123.0, 136.3, 136.5, 136.6, 148.9, 149.0, 149.1, 155.5, 155.7, 158.3, 158.8, 159.4, 159.5 ppm. HRMS (ESI): m/z calcd for C25H32N5O2 ([M+H]+): 434.2551; found: 434.2553. Compound 2 (167 mg, 0.39 mmol) and 2(bromomethyl)phenylboronic acid (373 mg, 1.74 mmol, 4.5 equiv.) were dissolved in 6 mL THF, then the mixture was refluxed at 65°C for 48 h and a faint yellow solid was obtained. After cooling to room temperature, the supernatant was removed, and the resulting solid was washed with THF, followed by purification through reversed-phase column chromatography (SiliaSphere C18) eluting with a MeOH/H2O (1:9) system. MeOH and H2O in the eluate were respectively removed via rotary evaporation and lyophilization to afford compound 3 as a white flocculent solid (187 mg, 45.0%). 1H NMR (400 MHz, D2O): δ = 1.05 (s, 5H), 1.13 (s, 4H), 2.46 (bs, 1H), 2.54 (bs, 1H), 3.26 (bs, 2H), 3.85 (s, 2H), 3.92 (s, 2H),

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

4.51 (d, J = 9.7 Hz, 2H), 5.73-5.88 (m, 6H), 6.51 (d, J = 6.9 Hz, 1H), 6.56-6.70 (m, 2H), 7.18-7.34 (m, 6H), 7.52-7.94 (m, 9H), 8.24 (t. J = 7.6 Hz, 1H), 8.34 (t, J = 7.3 Hz, 1H), 8.408.72 (m, 4H) ppm; 13C NMR (100 MHz, D2O): δ = 27.4, 27.5, 44.5, 45.2, 48.1, 48.4, 52.3, 52.6, 54.6, 61.1, 83.6, 126.0, 126.5, 126.7, 126.8, 127.0, 127.2, 127.5, 127.6, 128.0, 128.1, 128.5, 128.78, 128.83, 131.3, 131.5, 135.2, 135.8, 136.1, 146.4, 146.8, 146.9, 147.1, 147.3, 153.8, 154.2, 155.8, 156.4 ppm. ESI-MS for C46H55B3Br3N5O8: m/z 1024.4 [MBr+2CH3OH-2H2O]+, 1038.4 [M-Br+3CH3OH-3H2O]+. Compound 3 (77 mg, 0.072 mmol) was dissolved in 3 mLtBuOH-H2O (v/v = 1:1) mixed solvent was refluxed at 100 °C for 12 h, the starting material had been completely consumed as detected by TLC (nBuOH/H2O=4/1, stained with ninhydrin). After cooling to room temperature, the solvent was removed directly via lyophilization to afford FBA3 as a white flocculent solid (67 mg, 95.8%). 1H NMR (400 MHz, D2O): δ = 2.69 (t, J = 6.7 Hz, 2H), 2.84 (t, J = 6.4 Hz, 2H), 3.96 (s, 4H), 4.14 (s, 2H), 5.82 (s, 6H), 6.53 (d, J = 7.6 Hz, 2H), 6.74-6.83 (m, 1H), 7.16-7.30 (m, 6H), 7.55 (d, J = 6.8 Hz, 3H), 7.73-7.82 (m, 3H), 7.90 (d, J = 7.9 Hz, 3H), 8.29 (t, J = 7.8 Hz, 2H), 8.36 (t, J = 7.9 Hz, 1H), 8.46-8.55 (m, 3H) ppm; 13C NMR (100 MHz, D2O): δ = 45.7, 48.5, 52.5, 60.9, 127.0, 127.1, 127.2, 128.0, 128.2, 128.8, 129.0, 131.20, 131.23, 135.1, 135.2, 135.7, 136.0, 146.46, 146.51, 146.6, 147.0, 152.7, 153.8 ppm. ESIMS for C41H47B3Br3N5O6: m/z 938.4 [M-Br+3CH3OH-3H2O]+, 429.7 [M-2Br+3CH3OH-3H2O]2+. N,N,N’,N’-tetrakis[(2-boronic acid)benzyl pyridinium-2ylmethyl]-ethane-1,2-diamine tetrabromide (FBA4). N,N,N',N'-tetrakis(2-pyridylmethyl)ethylenediamine (85 mg, 0.2 mmol) and 2-(bromomethyl)phenylboronic acid (258 mg, 1.2 mmol, 6.0 equiv.) were dissolved in 3 mL THF, then the mixture was refluxed at 65°C for 48 h and a yellow solid was obtained. After cooling to room temperature, the supernatant was removed, and the resulting solid was washed with THF, followed by purification through reversed-phase column chromatography (SiliaSphere C18) eluting with a MeOH/H2O (1:9) system. MeOH and H2O in the eluate were respectively removed via rotary evaporation and lyophilization to afford FBA4 as a white flocculent solid (97 mg, 37.9%). 1H NMR (400 MHz, D2O): δ = 2.52 (s, 4H), 3.75 (s, 8H), 5.79 (s, 8H), 6.38 (d, J = 7.7 Hz, 4H), 7.12 (t, J = 7.6 Hz, 4H), 7.21 (t, J = 7.4 Hz, 4H), 7.51 (d, J = 7.3 Hz, 4H), 7.67 (d, J = 7.9 Hz, 4H), 7.79 (t, J = 6.8 Hz, 4H), 8.23 (t, J = 7.7 Hz, 4H), 8.55 (d, J = 6.0 Hz, 4H) ppm; 13C NMR (100 MHz, D2O): δ = 51.7, 54.8, 61.0, 126.4, 127.1, 128.4, 128.6, 131.1, 135.1, 136.1, 146.6, 147.4, 153.4 ppm. ESI-MS for C54H60B4Br4N6O8: m/z 589.3 [M-2Br+4CH3OH-4H2O]2+, 596.3 [M-2Br+5CH3OH-5H2O]2+. Poly(phenylene-ethynylene)s (PPEs). The fluorescent indicators IA and IB were synthesized following the reported protocols,17,28,42,43 and the synthetic route was shown in Scheme S2. The absolute fluorescence quantum yields of IA and IB were 0.066 and 0.086 (in pure water), respectively, which were consistent with the reported values28,42. Preparation of samples. Titration experiments. Stock solution of IA or IB (5 mM in H2O) was added into 3 mL phosphate buffer (10 mM, pH 7.4), the final monomer concentration of fluorescent indicator was 10 µM. Upon addition of FBA stock solution (1 mM in H2O), a two-component sensing ensemble was obtained with the final concentration of FBA at

Page 6 of 9

5 µM (interacted with IA) or 3 µM (interacted with IB). A series of stock solutions of a ginsenoside at different concentrations (5, 10, 15, 20, 25, 30, 40, 50, 75, 100 mM in DMSO) were prepared beforehand, and 60 µL stock solution of the analyte was added to form an aqueous mixture with 2% DMSO. As results, the final concentrations of ginsenoside were 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.8, 1.0, 1.5, 2.0 mM, respectively. For the titration of sucrose, 10 mg solid was added directly into the aqueous solution of sensing ensembles. The data of fluorescence emission, UV-vis absorption and effective diameters of particles were collected after aging for 2 h. Buffer solution used in DLS analysis was preprocessed by 220 nm filtration membrane. Collection of fluorescence signals for LDA/PCA/QDA. The aqueous solutions of the sensing ensemble with fluorescent indicator PPE (10 µM) and boronic acid FBA (5 µM or 3 µM) were prepared using 10 mM phosphate buffer as the medium, and then mixed with 4 µL ginsenoside stock solutions (in DMSO) at different concentrations. The total volume in each well of 96-well plate was 200 µL, and the final concentrations of the ginsenoside were 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.8, 1.0, 1.5, 2.0 mM, respectively (positions of the experimental groups: 3A~12G). Seven replicates were set for each case. The control groups were set to be PPE (10 µM) only (positions: 1A~1G) and PPE (10 µM)/FBA (5 µM or 3 µM) ensemble (positions: 2A~2G). The blank group contained buffer solution only (positions: H1~H12). All of the tested samples contained 2% DMSO. Data were collected on a fluorescence plate reader after shaking for 3 min and aging for 2 h. For the titration process of a certain ginsenoside, all of the 70 analyte samples (10 concentrations×7 replicates) were set on a 96-well plate loaded with the same sensing ensemble. Classification of unknown samples. The aqueous solutions of the sensing ensemble with fluorescent indicator PPE (10 µM) and boronic acid FBA (5 µM or 3 µM) were prepared using 10 mM phosphate buffer as the medium, and then mixed with 4 µL unknown ginsenoside solutions (in DMSO) at different concentrations. The total volume in each well of 96-well plate was 200 µL, and the final concentrations of the ginsenosides were 0.5, 1.0, 1.5 mM, respectively. Three replicates were set for each case. The control groups were set to be PPE (10 µM) only and PPE (10 µM)/FBA (5 µM or 3 µM). The blank group contained buffer solution only. All of the tested samples contained 2% DMSO. Data were collected on a fluorescence plate reader after shaking for 3 min and aging for 2 h. All of the 45 analyte samples (5 analytes×3 concentrations×3 replicates) were set on a 96-well plate with the same sensing ensemble. Data analysis. Quenching process. The Stern-Volmer quenching constants were calculated by fitting the data with the linear function as below:30 F0/F = 1+Ksv[Q] where F0: initial fluorescence intensity of PPE; F: fluorescence intensity upon addition of quencher; [Q]: concentration of the quencher; Ksv: Stern-Volmer quenching constant. Allosteric indicator displacement (AID) process. The apparent binding constants were calculated by fitting the data with the function as below.33,34 The stoichiometric ratios were uniformly presumed to be 1:1 to facilitate the calculation and comparison.

ACS Paragon Plus Environment

Page 7 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

F/F0 = (1+(Fmax/F0)Kb[A])/(1+Kb[A]) F0: fluorescence intensity of PPE/FAB ensembles; F: fluorescence intensity upon addition of the ginsenoside; Fmax: the calculated maximum fluorescence intensity, where the platform of the titration curve would exhibit; Kb: apparent binding constant; [A]: concentration of the analyte.

ASSOCIATED CONTENT Supporting Information. Detailed synthetic routes of FBAs and PPEs; illustration of the quenching events of PPEs upon addition of FBAs using fluorescence spectra, UV-Vis absorption spectra, dynamic light scattering and transmission electron microscope; titration plots of ginsenosides using different sensing ensembles; illustration of PCA and LDA; tables showing the detection limits, apparent binding constants, the results analyzed by LDA, QDA, multi-layered LDA; the identification results of unknown samples; 1H and 13C NMR spectra of FBAs. This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding Author * [email protected]

Notes The authors declare no competing financial interest.

ACKNOWLEDGMENT We gratefully acknowledge the financial support from the National Natural Science Foundation of China (21272027).

REFERENCES (1) Zhou, L. G.; Wu, J. Y. Development and Application of Medical Plant Tissue Cultures for Production of Drugs and Herbal Medicinals in China. Nat. Prod. Rep. 2006, 23, 789–810. (2) Radad, K.; Gille, G.; Liu, L. L.; Rausch, W. D. Use of Ginseng in Medicine with Emphasis on Neurodegenerative Disorders. J. Pharmacol. Sci. 2006, 100, 175–186. (3) Yap, K. Y.-L.; Chan, S. Y.; Chan, Y. W.; Lim, C. S. Overview on the Analytical Tools for Quality Control of Natural Product-Based Supplements: a Case Study of Ginseng. Assay Drug Dev. Technol. 2005, 3, 683–699. (4) Sugimoto, S.; Nakamura, S.; Matsuda, H.; Kitagawa, N.; Yoshikawa, M. Chemical Constituents from Seeds of Panax Ginseng: Structure of New Dammarane-type Triterpene Ketone, Panaxadione, and HPLC Comparisons of Seeds and Flesh. Chem. Pharm. Bull. 2009, 57, 283–287. (5) Lee, S. M.; Shon, H. J.; Choi, C. S.; Hung, T. M.; Min, B. S.; Bae, K. Ginsenosides from Heat Processed Ginseng. Chem. Pharm. Bull. 2009, 57, 92–94. (6) Fuzzati, N. Analysis Methods of Ginsenosides. J. Chromatogr. B 2004, 812, 119–133. (7) Jiang, Y.; David, B.; Tu, P.; Barbin, Y. Recent Analytical Approaches in Quality Control of Traditional Chinese Medicines–A Review. Anal. Chim. Acta 2010, 657, 9–18. (8) Wong, M. Y.-M.; So, P.-K.; Yao, Z.-P. Direct Analysis of Traditional Chinese Medicines by Mass Spectrometry. J. Chromatogr. B 2016, 1026, 2–14. (9) Malnic, B.; Hirono, J.; Sato, T.; Buck, L. B. Combinatorial Receptor Codes for Odors. Cell 1999, 96, 713–723. (10) Askim, J. R.; Mahmoudi, M.; Suslick, K. S. Optical Sensor Arrays for Chemical Sensing: the Optoelectronic Nose. Chem. Soc. Rev. 2013, 42, 8649–8682.

(11) Bunz, U. H. F.; Rotello, V. M. Gold Nanoparticle-Fluorophore Complex: Sensitive and Discerning “Noses” for Biosystems Sensing. Angew. Chem. Int. Ed. 2010, 49, 3268-3279. (12) Elci, S. G.; Moyano, D. F.; Rana, S.; Tonga, G. Y.; Phillips, R. L.; Bunz, U. H. F.; Rotello, V. M. Recognition of Glycosaminoglycan Chemical Patterns Using an Unbiased Sensor Array. Chem. Sci. 2013, 4, 2076-2080. (13) Han, J.; Bender, M.; Seehafer, K.; Bunz, U. H. F. Identification of White Wines by Using Two Oppositely Charged Poly(pphenyleneethynylene)s Individually and in Complex. Angew. Chem. Int. Ed. 2016, 55, 7689-7692. (14) Boronic Acids: Preparation, Applications in Organic Synthesis and Medicine; Hall, D. G. Ed.; John Wiley & Sons, 2006. (15) Ni, N.; Laughlin, S.; Wang, Y.; Feng, Y.; Zheng, Y.; Wang, B. Probing the General Time Scale Question of Boronic Acid Binding with Sugars in Aqueous Solution at Physiological pH. Bioorg. Med. Chem. 2012, 20, 2957–2961. (16) Zhang, X.-T.; Wang, S.; Xing, G.-W. Novel Boronlectins Based on Bispyridium Salt with a Flexible Linker: Discriminative Sensing of Lactose and Other Monosaccharides and Disaccharides in Aqueous Solution. Chem. Asian. J. 2015, 10, 2594–2598. (17) Zhou, Q.; Swager, T. M. Fluorescent Chemosensors Based on Energy Migration in Conjugated Polymers: the Molecular Wire Approach to Increased Sensitivity. J. Am. Chem. Soc. 1995, 117, 12593– 12602. (18) Chen, L.; McBranch, D. W.; Wang, H.-L.; Helgeson, R.; Wudl, F.; Whitten, D. G. Highly Sensitive Biological and Chemical Sensors Based on Reversible Fluorescence Quenching in a Conjugated Polymer. Proc. Natl. Acad. Sci. USA 1999, 96, 12287–12292. (19) DiCesare, N.; Pinto, M. R.; Schanze, K. S.; Lakowicz, J. R. Saccharide Detection Based on the Amplified Fluorescence Quenching of a Water-Soluble Poly(phenylene ethynylene) by a Boronic Acid Functionalized Benzyl Viologen Derivative. Langmuir 2002, 18, 7785–7787. (20) Going Beyond Continuous Glucose Monitoring with Boronic Acid Appended Bipyridinium Salts. Schiller, A.; Vilozny, B.; Wessling, R. A.; Singaram, B. In Reviews in Fluorescence 2009, (Ed.: Geddes, C.), Springer, New York, 2011, pp. 155–191. (21) Cordes, D. B.; Singaram, B. A Unique, Two-Component Sensing System for Fluorescence Detection of Glucose and Other Carbohydrates. Pure Appl. Chem. 2012, 84, 2183–2202. (22) Wu, X.; Li, Z.; Chen, X.-X.; Fossey, J. S.; James, T. D.; Jiang, Y.-B. Selective Sensing of Saccharides Using Simple Boronic Acids and Their Aggregates. Chem. Soc. Rev. 2013, 42, 8032–8048. (23) Sun, X.; James, T. D. Glucose Sensing in Supramolecular Chemistry. Chem. Rev. 2015, 115, 8001–8037. (24) Wu, Y.; Guo, H.; Zhang, X.; James, T. D.; Zhao, J. Chiral Donor Photoinduced-Electron-Transfer (d-PET) Boronic Acid Chemosensors for the Selective Recognition of Tartaric Acids, Disaccharides, and Ginsenosides. Chem. Eur. J. 2011, 17, 7632–7644. (25) Hargrove, A. E.; Reyes, R. N.; Riddington, I.; Anslyn, E. V.; Sessler, J. L. Boronic Acid Porphyrin Receptor for Ginsenoside Sensing. Org. Lett. 2010, 12, 4804–4807. (26) Zhang, X.; You, L.; Anslyn, E. V.; Qian, X. Discrimination and Classification of Ginsenosides and Ginsengs Using Bisboronic Acid Receptors in Dynamic Multicomponent Indicator Displacement Sensor Array. Chem. Eur. J. 2012, 18, 1102–1110. (27) Tan, C.; Atas, E.; Müller, J. G.; Pinto, M. R.; Kleiman, V. D.; Schanze, K. S. Amplified Quenching of a Conjugated Polyelectrolyte by Cyanine Dyes. J. Am. Chem. Soc. 2004, 126, 13685–13694. (28) Jiang, H.; Zhao, X.; Schanze, K. S. Amplified Fluorescence Quenching of a Conjugated Polyelectrolyte Mediated by Ca2+. Langmuir 2006, 22, 5541–5543. (29) Jiang, H.; Zhao, X.; Schanze, K. S. Effects of Polymer Aggregation and Quencher Size on Amplified Fluorescence Quenching of Conjugated Polyelectrolytes. Langmuir 2007, 23, 9481–9486.

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(30) Jiang, H.; Taranekar, P.; Reynolds, J. R.; Schanze, K. S. Conjugated Polyelectrolytes: Synthesis, Photophysics, and Applications. Angew. Chem. Int. Ed. 2009, 48, 4300–4316. (31) Wang, D.; Wang, J.; Moses, D.; Bazan, G. C.; Heeger, A. J. Photoluminescene Quenching of Conjugated Macromolecules by Bipyridinium Derivatives in Aqueous Media: Charge dependence. Langmuir 2001, 17, 1262–1266. (32) Tan, C.; Pinto, M. R.; Schanze, K. S. Photophysics, Aggregation and Amplified Quenching of a Water-Soluble Poly(phenylene ethynylene). Chem. Commun. 2002, 446–447. (33) Boens, N.; Leen, V.; Dehaen, W. Fluorescent Indicators Based on BODIPY. Chem. Soc. Rev. 2012, 41, 1130–1172. (34) Gamsey, S.; Miller, A.; Olmstead, M. M.; Beavers, C. M.; Hirayama, L. C.; Pradhan, S.; Wessling, R. A.; Singaram, B. Boronic Acid-based Bipyridinium Salts as Tunable Receptors for Monosaccharides and α-Hydroxycarboxylates. J. Am. Chem. Soc. 2007, 129, 1278–1286. (35) Jurs, P. C.; Bakken, G. A.; McClelland, H. E. Computational Methods for the Analysis of Chemical Sensor Array Data from Volatile Analytes. Chem. Rev. 2000, 100, 2649–2678. (36) Palacios, M. A.; Wang, Z.; Montes, V. A.; Zyryanov, G. V.; Anzenbacher, P. Jr. Rational Design of a Minimal Size Sensor Array for Metal Ion Detection. J. Am. Chem. Soc. 2008, 130, 10307–10314. (37) Chen, W.; Li, Q.; Zheng, W.; Hu, F.; Zhang, G.; Wang. Z.; Zhang, D.; Jiang, X. Identification of Bacteria in Water by a Fluorescent Array. Angew. Chem. Int. Ed. 2014, 53, 13734–13739.

Page 8 of 9

(38) Nelson, T. L.; Tran, I.; Ingallinera, T. G.; Maynor, M. S.; Lavigne, J. J. Multi-Layered Analyses Using Directed Partitioning to Identify and Discriminate Between Biogenic Amines. Analyst 2007, 132, 1024–1030. (39) Bicker, K. L.; Sun, J.; Harrell, M.; Zhang, Y.; Pena, M. M.; Thompson, P. R.; Lavigne, J. J. Synthetic Lectin Arrays for the Detection and Discrimination of Cancer Associated Glycans and Cell Lines. Chem. Sci. 2012, 3, 1147–1156. (40) Kumar, P.; Kalita, A.; Mondal, B. Nitric Oxide Reactivity of Cu(Ⅱ) Complexes of Tetra- and Pentadentate Ligands: Structural Influence in Deciding the Reduction Pathway. Dalton Trans. 2013, 42, 5731–5739. (41) Hureau, C.; Groni. S.; Guillot, R.; Blondin, G.; Duboc, C.; Anxolabéhère-Mallart, E. Syntheses, X-ray Structures, Solid State HighField Electron Paramagnetic Resonance, and Density-Functional Ⅱ Theory Investigations on Chloro and Aqua Mn Mononuclear Complexes with Amino-Pyridine Pentadentate Ligands. Inorg. Chem. 2008, 47, 9238–9247. (42) Kim, I.-B.; Dunkhorst, A.; Gilbert, J.; Bunz, U. H. F. Sensing of Lead Ions by a Carboxylate-Substituted PPE: Multivalency Effects. Macromolecules 2005, 38, 4560–4562. (43) Kim, J.; McQuade, D. T.; McHugh, S. K.; Swager, T. M. IonSpecific Aggregation in Conjugated Polymers: Highly Sensitive and Selective Fluorescent Ion Chemosensors. Angew. Chem. Int. Ed. 2000, 39, 3868-3872.

ACS Paragon Plus Environment

Page 9 of 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

ACS Applied Materials & Interfaces

ACS Paragon Plus Environment