Pattern Recognition of Monosaccharides via a Virtual Lectin Array

Subscriber access provided by University of Colorado Anschutz Medical Campus Health Sciences Library

Article

Pattern Recognition of Monosaccharides via Virtual Lectin Array Constructed by Boronate Affinity-Based pH-Featured Encoding Xiaodong Bi, Daojin Li, and Zhen Liu Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.5b01034 • Publication Date (Web): 27 Mar 2015 Downloaded from http://pubs.acs.org on March 31, 2015

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.

Analytical Chemistry 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 28

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

Analytical Chemistry

Pattern Recognition of Monosaccharides via Virtual Lectin Array Constructed by Boronate Affinity-Based pH-Featured Encoding

Xiaodong Bi, Daojin Li and Zhen Liu*

State Key Laboratory of Analytical Chemistry for Life Science, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing 210093, China

* Corresponding author: [email protected]

1

ACS Paragon Plus Environment


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

Abstract: Lectin array is an important tool in the fields of carbohydrate chemistry, glycobiology and glycomics. Because natural lectins are associated with some apparent disadvantages such as tedious purification and easy loss of activity, artificial materials are applied to overcome such shortages by mimicking and replacing lectins in an artificial lectin array, amongst which boronate affinity-based materials are very outstanding and widely used. However, complicated synthetic works are often involved to design and create boronate affinity-based lectin-mimics. In this work, a facile and novel method was proposed to establish virtual lectin array based on boronate affinity-based pH-featured encoding for discrimination of monosaccharides by pattern recognition. The dependence of boronate affinity on environmental pH was selected to encode each monosaccharides for feature generation, and the pH-featured encoding was used to construct the virtual lectin array. On the basis of the virtual array, pattern recognition algorithms were applied for data analysis. Monosaccharides were discriminated by principal component analysis and the relations in the virtual lectin array were unraveled by cluster analysis. In this proof-of-concept work, without complicated synthesis or preparation, the proposed method was successful in mimicking lectin array and discriminating nine elementary monosaccharides found in the nature, and it was also a new way of encoding in expanding the applications of boronate affinity-based materials and methods in the field of biomimetics.

2


Page 2 of 28

Page 3 of 28

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


Keywords: Lectin, monosaccharides, boronate affinity, encoding, array, pattern recognition

INTRODUCTION SECTION

Lectins,

due to their specific and reversible binding with monosaccharides or

oligosaccharides,1 have been of great importance in many important areas such as carbohydrate chemistry, glycobiology and glycomics.2-5 Lectin-based methods have attracted increasing attention in analyzing and investigating glycan-related structures, functions and bio-processes.6-8 Amongst all lectin-based methods, lectin array, due to its integration of the diversity of lectins in their specificity toward sugars and biological activity, high-throughput array techniques and bioinformatics, has become the most promising and powerful researching tool to rapidly provide multiplexed and comprehensive data in large amounts.9-13

The recognition element or functional unit in lectin array is various lectins. However, due to their intrinsic protein nature, lectins are associated with apparent drawbacks such as tedious purification from plants and easy loss of activity, many efforts have been dedicated to exploit artificial materials with specific sugar-binding capability to mimic and replace lectins.11 The lectin-mimics include two major categories11 based on the kinds of their binding force with sugars: 1) noncovalent

3



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 28

systems such as supermolecules and aptamers; 2) and covalent systems mainly represented by boronate affinity-based covalent systems. The noncovalent systems often require complicated designs such as hydrogen bond groups, while the boronate affinity-based covalent systems possess known and strong binding forces as the straightforward design basis.11 The boronate affinity ligands, with their pH-controlled and reversible binding with cis-diol compounds,14,

15

are naturally suitable as the

building blocks in synthetic lectin-mimics. To this end, many boronate affinity-based materials have been developed, which can be classified into three categories: 1) boronate affinity ligands-modified materials for the separation of sugars,16 drug delivery17-21and cellular imaging or locating;22 2) synthetic organic probes with specially designed frameworks to regulate the spatial distance or electron deficiency of the chosen boronate affinity ligands in order to realize the specific recognition toward glycans or sugars of interest;23-26 3) boronate affinity-based molecularly imprinted polymers (MIPs) which combined nanoscale imprinting cavities with boronate affinity to provide enhanced specificity toward the sugar templates.27-31 Because the foregoing work all involved complicated synthesis or preparation, it is still of vital importance to develop facile and novel boronate affinity-based materials or methods for sugar recognition.

The property of recognition elements determines the category and function of arrays. When recognition elements are of natural specificity, the arrays are of physical

4


Page 5 of 28

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


array body, for example, a 10 × 10 biochips composed of antibodies.9, 32-35 Such array can be applied for qualitative and quantitative analysis of complex samples. When recognition elements are artificial materials, the tailor-made specificity is usually incomparable to natural specificity. To compensate this point, data processing techniques such pattern recognition are often applied to analyze the original data to discriminate different targets in the final array-like results. 29, 36-42 In this case, the array is of no physical body; that is, the obtained array is conceptual and virtual. In such array, the design of recognition elements is the foremost consideration and the virtual array can be used for qualitative analysis such as to discriminate similar compounds

36-42

or

the kinds of complex sample.29 Pattern recognition is the process during which observed samples are classified into categories or classes by means of certain algorithms, unraveling the internal relationship in the data obtained and to visually demonstrate the results thereof.43-46 The process of pattern recognition usually includes feature generation, feature selection and classification.46 The strength of boronate affinity towards sugars can be adjusted by many features. 14,47 This adjustability constitutes the foundation for pattern recognition. 46 So far, the features that could be used to adjust the strength of boronate affinity towards sugars were changes of boronate affinity ligands23-26 and boronate affinity-based molecularly imprinting, 29 which both required dedicated and complex synthetic or preparative work. Therefore, it is necessary and inspiring to develop facile and simple methods to mimic lectins or lectin array based on

5



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

other features that function to adjust the strength of boronate affinity and meanwhile require less synthetic work.

In this work, we constructed a virtual lectin array by boronate affinity-based pH-featured encoding for pattern recognition of monosaccharides. The principle of the proposed approach is shown in Figure 1, which included the following three sequential sections: 1) boronate affinity-based pH-featured encoding, which included two sub-parts: a) input: based on the dependence of boronate affinity on environmental pH, three pHs were chosen to represent different binding capabilities of the chosen boronate affinity ligand with monosaccharides; b) processing: at each pH, the signal was produced in the form of descending extent of fluorescence intensity caused by the displacement of each monosaccharides with molecules of a dye called Alizarin Red S. (ARS) which produced strong fluorescence signal when binding to boronate affinity ligand while weak fluorescence signal when dissociated48; 2) construction of virtual lectin array: the group of three data obtained constituted the codes for each monosaccharide (pH-featured encoding), and then all the codes were output and transformed to construct the virtual lectin array, in which area and color depth of each circle represented different binding capacity of each monosaccharides with the chosen boronate affinity ligand; 3) analysis by pattern recognition: herein two commonly used pattern recognition algorithms, principal component analysis and cluster analysis, were applied for data analysis. For this proof-of-concept work, nine elementary

6


Page 6 of 28

Page 7 of 28

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


monosaccharides found in natural glycans were chosen as the test compounds because they are pure and neutral carbohydrates in the form of monosaccharide hexoses or pentoses without any other substitutional groups.11 The pH-dependence of boronate affinity on environment was successfully employed as a feature to encode monosaccharides, on the basis of which virtual lectin array was constructed. Then different monosacharides were discriminated by pattern recognition, and the relations in the virtual array were further unraveled by cluster analysis. By this novel approach, the nine elementary monosaccharides were successfully discriminated. The construction of virtual lectin array by boronate affinity-based pH-featured encoding was proved to be a facile and effective method to mimic lectin array with simply constructed boronate affinity materials and less synthetic work.

EXPERIMENTAL SECTION Reagents and materials. D-(+)-Mannose was purchased from Alfa Aesar (Tianjin, China). ARS, L-(-)-fucose and 3-aminopropyltriethoxysilane (APTES) were purchased from Aladdin Reagent (Shanghai, China). 4-Formylphenylboronic acid, D-(+)-glucose, D-(+)-xylose, D-(+)-galactose and sodium cyanoborohydride were from J&K Scientific (Beijing, China). L-(-)-Sorbose, D-(-)-arabinose, D-(-)-ribose, D-(-)-fructose, anhydrous ethanol, H2SO4 (98 %), H2O2 (30%), NaH2PO4 and Na2HPO4 were of analytical grade and purchased from Nanjing Reagent Company. Qualitative filter papers were 7



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

purchased from WoHua Filter Paper Company (Hangzhou, China). Ultrapure water, purified with a Milli-Q Advantage A10 (Millipore, Milford, MA, USA), was used to prepare all solutions.

Instruments. Synergy Mx microplate reader from BioTek (Winooski, VT, USA) was used for the detection. Scanning electron microscopic (SEM) characterization was performed on a FE-SEM S-4800 system (Hitachi, Tokyo, Japan). Polystyrene (PS) 96-well microplates including the lids (LOT No.165305) from Thermo Fisher Scientific (San Jose, CA, USA) were used for all experiments.

Preparation of 4-formylphenylboronic acid-functionalized filter paper. The boronic acid-functionalization process was adopted from a literature method49 with some major modifications. Briefly, a piece of filter paper was immersed into 20-fold water-diluted solution of 7:3 (v/v) H2SO4 (98%): H2O2 (30%) mixture for 30 min at room temperature. After being washed with water to neutral pH, the filter paper was washed with anhydrous ethanol to remove water and dried at 60 °C. Then the filter paper was immersed into an anhydrous ethanol solution containing 10% (v/v) APTES and the container was sealed and kept at 60 °C for 12 h. After that, the filter paper was washed with anhydrous ethanol to remove reactants and dried at 60 °C. Then the filter paper

8


Page 8 of 28

Page 9 of 28

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


was

immersed

into

an

anhydrous

ethanol

solution

containing

5

mg/mL

4-formylphenylboronic acid and 5 mg/mL sodium cyanoborohydride, and the container was sealed and kept at 60 °C for another 12 h. All the reaction solutions were excessive to swamp the filter paper. After being washed with anhydrous ethanol to remove reactants, the filter paper was dried and cut into circular pieces with diameter of 12 mm. Circular filter paper pieces of different modifications were all collected and kept sealed for later use.

Characterization of phenylboronic acid-functionalized filter paper. ARS was used to characterize the modification of phenylboronic acid onto the filter paper for its switchable fluorescence signal by virtue of the binding with boronic acid.48 Briefly, one circular piece of phenylboronic acid-functionalized filter paper, one circular piece of APTES-modified filter paper and one circular piece of unmodified filter paper were respectively immersed into 1 mL ARS solution of 1 × 10-5 M in 50 mM phosphate buffer (pH 9.4). After slightly shaking at room temperature for 5 min, the wet paper pieces were washed by water and dried by repeatedly touching with dry intact filter paper. Then the filter paper pieces were placed onto the well-shaped grids on the lid of a microplate (See Figure S1 in the Supporting Information), and fluorescence detection (excitation wavelength, 490 nm; emission wavelength, 590 nm; the slit width, 9 nm; the gain, 100; all the settings are the same hereinafter) was performed immediately. All the

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

experiments were performed for three times. Besides, the porous morphology of modified filter papers was characterized by SEM.

Detection of monosaccharides by dye displacement method. Phenylboronic acid-functionalized filter paper pieces were immersed into 1 mL of 60 mM monosaccharide solutions containing 1×10-5 M ARS in 50 mM phosphate buffer of different pH values (7.4, 8.5 and 9.4). After slightly shaking at room temperature for 5 min, the wet filter paper pieces were washed with water and dried by repeatedly touching with dry intact filter paper. Then the filter paper pieces were placed onto the well-shaped grids on the lid of a microplate, and fluorescence detection was performed immediately. All the experiments were performed for three times. For signal normalization, 1 mL of 1 × 10-5 M ARS solutions without monosaccharides in 50 mM phosphate buffer of different pH values (7.4, 8.5 and 9.4) were used to give the initial signals at corresponding pH which reflected the binding of ARS onto the phenylboronic acid-functionalized filter paper pieces.

Construction of virtual lection array and data analysis. The descending extent of fluorescence signal of each monosaccharide at each pH was represented by binding ratio (BRpH) through the following equation using the resulting fluorescence signal after

10


Page 10 of 28

Page 11 of 28

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


ARS was displaced by each monosaccharide (FpH) and the initial fluorescence signal ( F0pH ) of ARS at corresponding pH: BR pH = 1 − F pH / F0pH (pH = [7.4 8.5 9.4])

(1)

Then the data matrix of (BRpH × species) (pH = [7.4 8.5 9.4]) was used to construct the virtual lectin array. Principal component analysis (PCA) and cluster analysis were used for data analysis, which were carried out using the pattern recognition algorithms with the help of Matlab software (version 7.13, MathWorks, Natick, MA).

RESULTS AND DISCUSSION Characterization of phenylboronic acid-functionalized filter paper. The successful modification of phenylboronic acid onto the filter paper was confirmed using the turn-on fluorescence of ARS due to its binding to phenylboronic acid.47 As shown in Figure 2, the phenylboronic acid-modified filter paper demonstrated a much stronger fluorescence signal than the APTES-modified and the unmodified filter papers when binding to ARS, which confirmed the successful modification of phenylboronic acid. Besides, SEM characterization revealed that the modified filter paper still possessed porous structure (Figure S2, the Supporting Information), which is favorable to rapid mass transfer.

11



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

Theoretical considerations for boronate affinity-based pH-featured encoding. The binding of boronic acid ligands with cis-diol containing compounds is pH-dependent. 14 Experimental data, such as titration measurement of binding constant of boronic acid ligand with hydrogen ion or cis-diol compound to give information concerning pKa or dissociation constant, have been the factual confirmations that there exists a jumping change around the pKa.15, 23, 24, 26, 47, 50-53 And in the vicinity of the pKa value, different pH values could to the largest extent alter the binding force between the boronic acid ligand and the cis-diol compound. And the pH-adjusted binding strength could constitute the feature generation required in pattern recognition; that is, the different binding data from individual pHs could be used to encode the binding events as the selected features and then to discriminate the binding between cis-diol compounds with a boronic acid ligand by the pattern recognition algorithms.

Construction of virtual lectin array. For this proof-of-concept work, we chose one simply-structured boronate affinity ligand, 4-formylphenylboronic acid, for it contains a formyl group for chemical attachment onto a supporting material and no other substitutional groups that may change the pKa value.14 ARS was used as a signal reporter due to its turn-on fluorescence when binding to phenylboronic acid.48 Since the dye displacement could be triggered by different monosaccharides to give discriminative results,29,52,53 the leveling effect of ARS towards monosaccharides in the

12


Page 12 of 28

Page 13 of 28

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


binding of phenylboronic acid could be neglected. First, the original data of spectra at the three pH values (first three graphs in Figure S3 A-I) and the binding ratios (the fourth graph in Figure S3 A-I) were collected according to procedures described in experimental section. And then as shown in Figure 3 (taking glucose as the example), the data were transmitted into the plane array form by linear mapping the data value into the RGB color space and also by expressing the data value as the radius through the circular area in the array. The resulting array demonstrates the pH-featured encoding of different saccharides, in which a larger circular area or a darker color means a higher binding ratio and vice versa. The control means a zero binding ratio as shown in dash line and no filling. Figure 4 shows all the transformed pH-featured encoding in a recombinant array, the virtual lectin array, and the results are clear in a whole sight. Under the experimental conditions, in the virtual lectin array the fructose ranked the strongest binding at all pHs and followed by sorbose. Then arabinose and fucose seemed to have the similar pH-featured encoding, and the rest appeared to be another group. For further discrimination and better understanding of the virtual lectin array, pattern recognition was applied.

Principal component analysis. On the basis of the pH-featured encoding of monosaccharides in the virtual lectin array, principal component analysis (PCA) was applied for further discrimination. PCA is a common algorithm in pattern recognition

13



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

usually for feature extraction and building feature space out of the given data.43-46 As shown in Figure 5, the first two PCA factors total to more than 99 %, which means that the 2-D PCA could to a great extent cover and express the interrelation of the analyzed object. And after the PCA process, all the samples are far from the control group in the feature space, and each monosaccharide is classified into one compact region in the feature space; that is, in the feature space monosaccharides could be distinguished from one another and the choice of the pH-featured encoding as the feature generation is successful. Ribose, glucose and fucose are near with each other due to their close patterns in the virtual lectin array, which is mainly contributed to the similarity of behaviors of these monosaccharides under the experimental conditions. While others with less close patterns or much different patterns are mutually distant, which means that under the experimental settings the behaviors of these monosaccharides are discriminatable to give qualitative results. To sum up, the boronate affinity-based pH-featured encoding acted as a good feature for discriminating monosaccharides.

Cluster analysis. The cluster analysis algorithm was also applied to analyze the selected feature by PCA. Cluster analysis is an unsupervised classification method based on the similarities in the given data set.43-46 More similar objects tend to be clustered apart from less similar ones; that is, the cluster is an underlying pattern. The deviation of monosaccharides from the control decides the cluster belongings, and such

14


Page 14 of 28

Page 15 of 28

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


deviation was contributed to the similar behavior of the monosaccharides under the experimental conditions. As shown in Figure 6, the cluster analysis is in a hierarchical manner. The most similar monosaccharides gathered at the first level, and then the grouped ones found their closest similar counterpart at the second level and so forth. Sorbose and galactose, glucose and fructose, ribose and arabinose, xylose and fucose met with each other at the first level, which meant that the two in the pair were calculated to be the most similar in their deviation from the control. At the second level, mannose was classified with the xylose and fucose pair, and the sorbose and galactose pair and the glucose and fructose pair converged. And then the ribose and arabinose pair met with the mannose-xylose-fucose group. The cluster analysis qualitatively reflected the intrinsic relation in the virtual lectin array, and its result could be validated by those of PCA. Although the relevance between the distinct behaviors of monosaccharides and structures or thermodynamics thereof was not fully exploited for direct correlation of the selective feature (the pH-featured encoding) to the structures or thermodynamics was not the first design and consideration, the proof-of-concept work did provide a methodological feasibility to construct the virtual lectin array and also to expand the applicability of boronate affinity in the field of biomimetics as well.

15



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

CONCLUSION In this proof-of-concept work, a virtual lectin array was established by boronate affinity-based pH-featured encoding for pattern recognition of monosaccharides. Without complicated synthetic work, simply-structured boronate affinity material was used, and the pH-dependent property of boronate affinity was used to generate codes for each monosaccharides. For signal production, ARS dye displacement was applied to give the encoding data. The pH-featured encoding results were then used to construct the virtual lectin array. Principal component analysis was introduced for detailed discrimination of monosaccharides, and cluster analysis was also applied to unravel the internal relations in the virtual lectin array. Nine elementary monosaccharides found in the nature were successfully discriminated. The virtual lectin array proposed in this work can be used for qualitative analysis of monosaccharides in simple samples. For further development, if a new recognition element that contains structural or functional information is introduced, the virtual lectin array will be able to provide more in-depth qualitative results for monosaccharides or other more complex carbohydrates.

16


Page 16 of 28

Page 17 of 28

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


ACKNOWLEDGEMENTS We acknowledge the financial support of the National Science Fund for Distinguished Young Scholars (No. 21425520) and the general grant (No. 21275073) from the National Natural Science Foundation of China, and the Key Grant of 973 Program (No. 2013CB911202) from the Ministry of Science and Technology of China.

SUPPORTING INFORMATION This material is available free of charge via the Internet at http://pubs.acs.org.

17



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

REFERENCES [1] H. Rudiger; H.J. Gabius. Glycoconjugate J. 2001, 18, 589-613. [2] R.A. Dwek. Chem. Rev. 1996, 96, 683-720. [3] C.R. Bertozzi; L.L. Kiessling. Science 2001, 291, 2357-2364. [4] P.M. Rudd; T. Elliott; P. Cresswell; I.A. Wilson; R.A. Dwek. Science 2001, 291, 2370-2376. [5] S.L. Flitsch; R.V. Ulijn. Nature 2003, 421, 219-220. [6] D. Clark; L. Mao. Dis. Markers 2012, 33, 1-10. [7] P. Gemeiner; D. Mislovicova; J. Tkac; J. Svitel; V. Patoprsty; E. Hrabarova; G. Kogan; T. Kozar. Biotechnol. Adv. 2009, 27, 1-15. [8] J. Hirabayashi. J. Biochem. 2008, 144, 139-147. [9] H. Li; L.M. Wei; P. Fang; P.Y. Yang. Anal. Methods 2014, 6, 2003-2014. [10] S. Park; J.C. Gildersleeve; O. Blixt; I. Shin. Chem. Soc. Rev. 2013, 42, 4310-4326. [11] J. Hirabayashi; M. Yamada; A. Kuno; H. Tateno. Chem. Soc. Rev. 2013, 42, 4443-4458. [12] G. Gupta; A. Surolia; S.G. Sampathkumar. Omics 2010, 14, 419-436. [13] J.A. Maynard; R. Myhre; B. Roy. Curr. Opin. Chem. Biol. 2007, 11, 306-315. [14] H.Y. Li; Z. Liu. TRAC-Trends Anal. Chem. 2012, 37, 148-161. [15] Y. Lu; Z.J. Bie; Y.C. Liu; Z. Liu. Analyst 2013, 138, 290-298. [16] Z.F. Xu; K.M. A. Uddin; T. Kamra; J. Schnadt; L. Ye. ACS Appl. Mater. Inter. 2014, 6, 1406-1414. [17] W. Chen; S.C. Mehta; D.R. Lu. Adv. Drug Deliv. Rev. 1997, 26, 231-247. [18] Y. Zhao; B.G. Trewyn; I.I. Slowing; V.S.Y. Lin. J. Am. Chem. Soc. 2009, 131, 8398-8400. [19] W. Chen; Y. Cheng; B.H. Wang. Angew. Chem. Int. Ed. 2012, 51, 5293-5295. [20] G.A. Ellis; M.J. Palte; R.T. Raines. J. Am. Chem. Soc. 2012, 134, 3631-3634. [21] Y. Li; W. Xiao; K. Xiao; L. Berti; J. Luo; H.P. Tseng; G. Fung; K.S. Lam. Angew. Chem. Int. Ed. 2012, 51, 2864-2869. [22] X. Chen; P. Wu; M. Rousseas; D. Okawa; Z. Gartner; A. Zettl; C.R. Bertozzi. J. Am. Chem. Soc. 2009, 131, 890-891.

18


Page 18 of 28

Page 19 of 28

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


[23] J. Yan; H. Fang; B.H. Wang. Med. Res. Rev. 2005, 25, 490-520. [24] R. Nishiyabu; Y. Kubo; T.D. James; J.S. Fossey. Chem. Commun. 2011, 47, 1106-1123. [25] J. Arnaud; A. Audfray; A. Imberty. Chem. Soc. Rev. 2013, 42, 4798-4813. [26] X. Wu; Z. Li; X.X. Chen; J.S. Fossey; T.D. James; J.S. Fossey. Chem. Soc. Rev. 2013, 42, 8032-8048. [27] G. Wulff; S. Schauhoff. J.Org.Chem.1991, 56, 395-400. [28] G. Vozzi; I. Morelli; F. Vozzi; C. Andreoni; E. Salsedo; A. Morachioli; P. Giusti; G. Ciardelli, Biotechnol. Bioeng. 2010, 106, 804-817. [29] J. Tan; H.-F. Wang; X.-P. Yan. Anal. Chem. 2009, 81, 5273-5280. [30] P. Parmpi; P. Kofinas. Biomaterials 2004, 25, 1969-1973. [31] T. Miyahara; K. Kurihara. Chem. Lett. 2000, 12, 1356-1357. [32] P. Cullen; S. Lorkowski. Expert Opin. Ther. Patents. 2002, 12, 1783-1794. [33] H. Eickhoff; Z. Konthur; A. Lueking; H. Lehrach; G. Walter; E. Nordhoff; L. Nyarsik; K. Bussow. Adv. Biochem. Eng./Biotechnol. 2002, 77, 103-112. [34] J. Hudson; M. Altamirano. J. Ethnopharmacol. 2006, 108, 2-15. [35] L.J. Kricka; K. Imai; P. Fortina. Clin. Chem. 2010, 56, 1797-1803. [36] D. Shibata. Am. J. Pathol. 1999, 154, 979-980. [37] K.J. Albert; N.S. Lewis; C.L. Schauer; G.A. Sotzing; S.E. Stitzel; T.P.Vaid; D.R.Walt. Chem. Rev. 2000, 100, 2595-2626. [38] J.H. Lu; C. Teh; U. Kishore; K.B.M. Reid. Biochim. Biophys. Acta-Gen. Subj. 2002, 1572, 387-400. [39] B. Adhikari; S. Majumdar. Prog. Polym. Sci. 2004, 29, 699-766. [40] B.K. Lavine; C.E. Davidson; W.S. Rayens. Comb. Chem. High Throughput Screen 2004, 7, 115-131. [41] S.M. Scott; D. James; Z. Ali. Microchim. Acta 2006, 156, 183-207. [42] Y. Liu; L. Ding; Y. Cao; Y. Fang. Prog. Chem. 2012, 24, 1915-1927. [43] A.K. Jain; R.P.W. Duin; J.C. Mao. IEEE Trans. Pattern Anal. Mach. Intell. 2000, 22, 4-37. [44] J.S. Suri. Pattern Anal. Appl. 2000, 3, 209-242.

19



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

[45] M.H. Yang; D.J. Kriegman; N. Ahuja. IEEE Trans. Pattern Anal. Mach. Intell. 2002, 24, 34-58. [46] S. Theodoridis; K.Koutroumbas. Pattern Recognition, 4th Edition; China Machine Press: Beijing, China , 2009. [47] C.C. Lü; H.Y. Li; H.Y. Wang; Z. Liu. Anal. Chem. 2013, 85, 2361-2369. [48] G. Springsteen; B.H. Wang. Chem. Commun. 2001, 17, 1608-1609. [49] X.D. Bi; Z. Liu. Anal. Chem. 2014, 86, 959-966. [50] G. Springsteen; B.H. Wang. Tetrahedron 2002, 58, 5291-5300. [51] J. Yan; G. Springsteen; S. Deeter; B.H. Wang. Tetrahedron 2004, 60, 11205-11209. [52] S. Shariki; O.T.L. Cox; D.A. Tickell; M.P.P. Morais; J.M.H. van den Elsen; T.D. James; S.E.C. Dale; S. Bending; F. Marken. J. Mater. Chem. 2012, 22, 18999-19006. [53] R. Zhang; J.P. Miao; C.H. Ge; M.Y. Zhang; L.X. Wang; X.D. Zhang. Sens. Actuator B-Chem. 2014, 198, 260-267.

20


Page 20 of 28

Page 21 of 28

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


Figure Captions

Figure 1. Schematic diagram of construction of virtual lectin array by boronate affinity-based pH-featured encoding for pattern recognition of monosaccharides.

Figure 2. Characterization of modification of phenylboronic acid by dye replacement.

Figure 3. Process of pH-featured encoding (with glucose as an example).

Figure 4. Construction of virtual lectin array.

Figure 5. Principal component analysis.

Figure 6. Cluster analysis.

21



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

Figure 1

22


Page 22 of 28

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


Fluorescence Intensity /104

Page 23 of 28

8

6

4

2

0

Paper

NH2@paper PhB(OH)2@paper

Figure 2

23



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

Figure 3

24


Page 24 of 28

Page 25 of 28

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


Figure 4

25



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

Figure 5

26


Page 26 of 28

Page 27 of 28

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


Figure 6

27



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

For TOC only

28


Page 28 of 28

Pattern Recognition of Monosaccharides via a Virtual Lectin Array

Recommend Documents