Functionalized gold nanoparticles as affinity nanoprobes for multiple lectins

Accepted Manuscript Title: Functionalized Gold Nanoparticles as Affinity Nanoprobes for Multiple Lectins Authors: Karuppuchamy Selvaprakash, Yu-Chie Chen PII: DOI: Reference:

S0927-7765(17)30758-0 https://doi.org/10.1016/j.colsurfb.2017.11.022 COLSUB 8973

To appear in:

Colloids and Surfaces B: Biointerfaces

Received date: Revised date: Accepted date:

6-6-2017 11-10-2017 7-11-2017

Please cite this article as: Karuppuchamy Selvaprakash, Yu-Chie Chen, Functionalized Gold Nanoparticles as Affinity Nanoprobes for Multiple Lectins, Colloids and Surfaces B: Biointerfaces https://doi.org/10.1016/j.colsurfb.2017.11.022 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Functionalized Gold Nanoparticles as Affinity Nanoprobes for Multiple Lectins Karuppuchamy Selvaprakash and Yu-Chie Chen*

SC RI PT

Department of Applied Chemistry, National Chiao Tung University, Hsinchu 300, Taiwan *

A

CC

EP

TE

D

M

A

N

U

Corresponding author E-mail: [email protected] Phone: +886-3-5131527 Fax: +886-3-5723764

1

SC RI PT U N A M D TE EP CC A Highlights 

Ovalbumin immobilized Au NPs were generated from one-pot reactions.



The Au NPs containing glycan-ligands have the capability to trap several lectins. 2



MALDI-MS is used as the detection tool.



The limit of detection of this method for model lectins is in low nM level.

SC RI PT

Abstract Glycan–lectin interactions are commonly observed in nature. Analytical methods, which are used to detect lectins that rely on the use of glycan ligand-modified nanoprobes as affinity probes, have been developed. Most of the existing methods are focused on the use of synthetic glycan ligands. Nevertheless, naturally available glycoproteins, such as ovalbumin in chicken egg

U

whites, are good sources for fabricating glycan-immobilized nanoprobes. In this study, we

N

generated functionalized gold nanoparticles (AuNPs) from a one-pot reaction by reacting

A

chicken egg white (cew) proteins with aqueous tetrachloroaurate. The generated Au@cew NPs

M

are mainly encapsulated by ovalbumin, in which the surface is decorated by abundant hybrid

D

mannose and Galβ(14)GlcNAc-terminated glycan ligands. Thus, the generated Au@cew NPs

TE

containing hybrid mannose and Galβ(14)GlcNAc have the capability to selectively bind with their corresponding lectins. Lectins including concanavalin A, banana lectin, and ricin B that

EP

have binding moieties toward specific sugars were used as the model samples. Our results showed that the generated AuNPs can be used as multiplex affinity probes for these model

CC

lectins. Lectins can be selectively released from the Au@cew NP-lectin conjugates by using specific sugars, such as mannose, glucose, and β-lactose, as the releasing agents to release

A

specific lectins from the conjugates. Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) was used as the tool to characterize the released species from the nanoprobes. The limit of detection of these model lectins using the current approach was low (in

3

nM). The feasibility of using the Au@cew NP-based affinity MALDI-MS to selectively detect specific lectins from complex samples was also demonstrated. Keywords: Lectins; concanavalin A; banana lectin; ricin; chicken ovalbumin

SC RI PT

Introduction

Lectins are proteins with specific carbohydrate recognition sites [1]; they exist in plants [2], animals [1,2], and microorganisms [3]. Concanavalin A (Con A) is a well-known plant lectin that is present in jack bean seeds. Con A contains specific binding moieties against glucose and mannose [4,5]. The applications of using the interactions between Con A and mannose/glucose

U

containing ligands have been extensively employed to biochemical research [6,7]. Con A has

N

also considered as a potential anticancer drug [8]. In addition, lectins, such as ricin, are highly

A

toxic [9]. Ricin is produced from the seeds of castor oil plant (Ricinis communis) [9] and can

M

inactivate the function of protein synthesis in ribosomes [10,11]. Ricin consists of A chain (MW

D

= ~32 KDa) and B chain (MW= ~34 KDa), which are linked by a disulfide bond [12]. Ricin A

TE

chain possesses catalytic activity and is responsible for cellular toxicity. Ricin B chain is a lectin, which can facilitate the binding of ricin onto glycoproteins and glycolipids that contains terminal

EP

galactose on the cell membrane of host cells [12–16]. Ricin has been used as a biological weapon

CC

[17] and has been listed as a potential bioterrorism risk group category B agent by the Center for Disease Control and Prevention (CDC) in the USA [18]. Banana lectin (BanLec), which is a

A

mannose/glucose binding lectin [19], has been considered as a potential agent for inhibition of the replication of human immunodeficiency virus type 1 (HIV-1) [20]. Recently, a drug engineered from BanLec has been demonstrated as the killing agent for viruses including hepatitis C and influenza [21]. Lectins such as ricin, may cause toxicity; other lectins, such as con A and BanLec may be used as medicines. Thus, to develop affinity probes that can be used 4

to detect or to purify these lectins from complex samples is significant. Furthermore, one affinity probe that is capable to selectively trap multiple lectins is even ideal. Gold nanoparticles (Au NPs) have been used as affinity probes for lectins because they can

SC RI PT

be easily synthesized and functionalized [22–25]. Synthetic glycan or antibody-coated Au NPs are commonly used as affinity probes to trap lectins, such as Con A [26–28], BanLec [29], and ricin B [30–32]. However, considering the preparation time and the cost of synthetic glycans and antibodies, alternative probe molecules should be explored. Furthermore, alternative probe molecules should also possess affinities toward multiplex lectins. The time spent in preparation

U

and the cost spent in obtaining different probes can be greatly reduced. To the best of our

N

knowledge, no reports have demonstrated that one nanoprobe can possess multiplex affinity

A

capabilities against different lectins, such as con A, BanLec, and ricin.

M

Protein-direct synthesized Au nanoclusters [33–38] and Au NPs [39–42] from one-pot reactions can readily immobilize reactant proteins onto the surface of the generated Au NPs. The

D

function of proteins on the surface of the generated particles may still remain [43]. The

TE

functional groups derived from the protein on the surface of the generated particles can be

EP

readily used to probe their target species. Ovalbumin is a glycoprotein and is abundant (~54%) in chicken egg white proteins [44]. Galβ(14)GlcNAc ligands on ovalbumin can be used to

CC

interact with lectins that have the specific pocket for this glycan ligand [45,46]. Nevertheless, other glycan ligands, such as hybrid mannose, are also present in the structure of ovalbumin [47].

A

Thus, these glycan ligands available on ovalbumin should be able to interact with their corresponding lectins, such as ricin B, Con A, and BanLec. It has been reported that ovalbumin immobilized Au NCs [42, 45] can be readily generated from one-pot reactions. In this study, we generated ovalbumin-immobilized Au NPs from one-pot reaction by reacting chicken egg white

5

proteins with tetrachloroacuric acid (Au@cew NPs). The feasibility of using the generated Au@cew NPs as affinity probes for multiple lectins including Con A, BanLec, and ricin B was demonstrated. The selectivity of this approach was implemented by using different saccharides

SC RI PT

as releasing agents to selectively release target lectins from the nanoprobes. Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS), which is a suitable analytical tools for large biomolecules, was used as the detection tool for characterization the released lectins.

Experimental Section

U

Reagents and materials

N

Tryptic soy broth (TSB) and yeast (Y) extract were purchased from Becton Dickinson

A

(Franklin Lakes, NJ, USA). Ammonium bicarbonate, sodium chloride, and urea were purchased

M

from J. T. Baker (Phillipsburg, NJ, USA.). Acetonitrile was acquired from Merck (Darmstadt, Germany). Ammonium sulfate, sodium hydroxide, and trifluoroacetic acid (TFA) (99%) were

D

obtained from Riedel-de Haën (Seelze, Germany). Hydrogen tetrachloroaurate (III) tetrahydrate

TE

was obtained from Showa (Tokyo, Japan). Calcium chloride, Con A, α-cyano-4hydroxycinnamic acid (CHCA), cytochrome c (from bovine heart), 4-(2-hydroxyethyl)-1-

EP

piperazineethanesulfonic acid (HEPES), -lactose, magnesium chloride, mannose, myoglobin

CC

(from horse heart), sinapinic acid, sodium phosphate dibasic, potassium phosphate monobasic, and trypsin (from bovine pancreas) were purchased from Sigma-Aldrich (St. Louis, MO, U.S.A.).

A

Musa paradisiaca lectin (BanLec) and ricin B chain were obtained from Vector Laboratories (Burlingame, CA, USA.). Human sera were donated by a healthy individual. Amicon Ultra-4 centrifugal filters were purchased from Millipore (Billerica, MA, USA). Pall acrodisc syringe filters (PN-4612; pore size, 0.2 µm) were obtained from Voigt Global Distribution (Lawrence, KS, U.S.A.). Escherichia coli O157:H7 (BCRC 13085) was purchased from the Bioresource 6

Collection and Research Center (Hsinchu, Taiwan). Fresh chicken eggs and ripe bananas were obtained from a local supermarket. Instrumentation

SC RI PT

Mass spectra were obtained using an Autoflex III Smartbeam MALDI-time-of-flight mass spectrometer (Bruker Daltonics, Bremen, Germany), which was equipped with a solid-state Nd:YAG laser (λ= 355 nm). The laser frequency was set to 100 Hz. Each mass spectrum was acquired from 1000 laser shots in linear positive ion mode for protein analysis. The MS parameter settings were listed as follows: ion source 1, 20.00 kV; ion source 2, 18.40 kV; lens,

U

6.85 kV. When operating in reflectron mode for peptide analysis, the voltages were set as

N

follows: ion source 1, 19.2 kV; ion source 2, 16.55 kV; lens, 8.75 kV, reflector 1, 21.0 kV;

A

reflector 2, 9.70 kV. Mass spectra were acquired at the range of m/z 1000−5000. The ion

M

suppression was set at m/z 900. Absorption spectra were obtained using a Varian Cary 50 UV-

Synthesis of Au@cew NPs

D

Vis absorption spectrophotometer (Melbourne, Australia).

TE

Chicken egg white protein-directed AuNP synthesis (Au@cew NPs) was conducted in one-

EP

pot reactions. That is, egg white was separated from chicken eggs followed by 100-fold dilution with deionized water. The diluted mixture was stirred for 15 min followed by centrifugation at

CC

3750g (rotor radius, 93 mm) for 20 min. The concentration of the chicken egg white protein solution was ~3 mg/mL that was estimated by lyophilizing a give volume of the resultant chicken

A

egg white protein solution. The resultant clear solution (~3 mg/mL, 5 mL) was stirred further with aqueous tetrachloroauric acid (0.1 M, 100 µL) for 5 min, and aqueous NaOH (2 N, 100 µL) was added to the mixture under constantly stirring for another 5 min. The mixture was then placed in a water bath maintained at 60 °C under stirring at 400 rpm for 6 h. After 6 h, the

7

resultant Au@cew NPs were centrifuged at 26,810g (4 °C) for 45 min, and the supernatant was removed. The particles at the bottom of the vial were further rinsed with deionized water (1 mL  3) followed by centrifugation at 26,810g (4 °C) for 45 min to remove supernatant. The settled

SC RI PT

Au@cew NPs were re-dispersed in deionized water (0.2 mL) and stored at 4 °C prior to use. Tryptic digestion of Au@cew NPs

The proteins on the surface of Au@cew NPs were characterized by combining trypsin digestion with MALDI-MS analysis. Au@cew NPs (2.6 mg mL−1, 5 µL) were mixed with trypsin (5 µL, 5 µM) prepared in ammonium bicarbonate buffer (50 mM, pH ~8). Microwave-

U

heating was used to accelerate the tryptic digestion [48,49]. That is, the mixture was placed into

N

a water bath (10 mL) followed by subjection in a microwave oven (power, 900 W) for 1 min.

A

Five heating cycles (1 min/cycle) were conducted, whereas water in the water bath was renewed

M

between heating cycles to prevent overheating. TFA (3%, 2 µL) was then added to the digest

D

mixture to stop the enzymatic reaction. The resultant tryptic digest (1 µL) was mixed with the

TE

MALDI matrix, i.e., CHCA (1 µL, 15 mg/mL), prepared in acetonitrile/0.1% TFA (2:1, v/v). The mixture was deposited on a MALDI plate. After solvent evaporation, the sample was ready for

EP

MALDI-MS analysis.

Using Au@cew NPs as the affinity probes for model lectins

CC

Lectins including Con A, BanLec, and ricin B were selected as model samples. Lectins were

initially prepared in HEPES buffer (20 mM, pH 7.2–7.5). HEPES buffer was added with

A

additional NaCl (50 mM), calcium chloride (1 mM), and magnesium chloride (1 mM). The binding experiments using Au@cew NPs as affinity probes towards the lectin samples were carried out by shaking (220 rpm) Au@cew NPs (2.6 mg/mL, 20 µL) with lectin samples (0.5 mL) in HEPES buffer (pH 7.5, pH 7.2 for BanLec) for 2 h. The resultant sample was centrifuged at

8

26810g for 10 min, and the supernatant was removed. The NP–lectin conjugates were further rinsed with HEPES buffer (0.5 mL  2) under centrifugation at 26810g for 10 min. The resultant conjugates were mixed with saccharides (50 mM, 10 µL), such as glucose, mannose,

SC RI PT

and β-lactose followed by pipetting the mixture in and out of the tip in a tube vigorously for 2 min. The resultant solution was centrifuged at 26810g for 10 min. The supernatant (2 µL) containing the released target species was mixed with MALDI matrix, i.e., sinapinic acid (25 mg mL−1, 2 µL), prepared in acetonitrile/0.1% TFA (2:1, v/v). The mixture (1 L) was quickly deposited on a MALDI plate. After solvent evaporation, the sample was ready for MALDI-MS

N

Preparation of crude extract from banana pulp

U

analysis.

A

The crude extract from banana pulp was prepared using the procedures from a previous study

M

[29]. The outer peel of a ripe banana was removed manually. Banana pulp (5 g) was

D

homogenized in phosphate buffer saline (PBS) (pH 7.2, 100 mM, 10 mL), which was prepared

TE

by mixing monobasic potassium phosphate (0.1 M) and dibasic sodium phosphate (0.1 M). PBS buffer used for homogenization was added with additional sodium chloride (450 mM) and 2-

EP

mercaptoethanol (10 mM). 2-Mercaptoethanol was used to prevent protein oxidation. The extract was stirred at room temperature (~25 °C) for 2 h followed by centrifugation at 3750g for 15

CC

min. The resultant supernatant was passed through a filter membrane (pore size, 0.2 µm). The

A

filtrate was added with aqueous ammonium sulfate (0.36 g/mL) to precipitate proteins. The resultant precipitates were collected by centrifuging the mixture at 26810g (4 °C) for 30 min. The collected precipitates were dissolved in deionized water (4 mL) followed by desalting steps. The solution was loaded on a filter membrane (cutoff mass, 3 kDa) followed by centrifugation at

9

3750g for 30 min. The resultant species (~0.2 mL) remained on the filter membrane was further added with HEPES buffer (pH 7.2, 0.8 mL) and stored at 4 °C prior to use. Selective enrichment of BanLec from banana extract

SC RI PT

Au@cew NPs (2.6 mg/mL, 20 L) were mixed with the crude banana extract (100 L) obtained above under shaking at 220 rpm for 2 h followed by centrifugation at 26810g for 10 min. After the supernatant was removed, the resultant precipitates containing the conjugates of Au@cew NP-target species were rinsed with HEPES buffer (pH 7.2, 0.1 mL  3) under centrifugation at 26810g for 10 min. The resultant conjugates were mixed with mannose (100

U

mM, 10 µL) in a vial and pipetted in and out within a tip vigorously for 2 min to release the

N

binding species from the conjugates. The resultant solution was centrifuged at 26810g for 10

A

min. The supernatant containing the released target species (2 µL) was mixed with sinapinic acid

M

(20 mg mL−1, 2 µL) prepared in acetonitrile/0.1% TFA (2:1, v/v). The mixture (1 L) was

TE

MALDI-MS analysis.

D

quickly deposited on a MALDI plate. After solvent evaporation, the sample was ready for

Tryptic digestion of the conjugates of Au@cew NP-target species from crude extract of

EP

banana pulp

CC

To confirm the identity of the target species trapped by Au@cew NPs from the crude banana extract, trypsin digestion combined with MALDI-MS was conducted. The proteins on the

A

conjugates were released by addition of mannose (50 mM, 10 L). After vigorously pipetting the mixture for 2 min, the resultant solution was centrifuged at 26810g for 10 min. The released species in the supernatant (10 L) were heated at 100 °C in a water bath for 15 min to denature the released species. The sample vial should be capped to avoid solvent evaporation. This heating step was necessary because lectins are usually resistant to enzymatic digestion prior to 10

undertaking denaturing steps. The resultant heated solution (5 L) was conducted with trypsin digestion by adding with trypsin (5 L, 5 M) prepared in ammonium bicarbonate buffer (pH ~8, 75 mM) followed by subjecting the mixture into a water bath (10 mL). The water bath was

SC RI PT

placed into a microwave oven and heated (power, 900 W) for 1 min. The heating cycle was repeated five times, and the water in the water bath was renewed between heating cycles. The resultant digest (2 µL) was mixed with MALDI matrix, CHCA (20 mg mL−1, 2 µL), prepared in acetonitrile/0.1% TFA (2:1, v/v). The mixture was quickly deposited on a MALDI plate. After solvent evaporation, the sample was ready for MALDI-MS analysis

U

Preparation of Shiga-like toxins containing cell lysate

N

E. coli O157:H7 is a risk group 2 pathogen; hence, the culture experiments involved the

A

use of E. coli O157:H7 were carried out in a biosafety level 2 laboratory. E. coli O157:H7 that

M

can naturally express Shiga-like toxins (SLTs) was cultured in a TSBY broth (10 mL) for 14 h.

D

TSBY was prepared by mixing TSB (12 g) and Y (2 g) in deionized water (400 mL). The

TE

resultant bacterial suspension was centrifuged at 3750g for 10 min. The settled bacterial pellets were rinsed with deionized water (8 mL × 2) under centrifugation at 3750g for 10 min. The

EP

rinsed bacterial pellets were added with urea (8 M, 1 mL) to lyse bacterial cells by incubating the mixture at 37 °C for 2 h. The resultant cell lysate was centrifuged at 3750g for 5 min. The

CC

supernatant was filtered through a Pall acrodisc syringe filter (pore size, 0.2 µm). The filtrate

A

was further treated by loading on a filter membrane (cutoff mass, 3 kDa) under centrifugation at 4500 rpm for 20 min. The species that remained on the membrane were rinsed by deionized water (1 mL × 5) under centrifugation at 3750g for 20 min to eliminate salts. The resultant cell lysate containing SLTs was stored in a freezer (−20 °C) prior to use.

11

Results and Discussion We used one-pot reactions to generate Au@cew NPs by reacting chicken egg white proteins with aqueous tetrachloroaurate in a water bath at 60 °C for 6 h. Figure 1A shows the ultraviolet-

SC RI PT

visible (UV-Vis) absorption spectrum of the generated Au@cew NPs. A maximum absorption band appeared at ~520 nm, indicating that Au@cew NPs were successfully generated. The inset in Figure 1A shows the photograph of the generated Au@cew NPs, which have ruby red color. Figure 1B shows the TEM image of the generated Au@cew NPs. The generated Au@cew NPs

U

possesses spherical shapes with an average size of 9.1 ± 1.3 nm (Figure 1C).

N

Given that ovalbumin is the major component in the chicken egg white proteins [42], we

A

believed that the surface on the generated Au@cew NPs was dominated by ovalbumin. To

M

clarify this point, we further conducted trypsin digestion of the intact Au@cew NPs. Figure S1A shows the resultant MALDI-MS spectrum of the tryptic digest of Au@cew NPs. Except for the

D

peak at m/z 2163.05 that was derived from trypsin, the rest of the peaks at m/z 1345.75, 1555.71,

TE

1581.71, 1687.83, 1773.89, 1859.92, and 2281.18 were derived from ovalbumin. Table S1 lists the theoretical and observed m/z of these peptide peaks with their corresponding sequences.

EP

These results showed that the generated Au@cew NPs were mainly dominated by ovalbumin.

CC

The involvement of thiol groups in the binding between Au core and its immobilized proteins have been observed in protein-directed synthesis Au nanoprobes [33,50]. Ovalbumin contains six

A

cysteine residues [51]; hence, involvement of these cysteines in the generation of Au@cew NPs is likely. Thus, the binding between the Au core and surface proteins is characterized by SALDIMS, which is an effective tool that can be used to verify Au–S binding between gold and protective shell [43,52]. Figure S1B shows the SALDI mass spectrum of intact Au@cew NPs obtained in reflectron negative ion mode. Two series of ion peaks with the mass difference of 12

m/z 197, i.e., the atomic mass of Au, appeared in the mass spectrum. The ion peaks (marked with red circles) appearing at m/z 788.31, 985.37, 1182.48, 1379.60, 1576.79, and 1773.89 corresponding to Au4−, Au5 −, Au6 −, Au7 −, Au8 −, and Au9 −, respectively, were attributed to the

SC RI PT

Au cluster ions. The ion peaks (marked with stars) appeared at m/z 820.62, 1017.81, 1214.64, 1411.46, and 1608.54, corresponding to Au4S−, Au5S−, Au6S−, Au7S−, and Au8S−, respectively. These two ion groups have the mass difference of ~32 Da between the adjacent peaks, indicating that Au–S bonding was involved in the formation of Au@cew NPs. The results confirmed that S–Au binding was involved in the generation of Au@cew NPs.

U

We then used the generated Au@cew NPs as affinity probes to trap lectins including Con A,

N

BanLec, and ricin B. Ubiquitin (MW = 8565 Da), cytochrome c (MW= 12,361 Da), and

A

myoglobin (MW = 16,952 Da) were selected as non-lectin proteins. Figures 2A, 2B, and 2C

M

show the direct MALDI mass spectra of the mixtures containing the non-lectin proteins mixed individually with Con A (MW = 25,574 Da), BanLec (MW = 14,555 Da), and ricin B (MW=

D

~32,000 Da), respectively, prior to selectively trapped by Au@cew NPs. Apparently, all the

TE

mass spectra were dominated by ubiquitin (Ubi+), cytochrome c (Cyt c+), and myoglobin (Myo+).

EP

However, after using Au@cew NPs as the affinity probes to trap lectins from the same protein mixtures followed by released by glucose (Figure 2D), mannose (Figure 2E), and lactose

CC

(Figure 2F), lectin peaks started to appear in the mass spectra. The ion peak at m/z ~25,570 was derived from Con A (Figure 2D), whereas the ion peak at m/z ~14,550 was derived from BanLec

A

(Figure 2E). The ion peaks at m/z ~16,000 and ~32,000 corresponded to doubly and singly charged ricin B (Figure 2F). Au@cew NPs can be used to selectively trap target lectins from a protein mixture. One may suspect that the binding interactions between Au@cew NPs and these lectins resulted from electrostatic interactions. The isoelectric point (pI) of Con A and ricin B are

13

4.5–5 [53] and 4.8 [54], respectively. Under the enrichment condition (pH 7.5), these two lectins should carry net negative charges. The pI value for BanLec is 7.2–7.5 [55], so BanLec barely carries charges at pH 7.5. The zeta potential of Au@cew NPs was estimated to be −26.2 mV at

SC RI PT

pH 7.5 based on our measurement. Therefore, the involvement of electrostatic interactions in the binding between Au@cew NPs and these lectins is unlikely. On the other hand, the pIs of ubiquitin, cytochrome c, and myoglobin are 6.7 [56], 10-10.5 [57], and 6.8-7.2 [58], respectively. Cytochrome c may be trapped by Au@cew NPs through electrostatic interactions. However, no ions derived from cytochrome c were observed the resultant mass spectra. It was because

U

specific saccharides were used as the releasing agents, non-lectin proteins would not be released

N

by these releasing agents. By combining the use of the generated Au@cew NPs with the

A

saccharide-based releasing agents, the selectivity of our approach toward target lectin is desirable.

M

In addition, the experimental parameters including enrichment time, pH, and temperature were optimized by using Con A as the model sample (Figure S2). The results indicated that optimized

D

time, pH, and the temperature were ~1.5 h, ~6.5, ~25-30 oC, respectively. Nevertheless, to

TE

reduce the non-specific binding from non-target species that have lower isoelectric points than 7,

EP

the pH in the samples was adjusted to be slightly higher, i.e. 7.2-7.5, than the optimized value. When generating the as-prepared Au@cew NPs, non-denatured chicken egg white proteins

CC

were used as the starting materials. We were wondering if the proteins on the as-prepared Au@cew NPs had been denatured or not during the synthesis processes. Thus, we used circular

A

dichroism spectroscopy to examine the confirmation of proteins on the as-prepared Au NPs. However, owing to the strong scattering effect resulting from the Au NPs, no apparent spectrum was obtained (blue curve, Figure S3). Nevertheless, we also examined if Au@cew NPs generated from denatured chicken egg white proteins treated by dithiothreitol and iodoacetic acid

14

as the reducing agent and the alkylating agent, respectively, still possess the trapping capacity toward these model lectins including Con A and BanLec. Figure S4A shows the MALDI mass spectrum of the sample containing trace Con A (~32 nM) obtained before enrichment. Figures

SC RI PT

S4B and S4C show the resultant MALDI mass spectra obtained after using Au@cew prepared from denatured and non-denatured egg white proteins as affinity probes to concentrate target species from the sample containing trace amount of Con A. The peaks derived from Con A dominated these two mass spectra. That is, Au@cew NPs prepared from either non-denature or denatured chicken egg white proteins possess similar trapping capacity toward their target

U

species.

N

We further conducted the binding experiments of Au@cew NPs toward the sample

A

containing the mixture of Con A, ricin B, and BanLec. Figure 3A shows the direct MALDI mass

M

spectrum of this mixture before enrichment. No ion peaks appeared in the mass spectrum because the concentrations of the sample were too low. Figures 3B, 3C, and 3D show the

D

MALDI mass spectra of the same sample obtained after enriched by using Au@cew NPs as the

TE

affinity probes followed by releasing the target lectin from the conjugates using mannose,

EP

glucose, and β-lactose, respectively. The ion peaks derived from Con A and BanLec appeared in the mass spectra obtained after using mannose (Figure 3B) and glucose (Figure 3C) as the

CC

releasing agents. This result was reasonable because both Con A and BanLec contain binding moieties toward mannose and glucose. The ion peak derives from ricin B was only observed in

A

the mass spectrum when β-lactose was used as the releasing agent to elute target lectins from the conjugates of Au@cew NP-target species (Figure 3D). Our nanoprobes possess multiplex affinity toward these model lectins that contain different binding pockets with specific

15

saccharides. Furthermore, different saccharides can be used to selectively release target lectins from the conjugates of Au@cew NP-lectin. The B subunit of Shiga-like toxin-1 (SLT-1B) is a lectin and possesses the binding moieties

SC RI PT

against Galα(14)Gal [59–61]. Although chicken ovalbumin contains galactose-terminated glycans, the glycan ligand is Galβ(14)GlcNAc [47,62]. Thus, Au@cew NPs is unlikely to be used as affinity probes against SLT-1B. To further verify the selectivity of our nanoprobes, SLT1B produced from the cell lysate of E. coli O157:H7 spiked with ConA, BanLec, and ricin B was used as the sample. Figures 4A shows the MALDI mass spectrum of the cell lysate sample

U

containing SLT-1B, ConA, BanLec, and ricin B. The ion peak at m/z 9744 was derived from

N

SLT-1B (MW= 9743 Da)\. However, no peaks appeared at the high mass region (inset in Figure

A

4A). After enrichment, the peak at m/z 9744 standing for SLT-1B disappeared in the resultant

M

mass spectrum obtained using β-lactose as the releasing agent (Figure 4B). Nonetheless, the peak represents ricin B at m/z ~32,000 appeared in the high mass region in the mass spectrum

D

(Figure 4C). When the Au NP-target species conjugates were treated with mannose, the ion

TE

peaks at m/z ~12750, m/z ~25,570, and m/z ~14,560 corresponding to doubly charged Con A

EP

(Con A2+), singly charged Con A (Con A+), and BanLec, respectively, appeared in the resultant mass spectrum (Figure 4D). The results showed that Au@cew NPs were able to selectively trap

CC

their target lectins from a very complex cell lysate sample. Furthermore, the binding selectivity of this approach toward their target lectins is quite good. Although SLT-1B contains binding

A

pocket against Galα(1→4)Gal, it cannot be selectively trapped by our Au@cew NPs that only contains Galβ(14)GlcNAc ligands. We further used a serum sample spiked with Con A, BanLec, and ricin B as the model sample to further examine the selectivity of our affinity probes toward their target lectins. Figure

16

S5A shows the MALDI mass spectrum obtained before enrichment. Two apparent peaks appeared at m/z ~33281 and ~66451, corresponding to the doubly and singly charged human serum albumin (HSA), which is quite abundant in human sera. No other ion peaks were observed

SC RI PT

in the same mass spectrum. Figures S5B and S5C show the resultant MALDI mass spectra obtained after using Au@cew NPs as affinity probes to selectively enrich target species from the as-prepared serum sample followed by eluting the target species using mannose and β-lactose as the releasing agents, respectively. A peak derived from doubly charged Con A dominated the mass spectrum in Figure S5B, whereas the ion peak at m/z ~14,550, derived from BanLec, was

U

observed in the same mass spectrum. The ion peak appeared at m/z ~32000, corresponding to

N

singly charged ricin B, dominated the mass spectrum in Figure S5C. It is understandable since

A

both Con A and BanLec are mannose-binding lectins and ricin B is a β-lactose-binding lectin.

M

These results confirmed again that our approach has high selectivity toward their target lectins and can selectively trap their target species from a very complex serum sample.

D

In addition, the limit of detection (LOD) of this approach toward lectins including Con A,

TE

BanLec, and ricin B was also investigated. Figures S6A, S6B, and S6C show the resultant

EP

MALDI mass spectra of direct analysis of Con A (~3.9 nM), BanLec (~7.8 nM), and ricin B (~31.3 nM), respectively. Because the concentrations of these lectins in the samples were quite

CC

low, no ion peaks appeared in the mass spectra. However, after using Au@cew NPs as the affinity probes to trap these lectins from the samples followed by proper releasing steps, lectin

A

peaks started to appear in the mass spectra (Figures S6D, S6E, and S6F). The ion peak at m/z ~25,570 was derived from Con A (Figure S6D). The ion peak at m/z ~14,550 was derived from BanLec (Figure S6E), whereas the ion peak at m/z ~32,000 was contributed by ricin B (Figure

17

S6F). The LODs for these model lectins using this approach were in low nM level. The LOD is comparable with the existing lectin detection methods [29, 63–65] (Table S2). To demonstrate the feasibility of using the current approach for real-world samples,

SC RI PT

BanLec-rich banana pulp was used as the sample. A crude banana extract was prepared from over ripened banana pulp by using the procedures given in Experimental Section. Figure 5A shows the direct MALDI mass spectrum of crude banana extract. Many ion peaks derived from banana appeared in the mass spectrum. Figure 5B shows the MALDI mass spectrum obtained after using Au@cew NPs as affinity probes to selectively enrich target species from the same

U

banana extract followed by using mannose as the releasing agent. Only the ion peaks at m/z

N

~14,555 and m/z ~29,099, presumably corresponding to monomer (BanLec+) and dimer ions

A

(BamLec2+) of BanLec, respectively, were observed in the resultant mass spectrum. To further

M

confirm the identity of the species enriched and released from the Au@cew NP-target species conjugates, the tryptic digestion of the released species was further conducted. Figure S7 shows

D

the MALDI mass spectrum of the tryptic digest of the species eluted from Au@cew NP-target

TE

species. The peptide peaks at m/z 1869.91, 2023.49, and 2569.79 corresponded to the peptide

EP

sequences derived from BanLec, whereas the ion peak at m/z at 2273.16 was derived from trypsin. Table S3 shows a list of the theoretical and observed m/z and their corresponding

CC

peptide sequences. These results indicate that our nanoprobes possess good selectivity toward BanLec. The developed method should be useful for selective enrichment of target lectins from

A

complex samples.

18

Conclusions We have successfully demonstrated that Au@cew NPs are effective affinity probes that possess the capability to trap multiple lectins for which contain binding pockets including

SC RI PT

mannose, glucose, and β-galactose. The selectivity of this approach towards different target lectins is boosted by using suitable saccharides as the releasing agents to elute target lectins from the Au@cew NP-target species conjugates. As demonstrated in this study, Au@cew NPs are mainly dominated by a glycoprotein, i.e., ovalbumin, immobilized Au NPs. The glycan ligands on the structure of ovalbumin reserve the binding affinity of the generated Au@cew NPs toward

U

different lectins that possess the binding moieties including mannose, glucose, and β-galactose.

N

One type of nanoprobe can be used to target multiple lectins is advantageous finding.

A

Furthermore, chicken eggs are inexpensive and easily available. In addition, the visible color of

M

the as-prepared Au NPs is helpful when handling the samples. Although Ag NPs possessing visible color may also be synthesized using the similar approach as proposed in this study. The

D

relatively high toxicity of Ag NPs is a concern when using such affinity probes for conducting

TE

the analysis. The main advantages of this approach include simplicity, low cost, high sensitivity,

EP

and high selectivity. To the best of our knowledge, this is the first report that demonstrates the feasibility of using Au@cew NPs as affinity probes to probe multiple lectins. Although we only

CC

used few lectins as our model samples, we believe that the generated Au@cew NPs should also possess the capability to probe other lectins that contain the binding pockets against mannose,

A

glucose, and β-galactose. Thus, the potential applications by using Au@cew NPs as sensing and enrichment probes for their target lectins can be further explored.

19

Acknowledgements We thank the Ministry of Science and Technology of Taiwan (MOST 105-2113-M-009022-MY3 and MOST 102-2627-M-009-002) for financial support of this work. KS thanks NCTU for providing him the NCTU International Student Scholarship. We also thank Prof. S.-Y

A

CC

EP

TE

D

M

A

N

U

SC RI PT

Lin for her kind help in obtaining CD spectra.

20

References [1] I. J. Goldstein, C. E. Hayes, The lectins: carbohydrate-binding proteins of plants and animals. Adv. Carbohydr. Chem. Biochem. 35 (1978) 127−340.

SC RI PT

[2] W. J. Peumans, E. J. van Damme, The role of lectins in plant defence. Histochem. J. 27, (1995) 253–271.

[3] A. Varki, R. D. Cummings, J. D. Esko, H. H. Freeze, P. Stanley, C. R. Bertozzi, G. W. Hart, M. E. Etzler, 2nd edition. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 2009.

U

[4] I. J. Goldstein, C. E. Hollerman, E. E. Smith, Protein-carbohydrate interaction. inhibition

N

studies on the interaction of concanavalin a with polysaccharides. Biochemistry 4 (1965)

A

876−883.

M

[5] H. Lis, N. Sharon, Lectins: Carbohydrate-Specific Proteins That Mediate Cellular

D

Recognition. Chem. Rev. 98 (1998) 637−674.

TE

[6] E. M. A. Pereira, M. R. Sierakowski, T. A. Jo, R. A. Moreira, A. C. O. Monteiro-Moreira, R. F. O. França, B. A. L. Fonseca, D. F. S. Petri, Lectins and/or xyloglucans/alginate layers as

EP

supports for immobilization of dengue virus particles. Colloids Surf. B Biointerfaces 66 (2008) 45−52.

CC

[7] C. A. S. Andrade, M. D. L. Oliveira, C. P. de Melo, L.C. B. B. Coelho, M. T. S. Correia, M. L. Nogueira, P. R. Singh, X. Zeng, Diagnosis of dengue infection using a modified gold

A

electrode with hybrid organic-inorganic nanocomposite and Bauhinia monandra lectin. J. Colloid Interface Sci. 362 (2011) 517−523.

21

[8] Z. Shi, J. Chen, C. –Y. Li, N. An, Z. –J. Wang, S. –L. Yang, K. -F. Huang and J. –K. Bao, Antitumor effects of concanavalin A and Sophora flavescens lectin in vitro and in vivo. Acta Pharmaco. Sin. 35 (2014) 248−256.

SC RI PT

[9] S. Olsnes, J. V. Kozlov, Ricin. Toxicon. 39 (2001) 1723−1728. [10] Y. Endo, K. Tsurugi, RNA N-glycosidase activity of ricin A-chain. Mechanism of action of the toxic lectin ricin on eukaryotic ribosomes. J. Biol. Chem. 262 (1987) 8128−8130.

[11] Y. Endo, K. Mitsui, M. Motizuki and K. Tsurugi, The mechanism of action of ricin and related toxic lectins on eukaryotic ribosomes. The site and the characteristics of the

U

modification in 28 S ribosomal RNA caused by the toxins. J. Biol. Chem. 262 (1987)

N

5908−5912.

A

[12] S. Olsnes, E. Saltvedt, A. Pihl, Isolation and comparison of galactose-binding lectins from

M

Abrus precatorius and Ricinus communis. J. Biol. Chem. 249 (1974) 803−810. [13] S. Olsnes, The history of ricin, abrin and related toxins. Toxicon. 44 (2004) 361−370.

D

[14] G. L. Nicolson, J. Blaustein, M. E. Etzler, Characterization of two plant lectins from

TE

Ricinus communis and their quantitative interaction with a murine lymphoma. Biochemistry

EP

13 (1974) 196–204.

[15] K. R. Challoner, M. M. McCarron, Castor bean intoxication. Ann. Emerg. Med. 19 (1990)

CC

1177−1183.

[16] K. Sandvig, B. van Deurs, Entry of ricin and Shiga toxin into cells: molecular mechanisms

A

and medical perspectives. EMBO J. 19 (2000) 5943−5950.

[17] R. A. Zilinskas, Iraq's biological weapons. The past as future? JAMA. 278 (1997) 418–424. [18] L. D. Rotz, A. A. Khan, S. R. Lillibridge, S. M. Ostroff, J. M. Hughes, Public health assessment of potential biological terrorism agents. Emerg. Infect. Dis. 8 (2002) 225–230.

22

[19] H. Mo, H. C. Winter, E. J. Van Damme, W. J. Peumans, A. Misaki, I. J. Goldstein, Carbohydrate binding properties of banana (Musa acuminata) lectin I. Novel recognition of internal alpha1,3-linked glucosyl residues. Eur. J. Biochem. 268 (2001) 2609–2615.

SC RI PT

[20] Swanson, M. D.; Winter, H. C.; Goldstein, I. J.; Markovitz, D. M. A lectin isolated from bananas is a potent inhibitor of HIV replication. J. Biol. Chem. 285 (2010) 8646–8655.

[21] M. D. Swanson, H. C. Winter, I. J. Goldstein and D. M. Markovitz, Engineering a therapeutic lectin by uncoupling mitogenicity from antiviral activity. Cell 163 (2015) 746– 758.

U

[22] D. C. Hone, A. H. Haines, D. A. Russell, Rapid, Quantitative Colorimetric Detection of a

N

Lectin Using Mannose-Stabilized Gold Nanoparticles. Langmuir 19 (2003) 7141–7144.

A

[23] C. S. Tsai, T. B. Yu and C. T. Chen, Gold nanoparticle-based competitive colorimetric assay

M

for detection of protein-protein interactions. Chem. Commun. 34 (2005) 4273–4275. [24] Y. K. Lyu, K. R. Lim, B. Y. Lee, K. S. Kim, W. Y. Lee, Microgravimetric lectin biosensor

D

based on signal amplification using carbohydrate-stabilized gold nanoparticles. Chem.

TE

Commun. 39 (2008) 4771–4773.

EP

[25] L. Dykman, N. Khlebtsov, Gold nanoparticles in biomedical applications: recent advances and perspectives. Chem. Soc. Rev. 41 (2012) 2256–2282.

CC

[26] C. C. Lin, Y. C. Yeh, C. Y. Yang, G. F. Chen, Y. C. Chen, Y. C. Wu, C. C. Chen, Quantitative analysis of multivalent interactions of carbohydrate-encapsulated gold

A

nanoparticles with concanavalin A. Chem. Commun. 23 (2003) 2920–2921.

[27] K. R. Lim, K. S. Ahn, W. Y. Lee, Detection of concanavalin A based on attenuated fluorescence resonance energy transfer between quantum dots and mannose-stabilized gold nanoparticles. Anal. Methods 5 (2013) 64–67.

23

[28] X. Jiang, A. Housni, G. Gody, P. Boullanger, M. T. Charreyre, T. Delair, R. Narain, Synthesis of biotinylated α-D-mannoside or N-acetyl β-D-glucosaminoside decorated gold nanoparticles: study of their biomolecular recognition with Con A and WGA Lectins.

SC RI PT

Bioconjugate Chem. 21 (2010) 521–530. [29] S. Nakamura-Tsuruta, Y. Kishimoto, T. Nishimura, Y. Suda, One-step purification of lectins from banana pulp using sugar-immobilized gold nano-particles. J. Biochem. 143 (2008) 833–839.

[30] C. L. Schofield, B. Mukhopadhyay, S. M. Hardy, M. B. McDonnell, R. A. Field, D. A.

U

Russell, D. A. Colorimetric detection of Ricinus communis Agglutinin 120 using optimally

N

presented carbohydrate-stabilised gold nanoparticles. Analyst 133 (2008) 626–634.

A

[31] H. Q. Yin, M. X. Jia, S. Yang, S. Q. Wang, J. G. Zhang, A nanoparticle-based bio-barcode

M

assay for ultrasensitive detection of ricin toxin. Toxicon. 39 (2012) 12–16. [32] T. Nagatsuka, H. Uzawa, K. Sato, S. Kondo, M. Izumi, K. Yokoyama and Y. Nishida,

D

Localized surface plasmon resonance detection of biological toxins using cell surface

TE

oligosaccharides on glyco chips. ACS Appl. Mater. Interfaces 5 (2013) 4173–4180.

EP

[33] J. Xie, Y. Zheng, J. Y. Ying, Protein-directed synthesis of highly fluorescent gold nanoclusters. J. Am. Chem. Soc. 131 (2009) 888−889.

CC

[34] H. Wei, Z. Wang, L. Yang, S. Tian, C. Hou, Y. Lu, Lysozyme-stabilized gold fluorescent cluster: Synthesis and application as Hg(2+) sensor. Analyst 135 (2010) 1406−1410.

A

[35] P. –H. Chan, Y. –C. Chen, Human serum albumin stabilized gold nanoclusters as selective luminescent probes for Staphylococcus aureus and methicillin-resistant Staphylococcus aureus. Anal. Chem. 84 (2012) 8952−8956. [36] H. Kawasaki, K. Hamaguchi, I. Osaka, R. Arakawa, ph‐Dependent synthesis of pepsin‐

24

mediated gold nanoclusters with blue green and red fluorescent emission Adv. Funct. Mater. 21 (2011) 3508−3515. [37] K. Selvaprakash, Y. –C. Chen, Using protein-encapsulated gold nanoclusters as

SC RI PT

photoluminescent sensing probes for biomolecules. Biosens. Bioelectron. 61 (2014) 88–94. [38] D. Joseph, K. E. Geckeler, Synthesis of highly fluorescent gold nanoclusters using egg white proteins. Colloids Surf. B Biointerfaces 115 (2014) 46−50.

[39] T. Yang, Z. Li, L. Wang, C. Guo, Y. Sun, Synthesis, characterization, and self-assembly of protein lysozyme monolayer-stabilized gold

nanoparticles.

U

10533−10538.

Langmuir 23 (2007)

N

[40] A. Housni, M. Ahmed, S. Liu, R. Narain, Monodisperse protein stabilized gold

A

nanoparticles via a simple photochemical process J. Phys. Chem. C 112 (2008) 12282–

M

12290.

[41] Y. Liu, L. Liu, M. Yuan, R. Guo, Preparation and characterization of casein-stabilized gold

D

nanoparticles for catalytic applications. Colloids and Surfaces A: Physicochem. Eng.

TE

Aspects 417 (2013) 18−25.

EP

[42] J. Yoshimoto, N. Tanaka, M. Inada, R. Arakawa, H. Kawasaki, Microwave-assisted synthesis of near-infrared-luminescent ovalbumin-protected gold nanoparticles as a

CC

luminescent glucose sensor. Chem. Lett., 43 (2014) 793−795. [43] W. -Y. Chen, J. -Y, Lin, W.-J. Chen, L. Lo, E. W.-G. Diau, Y.-C. Chen, Functional gold

A

nanoclusters as antimicrobial agents for antibiotic-resistant bacteria. Nanomedicine 5 (2010) 755−764.

[44] Y. Mine, Recent advances in the understanding of egg white protein functionality. Trends Food Sci. Technol. 7 (1995) 225−232.

25

[45] K. Selvaprakash, Y. -C. Chen, Detection of ricin by using gold nanoclusters functionalized with chicken egg white proteins as sensing probes. Biosens. Bioelectron. 92 (2017) 410– 416.

SC RI PT

[46] L. M. Hennigar, R. A. Hennigar, B. A. Schulte, Histochemical specificity of beta-galactose binding lectins from Arachis hypogaea (peanut) and Ricinus communis (castor bean). Stain. Technol. 62 (1987) 317−325.

[47] D. J. Harvey, D. R. Wing, B. Kuster, I. B. H. Wilson, Composition of N-linked carbohydrates from ovalbumin and co-purified glycoproteins. J. Am. Soc. Mass. Spectrom.

U

11 (2000) 564–571.

N

[48] W.-Y. Chen, Y.-C. Chen, Acceleration of microwave-assisted enzymatic digestion reactions

A

by magnetite beads. Anal. Chem. 79 (2007) 2394–2401.

M

[49] P.-H. Chan, Y. -C. Chen, Human serum albumin stabilized gold nanoclusters as selective luminescent probes for Staphylococcus aureus and methicillin-resistant Staphylococcus

D

aureus. Anal. Chem. 84 (2012) 8952–8956.

TE

[50] X. Le Guével, N. Daum, M. Schneider, Synthesis and characterization of human transferrin-

EP

stabilized gold nanoclusters. Nanotechnology 22 (2011) 275103. [51] E. O. Thompson, W. K. Fisher, Amino acid sequences containing half-cystine residues in

CC

ovalbumin. Aust. J. Biol. Sci. 31 (1978) 433−442. [52] H. -Z. Lai, S.-G. Wang, C.-Y. Wu, Y. -C. Chen, Detection of Staphylococcus aureus by

A

functional gold nanoparticle-based affinity surface-assisted laser desorption/ionization mass spectrometry. Anal. Chem. 87 (2015) 2114−2120.

[53] G. Entlicher, J. V. Koštíř, J. Kocourek, Studies on phytohemagglutinins. 8. Isoelectric point and multiplicity of purified concanavalin A. Biochim. Biophys. Acta. 236 (1971) 795−797.

26

[54] S. Olsnes, Ricin and ricinus agglutinin, toxic lectins from castor bean. Methods Enzym. 50 (1978) 330–335. [55] V. L. Koshte, W. van Dijk, M. E. van der Stelt, R. C. Aalberse, Isolation and

SC RI PT

characterization of BanLec-I, a mannoside-binding lectin from Musa paradisiac (banana). Biochem. J. 272 (1990) 721–726.

[56] K. D. Wilkinson, M. K. Urban, A. L. Haas, Ubiquitin is the ATP-dependent proteolysis factor I of rabbit reticulocytes. J. Biol. Chem. 255 (1980) 7529–7532.

[57] B. F. Van Gelder, E. C. Slater, The extinction coefficient of cytochrome c. Biochim.

U

Biophys. Acta 58 (1962) 593–595.

N

[58] T. Sakai, T. Sakamoto, Purification and some properties of a protopectin-solubilizing

A

enzyme that has potent activity on sugar beet protopectin. Agric. Biol. Chem. 54 (1990)

M

879–889.

[59] M. T. Tarrago-Trani, B. Storrie, A method for the purification of Shiga-like toxin 1 subunit

TE

170–176.

D

B using a commercially available galabiose–agarose resin. Protein Expr. Purif. 38 (2004)

EP

[60] N. Takahashi, K. H. Khoo, N. Suzuki, J. R. Johnson, Y. C. Lee, N-glycan structures from the major glycoproteins of pigeon egg white: predominance of terminal Galalpha(1)Gal. J.

CC

Biol. Chem. 276 (2001) 23230–23239. [61] F.-Y. Kuo, B.-Y. Chang, C.-Y. Wu, K.-K. T. Mong, Y.-C. Chen, Magnetic nanoparticle-

A

based platform for characterization of shiga-like Toxin 1 from complex samples. Anal. Chem. 87 (2015) 10513–10520.

[62] K. Yamashita, Y. Tachibana, A. Kobata, The structures of the galactose-containing sugar chains of ovalbumin. J. Biol. Chem. 253 (1978) 3862–3869.

27

[63] S. Watanabe, K. Yoshida, K. Shinkawa, D. Kumagawa, H. Seguchi, Thioglucose-stabilized gold nanoparticles as a novel platform for colorimetric bioassay based on nanoparticle aggregation. Colloids Surf. B Biointerfaces, 81 (2010) 570–577.

SC RI PT

[64] P. Babu, S. Sinha, A. Surolia, Sugar-quantum dot conjugates for a selective and sensitive detection of lectins. Bioconjug. Chem. 18 (2007) 146–151.

[65] C. L. Schofield, B. Mukhopadhyay, S. M. Hardy, M. B. McDonnell, R. A. Field, D. A. Russell, Colorimetric detection of Ricinus communis Agglutinin 120 using optimally

A

CC

EP

TE

D

M

A

N

U

presented carbohydrate-stabilized gold nanoparticles. Analyst, 133 (2008) 626–634.

28

Figure Legends Figure 1. (A) UV-Vis absorption spectrum of the generated Au@cew NPs. (B) TEM image of

SC RI PT

the generated Au@cew NPs (scale bar: 100 nm). (C) Particle size distribution of the generated Au@cew NPs obtained from the TEM image in panel B. ImageJ was used for estimating the size distribution.

Figure 2. MALDI mass spectra of the samples containing the mixtures of (A) Con A (0.5 µM),

U

ubiquitin (1 µM), cytochrome c (1 µM), and myoglobin (1 µM), (B) BanLec (0.5 µM), ubiquitin

N

(1 µM), cytochrome c (1 µM), and myoglobin (1 µM), (C) ricin B (0.5 µM), ubiquitin (1 µM),

A

cytochrome c (1 µM), and myoglobin (1 µM) obtained before enrichment. The MALDI mass

M

spectra shown in Panels D, E, and F were obtained after using Au@cew NPs (2.6 mg mL-1, 20 µL) as affinity probes to selectively enrich target species from the same samples (0.5 mL) as

D

used to obtain Panels A, B, and C, respectively, followed by treating with glucose (50 mM, 10

TE

µL), mannose (50 mM, 10 µL), and -lactose (50 mM, 10 µL), respectively. CHCA (20 mg mL-1)

EP

prepared in acetonitrile/0.1% TFA (2:1, v/v) was used as the MALDI matrix to obtain Panels A, B, D and E. Sinapinic acid (20 mg mL-1) prepared in acetonitrile/0.1% TFA (2:1, v/v) was used

CC

as the MALDI matrix in Panels C and F. Figure 3. MALDI mass spectra of the samples containing the mixture of Con A (15.6 nM),

A

BanLec (15.6 nM), and ricin B (125 nM) obtained (A) before and after enrichment by using Au@cew NPs (2.6 mg mL-1, 20 µL) as affinity probes followed by releasing lectins using (B) mannose (50 mM, 10 µL), (C) glucose (50 mM, 10 µL), and (D) -lactose (50 mM, 10 µL) as

29

the releasing agents. Sinapinic acid prepared in acetonitrile/0.1% TFA (2:1, v/v) was used as the MALDI matrix. Figure 4. (A) MALDI mass spectrum of the cell lysate (0.5 mL) derived from E. coli O157:H7

SC RI PT

containing SLT-1 mixed with Con A (15.6 nM), BanLec (15.6 nM), and ricin B (0.125 μM). MALDI mass spectra obtained after using Au@cew NPs (2.6 mg mL-1, 20 µL) as affinity probes to enrich target species from the same sample as used in Panel A followed by releasing steps using -lactose (50 mM, 10 µL) as the releasing agent obtained in (B) low mass region (m/z 4,000-10,000) and (C) high mass region (m/z 10,000-40,000). (D) MALDI mass spectrum

U

obtained by further treating the same sample using mannose (50 mM, 10 µL) as the releasing

N

agents. Sinapinic acid prepared in acetonitrile/0.1% TFA (2:1, v/v) was used as the MALDI

M

A

matrix.

Figure 5. (A) MALDI mass spectrum of the crude banana pulp extract. (B) MALDI mass

D

spectrum of the sample obtained after shaking Au@cew NPs (2.6 mg/mL, 20 μL) with the same

TE

crude banana extract (0.1 mL) used to obtain Panel (A) for 2 h followed by releasing the target species from the conjugates of the Au@cew NP-target species using mannose (100 mM, 10 μL)

A

CC

EP

as the releasing agent.

30

D

TE

EP

CC

A Figure 1

31

SC RI PT

U

N

A

M

D

TE

EP

CC

A

Figure 2

32

SC RI PT

U

N

A

M

D

TE

EP

CC

A Figure 3

33

SC RI PT

U

N

A

M

D

TE

EP

CC

A Figure 4

34

SC RI PT

U

N

A

M

D

TE

EP

CC

A Figure 5

35

SC RI PT

U

N

A

M