Branched polyethyleneimine-assisted boronic acid-functionalized silica nanoparticles for the selective enrichment of trace glycoproteins

Author’s Accepted Manuscript Branched polyethyleneimine-assisted boronic acidfunctionalized silica nanoparticles for the selective enrichment of trace glycoproteins Hongjun Xia, Shuangshou Wang, Lin Wang www.elsevier.com/locate/talanta

PII: DOI: Reference:

S0039-9140(18)30130-9 https://doi.org/10.1016/j.talanta.2018.02.021 TAL18340

To appear in: Talanta Received date: 12 September 2017 Revised date: 29 January 2018 Accepted date: 6 February 2018 Cite this article as: Hongjun Xia, Shuangshou Wang and Lin Wang, Branched polyethyleneimine-assisted boronic acid-functionalized silica nanoparticles for the selective enrichment of trace glycoproteins, Talanta, https://doi.org/10.1016/j.talanta.2018.02.021 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 galley proof before it is published in its final citable 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.

Branched polyethyleneimine-assisted boronic acid-functionalized silica nanoparticles for the selective enrichment of trace glycoproteins

Hongjun Xiaa,b*, Shuangshou Wangc, Lin Wanga,b a

The Key Laboratory of Rare Earth Functional Materials and Applications, Zhoukou

Normal University, Zhoukou 466001, Henan, P. R. China. b

International Joint Research Laboratory for Biomedical Nanomaterials of Henan,

Zhoukou Normal University, Zhoukou 466001, P. R. China. c

Anhui university of technology, School of chemistry and chemical engineering.

* Corresponding author. Tel.Fax/:+86-394-8178518. [email protected] Abstract Boronate affinity materials have attracted more and more attention in extraction, separation and enrichment of glycoproteins due to the important roles that glycoproteins take on in recent years. However, conventional boronate affinity materials suffer from low binding affinity mainly because of the use of single boronic acids. This makes the extraction of glycoproteins of trace concentration become rather difficult or impossible. Here we present a novel boronate avidity material, polyethyleneimine (PEI)-assisted boronic acid-functionalized silica nanoparticles (SNPs). Branched PEI was applied as a scaffold to amplify the number of boronic acid moieties. While 3-carboxybenzoboroxole, exhibiting high affinity and excellent water solubility toward glycoproteins, was used as an affinity ligand. Due to the PEI-assisted synergistic multivalent binding, the boronate avidity SNPs exhibited strong binding strength toward glycoproteins with dissociation constants of 10−7 M, which was the highest among reported boronic acid-functionalized materials that can

be applied for glycoproteomic analysis. Such a high avidity enabled the selective extraction of trace glycoproteins as low as 0.4 pg/mL. This feature greatly favored the selective enrichment of trace glycoproteins from real samples. Meanwhile, the boronate avidity SNPs was tolerant of the interference of abundant sugars. In addition, the PEI-assisted boronate avidity SNPs exhibited high binding capacity and low binding pH. The feasibility for practical applications was demonstrated with the selective enrichment of trace glycoproteins in human saliva. Keywords：Glycoproteins; Boronic acid; Dissociation constants; Silica nanoparticles; Branched polyethyleneimine

1. Introduction Glycoproteins, which occupy more than 50% of the total proteins in body, play extremely important roles in a variety of biological processes, such as molecular recognition, inter and intracellular signaling, immune response and so on [1]. In addition, the occurrence of many diseases is associated with the glycosylation state of related proteins, and thereby many glycoproteins have been used as disease biomarkers and therapeutic targets [2,3]. However, many glycoproteins with importance for research and clinical diagnosis have low ionization efficiency and exist in very low abundance in real samples. Besides, high-abundance interfering components usually co-exist in sample matrix with glycoproteins, seriously suppressing the mass spectrometric (MS) signal of glycoproteins with low abundance. Therefore, selective enrichment of glycoproteins/glycopeptides is essential prior to

MS analysis. To date, the tools that can be used for selectively enriching glycoproteins mainly include lectins [4–6], antibodies [7,8], hydrazide [9–11] and boronate affinity materials [12–37]. However, the former three are associated with apparent disadvantages. A certain lectin can bind only a narrow range of glycoproteins and therefore multiple lectins have to be used for abundant glycoproteins analysis. Antibodies suff er from some apparent disadvantages, such as hard to prepare and poor storage stability. Although the hydrazine chemistry-based method is effective for the enrichment of glycoproteins from complex samples, it involves multiple chemical derivatization steps. Boronate affinity materials can provide several significant advantages such as low cost, class selectivity, easy-to-manipulate capture/release (through pH switch) and good compatibility with mass spectrometry. However, conventional boronate affinity materials cannot provide high affinity mainly because the binding strength of single boronic acids toward cis-diol-containing compounds is relatively weak, with dissociation constants (Kd) between boronic acids and sugars or glycoproteins ranging from 10−1 -10−3 M. Thus, it is rather difficult or impossible to capture glycoproteins of very low concentration using conventional boronate affinity materials. Until now, there were several types of materials such as boronate affinity molecularly imprinted polymers (MIPs) [38-42], boronic acid-capped Mn-doped ZnS quantum dots [37] and boronate avidity materials [43] to address the above issues. In these means, synergistic effect of simultaneous multiple binding used in the

preparation of boronate avidity materials has been developed as an effective strategy for enhancing the binding strength of boronic acid-functionalized materials [43]. Through amplification of the number of boronic acid moieties using highly branched poly(amidoamine) (PAMAM) dendrimers, the boronate avidity materials exhibited significantly enhanced binding affinity toward glycoproteins with Kd values within 10−5–10−6 M. However, the highly branched PAMAM dendrimers suffer from apparent disadvantages, such as high cost and strong rigidity. Especially, the rigid structure is disadvantageous for effective enrichment and quick equilibration. The above disadvantages lead to much less binding strength than that provided by antibody. It is necessary to develop novel boronic acid-functionalized materials with higher affinity and specificity. As compared with PAMAM, the branched polyethyleneimine (PEI) can be used as a better scaffold to amplify the number of boronic acid moieties due to its flexible chains, hydrophilic properties, easy post-modification, plentiful amino groups and low cost. Nanoparticles have attracted considerable interest in separation, catalysis, and drug delivery. In nanoparticles, silica nanoparticles (SNPs) are the most widely used due to their good biocompatibility, low toxicity and easy preparation. Recently, SNPs have been used as supporting materials to prepare boronic acid-functionalized particles [42,44], which were well applied to quickly separate and enrich glycoproteins for proteomics. Thus, it is essential to develop polyethyleneimine (PEI)-assisted boronic acid -functionalized SNPs. In the study, we present a branched polyethyleneimine (PEI)-assisted boronic acid

-functionalized SNPs with high avidity for the selective enrichment of trace glycoproteins. The 3-carboxybenzoboroxole, which exhibited high affinity and hydrophilicity toward cis-diols at neutral or medium acidic pH condition [24], was used as an affinity ligand. Fig. 1 depicts synergistic multiple binding between a glycoprotein and PEI-amplified boronic acid-functionalized SNPs. Due to synergistic multiple binding of the PEI-amplified boronic acid sites, the binding strength of the obtained boronate avidity SNPs was significantly improved. The boronate avidity SNPs exhibited Kd values of 10−7 M toward glycoproteins under investigation, which was higher than already reported boronic acid-functionalized materials except boronic acid-based MIPs and boronic acid-capped Mn-doped ZnS quantum dots [37,38]. In addition, the boronate avidity SNPs exhibited lowered binding pH toward glycoproteins. These features were directly applied to the selective enrichment of glycoproteins of low concentration from real samples. As compared with other boronic acid–functionalized materials, the PEI-assisted boronate avidity SNPs exhibited the best performance.

2. Experimental 2.1. Materials Branched polyethyleneimine (PEI) (Mw = 600, 1800, 10,000 and 70,000), ribonuclease A (RNase A), ribonuclease B (RNase B), transferring (TRF), horseradish peroxidase (HRP), albumin from bovine serum (BSA), siapinic acid (SA), 3-carboxybenzoboroxole and protease cocktail inhibitor were purchased from

Sigma-Aldrich (St. Louis, MO, USA). Sodium cyanoborohydride, anhydrous methanol, galactose, xylose and fucose were from J&K scientific (Shanghai, China). Glucose, tetramethoxysilane (TMOS), ammonium hydroxide, ethylenediamine, and 3,3,5′,5′-tetramethylbenzidine

dihydrochloride

(3-aminopropyl)triethoxysilane

(TMB), (APTES),

1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and N-hydroxysuccinimide (NHS) were purchased from Alfa Aesar (Tianjin, China). All used reagents were of analytical grade. All commercially available reagents were used without further purification. Water used in all the experiments was purified by a Milli-Q Advantage A10 ultrapure purification system (Millipore, Milford, USA).

2.2. Apparatus and methods Transmission electron microscopy (TEM) was performed on a JEM-1010 system (JEOL, Tokyo, Japan). The adsorption isotherm measurements were performed with a U-4100 spectrophotometer (Hitachi, Japan). MALDI-TOF MS analyses were performed on a MALDI TOF/TOF Analyzer (Shimadzu/Kratos, Manches-ter, UK) with a pulsed nitrogen laser operated at 337 nm. The laser energy was adjusted to above the threshold to obtain good resolution and signal-to-noise ratio(S/N). Spectra were obtained in the positive linear ion mode. In all, 500 laser shots per spot were accumulated for each spectrum. The accelerating voltage was 20 kV. 10 mg/mL SA in 50% ACN containing 0.1% (v/v) TFA was used as the matrix. Equivalent amounts (1 L) of the sample and matrix were dropped onto the MALDI plate for mass

spectrometric analysis. Data were processed using Data Explorer Software Version 3.7 (AB Sciex).

2.3. Preparation of PEIPBA-SNPs The procedure for the synthesis of PEIPBA-SNPs was shown in Fig.S1A in the Supporting Information, which included four steps: (1) Synthesis of silica nanoparticles (SNPs), (2) functionalization with 3-aminopropyltriethoxysilane (APTES),

(3)

preparation

of

PEI modified

SNPs

(PEI-SNPs),

and

(4)

functionalization with boronic acid 3-carboxybenzoboroxole (PEIPBA-SNPs). The SNPs were synthesized based on a previously reported method [42] with minor modification. Briefly, 6 mL of TEOS was gradually added to the mixture of 100 mL of ethanol, 4 mL of deionized water and 3.2 mL of aqueous solution of 25-28% ammonium with vigorous stirring at 30 ℃ and the reaction was continued for 24 h. The resulting SNPs were rinsed with water and ethanol for 3 times each, and then dried at 60 °C in a vacuum overnight. Then, 1.0 mL of APTES was added dropwise to 20 mL anhydrous methanol containing the above obtained SNPs and then the coating of silica NPs with amino groups was obtained by stirring the mixture for 24 h. The resultant APTES -coated silica NPs. were purified by three cycles of centrifugation, decantation,

and

resuspension in

ethanol by ultrasonic.

The

as-prepared

APTES-coated silica NPs were dried at room temperature under vacuum for further use. After 200 mg APTES-coated silica NPs were added to 30 mL anhydrous methanol containing 5% glutaraldehyde, the mixture was stirred for 12 h at room

temperature. The resulting glutaraldehyde-activated SNPs were washed three times with anhydrous methanol and then dispersed in 30 mL anhydrous methanol containing 0.6 g PEI by ultrasound. The mixture was stirred for 12 h at room temperature. 60 mg/mL sodium cyanoborohydride was added (200 mg every 6 hours) during the course of reaction for 12 h. The PEI-SNPs were collected by centrifuge and washed with water and ethanol for 3 times each, and then dried at 60 °C overnight. 0.1 mL of freshly prepared EDC in 100 mM PBS (100 mM, pH 7.4) and 0.2 mL of NHS in 10 mM PBS (10 mM, pH 7.4) were mixed with 20 mL of PEI-SNPs suspension. After being shaken at 25 ℃ for 15 min, 4 mL of 3-carboxybenzoboroxole in 100 mM PBS (20 mg/mL, pH 7.4) was added, and the mixture was kept shaking at 25 ℃ for 2 h. Unbound 3-carboxybenzoboroxole were removed by centrifugation at 8000 rpm for 10

min

twice

to

obtain

the

3-carboxybenzoboroxole-modified

PEI-SNPs

(PEIPBA-SNPs). The resultant SNPs were collected and washed with water and ethanol and dried at 60 °C. The obtained 3-carboxybenzoboroxole-SNPs (PEIPBA-SNPs) were stored for further use. The procedure for the synthesis of PBA-SNPs was consisted of two steps as depicted in Fig. S1B: (1) Synthesis of APTES-SNPs, which is the same as in Fig. S1A. (2) Functionalization of the APTES-SNPs with 3-carboxybenzoboroxole.

2.4. Measurement of adsorption isotherm and Scatchard analysis The measurement of adsorption isotherm was performed according to the following process. Briefly, the PEIPBA-SNPs and PBA-SNPs (3 mg) in 500-μL

plastic microcentrifugal tubes were first equilibrated with 200 μL loading buffer (100 mM phosphate buffer, pH 7.4). After the SNPs were collected at the tube wall by centrifuge, 200 μL solutions of different concentrations of glycoprotein or adenosine were added to the above PEIPBA-SNPs and PBA-SNPs. The tubes were shaken on a rotator for 2 h at room temperature. Thereafter, PEIPBA-SNPs and PBA-SNPs were collected and washed with 200 μL100 mM sodium phosphate buffer (pH 7.4) for 3 times each. Finally, the components bound on the SNPs were eluted by 100 mM acetic acid solution. The eluent containing glycoprotein or adenosine adsorbed by the SNPs were measured with UV-vis spectrophotometer. UV absorbance was adopted at 260 nm for adenosine, 403 nm for HRP, 280 nm for TRF. The Scatchard analysis was implemented according to a previously reported method [43]. Dissociation constant (Kd) and apparent maximum binding capacity (Qmax) were evaluated according to the following Scatchard equation:

Qe Qmax Qe   Cs Kd Kd where Qmax and Kd is the saturated binding capacity and the dissociation constant, respectively, Qe is the binding amount of glycoprotein or adenosine to PEIPBA-SNPs and PBA-SNPs at equilibrium, Cs is the free concentration at adsorption equilibrium. The values for Kd and Qmax can be calculated from the slope and the intercept of plots of Qe/Cs versus Qe.

2.5. Measurements of the lowest extractable concentration of HRP Equivalent amounts of PEIPBA-SNPs (10 mg) were added to a series of 10-mL

plastic microcentrifugal tubes which contained 8 mL HRP solutions at concentration ranging from 0.2 pg/mL to 3 ng/mL. To measure the captured HRP in the eluates, 1 µL eluate was mixed with 100 μL TMB staining solution and incubated for 15 min. After terminalizing the reaction by adding 10 μL H2SO4 solution (1M), the obtained solutions were detected by absorbance at 450 nm within 30 min. A stock solution of TMB was prepared by dissolving 6.0 mg TMB in 1 mL dimethyl sulphoxide. A staining solution was prepared by adding 25 μL TMB stock solution and 1 μL H2O2 (30% v/v) to 2 mL of a mixture of 0.1 M Na2HPO4 and 0.05 M citric acid (1:1).

2.6. Effect of competing monosaccharides on the enrichment of glycoproteins by PEIPBA-SNPs The enriching, washing and elution procedures were identical to those in section 2.5 except for adding monosaccharides to solution. Equivalent of PEIPBA-SNPs (3 mg) were added to a series of solutions (1 mL) of 100 mM sodium phosphate (pH 7.4) containing 1 ng/mL HRP without or with 1 mg/mL different monosaccharides including mannose, xylose, fucose, galactose, Neu5Ac, glucose and fructose. HRP amount in the eluates was determined with TMB colorimetric detection as stated in section 2.5. The same procedure was performed for PBA-SNPs.

2.7. Selective enrichment of glycoproteins in saliva samples Whole saliva was collected from adults in the morning, more than 2 h after the last intake of food. The mouth was rinsed with water before the collection. The tube

was then placed on ice to collect whole saliva. Protease cocktail inhibitor (1 µL per mL of saliva) was added to the saliva immediately to reduce protein degradation. The obtained samples were then centrifuged at 10 000 rpm at 4 °C for 15 min. After that, the supernatant was collected and stored at -80 °C. Before analysis, the samples were thawed at 4 °C. After the pH values were adjusted to 7.4 using 250 mM NH4Ac containing 400 mM NaCl (pH 7.4) of equivalent volume to saliva samples, 3 mg of PEIPBA-SNPs were added to 300 µL the above solutions and the obtained mixtures were shaken on a rotator for 2 h at room temperature. After capturing and washing, the compounds extracted by the PEIPBA-SNPs were eluted with 100 µL 100 mM HAc solution for 1 h. The obtained eluates were submitted to MALDI-TOF MS analysis.

3. Results and discussion 3.1. Characterization of the silica nanoparticles (SNPs) The morphology of the prepared SNPs was characterized by transmission electron microscopy (TEM). As shown in Fig. 2A, TEM images revealed that these SNPs had good dispersibility and relatively homogeneous size distribution. The average diameter of the SNPs was estimated to be about 80 nm. To investigate if the dispersibility and homogeneous size distribution were changed by post-modifications, the PEIPBA-SNPs (PEI 10000) was selected as the representative materials. As depicted in Fig. 2B, the obtained PEIPBA-SNPs exhibited good dispersibility and relatively homogeneous size distribution. The average diameter of the PEIPBA-SNPs was evaluated to be 80 nm.

3.2. Selectivity The selectivity of the prepared PEIPBA-SNPs was first investigated using adenosine as a cis-diol-containing test compound while deoxyadenosine as non-cis-diol test compound. As shown in Fig. 3A, PEIPBA-SNPs selectively captured adenosine but excluded deoxyadenosine, indicating excellent selectivity. To further demonstrate the selectivity of PEIPBA-SNPs toward glycoproteins, HRP, TRF and RNase B were used as test glycoproteins while RNase A and BSA was used as non-glycoproteins. As depicted in Fig. 3B, HRP, TRF and RNase B were well captured by the PEIPBA-SNPs under neutral condition while RNase A and BSA was not retained at all. These results indicate that the PEIPBA-SNPs exhibited good selectivity toward cis-diol-containing compounds.

3.3. Optimization of the molecular weight of PEI To take advantage of synergistic effect to the greatest extent, the total number of the binding sites should be as large as possible. Generally, the larger the molecular weight of PEI is, the more amine groups it can provide. However, once the molecular weight of PEI exceeds a certain value, a portion of amine groups may become inaccessible. Thus, it is necessary to investigate the effect of the molecular weight of PEI on the number of the binding sites. The number of the binding sites can be reflected by binding capacity (Q, mg/g). To determine which molecular weight of PEI works the best, we investigated the effect of the chain length of PEI on the binding capacity. As depicted in Fig. S2, the binding capacity increased with the increase of

molecular weight of PEI from 600 to 10,000. However, when PEI 70,000 was used, the binding capacity is lower as compared with PEI 10,000. The, PEI 10,000-modified 3-carboxybenzoboroxole-functionalized SNPs were considered as the optimal boronate avidity SNPs for further investigations and application.

3.4. Determination of Kd and Qmax Binding affinity can determine how low the concentrations of glycoproteins can be enriched by boronate affinity materials. The dissociation constant (Kd) and binding capacity (Qmax) of PEIPBA-SNPs and PBA-SNPs with HRP (44 kDa) and transferrin (TRF) (77 kDa) as well as adenosine (267 Da), were evaluated using UV–vis spectrophotometry. We first investigated binding strength of PEIPBA-SNPs and PBA-SNPs toward adenosine. As shown in Fig. 4, according to binding isotherms and Scatchard plots analyses, Kd values of PEIPBA-SNPs and PBA-SNPs were calculated to be (1.09 ± 0.12) and (1.01 ± 0.11) × 10−3 M, respectively. Clearly, these Kd values are also comparable with those the previous boronate affinity materials provided [45, 46]. This possible explanation is that as adenosine contains only one cis-diol group, there is no synergistic effect on the PEIPBA-SNPs. Thus, PEIPBA-SNPs and PBA-SNPs provided comparable binding constants toward adenosine. We further investigated the binding strength of the PEIPBA-SNPs and PBA-SNPs toward the two glycoproteins (HRP, TRF). The binding isotherms and Scatchard plots for HRP and TRF were shown in Fig. 5 and Fig. S3, respectively. The Kd values obtained from the Scatchard

plot are summarized in Table 1. It can be observed form Table 1 that the PEIPBA-SNPs exhibited Kd value of 10−7 M for HRP or TRF, which is lower than that of the PBA-MNPs by 2 orders of magnitude. Obviously, the amplified number of boronic acid moieties by using PEI as a scaffold can improve greatly the binding strength of the boronic acid-functionalized SNPs toward glycoproteins due to the PEI-assisted synergistic effect, which was the highest among already reported boronic acid-functionalized materials except boronate affinity MIPs. The Kd value of the PEIPBA-SNPs for HRP was lower than that of PAMAM dendrimer-amplifed boronate avidity MNPs ((1.0 ± 0.1) × 10−6 M)) [43] by 1 order of magnitude. Because PEI is much more flexible than PAMAM dendrimer, it enhanced more effectively the binding strength toward glycoproteins. In addition, the estimated binding capacity (Qmax) can provide a deeper understanding of synergistic effect of multiple binding. As shown in Table 1, the Qmax value of the PEIPBA-SNPs was about 5-fold higher than that of the PBA-SNPs for adenosine because PEI amplified the number of boronic acid moieties. However, the Qmax values of the PEIPBA-SNPs were comparable to those of the PBA-SNPs for glycoproteins. This implies that the synergistic effect of multiple binding can enhance effectively the binding strength toward glycoproteins but fails to improve the binding capacity toward glycoproteins.

3.5. The lowest extractable concentration of HRP As we know, binding strength can determine how low extractable concentration of

glycoproteins. As the PEIPBA-SNPs exhibited significantly increased binding affinity and binding capacity towards glycoproteins, the PEIPBA-SNPs were able to extract trace glycoproteins. To investigate the lowest extractable concentration of HRP, UV-vis absorbance with 3,3,50,50-tetramethylbenzidine (TMB) staining, one of the most sensitive approaches for the detection of HRP, was used. As shown in Fig. 6, for HRP standard solution, the lowest detectable concentration was 0.04 ng/mL (S/N = 6.0). When the HRP standard solution was extracted by PEIPBA-SNPs, the lowest detectable concentration was lowered to 0.4 pg/mL (S/N = 12).

3.6. Effect of competing monosaccharides on the extraction of glycoproteins by PEIPBA-SNPs Because biological samples contained plenty of sugars, the extraction of glycoproteins by conventional boronate affinity materials in real samples can be greatly influenced. That’s, the competing binding of sugars and glycoproteins to boronic acids may result in reduced binding capabilities of boronate affinity materials towards glycoproteins. However, since PEIPBA-SNPs are able to provide much higher binding affinity towards glycoproteins as compared with their affinity towards monosaccharides, the PEIPBA-SNPs had stronger anti-interference ability for competing monosaccharides. As shown in Fig. 7, when concentration of a competing monosaccharide is 1 million-fold higher than that of HRP, the HRP amount extracted by the PEIPBA-SNPs was 70–88% of that in the absence of sugar. In contrast, the HRP amount extracted by the PBA-SNPs was only 15–28% of that in the absence of sugar.

3.7. Binding pH Binding pH is an important binding property of boronic acid-functionalized materials. As most commercially available boronic acids are generally weak acids with a pKa of 8–9, conventional boronic acid-functionalized materials require a basic binding pH (usually the pH should be > the pKa value of the boronic acid ligand). This is an apparent limitation because the pH of frequently used biological samples ranges from 4.5 to 8.0. Thus, the use of a basic pH will lead to not only inconvenience of pH adjustment but also the risk of degradation of labile compounds. To address this issue, relatively low binding pH (neutral conditions) was obtained by introducing boronic acid

ligands

with

electron-withdrawing

groups,

such

as

DFFPBA

and

3-carboxybenzoboroxole.[24,47] The prepared boronate avidity SNPs (PEIPBA-SNPs) using 3-carboxybenzoboroxole as a ligand and PEI as should provide a lower binding pH because higher affinity can result in lower binding pH. As depicted in Fig. 8A, the boronate avidity SNPs are able to well capture HRP and exhibit high binding capacity at

pH

≥

6.0.

As

a

comparison,

the

boronate

affinity

SNPs

using

3-carboxybenzoboroxole as a ligand exhibited relatively low binding capacity at pH 6.0 (Fig. 8B). The result further confirms the presence of the synergistic effect of multiple binding, which gives rise to lower binding pH.

3.8. Analysis of glycoproteins in a real sample To confirm the feasibility of the PEIPBA-SNPs for application to real samples, the selective enrichment of trace glycoproteins from human saliva was carried out. In

human saliva, a lot of glycoproteins are potential biomarkers for early diseases diagnosis [48]. However, ionization efficiency and abundance of these glycoproteins are low in real samples. Besides, high-abundance interfering components usually co-exist in sample matrix with glycoproteins, which seriously suppresses the mass spectrometric (MS) signal of glycoproteins. Therefore, selective enrichment is essential prior to the analysis of glycoproteins. A human saliva sample collected from an adult (pH 7.0) was first extracted using the PEIPBA-SNPs and PBA-SNPs, separately. Since the boronate avidity SNPs could bind glycoproteins at a low binding pH, the SNPs were directly applied to the sample without pH adjustment. The extracted compounds were eluted and the resulting solutions as well as the original sample were subject to MALDI-TOF MS analysis. As shown in Fig. 9, almost no glycoprotein peaks were observed for direct analysis due to interference of high-abundance components in saliva. When the sample was extracted by the PEIPBA-SNPs, 10 peaks were observed and identified as 7 glycoproteins according to the reported values in literature [49,50] (their identities are listed in Table 2). In contrast, when the saliva sample was treated by PBA-SNPs, only 1 glycoprotein was identified (1 was unknown). In addition, the performance of the PEIPBA-SNPs was better than that of PAMAM dendrimer-amplified boronic acid-functionalized SNPs [43]. The glycoprotein number found by PAMAM dendrimer-amplifed boronate avidity MNPs was less, being 4. The boronate avidity SNPs is more advantageous for the selective enrichment of trace glycoproteins from real samples, in terms of glycoprotein number extracted.

Conclusions In this study, the high boronate avidity SNPs were prepared for the selective enrichment of trace glycoproteins. By using branched PEI as the scaffold for amplification of boronic acid sites and 3-carboxybenzoboroxole as the affinity ligand, high boronate avidity was obtained owing to PEI-assisted synergistic multivalent binding. The obtained binding constant is 10-7 M, which was the highest among already reported boronic acid-functionalized materials except MIPs. Besides, PEI-assisted synergistic effect can lead to lower binding pH. Due to the enhanced binding strength and the lower binding pH, the boronate avidity MNPs can be directly applied to selective enrichment of trace glycoproteins from real samples. References [1] P.M. Rudd, T. Elliott, P. Cresswell, I.A. Wilson, R.A. Dwek, Glycosylation and the immune system, Science 291 (2001) 2370–2376. [2] J. A. Ludwig, J. N. Weinstein, Biomarkers in cancer staging, prognosis and treatment selection, Nat. Rev. Cancer 5 (2005) 845–856. [3] H. J. Gabius, Glycobiomarkers by glycoproteomics and glycan profiling: emergence of functionality, Biochem. Soc. Trans. 39 (2011) 399–405. [4] H. Kaji, H. Saito, Y. Yamauchi, T. Shinkawa, M. Taoka, J. Hirabayashi, K. Kasai, N. Takahashi, T. Isobe, Lectin affinity capture, isotope-coded tagging and mass spectrometry to identify N-linked glycoproteins, Nat. Biotechnol. 21 (2003) 667–672. [5] Z.B. Xu, X.W. Zhou, H.J. Lu, N. Wu, H.B. Zhao, L.N. Zhang, W. Zhang, Y.L.

Liang, L.Y. Wang, Y.K. Liu, P.Y. Yang, X.L. Zha, Comparative glycoproteomics based on lectins affinity capture of N-linked glycoproteins from human Chang liver cells and MHCC97-H cells, Proteomics 7 (2007) 2358–2370. [6] B.F. Mann, A.K.P. Mann, S.E. Skrabalak, M.V. Novotny, Sub 2-mu m macroporous silica particles derivatized for enhanced lectin affinity enrichment of glycoproteins, Anal. Chem. 85 (2013) 1905–1912. [7] P.C. Lin, P.H. Chou, S.H. Chen, H.K. Liao, K.Y. Wang, Y.J. Chen, C.C. Lin, Ethylene

glycol-protected

magnetic

nanoparticles

for

a

multiplexed

immunoassay in human plasma, Small 2 (2006) 485–489. [8] B.C. Kim, M.K. Ju, A. Dan-Chin-Yu, P. Sommer, Quantitative detection of HIV-1 particles using HIV-1 neutralizing antibody-conjugated beads, Anal. Chem. 81(2009) 2388–2393. [9] H.W. Morehead, K.W. Talmadge, D.J. O’Shannessy, C.J. Siebert, Optimization of oxidation of glycoproteins: an assay for predicting coupling to hydrazide chromatographic supports, J. Chromatogr. 587 (1991) 171–176. [10] H. Zhang, X.J. Li, D.B. Martin, R. Aebersold, Identification and quantification of N-linked glycoproteins using hydrazide chemistry, stable isotope labeling and mass spectrometry, Nat. Biotechnol. 21 (2003) 660–666. [11] T. Liu, W.J. Qian, M.A. Gritsenko, D.G. Camp II, M.E. Monroe, R.J. Moore, R.D. Smith, Human plasma N-glycoproteome analysis by immunoaffinity subtraction, hydrazide chemistry, and mass spectrometry, J. Proteome Res. 4 (2005) 2070–2080.

[12] Y.Q. Jiang, Y.F. Ma, A fast capillary electrophoresis method for separation and quantification of modified nucleosides in urinary samples, Anal. Chem. 81(2009) 6474–6480. [13] W. Zhou, N. Yao, G.P. Yao, C.H. Deng, X.M. Zhang, P.Y. Yang, Facile synthesis of aminophenylboronic acid-functionalized magnetic nanoparticles for selective separation of glycopeptides and glycoproteins, Chem. Commun. 557 (2008) 7–5579. [14] M. Chen, Y. Lu, Q. Ma, L. Guo, Y.Q. Feng, Boronate affinity monolith for highly selective enrichment of glycopeptides and glycoproteins, Analyst 134 (2009) 2158–2164. [15] L.B. Ren, Y.C. Liu, M.M. Dong, Z. Liu, Synthesis of hydrophilic boronate affinity monolithic capillary for specific capture of glycoproteins by capillary liquid chromatography, J. Chromatogr. A 1216 (2009) 8421–8425. [16] J. Tang, Y. Liu, P. Yin, G.P. Yao, G.Q. Yan, C.H. Deng, X.M. Zhang, Concanavalin Aimmobilized magnetic nanoparticles for selective enrichment of glycoproteins and application to glycoproteomics in hepatocelluar carcinoma cell line, Proteomics 10 (2010) 2000–2014 [17] L. Liang, Z. Liu, A self-assembled molecular team of boronic acids at the gold surface for specific capture of cis-diol biomolecules at neutral pH, Chem. Commun. 47 (2011) 2255–2257. [18] Y.C. Liu, L.B. Ren, Z. Liu, A unique boronic acid functionalized monolithic capillary for specific capture, separation and immobilization of cis-diol

biomolecules, Chem. Commun. 47 (2011) 5067–5069. [19] Z.A. Lin, J.L. Pang, H.H. Yang, Z.W. Cai, L. Zhang, G.N. Chen, One-pot synthesis of an organic–inorganic hybrid affinity monolithic column for specific capture of glycoproteins, Chem. Commun. 47 (2011) 9675–9677. [20] F. Yang, Z.A. Lin, X.W. He, L.X. Chen, Y.K. Zhang, Synthesis and application of a macroporous boronate affinity monolithic column using a metal–organic gel as a porogenic template for the specific capture of glycoproteins, J. Chromatogr. A 1218 (2011) 9194–9201. [21]Z.A. Lin, J.L. Pang, Y. Lin, H. Huang, Z.W. Cai, L. Zhang, G.N. Chen, Preparation and evaluation of a phenylboronate affinity monolith for selective capture of glycoproteins by capillary liquid chromatography, Analyst 136 (2011) 3281–3288. [22] Z.J. Liu, K. Ullah, L.P. Su, F. Lv, Y.L. Deng, R.J. Dai, Y.J. Li, Y.K. Zhang, Switchable boronate affinity materials for thermally modulated capture, separation and enrichment of cis-diol biomolecules, J. Mater. Chem. 22 (2012) 18753–18756. [23] Y.C. Liu, Y. Lu, Z. Liu, Restricted access boronate affinity porous monolith as a protein A mimetic for the specific capture of immunoglobulin G, Chem. Sci. 3(2012) 1467–1471. [24] H.Y. Li, H.Y. Wang, Y.C. Liu, Z. Liu, A benzoboroxole-functionalized monolithic column for the selective enrichment and separation of cis-diol containing biomolecules, Chem. Commun. 48 (2012) 4115–4117.

[25] L.T. Liu, Y. Zhang, L. Zhang, G.Q. Yan, J. Yao, P.Y. Yang, H.J. Lu, Highly specific revelation of rat serum glycopeptidome by boronic acid-functionalized mesoporous silica, Anal. Chim. Acta 753 (2012) 64–72. [26] Qu, YY; Liu, X; Liang, Z; Zhang, LH; Zhang, YK. Boronic Acid Functionalized Core-Shell Polymer Nanoparticles Prepared by Distillation Precipitation Polymerization for Glycopeptide Enrichment.Chem. Eur. J. 2012, 18, 9056. [27] Xu, ZF; Uddin, KMA; Ye, L. Boronic Acid Terminated Thermo-Responsive and Fluorogenic Polymer: Controlling Polymer Architecture for Chemical Sensing and Affinity Separation. Macromolecules 2012, 45, 6464. [28] Shen, XT; Xu, CG; Uddin, KMA; Larsson, PO; Ye, L. Molecular recognition with colloidosomes enabled by imprinted polymer nanoparticles and fluorogenic boronic acid. J. Mater. Chem. B 2013, 1, 4612. [29] F. Yang, J. Mao, X.W. He, L.X. Chen, Y.K. Zhang, Preparation of a boronatefunctionalized affinity hybrid monolith for specific capture of glycoproteins, Anal. Bioanal. Chem. 405 (2013) 5321–5331. [30] H. Li, Y.H. Shan, L.Z. Qiao, A. Dou, X.Z. Shi, G.W. Xu, Facile synthesis of boronatedecorated polyethyleneimine-grafted hybrid magnetic nanoparticles for the highly selective enrichment of modified nucleosides and ribosylated metabolites, Anal. Chem. 85 (2013) 11585–11592. [31] G.B. Xu, W. Zhang, L.M. Wei, H.J. Lu, P.Y. Yang, Boronic acid-functionalized detonation

nanodiamond for specific enrichment of glycopeptides

glycoproteome analysis, Analyst 138 (2013) 1876-1885.

in

[32] S.T. Zhang, X.W. He, L.X. Chen, Y.K. Zhang, Boronic acid functionalized magnetic nanoparticles via thiol–ene click chemistry for selective enrichment of glycoproteins, New J. Chem. 38 (2014) 4212-4218. [33] X.H. Zhang, J.W. Wang, X.W. He, L.X. Chen, Y.K. Zhang, Tailor-made boronic acid functionalized magnetic nanoparticles with a tunable polymer shell-assisted for the selective enrichment of glycoproteins/glycopeptides, ACS Applied Mater. Interfaces 7 (2015) 24576–24584. [34] J.W. Wang, X.W. He, L.X. Chen, Y.K. Zhang, Boronic acid functionalized magnetic nanoparticles synthesized by atom transfer radical polymerization and their application for selective enrichment of glycoproteins, RSC Adv. 6 (2016) 47055-47061. [35] X.Y. An, X.W. He, L.X. Chen, Y.K. Zhang, Graphene oxide-based boronate polymer brushes via surface initiated atom transfer radical polymerization for the selective enrichment of glycoproteins, J. Mater. Chem. B 4 (2016) 6125-6133. [36] J.A. Yang, X.W. He, L.X. Chen, Y.K. Zhang, Thiol-yne click synthesis of boronic acid functionalized silica nanoparticle-graphene oxide composites for highly selective enrichment of glycoproteins, J. Chromatogr. A 1513 (2017) 118-125. [37] L.F. Chang, X.W. He, L.X. Chen, Y.K. Zhang, Mercaptophenylboronic acid-capped Mn-doped ZnS quantum dots for highly selective and sensitive fluorescence detection of glycoproteins, Sensor Actuat. B-Chem. 243 (2017) 72–77. [38] D.J. Li, Y. Chen, Z. Liu, Boronate affinity materials for separation and molecular

recognition: structure, properties and applications, Chem. Soc. Rev. 44 (2015) 8097-8123. [39] Y.X. Li, M. Hong, M. Miao, Q. Bin, Z.Y. Lin, Z.W. Cai, G.N. Chen, Novel composites of multifunctional Fe3O4@Au nanofibers for highly efficient glycoprotein imprinting, J. Mater. Chem. B 1 (2013) 1044–1051. [40] W. Zhang, W. Liu, P. Li, H.B. Xiao, H. Wang, B. Tang, A fluorescence nanosensor for glycoproteins with activity based on the molecularly imprinted spatial structure of the target and boronate affinity, Angew. Chem. Int. Ed. 53 (2014) 12489–12493. [41] S.S. Wang, J. Ye, Z.J. Bie, Z. Liu, Affinity-tunable specific recognition of glycoproteins

via boronate affinity-based controllable oriented surface

imprinting, Chem. Sci. 5 (2014) 1135–1140. [42] Z.A. Lin, L.X. Sun, W. Liu, Z.W. Xia, H.H. Yang, G.N. Chen, Synthesis of boronic acid-functionalized molecularly imprinted silica nanoparticles for glycoprotein recognition and enrichment, J. Mater. Chem. B 2 (2014) 637–643. [43] H.Y. Wang, Z.J. Bie, C.C. Lü, Z. Liu, Magnetic nanoparticles with dendrimer-assisted boronate avidity for the selective enrichment of trace glycoproteins, Chem. Sci. 4 (2013) 4298–4304. [44] Z.F. Xu, K. M. A. Uddin, T. Kamra, J. Schnadt, L. Ye, ACS Appl. Mater. Interfaces 2014, 6, 1406−1414. [45] D.J. Li, Y. Li, X.L. Li, Z.J. Bie, X.H. Pan, Q. Zhang, Z. Liu, J. Chromatogr. A 1384 (2015) 88–96.

[46] Y. Chen, S.S. Wang, J. Ye, D.J. Li, Z. Liu, X.C. Wu, Insights into the effect of nanoconfinement on molecular interactions, Nanoscale 6 (2014) 9563–9567. [47] Q.J. Li, C.C. Lu, Z. Liu, J. Chromatogr. A 1305 (2013) 123–130. [48] S. Chiappin, G. Antonelli, R. Gatti, E.F. De Palo, Saliva specimen: a new laboratory tool for diagnostic and basic investigation, Clin. Chim. Acta 383 (2007) 30–40. [49] M. Gonzalez-Begne, B. Lu, L. Liao, T. Xu, G. Bedi, J. E. Melvin and J. R. Yates, J. Proteome Res., 2011, 10, 5031–5046. [50] P. Ramachandran, P. Boontheung, Y. M. Xie, M. Sondej, D. T. Wong, J. A. Loo, J. Proteome Res., 2006, 5, 493–1503.

Fig. 1 Principle of enhancing the binding strength of boronic acid-functionalized silica nanoparticles towards glycoproteins through PEI-assisted synergistic binding.

Fig. 2 TEM images for bare SNPs (A) and PEIPBA-SNPs (B).

A

3.5 3.0

Qe (mg/g)

2.5 2.0 1.5 1.0 0.5 0.0

Adenosine

Deoxyadenosine

B 25 Q (mg/g)

20 15 10 5 0 HRP

TRF

RNase B RNase A

BSA

Fig. 3 Comparison of the amount of different samples captured by PEIPBA-SNPs. Binding buffer: 100 mM sodium phosphate buffer containing 400 mM sodium chloride (pH 7.4); elution solution: 100 mM HAc (pH 2.7); samples: A) 1 mg/mL adenosine or deoxyadenosine; B) 1 mg/mL HRP, TRF, RNase B, RNase A, or BSA.

3.2

PEIPBA-SNPs PBA-SNPs

2.8

Qe (mg/g)

2.4 2.0 1.6 1.2 0.8 0.4 0.0 0.0

B 18 16

Kd = (1.09±0.12) × 10-3

14

Kd = (1.01±0.11) × 10-3

Qe[adensosine] (mL/g)

A 3.6

0.3

0.6

0.9

1.2

12 10 8 6 4 2 0 0.0

[adenosine] (mg/mL)

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Qe (mg/g)

Fig. 4 Binding isotherms (A) and Scatchard plots (B) for the binding of PEIPBA-SNPs and PBA-SNPs with adenosine.

B1600

PEIPBA-SNPs

Qe (mg/g)

20

1000

5 0.00 0.06 0.12 0.18 0.24 0.30

800 600

[HRP] (mg/mL)

Qe (mg/g)

10 8 6 4 2 0.8

4

D 18

PBA-SNPs

0.4

2

1.2

1.6

[HRP] (mg/mL)

2.0

6

8

10 12 14 16

Qe

Kd = (2.44±0.21) × 10-5 M R2 = 0.96

17 16 15 14 13 12 11 10

Qe/[HRP] (mL/g)

12

R2 = 0.96

1200

10

C 14

Kd = (5.12±0.48) × 10-7 M

1400

15

0

Qe/[HRP] (mL/g)

A 25

4

5

6

7

8

Qe

9

10 11

Fig. 5 Binding isotherms for the binding of PEIPBA-SNPs (A) and PBA-SNPs (C) with HRP and Scatchard plots for the binding of PEIPBA-SNPs (B) and PBA-SNPs (D) with HRP.

3.0

Absorbance (450 nm)

2.5

HRP standard HRP extracted by PEIPBA-SNPs

2.0 1.5 1.0 0.5 0.0 0.4 pg/mL 4 pg/mL 0.04 ng/mL 0.4 ng/mL 4 ng/mL

[HRP] Fig. 6 The lowest extractable concentration of HRP

A

2.8

Absorbance (450 nm)

2.4 2.0 1.6 1.2 0.8 0.4 0.0

without Fuc Man Xyl saccharide

Gal Neu5Ac Glu Fru

0.8

Absorbance (450 nm)

B 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

without Fuc Man Xyl Gal Neu5Ac Glu Fru saccharide

Fig. 7 Effect of competing monosaccharides on the HRP amount extracted by the PEIPBA-SNPs (A) and the PBA-SNPs (B). Samples: (A) 1 ng/mL HRP without or with 1 mg/mL monosaccharide; (B) 1 ng/mL HRP without or with 1 mg/mL fucose (Fuc), mannose (Man), xylose (Xyl), galactose (Gal), Neu5Ac, glucose (Glu) or fructose (Fru).

A24

PEIPBA-SNPs

Q (mg/g)

20 16 12 8 4 0 8.0

7.0

6.0

5.0

pH 24

B 21

PBA-SNPs

18

Q (mg/g)

15 12 9 6 3 0 8.0

7.0

6.0

5.0

pH Fig. 8 Comparison of the binding capacity of PEIPBA-SNPs (A) and PBA-SNPs (B) towards HRP at different pH values.

1500 9

Intensity

1000

4

2 1

500

3 1

5 6

7 8

10

2

b a

0

20000

c

40000

60000

80000

m/z

100000

Fig.

Fig. 9 MALDI-TOF MS spectra for human saliva (a), proteins extracted from saliva by the PBA-SNPs (b) and proteins extracted from saliva by the PEIPBA-SNPs (c). Peaks for glycoproteins are marked with Arabic numerals (1–10).

Table 1 Dissociation constants (Kd) and binding capacity (Qmax) of PEIPBA-SNPs or PBA-SNPs for glycoprotein and adenosine PEIPBA-SNPs Compound

Qmax Kd (M)

(μmol/g)

PBA-SNPs Qmax

R2

Kd (M)

(μmol/g)

R2

Adenosine

(1.09±0.12)×10-3

0.065±0.006

0.95

(1.01±0.11)×10-3

0.013±0.0014

0.92

HRP

(5.12±0.48)×10-7

0.76±0.062

0.96

(2.44±0.21)×10-5

0.51±0.019

0.96

TRF

(2.32±0.25)×10-7

0.35±0.02

0.92

(1.22±0.11)×10-5

0.27±0.01

0.96

Table 2 the glycoprotiens enriched by PEIPBA-SNPs and PBA-SNPs from saliva PEIPBA-MNPs

PBA-MNPs

Number

MW (Da)

1

2

3

4

22688

Name Neutrophil gelatinase associated Lipocalin precursor [M + H] +

27954

α-Amylase [M+2H]2+

29970

Phosphoglycerate mutase [M + H]+

33205

5

40753

6

45012

Galectin-3 binding protein precursor [M + 2H]2+ Alivary proline-rich protein 2 [M+H]+ Haptoglobin precursor [M+H]+

MW (Da)

Name

28708

Phosphoglycerate mutase [M + H]+

43024

unknown

7

8

51317

Hemopexin precursor [M + H]+

55778

α-Amylase [M + H]+

9

66024

10

81404

Galectin-3 binding protein precursor [M +H] + Alivary proline-rich protein 2 [2M+H]+

Highlights • PEI-assisted boronic acid-functionalized silica nanoparticles were prepared. • PEI was applied as a scaffold to amplify the number of boronic acid moieties. • 3-carboxybenzoboroxole was used as an affinity ligand. • The boronate avidity SNPs exhibited strong binding strength toward glycoproteins. • The prepared SNPs were successfully applied to the analysis of human saliva.

Graphical abstract