Introduction

Protein glycosylation, as a significant and common co-/post-translational modification, is recently attracting more and more attention. The study of glycoproteins has important biological and clinical significance because of their vital role in diverse physiological and biological processes, such as disease detection, cell division, tumorigenesis, inflammation, vaccine development, among others [14]. However, due to their relatively low-abundance, the determination and characterization of glycoproteins are severely hampered by other high-abundance proteins in complex biological samples. Efficient methods of isolation and purification of glycoproteins are indispensable in order to obtain an in-depth understanding of glycoproteins. To date, several enrichment methods of glycoproteins have been proposed for specific capture of glycoproteins. Besides lectin-based affinity chromatography [5], hydrazide chemistry [6], hydrophilic-interaction liquid chromatography [7], immunoaffinity chromatography [8], immobilized metal affinity chromatography [9] and size exclusion chromatography [10] the boronate affinity method has been developed for the isolation and enrichment of glycoproteins. Boronate affinity chromatography as a simple tool has been frequently applied in recognition of cis-diol-containing compounds because of its unique affinity property. The mechanism can be explained that the boronic acid can covalently bind with cis-diols to form five- or six-membered cyclic esters in a basic aqueous media; the reversible disassociation can be performed under acidic conditions. The excellent specificity makes boronate groups potential affinity ligands for specific recognition and isolation of cis-diol-containing compounds, including catechols, nucleotides, carbohydrates, nucleosides, glycopeptides, and glycoproteins [1114]. In addition, the moderate acidic elution condition make boronate affinity well coupled with mass spectrometry for analysis of cis-diol biomolecules in -omics [1416]. Different types of boronate affinity materials have been developed for specific capture of glycoproteins [1520].

Monolith is separation media that consist of a single, continuous, integrated interconnecting porous skeleton without interparticular voids. With the unique properties of simplicity of preparation, rapid mass-transfer rate, low backpressure, and versatile surface modification, monolithic materials have been more and more attractive and widely used in proteomic analysis [21, 22]. Since Malik et al. [23] first prepared a C18-incorprated hybrid monolith, the organic-silica hybrid monolith has drawn more attentions in the separation field [24, 25], which combines the advantages of both organic monolith and inorganic monolith, such as easily fabricated, less shrinkage and good pH stability. A “one-pot” approach was developed for preparation of organic-silica hybrid capillary monolithic columns by concurrently introducing organic monomers and alkoxysilanes. In this process, the polycondensation and copolymerization of the organic and inorganic monomers were carried out orderly by controlling at proper temperatures. The “one-pot” synthesis method is simple, convenient and time-saving, and it has been used to fabricate a series of organic-silica hybrid monolith with different organic monomers for the chromatographic analysis [2629]. These researches provided a facile approach for covalent incorporation of different organic moieties into the silica monolith to prepare various organic-silica hybrid monoliths.

Boronic acid functionalized monoliths have recently gained rising attentions [3046], and a lots of attempts have been made to design boronate affinity monolith with desired properties for glycoproteomic researches. A variety of boronate affinity monoliths have been developed for specific capture of glycopeptides and glycoproteins [3844]. Different from regular polymer-based boronate affinity monoliths, an inorganic–organic hybrid boronate affinity monolith synthesized via a “one-pot” process was brought into sight by Lin [44], and the monolith was successfully applied to isolation and enrichment of glycoproteins. A lowered binding pH not only is favorable for protecting the hybrid monolith, but also can avoid the risk of protein degradation when applied monolith to separations of glycoproteins. A number of strategies have been proposed to reduce the binding pH [34, 35, 41, 42], and the resulted boronate affinity monoliths exhibit excellent specificity toward cis-diol-containing biomolecules under neutral conditions. A hybrid boronate affinity monolith was subsequently reported by Liu [45], which exhibited selectivity of cis-diol containing compounds under neutral pH conditions. In addition, the organic-silica hybrid monoliths have the enhanced hydrophilic nature relative to conventional silica monoliths, they showed the better selectivity and basically eliminated the non-specific retention originating from hydrophobic interaction [30, 44, 45, 47]. More efforts are needed to explore new kind of organic-silica boronate affinity monolithic materials and expand their applications in the separation of glycoproteins.

Herein, we reported a facile “one-pot” process for the synthesis of 3-acrylamidophenylboronic acid (AAPBA)-silica hybrid affinity monolith for specific capture of glycoproteins. Compared with regular polymer-based boronate affinity monolith, a hybrid affinity matrix is rare and attractive. The monoliths were systematically characterized to evaluate the morphology, permeability, binding capacity and chromatographic properties. The hybrid affinity monolith showed its hydrophilic nature and no organic solvent was needed in the mobile phase. Excellent affinity to both cis-diol-containing small molecules and glycoproteins was achieved at a physiological condition, and the monolith was successfully applied to the isolation of transferrin from spiked biological sample.

Experimental details

Materials

3-Aminophenylboronic acid monohydrate (APBA) was purchased from Beijing Element Chem. Tech. Company (Beijing, China). Acryloyl chloride was purchased from Alfa Aesar (Ward Hill, MA, USA). Tetramethoxysilane (TMOS) was obtained from Chemistry Factory of Wuhan University (Wuhan, China). Vinyltrimethoxysilane (VTMS) was obtained from Acros (NJ, USA). Poly (ethylene glycol) (PEG, MW 10,000) and 2,2′-azobisisobutyronitrile (AIBN) was purchased from Tianjin Chemical Reagent Company (Tianjin, China), and AIBN was used after recrystallized with methanol. Adenosine, deoxyadenosine, bovine serum albumin (BSA), cytochrome c (cyt c), lysozyme (Lyz), ribonuclease A (RNase A), trypsin, ovalbumin (OVA), horseradish peroxidase (HRP), and transferrin (TF) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Deionized water was purified using a Milli-Q water purification system (Millipore, Milford, MA, USA). Other reagents utilized were of analytical grade or better. Stainless steel columns (100 mm × 4.6 mm i.d.) were supplied by Puxiang chromatogram equipment Co., Ltd. (Tianjin, China).

Instruments

A Shimadzu SPD-M20A HPLC system (Shimadzu, Japan) was applied to all chromatographic experiments, which consisted of two LC-20AD HPLC pumps with a DGU-20A5 online degasser, and a diode array detector. The data was acquired and processed with an LCsolution chromatographic workstation (Shimadzu, Japan). Fourier-transform infrared (FTIR) spectra were determined by using an IR instrument model FTS6000 (Bio-Rad, USA). Elemental analysis of the monolith was performed by a CHNOS elemental analyzer (Vario MICRO, Elementar, Germany). The scanning electron micrographic (SEM) images of the monoliths were examined by using a SS-550 scanning electron microscope (Shimadzu, Japan). The pore size distribution measurement was conducted on an AutoPoreIV 9500 mercury intrusion porosimetry (Micromertics, Norcross, GA, USA).

Synthesis of 3-acrylamidophenylboronic acid

AAPBA as organic monomer was synthesized according to the method mentioned in our previous work [43]. APBA (1.0 g, 7.3 mmol) was dissolved in NaOH solution (25 mL, 30 mmol), which was cooled to 0 °C in an ice-bath. Chilled acryloyl chloride (1.2 mL, 15 mmol) was added dropwise to the above mixture solution under intensive magnetic stirring for 15 min. After reacting for 1 h, the pH value of the reaction mixture was adjust to 1.0 by slowly adding a hydrochloric acid solution (1 M) to precipitate the product. The resulting white solid precipitate was filtered and washed several times with cold water. After the precipitate was dissolved in 20 % (v/v) aqueous ethanol (20 mL) and heated to 60 °C, the impurities were filtered off, and the filtrate was stored overnight at 0 °C to crystallize the product as white-needle crystals. The crystals were filtered and washed with chilled distilled water and were dried in a vacuum oven for later use. The chemical structure of AAPBA was confirmed by\( \begin{array}{*{20}c} 1\mathrm{H}\ \mathrm{NMR}\ \mathrm{spectrum},1\mathrm{H}\ \mathrm{NMR}\left( {\mathrm{HZ}} \right):\delta 10.07\left( {\mathrm{s},\ 1\mathrm{H},\ \mathrm{NH}} \right),\delta 8.02\left( {\mathrm{s},\ 2\mathrm{H},\ \mathrm{B}-\mathrm{OH}} \right),\delta 7.88\ \hfill \\ \left( {\mathrm{s},\ 1\mathrm{H},\ \mathrm{Ar}-\mathrm{H}} \right),\delta 7.81\left( {\mathrm{d},\ 1\mathrm{H},\ \mathrm{Ar}-\mathrm{H}} \right),\delta 7.49\left( {\mathrm{d},\ 1\mathrm{H},\ \mathrm{Ar}-\mathrm{H}} \right),\delta 7.28\left( {\mathrm{t},\ 1\mathrm{H},\ \mathrm{Ar}-\mathrm{H}} \right),\delta 6.46\left( {\mathrm{d}\mathrm{d},\ 1\mathrm{H},\ \mathrm{CH}} \right),\delta 6.27\ \hfill \\ \left( {\mathrm{d}\mathrm{d},1\mathrm{H},\mathrm{C}=\mathrm{CH}2} \right),\delta 5.74\left( {\mathrm{d}\mathrm{d},1\mathrm{H},\mathrm{C}=\mathrm{CH}2} \right). \hfill \\\end{array} \)

Preparation of the AAPBA-silica hybrid affinity monolith

The boronate-silica hybrid affinity monolith was synthesized via a one-pot process in a stainless steel column (100 mm × 4.6 mm i.d.). The optimum condition for the preparation of the hybrid affinity monolith was as follows: a prehydrolyzed mixture was prepared by mixing and vigorously stirring TMOS (1.8 mL), VTMS (0.6 mL), PEG (10,000 MW, 0.54 g) and acetic acid (HAC, 0.01 M, 5 mL) at 0 °C for 4 h to form a transparent solution. Subsequently, 33 mg mL−1 AAPBA and 1 wt % AIBN were introduced into the above mixture. After the mixture was sonicated for 1 min to degas, the homogeneous solution was carefully introduced into a stainless steel column using a syringe. Then the column with both ends sealed was incubated in an oven at 45 °C for 12 h, and then at 70 °C for 12 h. Finally, the prepared monolith was rinsed with water and methanol successively by a LC pump to remove PEG and other residues. The hybrid affinity monolith was pre-equilibrated with 0.2 M acetic acid over night before use.

As a control, a vinyl-containing hybrid silica monolith named “naked monolith” was synthesized by a traditional sol–gel process in the absence of AAPBA. The homogeneous prehydrolyzed solution of TMOS and VTMS was obtained as the same procedure described previously. Without adding organic monomer, the mixture was directly introduced into the stainless-steel column after sonication. After reacting at 45 °C for 12 h, the obtained monolith was flushed with water and methanol respectively to remove the residues.

Chromatographic conditions

Chromatographic separations of samples were achieved on the hybrid affinity monolith prepared via the one-pot procedure. The UV detection wavelength was set at 254 nm for nucleosides; while for the detection of proteins, the wavelength was set at 214 nm. The sample injection volume was set as 20 μL. The phosphate buffer (0.05 M, pH 7.0) containing 0.4 M NaCl and acetate acid (HAC, 0.2 M) were employed as mobile phase for the gradient elution. A flow rate of 0.5 mL min−1 was used throughout experiments. Protein samples were dissolved and diluted in phosphate buffer (0.05 M, pH 7.0) with the concentration of 1.0 mg mL−1 and stored at 4 °C prior to use. All mobile phases and sample solutions were filtered through a 0.45-μm membrane (Nihon Millipore Ltd) before use.

Determination of binding capacity of the hybrid affinity monolith

The specific binding capacity of the hybrid boronate affinity monolith was measured by frontal chromatography method. BSA as the nonglycoprotein was eluted without any retention on the monolith and was chosen as the dead time marker. OVA as the glycoprotein could be totally captured by the hybrid affinity monolith until the monolith reached to saturation. Monoliths were first equilibrated with loading buffer before measurements. Loading buffers containing 1 mg mL−1 OVA and 0.01 mg mL−1 BSA with pH ranging from 7.0 to 9.0 was pumped through the monoliths by a HPLC pump, respectively. The monolith was eluted by acetate acid (0.2 M) before next measurement.

Results and discussion

Preparation of AAPBA-silica hybrid affinity monolith

In this work, a facile and efficient approach for introduction of boronate affinity groups on hybrid monolith was employed. The AAPBA-silica hybrid affinity monolith was synthesized via a one-pot procedure, which mainly involved two major steps as mentioned in the literatures [2629]: the polycondensation between the hydrolyzed alkoxysilanes TMOS and VTMS to form vinyl-silica hybrid monolithic matrix at a relatively low temperature, and the copolymerization of vinyl functional groups on the vinyl-silica hybrid monolithic matrix and the organic monomer AAPBA at a proper higher temperature. The synthetic scheme of the AAPBA-silica hybrid affinity monolith is shown in Fig. 1. TMOS and VTMS as co-precursors were hydrolyzed with a weak acid-catalyzed process by stirring in ice-bath for 4 h. PEG as porogen to form macropore was set as 0.54 g as a classic choice mentioned in a relevant study [26]. To obtain the monolith with good property and performance, several crucial parameters such as the ratio of TMOS to VTMS, the reaction temperatures and the amount of AAPBA were optimized carefully in the preparation process.

Fig. 1
figure 1

The scheme of preparation of AAPBA-silica hybrid affinity monolith via a one-pot process

Since VTMS plays a role in influencing on the phase separation process, which is related to the skeleton size and through-pore size on macroporous monolith, the ratio of TMOS to VTMS may affect the morphology and porosity properties of the final materials [48], it is necessary to investigate the ratio carefully. A series of monoliths were synthesized with different ratio of TMOS to VTMS (v/v) from 2:1 to 5:1, as shown in Table 1 (corresponding to column 1–4). Monolith prepared by ratio of 5:1 was semitransparent and the skeleton was very fragile. When the content of VTMS increased, the monolith was fully filled in the column and the permeability was better. However, when the ratio increased to 2:1, the resulted monolith had a slack skeleton and could not tolerate a high flow rate on HPLC. At the ratio of 3:1, a homogeneous monolith could be obtained with good permeability, enough mechanical strength and denser matrix structure at the same time. Therefore, ratio of TMOS to VTMS 3:1 was regarded as the optimum ratio and used for further optimized experiments.

Table 1 Preparation conditions of the AAPBA-silica hybrid affinity monolith

The influence of reaction temperature on the formation of hybrid monolith was also investigated. The different polycondensation and copolymerization temperatures were examined during the preparation of monoliths. At a lower temperature as 40 °C, the polycondensation of alkoxysilane monomers was performed incompletely and the monolithic matrix was semitransparent. When increasing the polycondensation temperature to 45 °C, a white and opaque monolithic matrix could be obtained in 12 h. The copolymerization temperature was also investigated, when lower than 70 °C, the obtained monolith was detached from the inner wall of the column tube and residues were easily observed. When the copolymerization temperature was higher than 70 °C, the obtained monolith became homogeneous and was fully filled in the column. Considering these results, 45 and 70 °C were chosen as the optimum polycondensation and copolymerization temperature, respectively.

AAPBA as organic monomer is also a crucial parameter in the one-pot procedure. It significantly affects not only the morphology of the final monolith because of its coexistence with alkoxysilanes during the condensation process, but also the selectivity and retention behavior of the boronate affinity monolith. Different amounts of AAPBA were added into the pre-condensation solution to investigate the optimum addition (corresponding to column 5–7 in Table 1). Based on the observation of the resulted monoliths, it was found that different amounts of AAPBA did affect the formation of monolith. The higher content of AAPBA (50 mg mL−1, column 7) in the pre-condensation solution would result in an incomplete reaction or even no polymerization reaction happened. With a content of AAPBA of 40 mg mL−1 (column 6), the resulted monolith was not fully filled in the column tube. Addition of 36 mg mL−1 AAPBA (column 5) could result in a homogeneous monolith but slight shrinkage happened when pumping with mobile phase for hours. An optimal addition of 33 mg mL−1 AAPBA was adopted to get a homogenous and fully filled monolith. The as-prepared hybrid monolith under the optimal condition kept stable enough for continuous use of ∼3 months with a constant back pressure in the following experiments including binding experiment, optimization of mobile phase conditions and protein separation (Fig. S1, Supporting Information).

Characterization

The resulted hybrid affinity monoliths were characterized by FTIR spectroscopy (Fig. 2). The obvious differences can be readily identified from the spectra a (vinyl-silica hybrid monolith) and spectra b (monolith after copolymerization). The disappearance of the peak at 1,602 cm−1 in spectra b might be attributed to the decrease of vinyl groups on monolith after copolymerization, because of the polymerization reaction between vinyl groups. The strong absorption peaks at 1,669 and 1,541 cm−1 in spectra b could be attributed to amide group. The increasing intensity of peak at 1,349 cm−1 indicated the appearance of boronic acid groups (−B (OH)2). The absorption peaks observed at 1,602, 1,582, and 1,484 cm−1 were caused by stretching vibrations of benzene ring. The stretching band at 708 cm−1 was the characteristic of meta-substitution on benzene ring. In general, all of these results indicated that the phenylboronic acid groups had been successfully incorporated into the monolith.

Fig. 2
figure 2

FTIR spectra of (a) the vinyl-silica hybrid monolithic matrix and (b) the AAPBA-silica hybrid affinity monolith

The concentration of boronic acid groups in the monolithic column calculating from nitrogen content was estimated. First, the data of elemental analysis provide the concentration of vinyl functional groups on the hybrid monolith (Table 2). The increase of carbon and nitrogen content on the monolith after copolymerization demonstrated that AAPBA was incorporated into the monolith successfully during the one-pot procedure. The carbon content of vinyl-silica hybrid monolith was 11.65 % which was greater than that of theoretical calculated carbon content (8.80 %). This may be accounted for residual methyl from incomplete hydrolysis of siloxane precursors. The concentration of boronic acid groups in the column was estimated to be 2.6 mmol g−1 (calculated from nitrogen content). By comparison with theoretical calculation from the content of vinyl groups (3.65 mmol g−1), the lower content of boronic acid groups in the column can be ascribed that some of the AAPBA had not taken part in the one-pot procedure or they were not incorporated into the hybrid skeletons. Nevertheless, it was still satisfactory that ∼71.2 % boronic acid groups were immobilized on the monolith.

Table 2 The amount of C and N measured by elemental analysis and the concentration of vinyl groups on the hybrid monolith

Under the optimum preparation conditions, the morphology of the resultant monolith was characterized by SEM (Fig. 3). A hybrid monolith with a continuous network and uniform pore distribution was obtained, and the cross-section morphology of the monolith was homogenous. In contrast to Fig. 3a (the vinyl-silica monolith), Fig. 3b, c showed that globular particles were aggregated on the surface of interconnected silica skeleton to form large functional clusters after the polymerization; this phenomena was corresponding to the characteristic morphology of organic monolith. From SEM images, the flow-through macropore sizes were estimated to range from 2 to 5 μm. The existence of interconnecting macropore structure would decrease the mass-transfer resistance from mobile phase to stationary phase and improve the permeability. The pore structure of the monolith was measured by mercury intrusion porosimetry. The average pore diameter of the hybrid monolith was 108 nm, and the total surface area was 20.05 m2 g−1. This result represented the macroporous character of the prepared monolith and higher permeability could be expected.

Fig. 3
figure 3

SEM images of the vinyl-silica monolithic matrix (a) and AAPBA-silica hybrid affinity monolith under the optimized preparation conditions at different magnifications (b, c)

To evaluate the mechanical stability of the resulted AAPBA-silica hybrid affinity monolith, H2O and methanol were used to equilibrate the monoliths with the flow rate ranging from 0.1 to 5.0 mL min−1, respectively, and the changing tendency of backpressure was measured. It was found that with the increase of flow rate, the backpressures were linearly increased (y = 1.15x − 0.079 for water and y = 0.68x − 0.113 for methanol) (Fig. 4). It indicated that the mechanical stability of the resulted hybrid monolith was good. Using Darcy’s Law of permeability B 0 = P [49] (C is a constant related to flow rate and the size of column (in square meter per second), η is the viscosity of the mobile phase (in pascal second), and ΔP is the pressure drop of the column (in pascal)), the permeability of the AAPBA-silica hybrid affinity monolith under optimum conditions was calculated as 10.08 × 10−14 m2 for water (η = 1.005 cP, 20 °C) and 11.64 × 10−14 m2 for methanol (η = 0.580 cP, 20 °C), respectively. An ideal monolithic column used in proteome analysis should be stable enough in different separation buffers and not exhibit extra swelling or shrinking [50]. There was 13.4 % change of permeability on monolith when changing mobile phase, which indicated that only a slight solvent swelling happened. The advantage of less shrinkage could be attributed to organic-silica monolithic matrix. The prepared hybrid affinity monolith possessed good permeability and was stable in different solvents.

Fig. 4
figure 4

The mechanical stability of the AAPBA-silica hybrid affinity monolith

Optimization of mobile phase conditions

The specificity is a main concern in boronate affinity materials, while the existence of the electrostatic interaction always influences the selectivity of boronate affinity monolith [30]. An electrostatic interaction will happen between basic proteins and the monolith to generate undesirable secondary retention when boronic acid group is negatively charged, and it can be suppressed effectively by increasing the ionic strength of binding buffer. Since the main designed function of the boronate affinity monoliths is selective capture of glycoprotein, proteins are chosen as the test compounds to investigate the proper mobile phase conditions for the separation of glycoproteins. Adequate NaCl was added into mobile phase to suppress the non-affinity interaction between non-glycoproteins and boronate ligands. Trypsin (pI 10.5), Lyz (pI 10.7), cyt c (pI 9.8), and RNase A (pI 7.8) were chosen as basic test proteins to investigate the retention behavior on the hybrid affinity monolith by changing the concentration of NaCl in mobile phase (phosphate buffer, 0.05 M). As shown in Fig. 5a, the peak area of proteins increased when increasing the concentration of NaCl and reached to a plateau at 0.4 M NaCl addition. This result demonstrated that 0.4 M NaCl addition in mobile phase could maximally suppress the electrostatic interaction, and it was chosen as a proper salt concentration in mobile phase for further analysis of proteins.

Fig. 5
figure 5

Effect of salt concentration (a) and acetonitrile concentration (b) in the mobile phase on the chromatographic retention behavior of proteins on the AAPBA-silica hybrid affinity monolith

The existence of undesirable hydrophobicity on the resulted hybrid affinity monolith was investigated. Different concentrations of acetonitrile ranging from 0 to 50 % (v/v) were added into the mobile phase (phosphate buffer, 0.05 M, pH 7.0) and chromatographic retention behaviors of proteins were examined. As shown in Fig. 5b, a significant decreasing tendency of the peak areas was observed when the concentration of acetonitrile increased. No significant hydrophobic interaction was observed between the AAPBA-silica hybrid affinity monolith and analytes, and no organic solvent was needed in the loading and elution buffer. This property made the hybrid monolith more beneficial for analysis of biological samples. The hydrophilic characteristic of the resulted monolith could be attributed to not only the rather polar monomer AAPBA, but also the hydrophilic nature of the organic-silica hybrid monolith.

Binding capacity

The prepared AAPBA-silica hybrid affinity monolith exhibited excellent affinity to both cis-diol-containing small molecules and glycoproteins. To determine the specific binding capacity with glycoproteins on the AAPBA-silica hybrid affinity monolith, frontal chromatography method was employed. According to the principle of boronate affinity chromatography, BSA was eluted out while OVA could be totally captured until the monolith reached to saturation. With pH ranging from 7.0 to 9.0, the loading buffer containing test proteins was pumped through the monolith. The affinity dynamic binding capacity Q (in milligram per gram) was calculated according to the following equation [43]:

$$ {{{Q=\left( {{V_B}-{V_0}} \right)\times C}} \left/ {m} \right.} $$

where V B (in milliliter) is the 10 % breakthrough volume, V 0 is the dead volume, C is the concentration of OVA (in milligram per milliliter), and m is the weight of the dry monolithic rod (in gram). The results of Q measured under different pH conditions were listed in Table 3. It showed that Q increased from 2.5 to 7.9 (in milligram per gram) with the pH of mobile phase ranging from 7.0 to 9.0. The result was in agreement with the boronic acid chemistry that glycoproteins were easier to react with the boronic acid group at basic conditions. The concentration of boronic acid groups in the monolith is estimated to be 2.6 mmol g−1 by the data of the elemental analysis (calculated from nitrogen content), but the result did not represent the amount of boronic groups on the surface of the skeleton. The affinity binding capacity is related closely with the amount of AAPBA. In a sense, the calculation of binding capacity can be used as a helpful indirect parameter to estimate the accessible boronic acid groups. Compared with the value in relevant literatures (0.078 to 8.2 mg g−1) [38, 40, 43, 44], the binding capacity of the monolith was acceptable.

Table 3 Binding capacity of AAPBA-silica hybrid affinity monolith at different pH conditions

As discussed in relative literatures [30], a basic pH condition to operate boronate affinity becomes an apparent disadvantage because it can increase the risk of degradation of labile compounds. A neutral binding pH would be more suitable for physiological samples. It was worth noting that the specific binding between the AAPBA-silica hybrid affinity monolith and glycoprotein occurred at a neutral pH as 7.0. As for our knowledge, the loading capacity of the conventional boronate affinity monolith was reported no binding to glycoprotein at pH 7.0 [33, 39]. This result may be ascribed to the lower pK a of AAPBA (8.2), which was expected to explain that the resulted monolith can function at a slightly lowered pH [45]. This result indicated that the AAPBA-silica hybrid monolith had the ability to specifically capture glycoproteins at pH 7.0, and this monolithic material could be further used for biological analysis under physiological pH conditions.

Boronate selectivity

In order to evaluate the specificity of the prepared AAPBA-silica hybrid monolith toward cis-diol-containing biomolecules, adenosine was selected as test analyte and 2-deoxyadenosine (an analog of adenosine) was chosen as a contrast. The monolith was equilibrated with the phosphate buffer before use. The different chromatographic retention behavior of the two analytes was shown in Fig. 6. At a neutral mobile phase condition, 2-deoxyadenosine had no retention on the monolith and was eluted immediately near the void time, while adenosine containing the 1,2-cis-diol groups, was completely captured by the monolith at pH 7.0. The elution of adenosine was performed by switching to an acidic mobile phase (pH 2.5) at 6 min. This result was well in correspondence with the conclusion from boronate affinity principle. Obviously, the hybrid affinity monolith exhibited high affinity interaction and selectivity toward cis-diol biomolecules over other non-cis-diol analytes at neutral pH.

Fig. 6
figure 6

Chromatographic retention behavior of adenosine and 2-deoxyadenosine on the AAPBA-silica hybrid affinity monolith. Mobile phase: phosphate buffer (pH 7.0, 0.05 M, 0.4 M NaCl), switched to 0.2 M HAC at 6 min. Flow rate: 0.5 mL min−1. (1) 1 mg mL−1 2-deoxyadenosine; (2) 1 mg mL−1 adenosine; (3) the mixture of 1 mg mL−1 adenosine and 1 mg mL−1 2-deoxyadenosine

To evaluate the specificity toward glycoproteins, HRP, TF, and OVA were selected as test glycoproteins, and BSA, cyt c, RNase A (three non-glycoproteins) were selected as contrasts. As shown in Fig. 7, the chromatographic retention behaviors of proteins were remarkably different between glycoproteins and non-glycoproteins under the optimum mobile phase conditions. At neutral pH conditions, glycoproteins were completely trapped by the hybrid affinity monolith, while non-glycoproteins were quickly eluted out. When switching to an acidic mobile phase at 6 min, the reversible disassociation was performed and the glycoproteins were eluted. This result was in accordance with the principle of boronate affinity and clearly demonstrated the specific retention of glycoproteins on the hybrid affinity monolith at neutral pH. The baseline disturbance at ∼12 min was caused by the background subtraction operation, which was employed to eliminate the baseline drift caused by the changes of mobile phase conditions from neutral to acidic. The resulted monolith could also tolerate a high flow rate to achieve high-throughput and rapid separations. These properties of the hybrid monolith can easily satisfy the request to analyze biological samples in a rapid and efficient way under physiological pH.

Fig. 7
figure 7

Chromatographic retention behavior of glycoproteins and non-glycoproteins on the AAPBA-silica hybrid affinity monolith. Mobile phase: phosphate buffer (pH 7.0, 0.05 M, 0.4 M NaCl), switched to 0.2 M HAC at 6 min. Flow rate: 0.5 mL min−1

Application to biological sample analysis

The prepared monolith was further applied to the analysis of spiked biological samples. A bovine serum sample was diluted 50-fold with the 0.05 M phosphate buffer at different pH, and 1 mg mL−1 TF was dissolved into the sample. With the pH of mobile phase ranging from 7.0 to 9.0, the chromatographic retention behaviors of samples on the hybrid affinity monolith were shown in Fig. 8. The results of the chromatographic analysis showed that the bovine serum was quickly eluted near void time under alkaline or neutral conditions, while TF was captured by the hybrid affinity monolith. The mobile phase was switched to an acidic condition at 6 min, and TF was eluted and selectively separated from the bovine serum sample. The result was in accordance with the discussion of retention mechanism under neutral conditions in literatures [30]. As shown in Fig. 8, that specific capture of TF from the bovine serum sample had been achieved at different pH. At pH 9.0 (Fig. 8a) TF was eluted out at 14 min, while the elution time at pH 7.0 (Fig. 8c) was 10 min. The longer elution time of pH 9.0 could be explained that it would take more time to change the pH environment of the hybrid affinity monolith from alkaline condition to acidic condition. In case that the same results of selective separation TF from bovine serum sample could be obtained at different pH condition, the neutral pH as 7.0 has more advantages such as time-saving, suitable for analysis of proteins and closer to the physiological environment. The results further confirmed that the monoliths had the ability of capturing TF from complex biological samples at neutral pH.

Fig. 8
figure 8

Specific capture of TF from a bovine serum sample on the AAPBA-silica hybrid affinity monolith. Mobile phase: a phosphate buffer (pH 9.0, 0.05 M, 0.4 M NaCl), switched to 0.2 M HAC at 6 min. b phosphate buffer (pH 8.0, 0.05 M, 0.4 M NaCl), switched to 0.2 M HAC at 6 min. c phosphate buffer (pH 7.0, 0.05 M, 0.4 M NaCl), switched to 0.2 M HAC at 6 min. Flow rate: 0.5 mL min−1

Reproducibility of monolith preparation

The preparation reproducibility is an important factor to evaluate the effectiveness and practicability of the method. Under the optimized conditions, the reproducibility of the boronate-silica hybrid monolith was investigated by calculating the relative standard deviation (RSD) of the retention time of adenosine. The results showed that RSDs for the run-to-run, column-to-column and batch-to-batch were 1.15 % (n = 5), 4.77 % (n = 5), and 6.08 % (n = 3), respectively, indicating satisfactory reproducibility of the monolith preparation.

Conclusions

In summary, a facile one-pot approach was developed for preparation of boronate-silica hybrid affinity monolith in this work, and facile functionalization of hybrid silica monoliths could be achieved. By introducing an attractive ligand AAPBA and the hydrophilic hybrid monolith, the obtained AAPBA-silica hybrid affinity monolith exhibited excellent affinity towards glycoproteins at neutral pH conditions, without adding organic solvent in the mobile phase. In addition, the hybrid affinity monolith was successfully applied to separate transferrin from bovine serum sample at a physiological pH. The successful application suggests the prepared monolith may be a promising affinity material for glycoproteomic analysis.