Introduction

The level of modified nucleosides excreted into urine is directly related to the degradation degree of ribonucleic acids (RNAs) in the organism. When RNAs are degraded, normal nucleosides are reutilized to synthesize nucleic acids or undergo degradation. However, modified nucleosides are directly excreted into urine due to the lack of specific phosphorylases [1]. Transfer RNAs (tRNAs) exhibit a very different modification pattern in tumor cells [2], and irregular tRNA metabolism leads to a higher level of excreted modified nucleosides [3]. Therefore, modified nucleosides in urine have been studied as cancer biomarkers [4].

Nucleosides belong to cis-diols. Various analytical methods have been developed for the selective capture of cis-diols, including hydrazide chemistry [5], boronic acid chemistry [6–8], hydrophilic interaction liquid chromatography [9] and lectin-based method [10]. Among the above strategies, boronate affinity materials (BAMs) based on boronic acid chemistry have been widely used [11–15]. The principle is that boronic acids can covalently form five- or six membered cyclic esters with 1,2- or 1,3- cis-diols under basic conditions and these cyclic esters dissociate at acidic pH values. Binding pH and adsorption capacity are two important properties of BAMs. BAMs with the two features of low binding pH and high adsorption capacity are highly appreciated. The basic binding conditions result in inconvenience of operation and the degradation of labile molecules. Four strategies have been used to reduce binding pH: introducing an electron-withdrawing group into the phenyl ring [16], utilizing intramolecular B–N [17] or B–O coordination [18], and setting up a molecular team [19]. There are two methods to improve adsorption capacity. One is increasing the amount of binding sites on a given material by improving the graft efficiency of boronate ligand [20, 21]. The other is utilizing supporting materials with large specific surface area [22, 23]. However, most reported BAMs only possess one of the two features.

In the present work, a binary boronic acid denoted as DBA was prepared by reacting cyanuric chloride with 3-aminophenylboronic acid, and then it was grafted onto attapulgite (a fibrous aluminum-magnesium silicate). There existed a strong electron-deficient 1,3,5-triazine ring in DBA, making the functionalized attapulgite to bind cis-diols at lower pH values. Moreover, since attapulgite possessed large specific surface area and DBA exhibited two binding sites, the adsorbent displayed high adsorption capacity. Therefore, the prepared material possessed the two features of low binding pH value and high adsorption capacity. Finally it was successfully applied to selective extraction of nucleosides from human urine.

Experimental

Materials and chemicals

Attapulgite (ATTA) was provided by Jiangsu Jiuchuan Nano-material Technology Co. Ltd. (Jiangsu, China, http://www.chinaclay.net), and it was activated by 6 mol L−1 hydrochloric acid before use. Cyanuric chloride was purchased from J&K Chemical Co. Ltd. (Beijing, China, http://www.jkchemical.com/index.aspx). 3-Aminophenylboronic acid (APBA) monohydrate was purchased from Energy Chemical (Shanghai, China, http://www.energy-chemical.com.cn). 3-Aminopropyltrimethoxysilane (APTMS), N,N-diisopropylethylamine (DIPEA), cytidine, uridine, guanosine, adenosine and deoxyadenosine were purchased from Aladdin Chemistry Co. Ltd. (Shanghai, China, http://www.aladdin-e.com). All other reagents were of analytical grade (Tianjin, China, http://www.rionlon.com/yyhx.html). Acetonitrile (ACN) of HPLC grade was from Dikma Technology (VA, USA, http://www.dikma.com.cn). Purified water was provided by a Milli-Q system (Millipore, Bedford, MA, USA, http://www.millipore.bioon.com.cn) and used throughout the experiments. The standard solution of 0.1 mg mL−1 for each analyte was prepared in water and stored at 4 °C in the dark. With these standard solutions, sample solutions were spiked to the desired concentrations for experiments.

Preparation of boronate affinity adsorbent

Synthesis of 3 , 3 ′ - ( 6 - chloro - 1 , 3 , 5-triazine - 2 , 4 - diyl ) bis ( azanediyl ) bis ( 3 , 1 - phenylene ) diboronic acid ( DBA )

DBA was synthesized according to the method described in literature [24]. A solution of cyanuric chloride (4.5 mmol, 0.84 g) in 20 mL acetic acid and a solution of APBA monohydrate (9.0 mmol, 1.39 g) with sodium acetate (11.25 mmol, 0.92 g) in 10 mL acetic acid-water (v/v, 1:1) were combined at room temperature. Moderate cooling was applied to the resulting solution to maintain a temperature below 25 °C for 3 h. Precipitates were separated by filtration, and washed sequentially with acetic acid and water. The solid was then dried under vacuum at 80 °C overnight. IR (KBr, Fig. S1 of ESM): 696, 795, 852 (C–Cl), 1003, 1363 (B–O), 1536, 1599, 1637, 3365 cm−1. 1H NMR (400 MHz, DMSO-d6, Fig. S2 of ESM): δ 7.27 (t, 2H), 7.50 (d, 2H), 7.79 (s, 4H), 7.86 (s, 4H), 9.96 (s, 2H). HRMS (ESI, m/z) Calcd. for [C15H14B2ClN5O4 + H+]: 386.0999, found: 386.0990.

Preparation of amine - functionalized attapulgite ( ATTA - NH 2 )

ATTA-NH2 was synthesized in a manner similar to that described in literature [25]. ATTA (1.0 g) and APTMS (2.0 mL) were dispersed in toluene (30 mL) with ultrasonic agitation for 30 min. The mixture was then refluxed for 8 h under nitrogen atmosphere. After cooling to room temperature, the precipitates were separated by filtration and washed with ethanol. The solid was then dried under vacuum at 60 °C overnight.

Preparation of binary boronic acid - functionalized attapulgite ( ATTA - NH 2 -DBA )

ATTA-NH2-DBA was synthesized in a manner similar to that described in literature [26]. DBA (1.0 g) and DIPEA (435 μL) were dissolved in N,N-dimethylformamide (DMF, 40 mL). ATTA-NH2 (1.24 g) was dispersed in the above solution with ultrasonic agitation for 30 min. The mixture was then stirred at 90 °C for 8 h under nitrogen atmosphere. After cooling to room temperature, the precipitates were separated, and washed sequentially with DMF and ethanol. The solid was then dried under vacuum at 60 °C overnight.

The synthetic route is illustrated in Fig. 1.

Fig. 1
figure 1

The synthesis procedure for boronate affinity adsorbent

Full size image

Dispersive solid-phase extraction procedure

For each extraction, ATTA-NH2-DBA (50 mg) was dispersed in 2.0 mL of loading sample to carry out ultrasonic agitation for 3 min, and then it was collected by centrifugation for 5 min at 14,797 g. Subsequently, the adsorbent was rinsed sequentially with 0.5 mL water and 0.5 mL ACN-H2O (v/v, 1:1) with ultrasonic agitation for 3 min each time. Analytes were then eluted from the adsorbent with 3 × 0.5 mL of 100 mM formic acid with ultrasonic agitation for 3 min each time. The eluate was concentrated under a gentle stream of N2 with a Termovap sample concentrator (HP5016SY, Shanghai, China). The residue was dissolved in water (200 μL) for chromatographic analysis.

Selectivity evaluation

An aqueous solution containing adenosine (cis-diol, 1 μg mL−1) and deoxyadenosine (non-cis-diol, 1 μg mL−1) was used to investigate the selectivity of ATTA-NH2-DBA. This solution (2.0 mL) was incubated with ATTA-NH2-DBA (20 mg) for 3 min. Then the supernatant was collected by centrifugation for 5 min at 14,797 g. The initial solution and the supernatant were injected into the chromatographic system. The process was repeated with 10-fold, 100-fold and 1000-fold excesses of the interfering analog to determine the ability of ATTA-NH2-DBA to capture cis-diols from complex samples.

Adsorption capacity evaluation

ATTA-NH2-DBA (20 mg) was incubated with aqueous solutions (2.0 mL) of adenosine for 3 min. The concentrations of adenosine solutions ranged from 0 to 800 μg mL−1. Supernatants were collected by centrifugation for 5 min at 14,797 g and then injected for HPLC analysis. The equilibrium adsorption amount (q e , mg g−1) was calculated according to the following formula:

$$ {q}_e=\left({c}_0-{c}_e\right)V/m $$
(1)

where c 0 and c e (μg mL−1) are the concentration of adenosine solution before and after adsorption, respectively; V (mL) is the volume of the adenosine solution; and m (mg) is the mass of ATTA-NH2-DBA.

Urine sample preparation

Urine samples were collected from healthy volunteers and stored at −20 °C. After thawing, urine sample was diluted with water at a ratio of 1:9 (v/v), and then mixed with a vortex mixer. The obtained mixture was centrifuged for 5 min at 14,797 g, and the supernatant was used for the extraction procedure. Spiked sample was prepared by addition of nucleoside standards in urine.

Results and discussion

Choice of materials

Cyanuric chloride was chosen to synthesize boronate ligand due to its following features. (i) It is a trivalent molecule and its three chlorine atoms are easily displaced by various nucleophiles in a controlled manner. Therefore, it reacts with two boronic acid molecules to obtain a binary boronic acid. One remaining chlorine atom in the binary boronic acid can react with the nucleophilic groups on supporting materials, thus the binary boronic acid can be grafted onto supporting materials. There exist two binding sites on the binary boronic acid, thus adsorption capacity of the prepared adsorbent is improved. (ii) There exhibits a strong electron-deficient 1,3,5-triazine ring in the binary boronic acid, making the resulting material to bind cis-diols at lower pH values. APBA was selected to react with cyanuric chloride, because the amino group in APBA is nucleophilic and can react with the chlorine atoms in cyanuric chloride. We chose ATTA as the supporting material. Our previous work [23] has demonstrated that ATTA-based boronate affinity adsorbent exhibited high adsorption capacity due to the large specific surface area provided by ATTA.

Characterization

FT-IR spectra of ATTA, ATTA-NH2 and ATTA-NH2-DBA are shown in Fig. 2. C–H absorbance band at 2931 cm−1 in the FT-IR spectra of ATTA-NH2 suggested that APTMS was successfully grafted onto ATTA. Moreover, XPS confirmed the presence of C, N, O, Si and B elements in ATTA-NH2-DBA (Fig. 3). The existence of B element suggested that ATTA-NH2-DBA was synthesized successfully. Besides, ICP was used to characterize ATTA-NH2-DBA, and the result showed that the mass percent of B was 0.3528 % in ATTA-NH2-DBA. This further confirmed the successful synthesis of ATTA-NH2-DBA.

Fig. 2
figure 2

FT-IR spectra of ATTA (a) and ATTA-NH2 (b)

Full size image
Fig. 3
figure 3

X-ray Photoelectron Spectroscopy (XPS) for element analysis of ATTA-NH2-DBA, all elements (a) and B element (b)

Full size image

TEM images were obtained to investigate the morphology of the prepared materials. ATTA was shown to have a fibrous morphology that was unchanged after modification (Fig. S3 of ESM). Additionally, nitrogen adsorption-desorption measurement at 77 K showed that the BET surface area of ATTA-NH2-DBA was 114.7 m2 g−1.

Optimization of method

The following parameters were optimized: sample pH value and ionic strength, the equilibrium time of adsorption and desorption, and the number of elutions. The results are given in the ESM (Fig. S4 and S5). The binding pH of ATTA-NH2-DBA was 5.0 (Fig. S4). The low binding pH was attributed to the strong electron-withdrawing 1,3,5-triazine ring in DBA. As the pH of urine varies from 4.5 to 8.0 [27], the pH of urine was above 5.0 after ten-fold dilution with water. Therefore, nucleosides were directly extracted from ten-fold diluted urine sample without pH adjustment. Other optimized experimental conditions were as follows: (a) NaCl concentration in sample: 0 mM, (b) adsorption time: 3 min, and (c) desorption condition: 3 × 0.5 mL of 100 mM formic acid with ultrasonic agitation for 3 min each time.

Selectivity evaluation

The cis-diol, adenosine, and non-cis-diol, deoxyadenosine, were used as analytes to evaluate the selectivity of ATTA-NH2-DBA. As shown in Fig. 4, only adenosine was bound by ATTA-NH2-DBA. Moreover, the signal intensity of adenosine after adsorption was stable when the amount of deoxyadenosine increased. The above all suggested that ATTA-NH2-DBA had high selectivity for cis-diols.

Fig. 4
figure 4

Chromatograms of the mixture of adenosine (1) and deoxyadenosine (2) before (i) and after (ii) adsorption by ATTA-NH2-DBA with mass ratios (adenosine/deoxyadenosine) of 1:1 (a), 1:10 (b), 1:100 (c) and 1:1000 (d)

Full size image

Adsorption capacity evaluation

The data of adsorption experiments was analyzed by the Langmuir and Freundlich adsorption isotherm models (Fig. S6 of ESM) as expressed in Eqs. (2) and (3), respectively.

$$ {q}_e=b{q}_m{c}_e/\left(1+b{c}_e\right) $$
(2)
$$ {q}_e=k{c_e}^{1/n} $$
(3)

where q e and q m (mg g−1) are the equilibrium adsorption amount and the maximum adsorption amount, respectively; c e (μg mL−1) is the equilibrium concentration of adenosine in solution after adsorption; b (mL μg−1) is the adsorption equilibrium constant; k (mL μg−1) and n are Freundlich isotherm constant related to adsorption capacity and adsorption intensity, respectively.

Parameters from the fitting of Langmuir and Freundlich adsorption isotherm models are presented in Table S1 of ESM. It showed that Langmuir model was more appropriate to describe the isothermal adsorption behavior of adenosine on ATTA-NH2-DBA. The adsorption capacity of ATTA-NH2-DBA calculated from Langmuir model was 19.48 ± 1.13 mg g−1. The high adsorption capacity of ATTA-NH2-DBA was contributed to two factors. (a) One was the large specific surface area provided by ATTA. We prepared 2,4-difluoro-3-formyl-phenylboronic acid-modified magnetic ATTA in our previous work [23]. It has been demonstrated that the large specific surface area provided by ATTA made the adsorbent to exhibit high adsorption capacity (13.78 mg g−1 for adenosine). (b) The other was the binary boronic acid DBA. Compared with 2,4-difluoro-3-formyl-phenylboronic acid-modified magnetic ATTA, ATTA-NH2-DBA possessed higher adsorption capacity. This suggested that the prepared binary boronic acid improved adsorption capacity.

Binding pH and adsorption capacity are two important properties of BAMs. BAMs are expected to have the two features of low binding pH and high adsorption capacity. ATTA-NH2-DBA possessed the two features. Table 1 lists a comparison between ATTA-NH2-DBA and reported BAMs. Most reported adsorbents only had one of the two features: (a) low binding pH but with low adsorption capacity [16–19, 29], (b) high adsorption capacity but with high binding pH [20, 30]. Although the material in literature [31] possessed the two features, its binding pH was higher and adsorption capacity was lower compared with ATTA-NH2-DBA.

Table 1 Comparison between ATTA-NH2-DBA and reported BAMs in terms of binding pH and adsorption capacity
Full size table

Methodological investigation

Under the optimal conditions, the features of the method were investigated. As shown in Table S2 of ESM, the limits of detection (LODs) and limits of quantification (LOQs) were in the range of 4–17 ng mL−1 and 13–57 ng mL−1, respectively. In addition, the intra- and inter-day relative standard deviations (RSDs) were ≤7.7 % and ≤9.1 %, respectively (Table S3 of ESM). The recoveries were measured by analyzing the spiked human urine at three different concentrations ranging from 0.1 to 2.0 μg mL−1. The result showed that the recoveries were in the range of 85.0–112.9 % (Table S3 of ESM). Figure 5 shows the chromatograms of nucleosides standard and spiked urine sample. The chromatogram of spiked urine sample after extraction exhibited the peaks of non-target compounds. It was because there existed a variety of nucleosides in urine and those non-target nucleosides were also adsorbed by the adsorbent [21].

Fig. 5
figure 5

Chromatograms of nucleosides standard with each 1.0 μg mL−1 analyzed directly (a), and spiked urine sample (2.0 μg mL−1 for each nucleoside) before (c) and after (b) extraction. Peaks: 1, cytidine; 2, uridine; 3, guanosine; 4, adenosine

Full size image

A comparison between this method and reported methods is presented in Table 2. It was observed that this method had relatively low LODs and consumed less adsorbent. Moreover, our method did not need the procedure of pH adjustment, avoiding inconvenience of operation and the degradation of labile molecules.

Table 2 Comparison of various methods for determination of nucleosides with BAMs
Full size table

Conclusion

A 1,3,5-triazine-containing binary boronic acid (denoted as DBA) was synthesized. It was then grafted onto ATTA, generating ATTA-NH2-DBA. The applicable pH range of ATTA-NH2-DBA (≥ 5.0) covered the pH ranges of frequently used biosamples, such as blood, tear and saliva. This enabled ATTA-NH2-DBA to extract analytes from these biosamples without pH adjustment. Moreover, ATTA-NH2-DBA possessed high adsorption capacity and can capture more targeted molecules. Thus we expect the application of this material in the areas of proteomics, metabolomic and glycomics. However, ATTA-NH2-DBA had the limitation of separation. Centrifugation was applied to separate ATTA-NH2-DBA from sample solutions, which was inconvenient and time-consuming. We will further try to settle this issue and improve the method.