Introduction

The enzyme alkaline phosphatase (ALP) can induce specific hydrolysis of phosphate groups of a variety of molecules, including nucleotides, proteins, and small molecules under alkaline conditions [1]. In vivo studies indicate that the enzyme exists in many tissues and plays critical roles in many biochemical and physiological processes such as bone mineralization, embryonic development and regulation of lipid and phosphate transport [2, 3]. Many diseases are tightly related with abnormal expression of this enzyme. For example, high serum ALP levels are associated with various diseases including bone disease, liver dysfunction, prostatic cancer, and bile duct blockage, while aberrant low expression is usually linked with hypophosphatasia, osteoporosis, anemia, cretinism, et al. [4,5,6,7,8]. Thus, developing reliable, convenient, and sensitive methods for ALP detection is of great significance for better understanding the role of ALP in related biological processes and clinical diagnosis.

A number of techniques such as electrochemical, colorimetric, and various fluorescence-based strategies have been developed to quantify ALP activity in the past decades. For example, Miao et al. [9] developed a novel electrochemical biosensor for ALP sensing using complementary DNA probes coupled with λ exo. However, this method needs complex manipulations including electrode preparation and modification as well as the removal of surplus or adsorbed oligonucleotide, which increases the cost and overall assay time; moreover, the detection limit of this method (100 U/L) should be improved. Recently, Zhang et al. [10] demonstrated an improved immobilization-free electrochemical strategy for ALP assay to avoid the tedious and time-consuming steps of electrode modification. Moreover, the sensitivity was greatly improved using signal amplification (the detection limit is 0.1 U/L). Despite this, the drawbacks of this method reside in the use of multiple reagents, i.e., KF polymerase, dNTPs, and T7 exonuclease, which increased the cost and affected the repeatability of the assay. Although colorimetric methods without covalent immobilization of the nucleic acid and the introduction of excess amounts of salt have been reported for ALP assay [11], the low detection limit of the method (32 U/L) significantly limited their wide application. Yang et al. [12] developed a colorimetric method with a detection limit of 0.84 U/L for ALP assay on the basis of pyrophosphate-induced inhibition of peroxidase-like activity and used it for ALP quantitative determination in biosamples. However, the complexity of G20-Cu(II) limits its wide application. Recently, fluorescence analysis based on various nanomaterials has attracted wide attention owing to its excellent convenience, sensitivity, and environmental friendliness. For example, using β-CD-CQDs nanoprobes, Tang et al. [13] developed a convenient and sensitive assay for real-time monitoring reaction and quantitative evaluation of ALP activity. Liu et al. [14] developed a label-free fluorescence method for ALP assay through modulating the fluorescence of nitrogen-doped graphene quantum dots (NGQDs). Although these methods are capable of assaying ALP in biosamples, the disadvantages of narrow detection range and complicated manipulation steps should not be ignored. Furthermore, the need to prepare stable duplex under complicated conditions greatly increased the complexity of manipulation and probe design. Therefore, the development of a sensitive, convenient, and low-cost strategy for ALP assay is still urgently needed for clinical diagnosis and fundamental studies.

As a class of two-dimensional sp2 carbon nanomaterials, graphene oxide (GO) has been attracted tremendous attention in the areas of biosensors, bioimaging, drug/gene delivery, photodynamic therapy, and tissue engineering owing to its intrinsic physical and chemical properties [15]. In biosensing, using its ability to adsorb single-stranded DNA, GO was used to construct a fluorescence or electrochemical biosensor platform to detect various targets, including metal ions [16], DNA [17], RNA [18], and enzymes [19, 20]. However, reduced graphene oxide (rGO) with high solubility and stability simultaneously showed tighter adsorption and stronger fluorescence quenching capability than GO [17]. On the basis of these discoveries, we developed a novel ALP sensor and constructed a compound screening platform by combing rGO and λ exo enzyme. The principle of the assay for ALP is as follows: the substrate of HP1 is a hairpin DNA probe modified with a phosphate moiety and a fluorophore at the 5′-end. The presence of ALP can hydrolyze the 5′- phosphate moiety of HP1 and produce a new 5′-hydroxyl moiety. Since λ exo has negligible activity on non-phosphorylated DNA strands [20], the hairpin probe which maintains the integrity can be adsorbed by rGO and accordingly resulted in fluorescence signal quenching. In contrast, without the hydrolysis of HP1 probe by ALP, λ exo can cleave the DNA strand with a 5′- phosphate moiety and produce many single oligonucleotides. Due to the weak adsorbability of rGO to the single oligonucleotide, the release of single oligonucleotides causes the restoration of fluorescence FAM. Thus, ALP activity can be easily detected by monitoring the fluorescence signal change.

Experimental section

Materials and reagents

Alkaline phosphatase (ALP, M0290) and λ exonuclease (λ exo, M0262) were purchased from New England Biolabs (Beijing, China). Hairpin DNA probe (HP1) and dsDNA probe (HP2, containing two complementary strands F-DNA and P-DNA) were purchased from Takara Biotechnology Co. Ltd. (Dalian, China) and purified using high-performance liquid chromatography. Their sequences are as follows: HP1, FAM-5′-p-TAGCTTATAAAAATATGGAGCGGCATAAGCTA-3′; HP2, 5’-TCAACATCAGTCTGATAAGCTA-FAM-3′ (F-DNA), 5′-p-TAGCTTATCAGACTGATGTTGA-3′ (P-DNA). The buffer solutions used in this work were as follows: (1) 1 × Lambda exonuclease buffer solution: 67 mM glycine-KOH (pH 9.4), 2.5 mM MgCl2, 50 μg/mL BSA; (2) Tris-HCl: 50 mM Tris-HCl (pH 8.0), 100 mM NaCl, 5 mM MgCl2. All solutions were prepared using Milli-Q purified water.

Preparation of rGO

rGO was prepared from GO (XF NANO Co. Ltd., Nanjing, China) as follows: 1 mg/mL GO was mixed with 5 mg/mL NaOH and sonicated for 15 min at 25 °C. Then, the mixture was put into a water bath for 1 h at 40 °C and cooled to room temperature. Then, the mixture was put into a water bath at 90 °C with stirring for 1 h and incubated for 30 min more with vigorous stirring. The mixture was sonicated again for 15 min at 25 °C, filtered, and the insoluble rGO black solid was washed with deionized water. The rGO was redispersed in water to a concentration of 1 mg/mL and stored at 4 °C for further use.

Material characterizations

The sheet size and heights of rGO and GO were evaluated using a Bruker Dimension ICON atomic force microscope (Brucker, MA) at a concentration of 10 μg/mL. ζ-potentials of rGO and GO were determined by nanoparticle size and Zeta-potentiometer (Zetasizer Nano ZS, Britain) at a concentration of 1 mg/mL. The UV–vis absorption spectra of rGO and GO were measured from 200 to 700 nm by an ultraviolet–visible spectrophotometer (UV-1800, Shimadzu Company, Japan) at a concentration of 20 μg/mL. The dispersion stability of solution was observed at 4 °C for 30 days. The fluorescence spectra of rGO and GO solution were measured using a FL-2500 fluorescence spectrometer (Hitachi, Japan) at a concentration of 100 μg/mL and the parameter setting was Ex/Em = 450/521 nm.

Determination of ALP activity

For quantitative measurement of ALP activity, 0.5 μL HP1 (10 μM) was added as the substrate for ALP in a suitable volume of Tris-HCl buffer solution (pH 8.0) and incubated with different concentrations of ALP at 37 °C for 30 min (see Electronic Supplementary Material (ESM) Fig. S1). After that, 3 U λ exo was subsequently added into the mixture and allowed to incubate at 37 °C for 10 min. Then, 0.5 μL of rGO (1 mg/mL) was added to the mixture and incubated at 37 °C for 10 min. Finally, the fluorescence intensity was measured at Ex/Em = 450/521 nm.

ALP activity assay in diluted cell-free extracts

Tumor cells (BT-549 and SMMC-7721, purchased from the cell library of Xiangya Central Laboratory, Central South University) were cultured in DMEM supplemented with 10% (v/v) FBS, 1% penicillin-streptomycin at 37 °C in humidified air containing 5% CO2. Cell-free extracts were prepared as follows: 1 × 106 cells were harvested under trypsin treatment and centrifuged at 1500 rpm for 2 min. Cells were washed three times with 1 mL of cold PBS, centrifuged, and resuspended in 0.5 mL of ice-cold cell lysis buffer (cell signaling) on ice for 5 min. Cells were pulse-sonicated on ice five times for 5 s each. Then, cell-free extracts were centrifuged at 15,000 rpm for 20 min at 4 °C and supernatants were collected. Concentrations of cell-free extracts were quantited by measuring the absorbance at 595 nm using Coomassie blue protein reagent (Pierce, USA).

For the ALP assay in biosamples, the cell-free extracts were diluted 100 times with Tris-HCl solution (pH 8.0) and then different concentrations of ALP were added to the diluted cell-free extracts to prepare the spiked sample. HP1 (0.5 μL, 10 μM) was added in a suitable volume of Tris-HCl buffer solution (pH 8.0). Then, 1 μL of cell-free extract (protein concentrations of BT-549 and SMMC-7721 were each 1 mg/mL) was added to the Tris-HCl buffer solution and incubated at 95 °C for 10 min. After that, it was incubated with different concentrations of ALP at 37 °C for 30 min. Then, 3 U λ exo was added into this mixture and allowed to incubate at 37 °C for 10 min. Next, 0.5 μL rGO (1 mg/mL) was added into the mixture and incubated at 37 °C for 10 min. Finally, the fluorescence intensity was measured at Ex/Em = 450/521 nm.

ALP inhibitor screening

To evaluate the effect of ALP inhibitor, 0.5 μL HP1 (10 μM) was added as the substrate for ALP in a suitable volume of Tris-HCl buffer solution (pH 8.0) and then different concentrations of ATP and EDTA were mixed with 1 μL of ALP (10 U/mL) and incubated at 37 °C for 30 min. After that, 3 U λ exo was added into this mixture and allowed to incubate at 37 °C for 10 min. Next, 0.5 μL rGO (1 mg/mL) was added into the mixture and incubated at 37 °C for 10 min. Finally, the fluorescence intensity was measured at Ex /Em = 450/521 nm.

Natural compound screening

To investigate the effect of natural compounds on ALP, their effect on λ exo was researched first. Different concentrations of natural compounds were mixed with 3 U λ exo and incubated at 37 °C for 10 min in a suitable volume of Tris-HCl buffer solution (pH 8.0). Then, 0.5 μL HP1 (10 μM) was added and incubated at 37 °C for 10 min. Next, 0.5 μL rGO (1 mg/mL) was added to the mixture and incubated at 37 °C for 10 min. Finally, the fluorescence intensity was measured at Ex/Em = 450/521 nm. To investigate the effect of natural compounds on ALP, 2-μL volumes of natural compounds (10 μM) were mixed with 1 μL of ALP (10 U/mL) and incubated at 37 °C for 30 min in a suitable volume of Tris-HCl buffer solution (pH 8.0). Then, 0.5 μL HP1 (10 μM) was added as the substrate for ALP and incubated at 37 °C for 30 min. After that, 3 U λ exo was added into this mixture and allowed to incubate at 37 °C for 10 min. Next, 0.5 μL rGO (1 mg/mL) was added to the mixture and incubated at 37 °C for 10 min. Finally, the fluorescence intensity was measured at Ex/Em = 450/521 nm.

Gel electrophoresis

To verify the feasibility of the principle, polyacrylamide gel analysis was used to detect enzymatic products of ALP and λ exo. Using 1 × TBE (100 mM Tris-HCl, 83 mM boric acid,1 mM EDTA, pH 8.0) as running buffer, 20 μL of samples along with 2 μL loading buffer was loaded onto a 15% polyacrylamide gel and then run at 100 V for 70 min. The gels were stained with ethidium bromide (EB) for 30 min.

Fluorescence assay

An FL-2500 fluorescence spectrophotometer (Hitachi, Japan) was used to record the fluorescence spectra of a sample in a 700-μL quartz cuvette. The fluorescence intensity was monitored with Ex/Em = 450/521 nm and the scanning range was 495–580 nm. The slits for excitation and emission were set at 10 nm. The fitting curve of the experimental data was generated using the software Sigmaplot 12.0.

Results and discussion

Characterization of rGO

The characterization of nanomaterials is important to evaluate their physical and chemical properties [21]. Atomic force microscopy (AFM) images showed that the sizes and heights of rGO were 50–100 nm and 1–1.5 nm, respectively, which were significantly smaller than those of GO (the sizes and heights were 500–700 nm and 1–2 nm, respectively) (Fig. 1A). Compared to GO at the same concentration, rGO exhibited dramatically enhanced NIR absorbance and the maximum absorption peak λm is red-shifted from 222 nm to 236 nm. This is mainly due to the partial elimination of surface functional groups and the restoration of the largely delocalized π-electron structure on the GO [22]. No obvious precipitation was found in the GO and rGO solutions after storage for 30 days at 4 °C, which reflected the high stability of the two nanomaterials (Fig. 1B, insert). In addition, we compared the effect of reductive reaction on the fluorescence intensity of GO and found that the reductive reaction decreased the background fluorescence of GO (Fig. 1C). Considering that the rGO particle size is smaller than that of GO, compared with GO at the same mass/volume, the more numerous rGO particles have lower fluorescence intensity, suggesting that the background fluorescence intensity for every rGO particle in Fig. 1C will be much lower than that of each GO particle. Finally, we investigated the ζ-potential of the two nanomaterials. As we expected, the ζ-potential of rGO (|−34.4 eV|) decreased in magnitude compared with that of GO (|−38.7 eV|) (Fig. 1D) as a result of the partial reduction of the oxygen-containing functional groups on the GO surface.

Fig. 1
figure 1

Characterization of rGO and GO. (A) AFM images of rGO and GO; the size distribution was 50–100 nm and 500–700 nm, respectively. The concentrations of GO and rGO are 10 μg/mL. (B) UV–vis absorbance of rGO and GO at the same concentration (20 μg/mL); inserts are photographs of rGO and GO solutions. (C) Fluorescence spectra of rGO and GO with concentration of 100 μg/mL. (D) ζ-potentials of rGO and GO at the same concentration (1 mg/mL)

Full size image

Working principle of rGO-based ALP assay

The rGO-based strategy to detect ALP activity is illustrated as the Fig. 2. HP1, a hairpin probe modified with a phosphate moiety and a fluorophore at the 5′-end, is employed as the substrate for ALP. When the 5′- phosphate moiety is hydrolyzed by ALP, HP1 maintains the hairpin probe integrity since λ exo has a negligible activity on non-phosphorylated DNA strands; HP1 containing the 5′-terminal FAM fluorophore can then be efficiently adsorbed by rGO and cause fluorescence signal quenching. In the absence of ALP, the cleavage of the 5′- phosphate moiety by λ exo will produce small molecule fragments and result in the release of the 5′-terminal FAM tag into solution. These short FAM-labeled oligonucleotides, which can not be adsorbed by rGO, will emit fluorescence. Thus, ALP activity could be easily detected by monitoring the fluorescence signal change.

Fig. 2
figure 2

Schematic illustration of ALP activity assay based on the rGO and λ exo cleavage

Full size image

Feasibility analysis of rGO-based ALP assay

Firstly, we investigated the feasibility of the rGO-based ALP assay. Fig.3A and 3B indicate that the fluorescence intensity of HP1 decreased 90% after incubation with rGO. Similarly, the fluorescence intensity of HP1 decreased 89% after sequentially incubating with ALP, λ exo, and rGO. These results demonstrated that rGO can efficiently adsorb HP1 probe with phosphorylation and cause fluorescence quenching. As a control, in the absence of ALP, the fluorescence intensity of HP1 decreased 10% after incubating with λ exo and rGO in sequence respectively. This result demonstrates that λ exo can efficiently cleave the HP1 with 5′- phosphate moiety and cause fluorescence restoration due to the weak adsorbability of rGO to the FAM-labeled single oligonucleotides (enzymatic products). According to the change of fluorescence intensity of this system, we can conveniently monitor ALP activity.

Fig. 3
figure 3

Feasibility analysis of the strategy. (A) Fluorescence spectra of HP1 under different conditions. The concentrations of HP1, ALP, λ exo, and rGO are 50 nM, 100 U/L, 3 U, and 5 μg/mL, respectively. The pH of Tris-HCl buffer is 7.6. (B) Relative fluorescence intensity of HP1 under different conditions; F is the fluorescence intensity of HP1 under different conditions and F0 is the fluorescence intensity of HP1. (C) 15% PAGE analysis for verifying the principle. The concentrations of HP1, ALP, and λ exo are 50 nM, 100 U/L, and 0.5 U, respectively. The pH of Tris-HCl buffer is 7.6

Full size image

PAGE analysis was then applied to confirm the reliability of the rGO-based ALP assay. Samples 1 to 4 in Fig. 3C are HP1, HP1 + λ exo, HP1 + ALP, and HP1 + ALP + λ exo, respectively. This figure shows that the cleavage of λ exo to HP1 can produce debris (compared lane 2 with lane 1). The band of HP1 with ALP incubation (lane 3) migrated a little faster than that of HP1 (lane 1) owing to the decrease of molecular weight caused by dephosphorylation. Since λ exo has a negligible activity on non-phosphorylated DNA strands, the band in lane 4 is same as that of lane 3. These results further confirmed the reliability of the rGO-based ALP assay.

Optimization of experimental conditions

In an attempt to obtain optimal conditions and achieve high sensitivity, the pH of Tris-HCl buffer and the concentrations of rGO and λ exo were carefully investigated. Fig. 4A shows that the difference in fluorescence intensity between the solution added with ALP (HP1 + ALP + λ exo + rGO) and that without ALP (HP1 + λ exo + rGO) increased with increasing pH value. According to the ΔF data in Fig. 4A (insert), the highest ΔF occurred at a buffer pH of 8.0. Fig. 4B shows the effect of rGO concentration on the ALP assay. It can be found that the value of (F − F1)/F0 increased gradually with concentration of rGO and reached its highest value at [rGO] = 5 μg/mL. Thus, the buffer with pH 8.0 and 5 μg/mL rGO (final concentration) was used for the following experiments. We also investigated the effect of λ exo concentration on the assay. Fig. 4C clearly indicates the positive relationship between the λ exo concentration and the corresponding fluorescence intensity. The plateau obtained in the presence of 3 U λ exo or higher suggested that complete cleavage reaction had occurred. Thus, the optimal concentration of λ exo was chosen as 3 U.

Fig. 4
figure 4

Optimization of experiment conditions. (A) Effect of pH on the sensing system. Insert: fluorescence intensity difference of HP1 with and without ALP. (B) Effect of rGO concentration on the sensing system. F and F1 are the fluorescence intensity of HP1 with and without ALP (i.e., HP1 + ALP + λ exo + rGO and HP1 + λ exo + rGO, respectively), F0 is the fluorescence intensity of HP1. (C) Effect of λ exo concentration on the sensing system. The concentrations of HP1, ALP, and rGO are 50 nM, 10 U/L, and 5 μg/mL, respectively

Full size image

Quantitative measurement of ALP activity

The quantitative detection of ALP was evaluated by using this sensing system under the optimized conditions. As shown in Fig. 5A, the fluorescence intensity dramatically decreased as the ALP concentration increased. This result indicated that the dephosphorylation of HP1 caused by ALP maintained the hairpin probe integrity and the 5′-terminal FAM tag is adsorbed on the rGO even in the presence of λ exo, which has negligible activity on the non-phosphorylated DNA strand. Furthermore, a good linear relationship with a calibration equation y = 0.2579x + 0.2136 (R2 = 0.9795) was obtained with the linear region of 0.5–70 U/L, where y is the quenching efficiency (the quenching efficiency is (F − F1)/F , F is the fluorescence intensity of HP1 + λ exo + rGO,  F1 is the fluorescence of HP1 + ALP + λ exo + rGO) and x is lg[ALP]. The limit of detection (LOD) of ALP is estimated to be 0.01 U/L (Fig. 5B). The normal range of serum ALP in adults is about 20–140 U/L, but even more than 500 U/L for children and pregnant woman [7, 23, 24], which is higher than the LOD of the assay. This means that the serum sample can be used for ALP assay with small volumes, which fully demonstrates its great potential for diagnostic purposes.

Fig. 5
figure 5

(A) Fluorescence intensity of HP1 detection system under different concentrations of ALP. (B) Linear relationship between the quenching efficiency and ALP concentrations, each data point was determined from three independent experiments

Full size image

In addition, we compared the new method with a reported strategy for ALP assay [20], which is illustrated in Fig. S2 (see ESM). Referring to this article, we used the reported probe (HP2) and rGO in experiments. Firstly, we investigated the effect of rGO concentration on the fluorescence change of HP2 probe with or without ALP. Figure S3A (see ESM) shows the similar behavior to that of HP1. By comparing the difference of fluorescence intensity of the solution with or without ALP (ΔF), we found that the value of ΔF is the highest when the rGO concentration is 5 μg/mL (ESM Fig. S3B). Under the optimal conditions, we also found that the fluorescence intensity dramatically increased with the increase of ALP concentration (ESM Fig. S3C). A good linear relationship with a calibration equation y = 214.7292x + 299.8385 (R2 = 0.9606) was obtained in the linear region of 0.5–30 U/L, where y is the fluorescence intensity of HP2 + ALP + λ exo + rGO and x is lg[ALP]. The LOD of ALP is estimated to be 0.2 U/L (ESM Fig. S3D). Obviously, the LOD and the detection range of our method are better than the reported data [20]. Furthermore, by summarizing reported methods related to ALP, we found that the present method showed significant improvement in the LOD and detection range (ESM Table S1). Moreover, the strategy mentioned in this work also has the advantages of low cost, easy manipulation, and short time requirements.

ALP activity assay in diluted cell-free extracts

In order to evaluate the practicability of the proposed method, the detection of ALP activity spiked in diluted cell-free extracts was performed. 1% cell extracts of BT-549 and SMMC-7721 were added to the detection system. When the concentration of ALP increased from 0 to 50 U/L, the fluorescence intensity decreased (Fig. 6A, C) and the fluorescence quenching efficiency increased from 10.3% to 51.8% (SMMC-7721) and from 8.1% to 49.5% (BT-549) (Fig. 6B, D). The results showed that the fluorescence quenching efficiency increased with the increase of ALP concentration. Encouraged by the result, we concluded that the application of this sensor for detecting ALP in complicated samples is wholly feasible.

Fig. 6
figure 6

Fluorescence intensity of HP1 in the presence of ALP and 1% cell-free extracts of SMMC-7721 (A) or BT-549 (C). Fluorescence quenching efficiency of HP1 in the presence of ALP and 1% cell-free extracts of SMMC-7721 (B) or BT-549 (D)

Full size image

ALP inhibitor screening

Enzyme inhibitors are of great importance for drug development [25]. Besides its ability for ALP detection, this biosensor was further used to screen inhibitors of ALP. A previous study confirmed that ATP and EDTA are ALP inhibitors [26]. As the nucleotide pyrophosphatase was similar to ALP in terms of structure and catalytic activity, ALP could hydrolyze the nucleotide pyrophosphate bond [27]; thus ATP could compete with HP1 for ALP biding sites. The inhibition effect of ATP is reversible and competitive. In addition, Zn2+ and Mg2+ played important roles in the secondary structure formation of ALP. The interaction of EDTA with Zn2+ or Mg2+ can cause the inactivation of ALP and the inhibition effect of EDTA is irreversible and noncompetitive [28]. By using our new method, we investigated the effect of ATP and EDTA on the ALP activity. Fig. 7 shows that the fluorescence quenching efficiency accordingly decreased with increasing concentrations of ATP and EDTA and that the 50% inhibitory concentrations (IC50) of ATP and EDTA were 0.304 mM (Fig. 7A) and 5.253 mM (Fig. 7B), respectively. These results, which were consistent with previous reports, confirmed the reliability of our new method for inhibitor screening.

Fig. 7
figure 7

Relationship between the quenching efficiency of ALP-responsive system and the concentration of ATP (A) or EDTA (B)

Full size image

Natural compound screening

The positive correlation between the serum ALP levels and disease activity including rheumatoid arthritis (RA) has been widely reported [29,30,31,32,33,34,35,36,37,38,39,40]. Meanwhile, the dried roots of Kadsura coccinea have been used as a folk medicine for RA treatment [41]. Thus, we suspect that K. coccinea exerts its function against RA through regulating ALP activity. We therefore isolated several natural compounds from the roots of K. coccinea (Lem.) A. C. Sm. (Schisandraceae) (ESM Table S2) and investigated their effect on the ALP activity in vitro.

In order to avoid confusion, we first examined the effects of these compounds on λ exo and found that these compounds can inhibit the activity of λ exo at high concentrations (Fig. 8A). The concentration reduction of compounds from 10 μM to 0.2 μM accordingly increased the fluorescence intensity of the solution (HP1 + λ exo + rGO), which indicated the low inhibitory effect of these natural compounds on λ exo. Furthermore, the inhibitory effect of the compounds on λ exo can be neglected when the concentration is less than 0.2 μM; this finding was exploited when the method was used for ALP inhibitor screening. By investigating the effect of 0.2 μM compounds on ALP, we found that five natural compounds can inhibit ALP activity, and compound F showed the strongest inhibitory effect (Fig. 8B). It has been reported that the terpenoids reduced ALP activity in hepatocarcinoma [42, 43] and that lignans reduced ALP activity in chronic liver injury [44]. Interestingly, by further exploring the relationship between compound structure and ALP, we found that four of them are terpenoids and two are lignans. To the best of our knowledge, this is the first report about the relationship between specific terpenoids or lignans from K. coccinea and ALP activity in vitro. This finding will be very helpful for further exploring the antirheumatoid arthritis mechanism of K. coccinea.

Fig. 8
figure 8

(A) Effect of natural compounds on λ exo activity. Data are mean ± SD (n = 3 independent experiments). *P < 0.05, **P < 0.01. (B) Relative fluorescence quenching efficiency of HP1 in the presence of 0.2 μM natural compounds in the ALP detection system. Data are mean ± SD (n = 3 independent experiments). *P < 0.05, **P < 0.01

Full size image

Conclusion

We successfully designed a novel fluorescence turn-off sensing system for ALP activity that is based on the binding difference of rGO coupled with λ exo cleavage and achieved quantitative detection of ALP activity by monitoring the change of fluorescence signal. Under the optimal conditions, the LOD of ALP is estimated to be 0.01 U/L with the linear region of 0.5–70 U/L. Meanwhile, the proposed assay not only showed the capability of ALP detection in cell-free extracts but also for inhibitor screening in vitro. The results indicated that five compounds isolated from K. coccinea can differentially inhibit ALP activity. The new assay platform for detecting ALP can be applied for therapeutic drug monitoring (TDM) and rapid screening of natural compounds.