Introduction

Multi-enzyme cascade (MEC) reactions consist of at least two successive reactions [1], which provide an effective reaction system that can avoid an ineffective diffusion effect [2]. They have some significant advantages, such as less waste generation [3], higher yields, a shorter reaction time, and a more convenient operation process. Hence, more and more people are interested in MEC [4]. Although they have many advantages as mentioned, there are still two difficult problems in the MEC system: the spatial distance between the two enzymes is difficult to adjust, and the system lacks stability. Regulating the spatial distance of synergistic enzymes is the key to improving the efficiency and stability of MEC [5], so how to regulate the distance between enzymes is the major obstacle limiting the development of MEC. The conjugation of synergistic enzymes is commonly found in MEC reactions and has the advantages of efficient utilization of intermediates and improved yields compared to non-coupled reactions.

Natural enzyme, a kind of important biocatalyst, has high specificity for substrates and high catalytic efficiency. So it has been widely applied in production and disease diagnosis. Unfortunately, the development of a natural enzyme is limited due to its poor stability and high cost. Therefore, nanozymes have been studied and developed to replace natural enzymes. Nanozymes, a class of nanomaterials with natural enzyme-like activities, are a promising research area. Nanozymes not only have the catalytic activity of natural enzymes, but also can adjust their different activities by changing the preparation conditions, such as peroxidase-like activity, oxidase-like activity, superoxide-like activity, etc. [6]. Compared with natural enzymes, nanozymes overcome their disadvantages such as difficult storage, easy deactivation, and time-consuming preparation [7]. Since the discovery of Fe3O4 with peroxidase-like activity in 2007, various types of nanomaterials have been synthesized, such as carbon nanomaterials [8, 9], noble metal nanomaterials [10, 11], and metal oxide nanomaterials [12]. Compared with other types of nanozymes, Fe3O4 nanozymes have the advantages of low price, easy preparation, convenient classification, and recycling, so they have been widely studied and applied [13]. Despite the above-mentioned advantages, multiple complex activities of a nanozyme may exist simultaneously. This means that a reaction phenomenon that uses nanozymes may be the result of several different activities. Therefore, it is particularly important to explore the natural enzyme-like activities of nanozymes.

Graphene oxide (GO) is a two-dimensional nanomaterial with a high surface area and good biocompatibility, which can reduce background signals as a carrier [14]. The introduction of GO provides nanomaterials with good thermal and chemical stability and greatly improves the solubility and dispersion of the nanomaterials in polar solvents such as water [15]. Magnetic nanoparticles are more likely to agglomerate and aggregate because of their magnetic properties and high surface energy, which seriously affects the application and operation of nanozymes. Introducing a carrier is one of the ways to solve this problem; therefore, Li and his coworkers proposed a new strategy for cholesterol (CS) detection using Pd nanoparticles confined in ZIF-8 [7]. Some mesoporous materials are often used as carriers of nanozymes, but this may lead to the problem of slow molecular mass transfer due to the blockage of pore channels by nanozymes. GO, as a planar two-dimensional material, can overcome the above problem while having a high specific surface area [16]. Therefore, GO as a carrier to support nanozymes is an effective strategy.

The detection of human metabolite levels is of increasing significance, which is not only directly or indirectly related to the occurrence of many diseases but also plays a role in ensuring food safety. For example, uric acid (UA) and CS are two extremely important metabolites in the human body. The occurrence of gout [17], hyperuricemia, various cardiovascular diseases [18, 19], hypothyroidism, and other diseases are closely related to the levels of UA and CS in human body fluids. In addition, the detection of CS levels in milk can effectively help consumers to ensure the safety of the products they buy. At present, the common detection methods of UA and CS include the colorimetric method [20, 21], chromatography [22], the sensor method [23,24,25,26], etc., but these methods have disadvantages such as tedious operation, high cost, long time, and so on. A new detection method needs to be developed urgently.

In 2017, Yan proposed a new strategy to improve the activity of Fe3O4 nanoparticles by the introduction of histidine based on simulating the microenvironment of natural enzymes [27]. Because the action of the imidazole group and iron on the surface of the nanozyme could simulate horseradish peroxidase–like activity [27], the histidine-doped Fe3O4 (H-Fe3O4) exhibited significantly enhanced peroxidase-like activity. Inspired by this, we first fabricated GO-supported H-Fe3O4 (GO@H-Fe3O4) by a one-step solvothermal method, which not only improved the peroxidase-like activity of the nanozyme, but also solved the problem of easy aggregation of the nanozyme. Subsequently, GO@H-Fe3O4 was covalently linked with natural enzyme with hydrophilic poly(ethylene glycol) as the spacer arm. With uric acid oxidase (UAO) as the model natural enzyme, the GO@H-Fe3O4-linked UAO (GHFU) could perform multi-enzyme cascade reactions for the detection of UA. Briefly, UAO specifically catalyzed the oxidation of UA to generate H2O2 which then immediately oxidized the colorless 3,3′,5,5′-tetramethylbenzidine (TMB) to blue ox-TMB by the GO@H-Fe3O4. Eventually, the GHFU was successfully used in the detection of UA in human serum samples, and the universal applicability of the GHFU strategy was also evaluated by the detection of CS in milk samples with GO@H-Fe3O4-linked cholesterol oxidase (GHFC).

Experimental method

Chemicals and materials

GO was supplied by 3AMaterials (Anhui, China). CS; cholesterol oxidase (ChOx); 3,3′,5,5′-tetramethylbenzidine (TMB); UA; UAO; 2, 2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS); o-phenylenediamine (OPD); and boric acid were supplied by Rhawn (Shanghai, China). Histidine, ethylene diamine tetraacetic acid (EDTAN-ethyl-N′-(3-dimethylaminopropyl) carbodiimide (EDC), and N-hydroxysuccinimide (NHS) were obtained from Macklin (Shanghai, China). Triton x-100 was supplied by Elabscience Biotechnology Co., Ltd. Galactose, glucose, glycine, glutamic acid, tryptophan, and serine were purchased from Bide Pharmatech Ltd. (Shanghai, China). Bovine serum albumin and 2-morpholinoethanesulfonic acid (MES) were obtained from Meryer (Shanghai, China). Amino- and carboxyl-containing homobifunctional PEG molecules, including NH2-PEG9, COOH-PEG9, COOH-PEG136, and COOH-PEG456, were supplied by Tainjin Welcome Science and Technology Development Co., Ltd. FeCl3·6H2O, glycol, disodium hydrogen phosphate, sodium dihydrogen phosphate, sodium acetate, dimethyl sulfoxide (DMSO), sodium hydroxide (NaOH), acetic acid, potassium chloride (KCl), and hydrogen peroxide (H2O2) were supplied by Tianjin Jiangtian Chemical Reagent Technology Co., Ltd. Haihe cocoa milk (milk 1) and Mengniu pure milk (milk 2) were bought from the local supermarket.

Instrument

Transmission electron microscope (TEM) images were obtained from JEM-2100F TEM (JEOL) equipped with ESCALAB™ 250Xi + X-ray photoelectron spectroscopy (Thermo Fisher Scientific). Fourier transform infrared (FT-IR) spectrum analysis was carried out with a Nicolet 380 FT-IR spectrometer (Nicolet Instrument Co.). Ultraviolet and visible spectrophotometry (UV–vis) spectral analysis was performed on a YU-1901 spectrophotometer (PERSEE). Particle size and zeta potential were obtained from Zetasizer Nano ZS90 (Malvern). Thermogravimetic analysis (TGA) was achieved on a TG209F3 thermogravimetric analyzer (NETZSCH). Magnetization behavior was characterized by a HH-15 vibrating sample magnetometer (VSM, Nanjing NanDa Instrument Plant). Free radicals were detected by an EMXplus ESR spectrometer (Bruker).

Fabrication of GHFU

GO@H-Fe3O4 nanozyme was first synthesized by the solvothermal method according to a previous report with minor modification [16]. Briefly, 0.41 g FeCl3·6H2O, 1.8 g NaAc, 0.15 g sodium citrate, 0.5 g PEG10000, and 30 mg histidine were added to 20 mL ethylene glycol and completely dissolved by ultrasound; then, 0.1 g GO was added for stirring overnight. The mixture was placed in a teflon-sealed autoclave and reacted at 200 °C for 12 h. After the reaction, the production was washed three times with alcohol and water alternately and dried in an oven at 50 °C.

In order to connect the natural enzyme with the nanozyme, COOH-PEG9-COOH was firstly introduced into the GO@H-Fe3O4. Then, 5 mg COOH-PEG9, 15 mg EDC, and 30 mg NHS were dissolved in 5 mL MES buffer (pH = 6). COOH-PEG9 was activated by shaking at ambient temperature for 6.5 h. Next, 15 mg GO@H-Fe3O4 was added into MES buffer to continue reacting for 12 h. After the reaction, the unreacted substances were removed by washing with PBS (pH 7.4) three times. NHS 25 mg was dissolved in 5 mL MES buffer (pH = 6) and 5 mg of the above-mentioned GO@H-Fe3O4 was dispersed again. After shaking for 1 h, 5 mg UAO and 30 mg EDC were added into 5 mL PBS (pH 7.4), and the mixture was added to the MES buffer continuing shaking for 3 h. Finally, the GHFU was stored at 4 °C after washing three times with PBS (pH 7.4).

Oxidase-like, peroxidase-like activity and magnetism of GO@H-Fe 3 O 4

It was found that the amount of histidine added had a great influence on the properties of GO@H-Fe3O4. Therefore, the relationships between the addition of histidine and the oxidase- and peroxidase-like activity and magnetism of the GO@H-Fe3O4 were explored.

The oxidase-like activity of GO@H-Fe3O4 was evaluated by adding 100 μg GO@H-Fe3O4 and 2.8 mM TMB into 0.7 mL HAc-NaAc buffer (0.1 M, pH 3). After reacting at 40 °C for 10 min, the GO@H-Fe3O4 was separated by an external magnetic field. Then, the absorbance of supernatant at 652 nm was measured to evaluate the oxidase-like activity.

The peroxidase-like activity of GO@H-Fe3O4 was investigated by using H2O2 as the substrate. When H2O2 was used as substrate, the concentration of TMB was maintained at 800 μM, while the concentration of H2O2 was varied. The absorbance of the supernatant was measured at 652 nm after incubating at 40 °C for 10 min in pH 3 buffer. All the measurements were performed at least in triplicate, and the values were given as mean ± standard deviation (SD). The Michaelis–Menten constant (Km) and the maximum reaction rate (Vmax) were calculated from Lineweaver–Burk plots generated using Eq. (1):

$$v=\frac{{V}_{\mathrm{max}}\times \left[S\right]}{{K}_{\mathrm{m}}+\left[S\right]}$$
(1)

where ν, Vmax, [S], and Km are initial velocity, maximal reaction velocity, concentration of substrate, and Michaelis–Menten constant, respectively.

The magnetic strength of the GO@H-Fe3O4 was investigated by recording the time that a certain concentration of GO@H-Fe3O4 was concentrated on the bottle wall under an external magnetic field.

Catalytic mechanism of GO@H-Fe 3 O 4 for peroxidase-like activity

DMPO was used as the capture agent for the analysis of hydroxyl radicals (•OH) formed during the reaction. Ten micrograms per milliliter of GO@H-Fe3O4, 1.5 mM H2O2, and 100 mM DMPO were mixed with HAc-NaAc buffer (pH = 3, 0.1 M). The ESR spectrometer was used to analyze the mixture after incubation for 10 min at ambient temperature.

Colorimetric detection of UA based on GHFU

Some important parameters were optimized to improve the performance of detection, including reaction temperature, pH value of solution, amount of GHFU, and incubation time. The UA diluent was prepared by adding 0.01% Triton x-100 and 1 mM EDTA into borate saline buffer (pH = 8.5, 50 mM) and heated at 80 °C until dissolved. A certain amount of GHFU (50, 100, 150, 200, 250 μg mL−1) and 0.5 mL 200-μM UA solution (pH = 8, 8.25, 8.5, 8.75, 9) were mixed at different temperatures (20, 30, 40, 50, 60 °C) for 45 min. Then, 1.1 mM TMB and 250 μL HAc-NaAc buffer (pH = 3, 0.1 M) were added and incubated for different times (15–120 min) at 40 °C. Finally, the absorbance of the supernatant was measured at 652 nm by UV–vis absorption spectrophotometer at least in triplicate.

Performance of the GHFU-based method

Based on the optimized parameters of “Colorimetric detection of UA based on GHFU,” GHFU was performed to evaluate the UA concentration ranging from 5 to 800 μM at 652 nm. The limit of detection (LOD) and limit of quantity (LOQ) were valued according to Eq. (2) and Eq. (3):

$$\mathrm{LOD}=3\sigma /k$$
(2)
$$\mathrm{LOQ}=10\upsigma /k$$
(3)

where σ represents standard deviations of absorbance of HAc-NaAc buffer at 652 nm and k represents the slopes of the linear range of detection.

To evaluate the selectivity, the proposed GHFU-based method was used to test 100 μM UA and interfering substances, including bovine serum albumin (BSA), potassium chloride (KCl), glycine, glucose, glutamic acid, galactose, tryptophan, and serine. In order to better evaluate the selectivity of the GHFU-based method, the above interference substances with molar ratios of 1:1, 10:1, and 100:1 were evaluated.

To investigate the detection capability of the GHFU-based method toward UA in complex substrates, 1-, 10-, and 100-fold above interfering substances were added to 100 μM UA solutions, including BSA, KCl, glycine, glucose, glutamic acid, galactose, tryptophan, and serine. The above mixtures were evaluated by the GHFU-based method at 652 nm.

The storage stability of GHFU was also studied. One hundred–micromolar UA was evaluated every 5 days by GHFU which was stored at 4 °C.

Generality and practicality

The generality of GHFU was evaluated by linking ChOx with GO@H-Fe3O4 in the same way to fabricate GHFC. Subsequently, the concentrations of CS ranging from 4 to 400 μM were evaluated by GHFC.

To verify the practicability of GHFU and GHFC, they were performed to detect UA and CS in human serum and milk, respectively. The actual samples were tested after pretreatment. The protein was denatured and precipitated by heating the human serum with ethanol, and the supernatant was evaluated by GHFU after diluting 10 times. Milk obtained from the local supermarket was evaluated by GHFC after saponification.

Results and discussion

Characterization of GHFU

To demonstrate the successful fabrication of GHFU, GHFU was characterized by various techniques. Figure 1A shows the infrared spectra to investigate the preparation of GHFU. For GO@Fe3O4, the characteristic peak of GO was obtained at 3440 cm−1 of νOH and the peak of Fe3O4 was at 594 cm−1 of the Fe–O bond [28]. These peaks indicated that Fe3O4 was successfully synthesized on GO. Due to the introduction of histidine, the νOH vibration peak at 3440 cm−1 was enhanced, and the δNH vibration peak of the amino group appeared at 1630 cm−1. After COOH-PEG was added as the spacer arm, the νOH vibration peak at 3440 cm−1 was enhanced again and the νC=O and νC–O–C appeared at 1730 cm−1 and 1120 cm−1, respectively [29]. Finally, after grafting UAO, the δNH vibration peak at 1640 cm−1 was significantly enhanced, and the δ–CH3 vibration peak at 1460 cm−1 was also enhanced because of the large number of alkane chains in UAO [30]. To further investigate the fabrication of GHFU, the surface charge values of GHFU and its precursors were measured (Fig. 1B). The zeta potential of GO was − 32.0 mV. After synthesizing H-Fe3O4 on GO, the zeta potential increased to 36.8 mV because the histidine on the surface of Fe3O4 exposed a large number of positively charged amino groups. The zeta potential was significantly reduced to 0.38 mV after the introduction of COOH-PEG with negative charge. Since the isoelectric point of UAO was 5.4 and it was negatively charged in aqueous solution, the potential of UAO was further reduced after modification. These results suggested that GHFU was fabricated successfully.

Fig. 1
figure 1

A FT-IR spectra of GO@Fe3O4 (a), GO@H-Fe3O4 (b), GO@H-Fe3O4-PEG (c), and GHFU (d). B Zeta potentials of GHFU and its precursors. C XPS of GHFU. D TEM of GHFU. E Magnetization curves of GO@H-Fe3O4 (a) and GHFU (b); the photograph of the dispersion of GHFU in the absence (left) and presence (right) of an external magnet (top left corner). F Thermogravimetric analysis measurements of GO@Fe3O4 (a), GO@H-Fe3O4 (b), GO@H-Fe3O4-PEG (c), and GHFU (d)

Full size image

XPS characterization provided information on the surface chemistry of GHFU. As Fig. S2B shows, the characteristic peaks of GO@H-Fe3O4 were obtained at 725 eV for Fe2p1/2, 711 eV for Fe2p3/2, 530 eV for O1s, 399 eV for N1s, and 285 eV for C1s. After bioconjugation of GO@H-Fe3O4 and UAO, the intensity of the characteristic peaks of GHFU (Fig. 1C) made some changes. Due to the introduction of UAO, N element content increased significantly. Furthermore, detailed changes in elemental content are shown in Table S1. The morphology of H-Fe3O4, GO@H-Fe3O4, and GHFU was characterized by TEM. As shown in Fig. 1D, Fe3O4 nanoparticles were uniformly dispersed and attached to GO, and the particle size of the modified Fe3O4 nanoparticles was about 100 nm, which was significantly increased compared with that before the modification (Fig. S2A). It was also proved that the overall morphology of the nanozyme before and after the modification was not affected. On the contrary, H-Fe3O4 without the carrier would appear as an agglomerative accumulation phenomenon (Fig. S2B). The above images could prove that GO increased the dispersion and uniformity of H-Fe3O4. GHFU was composed of C, N, O, P, S, and Fe elements in the elemental mapping of GHFU (Fig. S1). The P and S elements were derived from UAO, and the Fe element was derived from Fe3O4.

Magnetization curves of GO@H-Fe3O4 and GHFU passed through the origin and were symmetrical with the origin as the center, while there was no hysteresis (Fig. 1E). The saturation magnetizations of GO@H-Fe3O4 and GHFU were 24.4 emu g−1 and 21.9 emu g−1, respectively. After modification, the magnetism of GHFU was decreased slightly but the dispersed GHFU could accumulate on the inner wall of the bottle within 15 s under an external magnetic field. Thermogravimetric analysis was also used to investigate the compositions of GHFU and its precursors. Within the range of 40–800 °C, the weight loss can be divided into three stages (Fig. 1F). The weight loss in the range of 40–200 °C was the loss of residual water and solvent; the weight loss was mainly attributed to the volatilization of the residual acetic acid group and PEG in 200–400 °C; above 400 °C, the decomposition of the components mainly included the weight loss of oxygen-containing groups in GO, histidine, and UAO. The maximum weight losses of GO@Fe3O4, GO@H-Fe3O4, GO@H-Fe3O4-PEG, and GHFU were 19.4%, 21.1%, 22.0%, and 23.3%, respectively. This result indicated that the grafted UAO accounted for 1.3% of the total mass.

Mechanism of peroxidase-like activity of GO@H-Fe 3 O 4

To further understand the mechanism of peroxidase-like activity of GO@H-Fe3O4, free radicals generated during the peroxidase-like activity of GO@H-Fe3O4 were detected. To investigate the generation of free radicals in reaction, ESR spectroscopy was performed to detect •OH. As Fig. 2A shows, when only GO@H-Fe3O4 or H2O2 was present alone, only weak characteristic peaks appeared. However, when GO@H-Fe3O4 and H2O2 coexisted, a typical fourfold characteristic peak with a relative signal intensity of 1:2:2:1 appeared. This result indicated that •OH produced by the co-existence of GO@H-Fe3O4 and •OH could be captured by DMPO to form a spin adduct, resulting in a fourfold characteristic peak of 1:2:2:1. This demonstrated that GO@H-Fe3O4 catalyzed H2O2 to form •OH, which then oxidized colorless TMB to blue ox-TMB (Scheme 3). In conclusion, the peroxidase-like activity of GO@H-Fe3O4 was the result of the generation of •OH.

Fig. 2
figure 2

A ESR spectra of DMPO and GO@H-Fe3O4 (a); DMPO and H2O2 (b); and DMPO, H2O2, and GO@H-Fe3O4 (c). B Parameters of different GO@H-Fe3O4 from steady-state kinetic with H2O2 as substrate. C Effect of linking strategies on the performance of GHFU. D UV–vis absorption spectra and the corresponding color of GO@H-Fe3O4-ABTS-H2O2 (a), ABTS-H2O2 (b), GO@H-Fe3O4-OPD-H2O2 (c), OPD-H2O2 (d), GO@H-Fe3O4-TMB-H2O2 (e), and TMB-H2O2 (f)

Full size image

Fabrication of GHFU

In recent years, there has been an increasing interest in the design of nanozyme active centers using bionic concepts. By studying the structure of the natural enzyme active center, a similar structure is simulated in the nanozyme structure. This design is able to greatly improve the activity of the nanozymes. Histidine is a very important amino acid, which is involved in the catalytic activity centers of many natural enzymes. The imidazole group in histidine can coordinate with many metals to produce peroxidase-like activity [31], oxidase-like activity [32], and catalase-like activity [33]. Inspired by this, the Fe3O4 with histidine was synthesized on GO, and GO was used as a carrier to improve the problem of easy adhesion of Fe3O4 and enhance biocompatibility. The introduction of histidine not only improved the Fe3O4 peroxidase-like activity, but also introduced amino and carboxyl groups as reactive groups on the surface of Fe3O4, providing the possibility of subsequent modifications. Unfortunately, the introduction of histidine also had an influence on the magnetism and the oxidase-like activity of GO@H-Fe3O4, which related to the separation of the GO@H-Fe3O4 and the specificity of detection. Therefore, it was necessary to investigate the relationship among the amount of histidine and the activity of peroxidase-like, oxidase-like, and magnetic properties of GO@H-Fe3O4.

To investigate the relationship between histidine content and oxidase-like activity, GO@H-Fe3O4 was used to directly oxidize TMB. As shown in Fig. S3A, with the increase of histidine, the oxidase-like activity of GO@H-Fe3O4 increased significantly. The excessive oxidase-like activity of GO@H-Fe3O4 would oxidize TMB directly but was not related to the amount of CS in the reaction system, which might reduce the specificity of the detection and increase the LOD. Further, the effect of the histidine content on the peroxidase-like activity of GO@H-Fe3O4 was evaluated. Using H2O2 as the substrate, in the presence of TMB, Km and Vmax were calculated using Eq. (1). With the increase of histidine content, the Km value decreased obviously and then remained unchanged, and the Vmax value increased dramatically and then reached equilibrium (Fig. 2B). Kcat was also calculated using Eq. (4):

$${K}_{\mathrm{cat}}=\frac{{V}_{\mathrm{max}}}{\left[E\right]}$$
(4)

where Kcat and [E] are the catalytic efficiency and concentration of GO@H-Fe3O4, respectively. Kcat also gradually increased with the increase of histidine, and reached the maximum when the amount of histidine added was 60 mg (Fig. S3B). In order to further explore the effect of histidine on the activity of nanozyme peroxidase, the performances of Fe3O4, H-Fe3O4, and GO@Fe3O4 were tested. The Km values of Fe3O4, H-Fe3O4, GO@Fe3O4, and GO@H-Fe3O4 were 13.4, 4.7, 65.4, and 10.2 mM, respectively (Fig. S3C). And the Vmax values of Fe3O4, H-Fe3O4, GO@Fe3O4, and GO@H-Fe3O4 were 35.2, 92.2, 18.8, and 44.2 mM, respectively (Fig. S3D). The peroxidase activities of H-Fe3O4 and GO@H-Fe3O4 were significantly stronger than those of Fe3O4 and GO@Fe3O4. These results could prove that the introduction of histidine on the surface of Fe3O4 could significantly improve the activity of peroxidase. The peroxidase activity of H-Fe3O4 decreased after adding GO. This is because nanozymes of the same mass were added in the activity measurement, and the introduction of GO led to a decrease in the proportion of H-Fe3O4, resulting in a weaker GO@H-Fe3O4 peroxidase activity than H-Fe3O4 at the same mass. Because the imidazole group of histidine formed in coordination with Fe can simulate the active center of horseradish peroxidase (HRP), the peroxidase-like activity was significantly enhanced after the addition of histidine. The good magnetic properties of GO@H-Fe3O4 were helpful for rapid separation in detection, so the magnetic properties of GO@H-Fe3O4 were investigated. With the increase of histidine, the magnetism of GO@H-Fe3O4 gradually reduced. This phenomenon might be because with the increase of the histidine proportion in GO@H-Fe3O4, the proportion of magnetic material decreased, which led to the weakness of magnetism (Fig. S3E). Considering the above factors, 30 mg histidine was chosen as the final addition amount, which benefited the following detection furthest.

The H-Fe3O4 was synthesized on GO by solvothermal method (Scheme 1A). The introduction of histidine not only greatly improved the activity of peroxidase-like but also introduced active groups including amino and carboxyl groups on the surface of the nanozyme. Therefore, to conjugate GO@H-Fe3O4 and UAO, COOH-PEG was introduced as a spacer arm (Scheme 1B), which could adjust the spatial distance by changing the molecular weight. The spatial distance between the UAO and the GO@H-Fe3O4 might affect its detection ability. A too-large spatial distance might cause H2O2 catalyzed by UAO to not be catalyzed by GO@H-Fe3O4 immediately, leading to a diffusion effect. On the contrary, a too-small spatial distance might result in steric hindrance to modify enough UAO on the GO@H-Fe3O4. Hence, different types and lengths of arm spacer were investigated. As expected, there was no difference between amino- and carboxyl-containing homofunctional PEG, but COOH-PEG9 showed the best performance (Fig. 2C).

Scheme 1
scheme 1

A Fabrication routes of GO@H-Fe3O4. B Bioconjugation of GO@H-Fe3O4 and UAO (GHFU)

Full size image

To further verify the peroxidase-like activity of GO@H-Fe3O4, ABTS, OPD, and TMB were used as substrates for detection. After incubation with GHFU, colorless ABTS-H2O2, OPD-H2O2, and TMB-H2O2 exhibited green, yellow, and blue, respectively, and the corresponding UV–visible absorption peaks were 415 nm, 450 nm, and 652 nm, respectively (Fig. 2D). To investigate the stability of GHFU, 100 μM UA solution was tested using GHFU every 5 days. Fortunately, GHFU activity decreased by only 3.2% after 1 month of storage at 4 °C (Fig. S3F).

Optimization of GHFU-based cascade reaction

As shown in Scheme 2, the detection of UA by GHFU was a cascade reaction, including two main steps. Firstly, UA and oxygen in the sample reacted to H2O2 and allantoic acid under the action of UAO on the surface of GO@H-Fe3O4. Subsequently, the resulting H2O2 and TMB in the reaction were then transformed into oxygen, water, and blue ox-TMB under the catalysis of GO@H-Fe3O4. The UA in the detected samples could be quantified by detecting absorbance at 652 nm. Since the detection of UA was a cascade reaction, it was necessary to take both reaction steps into account when optimizing the detection conditions.

Scheme 2
scheme 2

Schematic of the GHFU-based cascade reaction for the detection of UA

Full size image

Several key factors had a great impact on the cascade reaction, including temperature, pH, amount of GHFU, and incubation time. Because of the peroxidase-like and oxidase-like activities of GO@H-Fe3O4, the GHFU-based colorimetric method for UA detection was compared with the single GO@H-Fe3O4 method. To obtain the best performance of GHFU, the temperature was investigated firstly. As shown in Fig. 3A, at first, ΔA gradually increased with the increase in temperature, reached the maximum value at 40 °C, and then gradually decreased. The reason might be that the optimum temperature of UAO was relatively gentle, while the activity of GO@H-Fe3O4 was best at higher temperature. The effect of pH was also investigated in the range of 8 to 9. With the increase in pH, ΔA increased slightly and then decreased dramatically (Fig. 3B). The optimum pH of 8.5 might be due to the low solubility of UA under acidic conditions, while alkaline conditions were not conducive to the activity of UAO. In order to improve the detection efficiency, the amount of GHFU and incubation temperature were also investigated. With the increase in the amount of GHFU added and incubation time, ΔA gradually increased and reached saturation at 1.5 mg mL−1 (Fig. 3C) and 105 min (Fig. 3D), respectively. The key parameters obtained from the above optimization would be applied in subsequent tests.

Fig. 3
figure 3

A Effect of reaction temperature on the GHFU- and GO@H-Fe3O4-based UA detection method. B Effect of solution pH on the GHFU- and GO@H-Fe3O4-based UA detection method. C Effect of amount of nanozyme on the GHFU- and GO@H-Fe3O4-based UA detection method. D Effect of incubation time on the GHFU- and GO@H-Fe3O4-based UA detection method

Full size image

Performance of GHFU-based strategy

To verify the availability of GHFU, UA solutions with different concentrations were detected using GHFU and standard curves were established. With the increase in UA concentration, the color gradually changed from colorless to blue, and the response signal gradually increased at 652 nm (Scheme 3). As shown in Fig. 4A, the absorbance values at 652 nm were linearly and positively correlated with the concentration of UA in the range of 5–300 μM (y = 0.003x + 0.146, R2 = 0.997) and 300–800 μM (y = 0.0006x + 0.8835, R2 = 0.9918). This result was similar to that of a previous study [34]. The LOD and LOQ values were as low as 1.5 μM and 5 μM, respectively, which indicated that the GHFU-based strategy exhibited excellent performance.

Scheme 3
scheme 3

Schematic of the mechanism of peroxidase-like activity of GO@H-Fe3O4: generation of •OH

Full size image
Fig. 4
figure 4

A Plot of absorbance at 652 nm versus the concentration of UA. B Absorbance of the GHFU-based method for UA and control substances. C Absorbance of the GHFU-based method for 100 μM UA with different ratios of interfering substances. D Visible absorption spectrum of GHFC and TMB in the presence of different concentrations of CS (4–400 μM, from left to right)

Full size image

To further evaluate the selectivity of GHFU, UA and some interfering substances in serum were detected by using GHFU, including BSA, KCl, glycine, glucose, glutamic acid, galactose, tryptophan, and serine. It was obvious that the response signal of UA was higher than that of other interfering substances, which indicated the good specificity of GHFU (Fig. 4B). The reason might be that UAO could only catalyze UA but not other interfering substances to produce H2O2. To assess the responsiveness of GHFU in complex substrates, mixtures of UA and interfering substances were detected using GHFU. Although the concentration of the interfering substances was 1-, 10-, and 100-fold higher than that of the target UA, the change in response signal was still minimal (Fig. 4C). These results demonstrated that GHFU had good selectivity and specificity in detecting UA in complex matrix.

The general applicability of the GHFU-based strategy

To verify the general applicability of GHFU, ChOx was linked with GO@H-Fe3O4 to fabricate GHFC in the same way, which just replaced UAO with ChOx. Firstly, the reaction time of GHFC was optimized. As Fig. S4A shows, the absorbance at 652 nm gradually increased with the prolongation of the reaction time, and basically tended to be stable at 60 min. With the increase in CS concentration, the response signal and color gradually increased (Fig. 4D). The absorbance values at 652 nm were positively and linearly correlated with CS concentrations in the range of 4–400 μM (y = 0.004x + 0.3755, R2 = 0.995) (Fig. S4B). The LOD and LOQ values were 1.13 μM and 3.75 μM, respectively. This proved that GHFU had good general applicability and the potential to detect other metabolites.

Detection of UA in human serum and CS in milk

To demonstrate the potential of GHFU for applications of clinical testing and food safety, UA in the serum of healthy volunteers and CS in different brands of commercially available milk were measured using GHFU and GHFC, respectively. The concentrations of UA in the original serum and spiked serum samples with three levels of UA concentration (100 μM, 300 μM, and 500 μM) were measured by GHFU. The results are shown in Table 1; 220 μM UA was detected in the original serum, which was in the range of serum UA concentration in healthy people (149–416 μM) [35]. The recovery values ranged from 100.6 to 103.1%, and the standard deviation (SD) was in the range of 11.2 to 38.4. The concentrations of CS in milk and spiked milk samples with three levels of CS concentration (100 μM, 300 μM, and 500 μM) were also detected by GHFC. As listed in Tables 2 and 3, 355 μM and 269 μM CS were detected in two brands of milk. The recovery values ranged from 97.5 to 102.5%, and the SD were in the range of 5.6 to 31.5. As listed in Tables S2 and S3, the performances of GHFU- and GHFC-based strategy were comparable to or relatively better than those of other methods [25, 36,37,38,39,40,41,42,43,44,45,46,47,48,49]. These results demonstrated that the proposed GHFU- and GHFC-based strategies were applicable and feasible in detecting UA and CS in practical samples. Obviously, the GHFU-based strategy had great application prospects in clinical detection and food safety.

Table 1 The precision and accuracy of the GHFU-based strategy for the determination of UA in serum samples
Full size table
Table 2 The precision and accuracy of the GHFC-based strategy for the determination of CS in milk 1
Full size table
Table 3 The precision and accuracy of the GHFC-based strategy for the determination of CS in milk 2
Full size table

Conclusion

In summary, the peroxidase-like activity of GO@H-Fe3O4 was enhanced by the introduction of histidine. The peroxidase-like activity of GO@H-Fe3O4 was mainly attributed to the generation of •OH. The GHFU was constructed by bioconjugation of UAO and GO@H-Fe3O4 with COOH-PEG as the spacer arm. GHFU could specifically detect UA in complex samples by a one-step cascade reaction, and had a wide linear range (5–300 μM and 300–800 μM). LOD and LOQ were as low as 1.5 μM and 5 μM, respectively. Meanwhile, a good linear range and LOD of GHFC were also obtained. Good recoveries and SD of spiked UA in human serum and spiked CS in milk were obtained by GHFU and GHFC, respectively. Therefore, we believe that the GHFU-based strategy has great potential in clinical metabolite detection and food safety determination.