Introduction

Point-of-care (POC) immunoassays are beneficial for the detection of pathogenic bacteria or monitoring biomarkers related to human health, and can be used at home or in the field. POC immunosensors are especially suitable for use in resource-limited environments because of their simplicity, accuracy, and low cost [1,2,3]. Although POC testing has grown rapidly over the past few decades [4], only a limited number of products, such as personal glucose meters (PGMs) [5] and pregnancy test strips [6], are commercially available to the public. The portable POC immunosensor has attracted interest from scientists and engineers because it produces real-time signal readouts, especially in low-resource settings [7].

The manufacture of POC devices is challenging. POC testing systems usually comprise a target recognition component, transducer, and signal processor. The transducer plays an important role in POC testing, and converts target molecular recognition events into measured signals. In the last few decades, some portable devices (e.g., PGMs, thermometers, pH meters, and pressure meters) have been developed for POC detection depending on the principle of completely different signals. Multiple new approaches based on converting non-glucose targets into glucose have been developed so that the most widely used PGM is simply used to quantify targets (e.g., platelet-derived growth factor-BB, metal ions, toxins, and pathogens) [8,9,10]. In traditional applications, the pH meter is used for detecting the concentration of hydrogen ions. pH meters also have been successfully used for detection of targets, from pathogens [11], to aflatoxin B1 [12], human oncogenic protein [13], and telomerase activity [14]. Thermometers [15] and pressure meters [16] have been successfully used as the signal readers in bioassays. The use of portable devices that are widely applied in daily life opens a new door for signal transduction strategies. However, these devices for the detection platform also have some shortcomings. For example, the respective interference of endogenous glucose and pH value in PGM and pH meter readouts in clinical samples is inevitable [17]. Pressure reading is sensitive to temperature and elevation changes. Ambient temperature changes affect thermometer readouts.

Hydrogen (H2), as a clean energy source without carbon-based emissions and other pollutants, is a globally accepted clean energy carrier. The requirement for H2 is increasing and H2 is widely used in daily life [18]. For the safety of H2 application, H2 detectors have been universally used, especially portable hand-held devices. The portable hand-held H2 detector has many attractive features, such as ease of use, portability, and accurate, sensitive H2 detection. A major challenge in using a portable H2 detector to assay targets (e.g., pathogenic bacteria) other than H2 is to find an approach that can connect the detection target to a change of H2 so that one simply detects the target and measures its concentration using the detector [15]. The efficient transformation of target recognition into generating H2 is an important step for sensitive detection in such bioassays. In all existing reactions to generate H2, the decomposition of 19.6 wt% high H2 capacity ammonia–borane (NH3-BH3, denoted as AB) has gained a lot of attention. Remarkably, generating H2 gas also has a nontoxic advantage. In theory, under standard conditions, 1 mmol AB is decomposed to produce 67.2 mL H2. If the reaction system is in a sealing device, production of H2 gas can easily be detected by a H2 detector.

A number of metal nanoparticle (MNP) catalysts have shown excellent catalytic activity and recycling performance in AB dehydrogenation [19]. Among them, platinum (Pt) is one of the most widely studied catalysts because it has remarkable activity for hydrolytic dehydrogenation of AB [20]. MNPs need to be coupled with substances, such as antibodies, that have strong affinity to the detection target, and then formed protein–MNP conjugates are used as efficient signal-transduction tags for bioassays. It is important to construct high-performance protein–MNP conjugates in the POC immunosensor. Until now, multiple methods synthesizing high-performance protein–MNP conjugates have been developed. Complex preparation processes often sacrifice the protein recognition function of the protein–MNP conjugates. A type of protein and phosphate (e.g., Cu3(PO4)2, CaHPO4) hybrid nanocomposites has attracted interest because of one-step coprecipitation preparation and large surface–volume ratio. Hybrid nanocomposites have been developed as a matrix to immobilize biological molecules, whose activity and stability are correspondingly enhanced. We have observed that the concanavalin A-Cu3(PO4)2 nanocomposites embedded a large number of hemin particles, which has outstanding peroxidase activity [21].

In the present study, we synthesized peptide–Cu3(PO4)2 hybrid nanocomposites with embedded PtNPs (PPNs), encapsulating many Pt particles, and the PPNs were an ideal immunosensor label for hand-held H2 detector signal readouts, by transforming pathogenic bacteria recognition events into H2 signals. Our adaptation of a hand-held H2 detector for pathogenic bacteria POC detection by combining hybrid nanocomposites consisted of two steps (Scheme 1): 1) the peptide–inorganic hybrid nanocomposites was captured by immunomagnetic bead separation; and 2) the hand-held H2 detector was used as a signal readout. The process of synthesizing hybrid nanocomposites at room temperature mixed antimicrobial peptides (AMPs) and PtNPs into phosphate-buffered saline (PBS) solution containing Cu2+ ions. AMPs have received increasing interest as promising biorecognition molecules for bioassays since they can bind to the surface of bacterial cells with high affinity and specificity. Using AMPs as recognition molecules, several bacterial biosensor platforms have been developed because of their stability, simple acquisition, cost-effectiveness, and standard composite protocols. PPNs had the ability to dehydrogenate AB to generate H2, and can selectively recognize target bacteria through binding of peptide to the bacterial surface. In addition, the uniform size and regular shape of the nanocomposites promoted their consistent attachment to each of the bacterial cells, leading to good linearity in the assay of pathogenic bacteria, namely E. coli O157:H7. Integration of hybrid nanocomposites–dehydrogenated AB and immunomagnetic separation led to a selective and sensitive gas signal method of detecting E. coli O157:H7.

Scheme 1
scheme 1

Schematic representation of the immunosensor for pathogenic bacteria using a H2 detector as signal readout

Full size image

To the best of our knowledge, this study is the first to develop an assay based on hand-held H2 detector readouts and nanocomposites signal amplification for pathogenic bacteria detection. A sensitive and selective method for the detection of E. coli O157:H7 was developed. The limit of detection (LOD) of the POC immunosensor was 10 CFU·mL−1 in a linear range from 10 to 104 CFU·mL−1.

Materials and methods

Reagents

PtNPs were provided by Sigma–Aldrich (St. Louis, MO, USA, www.sigmaaldrich.com/china-mainland.html). NH3-BH3 were obtained from Hwrk Chem Co. Ltd. (Beijing, China, www.hwrkchemical.com). The AMPs (Magainin I: GIGKFLHSAGKFGKAFVGEIMKS) were designed as described previously [22] and obtained from Sangon Biotech Co. Ltd. (Shanghai, China, www.sangon.com). Magnetic beads (MBs) group-modified with streptavidin (average diameter 2.8 μm) were purchased from Invitrogen (Waltham, MA, USA, www.thermofisher.com). E. coli O157:H7 polyclonal antibodies coupled with biotin (AB20640) were obtained from Abcam (Hong Kong, China, www.abcam.cn). CuSO4∙5H2O were purchased from Sangon Biotech (Shanghai, China, www.sangon.com). PBS consisted of 0.01 M NaH2PO4, 0.01 M Na2HPO4 and 0.014 M NaCl (pH 7.4). The hand-held hydrogen detector, which operated on principles of electrochemistry and thermal conductivity (SKY2000-H2), was purchased from Unitec Technology (Shenzhen, China, www.szyuante.com/list-8.html). All chemicals were used without further purification. All aqueous solutions were made from ultrapure water (18.3 MΩ∙ cm, Milli-Q; Millipore, Billerica, MA, USA, www.merckmillipore.com). Strains of Salmonella sp. (CMCC 50115), Listeria monocytogenes (Lis, BNCC 128913), E. coli O157:H7 (CICC 21530) and Staphylococcus aureus (BNCC 186335) were provided by the Academy of Military Medical Sciences (Beijing, China).

Synthesis of hybrid nanocomposites

PPNs with a diameter of about 4 μm were prepared according to the strand method with some modification [23]. Magainin I peptide aqueous solution (0.01 mg·mL−1) and CuSO4∙5H2O aqueous solution (20 μL of 120 mM) were added to 1 mL PBS (pH 7.4, 10 mM). PtNPs (0.9 mg·mL−1) were introduced under mild stirring, and allowed to form the PPNs at room temperature for 24 h. The PPNs were obtained by centrifugation (8000 rpm), rinsed twice and dispersed in 1 mL 0.1 mM PBS solution (pH 6.0). The solution was stored at 4 °C for further use. The structure of the PPNs was determined by scanning electron microscopy (SEM; Hitachi SU8010, Tokyo, Japan).

Detection process of this assay

We designed an immunosensor for E. coli O157:H7, and the complete detection strategy is shown in Scheme 1. MBs labeled with antibody were synthesized according to protocol and used for separation and recognition of E. coli O157:H7 (via streptavidin–biotin interaction); mainly including 0.05 mg·mL−1 of E. coli O157:H7 polyclonal antibodies dissolved in 5 mg·mL−1 MBs. After 45 min incubation at 25 °C, the solution of MBs modified with E. coli O157:H7 antibody was washed three times with PBS to remove the unbound antibody. The antibody−MBs were finally kept at 4 °C until further use. MB-conjugated antibody was added to 10 μL of different concentrations of E. coli O157:H7, for recognition of the target E. coli O157:H7. The mixture was incubated at 35 °C for 80 min and magnetically separated by washing with PBS three times. The supernatants were removed to dispose of the excess nontarget bacteria. PPNs (4 μL) were added to the remaining MBs for 35 min at 35 °C. After addition of the PPNs, the MBs–E. coli O157:H7 bound to the PPNs. The unbound PPNs were removed by magnetic separation. After another four washes with PBS, the mixture was resuspended in 500 μL PBS (0.1 mM, pH 6.0) and placed in a 50-mL centrifuge tube. As shown in Scheme 1, PPNs, after antibody-labeled MBs and the target of E. coli O157:H7, formed an immune−sandwich complex, and the signal output immunosensor can be acquired through this procedure. To exert the maximum catalytic effect of the PPNs, 10 μL NaOH (4 M) was added to destroy the nanocomposites to release the PtNPs. NH3-BH3 (1 mL, 25 mM) was injected through the pipe orifice of the centrifuge tube and the centrifuge tube was closed immediately with a rubber seal (double cover: the first layer was a common centrifugal tube cover with a circular hollow in the middle, and the second layer was a thin rubber layer). The needle of the digital H2 detector was inserted into the rubber-sealed tube (the center of the centrifugal pipe was hollowed out to expose the rubber layer). The generation of H2, which was utilized as a signal readout to detect target bacteria in the sealed centrifuge tube, was accurately measured by a portable H2 detector and the digitally displayed value was read within 3 min. E. coli O157:H7 was quantitatively and qualitatively analyzed. Parts per million (ppm) was the unit of measurement of H2 detection. P/P0 represented the ratio of target E. coli O157:H7 in the presence and absence of 104 CFU·mL−1.

Preparation of milk samples

For processing E. coli O157:H7 in real samples of milk, sterile milk was spiked by target pathogens. E. coli O157:H7 was diluted in PBS to a concentration of 102, 103, and 104 CFU·mL−1. Milk was purchased from a local supermarket. The immune sandwich structure with the target bacteria was used as described above to calculate the recovery (%), on the calibration plot.

Statistical analysis

All data were analyzed using GraphPad Prism 5 and expressed as the mean ± standard deviation. All parameters were measured at least three times to validate repeatability and reproducibility.

Results and discussion

Synthesis and characterization of PPNs

The PPNs were observed using SEM (Fig. 1a, b). According to the SEM characterization (Fig. S1), the PPNs had an average diameter of about 4 μm. This was consistent with the previous traditional organic–inorganic hybrid nanocomposites structure [23, 24].

Fig. 1
figure 1

SEM image of the PPNs at a scale of 10 μm (a) and 2 μm (b). The average size of the PPNs was about 4 μm

Full size image

AB is a promising chemical H2 storage material, which can be released by hydrolysis. The acquired PPNs were tested for catalytic hydrolysis of AB. The reaction was activated by addition of aqueous solution of AB into the reaction tube, comprising the PPN catalyst at room temperature. Figure 2a shows the time-dependent changes in H2 detector and the concentration of H2 produced as a function of target bacterium concentration. This change was directly converted to the H2 detector readout. The linear relation between the volume of PPNs in AB and the final concentration of H2 was apparent (Fig. 2b). This catalytic dissociation activity demonstrated that the synthesized PPNs had excellent AB decomposition efficiency and can be successfully used in H2 detector assays.

Fig. 2
figure 2

a Time-dependent changes in pressure with different concentrations of AB (PPN = 2.5 μL) and (b) linear relationship between H2 signal and volume of PPNs in the catalytic dissociation of AB (5 mg·mL−1) for 5 min at room temperature

Full size image

Optimization of conditions

Different conditions were investigated: (a) number of PtNPs; (b) concentration of Magainin I; (c) reaction temperature; (d) reaction incubation time; (e) pH value; and (f) concentration of NaOH.

The following experimental conditions were able to achieve the best results: (a) umber of PtNPs, 0.9 mg·mL−1; (b) concentration of Magainin I, 0.01 mg·mL−1; (c) reaction temperature, 35 °C; (d) reaction incubation time, 35 min; (e) pH value, 6.0 and (f) concentration of NaOH, 4 M. Detailed steps and figures can be found in the Electronic Supporting Material.

Sensitivity of the assay

To assess the responses of the immunosensor based on the hand-held H2 detector, various concentrations of E. coli O157:H7 were further tested. The H2 concentration inside the confined space gradually increased due to dehydrogenation of AB, and H2 was accurately measured by the hand-held detector. The H2 concentration increased gradually with concentration of E. coli O157:H7, and there was a linear relationship in the range of 10 to 104 CFU·mL−1 between H2 concentration and target E. coli O157:H7 concentration (Fig. 3). The LOD was 10 CFU·mL−1. The linear regression equation was Y = 0.4412X + 1.112 (R2 = 0.989) where Y is the relative aerogenesis intensity, X represents the common logarithm of the target bacteria concentration, and R2 is the correlation coefficient. The comparative LOD between our established immunosensor and other reported E. coli O157:H7 methods is shown in Table 1. In contrast to other methods, the immunosensors achieved excellent performance for detecting E. coli O157:H7, with an LOD of 10 CFU·mL−1. The lower LOD of the immunosensor may be attributed to the production of more H2 by PtNPs catalyzing AB and high sensitivity of the hand-held H2 detector. Compared with the traditional methods, such as fluorescence, electrochemistry and chemiluminescence, the approach did not rely on professional and expensive instruments.

Fig. 3
figure 3

a Sensitivity of the assay using a H2 detector for detecting E. coli O157:H7 in the range of 10–107 CFU·mL−1. b Linear relationship between the intensity of H2 signal and log of E. coli O157:H7 concentration. Each value represents the mean of three independent experiments (n = 3)

Full size image
Table 1 Comparison of different POC techniques for detection of E. coli O157:H7
Full size table

Specificity

To estimate the specificity of the method, the target E. coli O157:H7 and other nontarget pathogenic bacteria such as Salmonella sp., S. aureus, and L. monocytogenes were compared. The P/P0 ratios of other bacteria were similar to those of the control group, while the P/P0 ratio of E. coli O157:H7 was (Fig. 4). It is obvious from the results of these experiments that the method based on nanocomposites and hand-held H2 detector was specific to E. coli O157:H7.

Fig. 4
figure 4

Selectivity of the assay for detection of different targets

Full size image

Reproducibility of this method

To verify applicability and selectivity capability of our method for detection of E. coli O157:H7, we tested the assay in spiked commercial milk samples. Milk samples were mixed with PBS containing different concentration of E. coli O157:H7. The recovery and relative standard deviation (RSD) of E. coli O157:H7 added to milk samples were in the range of 86–126% and 3–27%, respectively (Table S1).

Conclusions

We have demonstrated a specific, sensitive and immunosensor system approach based on H2 generation in PPNs and a hand-held H2 detector for portable detection of pathogenic bacteria. E. coli O157:H7 was detected quantitatively through a direct hand-held H2 detector reading. These results show that the described assay is suitable for detection of a broad range of pathogens in real samples. The outstanding performance of the immunosensor approach was due to the high catalytic activity of PPNs in the dehydrogenation of AB, high H2 capacity of AB, and high sensitivity of the hand-held H2 detector. However, the practical application of this pathogenic bacteria detection strategy is still challenging, and strategies that are more efficient are required to simplify the experimental procedure. Nevertheless, the sensing platform POC detection of pathogenic bacteria based on the hand-held H2 detector has potential for the future in terms of portable application in low-resource settings.