Introduction

As it is well known, food-borne pathogens contamination poses a threat to human health worldwide through diet [1]. Some bacteria like Vibrio parahaemolyticus, Salmonella typhimurium are not allowed to be detected in food samples by the national standard of China [2]. Simultaneous determination of them is an important criterion for monitoring and reflecting the degree of bacteria contamination in aquaculture waters [3]. At present, many immunological and molecular biology technologies have been developed and applied in monitoring of pathogenic bacteria in aquaculture, such as enzyme-linked immunosorbent assay (ELISA) [4], polymerase chain reaction (PCR) [5], loop-mediated isothermal amplification (LAMP) [6], rolling circle amplification (RCA) [7], and gene microarray [8]. The above methods could bring reliable and accurate results. However, some of them are time-consuming, requiring expensive instruments or complex manipulations [9, 10]. Therefore, rapid identification and quantification of multiple food-borne pathogenic bacteria by convenient assays are essential to food safety monitoring.

In many test cases, the samples contaminated by target bacteria often accounted for a part of all samples. If all samples were brought back to the laboratory for precise quantification of various bacteria, it would lengthen the analytical time and be labor intensive [11]. Furthermore, it would reduce the timeliness for on-site supervision because the bacteria could grow geometrically within a few hours, thus would cause serious infection to aquatic products and lead to a large scale of deaths [12]. It was possible to solve the above problems by adopting dual-mode detection strategy. The strategy normally includes two methods to detect same samples [13, 14]. Generally, a rapid and convenient assay was used to judge positive samples containing food-borne pathogenic bacteria, and negative samples were not tested again. Then another method was used to precisely quantify various bacteria from positive samples. Thus, the dual-mode detection strategy could greatly improve the detection efficiency, timeliness and save time. For example, Wu [15] fabricated a bimodal aptasensor for the quantitative determination of V. parahaemolyticus through the combination of a fluorescent method and colorimetric method, based on platinum-coated gold nanorods (AuNR@Pt) probe. This analytical method could detect 10 CFU/mL by fluorescence mode and 75 CFU/mL V.P by colormetric mode. Wang et al. [16] has developed a fluorescent and colorimetric dual-channel assay for E. coli identification based on the OPD-Cu2+ sensing system. Both the fluorescent and colorimetric assay have a good sensitivity, reporting LODs of 100 CFU/mL and 44 CFU/mL respectively. According to the above studies, it could be found the colorimetry was very suitable for rapid screening because of visualization and no need for professional instruments [17,18,19]. It was also found that many reported dual-mode assays for food-borne pathogenic bacteria focused on single analyte measurements. However, there are many kinds of pathogenic bacteria that could coexist in aquaculture waters. Therefore, it is necessary to qualify or quantify all of them simultaneously in one test. Microfluidic chip (MC) is an ideal high-throughput platform for multiple targets separation and detection [20]. It normally takes several minutes to finish tests and the injecting volume could be as low as nanoliter or picoliter. Recently, some studies have revealed that MC could play an important role in the detection of multiple food-borne pathogenic bacteria. Zhang et al. [21] developed a rapid and sensitive strategy for the simultaneous detection of E. coli, L. monocytogenes, and S. Typhimurium in artificially contaminated milk using microchip capillary electrophoresis (MCE) coupled with multiplex polymerase chain reaction (PCR). The levels of detection were as low as 82 CFU/mL for E. coli, 68 CFU/mL for L. monocytogenes, and 85 CFU/mL for S. Typhimurium. Luo et al. [22] developed a specific method using an aptamer probe combining with the duplex PCR and MC-LIF to detect S. Typhimurium and P. aeruginosa, the limits of detection (S/N = 3) were 15 CFU/mL for S. Typhimurium and 5 CFU/mL for P. aeruginosa. All the studies showed the MC methods combining with aptamer probes could provide an ideal platform for rapid and simultaneous identification of various bacteria with high efficiency. However, due to the injection volume of MC is normally as small as pL or nL, its sensitivity may be limited. Thus, it usually takes several hours for culturing the pathogenic bacteria or amplifying the signals before detection.

The stir bar sorption extraction (SBSE) which was originally developed at the end of 1990s, is a kind of simple and solvent-free sample preparation technology with high enrichment capacity, detection sensitivity, and reproducibility [23]. Therefore, SBSE method could be used to directly extract pathogenic bacteria in large volume of aquaculture water with high concentration rate. It could achieve the goal of direct detection without a previous culture process, and shorten the analytical time. With the above assumptions, we hoped to develop a dual-mode aptasensor based on colorimetry and MC method together with SBSE enrichment. The assay could be used for effective on-site screening and accurate quantification of pathogenic bacteria.

The strategy was designed as following: Firstly, Vibrio parahaemolyticus (V.P) and Salmonella typhimurium (S.T) were employed to testify the proof-of-concept. The specific preparation process of the aptamer-streptavidin encoded probes (Apt-SAEs) and detection process were shown in Scheme 1. The aptamer modified with biotin (aptamer(V.P)-biotin or aptamer(S.T)-biotin) was conjugated with streptavidin modified with horseradish peroxidase (streptavidin-HRP) to generate the Apt-SAEs probes through the specific binding between biotin and streptavidin. After then, a stir bar self-assembled with 4-mercaptophenylboronic acid (stir bar@MPBA) was employed to nonspecifically extract V.P or S.T or both of them in aquatic waters added Apt-SAE probes. It was reported that phenylboric acid (PBA) could bind to polysaccharides on the cell wall of bacteria, which achieved specific enrichment [24, 25]. Then, the specific binding event among the bacteria, MPBA and Apt-SAE would result in the formation of sandwich structure. The HRP on the sandwich complex could catalyze the chromogenic reaction of 3,3′,5,5′-Tetramethylbenzidine (TMB) and H2O2 to emit the signal visualized by naked eyes [26, 27]. The color depth could be used to judge or semi-quantify the existence of V.P or S.T or both of them. For the colorless negative samples, the test would be finished. While for positive samples, the Apt-SAE probes corresponding to different bacteria on the stir bar were eluted, separated, and determined by MC. According to the relationship between the signal intensity and the standard concentrations, the concentrations of V.P and S.T in the sample could be accurately quantified. The as-prepared aptasensor was then employed to detect several aquaculture waters with satisfied results.

Scheme 1
scheme 1

The mechanism of dual-mode aptasensor based on colorimetry and microfluidic chip for simultaneous detection of multiple food-borne pathogenic bacteria.

Full size image

Experimental

Materials and reagents

The oligonucleotide sequences (listed in Table S1) were purchased from Shanghai Sangon Biological Engineering Technology and Services Co., Ltd. (Shanghai, China). PBS buffer (0.1 M phosphate buffered saline, 0.1 M KCl, pH 5–8) and binding buffer (1.0 M phosphate buffered saline, 10 mM EDTA, 250 mM NaCl, 20 mM MgCl2, and pH 6–8) were prepared by standard methods. 3,3′,5,5′-Tetramethylbenzidine (TMB), 6-Mercapto-1-hexanol (MCH), HRP-conjugated Streptavidin, 4-mercaptophenylboronic acid (MPBA), hydrogen tetrachloroaurate (III) tetrahydrate (HAuCl4), sodium citrate were from Sangon Biotechn Co., Ltd. (Shanghai, China). (3-mercaptopropyl)-Tri-Methoxysilane (MPTMS) was purchased from Signal-Aldrich Co., Ltd. (Milan, Italy). Vibrio parahaemolyticus (V. parahaemolyticus, ATCC 17802), Salmonella typhimurium (S. Typhimurium, ATCC 14028), E. coli (E. coli, ATCC 25922), Staphylococcus aureus (S. aureus, ATCC 29213), and Listeria monocytogenes (L. M, ATCC 13932) were obtained from Luwei Microbial Sci. and Tech. Co., Ltd. (Shanghai, China). The bare gold bar (99.99%, 25.0 cm in length and 0.2 mm in diameter) was obtained from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, China). SYBR Gold (10000×) was obtained from Shimadzu Co, Ltd. (Tokyo, Japan). All of other chemicals were analytical grade and used without further purification. Double-distilled water was used throughout all this work.

Scanning electron micrographs (SEM) were obtained by using a S3400N scanning electron microscope from Hitachi Co, Ltd. (Tokyo, Japan). The UV-vis spectra were recorded by a UV-1800 spectrophotometer Shimadzu Co, Ltd. (Tokyo, Japan). MultiNA 202 system was purchased from Shimadzu Co, Ltd. (Tokyo, Japan).

Preparation of Apt-SAEs encoded probes

The Apt-SAE encoded probes were prepared in advance, 50 mL PBS buffer (pH 7.4) containing 10 μM aptamer-biotin and 1.0 μM streptavidin-HRP solution was shaken on a shaker (120 rpm) at room temperature for 2 h to obtain Apt-SAE (1.0 μM).

Preparation of the stir bar

Gold nanoparticles (AuNPs) were prepared by the citrate-reduction method [28]. Stir bar modified with AuNPs (stir bar@AuNPs) was synthesized according to literature [29]. The specific steps were shown in the supplementary S1. Subsequently, the 25.0 cm stir bar@AuNPs was incubated with MPBA (10 mM, 10 mL) solution at room temperature for 2 h. After the reaction, it was rinsed in PBS buffer (pH 7.4) three times, and then immersed in MCH (1.0 mM, 10 mL) solution for 2 h to block the unspecific adsorption sites. After the stir bar was rinsed three times with PBS, the preparation of the stir bar@AuNPs was completed.

Colorimetric measurement

Binding buffer (50 mL, pH 8.0) and Apt-SAEs (50 mL, 1.0 μM) were added into a flask, and 400 mL aquaculture water containing the V.P and S.T (10, 1 × 102, 1 × 103, 1 × 104, 1 × 105, 1 × 106, and 1 × 107 CFU/mL) were added to the solution. During the extraction process, the stir bar wrapped around a glass rod was immersed in water samples (Fig. S1A), and stirred (100 rpm) at room temperature for 15 min. Then the stir bar was taken out and rinsed with PBS buffer (pH 7.4) three times to remove the unconjugated V.P, S.T and Apt-SAEs. Therefore, a three-sandwich structure of Apt-SAE, V.P or S.T, and stir bar was obtained. Finally, the stir bar with three-sandwich structure was immersed in 400 μL TMB and H2O2 solution to trigger the catalytic chromogenic reaction, and then its color change could be observed after 5 min. If its color turned blue, it meant that the sample contained at least one kind of bacteria.

Quantitative detection of various bacteria by MC

For the positive samples, the stir bar with three-sandwich structure was immersed into the PBS buffer (400 μL, pH 6.0) and eluted for 20 min by shaking (100 rpm). The Apt-SAE encoded probes on the stir bar would release into the eluant. 150 pL of the eluent was injected into the sampling port of microfluidic chip (Fig. S1) by electrokinetic injection (7 s × 10 kV). The Apt-SAE in the eluate has fluorescence by adding (1×) SYBR gold to bind with DNA strands. The excitation wavelength was 495 nm and the emission wavelength was 537 nm. Then the samples containing eluted Apt-SAEs were separated and determined by the chip using the standard program providing by the Shimazu Co.Ltd. of MCE 202 MultiNA. The peaks at 65 s and 71 s were correspondent with Apt(V.P)-SAE and Apt(S.T)-SAE, respectively. The intensity of peak was employed for quantifying the target bacteria. After the analysis, the micro-channels of the MC were washed three times with water and the matching eluting buffer provided by Shimazu Co., Ltd.

Detection of V.P and S.T in aquatic samples by the dual-mode aptasensor

The 400 mL water sample was collected from the area to be tested. The water samples were filtered through a 140 μm membrane to remove solid particles such as stone and dirt. Fifty milliliters binding buffer (pH 8.0, 1.0 M PBS) and 50 mL Apt-SAEs (1.0 μM) were added to the sample. After 15 min of incubation, the stir bar was eluted in 400 μL PBS buffer (100 mM, pH 6.0) for 20 min. Then, the sample was injected (sampling volume is 150 pL) and detected by the MultiNA 202 system, and the concentrations of V.P and S.T were calculated according to the signal intensity of Apt(V.P)-SAE (65 s) and Apt(S.T)-SAE (71 s), respectively.

Results and discussion

Characterization of SBSE method

The morphology of V.P and S.T were characterized by scanning electron microscope (SEM) in Fig. 1A and B, respectively. As could be seen the length of V.P was about 2.5 μm and the width of it was 550 nm. And the length and the width of S.T was about 1.8 μm and 850 nm, respectively. From Fig. 1C, it could be found the surface of the stir bar was smooth, while the morphology of stir bar@AuNPs was rough with many spherical particles (Fig. 1D). It meant many AuNPs were successfully immobilized on the stir bar, which could provide many active sites for 4-mercaptophenylboronic acid. In addition, Fig. 1E depicted the surface of the stir bar which successfully captured V.P and S.T. The S element was detected by energy-dispersive spectroscopy (EDS) on stir bar@AuNPs-MPBA, which proved that MPBA has been successfully modified on the stir bar (Fig. 1F).

Fig. 1
figure 1

The SEM images of V.P (A), S.T (B), stir bar (C), stir bar@AuNPs (D), stir bar@ AuNPs-MPBA capturing V.P and S.T (E), EDS of stir bar@AuNPs-MPBA (F)

Full size image

Feasibility verification of the dual-mode aptasensor

Figure 2A and B showed the feasibility of dual-mode aptasensor for simultaneous detection of V.P and S.T. As shown in Fig. 2A, distinctive bands of aptamer (V.P)-biotin (1.0 μM) appeared in lane 1. As could be seen from the figure, the band of V.P (105 CFU/mL) was almost blank (lane 2). When streptavidin-HRP (0.2 μM, 50 μL) was mixed with aptamer (V.P)-biotin (2.0 μM, 50 μL), a new lagged band appeared in lane 3. It meant that the Apt(V.P)-SAE probes were assembled successfully. Lane 4 showed the complex of Apt(V.P)-SAE and V.P. When Apt(V.P)-SAE was incubated with VP, the darkness of the band representing Apt(V.P)-SAE became lighter comparing with lane 2, and a new band appeared above the Apt(V.P)-SAE band. These proved that the Apt(V.P)-SAE was combined with V.P successfully. A similar phenomenon related to S.T also appeared in Fig. 2B. All these illustrated the feasibility of MC detection for V.P and S.T.

Fig. 2.
figure 2

A The electropherograms of Apt(V.P)-SAE and Apt(V.P)-SAE@V.P; B The electropherograms of Apt(S.T)-SAE and Apt(S.T)-SAE@S.T; C the amplification effect of AuNPs and Apt-SAE; D The feasibility of colorimetric visualization detection; E The feasibility of combining bacteria and Apt-SAE (excitation wavelength: 495 nm, emission wavelength: 537 nm)

Full size image

At the same time, we studied the amplification effect of AuNPs on the stir bar and Apt-SAE encoded probes by MC. Gold nanoparticles (AuNPs) are widely used for color development with stable properties and controllable size. In addition, gold nanoparticles can also have strong bond cooperation with sulfhydryl and other groups [30], so 4-mercaptophenylboronic acid (MPBA) can be easily self-assembled on the surface of the stir bar to achieve functionalization. In Fig. 2C, the red curve depicted the sample containing V.P (107 CFU/mL) and S.T (107 CFU/mL) using aptamer and stir bar without AuNPs. The green curve reflected the results using aptamer and stir bar modified with AuNPs. The black curve depicted Apt-SAE and the stir bar modified with AuNPs. It could be clearly seen the signal increased gradually. The amplification factor was 10 times. The reasons were as following: AuNPs could obviously increase the specific area of stir bar and the amount of MPBA modified on it, which could extract and enrich low concentration of bacteria onto the stir bar. Thus, the extraction capacity of target bacteria could be increased largely compared with the stir bar without AuNPs. Moreover, it was well known that one streptavidin could bind up to four aptamer-biotin. Thus, one Apt-SAE probe was supposed to contain four aptamer, which was more capable of binding the target bacteria.

Figure 2D showed the feasibility of colorimetric detection. There were almost no absorption peaks from 400 nm to 700 nm in UV-vis spectrum when no target bacteria (black curve) or 107 CFU/mL other bacteria (Listeria monocytogenes) existed (green curve). This meant that even if many bacteria could be adsorbed on the phenylboronic acid modified stir bar, the aptamer-streptavidin encoded probes (Apt-SAEs) could only combine with the target bacteria other than other interfering bacteria because the aptamer could only exhibit high affinity toward the target analyte. When 105 CFU/mL V.P and S.T were incubated with the stir bar separately, the visible blue color could be observed by naked eyes and a strong characteristic absorption peak appeared at 650 nm (purple and blue curve). The reason was that the HRP on Apt-SAEs could catalyze the colorless 3,3′,5,5′-converting tetramethylbenzidine (TMB) into blue oxidized TMB (ox TMB) in the presence of H2O2 [31]. In addition, there was a stronger UV peak when 105 CFU/mL V.P and S.T were simultaneously incubated with the probes than that from only 105 CFU/mL V.P or S.T. The intensity of both V.P and S.T were almost same as the sum of their individual ones. The recovery was 95%. This meant the UV-vis spectrum could quantify the analytes.

Figure 2E showed the feasibility for simultaneous detection of V.P and S.T by the MC. The Apt(V.P)-SAE (0.1 μM) and Apt(S.T)-SAE (0.1 μM) were incubated with target bacteria (105 CFU/mL) separately in binding buffer (pH 8.0). It could be found that there were two obvious peaks in curve-a corresponding to Apt(V.P)-SAE while the concentration ratio of aptamer-biotin and streptavidin-HRP was 10 to 1. This was because a single streptavidin could bind one to four biotins, the multiple peaks represented different species of Apt(V.P)-SAE. The latter peak represented the 1:4 (streptavidin-HRP: aptamer-biotin) specie. The previous one maybe represented the 1:1 to 1:3 (streptavidin-HRP: aptamer-biotin) species. Even if 10 folds higher of aptamer-biotin were incubated with the streptavidin-HRP, there was also other species existed except 1:4 species. This could be due to that there were steric effects between aptamers while binding with the streptavidin. In the presence of V.P, the Apt(V.P)-SAE and V.P could form Apt(V.P)-SAE@V.P complex. So there was a new peak appeared which was corresponding to Apt(V.P)-SAE@V.P in curve-b and the peak intensity of Apt(V.P)-SAE was significantly reduced. Curve-c and curve-d showed the similar results of S.T. All of the above results proved that the MC could simultaneously and specifically detect V.P and S.T.

Optimization of the experimental conditions

The experimental parameters including time of stir bar sorption extraction (SBSE) and desorption for Apt-SAEs on the stir bar, extraction and desorption pH for SBSE, concentration of Mg2+ were optimized to achieve the best analytical performance. In the assay, the time for color reaction of TMB-H2O2 system was about 5 min, and that for MC was 3 min. Therefore, the pretreatment time by SBSE was the crucial factor for obtaining the largest signal response and shortest analytical time. 105 CFU/mL V.P or S.T was used for the following experiment, and the number of experiments (n = 5). From Fig. 3A, it could be seen that the colorimetric signal increased with the extraction time, and remained stable after 15 min. Thus, the optimal extraction time of SBSE was 15 min. The desorption time of Apt-SAEs from stir bar was also a key factor influencing the detection time by MC. After the stir bar extracted V.P or S.T together with the Apt-SAE encoded probes, the adsorbed probes were eluted by washing with PBS buffer (pH 6.0) and detected by MC. It could be seen that the MC signal increased with time and reached to a plateau at 20 min in Fig. 3B. As is well known that pH could affect the surface charge of bacteria and aptamer, then influenced the binding capacity of Apt-SAE toward the target bacteria. In this study, the pH values ranging from 6.0 to 8.0 were examined for extraction and detection by colorimetric method. The signal intensity reached to the maximum at pH 8.0 (Fig. 3C). The results were in consonance with that from Wang’s group who employed pH 8.0 for incubating the aptamer with bacteria [32]. We employed pH 8.0 in further analysis. The desorption pH of Apt-SAE binding with the target bacteria on the stir bar was also optimized. The pH values of elution PBS buffer ranging from 5.0 to 8.0 were investigated. It could be found that pH 6.0 could completely elute the Apt-SAEs from the stir bar (Fig. 3D). It has been reported that the covalent bond of phenylboronic acid and 1,2–diol compound could be decomposed at pH 6.0 [33]. Thus, the MBPA on the stir bar maybe released the adsorbed bacteria at pH 6.0 and then the Apt-SAEs were released into supernatant.

Fig. 3
figure 3

The samples contained 105 CFU/mL V.P or S.T, n = 5 A the incubation time for SBSE, B the desorption time for target bacteria, C the pH of SBSE, D the pH of desorption, E concentration of Mg2+, the experimental conditions were as follows: binding buffer (pH 8.0, 1 M PBS), eluent (pH 6.0, 100 mM PBS), incubation time (15 min), desorption time (20 min), and concentration of Mg2+ (2.0 mM)

Full size image

Mg2+ has been reported to play a prominent role on the binding capacity of aptamer to its target [34]. During the process of Apt-SAEs combining with the target bacteria, Mg2+ aids in the formation of stable tertiary structures of aptamer, which is crucial for selective interaction between aptamer and their target [32]. In this study, we increased the Mg2+ concentration of sample buffer from 0 mM to 5.0 mM. As shown in Fig. 3E, it revealed that the intensity reached to the maximum at 2.0 mM.

Cross reactivity experiment and signal amplification

Under the optimum conditions, the cross reactivity of S.T and V.P in same samples was further studied. Each Apt-SAE probe own four aptamer DNA which could combine with SYBR Gold for signal development. The excitation wavelength was 495 nm and the emission wavelength was 537 nm. Because different Apt-SAEs embedded different length of aptamer, they could be separated by the microfluidic chip which was installed with polyacrylamide gel electrophoresis channels. Therefore, the Apt-SAEs corresponding to different target bacteria could be separated and detected by the microfluidic chip for simultaneously qualifying and quantifying different bacteria. As shown in Fig. 4, when the concentrations of V.P (102 CFU/mL) remained unchanged together with different concentrations of S.T, the signal intensity of S.T was significantly increased from 102 to 106 CFU/mL, while the signal of V.P remained unchanged. Similarly, when the concentrations of S.T (102 CFU/mL) remained unchanged and different concentrations of V.P were added, the signal of V.P increased from 102 to 106 CFU/mL while that of S.T was stable. Upon the concentrations of S.T and V.P were increased, both signals enhanced accordingly. The above results indicated that there would be no cross-reactivity of two target bacteria, and the multiple assay could be employed.

Fig. 4
figure 4

The cross-reactivity of V.P and S.T at different concentrations (102 and 106 CFU/mL) by MC (A), the enrichment factor by SBSE (B), the interference of other substances (105 CFU/mL Staphylococcus aureus, 105 CFU/mL Listeria monocytogenes, and 105 CFU/mL E. coli, 0.16 g/L starch, 0.31 g/L agar) toward V.P (C), and toward S.T (D)

Full size image

The enrichment factor of SBSE was testified. The stir bar was used to enrich 400 mL sample containing V.P (104 CFU/mL) then eluted by PBS buffer (0.4 mL, pH 6.0). It could be seen from Fig. 4B, the colorimetric signal was as same as that by ELISA for sample containing 107 CFU/mL V.P. All these probed that the enrichment factor by SBSE was about 1000 times, which could obviously increase the sensitivity of the assay.

The phenylboric acid on the stir bar was combined with the polysaccharides on the cell wall of bacterial, which realized the capture of the bacteria. However, other nontarget bacteria and polysaccharides in actual water could also be combined with phenylboronic acid, which would interfere with the experimental results. As shown in Fig. 4C, the black curve depicted the MC signal intensity of 105 CFU/mL V.P in the absence of other impurities. The red curve depicted the MC signal intensity of 105 CFU/mL V.P in the presence of other substances (105 CFU/mL Staphylococcus aureus, 105 CFU/mL Listeria monocytogenes, and 105 CFU/mL E. coli, 0.16 g/L starch, 0.31 g/L agar). It is not hard to see that the signal intensity of the two cases were almost the same. The conclusion of Fig. 4D was similar to that of Fig. 4C. It could be concluded that other bacteria and polysaccharides in the environment would not affect the testing results.

The analytical performance for V.P and S.T

Under the optimum conditions, we evaluated the analytical performance of the dual-mode aptasensor toward S.T and V.P. S.T and V.P were diluted to different concentrations with PBS buffer (1 × 102, 1 × 103, 1 × 104, 1 × 105, 1 × 106 and 1 × 107 CFU/mL), and then conducted further analysis. As showed in Fig. 5A and B, the UV-vis absorption peak intensity at 650 nm increased corresponding to the concentrations of V.P and S.T. The visible color could be observed (the insets in Fig. 5A and B). Therefore, the aptasensor could realize self-quantification for concentrations of V.P and S.T by naked eyes through the depth of color.

Fig. 5
figure 5

The colorimetric detection of A V.P and B S.T range from 102 to107 CFU/mL; C The simultaneous detection of V.P and S.T range from 102 to 107 CFU/mL; D The calibration curve of the intensity for determination of V.P and S.T (excitation/emission wavelength: 495/537 nm), n = 5

Full size image

Figure 5C showed the intensity of MC increased with an increase in V.P and S.T concentrations. The Apt-SAE intensity exhibited good linear relationship with the logarithm of V.P and S.T ranging from 1 × 102 to 1 × 107 CFU/mL. However, when the concentration of target bacteria in the sample was greater than 107 CFU/mL, the MC signal intensity tended to be constant. It was indicated that the saturated extraction amount of VP on the stir bar is 1.6 × 1010 CFU/bar, and saturated extraction amount of ST is 2.1 × 1010 CFU/bar, respectively. The linear equation of V.P was y = 13.45Cv.p-19.80 (R2 = 0.9936), the limit of detection (LOD) was evaluated to be 35 CFU/mL (S/N = 3). And the linear equation of S.T was y = 14.27Cs.t-16.28 (R2 = 0.9951), the LOD was also 35 CFU/mL (S/N = 3). In addition, Table 1 listed some recent methods for detecting S.T or V.P. Compared with all the methods, the dual-mode aptasensor showed higher sensitivity and a wider detection range. It could realize the detection of multiple targets with lower LOD. Moreover, there were no report about dual-mode aptasensor being suitable for both on-site and lab analysis of target bacteria up to now.

Table 1 The comparison of methods to detect S.T and V.P
Full size table

Evaluation of specific experiments

In order to evaluate the stability and specificity of the dual-mode aptasensor, several other pathogenic bacteria (Staphylococcus aureus, Listeria monocytogenes, and E. coli) were chosen for interference experiments. The concentrations of all four bacteria were 107 CFU/mL. Figure 6A showed V.P (105 CFU/mL) exhibited a strong signal that was consistent with its own existence, while the response of other pathogenic bacteria was negligible. The specificity experiment of S.T (105 CFU/mL) was shown in Fig. 6B. The result also demonstrated that other bacteria did not interfere the measurement of S.T. Accordingly, the dual-mode aptasensor showed satisfied specificity.

Fig. 6
figure 6

Specificity experiments of 105 CFU/mL A V.P and B S.T with the existence of 107 CFU/mL Staphylococcus aureus, Listeria monocytogenes, and E. coli (n = 3)

Full size image

Application in aquatic samples

To further evaluate the feasibility and practical application performance of the dual-mode aptasensor by using real water samples, all the samples were spiked with different concentrations of V.P and S.T (102 CFU/mL, 105 CFU/mL), respectively. And the recovery was tested in spiked samples. The experimental data were listed in Tables 2 and 3, the recoveries of samples spiked V.P were in the range of 95.9–102% with the RSD values at 3.2–5.5%. And the recoveries of spiked S.T were in the range of 97.4–102% with the RSD values at 2.7–6.2%. These results showed that the proposed method by detecting V.P and S.T simultaneously in water samples was effective.

Table 2 Detection of V.P in real samples by the dual-mode aptasensor (n = 5)
Full size table
Table 3 Detection of S.T in real samples by the dual-mode aptasensor (n   =   5)
Full size table

Conclusion

In this study, a dual-mode aptasensor based on colorimetry and microfluidic chip was firstly constructed together with SBSE for on-site visual analysis and laboratory confirmation of multiple food-borne pathogenic bacteria (Vibrio parahaemolyticus and Salmonella typhimurium). Compared with the related reported literature, this strategy showed several merits: Firstly, the stir bar was employed for extraction of target bacteria with high enrichment times. Thus, there was no requirement for long-time culture before detection, which greatly saved time and improved the detection efficiency. Secondarily, the aptamer-streptavidin encoded probes (Apt-SAEs) achieved dual transduction of colorimetric and microfluidic chip signals by the analytes, and exhibited multiple amplification effects, thus, improved the detection sensitivity. Thirdly, this strategy combined the advantages of both colorimetry and MC, which not only realized rapid on-site screening, but also realized the simultaneous quantification of multiple bacteria. Finally, this strategy had the potential to detect other food-borne pathogenic bacteria in food by using corresponding aptamer, which opened up a new path for the sensitive and rapid detection of multiple bacteria in foods. But there were some limitations in strategy. The total time for pretreatment, incubation, desorption process, and detection of samples was as long as 1 h. Moreover, the MPBA modified stir bar could adsorb different kinds of bacteria which would compete with target bacteria. Thus, it would influence further enhancement of sensitivity. All these should be resolved in our future work.