**3. Results**

#### *3.1. Design and Preparation of MIPs*

The design and preparation of the MIPs is shown in Scheme 1. MIPs were prepared via a Pickering emulation polymerization, which was composed of a water phase and an oil phase. During imprinting, an NAC monomer containing enough free amino groups was first prepared by reacting acrylotl chloride with the amino groups in chitosan. Meanwhile, the CdTe QDs were functionalized with the carboxyl group from the TGA. NAC-QDs were obtained through the amide bond between the amino groups from the NAC and the carboxyl group from the CdTe QDs. The positively charged NAC-QD complex was easily bound with the negatively charged template *L. monocytogenes* via electrostatic interactions in the water phase [35]. The oil phase consisted of TRIM and DVB as co-cross-linkers and BPO as an initiator. After mixing the two phases, a stable emulation was obtained after vigorous hand shaking (see Figure 1), indicating the high efficiency of the self-assembled bacteria-NAC-QD network to construct a stable Pickering emulsion as a surfactant.

When polymerization was induced, the NAC-QD complex at the oil-water interface and the co-cross-linkers in the oil phase (TRIM and DVB) polymerized to form solid polymer beads; the template bacteria were located on the surface of the polymer beads. After removing the template bacteria, the imprinted sites were generated and were completely fitted with the template. Furthermore, the fluorescence intensity of MIPs decreased when the template was bound to the MIP beads; the intensity could recover after removing the template.

**Scheme 1.** Molecularly imprinted polymer (MIP) synthesis process (NAC, TRIM, DVB, BPO, and DMA relate to N-Acrylchitosan, trimethylolpropane trimethacrylate, divinylbenzene, benzoyl peroxide, and *<sup>N</sup>*,*<sup>N</sup>*-dimethylaniline).

**Figure 1.** Optical images of emulsions obtained by shaking the oil-water mixture of (**a**) MIPs and (**b**) non-imprinted polymers (NIPs).

#### *3.2. Characterization of Polymer Beads*

The polymer beads were carefully characterized to validate this preparation concept. Figure 2 displays TEM images of the CdTe QDs and the NAC-QD complex. The diameters of the NAC-QDs were much larger than those of the original QDs, indicating a successful introduction of the NAC and formation of the NAC-QD complex. The morphology and size of the MIPs and NIPs were characterized by SEM (see Figure 3a,b). Both the MIPs and NIPs exhibited a uniform spherical structure with a particle diameter of about 200 μm. Furthermore, the surface of the MIPs was rough and irregular due to their imprinted cavity; the NIPs were relatively smooth. These results were confirmed via magnified SEM images, which can be seen in Figure 3c,d. A large number of effective and tailor-made imprinted sites were found on the surface of the MIPs; there were none on the NIP surfaces.

FT-IR spectra were used to analyze the surface groups of the MIPs and NIPs. The remarkable peaks at 1633 cm<sup>−</sup><sup>1</sup> and 1463 cm<sup>−</sup><sup>1</sup> were attributed to C=C stretching vibrations and the benzene ring vibration of DVB, respectively [36,37]. Strong bands were observed near 1160 cm<sup>−</sup><sup>1</sup> and 1720 cm<sup>−</sup><sup>1</sup> which corresponded to the C–O–C stretch and C=O vibration from cross-linking TRIM [38]. Furthermore, there was no significant difference between the FT-IR spectra of the MIPs and NIPs, confirming that the template was removed completely.

Figure 3f shows the optical properties of the QDs, NAC-QDs, MIPs, and NIPs. Compared with the emission peak of the pure QDs, the spectra peak position of the NAC-QDs had a slight red-shift to ~340 nm due to the aggregation of QDs during the self-assembly of the NAC-QDs [39]. The fluorescence intensity of the NAC-QDs was much higher than those of the pure QDs because of the surface passivation of the QDs [40]. However, the comparative blue shift in the emission peaks of the MIPs and NIPs was obvious. This might be because of the reduction in the surface charge of the QDs because of electrostatic interactions, leading to a smaller Stokes shift.

**Figure 2.** Transmission electron microscopy (TEM) images of (**a**) quantum dots (QDs) and (**b**) the NAC-QD complex.

**Figure 3.** Characterization of polymers. Scanning electron microscopy (SEM) images of (**a**) MIPs and (**b**) NIPs which correspond to magnified SEM images of (**c**) MIPs and (**d**) NIPs, respectively; (**e**) Fourier transform infrared (FT-IR) spectra of MIPs and NIPs; and (**f**) fluorescence emission spectra of the QDs, NAC-QD complex, MIPs, and NIPs.

#### *3.3. Adsorption Performance of MIPs and NIPs*

#### 3.3.1. Bacterial Binding under Overloading Conditions

Equal amounts of MIPs and NIPs (5 mg) were incubated with 1 mL of *L. monocytogenes* suspension (OD600 = 0.1, almost 1.0 × 10<sup>8</sup> CFU mL−1) for 3 h. After removing the supernatant, the polymer beads were directly observed via SEM. Figure 4a,b show the surface morphology of the treated MIPs and NIPs with *L. monocytogenes*; the target bacteria are clearly visible on the surface of the polymer beads and the density of the cells absorbed on the MIPs was significantly greater than that of the NIPs due to the high adsorption efficiency of MIPs for *L. monocytogenes*.

**Figure 4.** SEM images of (**a**) MIPs and (**b**) NIPs after the adsorption of *L. monocytogenes* (1.0 × 10<sup>8</sup> CFU mL−1). CFU means the colony forming unit.

To further investigate the adsorption performance, equilibrium binding analysis, adsorption isotherms, and selectivity adsorption were also studied, as discussed below.

#### 3.3.2. Kinetics Adsorption of MIPs

The kinetic adsorption curves of the MIPs and NIPs were all evaluated to explore the adsorption rate (see Figure 5a). A high adsorption rate was performed, and the binding amount linearly increased with increasing contact time before reaching an adsorption equilibrium. At bacteria concentrations of 1.0 × 10<sup>5</sup> CFU mL−1, the equilibrium time of the polymer beads was almost 2 h, which could be shortened upon reducing the bacterial concentration. The fast adsorption rate could have occurred because the bacterial-NAC-QD complex in the water phase was used as the surfactant to construct the stable O/W Pickering emulsion during the imprinting process; thus, the template bacteria were located on the surface of the polymer beads. After removing the template, several e ffective and tailor-made imprinted sites remained on the MIP surface. These were easily accessible for the rebinding of template bacteria. In addition, a small reduction of adsorption amounts on the polymer beads occurred over an extended time. One possible reason for this observation is that the bacteria died through a nutrient deficiency in the PBS, resulting in a gradual decrease in *L. monocytogenes* viability.

**Figure 5.** (**a**) Kinetic data of MIPs and NIPs, (**b**) static isotherm curve of MIPs and NIPs, and (**c**) the selectivity adsorption of MIPs and NIPs on four bacteria mixing solutions.

#### 3.3.3. Adsorption Isotherm Analysis

To further investigate the binding performance of MIPs and NIPs, static adsorption data were collected at initial concentrations of *L. monocytogenes* ranging from 7.5 × 10<sup>1</sup> to 7.5 × 10<sup>5</sup> CFU mL−1. Figure 5b shows the adsorption isotherm curve: the binding capacity increased with increasing initial bacteria concentrations and the MIPs exhibited a significantly higher adsorption capacity than that of the NIPs across the tested concentration range (*p* < 0.05). These results were mainly due to the specific adsorption of imprinted sites, leading to superior adsorption of *L. monocytogenes*. A trace of bacteria was absorbed on the NIPs simply based on non-specific adsorption. At an initial concentration of 3.8 × 10<sup>4</sup> CFU mL−1, the adsorption capacities of the MIPs (355.6 CFU mg<sup>−</sup>1) were almost 4.57-fold that of the NIPs (77.8 CFU mg<sup>−</sup>1), while the imprinting factor (IF) value reduced to 2.04 when the initial concentration increased to 7.6 × 10<sup>5</sup> CFU mL−1. At high concentrations, the surface-bound bacteria were able to attract more bacteria cells to self-assemble and aggregated as a membrane-bound cluster on the surface polymer beads, leading to an increase in non-specific adsorption.

## 3.3.4. Selectivity Study

For visual observation of the selectivity of MIPs, *L. monocytogenes* and *S. aureus* were separately absorbed on MIPs and were directly observed via an optical microscope (Figure 6). Many *L. monocytogenes* were observed, indicating the high binding uptake for template bacteria. By contrast, only a few *S. aureus* were absorbed onto the MIPs, showing that the spherical *S. aureus* did not fit into the imprinted sites. Those that were generated were completely fitted with the rod *L. monocytogenes*. The low binding capacity for *S. aureus* was simply dependent on the non-specific adsorption.

**Figure 6.** The optical images of MIPs absorbing (**a**) *L. monocytogenes* and (**b**) *S. aureus*.

The interference experiment evaluated selective recognition of the template bacteria. Some commonly foodborne pathogens were selected as competitors, including *E. coli*, *Salmonella*, and *S. aureus*. Figure 5c shows the adsorption capacities of MIPs and NIPs on different bacteria. Under the same conditions, the binding capacities of the MIPs on *L. monocytogenes* and the other three bacteria were significantly different (*p* < 0.05). The imprinted polymers expressed the highest specific adsorption for *L. monocytogenes* with significantly lower adsorption of other bacteria. By contrast, the NIPs absorbed fewer template bacteria than the MIP beads because of the absence of selective imprinted recognition sites. Therefore, other than the electrostatic interactions, the recognition mechanism of the MIP beads was closely related to the complementary shape and size of binding sites that were produced by the template bacteria. The results indicate that the high recognition specificity of the imprinted polymer enabled the MIP beads to bind the target bacteria.

#### *3.4. Establishment of a Detection Method*

During MIP preparation, several imprinted sites remained on the surface of the MIPs containing QDs. As a result, the template bacteria were easily bound to the polymer beads and caused a significant change in the fluorescence intensity. To confirm the utility of these materials, the MIPs and NIPs were first incubated with 10<sup>3</sup> and 10<sup>5</sup> CFU mL−<sup>1</sup> concentrations of bacteria and the fluorescence intensity of the resulting polymer beads was directly observed using a fluorescence microscope under an ultraviolet lamp. Images are presented in Figure 7. The beads differed in brightness at different concentrations of *L. monocytogenes*, and the original MIPs and NIPs exhibited the brightest blue luminescence. With increasing bacteria concentrations ranging from 10<sup>3</sup> to 10<sup>5</sup> CFU mL−1, the color of the polymer beads became increasingly darker with a more pronounced change in the MIPs. Considering that microbial imprinting has not reached the level of precision that can be achieved by imprinting small molecules, the fluorescence MIPs sensor was applied for fast and qualitative detection of *L. monocytogenes* with a limit of detection (LOD) of 10<sup>3</sup> CFU mL−1.

**Figure 7.** Images of MIPs (for **a**–**c**, the concentrations of bacteria were 0, 103, and 10<sup>5</sup> CFU mL−1) and NIPs (for **d**–**f**, the concentrations of bacteria were 0, 103, and 10<sup>5</sup> CFU mL−1) adsorbing different concentrations of bacteria.

#### *3.5. Analysis in Real Samples*

Milk and pork samples were purchased from a local supermarket and were verified to be free of *L. monocytogenes* according to the National Standard GB/4789.30-2016. Thus, to evaluate the binding performance of the MIP beads in real samples, the milk and pork samples were analyzed by spiking them with two levels (10<sup>3</sup> CFU mL−<sup>1</sup> and 10<sup>5</sup> CFU mL−1). Figure 8 shows MIP beads irradiated under an ultraviolet lamp. Due to matrix effects, the original MIPs exhibited bright blue-green fluorescence. The color of the MIP beads clearly became much darker with increasing concentrations of *L. monocytogenes*, indicating that this method could be applied for the rapid detection of *L. monocytogenes* in real samples.

**Figure 8.** Images of MIPs in (**<sup>a</sup>**–**<sup>c</sup>**) milk and (**d**–**f**) pork. (for **a**–**c** and **d**–**f**, the concentrations of bacteria were 0, 103, and 10<sup>5</sup> CFU mL−1).
