3.1 Characterizations of LHPs and β-CD@AuNPs
Typically, N, N-dimethylformamide (DMF) is a good solvent for LHPs, our group has studied the AIE feature of LHPs in DMF/water mixtures (Qu, et al. 2021). As shown in Fig. 1a, LHPs are highly dispersed in nano-scale irregular shapes in Transmission electron microscopy (TEM) image. However, LHPs aggregate significantly in water, and some micrometer-sized sticks are observed in the TEM image (Fig. 1b). Meanwhile, the emission of LHPs in DMF is extremely low, while the aggregated ones in water emit strong fluorescence (Fig. 1c). The AIE behavior of LHPs in water is supported by these results.
Figure 1
On the other hand, through the one-step colloidal synthesis method, the β-CD@AuNPs with wine-red color are prepared (Zhao, et al. 2016; Zhang, et al. 2019). The Fourier transform infrared spectroscopy (FT-IR) spectra (Fig. 2a) display the characterized surface groups of β-CD@AuNPs and β-CD. The peak of β-CD at 1640 cm− 1 corresponds to the stretching vibrations of O-C-O (Zhang, et al. 2019). However, its blue shifts to 1624 cm− 1 in the FT-IR spectrum of β-CD@AuNPs. This reason is that the hydroxy groups in β-CD reduce Au3+ into Au0, and the hydroxy groups themselves are oxidized to carboxyl groups (Zhao, et al. 2016). Furthermore, the peaks at 3430 and 1160 cm− 1 are the stretching vibrations of O-H and C-O in the FT-IR spectra of β-CD and β-CD@AuNPs. Additionally, these nanoparticles are mostly spherical particles and highly dispersed in solution (Fig. 2b). The diameter of β-CD@AuNPs is mainly around 21 nm. These results illustrate that β-CD@AuNPs are successfully synthesized with good solubility in water.
Figure 2
3.2 The interaction between β-CD@AuNPs and LHPs
A high AIE is shown by the aggregations of LHPs. However, the AIE is quenched with an increasing concentration of β-CD@AuNPs. A good linear relationship (R2 = 0.998) between the quenching efficiency and the concentrations of β-CD@AuNPs is from 5.2 to 10 pM (Fig. 3b). As well known, a variety of substances, such as CQDs (Li, et al. 2019), rhodamine B (RB) (Zhao, et al. 2016), CuInS2 quantum dots (Hu, et al. 2017), and so on, can be quenched by β-CD@AuNPs. The above reports use the concentrations of the β-CD@AuNPs in the nanomolar range. However, only 10 pM β-CD@AuNPs can quench the AIE of LHPs by about 87%. This phenomenon illustrates the sensitive response of LHPs to β-CD@AuNPs.
The quenching mechanism of LHPs by β-CD@AuNPs is investigated. The surface plasmon peak of β-CD@AuNPs in the UV-vis spectrum (Fig. 3c) appears at 533 nm. This result overlaps well with the emission of the LHPs. Hence, the inner filter effect (IFE) or Förster resonance energy transfer (FRET) may be causing the effective fluorescence quenching of LHPs by β-CD@AuNPs (Shang, et al. 2009; Jares-Eruman, et al. 1997). Different from IFE, the fluorescence lifetime decreases in FRET (Koushik and Vogel 2008). Figure 3d suggests that after adding β-CD@AuNPs, the fluorescence decay time of LHPs is from 1.86 to 0.64 ns. As a result, the fluorescence quenching of LHPs by β-CD@AuNPs is mostly based on FRET rather than IFE. Moreover, FRET is distance-dependent and occurs when the donor and acceptor are brought close to one another (Varghese, et al. 2010). The TEM image of LHPs after adding β-CD@AuNPs (Fig. S1) shows that many β-CD@AuNPs are attached to the surface of LHPs. It is reasonable to speculate that by host-guest interactions the ligand dodecanamine of LHPs may enter into the macrocyclic cavity of β-CD. Thus, this close distance provides a favorable condition for FRET. Additionally, the FRET efficiency can be assessed by Eq. (1) (Biju, et al. 2006; Qu, et al. 2021):
where τ0 (τ0 = 1.86 ns) and τ (τ = 0.64 ns) are the fluorescence lifetimes of the LHPs in the absence and presence of the β-CD@AuNPs at the concentration of 12.4 pM. The result indicates that approximately 66% of the quenching is attributed to the FRET by calculating.
Figure 3
To further investigate the remaining quenching effect, static quenching effect (SQE) and dynamic quenching effect (DQE) are analyzed according to the Stern-Volmer Eq. (2) (Xiang, et al. 2007):
$${\text{F}}_{\text{0}}\text{/F=1+}{\text{K}}_{\text{sv}}\left[\text{Q}\right]$$
2
where F0, F are the fluorescence intensities of LHPs in the absence and presence of β-CD@AuNPs, respectively; Ksv is the Stern–Volmer quenching constant, and [Q] is the concentration of β-CD@AuNPs. At different temperatures, the concentration of β-CD@AuNPs remains linear with the fluorescence intensity ratios (\({\text{F}}_{\text{0}}\text{/F}\)) (Fig. S2). One type of quenching mechanism (SQE or DQE) is predominating by the liner Stern–Volmer plots (\({\text{F}}_{\text{0}}\text{/F}\) versus [Q]) revealing (Xiang, et al. 2007). During the excited state, DQE is ascribed to the collisional encounters between the quencher and the fluorophore. Higher temperature promotes molecular diffusion. This phenomenon leads to an increase in the Ksv. For SQE, between the fluorophore and quencher forms a nonfluorescent ground-state complex. The temperature increment causes a decrease in the stability of the complex, leading to a smaller value of Ksv (Legrand, et al. 2021). The obtained Ksv values from Stern-Volmer slopes in Fig. S2 at 277, 288, 298, and 313 K are 5.435 ⋅ 1010, 4.682 ⋅ 1010, 4.315 ⋅ 1010 and 3.698 ⋅ 1010 M− 1, respectively. With increasing temperature, these Ksv values show a tendency that is gradually decreasing. The result proves that another mechanism of the β-CD@AuNPs quenching LHPs is DQE instead of SQE. The ground-state complexes formed between β-CD@AuNPs and LHPs are validated by UV-vis absorption spectra. The characteristic absorption peaks of LHPs and the surface plasmon peak of β-CD@AuNPs disappear when β-CD@AuNPs and LHPs are mixed (Fig. S3). Therefore, β-CD@AuNPs induce significant fluorescence quenching of LHPs via a combination of FRET and SQE.
3.3 Detection of cholesterol
As shown in Fig. 4, there is no reaction between cholesterol and LHPs, because cholesterol does not affect the AIE of LHPs. However, β-CD@AuNPs quench effectively the AIE of LHPs. When cholesterol and β-CD@AuNPs are preincubated, cholesterol can enter the macrocyclic cavity due to the strong hydrophobicity of the cholesterol structure (Zhao, et al. 2016). Thus, the ligand dodecylamine of LHPs cannot enter the cavities of β-CD, resulting in the fluorescence recovery of LHPs. Under UV lamp at 365 nm (inset in Fig. 4), the above phenomenon is also observed. Scheme 1 illustrates the principle of detection of cholesterol based on this system.
Figure 4
Scheme 1
Experimental conditions have been optimized for the sensitive detection of cholesterol, including addition order and incubation time (Fig. S4 and S5). Under the optimized conditions, the AIE of LHPs is gradually restored with increasing concentration of cholesterol in the presence of β-CD@AuNPs (Fig. 5a). A good linear range of cholesterol is obtained from 0.35 to 20 nM (Fig. 5b). The LOD is determined to be as low as 0.12 nM. The expression which is \(\text{LOD=3σ/K}\) is used to calculate LOD, where σ is the standard deviation for the blank solution (n = 10), and K is the slope of the calibration curve (Wu, et al. 2007; Zi, et al. 2021). As summarized in Table 1, the LOD of this work is better than most reported methods involving enzyme and non-enzyme. The sensitive response of LHPs to β-CD@AuNPs may be causing the AIE of LHPs effectively quenched by β-CD@AuNPs, which the concentration of β-CD@AuNPs is only a few pM. Thus, very few amounts of cholesterol can be detected.
Table 1
Comparison of various methods for the detection of cholesterol
Method
|
System
|
Linear range
|
LOD
|
Refs.
|
Electrochemistry
|
ChOx / AuNPs
|
0.5–48 µM
|
0.26 µM
|
[27]
|
Colorimetry
|
ChOx / horseradish peroxidase
|
0–40 µM
|
0.6 µM
|
[28]
|
Colorimetry
|
ChOx / Prussian blue
|
4-100 µM
|
3 µM
|
[29]
|
Fluorescence
|
ChOx / CdTe CQDs
|
5-100 nM
|
0.89 nM
|
[30]
|
Fluorescence
|
ChOx / Ag nanocluster
|
0.06-15 µM
|
0.03 µM
|
[31]
|
Fluorescence
|
CQDs / β-CD@AuNPs
|
10–210 µM
|
0.34 µM
|
[8]
|
Fluorescence
|
β-CD@AuNPs / RB
|
0.32–4.80 µM
|
0.15 µM
|
[11]
|
Fluorescence
|
β-CD functionalized CQDs
|
3.5–110 µM
|
0.7 µM
|
[32]
|
Fluorescence
|
LHPs / β-CD@AuNPs
|
0.35-20 nM
|
0.12 nM
|
This work
|
Figure 5
Table 1
3.4 Selectivity
It is important to have good selectivity for real sample detection. Therefore, β-sitosterol (a kind of phytosterols), lactose, palmitic acid, vitamin B, ascorbic acid, vitamin D, oleinic acid, K+, Na+, Mg2+, Ca2+, and Zn2+ are selected as potential interference substances. As shown in Fig. 6, AIE recovery of LHPs may not be induced by these substances. In particular, ascorbic acid does not interfere with the detection of cholesterol, because there is no redox reaction in this strategy.
Nowadays, phytosterols are often added to milk powder, because phytosterols can inhibit the body absorption of cholesterol, reducing the risk of cardiovascular disease (Gylling, et al. 2020). Thus, it is important to determine cholesterol even when cholesterol and phytosterols co-exist in food. Up to now, no reports are available on the accurate detection of cholesterol in the presence of phytosterols. Herein, cholesterol can be detected when cholesterol coexists with β-sitosterol, suggesting the importance of this method. The good selectivity may be related to the different affinities of cholesterol and β-sitosterol to β-CD. The fluorescence responses of LHPs to cholesterol, dodecylamine, and β-sitosterol have been studied in the presence of β-CD@AuNPs. The fluorescence of LHPs can be restored by cholesterol and dodecylamine rather than β-sitosterol (Fig. S6). The structure of β-sitosterol has one more ethyl group than cholesterol (Fig. S7), so it is reasonable to speculate that the ethyl group increases the steric hindrance. Therefore, the affinity of β-sitosterol for β-CD is weakest compared with that of cholesterol or dodecylamine. It should be mentioned that this is the first case to accurately detect cholesterol in the presence of phytosterols.
Figure 6
3.5 Detection of cholesterol in real samples
This method is further applied to the analysis of cholesterol in foods. As illustrated in Table S1, the content of cholesterol per gram of egg is 3.62 mg; the content of cholesterol per milliliter of milk is 0.25 mg; the content of cholesterol per gram of pork is 1.46 mg. These results are similar to those obtained by ELISA as a reference method. It reveals that the detection of cholesterol based on this essay is reliable. After that, a series of cholesterol standard solutions with different levels are spiked to the food samples. And the result can verify the accuracy of the proposed method. As shown in Table S2, it is obtained that the range of the satisfactory recoveries of cholesterol is from 96.8–100.6%, revealing that the detection of cholesterol in foods is accurate based on the proposed method. The therapeutic lifestyle recommends that the intake of cholesterol in patients with high cholesterol should not exceed 200 mg/day. Hence, according to the test results, it is recommended that the daily intake of eggs does not exceed 55 g; milk does not exceed 800 g; pork does not exceed 135 g. These results can provide dietary guidance for people with high cholesterol.
It is worth noting that this method can accurately determine the cholesterol content when cholesterol coexists with phytosterols, so cholesterol is measured in milk powder, where phytosterols are added (Table S3). The cholesterol in this kind of milk powder is 1.02 mg/g, which is not much different from the labeled amount on milk powder cans (1.08 mg/g). It indicates that cholesterol can be accurately detected in presence of phytosterols by this assay.