Honey has health benefits as well as medicinal value and it is one of the most widely consumed nutritional products in China (Yang et al. 2020). China is a vast country with a wide variety of nectar plants. Common nectar source plants are platycodon, rapeseed, lychee, loquat, osmanthus, safflower. Toxic nectar plants like Gelsemium elegans, Tripterygium wilfordii, Veratrum nigrum L, Golden Wisteria, Macleaya cordata, and Stellera chamaejasme Linn are also found in China. At the end of the flowering period of non-toxic nectar plants or in excessively wet or dry conditions, toxic plants can become the dominant source of nectar, which results in the production of toxic honey (Kumaravelu and Gopal. 2015; Sun et al. 2019).
G. elegans is a poisonous plant that is mainly found in southern China (provinces like Zhejiang, Guangxi, Guangdong and Fujian), which is one of the poisonous nectar source plants. Indole alkaloids are the main toxic constituents of G. elegans, distributed throughout the plant, which has strong toxicity (Shen et al. 2020). In recent years, several cases of honey poisoning and illness have occurred in southern China due to the consumption of G. elegans-containing honeys; deaths have occurred in serious cases (Xu et al. 2015). Therefore, studies about methods for the identification of toxic honey are of great significance for preventing honey poisoning and ensuring the stability of the honey industry. Currently, the methods for identifying toxic honey include sensory analysis (Marcazzan et al. 2018), pollen identification (Kraaijeveld et al. 2015), gas chromatography (Kowalczyk et al. 2018), thin-layer chromatography(Islam et al. 2020), and liquid chromatography-tandem mass spectrometry (Koike et al. 2020). Mistakes can occur in sensory analysis and pollen identification due to interpretation bias. Chromatography is complex, time-consuming, and expensive. Therefore, there is an urgent need to develop a fast, simple, and accurate analytical method for the identification of toxic honey.
Near-infrared (NIR) spectroscopy is a fast, simple, and accurate analytical technique that is widely used in food analysis (Yakubu et al. 2020). In recent years, many studies around the world have demonstrated the feasibility of using NIR spectroscopy for honey quality inspections (Skaff et al. 2021). However, the use of NIR to detect G. elegans-containing toxic honeys has yet to be reported. Honey consists of water and a diverse variety of carbohydrates. Due to its diverse components and water absorption, the NIR spectra of honey contains overlapping bands. Moreover, as the toxic components of G. elegans are present in low concentrations, their spectral absorptions are easily masked by those of other compounds. This makes it difficult to identify toxic honeys based on spectral differences alone. To address the aforementioned problems and extract useful information from NIR spectra, it is necessary to develop an algorithm that can extract the spectral features of toxic honey and use them to identify toxic honeys.
Aquaphotomics is a new scientific discipline introduced by Tsenkova (Tsenkova et al. 2018). In this novel approach to NIR analysis, water molecules are treated as the “matrix” of the system, and their absorptions are the primary source of information. Other substrates (solutes) are viewed as perturbing factors. By using the “extended water mirror approach” (EWMA), aquaphotomics allows changes in NIR water absorptions to reflect on changes in molecules in the aqueous system. Currently, aquaphotomics is widely used in food analysis (Muncan and Tsenkova. 2019; Yakubu et al. 2020) and disease diagnosis (Li et al. 2020; Tsenkova 2006). However, the use of NIR aquaphotomics to identify toxic honeys has not been reported.
The aim of this study was to provide a rapid and accurate method for identifying toxic honey. To this end, NIR spectroscopy was combined with partial least-squares discriminant analysis (PLS-DA) to develop a model for the discrimination of toxic and non-toxic honeys. NIR aquaphotomics was then used to elucidate changes in the NIR water absorptions of toxic and non-toxic honey and aquagrams were used to analyze the differences in their water spectral patterns (WASPs). A new approach for the rapid and accurate identification of toxic honeys was established.