Immunomagnetic Separation Of Alicyclobacillus Species In Apple Juice And Identification By Vibrational Spectroscopy

China is the largest concentrated apple juice producing country and its yield increases rapidly. The safety issues, especially Alicyclobacillus species, are“green technical barriers”of international trade. Control of Alicyclobacillus is the key to development of apple juice industry, and the most important aspect is to establish rapid methods to detect and identify Alicyclobacillus species. Currently, rapid separation and enrichment of Alicyclobacillus species is the critical technical bottleneck; furthermore, the existing methods for detection of microorganisms have tremendous limitations.In order to establish efficient and rapid methods to separate, enrich and identify Alicyclobacillus species, polyclonal antibody against Alicyclobacillus acidoterrestris (DSM3922) was connected to magnetic microspheres to prepare immune magnetic microspheres, a method for separating and enriching Alicyclobacillus species in apple juice based on immunomagnetic microspheres was established; further, vibrational spectroscopy (namely, near infrared, mid-infrared and Raman) combined with multivariate analysis was employed to identify Alicyclobacillus species and identification models based on multivariate analysis were established. A rapid, simple and convenient method to identify Alicyclobacillus species was provided for apple juice industry.The main results of this thesis:(1) Polyclonal antibody against Alicyclobacillus acidoterrestris DSM3922 was obtained by performing immune tests on two rabbits: R1, 27.6 mg/mL, 4.6 mL, titer 1:40000; R2, 10.8 mg/mL, 5 mL, titer 1:10000. Chitosan magnetic microsphere was prepared by chemical embedding and activated, which could stabilize polyclonal antibody at 2.48 mg antibody/g microsphere; Polyclonal antibody was coupled with magnetic microsphere to prepare immune magnetic microsphere. Immune magnetic microsphere was used to separate Alicyclobacillus acidoterrestris from medium and apple juice, the absorption rate was 68.5±1.7% and 70.1±4.7%, respectively.(2) Bacterial powders of Alicyclobacillus strain, one yeast and five bacteria strains were prepared for Fourier transform near-infrared (FT-NIR) spectral collection. FT-NIR spectral determination was done using a diffuse reflection-integrating sphere. Reduction of data was performed by principal component analysis (PCA) and two identification models based on linear discriminant analysis (LDA) and artificial neural network (ANN) were established to identify bacterial strains. The reproducibility of the method was satisfied (Dy1y2: 1.61±1.05-10.97±6.65). The wavenumber of 5400-4000cm-1 is the information-rich range for the FT-NIR spectra of microorganism and high identification accuracy was achieved in both the LDA model (accuracy rate: 100%) and the ANN model (average relative error: 5.04%).(3) A simple and rapid sample preparation method using nitrocellulose membrane filter (NMF) and a single reflection horizontal attenuated total reflection (HATR) accessory was developed, mid-infrared (mid-IR) spectra of Alicyclobacillus strain and seven other representative bacterial strains were collected and two identification models based on LDA and ANN respectively were established to identify and distinguish Alicyclobacillus strain from others. The sample preparation method was feasible and the microorganisms studied were successfully separated into different groups by PCA. High identification accuracy was achieved in both LDA model (accuracy rate: 100%) and ANN model (average relative error: 1.32%). In addition, Fourier transform infrared (FT-IR) spectroscopy was used and tested on eight Alicyclobacillus strains. The stains could be clearly separated into different groups by PCA. High identification accuracy (93.75%) was achieved using LDA model.(4) Raman spectra of Alicyclobacillus strain and five other representative bacterial strains were collected using a Raman microspectrometer. Reduction of data was performed by PCA and an identification model based on LDA was established to identify bacterial strains. Results showed that the main bands found in Raman spectra of microorganisms originated from proteins and carbohydrates, only several weak peaks were from nucleic acids and fatty acids. The minimal bacterial concentration for collecting Raman was 1011 CFU/mL and the S/N was higher than 250. The reproducibility of the method was satisfied (Dy1y2: 20.91±16.17-39.50±59.26) and high identification accuracy was achieved in LDA model (accuracy rate: 90%). Further, Raman spectroscopy was used and tested on eight Alicyclobacillus strains. The stains could be separated into different groups by PCA successfully. High identification accuracy (85%) was achieved using LDA model.