Rapid Detection Method for Pathogenic Candida Captured by Magnetic Nanoparticles and Identified Using SERS via AgNPs+

Shan Hu,1– 3,* Haiquan Kang,4,* Feng Gu,2,* Chongwen Wang,1,5 Siyun Cheng,3 Wenjing Gong,1 Liping Wang,1 Bing Gu,3,4 Ying Yang1 1Department of Biotechnology, Beijing Institute of Radiation Medicine, Beijing Key Laboratory of New Molecular Diagnosis Technologies for Infectious Diseases, Beijing, 100850, People’s Republic of China; 2Department of Laboratory Medicine, Xuzhou Tumor Hospital, Xuzhou, 221005, People’s Republic of China; 3Xuzhou Key Laboratory of Laboratory Diagnostics, Medical Technology School of Xuzhou Medical University, Xuzhou, 221004, People’s Republic of China; 4Department of Laboratory Medicine, The Affiliated Hospital of Xuzhou Medical University, Xuzhou, 221006, People’s Republic of China; 5College of Life Sciences, Anhui Agricultural University, Hefei, 230036, People’s Republic of China*These authors contributed equally to this workCorrespondence: Ying Yang; Bing Gu Email y_ying_77@163.com; gb20031129@163.comPurpose: Candidemia infection is common in the clinic and has a high mortality rate. Candida albicans, Candida tropicalis, and Candida krusei are very important and common pathogenic species. Candida is difficult to isolate from clinical samples and culture, and immunological detection cannot distinguish these related strains. Furthermore, Candida has a complex cell wall, which causes difficulties in the extraction of DNA for nucleic acid detection. The purpose of this study was to establish a protocol for the direct identification of Candida from serum.Materials and Methods: We synthesized Fe3O4@PEI (where PEI stands for polyethylenimine) magnetic nanoparticles to capture Candida and prepared positively charged silver nanoparticles (AgNPs+) as the substrate for surface-enhanced Raman scattering (SERS). Candida was directly identified from serum by SERS detection.Results: Orthogonal partial least squares discriminant analysis (OPLS-DA) was used as the multivariate analysis tool. Principal component analysis confirmed that this method can clearly distinguish common Candida. After 10-fold cross-validation, the accuracy of training data in this model was 100% and the accuracy of test data was 99.8%, indicating that the model has good classification ability.Conclusion: The detection could be completed within 40 minutes using Fe3O4@PEI and AgNPs+ prepared in advance. This is the first time that Fe3O4@PEI was used in the detection of Candida by SERS. We report the first rapid method to identify fungi directly from serum without breaking the cell wall to extract DNA from the fungi.Keywords: capture, surface-enhanced Raman scattering, positively charged silver nanoparticles, orthogonal partial least squares discriminant analysis, 10-fold cross-validation