Abstract
Prunus mume is a traditional ornamental plant, which owed a unique floral scent. However, the diversity of the floral scent in P. mume cultivars with different aroma types was not identified. In this study, the floral scent of eight P. mume cultivars was studied using headspace solid-phase microextraction (HS-SPME) and organic solvent extraction (OSE), combined with gas chromatography-mass spectrometry (GC-MS). In total, 66 headspace volatiles and 74 endogenous extracts were putatively identified, of which phenylpropanoids/benzenoids were the main volatile organic compounds categories. As a result of GC-MS analysis, benzyl acetate (1.55-61.26%), eugenol (0.87-6.03%), benzaldehyde (5.34-46.46%), benzyl alcohol (5.13-57.13%), chavicol (0-5.46%), and cinnamyl alcohol (0-6.49%) were considered to be the main components in most varieties. However, the volatilization rate of these main components was different. Based on the variable importance in projection (VIP) values in the orthogonal partial least-squares discriminate analysis (OPLS-DA), differential components of four aroma types were identified as biomarkers, and 10 volatile and 12 endogenous biomarkers were screened out, respectively. The odor activity value (OAV) revealed that several biomarkers, including (Z)-2-hexen-1-ol, pentyl acetate, (E)-cinnamaldehyde, methyl salicylate, cinnamyl alcohol, and benzoyl cyanide, contributed greatly to the strong-scented, fresh-scented, sweet-scented, and light-scented types of P. mume cultivars. This study provided a theoretical basis for the floral scent evaluation and breeding of P. mume cultivars.