The purpose of this study was to develop molecular markers suitable for the identification of grape aroma phenotypes, so as to provide a theoretical basis for grape molecular assisted breeding.

The aroma components and content of 45 grape varieties were determined by solid-phase microextraction combined with gas-phase mass spectrometry, and the structural variation and SNP variation of VvAAT gene were analyzed by Sanger sequencing and amplicon sequencing, respectively.

A total of 65 aroma components were found in 45 grape varieties, which could be divided into four types: esters, alcohols, terpenes and aldehydes. Among them, the content of esters showed significant variety differences, and 20 grape varieties such as Kyoho were rich in volatile esters, but little volatile esters was detected in other 25 grape varieties such as 87-1. A total of 5 structural variants have been found in the VvAAT gene. Types I, II, IV and V could not be translated correctly due to premature termination codons or fragment insertion mutations. Only type III could be translated aright. According to the results of the phylogenetic tree analysis of their amino acid sequences, Type III could be divided into two types: III.1 and III.2. Combining the aroma data, it could be inferred that III.1 was functional, while the rest were non-functional. Amplicon sequencing and bioinformatics analysis revealed 8 SNP sites (T32C, A69T, G436C, A1247G, A1818T, G1929T, A1959G, and C1975G) located in the exon region of the VvAAT gene, all of which caused mutations in the coding amino acids. It could accurately distinguish between ester-rich varieties and ester-poor varieties, with an accuracy rate of 97.8%.

VvAAT gene locus had abundant genetic variation, including gene structure variation and SNP variation; 8 SNP loci located in the coding region of VvAAT gene could accurately determine the phenotype of volatile esters, which could be applied to grape molecular assisted breeding.