Tannins may have detrimental effects on digestion, for example by inhibiting the digestive enzymes of the oro-gastro-intestinal tract [1]. Besides, some salivary proteins (such as PRPs or histatins) have a high affinity for tannins which is considered as a protective function of saliva against the deleterious impact of excessive tannins [2]. This research aims to describe the interactions between saliva proteins and tannins from apples and to follow the fate of the complexes during in vitro digestion. In the experiments, we used on one hand a polyphenol mixture extracted from the apple cider variety Dous Moën containing 730 g/kg total polyphenols of which 44% were condensed tannins. On the other hand, clarified human saliva at a protein concentration of 0.46 mg/ml was used. First, polyphenols and saliva were mixed at different ratios of tannins to saliva proteins (w/w) from 0.1 to 1.1. We measured the turbidity of the resulting samples by spectrophotometry at 400nm. Samples were then centrifuged at 10000 g, 4°C for 10 minutes in order to separate the larger aggregates in the pellet from soluble compounds and smaller aggregates in the supernatant. The concentration of tannins in the supernatants was measured by the Butanol-HCl method. Protein concentration in the pellets (resuspended in urea buffer) and in the supernatants was measured by the Bradford method, and protein profiles in both fractions were obtained by SDS-PAGE and compared to that of saliva alone. When the ratio of tannins to saliva protein increased, turbidity in whole samples increased, suggesting the formation of increasingly larger saliva proteins-tannins aggregates. In accordance, the protein content in pellets increased while the protein content in supernatants decreased. Results of tannins in supernatants showed that 34 to 50 % of tannins remained soluble. Protein profiles revealed that some proteins were enriched in the pellets, particularly small proteins with molecular weights estimated at around 10-15kDa. This confirmed that tannins interact with and precipitate preferentially low molecular weight saliva proteins. In a second step, polyphenol extracts-saliva mixtures were subjected to the INFOGEST static in vitro digestion model. Samples taken at the end of the gastric and intestinal phases of digestion were centrifuged and SDS-PAGE electrophoresis was performed on the pellets and supernatants. Apart from pepsin and trypsin, no proteins were detectable in gastric and intestinal supernatants respectively. In contrast, intestinal pellets showed a very clear band at around 10kDa on electrophoretic gels, suggesting the resistance of this small salivary protein to the digestive process. Additional experiments are currently being conducted to determine whether interaction with tannins plays a role in this persistence.