Cuticle function is closely related to the structure of the cutin polymer. However, the structure and formation of this hydrophobic polyester of glycerol and hydroxy/epoxy fatty acids has not been fully resolved. An apoplastic GDSL-lipase kown as cutin synthase 1 (CUS1) is required for cutin deposition in tomato fruit exocarp. In vitro, CUS1 catalyzes the self-transesterification of 2-monoacylglycerol of 9(10),16-dihydroxyhexadecanoic acid, the major tomato cutin monomer. This reaction releases glycerol and leads to the formation of oligomers with the secondary hydroxyl group remaining non-esterified. To check this mechanism in planta, a benzyl etherification of non-esterified hydroxyl groups of glycerol and hydroxy fatty acids was performed within cutin. Remarkably, in addition to a significant decrease in cutin deposition, midchain hydroxyl esterification of the dihydroxyhexadecanoic acid was affected in tomato RNAi and EMS-cus1 mutants. Furthermore, in these mutants, esterification of both sn-1,3 and sn-2 positions of glycerol was impacted and their cutin contained a higher molar glycerol to dihydroxyhexadecanoic acid ratio. Therefore in planta, CUS1 can catalyze the esterification of both primary and secondary alcohol groups of cutin monomers, and another enzymatic or non-enzymatic mechanism of polymerization may coexist with CUS1-catalyzed polymerization. This mechanism is poorly efficient with secondary alcohol groups and produces polyesters with lower molecular size. Confocal Raman imaging of benzyl etherified cutins showed that the polymerization is heterogeneous at the fruit surface. Finally, by comparing tomato mutants either affected or not in cutin polymerization, we concluded that the level of cutin crosslinking had no significant impact on water permeance.