The tested hypotheses of this study were that dentin enamel bonding agents (DBAs) proposed for compomers create a hybrid layer (HL) that seals the margins of Class V restorations and HL is free from voids or gaps on both enamel and dentin margins. For purposes of this study, Class V restorations (n = 70) were made in vitro at the CEJ in extracted third molars. Different systems (bonding agent + compomer) were selected. After finishing with discs, each margin was polished with diamond polishing paste for one minute, treated with a 2.5% NaOCl gel for 10 seconds and washed with deionized water to remove polishing debris and non-infiltrated collagen. All restorations were immersed in dye solution for 24 hours, then inspected along the margins. SEM analysis was used to evaluate the morphology of the marginal HL and microleakage tests to evaluate their ability to seal the margins of restorations. Marginal leakage was observed along the dentin and enamel margin. A thin marginal HL (0.5-1.2 microm) was detected only along the dentin margin of several bonding systems but not along the enamel margin. Porosities and gaps were detected along margins when no HL was observed. The results demonstrated that the tested bonding agents for compomers produced a thin marginal hybrid layer, especially along the dentin margin. Microleakage had a relationship with the morphology (gap, porosities and thickness) of this hybrid smear layer. In conclusion, the DBAs tested specifically developed for compomers did not ensure an intimate interfacial adaptation, because microleakage was detected along the enamel and dentin interfaces and the marginal hybrid layer was only partially homogeneous.