Using 8-10-m class telescopes and their Adaptive Optics (AO) systems, we conducted a long-term adaptive optics campaign initiated in 2003 focusing on four binary asteroid systems: (130) Elektra, (283) Emma, (379) Huenna, and (3749) Balam. The analysis of these data confirms the presence of their asteroidal satellite. We did not detect any additional satellite around these systems even though we have the capability of detecting a loosely-bound fragment (located at 1/4×R) ∼40 times smaller in diameter than the primary. The orbits derived for their satellites display significant eccentricity, ranging from 0.1 to 0.9, suggesting a different origin. Based on AO size estimate, we show that (130) Elektra and (283) Emma, G-type and P-type asteroids, respectively, have a significant porosity (30-60% considering CI-CO meteorites as analogs) and their satellite's eccentricities ( e∼0.1) are possibly due to excitation by tidal effects. (379) Huenna and (3749) Balam, two loosely bound binary systems, are most likely formed by mutual capture. (3749) Balam's possible high bulk density is similar to (433) Eros, another S-type asteroid, and should be poorly fractured as well. (379) Huenna seems to display both characteristics: the moonlet orbits far away from the primary in term of stability ( 20%×R), but the primary's porosity is significant (30-60%).