The present theoretical approaches for the single electron spectra from heavy meson decay, measured at RHIC by the STAR and the PHENIX collaborations, which are based on perturbative QCD (pQCD) underpredict the collisional energy loss by a large factor and are not able to reproduce the azimuthal distribution. They suffer in addition from the fact that neither the value of the (fixed) coupling constant nor that for the infrared (IR) regulator are well determined. We investigate the consequences of a physical running coupling constant and of a IR regulator, determined by a hard thermal loop (HTL) approach. Both together increase the collisional energy loss substantially and bring the out-of-plane $(v_2)$ anisotropy as well as the quenching to values close to the experimental values without excluding a contribution from radiative energy loss.