Cytosolic and mitochondrial lysyl-tRNA synthetases (LysRS) are encoded by a single gene and can be distinguished only according to their very N-terminal sequences. It was believed that cytosolic LysRS is packaged into HIV-1 virions via its association with Gag. Using monospecific antibodies, it was later shown that only the mitochondrial LysRS is taken up in viral particles along with tRNA(3)(Lys), the primer for reverse transcription of the HIV-1 genome. In this work, we re-analyzed the interaction between LysRS and GagPol to determine whether the particular N-terminal sequence of mitochondrial LysRS triggers a specific recognition with GagPol, or if differential routing of the two LysRS species in vivo could explain specific and exclusive packaging of the mitochondrial species. Here, we show that LysRS associates with the Pol domain of GagPol. More specifically, the transframe (TF or p6) and integrase (IN) domain proteins of Pol interact with the catalytic domain of LysRS. A model of the assembly of the LysRS-tRNA(3)(Lys)-GagPol packaging complex is proposed, which is consistent with the release of its different components after maturation of GagPol in the virions. The cytoplasmic and mitochondrial LysRS species share an identical catalytic domain. Accordingly, we found that both enzymes have the intrinsic capacity to bind to GagPol in vitro. In addition, both enzymes interact with p38 in vitro, the scaffold protein of the cytoplasmic multi-aminoacyl-tRNA synthetase complex, even though only the cytoplasmic species of LysRS is a bona fide component of this complex. These results suggest that the different LysRS species are strictly targeted in vivo, and open new perspectives for the search of a new class of inhibitors of the HIV-1 development cycle that would block the packaging of tRNA(3)(Lys) into viral particles.