Human Immunodeficiency Virus type 1 (HIV-1) integrase (IN) is an attractive target for the development of new antiviral therapies. Recently, several HIV-1 recombinant IN (rIN) in vitro inhibitors have been described. However, the great majority of them failed to block the virus replication in cell-based assays, suggesting the inadequacy of the in vitro assay systems used for inhibitor screening. To improve these systems, we designed a 40(mer) duplex DNA reaction substrate consisting of recognition sequences from both U3 and U5 HIV-1 long terminal repeat (LTR) termini. The HIV-1 rIN was able to catalyze its enzyme activities recognizing both ends of the 40(mer) dsDNA. Using this substrate we assayed the effects on rIN catalysis of different classes of compounds which inhibit the HIV-1 rIN in vitro when the reaction substrate is the standard 21(mer) U5 dsDNA, and that are either active or inactive on the HIV-1 replication. We also compared the efficacy of these compounds when added to the reaction before or after the formation of the rIN-dsDNA complex. In this system, the enzyme preincubation with the two-ended 40(mer) dsDNA before the addition of the compounds allowed a strong correlation between the effects of hydroxylated aromatics derivatives on rIN activity in cell-free assays and their effects on viral replication in cell-culture assays. This increase in drug selectivity of the rIN in vitro assay was explored by investigating whether it was due to the length of the 40(mer), longer than the standard 21(mer), or to presence of both viral ends, versus only one viral end. To this purpose we designed four 40(mer) oligonucleotides containing either only one viral end or two-repetitive ends, finding that the architecture of the rIN-dsDNA complex and its compound susceptibility is significantly influenced by the sequence of the dsDNA substrate.