Here, we use the Spatial Object Recognition (SOR) task to study the mechanism underlying the persistence of long term memory (LTM) after spaced learning. A weak SOR (wSOR) training induced short but not LTM, whereas a strong SOR (sSOR) training formed a 1 but not 7 days LTM. Significantly, adding a wSOR-retraining session 1 day after sSOR training, promotes SOR-LTM persistence tested at the 7th day. We observed that the retraining session is only effective when the memory is liable to be expressed. Emetine, Rapamycin and the ERKs1/2 inhibitor U0126 blocked SOR-LTM expression administered before a test session, and these drugs infusion in the dorsal hippocampus before retraining impaired the LTM persistence. The amnesia induced by Eme and Rapa, but not by U0126, could be reverted by an open field (OF) exposure after retraining. We analyzed these results under the Behavioral Tagging hypothesis (BT) which postulates that the formation of lasting memories relies on the setting of a learning tag and the synthesis of plasticity-related proteins (PRPs). Thus, our results suggest that blocking ERKs activity at retraining impaired the tag setting, because PRPs provided by OF exposure did not rescue LTM. We postulate that retraining will mainly reactivate the sites labeled by the original learning, where the PRPs needed for memory expression and/or induced by retrieval, would be used to sustain the synaptic plasticity for establishing a persistent mnemonic trace.