In order to evaluate the role of persisting virus replication during occult phase immunisation in the live attenuated SIV vaccine model, a novel SIVmac239Δnef variant (SIVrtTA) genetically engineered to replicate in the presence of doxycycline was evaluated for its ability to protect against wild-type SIVmac239. Indian rhesus macaques were vaccinated either with SIVrtTA or with SIVmac239Δnef. Doxycycline was withdrawn from 4 of 8 SIVrtTA vaccinates before challenge with wild-type virus. Unvaccinated challenge controls exhibited ~107 peak plasma viral RNA copies/ml persisting beyond the acute phase. Six vaccinates, four SIVmac239Δnef and two SIVrtTA vaccinates exhibited complete protection, defined by lack of wild-type viraemia post-challenge and virus-specific PCR analysis of tissues recovered post-mortem, whereas six SIVrtTA vaccinates were protected from high levels of viraemia. Critically, the complete protection in two SIVrtTA vaccinates was associated with enhanced SIVrtTA replication in the immediate post-acute vaccination period but was independent of doxycycline status at the time of challenge. Mutations were identified in the LTR promoter region and rtTA gene that do not affect doxycycline-control but were associated with enhanced post-acute phase replication in protected vaccinates. High frequencies of total circulating CD8+T effector memory cells and a higher total frequency of SIV-specific CD8+ mono and polyfunctional T cells on the day of wild-type challenge were associated with complete protection but these parameters were not predictive of outcome when assessed 130 days after challenge. Moreover, challenge virus-specific Nef CD8+ polyfunctional T cell responses and antigen were detected in tissues post mortem in completely-protected macaques indicating post-challenge control of infection. Within the parameters of the study design, on-going occult-phase replication may not be absolutely required for protective immunity.