Severe acute respiratory syndrome (SARS) is a zoonotic disease caused by SARS-related coronavirus (SARS-CoV) that emerged in 2002 to become a global health concern. Although the original outbreak was controlled by classical public health measures, there is a real risk that another SARS-CoV could re-emerge from its natural reservoir, either in its original form or as a more virulent or pathogenic strain; in which case, the virus would be difficult to control in the absence of any effective antiviral drugs or vaccines. Using the well-studied SARS-CoV isolate HKU-39849, we developed a vaccinia virus-based SARS-CoV reverse genetic system that is both robust and biosafe. The SARS-CoV genome was cloned in separate vaccinia virus vectors, (vSARS-CoV-5prime and vSARS-CoV-3prime) as two cDNAs that were subsequently ligated to create a genome-length SARS-CoV cDNA template for in vitro transcription of SARS-CoV infectious RNA transcripts. Transfection of the RNA transcripts into permissive cells led to the recovery of infectious virus (recSARS-CoV). Characterization of the plaques produced by recSARS-CoV showed that they were similar in size to the parental SARS-CoV isolate HKU-39849 but smaller than the SARS-CoV isolate Frankfurt-1. Comparative analysis of replication kinetics showed that the kinetics of recSARS-CoV replication are similar to those of SARS-CoV Frankfurt-1, although the titers of virus released into the culture supernatant are approximately 10-fold less. The reverse genetic system was finally used to generate a recSARS-CoV reporter virus expressing Renilla luciferase in order to facilitate the analysis of SARS-CoV gene expression in human dendritic cells (hDCs). In parallel, a Renilla luciferase gene was also inserted into the genome of human coronavirus 229E (HCoV-229E). Using this approach, we demonstrate that, in contrast to HCoV-229E, SARS-CoV is not able to mediate efficient heterologous gene expression in hDCs.