Due to its strong immune stimulatory effects through TLR9, CpG-containing oligodeoxynucleotides (CpG ODN) have been tested in multiple clinical trials as vaccine adjuvant for infectious diseases and cancer. However, immune suppression induced by systemic administration of CpGs has been reported recently. In this study, we evaluated the impact of CpGs in an acute rickettsiosis model. We found that systemic treatment with type B CpG (CpG-B), but not type A CpG (CpG-A), at 2 days after sublethal R. australis infection induced mouse death. Although wild-type (WT) B6 and IDO(-/-) mice showed similar survival rates with three different doses of R. australis infection, treatment with CpG-B after sublethal infection consistently induced higher mortality with greater tissue bacterial loads in WT but not IDO(-/-) mice. Also, CpG-B treatment promoted the development of higher serum concentrations of proinflammatory cytokines/chemokines through IDO. Furthermore, while T cell-mediated immune responses enhanced by CpG-B were independent of IDO, treatment with CpG-B promoted T cell activation, PD-1 expression and cell apoptosis partially through IDO. A depletion study using anti-mPDCA-1 mAb indicated that plasmacytoid dendritic cells (pDC) were not required for CpG-B-induced death of R. australis-infected mice. Additionally, the results in iNOS(-/-) mice suggested that nitric oxide (NO) was partially involved in CpG-B-induced death of R. australis-infected mice. Surprisingly, pre-treatment with CpG-B before administration of a lethal dose of R. australis provided effective immunity in WT, IDO(-/-) and iNOS(-/-) mice. Taken together, our study provides evidence that CpGs exert complex immunological effects by both IDO-dependent and -independent mechanisms, and that systemic treatment with CpGs before or after infection has a significant and distinct impact on disease outcomes.