ABSTRACT: BACKGROUND: Cassava mosaic disease is caused by several distinct geminivirus species, including South African cassava mosaic virus-[South Africa:99] (SACMV). To date, there is limited gene regulation information on viral stress responses in cassava, and global transcriptome profiling in SACMV-infected cassava represents an important step towards understanding natural host responses to plant geminiviruses. RESULTS: A RNA-seq time course (12, 32 and 67 dpi) study, monitoring gene expression in SACMV-challenged susceptible (T200) and tolerant (TME3) cassava landraces, was performed using the Applied Biosystems (ABI) SOLiD next-generation sequencing platform. The multiplexed paired end sequencing run produced a total of 523 MB and 693 MB of paired-end reads for SACMV-infected susceptible and tolerant cDNA libraries, respectively. Of these, approximately 50.7% of the T200 reads and 55.06% of TME3 reads mapped to the cassava reference genome available in phytozome. Using a log2 fold cut-off (p<0.05), comparative analysis between the six normalized cDNA libraries showed that 4181 and 1008 transcripts in total were differentially expressed in T200 and TME3, respectively, across 12, 32 and 67 days post infection, compared to mock-inoculated. The number of responsive transcripts increased dramatically from 12 to 32 dpi in both cultivars, but in contrast, in T200 the levels did not change significantly at 67 dpi, while in TME3 they declined. GOslim functional groups illustrated that differentially expressed genes in T200 and TME3 were overrepresented in the cellular component category for stress-related genes, plasma membrane and nucleus. Alterations in the expression of other interesting genes such as transcription factors, resistance (R) genes, and histone/DNA methylation-associated genes, were observed. KEGG pathway analysis uncovered important altered metabolic pathways, including phenylpropanoid biosynthesis, sucrose and starch metabolism, and plant hormone signalling. CONCLUSIONS: Molecular mechanisms for TME3 tolerance are proposed, and differences in patterns and levels of transcriptome profiling between T200 and TME3 with susceptible and tolerant phenotypes, respectively, support the hypothesis that viruses rearrange their molecular interactions in adapting to hosts with different genetic backgrounds.