cDNA copies of the coat protein (CP) gene of Indian peanut clump virus (IPCV)-H were introduced into cells of Nicotiana benthamiana or Escherichia coli by transformation with vectors based on pROKII or pET respectively. In both plant and bacterial cells, IPCV CP was expressed and assembled to form virus-like particles (VLP). In plant extracts, the smallest preponderant particle length was about 50 nm. Other abundant lengths were about 85 and about 120 nm. The commonest VLP length in bacterial extracts was about 30 nm. Many of the longer VLP appeared to comprise aggregates of shorter particles. The lengths of the supposed 'monomer' VLP corresponded approximately to those expected for encapsidated CP gene transcript RNA. Immunocapture RT-PCR, using primers designed to amplify the CP gene, confirmed that the VLP contained RNA encoding IPCV-H CP. The results show that encapsidation does not require the presence of the 5'-terminal untranslated sequence of the virus RNA and suggest that if there is an 'origin of assembly' motif or sequence, it lies within the CP gene. When transgenic plants expressing IPCV-H CP were inoculated with IPCV-L, a strain that is serologically distinct from IPCV-H, the virus particles that accumulated contained both types of CP.