respectively. Of the two mutants carrying a TPOP insertion element in the pduC gene, strain BE34, was still capable of producing polyhedra. It appeared that this was due to the more distal location of the insertion mutation in thepduC gene in strain BE34 as compared to strain BE35, which was unable to form polyhedra. Subsequent tests on strain BE87, containing a precise deletion of the pduCDE genes, indicated that thepduC gene was not required for the formation of polyhedral organelles substantiating the results seen in strain BE34. Apparently, downstream expression from the TPOP insertion in strain BE35 is not high enough to support polyhedral organelle formation. These results illustrate the difficulties in studying structures in bacteria, which may require strict gene dosage for proper formation.

      Complementation studies conducted on strains GH4 (pduA:: TPOP/plac22-pduA) and GH42 (pduB::TPOP/plac22-pduB) to determine if the respective mutations in the pduA and pduB genes could be complemented using plasmid-encoded sources of the genes suggested that both pduA and pduB are involved in polyhedral organelle formation since aberrant structures were observed in both strains upon induction of plasmid clones with IPTG. Possible reasons for the inability to correctly form polyhedral organelles could be the requirement of thepduA andpduB gene products in specific relative amounts, translational coupling of thepduA and pduB genes, or altered downstream expression from the tetracycline inducible promoters of the TPOP insertions.

      Formerly, Orus et al. investigated the biogenesis of carboxysomes using electron microscopy (Orus et al. 1995). They observed what they described as different stages of carboxysome formation. Based on these findings they postulated that the carboxysome shell is made first and RuBisCO is subsequently inserted into the carboxysome through