A HYPOTHESIS FOR IMPORT OF THE NUCLEAR‐ENCODED PsaE PROTEIN OF PAULINELLA CHROMATOPHORA (CERCOZOA,RHIZARIA) INTO ITS CYANOBACTERIAL ENDOSYMBIONTS/PLASTIDS VIA THE ENDOMEMBRANE SYSTEM1

The cyanobacterial endosymbionts of Paulinella chromatophora can shed new light on the process of plastid acquisition. Their genome is devoid of many essential genes, suggesting gene transfer to the host nucleus and protein import back into the endosymbionts/plastids. Strong evidence for such gene transfer is provided by the psaE gene, which encodes a PSI component that was efficiently transferred to the Paulinella nucleus. It remains unclear, however, how this protein is imported into the endosymbionts/plastids. We reanalyzed the sequence of Paulinella psaE and identified four potential non‐AUG translation initiation codons upstream of the previously proposed start codon. Interestingly, the longest polypeptide, starting from the first UUG, contains a clearly identifiable signal peptide with very high (90%) predictability. We also found several downstream hairpin structures that could enhance translation initiation from the alternative codon. These results strongly suggest that the PsaE protein is targeted to the outer membrane of Paulinella endosymbionts/plastids via the endomembrane system. On the basis of presence of respective bacterial homologs in the Paulinella endosymbiont/plastid genome, we discuss further trafficking of PsaE through the peptidoglycan wall and the inner envelope membrane. It is possible that other nuclear‐encoded proteins of P. chromatophora also carry signal peptides, but, alternatively, some may be equipped with transit peptides. If this is true, Paulinella endosymbionts/plastids would possess two distinct targeting systems, one cotranslational and the second posttranslational, as has been found in higher plant plastids. Considering the endomembrane system‐mediated import pathway, we also discuss homology of the membranes surrounding Paulinella endosymbionts/plastids.