Computer-assisted dissection of rolling circle DNA replication

Abstract

A comparative analysis of the proteins involved in initiation and termination of rolling circle replication (RCR) was performed using computer-assisted methods of data based screening, motif search and multiple amino acid sequence alignment. Two vast classes of such proteins were delineated, one of these being associated with RCR proper, and the other with mobilization (conjugal transfer) of plasmid DNA. The common denominator of the two classes was found to be a conserved amino acid motif that consists of the sequence HisUHisUUU (U — bulky hydrophobic residue; hereafter HUH motif). Based on analogies with metalloenzymes, it is hypothesized that the two conserved His residues this motif may be involved in metal ion coordination required for the activity of the RCR and mobilization proteins. The proteins of the replication (Rep) class contained two additional conserved motifs, with the motif around the Tyr residue(s) forming the covalent link with nicked DNA being located C-proximally of the HUH motif. This class further split into two large superfamilies and several smaller families, with the proteins belonging to a single but not to different (super)families demonstrating statistically significant similarity to each other. Superfamily I, prototyped by the gene A proteins of small isometric single-stranded (ss) DNA bacteriophages, included also Rep proteins of P2-related double-stranded (ds) DNA bacteriophages, the small phage-plasmid hybrid phasyl, and several cyanobacterial and archaebacterial plasmids. These proteins contained two invariant Tyr residues separated by three partially conserved amino acids, suggesting that they all may share the cleavage-ligation mechanism proposed for φX174 A protein and involving alternate covalent binding of both tyrosines to DNA (Van Mansfeld, A.D., Van Teeffelen, H.A., Baas, P.D., Jansz, H.S., 1986. Nucl. Acids Res. 14, 4229–4238). Superfamily II included Rep proteins of a number of ssDNA plasmids replicating mainly in gram-positive bacteria that unexpectedly were shown to be related to the Rep proteins of plant geminiviruses. Conservation of the “HUH” motif and a motif around the putative DNA-linking Tyr residue was observed also in the Rep proteins of animal parvoviruses containing linear ssDNA with a terminal hairpin and replicating via the rolling hairpin mechanism. The class of plasmid mobilization (Mob) proteins was characterized by the opposite orientation of the conserved motifs, with the (putative) DNA-linking Tyr being located N-proximally of the “HUH” motif. This class also separated into several distinct families, the largest of which was comprised by the Mob (Tral) proteins of promiscuous IncP and IncI plasmids, VirD2 endonucleases of Agrobacterium TI plasmids, and Mob proteins of a group of gram-positive bacterial ssDNA plasmids. The majority of ssDNA plasmid Mob proteins constituted another family, whereas the Mob domains of TraI proteins of F factor and related plasmids formed a separate group that was only distantly related to the former two families. Additionally, a family of plasmid Rep proteins was analyzed that are unrelated to the above two classes and do not contain the HUH motif but possess instead several distinct conserved motifs. A protein encoded by an archaebacterial virus gene was shown to be distantly related to this family, with significant sequence conservation observed around the putative DNA-linked tyrosine residue. This analysis allowed the prediction of the amino acid residues involved in DNA nicking, which is required for the initiation of RCR or conjugational transfer of ssDNA, in the Rep and Mob proteins encoded by a number of replicons of highly diverse size, structure and origin. It is conjectured that recombination has played a major part in the dissemination of genes encoding related Rep or Mob proteins among these replicons. It is speculated that the eukaryotic replicons encoding proteins with the conserved RCR motifs and replicating via RCR-related mechanisms, such as geminiviruses and parvoviruses, may have evolved from eubacterial ssDNA replicons. Analysis of the nucleotide sequences of the replication and transfer origins ( ori) of various replicons allowed the tentative identification of several previously uncharacterized ori sites but showed that only partial correlation exists between the sequence conservation in the replication (transfer) initiation proteins, and in ori sites.