The replication protein of the archaeal plasmid pRN1 is a multifunctional enzyme which appears to carry out several steps at the plasmid replication initiation. We recently determined the structure of the minimal primase domain of the replication protein and found out that the primase domain consists of a catalytic primase/polymerase domain and an accessory helix-bundle domain. Structure-guided mutagenesis allowed us to identify amino acids which are important for template binding, dinucleotide formation and a step before primer extension. On the basis of functional and structural data, we propose a model of the catalytic cycle of primer synthesis by the pRN1 replication protein.
The replication protein of the plasmid pRN1, an unusual multifunctional enzyme
The plasmid pRN1 was first isolated in 1994 by Zillig et al.  from a Sulfolobus islandicus strain. Interestingly, the original strain REN1-H1 contained two plasmids: pRN1 and pRN2. Sequencing revealed that these two plasmids are similar and, together with other archaeal plasmids, such as pHEN7, pDL10 and the virus-hybrid plasmid pSSVx, form a plasmid family distinct from other known groups of archaeal and bacterial plasmids [2,3].
My group is specifically interested in the replication of these pRN1-type plasmids and also in the application of these plasmids as shuttle vectors in order to have genetic access to Sulfolobus, one of the emerging model organisms of the archaeal domain. Meanwhile, a number of shuttle plasmids based on pRN1-type plasmids have successfully been constructed by several groups [4–6].
The highlight of the pRN1-type plasmids is a large replication protein with up to ~1000 amino acids, whose gene occupies, in the case of pRN1, approximately half of the coding capacity of the plasmid. This protein is a multifunctional replication protein with primase, DNA polymerase, ATPase and helicase activities .
The replication enzyme has an SF (superfamily) 3 helicase domain
When the plasmid pRN1 was first sequenced , the authors noted that the C-terminal part of the pRN1 replication protein (Figure 1) has sequence similarity to the C-terminal part of the bacteriophage P4 α protein, which is also a multifunctional protein with origin-binding, helicase and primase activities . In fact, these parts of the pRN1 replication protein and the P4 α protein constitute a helicase of SF3 . The SF3 helicases are overrepresented in plasmid, viral and bacteriophage genomes. Like the SF4 helicases and the MCM (minichromosome maintenance) proteins, it forms hexameric rings around DNA, but, unlike the helicases of SF4, it translocates in the 3′→5′ direction. A detailed analysis of the helicase activity of the pRN1 replication protein revealed that the helicase activity is rather low, suggesting that the replication protein is not a replicative helicase. The discovery of a winged-helix DNA-binding domain at the very C-terminus of the replication protein (Figure 1) and the observation that the ATPase activity of replication protein is especially well stimulated by double-stranded, but not single-stranded, DNA make it possible that the motor activity of the SF3 helicase domain is responsible for the unwinding of the plasmidal replication origin . Direct experimental proof for this replication initiating activity is, however, lacking.
Domain structure of the multifunctional pRN1 replication protein
The replication protein has a novel prim/pol (primase/polymerase) domain
By investigating the recombinant replication protein using various biochemical assays, we also discovered that the replication protein has primase and DNA polymerase activity, which was surprising, since the amino acid sequence of the replication protein did not reveal any similarity with other primases and DNA polymerases. Deletion mutants allowed the mapping these activities to the N-terminal half of the protein (Figure 1). When the structure of the N-terminal domain was solved, it turned out that amino acids 40–255 folded into a compact domain (Figure 2) with structural resemblance to the RRM (RNA-recognition motif) which is also present in the archeao-eukaryotic primases . This structural resemblance was completely unexpected, since the amino acid sequence of the pRN1 replication protein bears no similarity to the archeao-eukaryotic primases. In fact, there are only a few conserved amino acid positions mostly from the active site.
Structure of the minimal primase domain of the pRN1 replication protein
Archaeal primases are known to perform primer synthesis and elongation with ribonucleotides and deoxyribonucleotides bridging the gap between primases (template-based de novo synthesis of a ribo-oligonucleotide) and DNA polymerases (template-based elongation with deoxynucleotides). A rather unusual nucleotide specificity is observed for the pRN1 replication enzyme. The replication enzyme requires and only accepts dNTPs for elongation; however, for primer synthesis, the first base has to be a ribonucleotide . Thus the replication protein starts primer synthesis with a single ribonucleotide which is then elongated with deoxynucleotides. In the course of these studies, it was also discovered that the DNA polymerase activity is carried out by a deletion mutant of amino acids 40–255 as efficiently as the full-length protein. In contrast, the primase activity requires amino acids 40–370; no primer synthesis is observed when this deletion mutant is shortened further. This observation points to the critical involvement of amino acids 256–370 for primer synthesis. In contrast with this apparent important involvement for primer synthesis, this part of the protein is only conserved in the close homologues of the pRN1 replication proteins. We therefore determined the structure of the primase minimal domain of the replication protein (amino acids 40–370) in order to get more insight into the function of this domain.
Structure of the primase minimal domain
The structure of the primase minimal domain was solved at 1.85 Å (1 Å=0.1 nm) resolution and revealed a two-domain structure with the catalytic prim/pol domain which has already been determined  and a novel helix-bundle domain with six α-helices . Both domains are connected by a linker encompassing amino acids 250–255 which could not be resolved, indicating flexibility (Figure 2). The three C-terminal helices of the helix-bundle domain superimpose reasonably well (Dali Z-score=5.6) with the C-terminal helix-bundle domain of the primase RepB′ from the plasmid RSF1010 which has been determined recently . Interestingly, the RepB′ protein, although it stems from a bacterial plasmid, also contains a deviant archaeo-eukaryotic primase domain (Figure 1).
To understand the molecular basis of primer synthesis, the structure of a binary, tertiary or quaternary complex of a primase with template and/or nucleotides would be extremely valuable. Unfortunately, the structure of such a complex has not yet been determined, probably reflecting the transient nature of such a complex.
We therefore opted for a structure-guided mutational study and selected a number of conserved aromatic residues and mutated them to alanine. All mutants are active in primer template elongation, stressing that the helix-bundle domain is not involved in the chemistry of the elongation reaction which is carried out solely by the prim/pol domain. However, we discovered that Tyr352 and Trp314 are important for primer synthesis. Proteins with either the Y352A or the W314A mutation are unable to synthesize a primer. The defect of the W314A mutant could be explained by its lower ability to bind template DNA. In addition, the Y352A mutant appears to be involved in a step before dinucleotide formation as no dinucleotide is formed.
Next a deletion mutant devoid of the linker was constructed. Interestingly, this mutant is able to support the synthesis of a dinucleotide; however, the extension of the dinucleotide is not catalysed by this mutant, suggesting that the conformational change which is restricted by the linker deletion is somehow required for the progression of the primer synthesis. More specifically, I suggest that the repositioning of the template–dinucleotide after dinucleotide formation is not possible without the active involvement of the helix-bundle domain (Figure 3).
Hypothetical catalytic cycle of the pRN1 primase
In summary, the pRN1 replication protein is a highly interesting multifunctional replication protein which could carry out a concerted series of reactions during replication initiation. Potentially, the winged-helix DNA-binding domain could bind to the origin, next the helicase domain could unwind the replication origin and deliver the single-stranded DNA to the primase domain for primer synthesis. Structural information is now available for the minimal primase domain. It appears that conformational changes between the catalytic prim/pol domain and the helix-bundle domain aid in dinucleotide formation and primer/template repositioning.
Molecular Biology of Archaea II: A Biochemical Society Focused Meeting held at Robinson College, Cambridge, U.K., 16–18 August 2010. Organized and Edited by Stephen Bell (Oxford, U.K.) and Finn Werner (University College London, U.K.).
This work is supported by the Deutsche Forschungsgemeinschaft [grant numbers Li913/4 and Li913/6].