The pRN1 plasmid is a rather small multicopy plasmid which was isolated from a Sulfolobus islandicus strain in 1993 by Wolfram Zillig and co-workers. Sequence analysis of the genome sequence suggested that three conserved genes are important for plasmid replication. These genes code for two sequence-specific DNA-binding proteins (ORF56 and ORF80) and for a large multifunctional replication protein (ORF904). The protein ORF904 has primase, DNA polymerase and helicase activity. Remarkably, the primase activity is highly sequence specific, and primers are only efficiently synthesized on templates with the motif GTG. This protein could initiate the plasmid replication by melting the double-stranded DNA at the origin of replication and by synthesizing the first primers at the replication bubble. The protein ORF56 is a repressor, and combined biochemical and genetic evidence shows that this protein is involved in regulating the copy number of the plasmid. The function of the third conserved protein, ORF80, is still mysterious. Although this protein is highly conserved, it is not essential for replication, since shuttle vectors with a deleted orf80 gene are still able to replicate in Sulfolobus. Interestingly, plasmids lacking the orf80 gene display reduced plasmid retention under non-selective conditions, raising the possibility that ORF80 is involved in plasmid partitioning or has an accessory role in plasmid replication.
The pRN1 plasmid and the plasmid family pRN
The pRN1 plasmid is a small multicopy plasmid which was isolated from a Sulfolobus strain which was sampled by Zillig et al.  in solfataric fields of Iceland in the early 1990s. This plasmid was one of the first archaeal plasmids to be isolated and the first crenarchaeal plasmid which was sequenced . In the original publication, five rather small open reading frames (less than 100 amino acids) and one large open reading frame (904 amino acids) were proposed. The sequence analysis also revealed that pRN1 was distinct from all known bacterial plasmids. Nevertheless, by indirect evidence, it was suggested that pRN1 is a plasmid which replicates via a rolling circle .
Later on, additional crenarchaeal plasmids (e.g. pRN2, pDL10 and pHEN7) were sequenced, and it turned out that these plasmids formed a novel plasmid family, the pRN plasmid family . Plasmids of this family share a core of two highly conserved genes and an additional less-conserved gene, which are their hallmarks. In the case of pRN1, they are called orf904, orf80 and orf56, with the numbers reflecting the length of the proteins which they encode (Figure 1). In addition, the pRN plasmids contain a number of small and not conserved open reading frames which are probably not translated into proteins (e.g., for pRN1, these are orf72, orf90a and orf90b). Currently, it is not known whether the pRN plasmid genomes harbour genes for regulatory RNA. It is noteworthy that, besides the extrachromosomal plasmids, some pRN-type plasmids have been detected integrated into the genome of Sulfolobus solfataricus and Sulfolobus tokodaii [3,4].
Plasmid map of pRN1
My group has studied the pRN1 plasmid. In order to understand the replication of this novel plasmid family, we characterized the three proteins encoded by pRN1. The proteins were overexpressed in Escherichia coli and purified, and their activities were analysed.
Biochemical characterization of the three conserved proteins
The protein ORF56 has low sequence similarity with several strand–helix–turn–helix DNA-binding proteins. These proteins with the prototypical arc repressor are dimers and contact the major grove of the DNA through an N-terminal β-strand formed by both subunits of the dimer.
The ORF56 DNA-binding site was identified and found to be an inverted repeat upstream of its own gene. ORF56 assembles as a tetramer on its target binding site . In addition, reporter gene experiments suggested that ORF56 acts as a repressor . Since the genes orf56 and orf904 are co-transcribed and the latter gene encodes the replication protein, this type of negative-feedback control could be part of the copy-control mechanism of the plasmid. We therefore suggested that the physiological function of the protein ORF56 is to regulate the copy number of pRN1. This protein has also been named CopG in analogy to the copy number control proteins of rolling circle plasmids, such as the CopG protein from the plasmid pLS1 
Sequence analysis of the ORF904 protein of the pRN1 plasmid immediately reveals that the C-terminal half of the protein encompasses a conserved superfamily 3 helicase domain (Figure 2). These helicases form a distinct group of helicases and are often found in bacteriophages and viruses. The most prominent member is the large T-antigen of the SV40 (simian virus 40). The superfamily 3 helicases form hexameric rings and may unwind DNA by steric exclusion: only one strand of the double-stranded DNA strand is directionally transported through the central pore. By this mechanism, the two strands of the DNA are separated. However, the molecular mechanism of unwinding is a matter of debate, and it is also possible that double-stranded DNA is transported through the central pore and that the unwinding takes place in the context of a double hexameric assembly .
Domain structure of the replication protein ORF904 and representative homologous structures
Biochemical analysis of recombinant ORF904 has demonstrated that ORF904 has a weak unwinding activity and is able to translocate 3′→5′ on single-stranded DNA (M. Sanchez and G. Lipps, unpublished work).
In addition to the C-terminal superfamily 3 helicase domain, ORF904 appears to have two additional domains: an N-terminal prim/pol (primase/polymerase) domain and a C-terminal winged-helix DNA-binding domain (Figure 2).
Although the N-terminal half of ORF904 is highly conserved and probably carries out an important function, it was impossible to predict the function of this part of the protein owing to a lack of similarity to characterized proteins. This region of the protein has only marginal sequence similarity to several uncharacterized bacteriophage proteins. Surprisingly, we could demonstrate that this part of the protein has primase and DNA polymerase activity. Interestingly, the primase activity strongly prefers dNTPs for primer synthesis, which is quite unusual . Later on, a more detailed analysis revealed that a ribonucleotide is incorporated as the first base in the primer. Therefore the primers synthesized by ORF904 consist of a ribonucleotide followed by seven deoxynucleobases. Moreover, the primase activity of ORF904 is highly sequence-specific and requires the motif GTG in the template in order to synthesize a primer . This high sequence specificity is also rather unusual. The domain responsible for the primase and DNA polymerase activities encompasses the first 255 amino acids of the replication protein ORF904 and has been termed a prim/pol domain (Pfam PF09250). Unexpectedly the crystal structure revealed that this domain has an RRM (RNA recognition motif) fold and that the core of the domain is shared with the small catalytic subunit of the archaeo-eukaryotic primases and the polymerase domain of the bacterial LigD proteins involved into non-homologous end-joining repair of double-strand breaks. A study by Koonin and co-workers  confirmed our previous suggestion that the prim/pol domain and archaeal primase domain have a common ancestor. Although the sequences of both domains appear to be unrelated, the structures suggest that both domains are homologues.
A third domain of the replication protein ORF904 was identified recently by us. At the C-terminus, a region of about 100 amino acids appears to fold into a winged-helix DNA binding domain. Our results show that the isolated winged-helix domain of ORF904 has weak DNA-binding activity (M. Sanchez and G. Lipps, unpublished work). It is currently unknown whether this domain also has a sequence-specific DNA-binding activity.
The third conserved gene of the plasmid family pRN is the orf80 gene. This gene is found not only within the pRN plasmid family, but also on viruses and conjugative plasmids of the Crenarchaeota. We could demonstrate that this protein is a sequence-specific DNA-binding protein. The binding site of this protein was identified and found to be the palindromic motif TTAAN7TTAA . On pRN1, as well as the other genetic elements encoding an ORF80 homologue, two highly similar binding sites with a distance of approx. 60 bp are present. Possibly, the protein assembles into a larger nucleoprotein complex encompassing both binding sites. The exact functional role of this highly conserved protein remains obscure, however. We have suggested that ORF80 recognizes the replication origin and is therefore required for replication initiation. Recent results, however, have demonstrated that ORF80 is not essential for plasmid replication and therefore negate an important role for plasmid replication .
Altogether, the biochemical characterizations of the proteins encoded by the pRN1 plasmid suggest that ORF56 participates at the plasmidal copy control and that ORF904 is a multifunctional replication enzyme. This enzyme has helicase activity and a highly sequence-specific primase activity. With these two activities, the replication enzyme could melt the replication origin and synthesize the primers at the replication bubble. However, the detailed molecular mechanisms of origin recognition and origin unwinding and the location of the replication origin are not known. Typically, plasmidal replication origins are AT-rich and contain a number of iterons. On the basis of these criteria we were, however, not successful in identifying a replication origin. It is also possible that ORF904 has a function beyond replication initiation. With its helicase and DNA polymerase activities, ORF904 could, in principle, possess a function at the moving replication fork.
Genetic investigations on the pRN1 plasmid
For a more detailed study of the plasmid replication, more sophisticated methods than pure in vitro biochemical analysis are required. One approach to study the replication of the pRN1 plasmid is to introduce mutated plasmids into Sulfolobus cells and to study the effect of the mutation. This reverse genetic approach would allow the study of replication in vivo, but requires the successful construction of a Sulfolobus–E. coli shuttle vector which was not achieved until recently. The shuttle vectors constructed by our group replicate in E. coli under ampicillin selection and in a Sulfolobus acidocaldarius uracil auxotroph mutant under selection with uracil-free growth medium .
In the course of shuttle vector development, we constructed a number of plasmids which differed with respect to the integration site of the E. coli replicon and the Sulfolobus selection marker within the pRN1 plasmid. Integration within the orf56/orf904 operon was not tolerated as expected, but, surprisingly, the interruption of the highly conserved orf80 gene was allowed (Figure 3). Thus this gene is not essential for the plasmid (also see above).
Essential region of the pRN1 plasmid
These shuttle vectors now allow the construction of variants in E. coli and for them to be transformed into Sulfolobus cells for functional studies. In order to define the minimal replicon of pRN1, we constructed various deletion mutants and found out that more than a third of the native pRN1 sequence can be deleted (G. Lipps, unpublished work) (Figure 3). The minimal replicon of pRN1 appears to consist of only the operon orf56/orf904. However, a more detailed analysis suggested that a potential stem–loop region just downstream of the replication gene is also required (S. Berkner and G. Lipps, unpublished work). Possibly, this part of the plasmid could be the replication origin of pRN1.
To investigate the copy-control regulation of pRN1, we also introduced mutants into the orf56 promoter and the orf56 gene sequence. Preliminary experiments suggest that mutations destroying the ORF56-binding site and the deletion of the orf56 gene are not tolerated. In contrast, less severe mutation within the orf56 promoter, the ORF56-binding site and the orf56 gene appear to modulate the copy number of the plasmid (C. Lindemann and G. Lipps, unpublished work).
Plasmids are fundamentally important to investigate the molecular biology of various organisms and are instrumental in conducting genetic studies. Shuttle vectors based on the Haloferax volcanii plasmid pHV2 and pHH1 replicate in Halobacterium species, Haloferax volcanii and Haloarcula marismortui . The plasmid pC2A from Methanosarcina acetivorans was the backbone for the construction of shuttle vectors for several Methanosarcina species. Recently, progress has been made in constructing shuttle vectors for hyperthermophiles. A construct based on the plasmid pTN1 from Thermococcus nautilus was successfully applied to transform Thermococcus kodakaraensis [15,16]
Notwithstanding, there is a clear lack of understanding of the molecular biology of archaeal plasmids. For most plasmids, it is not known how the plasmids replicate and how their replication is regulated. In addition, for most archaeal plasmids, the benefit for the host organism remains to be identified.
My work is funded by the Deutsche Forschungsgemeinschaft [grant numbers Li913/3–Li913/6].
Molecular Biology of Archaea: Biochemical Society Focused Meeting held at University of St Andrews, U.K., 19–21 August 2008. Organized and Edited by Stephen Bell (Oxford, U.K.) and Malcolm White (St Andrews, U.K.).