Archaeal microorganisms, though prokaryotic, are phylogenetically distinct from bacteria 1 and exhibit considerable similarities to eukaryotes in their macromolecular biosynthetic machinery, particularly with respect to their DNA replication system. Among the Archaea, Halobacterium sp. NRC-1 provides an excellent model system to address questions of fundamental DNA replication biology using bioinformatic, genomic, and genetic approaches 2. The genome is relatively small, comprised of a 2 Mbp large chromosome and two minichromosomes, pNRC200 (365 kbp) and pNRC100 (191 kbp), and codes 2,682 putative genes. Of these, only 2,532 genes are unique, due to duplication of 145,428 bp between the two extrachromosomal replicons 3. Halobacterium sp. NRC-1 is easily cultured in the laboratory in hypersaline media containing 4.3 M NaCl and has well-developed genetic methodology, including a facile transformation system, plasmid shuttle vectors, selectable markers, and a directed gene knockout/replacement system 45.

For gene knockouts in the Halobacterium sp. NRC-1 system, we developed a method employing the selectable and counterselectable ura3 gene (Fig. 1) 67. The system also utilizes a suicide plasmid vector with two essential elements, a wild-type copy of the Halobacterium sp. NRC-1 ura3 gene plus its native promoter, and at least 500 bp of 5' and 3' DNA flanking the targeted gene. Transformation of an isogenic Halobacterium sp. NRC-1 strain containing a deletion of the chromosomal ura3 gene with the suicide vector, followed by selection for uracil prototrophy results in an integrated copy of the suicide vector at the genomic locus homologous to the targeted gene. Counterselection for suicide vector loss is accomplished by selection for 5-fluoroorotic acid (Foa) resistance and colonies are then screened via polymerase chain reaction (PCR) to discriminate between knockout and wild-type alleles. Excision of the suicide plasmid vector can occur on the same side as the integration, yielding restoration of the wild-type allele, or excision can occur on the opposite side of the integration, yielding replacement of the wild-type gene with a deletion of the targeted gene. In cases of essential genes, a functional copy of the targeted gene must be provided on a replicating plasmid to recover deletants. This gene knockout system has been successfully employed for studies of several gene clusters in Halobacterium sp. NRC-1 25.

In addition to the gene knockout system, a genetic screen for the isolation of autonomously replicating sequences (ARS) was established for Halobacterium sp. NRC-1. Earlier genetic work identified two likely replication origins in Halobacterium sp. NRC-1 via cloning of ARS elements, one on the large chromosome, and another located within the common region of pNRC100 and pNRC200 89. Sequence analysis of the pNRC minimal replicon showed the requirement of a unique gene, repH, and an AT rich region 5' to the gene. Mutation or deletion of either the AT rich sequence or the repH gene was found to abolish autonomous replication ability of plasmids 9. For the large chromosome, the ARS element was found directly 5' to orc7, one of ten orc/cdc6 genes in the genome, in a region of GC skew polarity switch 10 and global minimum in Z curve analyses 11. However, regions proximal to two other chromosomal orc/cdc6 genes, orc6 and orc8, could not confer autonomous replication ability. The chromosomal ARS region contained unusual sequence elements: a large (33 bp) inverted repeat flanking an AT rich region of 189 bp plus the orc7 gene. Genetic analysis showed that both the inverted repeats, the AT rich region, as well as the orc7 gene were required for autonomous replication ability 8. Work in other archaeal organisms identified chromosomal DNA replication origin(s) comprised of similar sequence elements proximal to orc7 homologs in the genomes of Pyrococcus abyssi 121314 and Sulfolobus spp. 1516.

In addition to genetic studies, predicted replisome components of haloarchaea have been identified via bioinformatic analysis 17. One of the most interesting findings was the presence of a large family of orc/cdc6 genes in Halobacterium sp. NRC-1 and other haloarchaea, homologous to eukaryotic origin recognition complex (ORC) proteins 1, 4, and 5 as well as to the eukaryotic replicative helicase loader Cdc6 (Fig. 2) 818. This finding suggested that multiple Orc proteins in Halobacterium sp. NRC-1 may be required for replication, perhaps through formation of heteromeric protein complexes for origin recognition. Many additional genes coding eukaryotic-type DNA replisome components have also been found, with homology to replicative helicase proteins (MCM), ssDNA binding proteins (RFA), processivity clamp loader proteins (RFC), processivity clamp protein (PCNA), primase proteins, Okazaki fragment maturation proteins (Rad2 and RNaseH), ATP dependant DNA ligase, DNA polymerases (B family), and type IIB topoisomerase (Top6A and B). The novel heterodimeric family D DNA polymerase found only in the euryarchaea is also present in Halobacterium sp. NRC-1 19. A few genes for bacterial-type replication proteins, e.g. a primase (DnaG), and type IA (TopA) and IIA DNA topoisomerases (GyrA and B), are also present 17.

With the availability of an inventory of replication factors likely acting at haloarchaeal DNA replication origins and a facile gene knockout system, we sought to answer basic questions regarding the essentiality of DNA replication gene assignments in an archaeon. The inability to recover deletion mutants indicates the requirement of genes coding two Orc/Cdc6 proteins, two different replicative DNA polymerases, a replicative helicase, a eukaryotic-type primase, a DNA polymerase sliding clamp, and the flap endonuclease. Eight of the orc/cdc6 genes and a polB gene are dispensable to cells. This study shows the first in vivo evidence for genes likely to be critical for DNA replication in Archaea.

Analysis of DNA replication components in archaeal systems has been restricted primarily to bioinformatic analysis and in vitro biochemical characterization. However, in our investigations, we have utilized the power of genetics in Halobacterium sp. NRC-1, to study DNA replication in this model Archaeon. Previously, we defined the cis acting elements required for chromosomal and pNRC100/200 DNA replication 89. In the current study, we have examined the in vivo essentiality of nineteen genes for predicted components of DNA replication initiation and elongation. Ten genes are most likely required, encoding two Orc/Cdc6 origin recognition proteins, two DNA polymerases (one B and both subunits of the D family), four accessory proteins, the replicative helicase protein MCM, primase proteins Pri1/Pri2, processivity clamp protein PCNA, and Okazaki fragment maturation protein Rad2. Taken together, our results provide a better view of the likely in vivo requirements for DNA replication in Halobacterium sp. NRC-1.

Significantly, our study has targeted the largest number of genes for deletions in any archaeon to date 6721222324252627282930313233. For the first time, we have used statistical analysis of gene knockout frequencies and in several cases complementation analysis to critically evaluate the essentiality of genes for which deletions could not be recovered. Statistical analysis showed that the probability of recovering knockout mutants is > 99.999 % in all cases where 40 potential candidates were screened. Where no mutants were observed (orc2, orc10, polD1, polD2, polB1, mcm, pri1, pri2, pcn, and rad2), we have very strong evidence for the requirement of these genes for cell viability. In five cases tested by complementation analysis (polD1, polB1, mcm, pri2, and rad2), knockouts were recovered when a functional copy of the gene was present on a replicating plasmid, confirming that the genes were essential to cells and also dominant in trans. These results provide a genetic system for further analysis of essential DNA replication genes in Halobacterium sp. NRC-1.

Interestingly, we found that only two of ten orc genes encoded in Halobacterium sp. NRC-1 are essential. We had previously hypothesized that orc7 and likely orc6 would be essential for viability, based upon our previous genetic work showing the requirement of orc7 for autonomous replication ability of a minichromosome plasmid replicon 8. Biochemical work performed on an Orc7 ortholog in S. solfataricus 15 and a chromatin immunoprecipitation study in Pyrococcus abyssi 13 are also consistent with the function of Orc7 proteins in chromosomal origin binding proteins in Archaea. However, we found the orc7 gene of Halobacterium sp. NRC-1 to be dispensable under standard growth conditions. Because NRC-1 contains ten orc/cdc6 homologs, it is possible that another gene may be functionally redundant to orc7 in this archaeon. In contrast, Orc7 orthologs are found in a single gene copy in most other Archaea, with the exception of Sulfolobus spp. which have two orc7 orthologs linked to two chromosomal DNA replication origins 1516.

Most Archaea encode an orc6 gene ortholog in their genomes 8, but our genetic analysis shows this gene is also not essential to Halobacterium sp. NRC-1. Sulfolobus spp. Orc6 proteins have been found to bind origin DNA sequences, although in partially synchronized cultures, expression of the Orc6 ortholog appears to be in G₂ phase cells 15. It is possible that Orc6 orthologs act as negative regulators of DNA replication initiation, preventing re-replication by binding to origin sequences and blocking binding of replication initiation factors. Both the Orc7 and Orc6 orthologs from Methanothermobacter thermoautotrophicus have also been shown to interact with MCM and inhibit helicase activity, with the Orc6 ortholog being a more potent inhibitor 3435. It is also possible that Orc6 orthologs in Archaea act as Cdc6 does in eukaryotes, recruiting the replicative helicase complex to DNA replication origins. In Halobacterium sp. NRC-1, the orc6 gene is not essential for viability and no discernable phenotypes are observed when it is deleted, possibly as a result of functional redundancy.

Surprisingly, we found orc10 on the large chromosome, and orc2 on pNRC200 are essential. Although these genes are not found to be conserved in the genomes of non-halophilic Archaea, there are likely orthologs and paralogs found in all halophilic Archaea. Orc10 shares 50% sequence similarity to the non-essential Orc8 protein from Halobacterium sp. NRC-1. It also shares sequence similarity to Cdc6-3 from N. pharaonis and eight homologs from H. marismortui, including a previously unrecognized Orc/Cdc6 homolog on the pNG500 replicon (Fig. 2), and at least three homologs from Haloferax volcanii (data not shown). Interestingly, the orc10 genetic locus harbors an ISH12 element 100 bp from the orc10 predicted translational start codon and is also an area of the large chromosome with extrachromosomal characteristics, e.g. an increased AT% and a higher concentration of IS elements 10. Orc2 is over 90 % identical in amino acid sequence to Orc4 and shares sequence homology with Orc5 and Orc3 from Halobacterium sp. NRC-1, and forms a clade with Cdc6-5 from N. pharaonis, four homologs from H. marismortui, and at least seven homologs from H. volcanii (data not shown). At this time we cannot strictly state that the orc10 and orc2 genes are essential for DNA replication, only that they are essential for viability of Halobacterium sp. NRC-1, although their homology to the other haloarchaeal, archaeal, and eukaryotic orc/cdc6 genes would strongly indicate that they are involved in some essential and thus far uniquely haloarchaeal role in DNA replication (Fig. 2). It is tempting to speculate that these two orc gene products play an important role in coordinating cell cycle and DNA replication of the chromosome and extrachromosomal replicons in Halobacterium sp. NRC-1. It is possible that they function as the origin binding proteins for the large chromosome and pNRC200, respectively, or they may be required to recruit the replicative helicase, or additional replisome components in haloarchaea. Moreover, our recent unpublished work has shown that the orc10 and orc2 genes are essential in mutants harboring multiple orc gene knockouts, while also indicating that some orc gene products are non-essential even in strains already having knockouts of other orc genes.

All sequenced haloarchaea to date contain at least one homolog in each of the five Orc/Cdc6 phylogenetic clades (Fig. 2). The large haloarchaeal orc/cdc6 gene family may therefore represent an evolutionary scenario, similar to eukaryotes, in which gene duplication events followed by functional divergence have led to evolution of heteromeric protein complexes for origin recognition. With discrimination of essential vs. non-essential orc genes, it will be interesting to determine if heteromeric Orc/Cdc6 complexes form in Halobacterium sp. NRC-1 and to identify specific functions and interactions of individual gene products.

Our results also show that two replicative-type DNA polymerases are absolutely required for Halobacterium sp. NRC-1. Both of the chromosomally encoded DNA polymerases, the B family polB1 polymerase, and the D family polD1/polD2 polymerase, are essential. From in vitro biochemical characteristics determined with the Pyrococcus B and D family DNA polymerases 3637, it would appear that the euryarchaeal specific heterodimeric D family polymerase, PolD1/PolD2, may act at the lagging strand and the B family polymerase, PolB1 may act at the leading strand. The B family polymerase can only use DNA primers for extension, while the D family polymerase can use either RNA or DNA primers for extension, though it requires PCNA for efficient DNA synthesis 38. However, these points are speculative and require more direct genetic and biochemical experiments to confirm.

The non-essentiality of polB2 is also interesting. PolB2 contains the ten conserved polymerase and exonuclease motifs of archaeal B family DNA polymerases (data not shown), so it would appear to be a functional DNA polymerase. A PolB2 homolog is also found in the genome of the distantly related halophile, H. marismortui, on extrachromosomal replicon pNG600. Of interest, as well, is the fact that in both Halobacterium sp. NRC-1 and H. marismortui, the polB2 gene is divergently oriented with respect to an Orc5 clade member gene 17. The function of this evolutionarily conserved genetic linkage between polB2 and an Orc5 clade member gene in these two haloarchaea is currently unknown, but in Halobacterium sp. NRC-1 both orc4 and polB2 are non-essential genes. While much in vitro work has been directed at determining the properties of archaeal DNA polymerases, especially since the discovery of a novel DNA polymerase family in euryarchaea 19, no in vivo analysis had previously been performed to determine whether these DNA polymerase family members were essential, consistent with a requirement for DNA replication.

For the other five accessory genes examined here, whose products comprise four protein complexes, the results were as expected: mcm, pri1, pri2, pcn, and rad2 are essential for normal growth of Halobacterium sp. NRC-1. The in vitro biochemical work done on these various gene products had indicated that it was likely that they would function in an analogous manner to their eukaryotic homologs. Though no biochemical work has been done on the haloarchaeal MCM, our genetic analysis is consistent with its predicted function as a replicative helicase. With Pri1/Pri2 (homologs of the eukaryotic p48 and p58 proteins), the archaeal complex likely acts as the DNA-dependent RNA primase for DNA replication. The finding of the essential nature of the pri1 and pri2 genes in Halobacterium sp. NRC-1 is consistent with their role as a replicative primase. In contrast, the function of the bacterial-type primase, DnaG, coded by most archaeal genomes, including Halobacterium sp. NRC-1 is unknown, although in S. solfataricus it has been reported to be associated with the archaeal exosome 39. For PCNA, the function is likely to be as a DNA polymerase sliding clamp. While most Archaea possess a single gene for pcn, similar to eukaryotes, two crenarchaea, S. solfataricus and Aeropyrum pernix, are exceptions, with three pcn genes each, reminiscent of the eukaryotic 9-1-1 complex 4041. In Halobacterium sp. NRC-1, we have found that the single pcn gene is essential, consistent with PCNA acting as the homotrimeric DNA polymerase sliding clamp. Rad2 family flap endonucleases are important in both the processes of DNA replication, (during Okazaki fragment maturation), and repair (in nucleotide excision repair). Organisms can possess multiple homologs, although just a single flap endonuclease gene was detected in the genome of Halobacterium sp. NRC-1 17. Genetic studies in yeast indicate that rad27, the rad2/FEN1 homolog in S. cerevisiae, is not essential unless a recombination gene (e.g. rad51 or exo1) is also deleted 42. In the present investigation, we have shown that the rad2 gene is essential for viability of Halobacterium sp. NRC-1. This finding is consistent with flap endonucleases being required for DNA replication via their role in Okazaki fragment maturation in this archaeon.

The results obtained in this and a previous investigation 8 are relevant to most other archaeal organisms, with the large orc gene family representing a unique aspect of DNA replication in haloarchaea. In our emerging model, archaeal chromosomal DNA replication origins are comprised of a large inverted repeat flanking an AT rich DNA sequence proximal to the gene encoding an origin binding protein, an orc/cdc6 gene that is an orc7 ortholog. These large inverted repeats likely serve as binding sequences for the origin binding protein, probably Orc7, although a multimeric ORC complex or other Orc proteins, especially the orc2 and orc10 gene products cannot be ruled out. Binding of origin recognition protein(s) would lead to local DNA helix destabilization of the intervening AT rich region allowing for recruitment of the essential mcm gene-coded replicative helicase complex, potentially by the orc6 gene product, followed by association of other replisome components, such as the essential eukaryotic-type primase (pri1/pri2 gene products). Once the primase lays down an RNA primer at the origin, the essential pcn gene product may be loaded onto the primed template and essential B (polB1) and D (polD1/polD2) family replicative DNA polymerases. The rad2 gene product encodes the likely flap endonuclease which helps to mature Okazaki fragments. During the replication process, the polB1 gene product coding the B family DNA polymerase may act as the leading strand DNA polymerase and the polD1 and polD2 gene products coding the D family DNA polymerase may act as the lagging strand DNA polymerase for processive and faithful duplication of the genome.

By utilizing a well developed in-frame gene knockout system in Halobacterium sp. NRC-1, we have established a foundation on which to explore further the in vivo roles of these DNA replication genes. With facile genetics, complete genome sequence, and established post-genomic methodologies, Halobacterium sp. NRC-1 provides an excellent model system to further study the characteristics of archaeal DNA replication. In addition, the gene knockout and complementation methodology used for studying DNA replication in Halobacterium sp. NRC-1 may be applied to the investigation of many other aspects of archaeal biology 2.

The authors declare that they have no competing interests.

BRB performed research, with assistance from PD, and drafted the manuscript. SD supervised the research, including design, data analysis, and finalized the manuscript, with assistance from PD.