Introduction

Chromosome replication starts at specific sites known as origins, where initiator proteins bind and recruit replication machinery components1,2. Bacteria use a single origin, which is usually adjacent to the initiator gene dnaA, to replicate their chromosome, whereas eukaryotes perform chromosome replication with multiple replication origins that are recognized by origin recognition complexes2,3. Dormant origins are replication origins that are not used in a normal cell cycle4. If DNA encounters replication stress or if adjacent origins fail to fire, dormant origins can be activated to help complete chromosome replication4,5,6, providing an important mechanism for cells to deal with replication stress and to maintain genomic stability7. For organisms with multiple replication origins on a chromosome, the coordination of origin utilization is vital to ensure complete and accurate genome duplication.

Chromosome replication using multiple replication origins has been widely recognized in archaea. The combination of an autonomously replicating sequence (ARS) assay or two-dimensional gel electrophoresis with marker frequency analysis (MFA) has led to the identification of multiple replication origins in Sulfolobus species8,9,10, haloarchaea11,14, Aeropyrum pernix15 and Pyrobaculum calidifontis16, thus greatly expanding our knowledge of archaeal replication origins. Archaeal replication origins generally contain an AT-rich region flanked by conserved repetitive DNA sequences designated as origin recognition boxes (ORBs)14,22, similar to the presence of dormant origins in eukaryotes. However, no dormant origin has been experimentally identified in archaea.

The origin usage in one cell cycle was investigated in Sulfolobus acidocaldarius, which possesses three origins on its chromosome. The three replication origins fire once per cell cycle, with oriC1 and oriC3 firing slightly earlier than oriC2 (ref. 23). For the three origins in S. islandicus and the two origins on the chromosome of H. hispanica, gene knockout experiments demonstrated that no single origin was essential, but at least one orc1/cdc6 gene (for S. islandicus) or one origin (for H. hispanica) was required for chromosome replication10,11,22, but no dormant origins have been reported in previous investigations. In the case of H. mediterranei, the two identified dormant origins oriC4-cdc6H and oriP1-cdc6L displayed the common features that both initiator genes were adjacent to the inverted ORB motifs and that no peaks of replication initiation emerged at the corresponding locations in the replication profiles of the wild-type strain. When these features were used to evaluate the potential origins in the H. volcanii main chromosome and in pHV4, which is integrated into the main chromosome of H. volcanii H26 (ref. 22), three orc1/cdc6 genes associated with adjacent intergenic regions were found to fulfil these criteria: orc4 on the main chromosome, and orc13 and orc7 on pHV4 (refs 11, no transformants were observed for the corresponding orc4 fragment transformation. However, the other three fragments, oriP1-orc3, oriP2-orc13 and oriP3-orc7, had autonomous replicating ability in H. mediterranei (Fig. 7a,b).

Figure 7: The dormant origin oriP2-orc13 from H. volcanii can initiate the replication of the entire H. mediterranei chromosome.
figure 7

(a) ARS assays for H. volcanii oriP1-orc3, orc4+ and the potential dormant origins oriP2-orc13 and oriP3-orc7. Plasmids pBI501-oriP1-orc3, pBI501-oriP2-orc13, pBI501-oriP3-orc7and pBI501-orc4+ (Supplementary Table 1) were transferred into DF50ΔoriP2-cdc6M. Fragment orc4+ refers to the orc4 gene plus the adjacent upstream intergenic region. Colonies were observed after 3 days at 37 °C. (b) Southern blot analysis using the bla gene as the probe. Lane P contains the plasmid pBI501-oriP3-orc7 isolated from E. coli, whereas the other lanes contain crude DNA extracted from the corresponding H. mediterranei transformants. (c) Schematic diagram of the replication origin of the H. mediterranei H13 chromosome. H13 was constructed by replacing the oriC4-cdc6H region in DFA50ΔoriC1ΔoriC2ΔoriC3 (DFA50Δ123) with the oriP2-orc13 region from H. volcanii. (d) The origin-replaced mutant H13 was confirmed by PCR analyses, with DF50 and DF50ΔoriC1ΔoriC2ΔoriC3 (Δ123) as controls. The primers were specific for oriC1, oriC2, oriC3, oriC4, oriP2-orc13 and 7S; 7S was used as an internal control. (e) The growth phenotype of H13 compared with DF50ΔoriC1ΔoriC2ΔoriC3 (Δ123). Serial dilutions of stationary phase cells (equal amounts) were spotted onto AS-168 plates supplemented with uracil. (f) Marker frequency distribution for the H13 chromosome. High-throughput sequencing was performed for the total DNA from the exponential phase of H13, and the number of reads was calculated for every 1-kb window and plotted against the chromosome position. The position of the replication origin oriP2-orc13 is marked with an arrow.

Full size image

To test the initiating activity of the putative dormant origins from H. volcanii for chromosome replication in H. mediterranei, we attempted to replace the only origin (oriC4-cdc6H) on the chromosome of H. mediterranei DFA50ΔoriC1ΔoriC2ΔoriC3 with the H. volcanii oriP2-orc13 origin. Considering the autonomous replicating ability of oriP2-orc13, we expected that knocking in the oriP2-orc13 fragment into the chromosome via a plasmid vector would be difficult. Thus, a linear DNA fragment containing oriP2-orc13 and a trpA marker (for positive selection of the replacement) was used for gene replacement. Replacement of the oriP2-orc13 origin at the oriC4-cdc6H position (Fig. 7c) was successfully conducted, as confirmed by both PCR (Fig. 7d) and Southern blot analysis (Supplementary Fig. 6), and the mutant strain was designated H. mediterranei H13 (Fig. 7c). Growth of the H13 strain containing only the oriP2-orc13 replication origin was comparable to that of DF50ΔoriC1ΔoriC2ΔoriC3 (Fig. 7e). Notably, the genome-wide replication profile of H13 revealed that chromosome replication was initiated only from oriP2-orc13 (Fig. 7f). Thus, the putative dormant origin oriP2-orc13 from H. volcanii can be used to efficiently replicate the entire chromosome of H. mediterranei in the absence of the other active origins.

Discussion

The initiation of chromosome replication from multiple origins is very common in archaea. Based on the conserved characteristics of archaeal origins, multiple cdc6-associated replication origins have been predicted in 15 fully sequenced haloarchaeal genomes14,22. Therefore, it is likely that dormant replication origins are widely distributed in haloarchaea. Because dormant origins can be activated for replicon replication in the absence of other active origins, they can act as backup sites for replication initiation when surrounding origins are inactivated or when replication forks stall, as has been demonstrated for dormant origins in eukaryotes27,28. In contrast to active origins, dormant origins in haloarchaea were proposed to have been recently acquired (Fig. 4), possibly from environmental plasmids, viruses or other haloarchaea. Interestingly, similar to our results that the recently acquired dormant origin is competent to replicate the entire chromosome of the origin-deleted H. mediterranei, suppression of the initiation defect of chromosome replication (for example, in the dnaA or oriC-deleted mutants) by plasmid or prophage integration has also been reported in bacteria29,30. As horizontal gene transfer among haloarchaea is common and the acquisition of replication origins is favourable for the transfer of new genetic content31,32, we propose that the acquired origins that accompanied foreign genetic contents are important for haloarchaea to shape the chromosomal structure and to adapt to harsh and variable environments. In addition, these recently acquired origins can be active or dormant as a result of different intracellular and extracellular conditions; this may act as an additional adaptive feature for these haloarchaea.

Interestingly, oriC4 is dormant in wild-type H. mediterranei but is significantly activated in DF50ΔoriC1ΔoriC2ΔoriC3. In eukaryotes, competition among origins for limiting replication factors such as CDC45, Sld2 and Sld3 determines their initiation time and efficiency33,34. Because the H. mediterranei chromosome is polyploid25, as observed in other haloarchaea35, and all origins share the common helicase MCM, MCM may be one of the limiting factors, as previously proposed for H. volcanii22. Therefore, we can speculate that the ability of oriC4-cdc6H to recruit MCM may be weaker than that of other active origins, resulting in the dormant state of oriC4 in the wild-type strain. The deletion of active origins would liberate more MCM molecules, allowing Cdc6H to recruit MCM and initiate replication at oriC4. The gradually increased activation of oriC4-cdc6H in the origin deletion strains supports this hypothesis (Fig. 2d). In particular, the firing ability of oriC4 appears to be suppressed by oriC1, the most efficient origin on the chromosome, as oriC4 could be slightly activated in the absence of oriC1. Interestingly, according to several phylogenetic analyses of Cdc6, Cdc6A is the most conserved Cdc6 and is distributed among all haloarchaeal genomes as well as other archaeal genomes11 were selected as the ORB elements. Only the intergenic regions with inverted ORB elements were considered as replication origins. To analyse the distribution of H. mediterranei-type replication origins in the Haloferax species, the BlastP programme (BLOSUM62 matrix) was used, comparing origin-associated Cdc6 proteins from H. mediterranei with those from certain Haloferax genomes (http://blast.ncbi.nlm.nih.gov/)49. Only those Cdc6 proteins with identities >80% (ref. 12) were recognized.

Genome resources and comparative genomics

Two completed genomes, H. mediterranei ATCC 33500 and H. volcanii DS2 (ref. 50), and the contigs of 13 draft genomes, H. sulfurifonti ATCC BAA 897 (ref. 51), H. mucosum ATCC BAA 1512 (ref. 51), H. denitrificans ATCC 35960 (ref. 51), H. prahovense DSM 18310, H. larsenii JCM 13917, H. gibbonsii ATCC 33959, H. elongans ATCC BAA-1513, H. alexandrinus JCM 10717, H. lucentense DSM 14919, Haloferax sp. ATCC BAA-644, Haloferax sp. ATCC BAA-645, Haloferax sp. ATCC BAA-646 and Haloferax sp. BAB-2207, were available through NCBI (http://www.ncbi.nlm.nih.gov/genome/). All comparative genomic analyses were performed with the CGView Server52 using the default parameters (http://stothard.afns.ualberta.ca/cgview_server/).

Additional information

Accession codes: The DNA microarray and high-throughput sequencing data have been deposited in the NCBI GEO library under accession numbers GSE70597 and GSE70592.

How to cite this article: Yang, H. et al. Activation of a dormant replication origin is essential for Haloferax mediterranei lacking the primary origins. Nat. Commun. 6:8321 doi: 10.1038/ncomms9321 (2015).