Recently, a novel gene-deletion method was developed for the crenarchaeal model Sulfolobus islandicus, which is a suitable tool for addressing gene essentiality in depth. Using this technique, we have investigated functions of putative DNA repair genes by constructing deletion mutants and studying their phenotype. We found that this archaeon may not encode a eukarya-type of NER (nucleotide excision repair) pathway because depleting each of the eukaryal NER homologues XPD, XPB and XPF did not impair the DNA repair capacity in their mutants. However, among seven homologous recombination proteins, including RadA, Hel308/Hjm, Rad50, Mre11, HerA, NurA and Hjc, only the Hjc nuclease is dispensable for cell viability. Sulfolobus encodes redundant BER (base excision repair) enzymes such as two uracil DNA glycosylases and two putative apurinic/apyrimidinic lyases, but inactivation of one of the redundant enzymes already impaired cell growth, highlighting their important roles in archaeal DNA repair. Systematically characterizing these mutants and generating mutants lacking two or more DNA repair genes will yield further insights into the genetic mechanisms of DNA repair in this model organism.
- base excision repair (BER)
- genetic manipulation
- homologous recombination (HR)
- nucleotide excision repair (NER)
- Sulfolobus islandicus
Organisms belonging to the genus Sulfolobus are aerobic heterotrophs thriving in terrestrial acidic hot springs worldwide. Although requiring very high temperatures (75–85°C) for growth, these organisms are genetically tractable as they form colonies on solid media solidified with Gelrite. Consequently, Sulfolobus species are model organisms for studying biological principles in Archaea, the third domain of life, as well as for studying molecular mechanisms at extremely high temperatures. Furthermore, Sulfolobus organisms dominate in terrestrial hot springs that are well separated from one another geologically on Earth, rendering them good models for studying geomicrobiology and evolution.
Sulfolobus is one of the archaeal genera that have been characterized extensively by genome sequencing. The first three complete genomes, i.e. Sulfolobus solfataricus  Sulfolobus tokodaii  and Sulfolobus acidocaldarius , provide blueprints for biochemical characterization of prominent archaeal enzymes. Then the genomes of several Sulfolobus islandicus strains isolated from hot springs in Russia and U.S.A. were sequenced . This was followed by the publication of S. islandicus strains REY15A and HVE10/4 isolated from hot springs in Iceland , both of which are useful hosts for studying archaeal genetic elements [6,7].
For the last decade, there have been consistent efforts in the Sulfolobus research community to develop genetic manipulation systems for three Sulfolobus species including S. solfataricus, S. islandicus and S. acidocaldarius. The initial genetic systems were based on insertion sequence-induced auxotrophic mutants of either pyrE/F genes encoding orotidine-5′-monophosphate pyrophosphorylase and orotidine-5′-monophosphate decarboxylase respectively, or lacS gene encoding β-glycosidase, but exploiting these systems for genetic studies yielded inconsistent results in different laboratories . We worked on developing efficient genetic systems for S. islandicus REY15A. Screening of a large number of spontaneous pyrEF mutants led to the isolation of three mutants carrying a deletion mutation, which were employed as genetic hosts to test for Sulfolobus–Escherichia coli shuttle vectors . We found that employing deletion mutants as genetic hosts for transformation eliminated false positive transformants due to the reverse mutation occurred for host cells. Subsequently, very versatile genetic systems have been established for S. islandicus REY15A, including conventional and novel genetic manipulation methods [10,11], a reporter gene system , plasmid interference assay to study a novel antiviral system named CRISPR (clustered regularly interspaced palindromic repeats)  and a protein-overexpression system , the last of which has been widely used in recombinant protein production in Sulfolobus for research [15–18]. In the present paper, we review the established methods of genetic manipulation and the progress in using them to study the functions of DNA repair genes.
Conventional gene deletion methods for S. islandicus
First, three conventional gene-knockout methods have been established for S. islandicus, including: (i) AR (allelic replacement) of a target gene with the pyrEF marker gene, (ii) the PIS (plasmid integration and segregation (also named pop-in/pop-out) method, and (iii) the MRL (marker replacement and looping out) method. The last two approaches produce mutants with unmarked deletion of target genes such that genetic complementation can be conducted with the same selectable marker used for genetic manipulation [10,19]. These methods were employed to construct knockouts of a few non-essential genes for which mutants were obtained (Table 1). More recently, simvastatin, an antibiotic, was developed as a general selection marker for S. islandicus . As this application no longer requires an auxotrophic mutant to be used as the host, genetic complementation of AR mutants where the pyrEF marker is remained can be now conducted with this antibiotic marker. In fact, this marker has been employed for gene knockout in another S. islandicus strain . It is important to point out that, as this selection is based on overexpression of the gene encoding a 3-hydroxy-3-methylglutaryl-CoA, this selection is much less stringent than the complementation of uracil deficiency from expression of pyrEF genes.
In the attempts to construct knockouts of topR genes encoding a reverse gyrase, we encountered the problem that mutants were not obtainable in S. islandicus. No transformants were yielded from AR, whereas only wild-type cells were recovered from counterselection of PIS recombinants, although the mutant and wild-type strains should each comprise theoretically 50% of colonies formed on selective plates. For genetic manipulation with MRL, the marker gene should replace the target gene at the first step as for the AR procedure. Should the target gene be essential, no transformants are to be obtained. Unexpectedly, transformants were obtained from MRL. Analysing these transformants showed that their recombinant alleles were formed by single-crossover recombination, devoid of the designed looping out arm and, consequently, these transformants no longer had the potential to generate the designed mutant . This peculiar recombinant allele, together with the recombinant allele formed from double-crossover recombination, was observed in the construction of the lacS deletion mutant with MRL and we have demonstrated that only the transformant of the latter category yielded lacS deletion mutants . In principle, the same scenario could occur for genetic manipulation of any other microbes with MRL, although there have not yet been any reports of this phenomenon in the current literature. In any case, none of the above methodology can demonstrate unambiguously that S. islandicus reverse gyrase genes are essential. Interestingly, topR is not essential in the hyperthermophilic euryarchaeon Thermococcus kodakarensis . Since reverse gyrase is the only enzyme conserved among hyperthermophiles  and it was predicted to be very important for any hyperthermophilic organism , it was very tempting to investigate whether the Sulfolobus reverse gyrase is also dispensable for cell viability.
Novel gene-knockout method and construction of gene-deletion mutants of S. islandicus
We designed a novel gene-deletion scheme, namely MID (marker insertion and target gene deletion)  for constructing an S. islandicus mutant lacking a reverse gyrase gene. The MID recombination scheme is shown in Figure 1. Donor DNA, either a linear plasmid or PCR product, carries carefully designed homologous arms such that the target gene in the recombinant cells remains active and therefore transformants are deemed to form colonies on uracil-free selective plates. When growing in uracil-free media that select for transformant cells, the marker cassette loops out constantly from the chromosomes of transformant cells, generating deletion mutant cells, which, although unable to grow, remain in the culture as the pyrEF selection is not based on cell killing. When the counterselection is enforced, mutant cells will grow and form colonies on the selective plates, whereas cells carrying the recombinant target gene allele will be selectively killed by 5-FOA (5-fluoro-orotic acid). The experimental procedure is facilitated by including lacS in the marker cassette that allows phenotypic identification of transformants with X-gal (5-bromo-4-chloroindol-3-yl β-D-galactopyranoside). Recently, the MID strategy has been extended to genetic manipulations of S. acidocaldarius  and, in theory, can be applied to genetic manipulation of any micro-organisms that are proficient in HR (homologous recombination).
In our early experiments, several mutants lacking the gene coding for one of the putative uracil DNA glycosylases, an error-prone DNA polymerase or a topoisomerase III were obtained with the allelic exchange methods (Table 1). Although further study of gene functions with these mutants was initially not possible, the recent development of a general antibiotic marker for S. islandicus  allows the genetic complementation to be conducted. This selective marker has been used to complement the deficiency of an herA gene encoding a helicase implicated in HR [26,27], opening new possibilities for Sulfolobus genetic study.
This method has been employed to construct mutants for selecting sets of genes encoding DNA replication and repair proteins as well as a few genes of other categories. Table 1 summarizes all DNA repair genes for which the gene-knockout analysis has been conducted in S. islandicus REY15A in our laboratory, regardless of whether mutants have been isolated.
Genetic analysis of DNA repair genes
As a hyperthermophile, cells of S. islandicus suffer from extensive deamination of thymine at optimal growth temperature. The yielded uracil bases on DNAs should be corrected rapidly and this is to be conducted by a BER (base excision repair) mechanism initiated by an enzyme called UDG (uracil DNA glycosylase). It has been shown that Sulfolobus UDG could have developed a mechanism to facilitate the DNA repair efficiency by interacting with PCNA (proliferating-cell nuclear antigen) . S. islandicus encodes two UDG homologues and two homologues of apurinic/apyrimidinic lyase. Attempts were made to isolate AR mutants for each of the four genes and it was successful for SiRe_0084 encoding a UDG homologous with the S. solfataricus UDG interacting with PCNA, whereas other attempts failed to yield any transformants. Then, MID was employed in the analysis and transformants were obtained for SiRe_1884 encoding a putative apurinic/apyrimidinic lyase from which mutants were also isolated. The obtained mutants were then characterized. Both mutants showed an impaired growth in the absence of any DNA-damaging agent, whereas their cell cycle profiles remained the same as that of the wild-type strain, suggesting that they could play very important roles in removing uracil on DNA. Conducting double/triple gene-knockout analysis of these BER genes will reveal functional redundancy and synthetic lethality of these DNA repair enzymes.
Whereas Haloarchaea and some methanogens contain bacterial and eukaryal NER (nucleotide excision repair) pathways, Crenarchaea only encode homologues of eukaryotic NER proteins. Homologues of eukaryal NER enzymes including two nucleases [XP (xeroderma pigmentosum) F and Fen1 (flap endonuclease 1)/XPG] and two helicases (XPB and XPD) are identified, whereas the protein or protein complex that recognizes distorted DNA for repair has not been identified from bioinformatic analysis. We conducted gene-knockout analysis for all identified NER genes and mutants lacking each of XPD, XPB1, XPB2, XPF and the archaea-specific nuclease Bax1 (binds archaeal XPB 1) were obtained, whereas Fen1/XPG was found to be essential (Table 1). Our results are consistent with the results obtained for the hyperthermophilic euryarchaeon T. kodakarensis where it was found that none of eukarya-type NER genes is essential except for fen1/xpg . As Fen1 is the enzyme responsible for maturation of Okazaki fragments in DNA replication , it is expected that this gene is essential. Furthermore, mutants of these apparent eukarya-homologous NER genes were as resistant to DNA-damaging agents as the wild-type strain (, and S. Li and Q. She, unpublished work), suggesting that they do not play a role in DNA repair. Analysing these mutants by flow cytometry indicated that the helicases do not play a role in DNA replication, but depleting the cells of the XPF enzyme resulted in the accumulation of S-phase cells, indicating that it either plays a direct role in DNA replication or functions in the restart of stalled replication forks. Since biochemical characterization of the S. solfataricus XPF has revealed that the enzyme requires PCNA for the nuclease activity , it is highly likely that this enzyme plays a role in DNA replication. Furthermore, genetic analysis of the euryarchaeal counterpart Hef also suggests that it functions in DNA replication . Taken together, it is unlikely that these eukaryotic NER homologues play a role in nucleotide excision repair in Archaea.
Sequence analysis has predicted that archaeal HR machinery is of eukaryotic type and the homologous proteins include RadA, Hel308 (also called Hjm), Rad50 and Mre11. Genes encoding the last two proteins form an operon with archaea-specific genes herA and nurA, and this operon structure is conserved in many archaeal organisms . In our genetic analysis, each of the six genes was found to be essential (Table 1). In particular, gene essentiality has been assessed for herA and hjm by examining duplication of mutant cells presented in MID transformant cultures [18,34]. This approach is in strict contrast with the genetic analysis conducted for any other microbial models where gene essentiality is deduced from incapability of colonization by mutant cells. Thus our results firmly establish that HR activity is essential in S. islandicus. Previously, genetic analysis with Haloferax volcanii and Halobacterium sp. NRC-1, two mesophilic Euryarchaea, has revealed that cells lacking rad50 and mre11 are viable [35,36]. However, analysis of the six genes in T. kodakarensis revealed that five of them could be essential, but hjm is not. Thus, although these results support the theory that HR activity is essential in hyperthermophiles, they also show an interesting difference between functional roles of Hjm in the two hyperthermophiles since the gene is only essential in S. islandicus.
Furthermore, mutants have been constructed for the gene coding for a putative DNA photolyase (phrB), which mediates direct reversal of UV photoproducts on DNAs. Characterization of the mutant cells by flow cytometry indicated that the photoactivated DNA damage reversal is strongly impaired or diminished in the mutant (C. Zhang and Q. She, unpublished work). This result is consistent with that the genetic analysis of a DNA photolyase of S. acidocaldarius Saci_1227 where gene disruption leads to negligible photoreactivation .
Taken together, our current knowledge of genetic mechanisms of archaeal DNA repair is mainly derived from case studies of a few model organisms and is therefore quite fragmentary. Nevertheless, we attempt to draw some preliminary conclusions with the aim of facilitating the research in the field. First, BER must play very important roles in DNA repair in hyperthermophiles because inactivation of one of the apparently redundant genes impairs the growth of mutant cells. Secondly, the eukarya-type NER repair system does not appear to work in all studied archaeal species. Thirdly, HR activity appears to be essential in hyperthermophilic archaea, although it is dispensable from cell viability in mesophilic archaea. Fourthly, interesting differences are emerging between the two hyperthermophiles studied, i.e. S. islandicus and T. kodakarensis: the former encodes two reverse gyrases, each of which is essential and, this is in strict contrast with the non-essential reverse gyrase gene present in the latter , suggesting that reverse gyrases can play different roles in different hyperthermophiles. In conclusion, genetic analysis of DNA repair genes should be conducted in different archaeal organisms to test how widely these preliminary conclusions are applicable.
This research was supported by Danish Council for Independent Research: Technology and Production Sciences [grant numbers 09-062332 and 11-106683] and by a special grant from Huazhong Agricultural University.
Molecular Biology of Archaea 3: An Independent Meeting held at the Max Planck Institute for Terrestrial Microbiology, Marburg, Germany, 2–4 July 2012. Organized and Edited by Sonja-Verena Albers (Max Planck Institute for Terrestrial Microbiology, Germany), Bettina Siebers (University of Duisberg-Essen, Germany) and Finn Werner (University College London, U.K.).
Abbreviations: AR, allelic replacement; BER, base excision repair; Fen1, flap endonuclease 1; 5-FOA, 5-fluoro-orotic acid; HR, homologous recombination; MID, marker insertion and unmarked target gene deletion; MRL, maker replacement and looping out; NER, nucleotide excision repair; PCNA, proliferating-cell nuclear antigen; PIS, plasmid integration and segregation (also called pop in and pop out); UDG, uracil DNA glycosylase; X-gal, 5-bromo-4-chloroindol-3-yl β-D-galactopyranoside; XP, xeroderma pigmentosum
- © The Authors Journal compilation © 2013 Biochemical Society