Sulfolobus belongs to the hyperthermophilic archaea and it serves as a model organism to study archaeal molecular biology and evolution. In the last few years, we have focused on developing genetic systems for Sulfolobus islandicus using pyrEF as a selection marker and versatile genetic tools have been developed, including methods for constructing gene knockouts and for identifying essential genes. These genetic tools enable us to conduct genetic analysis on the functions of the genes involved in DNA replication and repair processes in S. islandicus and they should also facilitate in vivo analysis of functions of other genes in this model organism.
Sulfolobus is a genus of hyperthermophilic acidophiles that thrives in terrestrial acidic hot springs worldwide. These organisms are aerobic heterotrophs, growing optimally in a relatively simple medium at 75–80°C and forming colonies on Gelrite plates. Thus they are among the archaeal species that can be manipulated relatively easily in the laboratory. Consequently, Sulfolobus species serve as models to study biological principles in hyperthermophiles and archaea.
At the present time, complete genome sequences are available for three Sulfolobus species, namely, Sulfolobus acidocaldarius , Sulfolobus solfataricus  and Sulfolobus tokodaii . However, another Sulfolobus species, collectively named Sulfolobus islandicus, is more widely present in the world; strains belonging to this species have been isolated from hot springs in Russia, U.S.A. and Iceland, and they have been used as models to study mechanisms of archaeal evolution [4,5]. As a result, ten genomes of S. islandicus strains have been sequenced, eight from the U.S. DOE (Department of Energy) Joint Genome Institute (http://www.jgi.doe.gov/genomeprojects/) and two from the joint efforts of Danish, Chinese and French laboratories. Furthermore, many genetic elements including diverse viruses and plasmids have been isolated from S. islandicus and characterized [6–14], providing useful tools for studying molecular biology with this model.
We have focused on developing genetic tools for S. islandicus for the last 4 years. Our work includes the following aspects: (i) isolating and characterizing pyrEF-deletion mutants and using them for developing host–vector systems; (ii) developing different gene-knockout methodologies and using them for constructing ΔpyrEFΔlacS double-deletion mutants; and (iii) using ΔpyrEFΔlacS mutants for developing gene reporter system as well as optimizing the established genetic tools. In the present paper, we summarize the progress in genetic study on this model.
pyrEF confers efficient selection on the Sulfolobus pyrEF-deletion mutant
Selection based on the genetic complementation of a mutant phenotype has been widely used in microbial genetics, but reverse mutation in the mutant often reduces marker selection efficiency. However, when a deletion mutant is used as a host for transformation, selection efficiency based on genetic complementation will be greatly increased since deletion mutation can never be reversed. Thus we started our experiments by isolating spontaneous pyrEF-deletion mutants.
S. islandicus REY15A, a strain free from any detectable genetic elements , was used for isolating colonies resistant to 5-FOA (5-fluoro-orotic acid), from which three large pyrEF-deletion mutants have been identified by analysing the sizes of their mutant pyrEF alleles (L. Deng, H. Zhu, Z. Chen, Y.X. Liang and Q. She, unpublished work). To test whether these mutants are suitable hosts for genetic study of Sulfolobus, Sulfolobus–Escherichia coli shuttle vectors have been constructed. The Sulfolobus replicons of these shuttle vectors are derived from Sulfolobus virus SSV2 , satellite virus pSSVi  or the cryptic plasmid pRN1 or pRN2 [18,19]. The constructed integrative shuttle vectors are based on the integrase gene of SSV2, pSSVi or the conjugative plasmid pNOB8 [20,21]. Furthermore, pyrEF has been cloned from S. solfataricus and used as a marker for these shuttle vectors. When used to transform the above mutants, the marker gene expression has successfully complemented the uracil auxotroph in two of the deletion mutants and transformants grow up and form colonies on uracil-free plates. However, these shuttle vectors have failed to complement the third deletion mutant. Analysing these mutant pyrEF alleles by DNA sequencing has revealed the reason: the first two mutants have deletion mutation within pyrEF, but the deletion in the third affects three genes pyrB, pyrE and pyrF (L. Deng, H. Zhu, Z. Chen, Y.X. Liang and Q. She, unpublished work). Thus complementing pyrBEF-deletion mutation requires not only pyrEF, but also pyrB.
There have been earlier efforts to develop pyrEF as a genetic marker for S. solfataricus and S. islandicus. These works have revealed that pyrEF marker can be used to generate S. solfataricus cultures containing virus-derived vectors , but it fails to yield a proper selection of the same S. solfataricus pyrEF mutant when transformed with pRN1-based plasmid shuttle vectors, since single colonies cannot be obtained under uracil-dropout selection . However, the hosts used in these studies were the pyrEF mutants that carried either a point mutation or IS (insertion sequence)-generated mutation in which reversible mutation may frequently occur. Consequently, pyrEF revertants formed from these mutants also grew up under the selection, rendering the transformants indistinguishable from the revertants. More recently, it has been reported that pRN1-based plasmid shuttle vectors confer efficient selection on a pyrEF-deletion mutant of S. acidocaldarius , reinforcing the above argument. Possibly, the pyrEF marker confers efficient uracil dropout selection on any host that carries a pyrEF-deletion mutation.
Another useful marker in Sulfolobus genetics is lacS encoding β-glycosidase, which enables the transformed Sulfolobus cells to grow in minimal lactose medium. Since Sulfolobus cells grow only very slowly in this medium, this prevents the transformants from forming colonies on minimal plates under lactose selection. Thus, in order to obtain transformants selected from the lacS marker, transformed cells are enriched with lactose minimal medium before plating for single colonies on rich plates. Sulfolobus colonies are then stained with X-gal (5-bromo-4-chloroindol-3-yl β-D-galactopyranoside) to identify the colonies of transformants since they appear as blue colonies, whereas those formed from non-transformants remain white. Using this approach, host–vector systems based on lacS selection have been developed for S. solfataricus  and a few gene knockouts have been constructed [25–29].
Construction of S. islandicus gene-knockout mutants
Since obtaining gene-knockout mutants is a very important step in the genetic study of any model organism, different strategies have been developed for mutant construction in diverse microbial models. We have adopted some of these methodologies for constructing Sulfolobus lacS knockouts via homologous recombination (L. Deng, H. Zhu, Z. Chen, Y.X. Liang and Q. She, unpublished work). One of them is allelic replacement that is widely used in bacterial genetics. The knockout plasmid constructed for this method contains a marker gene that is flanked by two homologous sequences called L-arm (left arm) and R-arm (right arm) (Figure 1, I) and the arms are identical in sequence with the flanking sequence of the target gene. Upon transformation, two successive events of recombination occur at the two homologous sequences between the knockout plasmid and the chromosomal target gene locus (double cross-over), resulting in the replacement of the target gene with the marker in the mutants.
Schematic diagram of gene-knockout plasmids
Using pyrEF as the marker and the allelic replacement protocol, we have constructed knockouts of several S. islandicus genes, including lacS, dpo4 and udg genes that encode β-glycosidase, DNA polymerase IV and uracil DNA glycosylase respectively. The udg mutants obtained, for example, exhibit an impaired growth phenotype, whereas dpo4 mutants grow as rapidly as the wild-type (K. Dalgaard, S. Gocke and Q. She, unpublished work). These mutants are being characterized further.
However, many genes are present in multiple copies in the Sulfolobus genome. For example, there are four udg genes in the genomes of all known Sulfolobus species. Thus all copies of the same gene have to be mutated in order to study the gene essentiality, and the only way of doing that in Sulfolobus is to reuse the same marker and conduct multiple rounds of gene deletion. To reuse the pyrEF marker in S. islandicus, the pyrEF marker should be removed from the mutant chromosome again in the process of mutant construction. This protocol is also called markerless gene knockout.
Two markerless gene-knockout methods have been developed, both of which involve a two-step gene-knockout procedure. The first method comprises: (i) integration of a knockout plasmid, and (ii) segregation of the resultant merodiploid form of the mutant allele. Although this method has been proven to be very effective in bacteria and in the mesophilic archaeal species tested [30,31], there is mixed fortune for the application of this markerless gene-knockout method in hyperthermophilic archaea; this method failed to yield any expected mutants in Thermoccocus , and is unlikely to do so in S. acidocaldarius , but it does yield gene-deletion mutants in S. solfataricus . The organization of such a knockout plasmid is illustrated in Figure 1 (II).
When used to delete the lacS gene in S. islandicus, lacS-deletion mutants have been obtained from the S. islandicus pyrEF mutant by the integration and segregation markerless gene-knockout method (L. Deng, H. Zhu, Z. Chen, Y.X. Liang and Q. She, unpublished work). Thus this gene-knockout method also functions in S. islandicus, just as for S. solfataricus. However, the transformation rate with this type of knockout plasmid is lower than those obtained with other knockout plasmids. Moreover, Southern blot analyses of the transformants of the lacS-knockout plasmid have revealed several unexpected hybridization signals, which probably result from the integration of multiple copies of the same knockout plasmid at the chromosomal target (L. Deng, W. Liu and Q. She, unpublished work). Furthermore, the outcomes from merodiploid segregation cannot be selected, and this increases the difficulty of recovering any mutants that exhibit a retarded growth phenotype.
The second markerless gene-knockout method employs selection at both steps. Selection at the second step is often achieved by employing another marker. Although there is only one efficient marker, namely pyrEF, in S. islandicus, this marker can be exploited as a selection/counterselection marker: the presence of pyrEF is selectable by uracil dropout, whereas its absence is selectable with 5-FOA. In order for pyrEF to function as such, the corresponding knockout plasmid should be modified so that the marker gene is flanked by a homologous sequence. This homologous arm can either be L-arm or R-arm (Figure 1, III). A flowchart of this method is shown in Figure 2.
Two-step markerless gene deletion using pyrEF marker
Using this method, lacS-deletion mutants have been constructed from the S. islandicus pyrEF-deletion mutant (L. Deng, H. Zhu, Z. Chen, Y.X. Liang and Q. She, unpublished work). Since removing pyrEF from the mutant chromosome at the second step of recombination restored the pyrEF-deletion genotype, this yielded lacS pyrEF double-deletion mutants. To explore the double-deletion mutant in genetic study, a lacS gene was cloned onto a pRN2-based shuttle vector, and the resulting plasmid was transformed into the mutant. We found that colonies of the transformants become blue in the presence of X-gal, indicating that lacS can be used as a second marker in S. islandicus (L. Deng, H. Zhu, Z. Chen, Y.X. Liang and Q. She, unpublished work). Currently, gene knockouts of several DNA-repair genes, including radA, radB, udgs and dpo4, are under construction using this methodology. Once obtained, the mutants will be used to study genetic complementation.
However, all of the above methods are not suitable for studying gene essentiality. Moreover, if the target genes are important for growth, the desired mutants can suffer major growth retardation, causing transformants to grow so slowly that colonies of transformants are unobtainable. Thus a new method is required to study the functions of these two categories of gene. By replacing one of the repeated homologous arms with a partial or complete target gene sequence in the knockout plasmid shown in Figure 1 (III), we constructed a novel knockout plasmid (Figure 1, IV), and gene knockout based on this plasmid allows us to identify truly essential genes.
The principle behind this experimental strategy is summarized as follows. After transforming the knockout plasmid into the host cell, the plasmid DNA fragment containing R-arm–L-arm and marker is inserted between the L-arm and R-arm and the target gene during the first homologous recombination. Since this recombination yields a functional target gene, transformants are readily obtainable. Then, the marker and the target gene are removed together from the host chromosome only at the second step by looping out at R-arm or L-arm. Because the second recombination starts with culture of the transformant, which does not involve transformation, this allows the deletion mutants to be recovered at a much higher rate. Currently, we use this methodology to investigate important replication/repair genes, such as sliding-clamp PCNA (proliferating-cell nuclear antigen), replication initiator Orc1/Cdc6 and reverse gyrases, that are apparently essential, as found in our early experiments (C. Zhang, Y. Xu and Q. She, unpublished work).
Gene reporter system and promoter analysis in S. islandicus
The availability of the S. islandicus ΔpyrEFΔlacS double-deletion mutant and E. coli–Sulfolobus shuttle vectors allows us to test lacS as a reporter gene in this model organism. Different promoters have been inserted upstream of the lacS reporter gene, yielding gene reporter plasmids in which lacS expression is under the control of these promoters. Thus LacS activity produced by the transformants of these reporter plasmids reflects the activity of the respective promoters.
Furthermore, this reporter system has also been used to identify the cis-acting element of the araS promoter. Reporter plasmids containing a series of promoter deletion mutants as well as the substitution mutants generated by PCR have been constructed. A functional full promoter for araS has been determined as a 55 bp sequence upstream of the transcription start site of the araS gene, and the main controlling elements include the ara box that has been identified previously using bioinformatic analysis , the putative BRE (transcription factor IIB-recognition element) and the TATA box (N. Peng, Q. Xia, S. Li and Q. She, unpublished work).
Versatile genetic systems have been established for S. islandicus, which provide a platform for developing further methods for generating gene-deletion mutants for both essential and non-essential genes. Studying gene function can then be conducted by complementing the mutant via regulated expression of the target gene either from a plasmid shuttle vector or an integrated vector. Furthermore, tagged protein expression in Sulfolobus cells will greatly facilitate biochemical investigation of protein–protein interactions in this model organism. Once powered with proteome and transcriptome analyses, the developed genetic methodologies will greatly facilitate the studies on this crenarchaeal model to investigate the molecular mechanisms of cellular processes in archaea.
Our work is supported by the Danish Forskningsrådet for Teknik og Produktion [grant number 274-07-0116] and Forskningsrådet for Natur og Univers [grant number 272-05-400] and by Huazhong Agricultural University Funds.
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.).