SEARCH

SEARCH BY CITATION

Keywords:

  • archaea;
  • genetics;
  • methanogens;
  • Sulfolobales;
  • Thermococcales;
  • halophiles

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References

The tree of life is split into three main branches: eukaryotes, bacteria, and archaea. Our knowledge of eukaryotic and bacteria cell biology has been built on a foundation of studies in model organisms, using the complementary approaches of genetics and biochemistry. Archaea have led to some exciting discoveries in the field of biochemistry, but archaeal genetics has been slow to get off the ground, not least because these organisms inhabit some of the more inhospitable places on earth and are therefore believed to be difficult to culture. In fact, many species can be cultivated with relative ease and there has been tremendous progress in the development of genetic tools for both major archaeal phyla, the Euryarchaeota and the Crenarchaeota. There are several model organisms available for methanogens, halophiles, and thermophiles; in the latter group, there are genetic systems for Sulfolobales and Thermococcales. In this review, we present the advantages and disadvantages of working with each archaeal group, give an overview of their different genetic systems, and direct the neophyte archaeologist to the most appropriate model organism.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References

In his ‘An Essay on Man’, the English poet Alexander Pope exhorts us to ‘Know then thyself, presume not God to scan; The proper study of Mankind is Man’. The resounding success of biomedical research using model organisms gives us reason to doubt the wisdom of Pope's words. Most of our knowledge of fundamental biological processes has come from work on simple and experimentally tractable species such as Escherichia coli, Saccharomyces cerevisiae, and Caenorhabditis elegans. Over time, these basic principles have been verified in complex species such as man, but simple model organisms remain vital to further medical discoveries (Fields & Johnston, 2005). Nevertheless, there is danger in such a reductionist approach – not everything in biology can be learned from E. coli. In order to witness the ‘grandeur in this view of life’ (apologies to Charles Darwin), we must expand our repertoire of model organisms to include representatives of the domain Archaea.

The choice of model organism is critically important. It should be easy to grow, have a short generation time, and be amenable to experimental manipulation. For microbial geneticists, the minimal specification is the ability to grow in isolation on solid media. Robert Koch first recognized that a colony formed on an agar plate represents the clonal expansion of a single cell, and this unassuming mound of cells has always been the cornerstone of microbiology. The ability to generate mutants is another part of the foundation of microbial genetics. Traditionally this was carried out by random mutagenesis (forward genetics), but since the molecular biology revolution the preferred method has been targeted mutation of a specific gene (reverse genetics). The latter would not be possible without methods for transformation, selectable markers, plasmid vectors, and systems for gene knockout by homologous or site-specific recombination. Of late, reverse genetics has been made significantly easier by whole genome sequencing, and is now taken for granted in the genetic toolbox.

Archaea make up one of the three main branches of the evolutionary tree; they are as different from eukaryotes as they are from bacteria (Garrett & Klenk, 2007). The distinct status of archaea was revealed in the late 1970s, when Carl Woese and colleagues seized upon the emerging technology of nucleic acid sequencing to tackle the problem of prokaryotic phylogeny. Woese chose small-subunit rRNA as a molecular chronometer; rRNA is an essential component of all self-replicating organisms and shows remarkable sequence conservation. The tree he constructed showed that an unusual group of methane-producing microorganisms were not bacteria, but formed a separate domain. Woese termed these organisms ‘archaebacteria’, but later changed this name to archaea (Woese & Fox, 1977; Woese et al., 1990). The tree suggested a closer relationship between archaea and eukaryotes, compared with bacteria, and reflected the common heritage of information-processing systems found in the archaeo-eukaryal lineage. This had been noted by Wolfram Zillig and colleagues in the 1980s, when they found that archaeal DNA-dependent RNA polymerase is strikingly similar to its eukaryotic counterpart, in terms of both complexity and subunit composition (Huet et al., 1983). Genome sequencing in the 1990s confirmed that archaea are a genetic mosaic – their information processing systems show significant homology to eukaryotic counterparts, while most operational (housekeeping) functions have a bacterial aspect (Olsen & Woese, 1996; Rivera et al., 1998; Yutin et al., 2008).

Over time, many halophilic and thermophilic microorganisms have found their home in the archaeal domain. Several of these species had been studied long before the third domain of life was proposed. For example, bacteriorhodopsin had been discovered in Halobacterium salinarum in 1971 (Oesterhelt & Stoeckenius, 1971), Thomas Brock had isolated Sulfolobus acidocaldarius from acidic mud ponds in Yellowstone National Park in 1972 (Brock et al., 1972), and in the 18th century Alessandro Volta had unwittingly unearthed methanogenic archaea in the swamps of northern Italy. What appeared to bind together all these exotic microorganisms was their love of habitats that had previously been considered uninhabitable. However, cultivation-independent analyses of microbial biodiversity have since revealed that archaea are surprisingly abundant in ‘normal’ environments (DeLong, 1998; Karner et al., 2001; Robertson et al., 2005). Unfortunately these archaea are not fit for genetics, because with the exception of the recently isolated Nitrosopumilus maritimus (Konneke et al., 2005), they cannot yet be cultured in isolation (Hugenholtz et al., 1998; Schleper et al., 2005). We are therefore left with four groups of archaea for which genetic systems have been developed: methanogens and halophiles (both euryarchaea), as well as thermophilic euryarchaea (Thermococcales) and crenarchaea (Sulfolobales) (Fig. 1).

image

Figure 1.  Phylogenetic tree showing key archaeal species with genetic systems. Phylogenetic tree based on 16S rRNA gene sequences of selected archaeal species whose genome sequences are available. Organisms indicated with solid red stars are key species that have been the focus of genetic development; open stars indicate species where genetics has been applied or where there is potential for genetics. Sequence alignments were performed using Clustal W and the tree was constructed by neighbor joining; branches with bootstrap values of <50% are not shown.

Download figure to PowerPoint

Each of these archaeal groups has its own unique selling point. Haloarchaea are renowned for the comparative sophistication of their genetic systems, the development of which was made possible by early work on transformation protocols. In addition, haloarchaea are easy to cultivate because they grow at moderate temperatures. Methanogenic archaea are also mesophilic, but unlike haloarchaea, their cytoplasm is not hypersaline. This has permitted the direct adaptation of many tools from bacterial genetics to methanogens; bacterial antibiotics remain the exception, their targets are generally not found in archaea. Thermophilic archaea of kingdoms Euryarchaeota and Crenarchaeota have long been of interest to biochemists and structural biologists, owing to their thermostable enzymes. They offer significant potential for biotechnology, and for researchers wishing to use a multidisciplinary approach that combines genetics with biochemistry.

In this review, we offer guidance to microbiologists who wish to convert to the third domain, and reassure them that archaeal genetics is not difficult or unusual. The first step on this road is to choose the most appropriate model organisms. We deal in turn with each archaeal group, highlighting their advantages and disadvantages in terms of the scientific questions that can be addressed, and the tools available to answer these questions. Our hope is that more microbiologists will work on archaea, and those who already do so will venture beyond the safe environs of biochemistry and structural biology. Only when genetics has found a place in every archaeal laboratory will the third domain of life rank alongside its eukaryotic and bacterial cousins.

Methanogens

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References

Introduction to methanogens, an ecologically and biochemically distinctive group

In 1977, a collaboration between the laboratories of Carl Woese and Ralph Wolfe resulted in the finding that the methanogens were ‘only distantly related to typical bacteria’ (Fox et al., 1977). Thus, the methanogens became the first known Archaea. They are now known to comprise five orders of the Euryarchaeota: Methanococcales, Methanosarcinales, Methanobacteriales, Methanomicrobiales, and Methanopyrales (Liu & Whitman, 2008). Genetic tools are available for certain species of the first two orders.

The methanogens are those organisms that generate methane as a catabolic end-product (Wolfe, 1996). Biological methanogenesis occurs in a variety of anaerobic habitats, including marine and freshwater sediments, rice paddies, bioreactors and sewage sludge digesters, landfills, animal digestive tracts, and hydrothermal vents (Wolfe, 1996). Most of these habitats contain an anaerobic ecosystem in which methanogenesis is the final step in the decomposition of organic matter. However, in habitats such as hydrothermal vents the substrates for methanogenesis, H2 and CO2, are presumably of geochemical origin. Much of the methane that is generated is reoxidized to CO2 or becomes sequestered in methane hydrates. However, a significant amount of methane is emitted into the atmosphere where it becomes a major greenhouse gas (Liu & Whitman, 2008). The product of methanogenesis is also of obvious importance as a fuel.

Methanogenesis is a kind of anaerobic respiration where single-carbon (C-1) units, most notably CO2, serve as the electron acceptor. Thermodynamics dictates that methanogenesis will occur only when more favorable electron acceptors are absent. Hence, methanogenesis is most prevalent when sulfate, nitrate, oxidized metals, and, especially, oxygen are absent. Because CO2 is the only electron acceptor that does not owe its abundance to photosynthesis, methanogenesis is favored as an early metabolism on earth, predating photosynthesis and other forms of respiration (Kasting & Siefert, 2002).

Substrates for methanogenesis are relatively restricted (Whitman et al., 2001). Nearly all species in the orders Methanococcales, Methanobacteriales, Methanomicrobiales, and Methanopyrales are hydrogenotrophic, using H2 and CO2. Many of these species can also use formate. In contrast, the Methanosarcinales is comprised of methylotrophic species, which use methyl compounds such as methanol and methylamines; some can use H2 and CO2 as well. In addition, Methanosarcina and Methanosaeta, members of the Methanosarcinales, use acetate. Most recognized species of methanogens are mesophilic, but hyperthermophilic and psychrotolerant species are also well known. To date, genetic tools have been developed only for certain mesophilic species.

The methanogens are biochemically distinctive. The enzymes and unique coenzymes of methanogenesis are known thanks largely to the work of Ralph Wolfe, Rolf Thauer, Godfried Vogels, and Gerhard Gottschalk. Our understanding of how methanogenesis is coupled to energy conservation has been slower to develop. As for all respirers, energy conservation is fundamentally chemiosmotic. A methyl transfer step plays a central role in most methanogenic pathways and directly drives the export of sodium ions. Other components of the energy conservation apparatus appear to differ in the methylotrophic and hydrogenotrophic methanogens. Methylotrophic methanogens have cytochromes and a proton-translocating electron transport chain, which they use to conserve energy in the last, exergonic step in methanogenesis. These components are lacking in hydrogenotrophic methanogens, making it unclear how these organisms achieve a net positive gain in energy conservation, because the first step in methanogenesis from CO2 is endergonic. A recently proposed mechanism involving electron bifurcation, where exergonic electron flow directly drives endergonic electron flow, could explain this conundrum (Thauer et al., 2008).

Why study methanogens?

The methanogenic pathway itself has captured the curiosity of many for decades. Recently genetic approaches have begun to fill some gaps in our knowledge left from biochemical approaches, as mentioned below. In addition, methanogens have been chosen for many studies of the molecular biology and physiology of Archaea. Methanocaldococcus jannaschii became the first species of Archaea to be subjected to genome sequencing in 1996 (Bult et al., 1996), and many studies followed. Genetic tools have not been developed for M. jannaschii, but many questions can be addressed using the genetic tools for its relatives in the genus Methanococcus (Tumbula & Whitman, 1999). Methanogens are models for archaeal replication (Walters & Chong, 2009), transcription, regulation (Geiduschek & Ouhammouch, 2005), osmoregulation (Spanheimer & Muller, 2008), and protein structure. The role of methanogens in nature leads directly to questions of syntrophy, the associations between organisms that facilitate the transfer of nutrients (Shimoyama et al., 2009). The discovery that close relatives of the Methanosarcinales as well as sulfate reducers are involved in anaerobic methane oxidation has broadened the importance of these organisms in the global carbon and sulfur cycles (Knittel & Boetius, 2009).

Methanogens are strict anaerobes that require special measures for their growth in the lab. However, in the late 1970s the relatively tricky Hungate technique was replaced with the technique of Balch and Wolfe (Balch et al., 1979), and the requirements for anaerobiosis are easily achieved with a modest expenditure on equipment and minimal training.

Key species of methanogens that have genetic systems

Species for which genetic tools have been developed belong to the genera Methanococcus and Methanosarcina. Thus, there are genetic systems for representatives of the two metabolic types, hydrogenotrophic and methylotrophic methanogens. Both genera have well-developed tools, but each has its intrinsic advantages.

Methanococcus species grow relatively fast (doubling times around 2 h) and liquid cultures grow to high densities overnight. For Methanococcus maripaludis, colonies of useful size often form in 2 days after inoculation on agar plates. In addition, for M. maripaludis a robust system for continuous culture in chemostats has been established and used effectively in studies of global regulation (Haydock et al., 2004; Hendrickson et al., 2008). The genomes of Methanococcus species are small (1.6–1.8 Mbp), streamlining annotation and transcriptomic and proteomic analyses. Genome sequences for four species [M. maripaludis (four strains), Methanococcus voltae, Methanococcus vannielii, and Methanococcus aeolicus] are available currently. The relatively restricted substrate range for methanogenesis in hydrogenotrophic species limits the utility of genetics in Methanococcus to study methanogenesis itself. Nevertheless, genetics demonstrated the role in vivo of an alternative pathway for the reduction of the electron carrier coenzyme F420 (Hendrickson & Leigh, 2008), and the role of the energy-conserving hydrogenase Ehb in carbon fixation (Porat et al., 2006).

Methanosarcina species grow more slowly (doubling times around 8 h), and the formation of colonies requires about 14 days of incubation. The genomes of Methanosarcina species are relatively large, ranging from 4.1 to 5.8 Mbp. Genome sequences for three species (Methanosarcina acetivorans, Methanosarcina barkeri, and Methanosarcina mazei) are available currently. Despite their slower growth, the metabolic versatility of Methanosarcina allows more possibilities for the study of the methanogenic pathway. For example, mutants in oxidation/reduction steps between the formyl and methyl levels lost the ability to grow on H2 and CO2 or methanol alone, but grew well on H2 and methanol (Welander & Metcalf, 2008).

Genetic tools for methanogens

The basic elements of the genetic toolbox consist of a means of DNA delivery, selection for that DNA, and a way for the DNA to replicate. Reliable plating of single cells, which grow into clonal colonies, is also needed for mutant screening. In methanogens a genetic manipulation of this kind was first achieved in 1987 when Bertani (of Luria–Bertani medium fame) and Baresi transformed auxotrophs of M. voltae to prototrophy (Bertani & Baresi, 1987). Selection by antibiotic resistance was initiated when A. Klein transformed M. voltae by expressing a puromycin resistance marker from Streptomyces (Gernhardt et al., 1990). In addition, W. Whitman devised a strategy for the enrichment of auxotrophic mutants in M. maripaludis (Ladapo & Whitman, 1990). Since then, the genetic tools for methanogens have been expanded and improved. Genetics became feasible in Methanosarcina, which normally grows in multicellular packets, when conditions for growth as single cells were found (Sowers et al., 1993), and W. Metcalf documented the high-efficiency transformation of Methanosarcina using liposomes (Metcalf et al., 1997). Table 1 outlines the genetic tools available for methanogens. Most of these techniques were worked out for M. maripaludis and M. acetivorans, but many of them have also been applied successfully in M. voltae, M. barkeri, and M. mazei. It has been possible to adapt many tools from standard bacterial genetics to the methanogens because they grow at moderate temperatures and salt concentrations.

Table 1.   Genetic tools for methanogens
 MethanococcusMethanosarcina
Negative enrichmentEnrichment of auxotrophic mutants (Ladapo & Whitman, 1990) 
DNA deliveryPEG-mediated transformation (Tumbula et al., 1994), conjugation (Dodsworth et al., 2010)Liposome-mediated transformation (Metcalf et al., 1997)
Replicative shuttle vectorsGardner & Whitman (1999)Metcalf et al. (1997)
Positive selectionPuromycin (Gernhardt et al., 1990), neomycin (Argyle et al., 1996)Puromycin (Gernhardt et al., 1990), pseudomonic acid (Boccazzi et al., 2000)
Counterselectionhpt (8-azahypoxanthine), upt (6-azauracil) (Moore & Leigh, 2005)hpt (8-aza-2,6-diaminopurine) (Pritchett et al., 2004)
Markerless genetic exchange (pop-in/pop-out gene replacement)Moore & Leigh (2005)(Pritchett et al., 2004)
Ectopic integrationInto hpt or upt (Moore & Leigh, 2005)Enhanced with φC31 site-specific recombination system (Guss et al., 2008)
Transposon insertionIn vitro (Porat & Whitman, 2009)In vivo (Zhang et al., 2000)
Reporter geneslacZ (β-galactosidase) (Lie & Leigh, 2002), uidA (β-glucuronidase) (Beneke et al., 1995)uidA (β-glucuronidase) (Pritchett et al., 2004)
Regulated gene expressionnif promoter (Lie & Leigh, 2002; Chaban et al., 2007)Tetracycline-responsive promoters (Guss et al., 2008)
DNA delivery, positive selection, shuttle vectors, and insertional gene disruption

DNA is introduced into Methanococcus and Methanosarcina species by transformation and plating under anaerobic conditions. In M. maripaludis the polyethylene glycol (PEG)-mediated transformation of spheroplasts results in frequencies near 105 transformants μg−1 DNA and 10−5 transformants CFU−1. In M. acetivorans a liposome-mediated method achieves higher frequencies, as high as 108 transformants μg−1 DNA and 20% of the CFU. In both cases, one easily obtains thousands of colonies in a single experiment. Puromycin transacetylase from Streptomyces works well (Gernhardt et al., 1990), and selection in both genera is most commonly achieved using puromycin. In M. maripaludis, neomycin resistance is also achieved using aminoglycoside phosphotransferase genes (Argyle et al., 1996). Replicative shuttle vectors have been devised for both genera, using replicative elements from naturally occurring plasmids of strains of each genus. DNA lacking autonomous replication can integrate into the chromosome via homologous recombination, which appears to occur at a particularly high frequency in M. maripaludis (Tumbula et al., 1997). This allows for gene replacement and disruption by insertion of the selectable marker (Fig. 2a), and other genetic manipulations described below. In addition to transformation, a conjugation system from E. coli to M. maripaludis has been described recently (Dodsworth et al., 2010). It is less laborious than transformation and may be useful for routine genetic manipulation of this methanogen.

image

Figure 2.  Gene knockout methods used in archaeal genetics. Specific details (such as selectable markers) are illustrative and do not necessarily apply for all archaeal groups. See text for further details. (a) Gene replacement with selectable marker (in this case trp), by recombination between flanking regions of the gene and a chromosomal target, uses linear DNA. (b) Pop-in/pop-out deletion method uses circular DNA. Integration of the deletion construct (pop-in) is selected by transformation to uracil prototrophy. Intramolecular recombinants that have lost the plasmid (pop-out) are counterselected using 5-FOA. In methanogens, two different markers are used for selection and counterselection, respectively. (c) Variant of the pop-in/pop-out method for gene deletion, where the gene is replaced with a marker allowing direct selection (in this case trp). Used for deletion of genes that are important for cell viability. (d) Refinement of the pop-in/pop-out gene replacement method, where the gene function is complemented in trans from a shuttle vector. Loss of the shuttle vector (plasmid shuffling) and gene deletion is ensured by counterselection with 5-FOA. (e) Ectopic integration at the ura locus. Pop-in of a construct bearing the point mutation (of experimental gene, not ura) is selected by transformation to thymidine prototrophy. Counterselection with 5-FOA ensures that the ura gene is replaced with the point mutation.

Download figure to PowerPoint

Markerless genetic exchange

In both Methanococcus and Methanosarcina, systems have also been devised for markerless genetic exchange. Using these systems, mutations, including in-frame deletions, can be made in which the selectable marker is removed, allowing its re-use for subsequent manipulations. Vectors for this purpose contain both selectable and counterselectable markers. hpt and upt, encoding hypoxanthine and uracil phosphoribosyltransferase, respectively, allow for counterselection in the presence of nucleobase analogs in genetic backgrounds from which these genes have been deleted. In this approach, often termed pop-in/pop-out gene replacement, the construct to be exchanged into the genome is cloned with homologous flanking DNA on both sides (Fig. 2b). After transformation, positive selection results in a merodiploid in which integration of the entire plasmid has occurred in a single recombination event. Counterselection results in a second recombination event, removing the vector. Because the second recombination event can occur on the same or the opposite side from the first, the result is either a wild type or a mutant locus, which must be distinguished by screening. For Methanosarcina, vectors have been equipped with recognition sites for the Flp site-specific recombination system, allowing more expedient removal of the marker (Rother & Metcalf, 2005).

Ectopic integration

Integration of constructs into the genome is desirable not only for gene disruption or for modification but also in cases where artifacts due to multiple copies on a plasmid are to be avoided. In M. maripaludis this has been achieved by ectopic incorporation of constructs into the sites of the counterselectable genes hpt or upt (Fig. 2e) (Moore & Leigh, 2005). For Methanosarcina a system has been devised that uses the φC31 site-specific recombination system to considerably increase the efficiency of integration (Guss et al., 2008).

Overexpression and controlled expression

Replicative vectors for Methanococcus and Methanosarcina are equipped with strong promoters that allow overexpression of genes. These vectors have been used to overexpress His-tagged proteins in their native species for subsequent purification (Dodsworth & Leigh, 2006). A tetracycline-inducible promoter has been constructed for Methanosarcina by combining a strong promoter from M. barkeri with binding sites for the bacterial TetR protein, and used in a test for gene essentiality (Guss et al., 2008). This system also has promise for the induction of gene expression for the purpose of protein production and purification. Attempts to adapt the tetracycline induction system for Methanococcus have not been successful. However, the nif (nitrogen fixation) promoter has been used in M. maripaludis for differential controlled expression (Lie & Leigh, 2002; Chaban et al., 2007) and has potential application in tests for gene essentiality.

Transposon insertion

In vivo transposon insertion has been devised for M. acetivorans. The transposon system is derived from a mini-mariner element and inserts randomly into the genome at high frequency. The transposon contains selectable markers for E. coli as well as for Methanosarcina, and contains an E. coli origin of replication, facilitating cloning of the transposon insertion sites (Zhang et al., 2000). This system works in M. maripaludis only at low frequency. However, transposition in vitro (Porat & Whitman, 2009) and into an E. coliλ lysate (Blank et al., 1995) have been used successfully to generate insertions in a cloned M. maripaludis gene cluster, followed by transformation into M. maripaludis. These approaches have potential for development into an efficient transposon insertion system for Methanococcus.

Reporter genes

Genes used in reporter gene fusions are uidA (encoding β-glucuronidase) for Methanosarcina (Pritchett et al., 2004) and uidA (Beneke et al., 1995) and lacZ (β-galactosidase) (Lie & Leigh, 2002) for Methanococcus. Most applications measure reporter enzyme activity in cell extracts. The use of reporter genes for in vivo screening is more limited, because color development requires oxygen. However, M. maripaludis colonies have been exposed to air and sprayed with X-gal. Color development occurs before loss of viability, and colonies can be returned to anaerobic conditions and picked. This approach was used to identify super-repressor variants of the transcriptional repressor NrpR (Lie & Leigh, 2007).

Discoveries and recent progress

Genetic approaches have been particularly useful in Methanosarcina for filling gaps in our knowledge of the methanogenic pathway. For example, the energy-conserving hydrogenase Ech was shown to be required for methanogenesis and carbon fixation (Meuer et al., 2002) in Methanosarcina. In another study, hydrogen cycling, a fundamental strategy in chemiosmotic energy conservation, was shown to occur (Kulkarni et al., 2009). A striking feature in Methanosarcina is that the methylamine methyltransferases contain pyrrolysine, the 22nd genetically encoded amino acid. Genetic studies have helped determine the function of a dedicated tRNA and aminoacyl-tRNA synthetase in pyrrolysyl-tRNA synthesis (Mahapatra et al., 2006, 2007).

Genetic studies in Methanococcus have addressed a wide range of questions. A number of regulatory mechanisms have been studied, and regulatory factors have been identified that govern the expression of genes for hydrogen metabolism (Sun & Klein, 2004) and nitrogen assimilation (Lie & Leigh, 2003; Lie et al., 2005). A novel mechanism for the regulation of nitrogenase activity was discovered in M. maripaludis, and evidently exists in a variety of diazotrophic Archaea and Bacteria (Dodsworth et al., 2005; Dodsworth & Leigh, 2006). Genetic studies in M. maripaludis have led to the identification of components of the archaeal flagellum system (Chaban et al., 2007), tRNA-dependent cysteine biosynthesis (Stathopoulos et al., 2001; Sauerwald et al., 2005), and requirements for selenocysteine biosynthesis (Rother et al., 2001; Yuan et al., 2006; Stock et al., 2010).

Numerous studies of global regulation at the transcriptomic and proteomic levels have been carried out with Methanococcus and Methanosarcina species, showing the global responses to alternative substrates, salt stress, and availabilities of nutrients including hydrogen and nitrogen (Hovey et al., 2005; Li et al., 2005a, b; Lessner et al., 2006; Veit et al., 2006; Xia et al., 2006, 2009; Hendrickson et al., 2007, 2008; Pfluger et al., 2007; Jager et al., 2009).

Halophiles

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References

Introduction to haloarchaea, the heterotrophic, aerobic halophiles of the Euryarchaeota

Halophilic archaea inhabit the most saline environments on earth, including solar salterns and natural salt lakes. Like many other habitats where archaea are found, salt lakes were once thought devoid of life. In 1936, Benjamin Elazari-Volcani published the first report of microbial life in the Dead Sea (Elazari-Volcani, 1936). His work was commemorated by the naming of Haloferax volcanii, which was isolated from Dead Sea mud in 1977 (Mullakhanbhai & Larsen, 1975). In fact, the discovery of halophilic archaea predates the proposal of the domain Archaea by Carl Woese in the late 1970s (Woese & Fox, 1977; Woese et al., 1990). For instance, H. salinarum was unwittingly discovered in 1922 as a red discoloration of salted fish (Harrison & Kennedy, 1922). All halophilic archaea are members of the Euryarchaeota and somewhat confusingly, the family Halobacteriaceae (Oren et al., 2009). In this review, we will use instead the term haloarchaea unless when referring to the genus Halobacterium.

Archaea are not alone in hypersaline environments; they share this habitat with bacteria, fungi and algae. In contrast to most halophilic bacteria and eukaryotes, haloarchaea maintain an osmotic balance with their medium by accumulating equimolar salt concentrations in the cytoplasm (Christian & Waltho, 1962; Oren, 2008). This ‘salt-in’ strategy predominantly uses potassium because it attracts less water than sodium. The converse ‘salt-out’ strategy favored by halotolerant bacteria excludes salt from the cytoplasm and uses organic solutes such as glycerol or glycine betaine to maintain an osmotic balance. The salt-out strategy is energetically costly and less suitable at saturating salt concentrations, which is probably why haloarchaea predominate under hypersaline conditions (Oren, 1999). There are notable exceptions: the bacterium Salinibacter ruber uses the archaeal salt-in strategy and coexists with haloarchaea at near-saturating salt concentrations (Oren et al., 2002).

Because of the salt-in strategy, haloarchaeal proteins are adapted to function in molar salt concentrations and commonly denature in low-salt solutions. The adaptation to salt relies on several different strategies (Lanyi, 1974; Mevarech et al., 2000).

  • A reduction in overall hydrophobicity, by replacing large hydrophobic residues on the protein surface with small hydrophilic residues. This strategy is used by the dihydrofolate reductase of Hfx. volcanii, which requires higher salt concentrations for correct folding than the E. coli enzyme (Wright et al., 2002).
  • An increase in acidic residues. A high density of negative charges coordinates a network of hydrated cations, which maintain the protein in solution (Lanyi, 1974). For example, glucose dehydrogenase of Haloferax mediterranei is very acidic (Britton et al., 2006), and the median pI of the Hbt. salinarum proteome is predicted to be 4.9 (Kennedy et al., 2001). By altering the total charge, it is possible to interconvert halophilic and mesophilic forms of a protein (Tadeo et al., 2009).
  • An additional domain that is not found in mesohalic proteins, as seen in the ferredoxin of Hbt. salinarum (Marg et al., 2005). The latter features a 30-amino acid insertion near its N-terminus, which is extremely rich in acidic amino acids and is essential for correct protein folding at high salt concentrations.

Besides their adaptation to salt, haloarchaea have other characteristic features. They are aerobic heterotrophs, some of which have the potential for anaerobic growth (Falb et al., 2008). They are slightly thermophilic with an optimum temperature of 40–50 °C, can withstand up to 60 °C, and grow reasonably well at 37 °C (Robinson et al., 2005). Even Halorubrum lacusprofundi, which was isolated in Antarctica, grows best at 36 °C. Haloarchaea are generally pigmented with C-50 bacterioruberins and some species contain retinal proteins such as bacteriorhodopsin. It is unlikely that these pigments play a significant role in protection against UV. While Hbt. salinarum can withstand very high doses of UV (McCready & Marcello, 2003), other pigmented species such as Hfx. volcanii are no more resistant than the model bacterium E. coli (Delmas et al., 2009).

To date, 12 haloarchaeal genomes have been sequenced [Halalkalicoccus jeotgali B3(T), Haloarcula marismortui ATCC 43049, Halobacterium sp. NRC-1, Hfx. volcanii DS2, Halogeometricum borinquense DSM 11551, Halomicrobium mukohataei DSM 12286, Haloquadratum walsbyi DSM 16790, Halorhabdus utahensis DSM 12940, Halorubrum lacusprofundi ATCC 49239, Haloterrigena turkmenica DSM 5511, Natronomonas pharaonis DSM 2160, Natrialba magadii ATCC 43099] and many more are underway. They usually consist of one main chromosome and a number of megaplasmids (Pfeiffer et al., 2008a; Soppa et al., 2008). Structural differences in their respective megaplasmids underlie the distinction between Hbt. salinarum and Halobacterium sp. NRC-1, but they are essentially the same species (we refer to both as Hbt. salinarum) (Ng et al., 2000; Pfeiffer et al., 2008b). Polyploidy is a signature of haloarchaea; there are 15–30 genome copies in Hfx. volcanii and Hbt. salinarum (Breuert et al., 2006). Haloarchaeal genomes are characterized by a high G+C content (∼65%) (Soppa et al., 2008); the one known exception is Haloquadratum walsbyi (45% G+C) (Bolhuis et al., 2006). It is often stated that the high G+C content is linked to the acidic proteome of haloarchaea (average pI of ∼4.4), but it is more probably due to evasion of insertion sequence (IS) elements that target A+T-rich sequences (Pfeifer & Betlach, 1985; Cohen et al., 1992; Hartman et al., 2010). Interestingly, IS elements in H. walsbyi have a higher GC content than the rest of the genome (Bolhuis et al., 2006). While all halophiles are infested with IS elements (Brugger et al., 2002), their activity varies between species. Halobacterium salinarum exhibits genome instability due to frequent IS-mediated rearrangements (Sapienza et al., 1982; Simsek et al., 1982), but this is much less of a problem in Hfx. volcanii, where IS elements are confined to nonessential regions on the megaplasmids (Cohen et al., 1992; Lopez-Garcia et al., 1995).

In common with methanogens, the genomes of haloarchaea encode multiple isoforms of genes that are present as a single copy in other organisms. For instance, there are 16 orc1/cdc6 genes for the DNA replication initiator in Hfx. volcanii and 10 in Hbt. salinarum (Berquist et al., 2007; Norais et al., 2007; Hartman et al., 2010). This might be due to a requirement for regulatory and metabolic flexibility in haloarchaea (Facciotti et al., 2007), but it is also possible that these redundant homologs have accumulated as a result of lateral gene transfer (LGT). There is ample evidence for large-scale LGT, often with bacteria, and it has been proposed that haloarchaea originally descended from methanogens that had acquired the genes for aerobic respiration from bacteria (Boucher et al., 2003). Underpinning this LGT is a system for mating and gene transfer in Hfx. volcanii (Rosenshine et al., 1989) and a wide variety of haloarchaeal viruses (Dyall-Smith et al., 2003).

Why study halophiles?

A significant motivation for working on haloarchaea is the sophistication of their genetic systems. They are easy to cultivate in the laboratory, have fast growth kinetics (Robinson et al., 2005), and are resistant to contamination by nonhalophilic microorganisms (of course, cross-contamination of haloarchaeal strains within the laboratory remains a constant threat). Halophiles were the first archaeal group in which routine transformation with foreign DNA was possible (Charlebois et al., 1987; Cline et al., 1989). The ease, efficiency, and broad applicability of PEG-mediated transformation has ensured that haloarchaea have remained at the forefront of genetic tool development (Soppa, 2006). Perhaps the greatest testament to the popularity of haloarchaea is the Halohandbook, an invaluable compendium of methods for working with halophiles, which is diligently curated by Mike Dyall-Smith (Dyall-Smith, 2009).

Besides sophisticated genetics, there are many other reasons for working on halophiles. Haloferax volcanii and Hbt. salinarum can grow over a range of salinities and have been exploited to uncover genes involved in osmotic stress (Bidle et al., 2007, 2008; Coker et al., 2007). Because halophilic proteins function under conditions of low water availability, they offer distinct advantages for structural biology and biotechnology. For example, the structure of the ribosome was solved using the complex from Har. marismortui, leading to the Nobel Prize for Chemistry in 2009 (Ban et al., 2000). In biotechnology, there are few success stories that can match bacteriorhodopsin. This purple membrane protein from Hbt. salinarum was identified by Walther Stoeckenius in 1971 (Oesterhelt & Stoeckenius, 1971) and has been used in countless photochemical applications (Margesin & Schinner, 2001; Oren, 2010).

Haloarchaea are an excellent choice to address the ‘prokaryotic species question’: do prokaryotic organisms form genomic and phenomic clusters that are sufficiently cohesive that we might legitimately call them species (Doolittle & Zhaxybayeva, 2009)? Haloarchaea are physiologically diverse and inhabit distinct ecological niches (Oren, 2008), they have dynamic genomes with systems for gene exchange and show evidence for LGT. Work on isolates of the genus Halorubrum from solar salterns and natural salt lakes has shown that haloarchaea exchange genetic information promiscuously, leading to the suggestion that there is no nonarbitrary way to define a prokaryotic species (Papke et al., 2004, 2007). Genetics has been used to answer the question of whether genes acquired by LGT can supplant an endogenous function. The UvrABC complex, which is of bacterial origin, is functional in the repair of UV-induced DNA damage in both Hbt. salinarum and Hfx. volcanii (Crowley et al., 2006; Lestini et al., 2010).

Key species of halophiles that have genetic systems

There are two haloarchaeal model organisms: Hbt. salinarum and Hfx. volcanii. Both species have mature genetic systems, but each has intrinsic advantages. Halobacterium salinarum is the traditional choice for haloarchaeal cell biology; it was isolated many years ago and quickly became a popular choice owing to its purple membrane protein, bacteriorhodopsin (Oesterhelt & Stoeckenius, 1971; Margesin & Schinner, 2001). The genome of Hbt. salinarum NRC-1 was published in 2000 (Ng et al., 2000), a second annotation with a different assembly of the megaplasmids was published for Hbt. salinarum R1 in 2008 (Pfeiffer et al., 2008b). The first archaeal method for gene knockout using a counterselectable marker was also published in 2000, using the ura3 gene of Hbt. salinarum (Peck et al., 2000) (Fig. 2). Availability of the genome sequence coincided with the growth of transcriptomics, and consequently Hbt. salinarum has been a favorite model for systems biology (DasSarma et al., 2006). DNA repair has been a fruitful topic, because Hbt. salinarum is extremely resistant to UV and ionizing radiation (Baliga et al., 2004; Whitehead et al., 2006). However, for those wishing to carry out traditional genetics, Hbt. salinarum is perhaps not ideal. It grows slowly, its genome is unstable due to frequent IS-mediated rearrangements (Sapienza et al., 1982; Simsek et al., 1982), and the range of selectable markers is somewhat limited.

Haloferax volcanii is better suited to traditional genetics. It has a generation time of ∼2 h (Robinson et al., 2005), its genome is stable (Lopez-Garcia et al., 1995), and it grows on synthetic media (Mevarech & Werczberger, 1985). PEG-mediated transformation protocols were originally developed for Hfx. volcanii (Charlebois et al., 1987; Cline et al., 1989), and methods for gene knockout have incorporated additional selectable markers (Bitan-Banin et al., 2003; Allers et al., 2004), thus facilitating the construction of multiply-mutated cells (Fig. 2). A Gateway system for deletion construction (El Yacoubi et al., 2009), several reporter genes (Holmes & Dyall-Smith, 2000; Reuter & Maupin-Furlow, 2004), an inducible promoter (Large et al., 2007), and a system for protein overexpression (Allers et al., 2010) have been developed in the last decade. Haloferax volcanii has a long history of genome research going back to 1991, when a physical map of overlapping genomic clones was published (Charlebois et al., 1991). Publication of the annotated genome sequence in 2010 (Hartman et al., 2010) has spurred the development of whole-genome microarrays (S. Chimileski & T. Papke, pers. commun.), which will allow faster data analysis than the shotgun DNA microarrays available previously (Zaigler et al., 2003).

Rudimentary genetic systems are available for other haloarchaea. Haloarcula marismortui can be transformed with shuttle plasmids from Hbt. salinarum and Hfx. volcanii (Cline & Doolittle, 1992) and gene replacement studies have been used to investigate the Har. marismortui ribosome (Tu et al., 2005). However, this species harbors restriction/modification systems that reduce the efficiency of transformation by 104-fold (Cline & Doolittle, 1992). Systems for Halorubrum are currently under development (S. Chimileski & T. Papke, pers. commun.). Haloferax mediterranei is closely related to Hfx. volcanii; they differ in that Hfx. mediterranei produces gas vesicles. These gas-filled proteinaceous particles are used by cells to increase their buoyancy and float to the surface of the brine; they have been studied intensively by the laboratory of Felicitas Pfeifer, often using Hfx. volcanii as a heterologous host (Pfeifer et al., 2002; Hechler & Pfeifer, 2009). This is one of the great strengths of haloarchaeal genetics: because there are several organisms with mature genetic systems, complementation by genes from a related species can be used to identify random mutations and thereby isolate novel enzymes. This approach was used to find a novel thymidylate synthase and an alternative pathway for reduced folate biosynthesis in Hbt. salinarum, using genes from Hfx. volcanii (Giladi et al., 2002; Levin et al., 2004). More recently, the essential pitA gene of Hfx. volcanii was replaced by an ortholog from the haloalkaliphile Natronomonas pharaonis; the latter lacks the histidine-rich linker region found in Hfx. volcanii PitA and does not copurify with His-tagged recombinant proteins (Allers et al., 2010).

Genetic tools for halophiles

Transformation

Modern genetics takes for granted a method for introducing DNA into cells, and a means to select for cells that have taken up the DNA. The development of transformation protocols is intimately linked with selectable markers. This was an acute problem in the early days of archaeal genetics, because bacterial antibiotics are largely ineffective against archaea (Hilpert et al., 1981). Cline and Doolittle overcame this hurdle by assaying for transfection of Hbt. salinarum with naked DNA from halovirus ΦH, and scoring for plaques on a lawn of cells (Cline & Doolittle, 1987). This allowed them to develop the PEG transformation protocol that is used today (Cline et al., 1989): the glycoprotein cell surface layer, which depends on Mg2+, is removed by treatment with EDTA and DNA is introduced into spheroplasts using PEG 600, after which cells recover in rich broth before plating on selective medium. This protocol yields up to 107 transformants μg−1 DNA, depending on restriction/modification systems. Haloferax volcanii has two such systems: one that targets unmethylated 5′-CTAG-3′ sites (Charlebois et al., 1987) (T. Allers, unpublished data) and another that restricts methylated 5′-GmeATC-3′ DNA (Holmes & Dyall-Smith, 1991). The latter results in a 102-fold drop in transformation efficiency and has been circumvented by passaging DNA through an E. coli dam mutant, which lacks the methylase that modifies 5′-GATC-3′ sites. This is no longer necessary, because anΔmrr mutant of Hfx. volcanii was recently shown to lack the restriction enzyme that targets 5′-GmeATC-3′ DNA (Allers et al., 2010).

Shuttle vectors

Many replicative shuttle vectors have been developed, using origins of DNA replication taken from indigenous haloarchaeal plasmids (Table 2). For Hbt. salinarum there are plasmids based on pGRB1, pHH1, and pNRC100 origins (Blaseio & Pfeifer, 1990; Krebs et al., 1991; DasSarma, 1995), while for Hfx. volcanii there are vectors based on pHK2, pHV2, and pHV1/4 origins (Lam & Doolittle, 1989; Holmes et al., 1994; Allers et al., 2004; Norais et al., 2007). Some of these origins are broad range: for example, pHV2-based vectors replicate in both species (Blaseio & Pfeifer, 1990). Interestingly, pHV2-based plasmids do not function in Hfx. volcanii mutants deficient in the RadA recombinase, where pHK2-based vectors are used instead (Woods & Dyall-Smith, 1997). For a comprehensive list of plasmid vectors, see Allers & Mevarech (2005) or Berquist et al. (2006).

Table 2.   Genetic tools for haloarchaea
 Hbt. salinarumHfx. volcanii
  • *

    M. Dyall-Smith, pers. commun.

Synthetic mediaNo, lacks biosynthetic capability for five amino acids (Falb et al., 2008)Yes (Mevarech & Werczberger, 1985)
DNA deliveryPEG-mediated transformation (Cline et al., 1989)PEG-mediated transformation (Cline et al., 1989)
Restriction barrierSome restriction of methylated CTTCCT DNA*, lacking in some strains (Schinzel & Burger, 1986; Blaseio & Pfeifer, 1990)Severe restriction of GmeATC DNA, eliminated inmrr mutant (Holmes & Dyall-Smith, 1991; Allers et al., 2010)
Replicative shuttle vectorsBased on pGRB1, pHH1 and pNRC100 origins (Blaseio & Pfeifer, 1990; Krebs et al., 1991; DasSarma, 1995; Berquist et al., 2006)Based on pHK2, pHV2 and pHV1/4 origins (Lam & Doolittle, 1989; Holmes et al., 1994; Allers et al., 2004; Allers & Mevarech, 2005; Norais et al., 2007)
Positive selectionMevinolin, novobiocin, uracil (ura3) (Blaseio & Pfeifer, 1990; Holmes & Dyall-Smith, 1991; Peck et al., 2000)Mevinolin, novobiocin, uracil (pyrE2), leucine (leuB), thymidine (hdrB), tryptophan (trpA) (Lam & Doolittle, 1989; Holmes & Dyall-Smith, 1991; Bitan-Banin et al., 2003; Allers et al., 2004). Also histidine (hisC) and methionine (metX) (M. Mevarech, pers. commun.)
Counterselection5-FOA (ura3) (Peck et al., 2000)5-FOA (pyrE2) (Bitan-Banin et al., 2003; Allers et al., 2004)
Random mutagenesisUsing UV and X-rays (Soppa & Oesterhelt, 1989)Using ethyl methanesulphonate (EMS) (Mevarech & Werczberger, 1985)
Negative enrichmentUsing 5-bromo-2′-deoxyuridine (BrdU) (Soppa & Oesterhelt, 1989)Using 5-bromo-2′-deoxyuridine (BrdU) (Soppa & Oesterhelt, 1989; Wanner & Soppa, 1999)
Markerless gene knockout or replacementUsing ura3 (Peck et al., 2000)Using pyrE2 (Bitan-Banin et al., 2003; Allers et al., 2004). Gateway system available (El Yacoubi et al., 2009)
Ectopic integrationAt ura3 (Peck et al., 2000)At pyrE2 (T. Allers, unpublished data)
Natural genetic exchangeNot observed (Tchelet & Mevarech, 1994)Involves cell-cell contact (Rosenshine et al., 1989)
Reporter genesβ-Galactosidase (bgaH) and GFP (Nomura & Harada, 1998; Patenge et al., 2000)β-Galactosidase (bgaH) and GFP (Holmes & Dyall-Smith, 2000; Reuter & Maupin-Furlow, 2004)
Regulated gene expressionNoTryptophan-inducible p.tnaA promoter (Large et al., 2007)
Protein overexpressionpBBEV1 expression plasmid with constitutive bop promoter (Berquist et al., 2006), pNBPA expression plasmid with constitutive fdx promoter (Facciotti et al., 2007)pitANph gene replacement strain H1209 and tryptophan-inducible expression plasmid pTA963, for His-tagged proteins (Allers et al., 2010)
Selectable markers

Bacterial antibiotics that are safe to use in eukaryotes are largely ineffective against archaea, because the targets of these drugs (e.g. peptidoglycan cell walls) are not encountered in archaeal or eukaryotic cells. There are some exceptions (Hilpert et al., 1981); they have been exploited to develop selectable markers for haloarchaeal genetics. Novobiocin is an inhibitor of DNA gyrase (gyrB), an essential enzyme found in both bacteria and archaea, and a resistant form of the gyrB gene was isolated from Haloferax strain Aa2.2 (Holmes & Dyall-Smith, 1991). Novobiocin has since become the most widely used haloarchaeal antibiotic. An alternative is mevinolin (simvastatin), which inhibits HMG-CoA reductase. In humans, it is prescribed as a cholesterol-lowering drug, while in archaea it inhibits membrane synthesis. A mutant allele of the Hfx. volcanii hmgA gene that leads to overexpression of the enzyme has been harnessed as a mevinolin-resistant marker for both Hfx. volcanii and Hbt. salinarum (Blaseio & Pfeifer, 1990; Lam & Doolittle, 1992).

The last decade has seen a shift away from antibiotics and towards auxotrophic selectable markers, where genes involved in amino acid or nucleotide biosynthesis are used to complement chromosomal mutations. Deletion of the gene circumvents the instability that results from homologous recombination of the marker with a chromosomal allele; this is a problem for antibiotics, because both novobiocin- and mevinolin-resistance markers have homology to essential chromosomal genes. The first auxotrophic marker to be developed was the ura3 gene for uracil biosynthesis in Hfx. salinarum (Peck et al., 2000); it was followed by a similar system based on the pyrE2 gene of Hfx. volcanii (Bitan-Banin et al., 2003). These markers are particularly useful for gene knockout or replacement (Fig. 2b), because they can be counterselected using 5-fluoroorotic acid (5-FOA). Integration of a deletion construct is selected by transformation to uracil prototrophy and loss of the construct (resulting in gene deletion) is counterselected with 5-FOA, which is converted to toxic 5-fluorouracil in ura+ (but not ura) cells. Selection for 5-FOA-resistance is also used to bring about ectopic integration, where a gene construct is used to replace the ura3 or pyrE2 gene (Fig. 2e).

In Hfx. volcanii there are five additional auxotrophic markers. The trpA gene for tryptophan and leuB gene for leucine biosynthesis are often used in conjunction with pyrE2 (Allers et al., 2004), for gene replacement with a selectable marker (Fig. 2c). This has proved essential for the construction of some mutants, where the gene is important for cell viability and therefore difficult to delete. The hdrB marker for thymidine biosynthesis is well suited for shuttle plasmids, because it allows them to be maintained in rich media based on yeast extract (which lacks thymidine) (Allers et al., 2004). For example, a shuttle vector with the hdrB marker was used for in trans complementation of radA (Fig. 2d) to generate a radA mre11 rad50 mutant (Delmas et al., 2009). This approach, known as plasmid shuffling, was also used to demonstrate that Hfx. volcanii mutants lacking both the Holliday junction resolvase Hjc and the Xpf homolog Hef are synthetically lethal (Lestini et al., 2010). Additional selectable markers based on the hisC and metX genes of Hfx. volcanii (involved in histidine and methionine biosynthesis, respectively) have recently been developed (M. Mevarech, pers. commun.). This approach has been less fruitful for Hbt. salinarum, because it lacks the biosynthetic capability for five amino acids and cannot grow on synthetic media (Falb et al., 2008).

Regulated gene expression

The ability to regulate gene expression via a tightly repressed promoter is taken for granted in bacterial and eukaryotic systems. Heat-inducible chaperonin promoters have been available for some time in Hfx. volcanii (Kuo et al., 1997), but given the pleiotropic effects of heat-shock, they are far from ideal. In 2007 the p.tnaA promoter from the tryptophanase gene of Hfx. volcanii was characterized; it is tightly repressed in the absence of tryptophan and shows rapid, strong induction upon addition of ≥1 mM tryptophan (Large et al., 2007). It has been used to construct a depletion mutant of the essential cct1 gene (Large et al., 2007) and to generate an overexpression system for hexahistidine-tagged halophilic proteins (Allers et al., 2010). This overexpression system features a Hfx. volcanii host strain with deletion of the mrr gene, allowing high-efficiency transformation with DNA isolated from E. coli, and replacement of the pitA gene by its ortholog from N. pharaonis, preventing copurification with His-tagged recombinant proteins (Allers et al., 2010).

Reporter genes

A number of biosynthetic genes have been used as reporters in haloarchaea: for example the dhfr gene has been used to study transcription and translation in Hfx. volcanii (Danner & Soppa, 1996; Hering et al., 2009). Two colorimetric reporter genes are available for monitoring gene expression in haloarchaea: β-galactosidase and green fluorescent protein (GFP). They succinctly illustrate the ‘halophilic adaptation’ of existing genetic tools, which takes into account the high intracellular salt concentrations of haloarchaea. Because the lacZ gene product of E. coli is not active at high salt concentrations, the laboratory of Mike Dyall-Smith isolated a β-galactosidase gene bgaH from Haloferax alicantei that develops a blue color from X-gal (Holmes & Dyall-Smith, 2000). It has been used extensively as a transcriptional reporter in Hfx. volcanii and Hbt. salinarum, both of which lack detectable β-galactosidase activity (Patenge et al., 2000). A different approach was used to generate a halophilic version of GFP. Because GFP is a stable protein that is resistant to many denaturants, it was possible, by introducing just four mutations, to generate a soluble modified derivative that exhibits fluorescence in Hfx. volcanii cells (Reuter & Maupin-Furlow, 2004).

Discoveries and recent progress

The last 5 years have seen considerable progress in the area of ‘whole genome’ studies of Hbt. salinarum. Systems biology has taken advantage of the genome sequence since its publication in 2000 (Ng et al., 2000), leading to insights into the cellular response to anaerobic growth, phototrophy, X-ray and UV irradiation, salinity and temperature shifts, heavy metal resistance, and phosphate limitation (Baliga et al., 2004; McCready et al., 2005; Muller & DasSarma, 2005; Kaur et al., 2006; Whitehead et al., 2006; Coker et al., 2007; Twellmeyer et al., 2007; Wende et al., 2009). Now that the Hfx. volcanii genome sequence is published (Hartman et al., 2010), we hope to see similar progress in the other model haloarchaeon. This species holds considerable potential for combining mutant strain construction by reverse genetics with ‘-omics’ technology, for example proteomics and metabolomics (Kirkland et al., 2008; Sisignano et al., 2010).

One of the most fascinating aspects of halophiles is how their proteins can function in molar salt concentrations. The stability of halophilic proteins is partly due to their unusually hydrophilic surfaces, which leads to a requirement for efficient protein folding. This poses problems for secreted proteins, owing to a lack of ATP-dependent chaperones in the extracytoplasmic space. In halophiles, an unusually large number of proteins are rerouted via the twin-arginine transport (Tat) pathway, which allows cytoplasmic folding of proteins before their secretion (Rose et al., 2002; Hutcheon & Bolhuis, 2003). Haloarchaea have also yielded surprising insights into the evolutionary origin of cellular processes that were once thought to be exclusive to eukaryotes or bacteria. Ubiquitin-like proteins have been found in Hfx. volcanii and shown to act in protein conjugation, probably targeting these proteins for proteasome-mediated proteolysis (Humbard et al., 2010). A caspase-like activity has also been found; its expression is induced by salt stress and may play a role in programmed cell death (Bidle et al., 2010). N-linked glycosylation has been shown to be widespread in archaea, with a wider variety of sugar subunits than seen in eukaryal or bacterial glycoproteins it is commonly encountered in S-layer proteins and flagellins (Calo et al., 2010; Jarrell et al., 2010). Archaeal flagellins assemble to form a structure with superficial similarity to bacterial flagella, but on a molecular level they are unrelated (Ng et al., 2006). Studies on Hbt. salinarum have shown how a chemotaxis signal transduction system consisting of bacterial-like proteins, and proteins unique to archaea, is used to modulate rotation of the flagellum (Schlesner et al., 2009).

Thermococcales

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References

Introduction to Thermococcales, the heterotrophic, sulfur-reducing hyperthermophiles of the Euryarchaeota

The Euryarchaeota include an abundant number of hyperthermophiles that exhibit considerable diversity in terms of metabolism. Hyperthermophiles are found in the orders Archaeoglobales (sulfate reducers), Methanococcales/Methanopyrales (methanogens), and Thermococcales (sulfur reducers). The Thermococcales consist of three genera: Pyrococcus (Fiala & Stetter, 1986), Thermococcus (Zillig et al., 1983), and Palaeococcus (Takai et al., 2000). Although several exceptions are present such as the freshwater Thermococcus waiotapuensis (Gonzalez et al., 1999), the majority of the Thermococcales have been isolated from shallow marine thermal springs or deep-sea hydrothermal vents. They are all considered to be obligate anaerobes and heterotrophs that assimilate amino acids, peptides, pyruvate, and oligosaccharides, coupled with sulfur reduction or hydrogen fermentation (Amend & Shock, 2001). Among the three genera, Thermococcus contains the highest number of characterized isolates, and environmental studies have indicated that the members of Thermococcus are ubiquitously present in deep-sea hydrothermal vent systems (Orphan et al., 2000; Holden et al., 2001).

Most members of the Thermococcales exhibit optimal growth temperatures >80 °C. Members of the genus Thermococcus exhibit growth temperature ranges that fall between 50 and 100 °C, with optimal temperatures between 80 and 90 °C. Members of Pyrococcus exhibit slightly higher growth temperatures, with growth observed between 65 and 105 °C, with optimums between 95 and 105 °C. Palaeococcus ferrophilus, the only member of the genus studied in detail, grows between 60 and 88 °C, with an optimum of 83 °C (Takai et al., 2000). As the members of the Thermococcales grow at these high temperature ranges, all of their cell components, including membrane lipids, nucleic acids, and proteins, must display extreme thermostability in order to function at these high temperatures. Due to both fundamental and application-based interests, an abundant number of studies have focused on the structures of the thermostable proteins from hyperthermophiles.

In most cases, enzymes from hyperthermophiles are highly similar to their counterparts from mesophilic organisms in terms of both primary and tertiary structure, and share common catalytic mechanisms. Studies on enzymes from (hyper)thermophilic organisms have helped us understand the effects of temperature on enzyme activity (Daniel et al., 2010), and the structural characteristics that lead to protein thermostability. Enhancement of protein thermostability is brought about by several general strategies and their degrees of contribution vary according to the protein. One major strategy is the presence of extensive ion pair networks formed by acidic and basic amino acid residues. Glutamate dehydrogenases from a variety of Thermococcales species have been studied as model proteins, confirming the importance of ion pair networks in protein thermostability (Yip et al., 1995, 1998; Rice et al., 1996; Rahman et al., 1998; Vetriani et al., 1998). Another factor is increased packing and loop shortening. An increase in the number of buried atoms and a decrease in internal cavity volume is observed in the citrate synthase from Pyrococcus furiosus compared with its counterparts from mesophiles. The enzyme also has six loops that are shorter than those found in the pig citrate synthase (Russell et al., 1997; Arnott et al., 2000). These studies on citrate synthases have also revealed the importance of subunit interaction towards thermostability. The third major strategy is an increase in hydrophobic interactions in the protein core. O6-methylguanine DNA methyltransferase from Thermococcus kodakarensis harbors increased aromatic amino acids in the enzyme core compared with its counterpart from E. coli (Hashimoto et al., 1999).

Why study Thermococcales?

Other than the fact that they are obligate anaerobes, the Thermococcales can be grown on simple, organic media and exhibit cell yields sufficient for biochemical analyses. In addition, as complete genome sequences of three Pyrococcus species became available at a relatively early stage, members of the Thermococcales have been utilized to study a wide range of archaeal biology. These include DNA replication and repair (Hopkins & Paull, 2008; Williams et al., 2008; Yoshimochi et al., 2008; Kiyonari et al., 2009; Mayanagi et al., 2009; Nishida et al., 2009), transcription and its regulation (Vierke et al., 2003; Lee et al., 2005, 2007; Goede et al., 2006; Kanai et al., 2007; Santangelo et al., 2007; Hirata et al., 2008a), carbon and energy metabolism (Sapra et al., 2003; Verhees et al., 2003; Siebers & Schönheit, 2005), CRISPR systems (Hale et al., 2009), and cellular responses to stress, such as oxidative (Jenney et al., 1999; Clay et al., 2003), osmotic (Neves et al., 2005; Rodrigues et al., 2007), and temperature stress (Laksanalamai & Robb, 2004; Danno et al., 2008; Fujiwara et al., 2008; Kanzaki et al., 2008; Kida et al., 2008). The abundant genome sequences have also promoted a wealth of ‘-omics’ research including transcriptomics (Schut et al., 2003; Lee et al., 2006; Trauger et al., 2008), proteomics (Menon et al., 2009), structural genomics (Hura et al., 2009), and other genome-based high-throughput strategies (Keese et al., 2010).

In terms of metabolism, the Thermococcales assimilate a wide range of organic compounds, in many cases via novel or modified metabolic pathways that have not been identified in other organisms. They have thus attracted much attention to those interested in carbon/energy metabolism. The Thermococcales are also known to efficiently utilize polymeric substrates, namely poly- and oligosaccharides and peptides, and are armed with a vast array of stable, polymer-degrading hydrolases, which are expected to be applicable in various fields of biotechnology (Atomi, 2005; Egorova & Antranikian, 2005).

A variety of thermostable poly(oligo)maltosaccharide-modifying enzymes have been identified and characterized, which include α-amylases (Laderman et al., 1993; Dong et al., 1997a), amylopullulanases (Dong et al., 1997b), cyclodextrin glucanotransferases (Tachibana et al., 1999; Rashid et al., 2002a), 4-α-glucanotransferases (Jeon et al., 1997), maltodextrin phosphorylases (Mizanur et al., 2008), cyclodextrinases (Hashimoto et al., 2001), and branching enzymes (Murakami et al., 2006). In addition, many enzymes that cleave β-1,3- or β-1,4-glycosidic bonds have been studied such as β-glucosidases, β-galactosidases, β-mannosidases, endo-β-1,3-glucanases, chitinases, and β-glucosaminidases (Kengen et al., 1993; Voorhorst et al., 1995; Bauer et al., 1996; Gueguen et al., 1997; Driskill et al., 1999; Matsui et al., 2000). Based on genome sequence predictions, members of the Thermococcales each harbor over 30 protease/peptidase-related genes (Ward et al., 2002), and many of their protein products have been examined (Halio et al., 1996; Voorhorst et al., 1996; Story et al., 2005). In terms of the active site nucleophile, hyperthermophilic proteases include serine proteases, cysteine proteases, the threonine-dependent proteasomes, metal-dependent proteases, and those of which the catalytic mechanisms have not been elucidated. It is worthy to note that a remarkable number of crystal structures have been reported for the proteases/peptidases from the Thermococcales (Du et al., 2000; Arndt et al., 2002; Maher et al., 2004; Yokoyama et al., 2006; Delfosse et al., 2009; Dura et al., 2009).

Intracellular sugar metabolism has been another major topic of interest in the Thermococcales. Glycolysis is carried out through a modified Embden–Meyerhof (EM) pathway. Studies with P. furiosus have revealed that sugar phosphorylation is carried out by novel ADP-dependent glucokinases and ADP-dependent phosphofructokinases that are structurally unrelated to the ATP-dependent enzymes of the classical EM pathway (Kengen et al., 1994, 1995; Tuininga et al., 1999). Fructose-1,6-bisphosphate aldolases are structurally distinct to the previously known enzymes from bacteria/eukaryotes (Galperin et al., 2000; Siebers et al., 2001; Imanaka et al., 2002), as is the case with the gluconeogenic enzyme fructose-1,6-bisphosphatase (Rashid et al., 2002b; Sato et al., 2004). Metabolism of glyceraldehyde 3-phosphate (GAP) involves a novel GAP:ferredoxin oxidoreductase (Mukund & Adams, 1995; van der Oost et al., 1998), in addition to the phosphorylating GAP dehydrogenase and phosphoglycerate kinase. In the final step from phosphoenolpyruvate to pyruvate, phosphoenolpyruvate synthase is the major enzyme rather than the well-known pyruvate kinase (Imanaka et al., 2006). Studies on this single glycolytic pathway have revealed the presence of enzymes with novel structures, novel activities, and novel metabolic roles.

An encouraging fact for those interested in studying the Thermococcales is the wealth of sequence information that has accumulated in recent years. Complete genome sequences have been reported for P. furiosus JCM8422 (Robb et al., 2001), Pyrococcus abyssi GE5 (Cohen et al., 2003), Pyrococcus horikoshii OT3 (Kawarabayasi et al., 1998), T. kodakarensis KOD1 (Fukui et al., 2005), Thermococcus gammatolerans EJ3 (Zivanovic et al., 2009), Thermococcus onnurineus NA1 (Lee et al., 2008), and Thermococcus sibiricus MM739 (Mardanov et al., 2009), and are publicly available for Thermococcus barophilus MP and Thermococcus sp. AM4. This will provide an advantage not only in structure–function studies of individual proteins, but also in predicting functional relationships of genes in various biological systems of the Thermococcales.

Genetic tools for Thermococcales

Genetic manipulation techniques have been developed for T. kodakarensis (Morikawa et al., 1994; Atomi et al., 2004a) by the groups of Tadayuki Imanaka and John Reeve. As will be described below, gene disruption, insertion, and replacement on the chromosome occur through homologous recombination. Thermococcus kodakarensisE. coli shuttle vectors have also been developed. Strong constitutive promoters have been identified and can be used for gene expression. A β-glycosidase gene and a chitinase gene have been utilized as reporter genes. As for members of Pyrococcus, shuttle vector-based transformation systems are available for P. abyssi (Lucas et al., 2002), and have recently been developed in P. furiosus (Waege et al., 2010). The current range of genetic tools for Thermococcales is shown in Table 3.

Table 3.   Genetic tools for Thermococcales
 T. kodakarensisP. abyssiP. furiosus
Defined mediaYes (Sato et al., 2003)Yes (Lucas et al., 2002)Yes (Blumentals et al., 1990)
DNA deliveryTransformation (Sato et al., 2003)PEG-mediated transformation (Lucas et al., 2002)Transformation (Waege et al., 2010)
Restriction barrierNoNoNo
Replicative shuttle vectorspLC70 (Santangelo et al., 2008a, b)pYS2 (Lucas et al., 2002)pYS3 (Waege et al., 2010)
Positive selectionSimvastatin (Matsumi et al., 2007) Simvastatin (Waege et al., 2010)
Counterselection5-FOA (pyrF) (Sato et al., 2005; Yokooji et al., 2009) 6-Methyl purine (TK0664) (Santangelo et al., 2010)  
Markerless gene knockout or replacementUsing pyrF (Sato et al., 2005; Yokooji et al., 2009) Using TK0664 (Santangelo et al., 2010)  
Ectopic integrationAt chitinase gene (TK1765) locus (Mueller et al., 2009)  
Reporter genesTK1761 (β-galactosidase) (Santangelo et al., 2008a, b, 2010)  
Regulated gene expressionfbp (TK2164) promoter (Hirata et al., 2008a, b)  
Protein overexpressiongdh (TK1431) and csg (TK0895) promoters (Matsumi et al., 2007; Mueller et al., 2009; Yokooji et al., 2009) gdh (PF1602) promoter (Waege et al., 2010)
Genetic systems using auxotrophic host strains in defined media

Uracil auxotrophs with mutations in the pyrE or pyrF genes of T. kodakarensis were positively selected in a medium containing uracil and 5-FOA. Homologous recombination was tested using a pyrF strain as the host strain and the wild-type pyrF gene as the marker. Specific gene disruption was achieved with plasmids designed for double-crossover recombination (Fig. 2a). Double crossover plasmids were used to delete the pyrF gene from the wild-type KOD1 strain, resulting in strain KU216 (ΔpyrF), and the pyrF marker gene was subsequently used to disrupt the trpE gene, leading to strain KW128 (ΔpyrF, ΔtrpEpyrF), a tryptophan auxotroph. The pyrF and trpE genes can be used as markers for gene manipulation in strains KU216 and KW128, respectively (Sato et al., 2003, 2005).

Thermococcus kodakarensis displays natural competency and cells collected from a routine culture can be used directly for transformation. However, frequencies are low (∼102 mg−1 DNA with homologous regions of 1000 bp), and the use of this methodology is limited to specific gene modifications, and is not applicable for experiments such as random mutagenesis/gene complementation. Frequencies become even lower when the length of the homologous region becomes shorter. No prototrophs are obtained with flanking regions of 100 bp (Sato et al., 2003, 2005).

The pop-in/pop-out strategy using the counterselectable pyrF gene is applicable in T. kodakarensis. Plasmids are designed so that two regions on the chromosome are directly fused on the plasmid, with the pyrF marker gene flanking this construct (Fig. 2b). Host cells (ΔpyrF) are transformed with the plasmid, and transformants that have undergone single-crossover recombination (pyrF+) at either one of the homologous regions are enriched in liquid medium depleted of pyrimidines. Cells are then spread on Gelrite solid medium containing uracil and 5-FOA, which allows only growth of cells that have undergone a second recombination event that removes the pyrF marker gene. If the second recombination event occurs at the same homologous region that was used in the first recombination, the genotype returns to that of the host strain, whereas recombination at the opposite region results in gene modification and marker removal. This strategy and other similar strategies utilizing counterselection with the pyrF gene have been used to construct double and triple auxotrophic host strains such as KUW1 (ΔpyrF, ΔtrpE) and KUWH1 (ΔpyrF, ΔtrpE, ΔhisD) (Sato et al., 2005), and various markerless gene knockout strains (Yokooji et al., 2009).

Counterselection is also possible using the TK0664 gene, annotated as hypoxanthine–guanine phosphoribosyltransferase. Deletion of this gene results in a strain that is resistant to 6-methyl purine. Therefore, by utilizing a host strain deleted of the TK0664 and trpE genes (T. kodakarensis TS517), a gene cassette consisting of the TK0664 and trpE gene under the control of strong promoters (trp-6MPS cassette) can be used for selection/counterselection (Fig. 2c) (Santangelo et al., 2010).

Genetic systems applicable in nutrient-rich media

Conventional antibiotic resistance marker genes cannot be used in hyperthermophiles due to the lack of thermostability in their protein products. However, a strategy based on inhibition of a particular endogenous protein by an antibiotic and relieving the inhibition by overexpressing the protein or by introducing a mutant protein insensitive to the antibiotic is feasible. As demonstrated in the halophiles, simvastatin, a specific inhibitor of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, displays severe effects on the growth of T. kodakarensis. Growth of the wild-type T. kodakarensis KOD1 was completely inhibited for 5 days in the presence of 4 μM simvastatin. Overexpression cassettes for the endogenous HMG-CoA reductase gene from T. kodakarensis (hmgTk) and the heterologous gene from P. furiosus (hmgPf) have been shown to be applicable as selection markers (Matsumi et al., 2007). The promoter applied for overexpression was the 5′-upstream region of the glutamate dehydrogenase gene (Pgdh) from T. kodakarensis. Transformants harboring the overexpression cassette display resistance and can be selected in the presence of 10 μM simvastatin. Although both hmgTk and hmgPf can be used, the hmgPf gene is recommended as it will prevent unintended recombination at the native hmg locus that can occur with the endogenous hmgTk. This system allows gene disruption in nutrient-rich media, and can be directly applied on the wild-type T. kodakarensis KOD1. Moreover, this system has proved applicable in several other Thermococcus species such as T. onnurineus NA1 (Kim et al., 2010).

Another system that has recently been reported utilizes a T. kodakarensis strain deleted of its arginine decarboxylase gene (pda). Arginine decarboxylase converts arginine to agmatine, a vital precursor for polyamine biosynthesis. Even in a nutrient-rich medium, the pda disruption strain can only grow when agmatine is supplemented to the medium (Fukuda et al., 2008). The pda gene can thus be used as a selectable marker in nutrient-rich medium (without agmatine) when a pda gene disruption strain (T. kodakarensis TS559) is used as a host (Santangelo et al., 2010). This system allows selection in nutrient-rich medium without the addition of antibiotics.

Shuttle vectors in Thermococcus and Pyrococcus

Shuttle vectors have been developed that replicate stably and express selectable phenotypes in both T. kodakarensis and E. coli (Santangelo et al., 2008b). A plasmid from Thermococcus nautilis (pTN1) was ligated to a commercial vector for E. coli, and the selectable markers trpE and the Pgdh-hmgPf overexpression cassette were added so that T. kodakarensis transformants could be selected by ΔtrpE complementation and/or mevinolin resistance. The plasmids are maintained in T. kodakarensis at a copy number of approximately three copies per chromosome. The use of these plasmids for gene expression in T. kodakarensis has also been shown (Santangelo et al., 2008b).

Shuttle vector-based transformation is also possible in P. abyssi (Lucas et al., 2002) and P. furiosus (Waege et al., 2010). In the P. abyssi system, strains with mutations in the pyrE gene were used as host cells. The shuttle vector pYS2, which harbors the pyrE gene from S. acidocaldarius as a selectable marker, can be introduced into P. abyssi cells by a PEG-spheroplast method. pYS2 is stably maintained in P. abyssi with a high copy number of 20–30 copies per chromosome (Lucas et al., 2002). In the P. furiosus system, pYS2 was modified so that the pyrE gene was replaced by an hmgPf overexpression cassette (pYS3). The promoter used was the gdh promoter from P. furiosus. Transformants can be selected based on their resistance towards 10 μM simvastatin. pYS3 is stable in P. furiosus, and the copy number of the vector was 1–2 copies per chromosome. Induced expression of the RNA polymerase subunit D has been achieved using this vector. It has also been mentioned that the introduction of mutations on the chromosome of P. furiosus is now possible using hmgPf as a selectable marker (Waege et al., 2010).

Reporter genes

Thermococcus kodakarensis harbors two nonessential genes that encode β-glycosidases (TK1761 and TK1827). As intracellular activity deriving from TK1827 was very low and remained constant; an in vivo gene reporter system was established based on TK1761 (Santangelo et al., 2008a). This initial system has been utilized to display the occurrence of polarity in archaea (Santangelo et al., 2008a) and to elucidate the transcription termination signal [oligo(T) sequence] recognized by archaeal RNA polymerases (Santangelo et al., 2009). Recently, a T. kodakarensis strain deleted of both TK1761 and TK1827 has been constructed as an improved host for the in vivoβ-glycosidase gene reporter system (Santangelo et al., 2010).

Gene expression

Several strong promoters have been identified for use in gene expression in T. kodakarensis. The glutamate dehydrogenase gene promoter (Pgdh) mentioned above is a 551-bp 5′-flanking region of TK1431 (Matsumi et al., 2007). A 200-bp 5′-flanking region of the TK0895 gene (Pcsg), which encodes a cell surface glycoprotein, has also been utilized for gene expression (Yokooji et al., 2009). In addition, a heterologous promoter of the histone-encoding hmtB gene from Methanothermobacter thermautotrophicus (PhmtB) supports high expression levels in T. kodakarensis (Santangelo et al., 2010). Heterologous expression of the α-1,4-glucan phosphorylase gene from Sulfolobus solfataricus, which could not be functionally expressed in mesophilic host cells, was achieved in T. kodakarensis using Pcsg (Mueller et al., 2009). Heterologous gene constructs can be integrated into the chitinase gene locus without any apparent decrease in growth rate and cell yield (Fig. 2e) (Mueller et al., 2009).

Discoveries and recent progress

Until recently, the use of genetics in the Thermococcales has been carried out mainly in T. kodakarensis. In addition to the studies described above, disruption of the reverse gyrase gene in T. kodakarensis has demonstrated that the enzyme provides a significant advantage for hyperthermophiles to grow at high temperatures, but is not essential for life at 90 °C (Atomi et al., 2004b). The use of genetics has also contributed in elucidating metabolic pathways unique to the archaea, such as pentose synthesis via the reverse flux of the ribulose monophosphate pathway (Orita et al., 2006), AMP degradation via a novel route involving Type III Rubiscos (Sato et al., 2007), and coenzyme A biosynthesis involving two novel enzymes pantoate kinase and phosphopantothenate synthetase (Yokooji et al., 2009). Further genetic studies have provided insight into how T. kodakarensis responds to various carbon sources and environmental stress (Kanai et al., 2007, 2010). Application of the simvastatin-based gene disruption system to T. onnurineus NA1 has contributed to the identification of genes involved in growth of this strain on formate (Kim et al., 2010).

Sulfolobales

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References

Introduction to the Sulfolobales, the aerobic thermoacidophiles of the Crenarchaeota

The first member of the Sulfolobales, S. acidocaldarius, was described by T. Brock in 1972 and isolated from a hot spring in Yellowstone National Park (Brock et al., 1972). Later on, different members of Sulfolobales were characterized from acidic hotsprings and mudholes all over the world, with S. solfataricus (Pozzuoli, Italy; Zillig et al., 1980) and Sulfolobus tokodaii (Japan; Suzuki et al., 2002) being the most commonly used strains in laboratories. Recently, seven genomes of Sulfolobus islandicus were sequenced from a variety of acidic hot springs in the United States, Iceland, and Russia (Whitaker et al., 2005), and have been used to describe the mechanism of archaeal genome evolution (Reno et al., 2009).

Sulfolobus solfataricus and S. islandicus are metabolically the most diverse species and are able to grow on a wide variety of peptide sources, amino acids mixtures, and minimal media containing only sugars such as arabinose, glucose, sucrose, trehalose, cellobiose, and others by aerobic respiration (Grogan, 1989). In contrast, although S. acidocaldarius and S. tokodaii contain the genes for the nonphosphorylated Entner–Doudoroff pathway and the partially overlapping alternative pathway that generates ATP (Siebers & Schönheit, 2005), they cannot grow in these media because they lack the wide variety of sugar and peptide uptake systems present in the previously mentioned strains (Elferink et al., 2001; Albers et al., 2004). Very recently, a systems biology approach was undertaken to understand the temperature adaptation of glucose metabolism in S. solfataricus (Albers et al., 2009). These studies initiated the development of standard operating procedures for a wide variety of techniques for Sulfolobales including fermentation, transcriptomics, proteomics, and metabolomics, which should facilitate the comparability of results once researchers use these standardized protocols (Zaparty et al., 2010).

The cell cycle in Sulfolobales is characterized by a short prereplication period and an extensive postreplication stage that accounts for up to 70% of the generation time (Lundgren et al., 2008). In stationary-phase Sulfolobus cultures, all cells contain two genome copies resulting in an increase in the average cellular DNA content relative to an exponentially growing culture (Bernander, 2007). In contrast to Euryarchaeota, Crenarchaeota do not exhibit FtsZ for septum formation during cell division. Recently, it was found that in S. solfataricus and S. acidocaldarius, homologs of the eukaryotic endosomal sorting (ESCRT) pathway are located at midcell before cell division (Lindas et al., 2008; Samson et al., 2008). However, the exact role of these proteins during division is unknown.

As the archaeal enzymes involved in DNA replication are more similar to their eukaryal counterpart than to their bacterial one, a common origin of the eukaryal and archaeal replication apparatus was further strengthened by the demonstration of three origins of replication in the genomes of S. solfataricus and S. acidocaldarius (Duggin et al., 2008). However, it is not yet clear whether all of these origins are used at the same time or are differentially induced.

As an important adaptation to their extreme habitat, membranes of Sulfolobus species contain a large amount of tetraether lipids (up to 98% of all membrane lipids) resulting in a monolayer membrane (Elferink et al., 1992). These lipids are highly proton-impermeable and therefore enable Sulfolobus to maintain an internal pH of 6.5 in a highly acidic surrounding (Moll & Schäfer, 1988; van de Vossenberg et al., 1995).

Full genome sequences have been determined for all commonly used Sulfolobus strains, and are characterized by a high A+T content (between 63% and 67%). The genome size ranges from 2 to 3 Mbp, with S. solfataricus having the largest genome (She et al., 2001). This is mainly due to the fact that 11% of the S. solfataricus genome consists of mobile elements including over 200 different IS elements whereas other Sulfolobus strains only contain very few mobile elements. Large genome arrangements are reported for S. solfataricus species (Redder & Garrett, 2006).

Sulfolobus species have been a source for the isolation and characterization of a large number of genetic elements such as viruses, plasmids, and conjugative plasmids (Zillig et al., 1996; Prangishvili et al., 1998, 2006; Peng et al., 2000; Greve et al., 2004). These genetic elements are intensively being studied and some of them have been used for the development of genetic tools (Berkner & Lipps, 2008).

Why study Sulfolobales?

Sulfolobales are the only representatives of the Crenarchaeota that are amenable for genetic manipulation so far. All of the above-mentioned strains have growth optima between 70 and 85 °C and grow at pH values of 2–3. A few Sulfolobus species are reported to grow chemolithoautotrophically (Huber & Stetter, 1991), but are easily cultivated aerobically under heterotrophic conditions in the laboratory and exhibit doubling times of 3–6 h. Sulfolobus spp. have developed into model organisms for studies on DNA translation, transcription, and replication, DNA repair, cell division, RNA processing, metabolism, and many other cellular aspects.

One reason for this popularity is that the proteins of hyperthermophiles such as Sulfolobus are very amenable for obtaining 3D structures, and the PDB database contains 501 crystal structures obtained from different Sulfolobus species (December 2010). Among these is the structure of its complete RNA polymerase (Hirata et al., 2008b; Korkhin et al., 2009), and understanding in eukaryotic DNA repair mechanisms has been achieved interpreting structures obtained from Sulfolobus (Liu et al., 2008).

As mentioned before, a systems biology project (SulfoSYS) was initiated leading to a wealth of metabolomic, transcriptomic, and proteomic data available and a set of standardized methods that will enable researchers to compare results from different laboratories better with each other (Albers et al., 2009; Zaparty et al., 2010). The first archaeal deep sequencing project was carried out with S. solfataricus providing detailed information of transcription start sites of all genes present in the genome, and identified 100 previously unknown expressed genes and antisense RNAs (Wurtzel et al., 2010).

Key species of Sulfolobales that have genetic systems

A major problem for the development of genetic tools in the Sulfolobales has been and still is that only two selectable markers, uracil auxotrophy and growth on lactose, are available. Moreover, the main species researchers worked on, namely S. solfataricus P1 and P2, until now have not yet shown to recombine foreign DNA into their genome. The breakthrough came when Worthington et al. (2003) used a natural mutant of the S. solfataricus strain 98/2 PBL2002 for the construction of the first directed deletion mutant. The strain had an insertion in the β-galactosidase gene (lacS) that is necessary for S. solfataricus to grow in minimal medium containing lactose. Selection for growth on lactose often resulted in revertants; therefore, a naturally occurring mutant of PBL2002, PBL2025, is now being used that exhibits a large deletion of 50 genes (Schelert et al., 2004). A disadvantage of working with the selection for growth on lactose is that this procedure is rather time consuming (7–14 days in liquid medium before plating; Albers & Driessen, 2008). PBL2025 has a different morphology to wild-type S. solfataricus cells, regularly showing very large cells in growing cultures. As the genome sequence of PBL2025 is not publicly available, it is unclear what causes the morphological changes and care should be taken when using this strain. For example, it has been demonstrated that PBL2025 shows a profound difference in the production of extracellular matrix once cells are attached to a variety of surfaces (Zolghadr et al., 2010).

The first strain that was used for the expression of proteins in Sulfolobus was the uracil auxotrophic S. solfataricus P1 strain PH1-16, which also contains an IS element in the lacS gene (Martusewitsch et al., 2000; Albers et al., 2006). However, so far it has not been possible to introduce any foreign DNA into the genomic DNA of this strain using analogous methods as for S. islandicus and S. acidocaldarius.

Very recently, MR31, a uracil auxotrophic S. acidocaldarius mutant (Reilly & Grogan, 2001), was used for the construction of deletion mutants (Wagner et al., 2009; Ellen et al., 2010). An advantage of this strain is the possibility of direct plating after electroporation and only 4–6 days growth until colonies appear on gelrite plates. A markerless deletion mutant can be obtained in <3 weeks, which makes this strain a prime candidate for deletion mutant studies (Wagner et al., 2009). A plasmid-based system has been developed for maltose-induced protein expression in MR31 (Berkner et al., 2010).

Another recently developed strain is the uracil auxotrophic S. islandicus E233S1 that has been used to obtain a markerless lacS mutant (Deng et al., 2009; She et al., 2009). A shuttle vector is available for this strain, which was used in a detailed study of the arabinose-binding protein promoter araS (Peng et al., 2009). Both the S. acidocaldarius and the S. islandicus strains have the advantage that 5-FOA can be used for counterselection (Fig. 2b).

Genetic tools for Sulfolobales

The transformation of Sulfolobus strains was already established in 1992 when Christa Schleper demonstrated that S. solfataricus could efficiently be transfected with SSV1 virus DNA by electroporation (Schleper et al., 1992). Because the lack of positive selection pressure made the use of plasmids for transformation impractical in early trials, E. coliSulfolobus shuttle vectors were constructed based on conjugative plasmids and viruses that would spread through a transfected culture (for a detailed review, see Berkner & Lipps, 2008). The most successfully used vector of this generation is pMJ0503, a shuttle vector based on SSV1 (Jonuscheit et al., 2003). This vector was used for promoter studies and adapted for the homologous and heterologous expression of tagged proteins in S. solfataricus (Jonuscheit et al., 2003; Albers et al., 2006). The vector has also been used for the complementation of deletion mutants in PBL2025 (Zolghadr et al., 2007; Frols et al., 2008). The in vivo overexpression with the virus vector of the translation elongation factor a/eIF2-γ demonstrated that this factor stabilizes mRNA in S. solfataricus (Hasenohrl et al., 2008).

Since the first deletion mutant was constructed in PBL2025 (Worthington et al., 2003), the development of genetic tools for Sulfolobales has gone into warp speed and the most recent methods will be discussed and summarized in Table 4.

Table 4.   Genetic tools for Sulfolobales
 S. solfataricus PBL2025S. islandicus E322SS. acidocaldarius
Defined mediaYes (Grogan, 1989)YesYes (Grogan, 1989)
DNA deliveryElectroporation (Schleper et al., 1992; Worthington et al., 2003; Albers & Driessen, 2008)Electroporation (Deng et al., 2009)Electroporation (Kurosawa & Grogan, 2005; Wagner et al., 2009)
Restriction barrierNoSuiI cuts at GCwGC (Sollner et al., 2006)SuaI, restricts unmethylated DNA at CCGG (Prangishvili et al., 1985)
Replicative shuttle vectorspEXSs (Cannio et al., 1998), pKMSD48 (Stedman et al., 1999), pMSS derivatives (Aucelli et al., 2006), pJlacS (Berkner et al., 2007), pMJ0503 derivatives (Jonuscheit et al., 2003)pRN2 based vectors (Deng et al., 2009)pCSV1 and pAG based vectors (Aagaard et al., 1996; Aravalli & Garrett, 1997), pCmalLacS (Berkner et al., 2010)
Positive selectionLactose (Worthington et al., 2003), uracil (Jonuscheit et al., 2003), hygromycin (Cannio et al., 1998)Uracil (Deng et al., 2009)Uracil (Wagner et al., 2009), alcohols (Aravalli & Garrett, 1997)
CounterselectionNo5-FOA (pyrEF) (Deng et al., 2009)5-FOA (pyrEF) (Wagner et al., 2009)
Markerless gene knockout or replacementUsing lacS (Schelert et al., 2004; Albers & Driessen, 2008)Using pyrE (Deng et al., 2009)Using pyrE (Wagner et al., 2009; Ellen et al., 2010)
Ectopic integration  At pyrEF and amyA (M. Wagner & S.-V. Albers, unpublished data)
Reporter geneslacS (β-galactosidase)(Jonuscheit et al., 2003)lacS (β-galactosidase)(Deng et al., 2009)lacS (β-galactosidase)
Regulated gene expressionaraS promoter (Albers et al., 2006)araS promoter (Peng et al., 2009)malE promoter (Berkner et al., 2010)
Protein overexpressionpMJ0503 virus vectors with arabinose inducible induction with His and Strep tags, pRN1-based vectors (Albers et al., 2006; Berkner et al., 2007)Based on pRN2 pHZ2lacS (Peng et al., 2009)Maltose inducible expression using pCmal (Berkner et al., 2010)
Insertional gene disruption and markerless genetic exchange

Marked gene disruptions in S. solfataricus PBL2025 were first obtained by single-crossover events using plasmid DNA and later by double-crossover events using linearized plasmids or PCR products (Fig. 2a and c) (Worthington et al., 2003; Schelert et al., 2004; Albers & Driessen, 2008; Wagner et al., 2009). Integration of plasmid DNA via single crossover leads to tandem integration (Wagner et al., 2009). Selection for positive transformants by growth on lactose is rather time consuming in this strain, and unfortunately direct plating of cells after electroporation does not yield colonies.

Very recently, in both S. islandicus and S. acidocaldarius, deletion mutants were obtained via homologous recombination via double-crossover and unmarked deletions via single-crossover events using uracil autotrophy for enrichment (Fig. 2b) (Deng et al., 2009; Wagner et al., 2009; Ellen et al., 2010).

Ectopic integration

To avoid complication by multiple copies during expression of genes or promoter fusion constructs in cells, ectopic integration was successfully used by inserting promoter fusion constructs into the amyA (α-amylase) locus of S. acidocaldarius (Fig. 2e) (M. Wagner & S.-V. Albers, pers. commun.).

Reporter genes, overexpression and controlled expression

lacS, the gene encoding the β-galactosidase, has been used in all plasmids to demonstrate expression, as its activity can easily be determined by an X-gal assay. In the virus-based plasmid pMJ0503 (Jonuscheit et al., 2003), the pRN1-based pA-pK (Berkner et al., 2007) and the pRN2-based pHZ2lacS (Deng et al., 2009) lacS was used as a reporter gene for promoter studies. In PBL2025 and S. islandicus the arabinose-inducible arabinose-binding protein promoter araS has been implemented for the expression of proteins and complementation of deletion mutants (Albers et al., 2006; Zolghadr et al., 2007). Because S. acidocaldarius does not contain an arabinose uptake system, the promoter of the maltose-binding protein was used for homologous and heterologous expression (Berkner et al., 2010). Unfortunately, the malE promoter is quite leaky even in the absence of maltose.

Homologous expression of proteins in Sulfolobus leads to the correct assembly of cofactors in recombinant proteins, and Histidine and Strep-tags are stably expressed on fusion proteins, which have been successfully isolated by affinity chromatography (Albers et al., 2006). The overexpression of a mutant protein was used to demonstrate a dominant negative effect in vivo (Samson et al., 2008) and expression vectors were also used to complement deletion mutants (Fig. 2d) (Zolghadr et al., 2007; Frols et al., 2008).

Discoveries and recent progress

Sulfolobus species are the only genetically tractable members of the Crenarchaeota. Various genetic tools are now available and have led to important new discoveries including a eukaryotic-like cell division apparatus (Samson et al., 2008), identification of a general stress response (Maaty et al., 2009), the identification of several transcriptional regulators (Schelert et al., 2006; Peeters et al., 2009), and new cell surface structures (Szabo et al., 2007; Zolghadr et al., 2007; Frols et al., 2008). A detailed mapping of the araS promoter was obtained, showing unexpected parts of the promoter to be involved in regulation (Peng et al., 2009), and in vivo studies showed that a/ELF2g counteracts 5′–3′ mRNA decay in Sulfolobus (Hasenohrl et al., 2008).

Conclusion and outlook

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References

To write a review encompassing an entire domain of life might seem overly ambitious, and archaea are indeed exceedingly diverse. On the other hand, the development of genetic technologies for each branch of archaea has involved common challenges. Characteristics that make archaea so interesting, such as thermophily, halophily, and strict anaerobiosis, pose challenges to routine laboratory culture. Furthermore, the immunity of the archaea to most conventional antibiotics has meant that other means of selection have had to be devised. This article describes how these challenges have now been overcome for various archaeal species. These developments allow us to more effectively characterize and exploit the very features that make archaea so fascinating as well as challenging. For instance, the Sulfolobales and Thermococcales host thermostable proteins, which, in addition to being of inherent interest, are very often used for biochemical studies, because they crystallize easily and are stable under laboratory conditions. The availability of genetic tools for Sulfolobus and Thermococcus species will now enable us to study these proteins not only in vitro but also in vivo. Among the methanogens, genetic approaches are starting to prove useful in elucidating the relatively constrained world of hydrogenotrophic methanogenesis as well as the more versatile methylotrophic methanogenesis (Costa et al., 2010). The halophiles have shown us how with a few cunning modifications, proteins can function in salt concentrations that were once thought to be incompatible with life.

Opportunities for studies in archaea have never been better. In addition to genetic tools, the number of genome sequences continues to expand. A glance at NCBI Microbial Genomes reveals a current repertoire of seven strains of Methanococcus, three of Methanosarcina, six of Thermococcus, three of Pyrococcus, 11 of Sulfolobus, two of Halobacterium, and one of Haloferax, to say nothing of taxa for which genetic tools are not yet available. Clearly, comparative studies are now feasible – they are important as well. The considerable diversity of archaea offers countless opportunities for interesting discoveries, but no single archaeal species can be representative of the domain as a whole, or even its own specific grouping. For example, Hbt. salinarum is known to be extremely resistant to UV irradiation, while Hfx. volcanii, a closely related species that grows in the same habitat, is not (Baliga et al., 2004; Delmas et al., 2009). Rather than extrapolating results from one archaeal model organism to another, it is better to study each organism in its own right. Encouragingly, genetic tools continue to expand into new taxa. For example, shuttle vector-based technology is available in P. abyssi and P. furiosus, and chromosome modification in the latter species is now possible. As the application of these tools to other members of this order has just begun, we will probably be able to witness the development of genetic systems in a wide range of Thermococcales in the near future. Systems biology is also making inroads. Numerous global regulation studies have been carried out in Hbt. salinarum and a predictive model for transcriptional control of physiology has been generated – the first in any domain of life (Bonneau et al., 2007). Transcriptomic and proteomic studies have been carried out with members of all the archaeal groups described here. As systems biology matures in these organisms, genetics will play an increasing role in testing the hypotheses that are generated.

To be sure, the genetic toolbox for archaea needs to expand and the number of selectable markers is still limited. A driving force has been reverse genetics, where a gene is precisely deleted or mutated. Unfortunately, there is only so much that we can learn from deletion mutants. Regulated promoters, already in place for some archaea and under development in others, will facilitate the study of essential genes that cannot otherwise be mutated. In addition, if we are to take advantage of the explosion in genomic data, we must carry out mutagenesis on a whole-genome level. Traditional protocols for mutagenesis with chemicals, UV or X-rays have been used in Hbt. salinarum and Hfx. volcanii, and negative enrichment for nondividing cells is possible using 5-bromo-2′-deoxyuridine (BrdU) (Soppa & Oesterhelt, 1989; Wanner & Soppa, 1999). Similarly, in M. maripaludis, after mutagenesis with ethyl methanesulfonate, growing cells were selectively killed upon exposure to the base analogs 6-azauracil and 8-azahypoxanthine (Ladapo & Whitman, 1990). However, with these methods it is difficult to isolate the resulting mutations. Insertion mutagenesis using recombinant transposons is a solution that allows for facile identification of the mutant allele. Transposon mutagenesis has been worked out for Methanosarcina, but is still lacking in other archaea; in halophiles, early attempts met with limited success (Dyall-Smith & Doolittle, 1994; Woods et al., 1999). An alternative is to use next generation sequencing to identify mutations; this is currently too expensive for routine usage, but costs will inevitably fall to within the reach of all researchers.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References

We thank Y. Liu, B. Lupa, W. Whitman, W. Metcalf, S. Delmas, M. Wagner, S. Berkner, G. Lipps, M. Mevarech, and M. Pohlschröder for helpful comments, and T. Nunoura for assistance in constructing the phylogenetic tree.

Note added in the proof

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References

During the final stages of the reviewing process, Lipscomb et al. (2011) reported the development of a gene disruption system in Pyrococcus furiosus (REF). A pyrF deletion strain was used as the host strain, and markerless gene disruption was demonstrated via selection and counterselection with the pyrF marker based on uracil prototrophy and resistance towards 5-FOA.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methanogens
  5. Halophiles
  6. Thermococcales
  7. Sulfolobales
  8. Conclusion and outlook
  9. Acknowledgements
  10. Note added in the proof
  11. References