Thermophiles gain increasing attention in biotechnological applications due to their potential to expand the thermal range of industrial biotechnology and their unique metabolic capabilities (Littlechild, 2015; Zeldes et al., 2015; Beeler and Singh, 2016; Donati et al., 2016; Basen and Müller, 2017; Straub et al., 2017). In this review, we focus on the well-characterized members of the phylum Crenarchaeota, the extreme thermoacidophilic Archaea belonging to the genus Sulfolobus. Natural habitats of these organisms are solfataric fields all around the world, including the United States, Costa Rica, Mexico, Russia, Japan, China, New Zealand, Germany, Italy, and Iceland. The outstanding characteristic of these organisms, which have been investigated since the 1970s (Supplementary Table S1), is their ability to thrive at extremely low pH and high temperature, unprecedented in Eukaryotes and Bacteria.

Since Sulfolobus spp. can be grown and manipulated under laboratory conditions, they are popular model organisms to study Archaea. Research has been focused on their biology and physiology. Currently, genomics (Bell et al., 2002; Dai et al., 2016), proteomics (Chong and Wright, 2005; Ellen et al., 2010; Pham et al., 2010; Kort et al., 2013), metabolomics (Ulas et al., 2012; Bräsen et al., 2014), composition and function of the archaeal membrane (Albers and Meyer, 2011) and the archaellum (Albers and Jarrell, 2015), as well as interaction with archaeal viruses (Prangishvili et al., 2006) are important fields of research. Nevertheless, there is also growing interest in the utilization of this genus in biotechnological applications and the development of engineered strains to exploit the organisms’ unique characteristics. Sulfolobus spp. are a source of unique enzymes (Littlechild, 2015), biomaterials (Benvegnu et al., 2009; Besse et al., 2015), and metabolic pathways (Bräsen et al., 2014). As most prominent examples, the branched Entner–Doudoroff (ED) pathway (Kouril et al., 2013b) as well as Weimberg and Dahms pathways for the degradation of hexoses and pentoses (Nunn et al., 2010) should be named. These diverse catabolic pathways present a promising field for the exploitation of novel products (Ahmed et al., 2005; Siebers and Schönheit, 2005).

Among the eight Sulfolobus species established in the literature, S. islandicus, S. solfataricus, and S. acidocaldarius are by far the best described members of the genus. While S. islandicus is used as a model organism for comparative genomics and genetics (Reno et al., 2009) and for host–virus interactions (Held and Whitaker, 2009), no type strain has been designated and strains are not commercially available yet. S. solfataricus is the metabolically most diverse species and many catabolic enzymes have been investigated in detail (Bräsen et al., 2014). Unfortunately, this diversity comes along with a significant genetic instability caused by the presence of several hundred mobile elements identified in its genome (Brügger et al., 2002). By contrast, the genome of S. acidocaldarius is much more stable (Chen et al., 2005). This makes S. acidocaldarius interesting for industrial applications, where strain stability is of utmost importance. A phylogenetic tree of the genus Sulfolobus is shown in Figure 1.

In this review, we give an overview of the current state of knowledge on carbon metabolism, genetic tools, and fermentation techniques of Sulfolobus spp., describe relevant products, and discuss potential future applications of this genus.

Sulfolobus spp. thrive at pH 2–3 and temperatures around 75–80°C. They are characterized by a chemoorganoheterotrophic lifestyle; however, chemolithoautotrophic growth using sulfur oxidation has been reported for some species (Huber et al., 1992; Schönheit and Schäfer, 1995). All Sulfolobus species exhibit an aerobic lifestyle and for S. solfataricus P2, a preferred growth at lower oxygen concentrations was reported (Grogan, 1989; Simon et al., 2009). The different Sulfolobus strains differ significantly in their metabolic potential. S. solfataricus possesses a broad substrate specificity and uses various sugars such as polysaccharides (e.g., cellulose, starch, dextrin), disaccharides (e.g., maltose and sucrose), hexoses (e.g., D-glucose, D-galactose, D-mannose, and L-fucose), pentoses (e.g., D-arabinose, L-arabinose, D-xylose), aldehydes, alcohols (e.g., ethanol, phenol), sugar acids as well as tryptone, peptides, and amino acids as carbon source (Grogan, 1989; Izzo et al., 2005; Brouns et al., 2006; Joshua et al., 2011; Comte et al., 2013; Wolf et al., 2016; Stark et al., 2017). For S. solfataricus, a genome scale model comprising 718 metabolic and 58 transport/exchange reactions and 705 metabolites was used to simulate growth on 35 different carbon sources (Ulas et al., 2012). While no such modeled data are published for S. acidocaldarius, traditional growth experiments suggest that this species is well adapted to proteolytic growth and can utilize only few other carbon sources such as dextrin, sucrose, D-glucose, D-xylose, and L-arabinose (Grogan, 1989; Joshua et al., 2011). The differences in the metabolic potential are also reflected by the respective genome size of 2.99 Mbp including 200 IS elements for S. solfataricus (She et al., 2001) and of 2.23 Mbp for S. acidocaldarius (Chen et al., 2005). In the following paragraphs, we sum up the knowledge on the central carbohydrate metabolism and give an illustration of these pathways in Figure 2 (hexose and pentose degradation as well as glycogen, trehalose, and pentose formation).

Like most aerobic bacteria Sulfolobus spp. rely on the ED pathway for carbon degradation; however, in contrast to the classical pathway found, for example, in Pseudomonas species (Entner and Doudoroff, 1952), the archaeal pathway is branched and omits the initial phosphorylation of D-glucose. Instead, the sugar is directly oxidized to D-gluconate and dehydrated to 2-keto-3-deoxygluconate (KDG) as the characteristic intermediate of the pathway. In S. solfataricus KDG is either directly cleaved by the bifunctional aldolase to pyruvate and glyceraldehyde in the non-phosphorylative (np) branch of the ED pathway or first phosphorylated to 2-keto-3-deoxy-6-phosphogluconate (KDPG) and cleaved to pyruvate and glyceraldehyde 3-phosphate (GAP) in the semi-phosphorylated (sp) branch of the ED pathway. In the npED branch, glyceraldehyde is further oxidized and phosphorylated by glyceraldehyde:ferredoxin oxidoreductase and glycerate kinase to 2-phophoglycerate, which enters the lower shunt of the Embden–Meyerhof–Parnas (EMP) pathway (Ahmed et al., 2005). In the spED, GAP is oxidized to 3-phosphoglycerate by a non-phosphorylating GAP dehydrogenase (GAPN), activated by glucose 1-phophate, replacing the classical GAP dehydrogenase (GAPDH) and phosphoglycerate kinase (PGK) couple (GAPDH/PGK). The pathway in S. solfataricus is promiscuous for D-glucose and D-galactose (Lamble et al., 2005). Metabolome analysis of the KDG kinase deletion strain revealed a major function of the spED pathway in providing GAP for gluconeogenesis (Kouril et al., 2013b).

Pyruvate is further oxidatively decarboxylated to acetyl-CoA via the pyruvate:ferredoxin oxidoreductase; the classical pyruvate dehydrogenase complex is absent in Archaea. Acetyl CoA enters the oxidative citric acid cycle and is finally completely oxidized to two molecules of CO₂. The substitution of the catabolic GAPDH and PGK couple by GAPN results in no net gain of ATP in the branched ED pathway. Only in the citric acid cycle, the succinyl-CoA synthetase is supposed to provide nucleoside triphosphate (NTP) by substrate level phosphorylation. Therefore, the major energy gain comes from aerobic respiration. The respiratory chain in several members of the Sulfolobales has been studied, and in S. solfataricus as well as in S. acidocaldarius, a branched electron transport chain with three terminal oxidases was reported (Schafer et al., 1999; Auernik and Kelly, 2008). For S. solfataricus the regulation at transcriptome level in response to different oxygen concentrations was demonstrated (Simon et al., 2009).

In Sulfolobus spp., the EMP pathway is only used for gluconeogenesis, although for glycolysis only a functional phosphofructokinase is missing (Kouril et al., 2013a). As key enzymes, especially the classical GAPDH and PGK are only active in the gluconeogenic direction. Further on, a bifunctional, gluconeogenic fructose bisphosphate aldolase/phosphatase (FBPA/ase) catalyzes the one-step formation of fructose 6-phosphate from GAP and dihydroxyacetone phosphate (Say and Fuchs, 2010; Kouril et al., 2013a; Bräsen et al., 2014). Glycogen is formed as carbon storage compound (König et al., 1982) and as source for trehalose formation via the TreY/TreZ pathway [i.e., maltooligosyltrehalose synthase and maltooligosyltrehalose trehalohydrolase (Maruta et al., 1996)]. Trehalose is the only compatible solute reported so far in Sulfolobus spp.

Thus, like in all Archaea, the central carbohydrate metabolism in Sulfolobus spp. is characterized by unusual pathways and enzymes that – moreover – also confer unique regulatory properties. In contrast to the classical bacterial and eukaryotic EMP pathway, the regulation is established at the level of triose phosphates, which seems to be a general feature in (hyper)thermophilic Archaea with optimal growth close to 80°C. Triose phosphates are labile at high temperatures and it was shown that the thermal degradation of these pathway intermediates is a crucial bottleneck for efficient substrate conversion (Kouril et al., 2013a).

In addition, the upper part of the EMP pathway seems to play an important function for pentose generation. In Sulfolobus species, as in most Archaea, the classical pentose phosphate pathway is absent and pentoses are formed from fructose 6-phosphate via the reversed ribulose monophosphate pathway (RuMP) (Soderberg, 2005). The RuMP pathway was previously reported as formaldehyde fixation pathway in methylotrophic bacteria.

Pentose degradation has been studied in S. solfataricus and S. acidocaldarius. For S. solfataricus the D-arabinose degradation was resolved and an oxidative pathway with formation of α-ketoglutarate, which directly enters the citric acid cycle, was demonstrated (Brouns et al., 2006). Later studies revealed that the transporter and degradation pathway is partially promiscuous for L-fucose utilization (Wolf et al., 2016). The D-arabinose and D-xylose pathway merge at the identical intermediates 2-keto-3-deoxy-D-arabionoate (D-KDA) and 2-keto-3-deoxy-D-xylonate (D-KDX). For D-xylose degradation, a branched pathway with an aldolase-dependent branch forming pyruvate and finally glyoxylate (Dahms pathway), which enters the glyoxylate bypass, and an aldolase-independent branch forming the citric acid cycle intermediate α-ketoglutarate (Weimberg pathway) were proposed for S. solfataricus (Nunn et al., 2010). Important for cellulosic biomass conversion the absence of diauxic growth on D-glucose and D-xylose was reported for S. acidocaldarius (Joshua et al., 2011).

In general, the availability of genome scale models, functional genomics, and systems biology approaches for Sulfolobales under different stress and growth conditions in combination with biochemical and genetic studies enabled an in depth insight into metabolism and cellular processes [e.g., growth on L-fucose and casamino acids compared to D-glucose in S. solfataricus (Wolf et al., 2016; Stark et al., 2017)]. The established knowledge forms an important prerequisite for the establishment of Sulfolobus spp. as thermoacidophilic, archaeal platform organisms using metabolic engineering, and synthetic biology approaches for future biotechnological applications.

The lack of genetic tools has been a major drawback for the establishment of archaeal model organisms for basic research and biotech industries. The major problem was that most of the traditionally used antibiotics and resistance cassette genes cannot be used in archaeal phyla and therefore auxotrophies have to be used as selectable markers. However, nowadays very well-developed genetic toolboxes exist for the euryarchaea Thermococcus kodakarensis, Pyrococcus furiosus, Haloferax volcanii, and a number of methanogenic Archaea (Leigh et al., 2011). For Pyrococcus, it has been demonstrated that large gene clusters can be introduced for the production of several compounds (Lipscomb et al., 2014). Also for the genus Sulfolobus, a number of genetic systems have been established (Leigh et al., 2011). Early in the 1990s, the first transformation protocols by electroporation were established for S. solfataricus strains and self-transmissible vectors based on a conjugative plasmid, pNOB8, and the virus SSV1 were developed (Schleper et al., 1992; Elferink et al., 1996). The virus vector-based pMJ0503 was successfully used for the overexpression of tagged proteins in S. solfataricus (Albers et al., 2006). For the expression of proteins in S. islandicus, the plasmid pSeSD1 proved to be very useful (Peng et al., 2012). The first targeted deletion mutants were obtained in a S. solfataricus 98/2 PBL2025, which had a large deletion of 50 kB in the genome including many genes coding for proteins involved in sugar metabolism. As this strain was unable to grow on lactose as single carbon source, the β-galactosidase LacS could be used as marker cassette (Worthington et al., 2003). However, in this case no counterselection could be used to remove the marker cassette and therefore double deletion mutants could not be obtained. In the meantime, three model systems have developed, namely two in S. islandicus strains and one in S. acidocaldarius, which use mainly uracil auxotrophy for the selection and counterselection of mutants (She et al., 2009; Wagner et al., 2012; Zhang and Whitaker, 2012). Whereas the two S. islandicus strains contain a large number of transposable elements, which can lead to large genome rearrangements, the S. acidocaldarius genome is remarkably stable (Chen et al., 2005), which was shown by sequencing several strains isolated from North America, Russia, and Japan (Mao and Grogan, 2012). For S. acidocaldarius currently two uracil auxotrophic mutants are being used, MW001 (Wagner et al., 2012) and MR31 (Reilly and Grogan, 2001). For MW001 a whole set of genetic tools has been established. This includes several plasmids for the construction of markerless deletion mutants or for the insertion of tags into the genome (Wagner et al., 2012). Using these, the glucose ABC transporter of S. solfataricus was ectopically integrated into the MW001 genome and successfully expressed (Wagner et al., 2012). Based on the cryptic plasmid pRN1 from S. islandicus (Zillig et al., 1993), Escherichia coli–Sulfolobus shuttle vectors and expression vectors were established, which enabled the homologous or heterologous expression of tagged proteins of interest (Berkner et al., 2007, 2010). The S. acidocaldarius MW001 genetic system has been successfully used in a number of laboratories and helped to establish S. acidocaldarius as a model crenarchaeon. In a recent achievement, it was possible to harness the endogenous CRISPR/Cas system of S. islandicus for targeted genome editing (Li et al., 2016). This is a great next step in the direction of facilitated and accelerated manipulation of the genus Sulfolobus. Table 1 gives an overview of robust and highly cited expression systems and tools for gene disruption/deletion and genomic integration for the genus Sulfolobus.

The availability of potent genetic tools (Wagner et al., 2012; Peng et al., 2017) makes the transfer of heterologous genes to Sulfolobus species possible, allowing to benefit from both the metabolic diversity of S. solfataricus and the stability of S. acidocaldarius. In fact, the simpler, less promiscuous catabolism of S. acidocaldarius is an advantage over S. solfataricus in biotechnological applications, making it much easier to partly knockout metabolic pathways with the aim to redirect substrate fluxes toward a desired product.

To date, extremophiles are exploited as source of thermostable enzymes, so-called extremozymes, for food and feed industry, textile and cleaning industry, pulp and paper industry, but also in scientific research and diagnostics. Starch-hydrolyzing (Elleuche and Antranikian, 2013), (hemi)cellulolytic (Beg et al., 2001; Kuhad et al., 2011), pectinolytic (Sharma et al., 2013), chitinolytic (Chavan and Deshpande, 2013), proteolytic (Li et al., 2013), and lipolytic (Hasan et al., 2006) enzymes are in high demand in industry (Elleuche et al., 2015). Enzymes of Sulfolobus spp. are especially interesting for such applications not only because of their great catalytic diversity, but also mainly due to their superior pH and temperature stability, which comes hand-in-hand with increased resilience toward organic solvents and resistance toward proteolysis (Daniel et al., 1982; Unsworth et al., 2007; Stepankova et al., 2013). However, also tetraether lipids, membrane vesicles with antimicrobial properties, the storage component trehalose, and novel β-galactooligosaccharides are gaining importance nowadays. The most important products are shortly described below and summarized in Table 2.

It is evident that Sulfolobus spp. accommodate a huge variety of high value-added products useful in different fields of research and industry. However, this resource has basically remained untapped until now, due to a lack of proper bioprocessing tools. Of course, many of these products can also be produced recombinantly in mesophilic hosts. Benefits of the heterologous production in mesophilic hosts are much faster growth rates, highly efficient expression, extremely well-developed process technology, and facilitated downstream processing of thermostable proteins, since a considerable amount of host cell proteins can be readily removed via heat precipitation. Nevertheless, the production of proteins difficult to express and products remaining inactive due to differences in the expression and folding machinery, call for protein production in the archaeal host (Eichler and Adams, 2005; Kim and Lee, 2006). Furthermore, certain products are native cell constituents of Sulfolobus spp. (e.g., archaeal membrane containing tetraether lipids), which underlines the need to generate biomass and thus of bioprocess technology.

We are convinced that thermophilic bioprocesses have the potential to compete with conventional bioprocesses, since the drawbacks of typically lower growth rates and protein expression rates can be outweighed by a number of advantages resulting from the elevated process temperature:

Although there is a steadily growing interest in the development of extremophilic bioprocesses, no industrial process utilizing Sulfolobus spp. has been developed yet. Doubling times of at least 5–8 h (Brock et al., 1972; Grogan, 1989) and low biomass titers in batch cultures [max. 2 g/L dry cell weight (Schiraldi et al., 1999)] are the main obstacles for establishing efficient bioprocesses. The low biomass titer not only is a severe hindrance for biotechnological applications, but also poses a limitation for basic research because biomass and enzyme production of Sulfolobus spp. in shake flasks is painfully inefficient. As a result, archaeal enzymes are still mainly produced recombinantly in mesophilic hosts like E. coli, despite the aforementioned limitations.

In order to realize a competitive bioprocess, high cell densities in a reasonable time and economically feasible space-time yields must be achieved. This can be done by genetic engineering, optimized nutrient supply, and adjustment of process parameters. On the other hand, for bioconversion reactions, the issue of a low growth rate is not necessarily a neck-breaking drawback, if it is possible to integrate a cell-retention system combined with continuous cultivating. In that case, rather the maximum cell density, which is proportional to the volumetric catalytic activity, is a critical process parameter. However, studies on bioreactor cultivations with Sulfolobus spp. are still scarce.

As shown in Table 3, a high cell density cultivation is only reported for S. shibatae B12. However, it is evident that a sophisticated bioreactor setup including a cell-retention system is needed to realize a competitive bioprocess with Sulfolobus spp. Such a bioreactor setup is exemplarily depicted in Figure 3.

Remarkably, in none of the fermentations reported to date, defined media were used. Nevertheless, this is of high importance for the generation of platform knowledge and science-based process development. Use of defined media does allow not only the characterization and comparison of the variety of strains, but also the generation of comprehensive process understanding enabling process control and prediction. Furthermore, the use of defined media facilitates the transfer of process knowledge and speeds up process development and optimization. Another aspect worth considering is that bioprocesses that follow good manufacturing practice guidelines call for defined media to avoid batch-to-batch variability. These aspects underline the importance of the substitution of complex carbon sources, like yeast extract or protein hydrolysates, for the application of Sulfolobus spp. in industrial biotechnology for the production of high value-added products. Summarizing, to move Sulfolobus spp. into industrial biotechnology, (1) sophisticated bioreactor solutions and (2) defined media must be available.

Besides being a native source of high value-added products like extremozymes, extreme thermoacidophiles are predestined for the task of sustainably converting lignocellulosic biomass into value-added products due to their resilience toward harsh process conditions and their hemicellulolytic and cellulolytic properties (Turner et al., 2007). S. solfataricus in particular can grow on a very broad range of carbon sources (Grogan, 1989) and harbors a variety of polymer-degrading enzymes such as cellulases (Kufner, 2011), glucoamylases (Kim et al., 2004), alpha-amylases (Haseltine et al., 1996), beta-glucosidases (Moracci et al., 1995), xylanases (Cannio et al., 2004), and xylosidases (Moracci et al., 2000). Optimal growth in a hot, acidic environment means perfect synergy with the state-of-the-art method of substrate pretreatment utilizing high temperature and low pH. Although a variety of concepts for substrate pretreatment exists, the most favored process is the one of dilute sulfuric acid hydrolysis where concentrations of 0.5–1.5% sulfuric acid and temperatures between 120 and 180°C are commonly used (Carvalheiro et al., 2008; Maurya et al., 2015). Thus, pretreated substrate can be utilized in biorefinery applications based on Sulfolobus spp. with little to no need of neutralization and cooling of the medium. During the pretreatment process, a mixture of sugar monomers (mainly D-xylose, D-glucose, D-mannose, and L-arabinose) is released. In contrast to mesophilic hosts like Saccharomyces cerevisiae or E. coli, S. acidocaldarius lacks carbon catabolite repression (Ulas et al., 2012), thus allowing the efficient simultaneous utilization of a variety of sugars.

The combination of broad substrate specificity, lack of carbon catabolite repression, expression of polymer degrading enzymes, and extreme growth conditions make Sulfolobus spp. promising candidates for biorefinery applications. Following this approach, waste streams of the chemical and pulp and paper industry can be converted into value-added products. These processes would greatly benefit from the increased substrate solubility due to high temperatures and low pH. The availability of genetic tools and a broad variety of different strains are the basis for an application of Sulfolobus spp. in the biorefinery – however, the challenge of realizing a competitive bioprocess remains.

There are several reasons to be optimistic with respect to the use of Sulfolobus spp. in biotechnology. Greatly reduced contamination risk, high substrate solubility, adaption to harsh substrate pretreatment conditions, facilitated removal of volatile products, and elimination of cooling costs are benefits of high-temperature processes with Sulfolobus spp. The genus is a source of a broad variety of temperature and acid stable enzymes as well as a producer of unique biomaterials and metabolites. A well-developed genetic toolset makes exploitation of these features possible and emergence of metabolically engineered production strains is reasonable in the near future.

However, there is still a great need for careful bioprocess development. No continuous processes are reported in the literature and sophisticated tools for monitoring and control, like on-line measurement techniques for assessing cell viability, are lacking completely. Furthermore, media development and optimization have largely been neglected. For the establishment of a competitive, long-lasting, or continuous bioprocess, it is mandatory to generate basic process knowledge to be able to understand and control the bioprocess. Thus, we will tackle this challenge to be able to add Sulfolobus spp. as key player in industrial biotechnology in the future.

OS conceived the idea for writing this review. JQ drafted the manuscript, while S-VA and BS contributed the chapters on genetic tools and central carbon metabolism, respectively. LS contributed the figure describing the central carbon metabolism. OS critically reviewed and corrected the manuscript and gave substantial input.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.