Every living cell is covered with a dense and complex array of covalently attached sugars or sugar chains. The majority of these glycans are linked to proteins via the so-called glycosylation process. Protein glycosylation is found in all three domains of life: Eukarya, Bacteria and Archaea. However, on the basis of the limit in analytic tools for glycobiology and genetics in Archaea, only in the last few years has research on archaeal glycosylation pathways started mainly in the Euryarchaeota Haloferax volcanii, Methanocaldococcus maripaludis and Methanococcus voltae. Recently, major steps of the crenarchaeal glycosylation process of the thermoacidophilic archaeon Sulfolobus acidocaldarius have been described. The present review summarizes the proposed N-glycosylation pathway of S. acidocaldarius, describing the phenotypes of the mutants disrupted in N-glycan biosynthesis as well as giving insights into the archaeal O-linked and glycosylphosphatidylinositol anchor glycosylation process.
As in the other two domains of life, archaeal proteins can undergo a variety of modifications, e.g. acetylation, phosphorylation, signal sequence cleavage, methylation, ubiquitination or glycosylation (for details, see ). Among all post-translational modifications, protein glycosylation is the prevalent one. It is proposed that more than half of the eukaryotic proteins are modified by the attachment of sugar molecules . Although the first archaeal N-glycoprotein in Halobacterium salinarium was described in 1977, only in the last 5 years, owing to the availability of genome sequences and the development of genetic tools for archaea, has substantial progress in describing the enzymes involved in archaeal N-glycosylation pathways been made. Today, glycosylation is thought to be conserved across all three major domains of life: Eukarya, Bacteria and Archaea. Although the N-lycosylation pathways of each domain seem to have certain characteristics in common, Archaea display a mosaic of features from the eukaryal and bacterial pathway. However, the study of the archaeal glycosylation is still restricted to three distinct Euryarchaeota: Methanococcus voltae [3–6], Methanococcus maripaludis [7–9] and Haloferax volcanii [10–13]. Recently, the N-glycosylation process of the thermoacidophilic crenarchaeon Sulfolobus acidocaldarius has been analysed ([14–16], and B.H. Meyer, E. Peyfoon, P.G. Hitchen, A. Dell and S.-V. Albers, unpublished work), which opened the field of N-glycosylation research in Crenarchaeota and contributes to the increasing data available on enzymes responsible for biosynthesis and assembly of archaeal N-glycans. The present review focuses on recent studies of the N-glycosylation process in S. acidocaldarius and include remarks on the so far not analysed O-glycosylation and GPI (glycosylphosphatidylinositol) anchor process in Crenarchaeota.
Three major classes of glycosylation exist: (i) O-linked glycosylation, in which nucleotide-activated sugar precursors are attached stepwise to the hydroxy group of serine and threonine residues of a target protein , (ii) N-linked glycosylation, in which an oligosaccharide is linked to the amide group of a selected aspartic acid residue within a recognition site of a target protein , and (iii) GPI anchoring, in which the C-terminal part of a protein is linked by an ethanolamine-phosphate bridge to lipid-anchored glycans  (Figure 1). All three types of glycosylation are either found or are predicted to occur in Crenarchaeota and are discussed in detail below.
N-glycosylation in Crenarchaeota
Archaeal N-glycans are highly diverse and composed of a variety of different sugar residues, such as Glc (glucose), Man (mannose), rhamnose, glucuronic acid, iduronic acid, N-acetylgalactosamine, GlcNAc (N-acetylglucosamine), galactofuranose, QuiS (sulfoquinovose) and galactouronic acid [3,5,8,14,21–28]. In addition, each of these sugar moieties can be modified, e.g. by methylation, sulfation and or the addition of threonine. This wide diversity in sugars is beyond that found in the bacterial and eukaryal N-glycosylation processes [29,30].
The N-glycan of S. acidocaldarius is the only crenarchaeal N-glycan characterized. Surface-exposed proteins of S. acidocaldarius, such as the S-layer , cytochrome b558/566 , and the archaellin FlaB  have been shown to be modified with a heterogeneous family of glycans, with the largest being a tribranched hexasaccharide. This tribranched hexasaccharide is composed of two GlcNAc residues, two Man residues, a sulfated sugar called QuiS and a Glc residue (Figure 2). Interestingly, S. acidocaldarius is the only archaeal species characterized so far whose glycans are linked via a chitobiose core, in which the two GlcNAc residues are connected by a β1–4 linkage. This chitobiose core is a common feature of all eukaryal N-linked glycans. Moreover, the tribranched topology of the Sulfolobus N-glycan is reminiscent of eukaryotic glycans which are usually multiantennary. So far branched N-glycans seem to be a common feature of thermophilic archaea, as the N-glycans of S. acidocaldarius , Pyrococcus furiosus  and Thermoplasma acidophilum  are branched, whereas the N-glycans of M. maripaludis , M. voltae , Hfx. volcanii  and Haloarcula marismortui  are linear. However, there might be one exception for the branching of thermophilic N-glycans, as the N-glycan of Methanothermus fervidus is proposed to be a linear heterosaccharide . Analysis of the glycosylation ratio of the large S-layer subunit SlaA from S. acidocaldarius revealed an astonishingly high degree of glycosylation averaging one modification for each stretch of 30–40 amino acid residues . This high glycosylation density has not been reported for any other archaeon so far, and might be an adaptation to the high temperature and acidic environment of S. acidocaldarius. Indeed, the prediction of glycosylation sites within the large S-layer protein of all Sulfolobales revealed also a high degree of possible glycosylation sites (Table 1), whereas other archaea displayed a lower number of N-glycosylation sites.
In the following sections, each biosynthesis step of the S. acidocaldarius N-glycan is described in the order of the assembly pathway (Figure 3), starting with the lipid carrier on which the N-glycan is assembled and ending with the transfer of the oligosaccharide on to the target protein.
The lipid carrier
In all three domains of life, the N-glycosylation process starts with the assembly of an oligosaccharide on a phosphopolyprenol carrier, either at the cytoplasmic membrane (Bacteria and Archaea) or at the ER (endoplasmic reticulum) (Eukarya). Similarly to eukarya , archaea use Dol (dolichol) as the lipid carrier [32,34], whereas bacteria are using undecaprenol as the lipid carrier . The search for the lipid carrier in the thermoacidophilic crenarchaeon S. acidocaldarius revealed unusually short and highly saturated DolP (dolichyl phosphate) species, not previously reported . Three distinct DolP species were detected in S. acidocaldarius, containing 40, 45 and 50 carbon atoms, with five saturated and three to five unsaturated isoprene units respectively. The C45 DolP from S. acidocaldarius is shown in Figure 4. These DolPs are unlike other archaeal or eukaryal counterparts as they are shorter than the C70–C110 DolP species found in eukarya  or the C55 and C60 DolP species observed in the archaeon Hfx. volcanii [32,36]. This shortness might be an adaptation to the cytoplasmic membrane, which consists of monolayer-forming tetraether lipids  with a fixed size of 40 carbon atoms, into which the DolP would fit perfectly. Furthermore, the saturation of the α- and ω-end distinguishes the archaeal Dol from the eukaryal counterparts, which are only saturated at the α-end. However, the high degree of saturation is beyond that seen in the other archaea so far, and may reflect the need for a stable molecule in the face of elevated temperatures and acidity, conditions naturally encountered by S. acidocaldarius DolP. The search for lipid-linked glycans revealed that all three existing DolP species are linked with three distinct hexose residues . However, in that study, no DolP species bearing higher-ordered glycans could be detected. The short DolP species in S. acidocaldarius are probably not sufficiently hydrophobic to partition into the chloroform-based lower phase formed in the extraction employed. Therefore it is still likely that the N-glycans are assembled on these short Dol species.
Initiation of the crenarchaeal N-glycosylation process
The first step of the N-glycosylation process starts with the transfer of a sugar residue on to the lipid carrier. It seems that all archaeal N-glycans that are linked by a hexose residue to the aspartic acid residue are transferred from DolP-linked glycans [33,38,39], whereas all N-glycans identified so far linked by a hexosamine are transferred from DolPP (dolichyl pyrophosphate)-linked glycans . In S. acidocaldarius, the initiation of the N-glycosylation process could be predicted based on the structure and composition of the N-glycan tree (see above). The assembly of the N-glycan starts with the transfer of GlcNAc phosphate from nucleotide-activated GlcNAc precursor on to the lipid carrier DolP, similar to the eukaryotic N-glycosylation pathway . A BLAST search identified the gene Saci0093, hereinafter named aglH, encoding a predicted UDP-GlcNAc-1-phosphate:dolichyl phosphate GlcNAc-1-phosphotransferase, which shows 27% amino acid sequence identity with both the eukaryal Alg7 (yeast) and the human DPAGT1, 31% with the archaeal AglH from M. voltae and 26% with WecA/Rfe of Escherichia coli (B.H. Meyer, H. Shams-Eldin, B. Zolghadr, C. Schäffer and S.-V. Albers, unpublished work). AlgH from S. acidocaldarius has ten predicted TM (transmembrane) helices, which are separated by five internal and four external hydrophilic loops. The topology is similar to the eukaryal Agl7 enzyme and to the bacterial WecA [42,43], whereas the archaeal homologue from M. voltae possess only seven predicted TMDs (TM domains) . The alignment of AglH with the eukaryal, bacterial and archaeal orthologues revealed conserved amino acids within each of the five cytoplasmic loops, indicating a possibly conserved function across the three domains of life. On the basis of the sequence similarity as well as similarities of the distinct topology profile of AglH, it was proposed that AglH indeed initiates the N-glycosylation process in S. acidocaldarius. Although every attempt to delete the gene failed, indicating the essential properties of this enzyme, complementation assays with AglH confirmed the function as an UDP-GlcNAc-1-phosphate:dolichyl phosphate GlcNAc-1-phosphotransferase (B.H. Meyer, H. Shams-Eldin, B. Zolghadr, C. Schäffer and S.-V. Albers, unpublished work). In this complementation assay, AglH was able to restore the functional loss of a conditional lethal Saccharomyces cerevisiae alg7 mutant (B.H. Meyer, H. Shams-Eldin, B. Zolghadr, C. Schäffer and S.-V. Albers, unpublished work). Alanine substitution of the conserved amino acids (Asp100, Phe220 and Phe264) within the cytoplasmic loops furthermore demonstrated the importance of these restudies for the enzymatic function for the complementation. On the basis of this complementation, sequence and topology similarity, it is most likely that the initiation of the N-glycosylation pathway in S. acidocaldarius occurs in the same way as shown in the eukaryal one, in which DolPP-linked GlcNAc is generated and used as a precursor for N-glycan assembly  (Figure 3).
Second N-glycan assembly step
In the second N-glycan assembly step, a second GlcNAc moiety is transferred on to the DolPP-linked GlcNAc, creating the DolPP-linked chitobiose core. This reaction is similar to the eukaryotic N-glycosylation process, in which the Alg13–Alg14 complex catalyses the transfer of GlcNAc from its UDP-GlcNAc donor on to DolPP-linked GlcNAc. Recently, it was shown that Alg13–Alg14 interacts with Alg7, the enzyme involved in the first step . This interaction tethers the soluble Alg13–Alg14 GTases (glycosyltransferases) to the ER membrane , which allows the enhanced biosynthesis of the N-glycan by clustering the GTase reactions. The interaction of Alg7 and Alg13–Alg14 is mediated by an N-terminal TMD of Alg14 . As S. acidocaldarius exhibits a chitobiose core, as the lipid-linking unit at the basis of its N-glycan as in Eukarya, a similar assembly process might be present in this archaeon. However, genomic analyses by the search for Alg13 and Alg14 orthologues in S. acidocaldarius did not lead to the identification of a GlcNAc transferase. Furthermore, this study revealed similarities of Alg13–Alg14 to the bacterial MurG, which catalyses the second step of peptidoglycan synthesis by creating UndPP (unadecaprenyl pyrophosphate)-MurNAc-(pentapeptide)-GlcNAc [46,47]. The search for a MurG homologue in the S. acidocaldarius genome identified Saci1262, a so far unknown predicted UDP-GlcNAc: GlcNAc transferase with a GT-B-type/MurG domain. However, every attempt to verify the predicted function of Saci1262 as an UDP-GlcNAc:GlcNAc transferase by deletion studies were unsuccessful and upcoming enzymatic assays will be used to elucidate the predicted function.
Third N-glycan assembly step
The GTase responsible for the third N-glycan assembly step has not been identified. However, on the basis of the defined N-glycan composition of the wild-type and agl3 deletion mutant (see below), the transfer of a Man residue connected by either an α1–4 or an α1–6 linkage to the second GlcNAc residue can be predicted .
Fourth N-glycan assembly step
In the fourth N-glycan assembly step, the nucleotide-activated QuiS has to be transferred on to the lipid-linked trisaccharide (Figure 3). This rare sugar is normally found as the head group of the non-phospholipid sulfoquinovosyldiacylglycerol, which is localized exclusively in the photosynthetic membrane of all higher plants, mosses, ferns and algae, as well as in most photosynthetic bacteria [48,49]. On the basis of sequence homologues of known bacterial sqdB or eukaryal sqd1 genes each encoding a UDP-sulfoquinovose synthase, the homologue Agl3 was identified in S. acidocaldarius . An activity assay of heterologously expressed Agl3, with its substrates UDP-Glc and sodium sulfite confirmed the function as a UDP-sulfoquinovose synthase . Furthermore, MS analyses of the N-glycan derived from the Δagl3 deletion mutant revealed the largest N-glycan to be a trisaccharide composed of Man1GlcNAc2. Interestingly, in addition to the expected missing QuiS and its linked Glc, also the second Man residue was absent from the N-glycan. This result demonstrated a highly ordered assembly process in which the QuiS and presumably its Glc residue have to be attached before the second Man residue can be transferred on to the glycan (Figure 3). Such a high-ordered N-glycan assembly process is also seen in Eukarya  and Bacteria [50,51]. In the gene locus of agl3, a gene coding for a predicted membrane-bound GTase family 39 member is found . Although the close localization to agl3 implied the function as a QuiS transferase, deletion of this gene did not result in the significant alteration of the N-glycan composition seen in the agl3 deletion mutant (B.H. Meyer and S.-V. Albers, unpublished work), and the GTase for the transfer of the QuiS remained unidentified.
Last N-glycan assembly step
The last assembly step of the N-glycan of S. acidocaldarius is catalysed by the GTase Agl16. Nano-LC–ESI–MS–MS (nano-liquid chromatography–electrospray ionization–tandem mass spectrometry) analyses of the SlaA glycopeptides from the agl16 deletion strain revealed the largest N-glycan to be a pentasaccharide, lacking one terminal hexose. The GTase Agl16, which is the first identified GTase involved in the crenarchaeal N-glycan biosynthesis, is a soluble 40.6 kDa protein which comprises a glycosyltransferase 1 domain (PF00534) spanning residues 167–330, whereas, outside this region, the protein contains no other known domain. The broad range of functions related to the glycosyltransferases 1 domain did not allow a precise prediction of the function of this GTase. A BLAST search of known protein structures of GT-B-folded GTase revealed a crystal structure (residues 172–338) of an archaeal GTase from Archaeoglobus fulgidus (PDB code 2F9F), is a homologue of the bacterial mannosyltransferase WbaZ from Salmonella enterica  and from E. coli K30 . However, HPLC analyses of the N-glycan revealed a significant change in the Glc/Man ratio, where the concentration of Glc in reduced in the Δagl16 strain (B.H. Meyer, E. Peyfoon, P.G. Hitchen, A. Dell and S.-V. Albers, unpublished work). This result indicates that Agl16 is probably a glycosyltransferase, catalysing the transfer of the terminal Glc residue on to the N-glycan tree.
Transfer of the N-glycan on to a target protein
The key step of N-glycosylation is the transfer of the fully assembled N-glycan on to a secreted target protein. In eukaryotes, this step is catalysed in the ER lumen by the multimeric OTase (oligosaccharyltransferase) complex. The OTase complex from S. cerevisiae is composed of eight non-identical membrane subunits , in which the Stt3p subunit contains the catalytic site. However, Stt3p alone is not sufficient for the OTase process [56–58]. The detailed function of the other subunits of the OTase complex is not fully understood, but they are thought to regulate and influence the modification of the N-glycosylation sites . In contrast with the huge eukaryal heteromeric OTase complex, only a single enzyme is needed for the OTase reaction in the prokaryotic system, which is an orthologue of the Stt3p [25,50] and no other OTase subunits have been detected in bacteria or in archaea. In contrast with bacteria, where orthologues of OTase key enzyme PglB have been found only in a few species of Deltaproteobacteria and Epsilonproteobacteria , the archaeal OTase AglB is found in almost all sequenced archaea [61,62], which underlines the broad distribution of the N-glycosylation process and the importance of this protein modification within the Archaea. Although AglB proteins exhibit a low overall sequence homology with the OTase orthologues, all AglB possess the highly conserved WWDYG motif [61,62]. The crystal structure of the C-terminal soluble domain of AglB from Pyrococcus furiosus was the first structure of an OTase . However, like the soluble part of the bacterial PglB , the soluble part of AglB was also non-functional . Analyses of the crystal structure and the alignment of AglB were used to identify a second conserved motif, DXXK , which lies in close structural proximity to the WWDYG motif and is thought to interact with this motif, thereby co-ordinating the target peptide [25,64]. On the basis of the sequence similarity of the highly conserved WWDYG and less conserved DXXK motif, the archaeal AglB can be grouped into three different OTase classes: the A (archaeal), B (bacterial) and E (eukaryal) types . The A-type is found in Halobacteria, Archaeoglobi and Methanomicrobia. The B-type is found in the Bacteria and in Archaea from the phyla Methanobacteria, Methanococci and Thermoplasmata. The E-type includes the eukaryal OTases as well as OTase from the thermophilic archaeal phyla Thermococci and Thermoprotei, including the Sulfolobales . In S. acidocaldarius, the OTase AglB (Saci1274) has a low sequence similarity to other OTases; however, the protein possesses a similar topology to other OTases identified, with 15 predicted TMDs and a soluble C-terminal domain, which is enclosed between TMD14 and TD15. Furthermore, the highly conserved WWDYG motif and the DXXK motif were found within the protein sequence. The presence of the conserved motifs strengthens the proposed function of Saci1274 as an OTase. According to Maita et al.  AglB from S. acidocaldarius can be grouped with the E-type OTases.
In contrast with the gene organization of Campylobacter jejuni and Hfx. volcanii, for which the gene coding for the OTase is found directly next to genes coding for GTases participating in the N-glycosylation process, no genes coding for a GTase are found near aglB (saci1274). The scattered localization of GTases distant to the OTases seems to be common in thermophilic archaea , which might be an adaptation to the extreme environment.
Furthermore, in contrast with euryarchaeal genomes, which possess up to three copies of AglB, such as in A. fulgidus , in the crenarchaeal genomes only a single copy of the OTase has been identified . Also multiple copies of OTase genes are also found within eukaryal and bacterial genomes [66,67]. In Eukarya, the different OTases are thought to have a different acceptor specificity, as in Trypanosoma brucei . Besides the multiple copies of AglB, GTase also clusters near each AglB homologue, implying multiple N-glycosylation pathways within the same genome. Indeed a different N-glycosylation process has been shown in H. volcanii which resulted in different N-glycans and glycosylation sites depending on the external salt concentration, which allow an adaptation to the environment .
It should be pointed out that not all euryarchaeal N-glycosylation processes studied are essential, as the deletion of AlgB resulted in a non-lethal phenotype [4,13]. These results are in contrast with S. acidocaldarius, as every attempt to create a deletion mutant of aglB in S. acidocaldarius was unsuccessful (B.H. Meyer and S.-V. Albers, unpublished work). Thus the fruitless attempts to delete the genes involved in the first two steps of the N-glycosylation process, and the drastic reduction of the fitness of the agl3 and agl16 deletion strains in S. acidocaldarius (see below) underline the importance of N-glycosylation in this crenarchaeon.
Effect of the phenotype by the agl deletion mutant
The agl deletion mutants, interfering in the biosynthesis of the N-glycan had significant effect on the S. acidocaldarius phenotype. On the basis of the high glycosylation density detected in the S-layer protein of Sulfolobus, N-glycosylation might have a greater influence to maintain cell integrity at high temperature, than in mesophilic archaea, which have only a lower glycosylation density (Table 1). Indeed, in the only report of thermoadaptation of S-layer proteins, the S-layer proteins of the hyperthermophilic Methanocaldococcus jannaschii and the Methanotorris igneus showed a higher number of potential N-glycosylation sites, compared with the mesophilic species  underlining the importance N-glycosylation might have for protein thermostability. This would explain why the N-glycosylation process in Sulfolobus is essential in contrast with the Euryarchaea studied.
Under standard growth conditions, no significant changes in the growth rates of the agl mutants compared with the background strain MW001 were observed (, and B.H. Meyer, E. Peyfoon, P.G. Hitchen, A. Dell and S.-V. Albers, unpublished work). The effect of the reduced glycan structure became evident when the strains were grown at higher salinities. Increased salinities lowered the growth rates of the deletion mutants in contrast with the wild-type strain in the agl16 deletion mutant 2- and in the agl3 deletion mutant 5-fold (, and B.H. Meyer, E. Peyfoon, P.G. Hitchen, A. Dell and S.-V. Albers, unpublished work).
This strong effect of the altered N-glycan lacking QuiS underlines the importance of QuiS, which probably plays a significant role in cell protection from the acidic and high-salt environment by its negative charge, which could co-ordinate multivalent cations. BLAST analyses with Agl3, responsible for the biosyntheses of UDP-QuiS (see fourth N-glycan assembly step) revealed the presence of Agl3 homologues in all thermoacidophilic Sulfolobales, in thermoacidophilic Crenarchaeota Acidilobus saccharovorans (pH 3.5–4.0), Caldivirga maquilingensis (pH 2.3–6.4) and Vulcanisaeta distributa (pH 4–4.5). Also in the halophilic Euryarchaeota Haloarcula marismortui, Haloquadratum walsbyi and in the Thermoplasmatales species Picrophilus torridus and T. acidophilum homologues of Agl3 were found. Indeed, a recent study of the structure and composition of the N-glycan from T. acidophilum confirmed the presence of a sulfated sugar which was identified as 6-deoxy-6-C-sulfo-D-galactose . All of these archaea, with the exclusion of the halophilic archaea, are thermoacidophilic with a high growth temperature (60–80°C) and at low pH optimum (2–4.5), implying an important adaptation to this extreme environment.
Besides the effect on the cell growth, the motility of the agl mutants is either totally impaired or significant reduced. In contrast with the background strain MW001, which showed strong motility after 9 days of incubation, the Δagl16 strain showed a drastically reduced swimming ability, whereas the ability to swim in the Δagl3 strain was totally abolished (B.H. Meyer, E. Peyfoon, P.G. Hitchen, A. Dell and S.-V. Albers, unpublished work). The complementation of Δagl16 restored motility, as the colony showed a similar swimming radius to that of the background strain MW001 (B.H. Meyer, E. Peyfoon, P.G. Hitchen, A. Dell and S.-V. Albers, unpublished work). Analysis of the cell surface by TEM (transmission electron microscopy) revealed, that in contrast with the background strain, the cells of the agl deletion mutants appeared non-archaellated, demonstrating that even a single missing sugar residue led to an instable archaella (B.H. Meyer, E. Peyfoon, P.G. Hitchen, A. Dell and S.-V. Albers, unpublished work). Defects in N-glycosylation also affected the motility, assembly or stability of the archaella in other archaea, e.g. Hfx. volcanii or M. maripaludis [4,69].
Besides the modification by the N-glycosylation process, archaeal proteins can also undergo O-glycosylation. In contrast with eukaryal and bacterial O-glycosylation, little is known about archaeal O-linked glycosylation. GTases initiating the O-glycosylation show a high degree of specificity for the polypeptide substrates. Thus the positions of O-glycan linkages in proteins are determined by the substrate specificity of each O-GTase [69,70]. Most of these O-glycosylated proteins possess a repeated stretch of serine or threonine residues. Also the small membrane-bound S-layer protein SlaB from S. acidocaldarius possesses such a stretch, which indicates that this protein might be O-glycosylated, as was shown for the S-layer protein of H. salinarium and Hfx. volcanii, for which regions have been shown to be O-glycosylated with galactose–Glc disaccharides [71,72]. Also in S. acidocaldarius, the membrane-bound cytochrome b558/566 have been shown to be modified by the O-linked Man residues . However, the O-glycosylation process in archaea is still an unsolved puzzle and prediction of this modification is very unreliable.
The GPI anchor is a glycolipid, which can tether secreted proteins to the membrane. In this post-translational modification, the glycan linker, which is glycosidically bound to the phosphatidylinositol group, is linked to the C-terminal end of proteins via an ethanolamine phosphate bridge (Figure 1). In eukarya, GPI-anchored proteins are found ubiquitously, and serve a variety of functions such as hydrolytic enzymes, adhesion molecules, complement regulatory proteins and receptors. All protein-linked GPI anchors share a common core structure, which consists of the ethanolamine phosphate-6-Man-(α1–2)-Man-(α1–6)-Man-(α1–4)-GlcN-(α1–6)-myo-inositol phosphate backbone. This backbone is to be further embraced by the different side chains and is linked to different lipids. In bacteria, the GPI anchor has not yet been found ; however, in archaea, GPI-anchor proteins are suggested. Analyses of archaeal genomes revealed a number of proteins showing a GPI signal sequence, and might therefore be substrates for the GPI anchor . Furthermore, homologues of the enzymes involved in the biosynthesis of GPI anchors have been detected in archaeal genomes [74,75]. So far there is only one experimental study of GPI-anchored proteins in archaea. In this study, three proteins from S. acidocaldarius have been identified, which incorporate radiolabelled GPI and caldarchaetidylinositol precursors . Although further experimental studies are needed to verify GPI-anchored proteins in archaea, it underlines the possible existence of archaeal GPI-anchored proteins.
Although the first studies allowed delineation of the N-glycosylation pathway in the crenarchaeon S. acidocaldarius, as proposed in Figure 3, many questions remained unsolved. Further studies will be undertaken to uncover the complete N-glycosylation pathway. It will be of special interest to establish whether the commonalities in the core structure between S. acidocaldarius glycans and those of Eukarya are mirrored in the same biosynthetic pathways by using orthologous GTases, which would give more insight into the evolution of N-glycosylation. Moreover, the identification of the glycosylation enzymes of S. acidocaldarius could lead to interesting biotechnological applications, as the core structure is equivalent to the eukaryal N-glycan.
B.H.M. was supported by the collaborative research centre 987 funded by the Deutsche Forschungsgemeinschaft (German Research Foundation) and S.-V.A. was supported by intramural funds of the Max Planck Society.
Molecular Biology of Archaea 3: An Independent Meeting held at the Max Planck Institute for Terrestrial Microbiology, Marburg, Germany, 2–4 July 2012. Organized and Edited by Sonja-Verena Albers (Max Planck Institute for Terrestrial Microbiology, Germany), Bettina Siebers (University of Duisberg-Essen, Germany) and Finn Werner (University College London, U.K.).
Abbreviations: A-type, archaeal type; B-type, bacterial type; Dol, dolichol; DolP, dolichyl phosphate; DolPP, dolichyl pyrophosphate; E-type, eukaryal type; ER, endoplasmic reticulum; Glc, glucose; GlcNAc, N-acetylglucosamine; GPI, glycosylphosphatidylinositol; GTase, glycosyltransferase; Man, mannose; OTase, oligosaccharyltransferase; QuiS, sulfoquinovose; TM, transmembrane; TMD, TM domain
- © The Authors Journal compilation © 2013 Biochemical Society