In Archaea, dolichol phosphates have been implicated as glycan carriers in the N-glycosylation pathway, much like their eukaryal counterparts. To clarify this relation, highly sensitive liquid chromatography/mass spectrometry was employed to detect and characterize glycan-charged phosphodolichols in the haloarchaeon Haloferax volcanii. It is reported that Hfx. volcanii contains a series of C55 and C60 dolichol phosphates presenting saturated isoprene subunits at the α and ω positions and sequentially modified with the first, second, third and methylated fourth sugar subunits comprising the first four subunits of the pentasaccharide N-linked to the S-layer glycoprotein, a reporter of N-glycosylation. Moreover, when this glycan-charged phosphodolichol pool was examined in cells deleted of agl genes encoding glycosyltransferases participating in N-glycosylation and previously assigned roles in adding pentasaccharide residues one to four, the composition of the lipid-linked glycans was perturbed in the identical manner as was S-layer glycoprotein N-glycosylation in these mutants. In contrast, the fifth sugar of the pentasaccharide, identified as mannose in this study, is added to a distinct dolichol phosphate carrier. This represents the first evidence that in Archaea, as in Eukarya, the oligosaccharides N-linked to glycoproteins are sequentially assembled from glycans originating from distinct phosphodolichol carriers.
N-glycosylation is a post-translational modification performed by all three domains of life, namely Eukarya, Bacteria and Archaea (Helenius and Aebi, 2004; Eichler and Adams, 2005; Szymanski and Wren, 2005; Weerapana and Imperiali, 2006). Presently, the pathway of N-glycosylation is best understood in higher Eukarya, where the oligosaccharide covalently linked to select Asn residues of a target protein is first assembled from seven soluble nucleoside-activated sugars sequentially added to a dolichol pyrophosphate carrier on the cytoplasmic face of the endoplasmic reticulum (ER) membrane. The charged lipid carrier is then flipped to face the ER lumen, at which point seven additional sugars, derived from individually charged and flipped phosphodolichol carriers, are added. Once assembled, the 14-meric oligosaccharide is transferred to the protein target (Burda and Aebi, 1999; Helenius and Aebi, 2004).
In contrast to the detailed delineation of the eukaryal N-glycosylation pathway, much less is known of this post-translational modification in Archaea. As in the ER membrane, glycan-charged phosphodolichol species have been detected in the archaeal plasma membrane. Moreover, evidence exists assigning glycan-charged phosphodolichols roles in archaeal N-glycosylation, in analogy to the function these lipids serve in the parallel eukaryal pathway (Burda and Aebi, 1999; Helenius and Aebi, 2004). For example, the identical methylated hexasaccharide moiety as attached to the Methanothermus fervidus surface (S)-layer glycoprotein is found on a dolichol pyrophosphate carrier in this species (Hartmann and Konig, 1989; Kärcher et al., 1993). Likewise, the sulphated polysaccharide moiety N-linked to the Halobacterium salinarum S-layer glycoprotein and flagellin was detected on dolichol phosphate intermediates, while C60 dolichol phosphate species bearing glucose, mannose and N-acetylglucosamine units have also been observed in this organism (Mescher et al., 1976; Wieland et al., 1980; 1985; Lechner et al., 1985a; Sumper, 1987). Pulse-chase radiolabelling of Hbt. salinarum cells revealed the gradual transfer of the radiolabel from a lipid precursor to the S-layer glycoprotein (Wieland et al., 1980), while the incorporation of radiolabelled glucose into Haloferax volcanii glycoproteins was shown to proceed through a glucose-containing phosphopolyisoprenol intermediate (Zhu et al., 1995). Indeed, Hfx. volcanii membranes were reported to contain C55 and C60 dolichol phosphate charged with an α-linked mannosyl-(β1-4)-galactosyl group, and, to lesser extents, with a sulphated or phosphorylated dihexose and with a tetrasaccharide (comprising the hexoses, mannose and galactose and the deoxyhexose, rhamnose), as well as with monosaccharides at radiochemical levels (Kuntz et al., 1997). Not all of these phosphodolichol-charged glycans have, however, been detected on Hfx. volcanii glycoproteins (Sumper et al., 1990; Mengele and Sumper, 1992; Abu-Qarn et al., 2007; Magidovich et al., 2010).
Thus, while glycan-charged phosphodolichols are implicated in archaeal N-glycosylation, numerous questions remain unanswered. Are those glycans found on dolichol phosphate carriers in Archaea, presumably destined to decorate target proteins, assembled from soluble, activated monosaccharides, from monosaccharides transferred from individual dolichol phosphate carriers or from both, as in Eukarya? If so, what is the relative contribution of each monosaccharide population in generating oligosaccharide-charged phosphodolichols? Is the protein-targeted oligosaccharide fully pre-assembled on a single phosphodolichol carrier in the archaeal cytoplasm or does assembly of the phosphodolichol-charged oligosaccharide involve steps that transpire on both sides of the membrane? Finally, one can also ask whether assembly of the N-linked glycan includes the addition of sugar subunits to a glycan already transferred from its phosphodolichol carrier to the protein target. Based on advances in describing the archaeal pathway of N-glycosylation made in the last 5 years (for review, see Yurist-Doutsch et al., 2008a; Calo et al., 2010), it may now be possible to address these and related questions.
Now, to address the involvement of glycan-charged phosphodolichols in the biosynthesis of the Hfx. volcanii N-linked pentasaccharide, liquid chromatography/mass spectrometry (LC/MS) was employed to define the sugar profiles of phosphodolichol-linked glycan carriers in a Hfx. volcanii parent strain as well as from cells deleted of different agl genes. It is reported that Hfx. volcanii contains a series of dolichol phosphate molecules sequentially modified with the one, two, three and four sugar subunits corresponding to the first four subunits of the pentasaccharide found on the S-layer glycoprotein. The fifth pentasaccharide subunit, mannose, is, in contrast, derived from its own phosphodolichol carrier. These findings thus not only provide the first direct evidence for the sequential assembly of an oligosaccharide on a dolichol carrier prior to the addition of that glycan to an archaeal protein but also that the assembly of N-linked oligosaccharides in Archaea involves glycans originating from distinct dolichol carriers, as occurs in Eukarya.
Hfx. volcanii contains a population of C55 and C60 dolichol phosphate molecules
In the present study, a total Hfx. volcanii lipid extract was subjected to normal-phase liquid chromatography coupled with mass spectrometry. Figure 1A shows the total ion chromatogram of the NPLC-ESI/MS performed in the negative ion mode. The mass spectrum averaged from those acquired during the retention time of 20–21 min (Fig. 1B) shows prominent ion peaks of m/z 849.695 (this and all reported values are for the monoisotopic ion peaks, unless otherwise stated) and m/z 917.757, corresponding to the [M–H]- ions of the C55 and C60 dolichol phosphates with two saturated isoprene units respectively. These measured ion masses are in agreement with the calculated values of m/z 849.690 for the [M–H]- ion of the C55 dolichol phosphate and m/z 917.752 for the [M–H]- ion of the C60 dolichol phosphate. In addition, a very minor peak corresponding to C50 dolichol phosphate (m/z 781.671) was observed. Tandem mass spectrometry (MS/MS) was performed on the [M–H]- ion at m/z 917.7 of C60 dolichol phosphate (Fig. 1C); the obtained fragmentation pattern is consistent with the chemical structure previously described (Kuntz et al., 1997), with the saturated isoprene units at both the α and the ω positions. The same saturation pattern also held true for C55 dolichol phosphate (not shown). This is in contrast to the C55 undecaprenol involved in N-glycosylation in Bacteria, where the α position is unsaturated and the longer dolichols involved in eukaryal N-glycosylation (C70–C110), where the α position is saturated (Jones et al., 2009). The presence of two saturated isoprene units in archaeal dolichol phosphate is quite remarkable, considering that Bacteria contain only unsaturated polyprenol phosphate, while in eukaryal dolichols, only the α-isoprenes are saturated (Burda and Aebi, 1999). Recently, the long-sought reductase for converting polyprenol to dolichol in eukaryotic cells has been identified (Cantagrel et al., 2010). At present, no archaeal polyprenol reductase has been described.
Apart from the complete pentasaccharide, Hfx. volcanii contains dolichol phosphates charged with the same glycan series as found N-linked to the S-layer glycoprotein
Earlier studies revealed the Hfx. volcanii S-layer glycoprotein to be modified by an N-linked pentasaccharide comprising a hexose, two hexuronic acids, a methyl ester of hexuronic acid and a final hexose (Abu-Qarn et al., 2007; Magidovich et al., 2010). In addition, S-layer glycoprotein-derived peptides have also been shown to be modified by glycans comprising the first, the first two, the first three and the first four sugar subunits of the N-linked pentasaccharide (Abu-Qarn et al., 2007; 2008; Yurist-Doutsch et al., 2008b; 2010; Magidovich et al., 2010). To assess whether similar glycans also decorate Hfx. volcanii dolichol phosphate, phosphodolichol-linked glycans were profiled by NPLC/MS in the total lipid extract of Hfx. volcanii. Figure 2 (left panels) reveals the presence of dolichol phosphate species modified by a glycan comprising one to four saccharides, while MS/MS analysis confirmed the nature of the sugars added to the C60 dolichol phosphate (Fig. 2, right panels). Specifically, the fraction eluting during the retention time of 16–16.5 min contains C55 and C60 dolichol phosphate species modified by a hexose (peaks at m/z 1011.724 and 1079.797 respectively; Fig. 2A). In addition, a major peak at the m/z 1055.714, corresponding to a previously described sulphoglycolipid (Sprott et al., 2003), was also observed. The fraction eluting during the retention time of 20.8–21.3 min contains C55 and C60 dolichol phosphate species modified by a hexose and a hexuronic acid (peaks at m/z 1187.793 and 1255.858 respectively; Fig. 2B), while the fraction eluting during the retention time of 26–27 min contains C55 and C60 dolichol phosphate species modified by a hexose and two hexuronic acids (peaks at m/z 1363.832 and 1431.895 respectively; Fig. 2C). Finally, the fraction eluting during the retention time of 35.5–36 min contains C55 and C60 dolichol phosphate species modified by a hexose, two hexuronic acids and a methyl ester of hexuronic acid (their doubly charged ions [M–2H]2− are observed at m/z 766.43 and 810.46 respectively; Fig. 2D). The structure of the C60 phosphodolichol-linked tetrasacchride was verified by MS/MS (Fig. 2D, right panel). The most prominent ion at m/z 207 in the MS/MS spectrum derived from the [M–2H]2− ion at m/z 810.46 is derived from the terminal (i.e. the fourth) sugar. Its methanol-less (32 Da) ion, shown at m/z 175, provides additional evidence for this sugar subunit being a methyl ester of hexuronic acid.
Finally, although the S-layer glycoprotein is ultimately modified by a N-linked pentasaccharide, no pentasaccharide-modified dolichol phosphate species was detected. Instead, only dolichol phosphate species sequentially charged with the first four saccharides comprising the pentasaccharide N-linked to the Hfx. volcanii S-layer glycoprotein were observed.
Hfx. volcanii cells lacking components of the N-glycosylation machinery present dolichol phosphates void of or bearing truncated glycans
To assess whether the various glycan-charged phosphodolichols described in the previous section are involved in the N-glycosylation of the S-layer glycoprotein, the dolichol phosphate-derived species from Hfx. volcanii cells deleted of aglG, aglI, aglE and aglD were considered. Previous efforts implicated the products of these genes, predicted glycosyltransferases, in the respective addition of the second, third, fourth and fifth saccharides of the pentasaccharide decorating the S-layer glycoprotein, although direct biochemical proof for such activity has yet to be provided (Abu-Qarn et al., 2007; 2008; Yurist-Doutsch et al., 2008b).
When the total lipid extract from cells lacking AglG was analysed by NPLC/MS/MS as above, only hexose-modified C55 and C60 phosphodolichols were detected (Fig. 3). Likewise, cells deleted of aglI or aglE presented dolichol phosphate species containing mono- and disaccharides and mono-, di- and trisaccharides respectively. Consistent with these results, a recent study (Kaminski et al., 2010) showed that in the absence of AglJ, involved in adding the first sugar subunit of the S-layer glycoprotein N-linked pentasaccharide, the level of a hexose-modified phosphodolichol species was significantly decreased, relative to what is seen in the parent strain.
When dolichol phosphate-derived species from cells lacking AglD, implicated in adding the fifth and final subunit of the S-layer glycoprotein N-linked pentasaccharide, were compared with those of the parent strain, a very different effect was seen than observed in cells lacking AglG, AglI or AglE. Hfx. volcanii membranes include three hexose-modified phosphodolichol species retained at 14.77, 16.06 and 16.92 min. Of these, only the hexose-modified phosphodolichol with a retention time of 16.92 min was eliminated in the aglD deletion strain (Fig. 4). This implies that AglD is a dolichol phosphate hexose synthase that catalyses the addition of a hexose residue to the lipid carrier. To identify the monosaccharide apparently added to the dolichol phosphate carrier by AglD, a mannose-charged C55 phosphodolichol standard was examined as above. The mannose-modified phosphodolichol peak eluted at 16.99 min, just as was that peak missing from cells lacking AglD. Thus, the fifth and final pentasaccharide subunit, added to its own dolichol phosphate carrier and previously reported to be a hexose (Abu-Qarn et al., 2007), is now identified as mannose.
In conclusion, it is proposed that those enzymes previously assigned roles in adding the first four saccharide subunits of the pentasaccharide N-linked to the Hfx. volcanii S-layer glycoprotein sequentially attach their respective sugar substrates to a common dolichol phosphate carrier. In contrast, AglD, previously implicated in adding the fifth saccharide (now identified as mannose) to the N-linked glycan decorating the S-layer glycoprotein, adds its substrate to a distinct dolichol phosphate carrier.
It has been long known that Archaea contain glycan-bearing phosphodolichols and, like their eukaryal counterparts (Burda and Aebi, 1999), roles for these lipids in N-glycosylation were postulated (Mescher et al., 1976; Mescher and Strominger, 1978; Wieland et al., 1980; 1985; Lechner et al., 1985a,b; Sumper, 1987). Accordingly, dolichol phosphates charged with either the identical or slightly modified versions of the glycans decorating glycoproteins in Hbt. salinarum and M. fervidus were reported (cf. Lechner and Wieland, 1989). Since little or nothing is known of the N-glycosylation process in these species, much related to the involvement of dolichol phosphates in N-glycosylation in Archaea remained a matter of speculation. The recent identification of a series of agl genes involved in the N-glycosylation of the Hfx. volcanii S-layer glycoprotein (for review, see Yurist-Doutsch et al., 2008a; Calo et al., 2010), however, now makes it possible to address the precise role of dolichol phosphates as putative glycan carriers in the archaeal version of this post-translational modification.
Kuntz et al. (1997) first reported the presence of C55 and C60 dolichol phosphates saturated at both the α and the ω positions in Hfx. volcanii, including glycan-modified species. The Hfx. volcanii dolichol phosphate pool was reportedly modified by mannosyl-galactosyl groups, and to a lesser extent, by sulphated or phosphorylated dihexosyl moieties and by a tetrasaccharide that includes mannose, galactose and rhamnose, sugars not detected as components of the N-linked glycans reported to decorate the S-layer glycoprotein at the time (Sumper et al., 1990; Mengele and Sumper, 1992). However, given the revision of the originally reported composition of the glycan N-linked to the Hfx. volcanii S-layer glycoprotein from a string of linear glucose residues (as well as a glucose-, idose- and galactose-containing polysaccharide) (Sumper et al., 1990; Mengele and Sumper, 1992) to a pentasaccharide comprising two hexoses, two hexuronic acids and a methyl ester of hexuronic acid (Abu-Qarn et al., 2007; Magidovich et al., 2010), the present study revisited the composition of glycans decorating dolichol phosphates in Hfx. volcanii in an attempt to link these glycan-charged lipids to the N-glycosylation process.
In recent work from our group (Kaminski et al., 2010), AglJ was shown to add the first hexose subunit of the N-linked S-layer glycoprotein pentasaccharide to a dolichol phosphate carrier. In the present report, it was revealed that the next three subunits of the pentasaccharide are sequentially added to that AglJ-generated monosaccharide-charged carrier, through the respective actions of AglG, AglI and AglE. Since no hexuronic acid-charged phosphodolichol was detected, it seems that pentasaccharide subunits two and three are added from soluble, activated species. Likewise, the methyl ester of hexuronic acid found at position four of the N-linked pentasaccharide is added to the existing trisaccharide-charged phosphodolichol from a soluble methylated hexuronic acid species, since neither a dolichol phosphate modified with only a methyl ester of hexuronic acid nor a tetrasaccharide-charged phosphodolichol bearing a hexuronic acid at position four was detected. These observations, moreover, offer support to the earlier assignment of the nucleoside-hexose dehydrogenase, AglM, shown to catalyse the in vitro conversion of UDP-glucose to UDP-glucuronic acid and likely involved in the biogenesis of pentasaccharide subunits two, three and four, and of AglP, the SAM-dependent methyltransferase responsible for modifying the fourth subunit of the pentasaccharide subunit, as being soluble enzymes (Magidovich et al., 2010; Yurist-Doutsch et al., 2010).
In contrast to the sequential assembly of the first four pentasaccharide subunits onto a common dolichol phosphate, the fifth subunit of the pentasaccharide, mannose, was detected on its own distinct lipid carrier. The finding that AglD, involved in the addition of the fifth pentasaccharide subunit, acts in a manner seemingly independent of the other Agl proteins involved in generating the oligosaccharide decorating the Hfx. volcanii S-layer glycoprotein is not unexpected, given that aglD is the only gene not found in the agl gene island present in the Hfx. volcanii genome (Yurist-Doutsch and Eichler, 2009). Thus, the observation that the same sugar subunits are found on both dolichol phosphate carriers and the S-layer glycoprotein, four of which are sequentially added to the lipid carrier in the same order as found on the modified protein, together with the fact that deletion of agl genes compromised dolichol phosphate glycosylation in a manner reminiscent of the effects of the same gene deletions on Hfx. volcanii S-layer glycoprotein N-glycosylation (Abu-Qarn et al., 2007; 2008; Yurist-Doutsch et al., 2008b), directly links dolichol phosphates, acting as mono- and oligosaccharide carriers, to the archaeal N-glycosylation process (Fig. 5).
The confirmed involvement of archaeal dolichol phosphate sugar carriers in Hfx. volcanii N-glycosylation provides novel insight into the mechanism of this post-translational modification. Earlier work had shown that Agl proteins involved in adding sugars found on the pentasaccharide N-linked to the Hfx. volcanii S-layer glycoprotein are membrane proteins oriented towards the cytoplasm, pointing to dolichol phosphate sugar charging as occurring within the confines of the cell (Plavner and Eichler, 2008). Hence, as no pentasaccharide-charged phosphodolichol species could be detected, it is possible that transfer of pentasaccharide subunit five occurs directly onto the S-layer glycoprotein-linked tetrasaccharide. Indeed, tetrasaccharide-modified S-layer glycoprotein-derived peptides have been observed (Abu-Qarn et al., 2007; Magidovich et al., 2010). Alternatively, transfer of the fifth sugar subunit from its own dolichol phosphate carrier to the lipid-linked tetrasaccharide and subsequent transfer to the S-layer glycoprotein may occur too rapidly to be detected here. This explanation for our inability to detect a Hfx. volcanii pentasaccharide-modified phosphodolichol species is unlikely, since such an entity was readily observed upon examination of the dolichol phosphate pool of another halophilic archaea originating from the Dead Sea, namely Haloarcula marismortui (Z. Guan, D. Calo and J. Eichler, in preparation). To determine whether the fifth and final subunit of the pentasaccharide is added to the tetrasaccharide-charged phosphodolichol to yield a potentially short-lived pentasaccharide-charged lipid carrier or directly to the tetrasaccharide-modified S-layer glycoprotein, additional biochemical studies are required. It is, however, clear that methylation of the fourth pentasaccharide subunit is important for addition of pentasaccharide subunit five, since no N-linked pentasaccharide is detected in the ΔaglP mutant, where methylation of the fourth pentasaccharide subunit fails to occur (Magidovich et al., 2010). On the other hand, the actions of AglP are not essential for modification of the S-layer glycoprotein by the tetrasaccharide formed in the absence of this methyltransferase.
In conclusion, despite considerable progress made in understanding archaeal N-glycosylation in recent years (Yurist-Doutsch et al., 2008a; Calo et al., 2010), many questions still remain unanswered. For instance, why does N-glycosylation in some Archaea, such as Hfx. volcanii, rely on dolichol phosphate while other species rely on dolichol pyrophosphate or both, as in the case of Hbt. salinarum (Lechner and Wieland, 1989)? What is/are the flippase(s) involved in N-glycosylation in Hfx. volcanii? Finally, do the glycan-modified phosphodolichol species originally reported by Kuntz et al. (1997) participate in any Hfx. volcanii post-translation modification? Continued examination of the Hfx. volcanii N-glycosylation pathway will likely provide answers to these and other outstanding questions.
The total lipid contents of the Hfx. volcanii parent strain and of Hfx. volcaniiΔaglG, ΔaglI, ΔaglE and ΔaglD cells were extracted as follows. Cells were harvested (8000 g, 30 min, 4°C) and frozen at −20°C until extraction was performed. At that point, the pelleted cells (15 g) were thawed, resuspended in 20 ml of double-distilled water (DDW) and DNase (1.7 µg ml−1; Sigma, St. Louis, MO) and stirred overnight at room temperature. Methanol and chloroform were added to the cell extract to yield a methanol : chloroform : cell extract ratio of 2:1:0.8. After stirring for 24 h at room temperature, the mixture was centrifuged (1075 g, 30 min, 4°C). The clarified supernatants were collected, combined and filtered through glass wool. Chloroform and DDW were added to the filtrate to yield a chloroform : DDW : filtrate ratio of 1:1:3.8, in a separating funnel. After separation, the lower clear organic phase, containing the total lipid extract, was collected into a round-bottom flask and evaporated in a rotary evaporator at 35°C. For analysis of the dolichol phosphate-derived species, the total lipid extracts were subjected to normal-phase LC/MS analysis without pre-fractionation.
LC/MS and MS/MS
Normal-phase LC-ESI/MS of lipids was performed using an Agilent 1200 Quaternary LC system coupled to a QSTAR XL quadrupole time-of-flight tandem mass spectrometer (Applied Biosystems, Foster City, CA). An Ascentis Si HPLC column (5 µm, 25 cm × 2.1 mm) was used. Mobile phase A consisted of chloroform/methanol/aqueous ammonium hydroxide (800:195:5, v/v/v). Mobile phase B consisted of chloroform/methanol/water/aqueous ammonium hydroxide (600:340:50:5, v/v/v/v). Mobile phase C consisted of chloroform/methanol/water/aqueous ammonium hydroxide (450:450:95:5, v/v/v/v). The elution programme consisted of the following: 100% mobile phase A was held isocratically for 2 min and then linearly increased to 100% mobile phase B over 14 min and held at 100% B for 11 min. The LC gradient was then changed to 100% mobile phase C over 3 min and held at 100% C for 3 min, and finally returned to 100% A over 0.5 min and held at 100% A for 5 min. The total LC flow rate was 300 µl min−1. The post-column splitter diverted ∼10% of the LC flow to the ESI source of the Q-Star XL mass spectrometer, with MS settings as follows: IS = −4500 V, CUR = 20 psi, GS1 = 20 psi, DP = −55 V and FP = −150 V. For MS/MS, collision-induced dissociation (CID) was performed with collision energy ranging from 40 V to 70 V (laboratory frame of energy) and with nitrogen as the collision gas. Data acquisition and analysis were performed using the instrument's Analyst QS software.
J.E. is supported by the Israel Science Foundation (Grant 30/07). The mass spectrometry facility in the Department of Biochemistry of the Duke University Medical Center and Z.G. are supported by the LIPID MAPS Large Scale Collaborative Grant No. GM-069338 from NIH. L.K. is the recipient of a Negev-Zin Associates Scholarship.