This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Plavner, N.
Right arrow Articles by Eichler, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Plavner, N.
Right arrow Articles by Eichler, J.

 Previous Article  |  Next Article 

Journal of Bacteriology, December 2008, p. 8045-8052, Vol. 190, No. 24
0021-9193/08/$08.00+0     doi:10.1128/JB.01200-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Defining the Topology of the N-Glycosylation Pathway in the Halophilic Archaeon Haloferax volcanii{triangledown}

Noa Plavner and Jerry Eichler*

Department of Life Sciences, Ben Gurion University, Beersheva 84105, Israel

Received 26 August 2008/ Accepted 2 October 2008


arrow
ABSTRACT
 
In Eukarya, N glycosylation involves the actions of enzymes working on both faces of the endoplasmic reticulum membrane. The steps of bacterial N glycosylation, in contrast, transpire essentially on the cytoplasmic side of the plasma membrane, with only transfer of the assembled glycan to the target protein occurring on the external surface of the cell. For Archaea, virtually nothing is known about the topology of enzymes involved in assembling those glycans that are subsequently N linked to target proteins on the external surface of the cell. To remedy this situation, subcellular localization and topology predictive algorithms, protease accessibility, and immunoblotting, together with cysteine modification following site-directed mutagenesis, were enlisted to define the topology of Haloferax volcanii proteins experimentally proven to participate in the N-glycosylation process. AglJ and AglD, involved in the earliest and latest stages, respectively, of assembly of the pentasaccharide decorating the H. volcanii S-layer glycoprotein, were shown to present their soluble N-terminal domain, likely containing the putative catalytic site of each enzyme, to the cytosol. The same holds true for Alg5-B, Dpm1-A, and Mpg1-D, proteins putatively involved in this posttranslational event. The results thus point to the assembly of the pentasaccharide linked to certain Asn residues of the H. volcanii S-layer glycoprotein as occurring within the cell.


arrow
INTRODUCTION
 
It has long been recognized that Archaea originating from a variety of diverse environments are able to N glycosylate numerous proteins (see reference 7 and references cited therein). Indeed, the first noneukaryal N-glycosylated protein described was derived from an archaeal source, i.e., the S-layer glycoprotein of the halophilic archaeon Halobacterium salinarum (15). Still, little is known about the pathway responsible for the covalent attachment of glycan moieties to certain Asn residues in archaeal glycoproteins. Of late, however, steps aimed at remedying this situation have been taken. In both the halophilic archaeon Haloferax volcanii and the methanogen Methanococcus voltae, genes involved in the N-glycosylation process have been defined (1-3, 5, 21, 28). Additional insight into the archaeal version of this posttranslational modification has come with the development of an in vitro assay to test Pyrococcus furiosus Stt3/AglB activity (11) and the recent solution of the three-dimensional structure of the C-terminal soluble domain of the protein (9), thought to comprise the sole unit of the archaeal oligosaccharide transferase (1, 5).

Thus, while progress is being made in identifying the different enzymes responsible for the various steps leading to archaeal N glycosylation, little is known about the topology of such reactions. Archaeal N glycosylation is thought to share steps with the parallel processes in Bacteria and Eukarya (1, 4, 9, 13, 27). In Campylobacter jejuni, the sole bacterial species for which the N-glycosylation pathway has been delineated, a heptasaccharide is assembled from soluble nucleotide-activated sugars onto a cytoplasmically oriented lipid carrier present in the plasma membrane. The oligosaccharide-charged lipid is then "flipped" across the membrane to face the cell exterior, where the glycan moiety is transferred to certain Asn residues in target proteins by the actions of PglB, the oligosaccharide transferase in this species (for reviews, see references 23 and 25). In Eukarya, such as Saccharomyces cerevisiae, N glycosylation also begins with the assembly of cytoplasmically located, soluble nucleotide-activated sugars into a heptasaccharide chain of defined composition on a lipid carrier associated with the cytoplasmic face of the endoplasmic reticulum (ER) membrane. Once assembled, the lipid-charged heptasaccharide is reoriented, such that the oligosaccharide now faces the ER lumen. Next, an additional seven sugar subunits, each derived from its own individual lipid carrier, charged on the cytoplasmic face of the ER membrane and flipped to face the ER lumen, are added to yield a 14-member oligosaccharide. This oligosaccharide is now transferred, en bloc, to certain Asn residues of a nascent polypeptide translocating into the ER, via the actions of the multimeric oligosaccharide transferase complex (for reviews, see references 4 and 8).

In the case of archaeal N glycosylation, comparatively less is known about the topology of the process. Based on studies following the modification of cell-impermeable peptide reporters of N glycosylation (14) or through the use of bacitracin, an antibiotic that interferes with the regeneration of the dolichol pyrophosphate oligosaccharide carrier presumably used in archaeal N glycosylation (26), the transfer of lipid-linked oligosaccharides to target proteins was assigned to the external surface of the archaeal plasma membrane. In contrast, the biosynthesis of nucleotide-activated sugars in Archaea, likely recruited for N glycosylation, has been shown to occur in the cytoplasm (17, 18). Thus, apart from the first and last phases of the process, virtually nothing is known about the topology of archaeal N glycosylation. Accordingly, the topologies of both proven (i.e., AglD and AglJ) and putative (i.e., Alg5-B, Dpm1-A, and Mpg1-D) components of the H. volcanii N-glycosylation pathway were considered experimentally in this study.


arrow
MATERIALS AND METHODS
 
Materials. Cellulose, novobiocin, and phenylmethylsulfonyl fluoride were obtained from Sigma (St. Louis MO). Proteinase K came from Boehringer (Mannheim, Germany). Yeast extract came from Pronadisa (Madrid, Spain), while tryptone came from USB (Cleveland, OH). 4-Acetoamido-4-maleimidylstilbene-2,2-disulfonic disodium salt (AMS) came from Invitrogen (Carlsbad, CA), while 14C-labeled N-ethylmaleimide (NEM) (20 to 40 mCi/mmol) came from Perkin-Elmer (Boston, MA). A 3.3 mM working solution of [14C]NEM was prepared by dilution into NEM prepared in ethanol.

Culture conditions. H. volcanii cells were grown in rich medium containing 3.4 M NaCl, 0.15 M MgSO4·7H2O, 1 mM MnCl2, 4 mM KCl, 3 mM CaCl2, 0.3% (wt/vol) yeast extract, 0.5% (wt/vol) tryptone, and 50 mM Tris-HCl (pH 7.2) at 40°C (16).

In silico topology analysis. To define protein topology, the HMMTOP (http://www.enzim.hu/hmmtop/), SOSUI (http://bp.nuap.nagoya-u.ac.jp/sosui/), TMHMM (http://www.cbs.dtu.dk/services/TMHMM-2.0/), TMpred (http://www.ch.embnet.org/software/TMPRED_form.html), and TopPred (http://bioweb.pasteur.fr/seqanal/interfaces/toppred.html) topology prediction programs, as found at www.expasy.ch, were consulted. The H. volcanii proteins considered by these algorithms included those previously identified as homologues of eukaryal or bacterial N-glycosylation proteins (1). For reasons described below, only those proteins predicted by the algorithms to possess an N terminus sequestered within the cytosol were considered for further analysis in the study.

Plasmid construction. The aglD gene was amplified from H. volcanii strain WR536 (H53) genomic DNA using the primers listed in Table 1, which were designed to introduce NdeI and KpnI restriction sites on the 5' and 3' ends, respectively, of the aglD coding region, and ligated into the pGemT-Easy vector (Promega). The aglD gene was then excised upon digestion with NdeI and KpnI and inserted into the pWL-CBD vector (10), which also predigested with the same restriction enzymes, resulting in a plasmid encoding the Clostridium thermocellum cellulose-binding domain (CBD) (GenBank accession number 2554722) fused to the 5' end of the AglD-encoding gene. Plasmids encoding CBD-AglJ (aglJ GenBank accession number, FM210664), CBD-Alg5-B, CBD-Dpm1-A, and CBD-Mpg1-D were similarly generated, using the appropriate primers.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Primers used in this study

Immunoblotting. Immunoblotting was performed as described previously (12), using polyclonal antibodies raised against H. volcanii SRP54 (1:1,000) (24) or the C. thermocellum CBD (obtained from Ed Bayer, Weizmann Institute of Science) (1:10,000). Antibody binding was detected using goat anti-rabbit horseradish peroxidase-conjugated antibodies (1:4,000; Bio-Rad, Hercules, CA) and an ECL enhanced chemiluminescence kit (Amersham, Buckingham, United Kingdom).

Determination of protease accessibility. To assess the protease accessibility of the CBD-tagged proteins, 1-ml aliquots of the transformed cells were challenged with proteinase K (1 mg/ml, 55°C). Aliquots were removed at time zero and at subsequent 30- to 60-min intervals and transferred to ice. The samples were then centrifuged (3,000 x g, 3 min, 4°C) and resuspended in 1 ml of lysis buffer (1% [vol/vol] Triton X-100, 1.8 M NaCl, 50 mM Tris-HCl, pH 7.2) containing 1 mM phenylmethylsulfonyl fluoride. The mixtures were rocked (10 min, room temperature [RT]), after which 50 µl of a 10% (wt/vol) solution of cellulose beads was added. After a 20-min rocking at RT, the suspension was centrifuged (3,000 x g, 3 min, RT), the supernatant was discarded, and the cellulose pellet was washed with 2 M NaCl-50 mM Tris-HCl, pH 7.2. This washing procedure was repeated twice. After the final wash, the cellulose beads were centrifuged (5,000 x g, 3 min, RT), the supernatant was removed, and the cellulose pellet was resuspended in 40 µl sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer. The samples were then boiled for 5 min and centrifuged (5,000 x g, 5 min) to release any cellulose-bound proteins, which were then examined by 10% SDS-PAGE and immunoblotting using anti-CBD antibodies.

Site-directed mutagenesis. To generate single- or reduced-cysteine-containing versions of AglD and Dpm1-A, site-directed mutagenesis was performed using the QuikChange (Stratagene) protocol according to the manufacturer's instructions, together with those plasmids encoding CBD-tagged versions of AglD or Dpm1-A as templates. The oligonucleotide primers used to introduce the various mutations are listed in Table 1. The introduction of mutations was confirmed by sequencing.

Cysteine modification. Cysteine modification was achieved using two different cysteine-reactive reagents, namely, membrane-permeable [14C]NEM and membrane-impermeable AMS. H. volcanii cells transformed to express the various single- or reduced-cysteine-containing versions of proven or putative N-glycosylation pathway proteins fused to CBD were challenged with [14C]NEM (15 µl of the working solution, 20 min, 30°C, with rocking), in some cases followed by AMS (5 mM final concentration, 30 min, 37°C). Alternatively, cells were first challenged with AMS and then incubated with [14C]NEM. In other cases, the cells were incubated with 1% Triton X-100 (5 min, RT) prior to incubation with the cysteine-reactive reagents. In all cases, cysteine modification was terminated by addition of dithiothreitol to a final concentration of 50 mM (10 min, 30°C). The CBD-based fusion proteins were then cellulose purified, separated by 10% SDS-PAGE, and visualized by fluorography.


arrow
RESULTS AND DISCUSSION
 
Predicted topology of proteins putatively involved or proven to participate in H. volcanii N glycosylation. As a first step in delineating the topology of components of the H. volcanii N-glycosylation machinery, both sequences of proteins putatively involved in the process (1) and sequences of proteins experimentally verified as participating in this posttranslational modification (2) were analyzed with predictive topology software. Only those proteins predicted by the algorithms to possess an N terminus sequestered within the cytosol were considered for further study (the reasons for this selection are expanded upon below). This list included Alg5-B (HVO_0704; http://archaea.ucsc.edu), Dpm1-A (HVO_2061), and Mpg1-D (HVO_A0586), all of which are H. volcanii homologues of eukaryal proteins known to participate in N glycosylation (1), as well as AglD (HVO_0798), which was previously verified as contributing to the N glycosylation of the H. volcanii S-layer glycoprotein (2). AglJ (HVO_1517), which was recently observed to participate in this posttranslational modification (M. Abu-Qarn, P. G. Hitchen, F. Battaglia, A. Dell, and J. Eichler, unpublished data), was also included. In each case, at least three of the topology prediction programs assigned the N terminus as being intracellular.

The various topology prediction programs were next employed to define the subcellular localization of the five proteins considered. Accordingly, Alg5-B was predicted to be a cytoplasmic protein by all of the programs consulted, while Mpg1-D was predicted to lie within the cytoplasm by three of the five topology programs. The other two programs predicted Mpg1-D to be a singly spanning membrane protein, with its N terminus found either inside or outside the cell. In contrast, AglD, AglJ, and Dpm1-A were all designated as membrane proteins, spanning the membrane at least twice, by all five algorithms.

Using proteinase K accessibility to define protein topology. To work toward experimentally verifying the various computer-based predictions, chimeras were generated at the DNA level between the H. volcanii protein of interest and the C. thermocellum CBD, with the CBD moiety being linked to the N terminus of each H. volcanii protein. The presence of the CBD moiety allows both for cellulose-based affinity purification in a reaction compatible with the molar salt concentrations in which H. volcanii exists and for identification of the chimeras using anti-CBD antibodies (10, 19). Moreover, by having the CBD entity attached to the presumed cytoplasmic N-terminal residue of the H. volcanii protein in question, one would expect that any possible complications arising from having to deliver the CBD moiety across the plasma membrane would be avoided. As revealed in Fig. 1, the CBD fusion proteins were all expressed by H. volcanii cells transformed with plasmids encoding the various chimeras, as reflected by cellulose-based purification from cell lysates followed by Coomassie blue staining (Fig. 1A) or immunoblotting using anti-CBD antibodies (Fig. 1B). Based on the deduced amino acid composition of the H. volcanii sequences in question, combined with the contribution of the attached 17-kDa CBD moiety, protein species with apparent molecular masses of 85, 50, 45, 61, and 61 kDa were predicted to be cellulose purified from extracts of cells transformed to express CBD-AglD, CBD-AglJ, CBD-Alg5-B, CBD-Dpm1-A, and CBD-Mpg1-D, respectively. Analysis of the Coomassie blue-stained gels and immunoblots indeed revealed bands of the expected size in the cases of CBD-AglD and CBD-Alg5-B. In the case of CBD-Alg5-B, however, some breakdown of the chimera was detected. In contrast, CBD-AglJ migrated as a band with an apparent molecular mass of 61 kDa, CBD-Dpm1A as a band with an apparent molecular mass of 72 kDa, and Mpg1-D as a band with an apparent molecular mass of 85 kDa, i.e., some 10 to 25 kDa heavier than predicted based on sequence considerations. The slower migration of halophilic proteins in SDS-PAGE, often leading to an overestimation of their molecular mass, is, however, a well-known phenomenon (see, for example, references 20 and 22).


Figure 1
View larger version (35K):
[in this window]
[in a new window]

  		FIG. 1. Cellulose-based purification of CBD-tagged H. volcanii enzymes proven to be or putatively involved in N glycosylation. (A) H. volcanii cells transformed to express CBD-AglD, -AlgJ, -Alg5-B, -Dpm1-A, or -Mpg1-D were lysed and the CBD-tagged proteins purified with cellulose. Following separation by 10% SDS-PAGE, the proteins were visualized by Coomassie blue staining. (B) Purified CBD-AglD, -AglJ, -Alg5-B, -Dpm1-A, and -Mpg1-D are recognized by anti-CBD antibodies in an immunoblotting protocol. Molecular mass markers (in kDa) are denoted on the right of each panel (A, with values shown only for the first panel) or of the set of panels (B).

To begin defining the topology of those chimeras containing the different H. volcanii proteins under consideration, cells expressing the various CBD-linked proteins were challenged with proteinase K under such conditions that only externally exposed domains of the protein were accessible for proteolytic digestion. To define such conditions, cells of the H. volcanii WR341 background strain were exposed to 1 mg/ml proteinase K for up to 3 h, with aliquots being removed every 30 to 60 min. The levels of markers for the external cell surface and the cytoplasm were then analyzed by SDS-PAGE and Coomassie blue staining or immunoblotting. Under such conditions, the externally facing S-layer glycoprotein, as revealed by Coomassie blue staining, was readily digested after 1 h. In contrast, the level of SRP54, a component of the H. volcanii signal recognition particle (24) that served as a cytoplasmic marker, was not affected by the added protease, as detected in immunoblotting using anti-SRP54 antibodies (Fig. 2A).


Figure 2
View larger version (34K):
[in this window]
[in a new window]

  		FIG. 2. Proteinase K digestion of CBD-incorporating chimeras. (A) In control experiments, H. volcanii cells were challenged with proteinase K (1 mg/ml, 55°C) for up to 3 h, and the amounts of surviving SRP54, a cytosolic marker, and of the S-layer glycoprotein (SLG), a membrane protein marker, were assessed by immunoblotting using anti-SRP54 antibodies and Coomassie blue staining, respectively, following separation by 10% SDS-PAGE. (B) H. volcanii cells transformed to express CBD-AglD, -AglJ, -Alg5-B, -Dpm1-A, or -Mpg1-D were similarly challenged with proteinase K, and the profile of CBD-tagged proteins over the period of digestion was obtained by immunoblotting using anti-CBD antibodies. The positions of molecular mass markers (in kDa) are shown on the right of each panel. (C) In control experiments, H. volcanii cells transformed to express CBD-AglD, -AglJ, -Alg5-B, -Dpm1-A, or -Mpg1-D were challenged with proteinase K (pk), in some cases following incubation with 1% Triton X-100 (Tx100), for 30 min.

H. volcanii cells expressing the various CBD-based chimeras were similarly challenged with proteinase K and subsequently probed in immunoblots using anti-CBD serum (Fig. 2B) after cellulose-based capture. In all cases, the CBD moiety attached to the H. volcanii protein of interest was protected from proteolysis (as reflected by maintained antibody labeling), pointing to the cytoplasmic localization of the N terminus of each H. volcanii protein considered. In the case of CBD-AglJ, CBD-Alg5-B, and CBD-Mpg1-D, the entire chimera was fully protected from the effects of proteinase K. In contrast, the apparent molecular masses of CBD-AglD and CBD-Dpm1-A were reduced following proteolytic treatment. In the case of CBD-AglD, the initial 85-kDa protein was transformed into a 71-kDa species over the course of the 3-hour protease challenge, while with CBD-Dpm1-A, the starting 61-kDa chimera was partially digested to yield a 56-kDa protein. Finally, to confirm that the proteinase K digestion patterns obtained did not reflect inherent protease resistance of the CBD-tagged H. volcanii proteins, the proteolysis was repeated, but this time after the transformed cells had been first incubated with 1% Triton X-100 so as to compromise membrane intactness. The results obtained reveal that proteinase K was able to fully digest each chimera if given free access (Fig. 2C).

The protection of CBD-linked Alg5-B from proteolysis lends support to the bioinformatics-based consensus prediction of Alg5-B being a cytosolic protein. Similarly, the failure of added protease to digest Mpg1-D or its N-terminally fused CBD group is in agreement with those algorithms that predicted the protein to also reside in the cytosol. It is thus not likely that Mpg1-D is a membrane protein (as assigned by two of the topology prediction programs considered), with an internal N terminus and an extracellular domain (residues 317 to 411). The observation that proteinase K treatment partially digested CBD-AglD and CBD-Dpm1A while apparently not compromising CBD integrity, as judged by the preserved ability of anti-CBD antibodies to recognize this moiety, also confirms the bioinformatics-based predictions assigning the N terminus of each of these membrane proteins to the cell interior. The observed shift in apparent molecular weight of CBD-AglD and CBD-Dpm1A following protease treatment is thus likely due to digestion of an externally exposed segment of the H. volcanii component of each chimera. Finally, the bioinformatics-based predictive tools assigned AglJ as a membrane protein with its N terminus either oriented toward the cell interior or exposed to the exterior. The failure of proteinase K to digest the CBD moiety of CBD-AglJ thus points to the N terminus of AglJ as residing within the cytosol. The failure of proteinase K to digest the extracellular segment of AglJ (residues 251 to 265) may be due to steric considerations.

Using cysteine-modifying reagents to define protein topology. To either confirm or distinguish between the various scenarios proposed above, as the case may be, the topologies of the different CBD-tagged H. volcanii N-glycosylation pathway candidates or participants were next considered by an alternative approach. In these experiments, the accessibility of cysteine residues for interaction with the membrane-permeating cysteine-modifying reagent [14C]NEM or the non-membrane-permeating cysteine-modifying reagent AMS was addressed. Since AMS reacts only with externally exposed cysteine residues, pretreatment with AMS would prevent any subsequent labeling of surface-exposed cysteine residues of the CBD-based chimeras by [14C]NEM. In contrast, [14C]NEM labeling of those cysteine residues found in the cytoplasmic portions of the CBD-based chimeras would not be affected by AMS pretreatment. As such, the ability of preincubation with AMS to preclude any subsequent [14C]NEM radiolabeling of single- or reduced-cysteine-containing versions of the CBD-fused H. volcanii proteins offers an elegant approach to delineate protein topology through the identification of those cysteine-containing regions of the protein exposed to the exterior. [14C]NEM radiolabeling of these CBD-based chimeras can be visualized, following detergent solubilization in the presence of excess dithiothreitol (serving as a quenching agent), by cellulose-based purification, SDS-PAGE, and fluorography.

Given our inability to generate a cysteine-lacking version of CBD, preliminary experiments were performed to determine whether the single cysteine residue of the CBD moiety (Cys-55) was accessible for labeling by [14C]NEM. The results show that the CBD that was cellulose purified from [14C]NEM-treated H. volcanii cells transformed to express this moiety, although present at a level that can be visualized by Coomassie blue staining (Fig. 3A, upper panel, lanes 2 to 6), was not modified by the radiolabeling reagent (lower panel, lane 2). Neither the presence of AMS nor incubation with 1% Triton X-100 altered this situation (Fig. 3A, lower panel, lanes 3 to 6). Control experiments, however, confirmed the ability of [14C]NEM, at the concentration and conditions used with the CBD-expressing cells as described above, to readily label other, native H. volcanii proteins (Fig. 3A, upper and lower panels, lane 1). These results confirm that any [14C]NEM labeling of the CBD-based chimeras in the transformed H. volcanii cells, under the conditions employed, would result from interaction of the modifying reagent with those cysteine residues present in the H. volcanii proteins under consideration rather than from binding of [14C]NEM to CBD Cys-55.


Figure 3
View larger version (50K):
[in this window]
[in a new window]

  		FIG. 3. Modification of cysteines as a tool for determining topology. (A) Incubation of H. volcanii cells (Coomassie blue panel, lane 1) with [14C]NEM resulted in the labeling of a large number of proteins (fluorography panel, lane 1). CBD from transformed H. volcanii cells was incubated with [14C]NEM, AMS, and/or 1% Triton X-100 in various combinations (fluorography panel, lanes 2 to 6). The cytoplasmically localized CBD moiety was not labeled by [14C]NEM (fluorography panel, lane 2), just as incubation with AMS either before (lane 3) or after (lane 4) [14C]NEM treatment did not result in CBD radiolabeling. Similarly, no radiolabeling was achieved when the cells were pretreated with 1% Triton X-100 and then challenged with AMS either before (lane 5) or after (lane 6) [14C]NEM treatment. Molecular mass markers on the left (170, 130, 100, 72, 55, 40, and 33 kDa) apply to lane 1, while those on the right (170, 130, 100, 72, 55, 40, 33, 24, 17, and 11 kDa) apply to lanes 2 to 6. (B) As described in Materials and Methods, H. volcanii cells transformed to express CBD-AglD, -AglJ, -Alg5-B, or -Mpg1-D, as well CBD-AglD(C61A), CBD-AglD(C176A), CBD-Dpm1-A(C362A), and CBD-Dpm1-A(C62A,C131,362A) mutants, were incubated with [14C]NEM alone (lane 1), or before (lane 3) or after (lane 2) treatment with AMS. In other experiments, the cells were incubated with 1% Triton X-100 (Tx100) prior to incubation with AMS and then [14C]NEM (lane 4) or vice versa (lane 5). Cysteine-mutated proteins were generated by site-directed mutagenesis, as described in Materials and Methods.

With this in mind, cysteine-based modification of CBD-tagged versions of the H. volcanii proteins of interest was next performed. In the case of Alg5-B and Mpg1-D, each containing a single cysteine residue (Cys-155 and Cys-197, respectively), labeling by [14C]NEM was observed (Fig. 3B, CBD-Alg5-B and CBD-Mpg1-D panels, lane 1). Prior treatment with AMS had no effect on the radiolabeling of either protein (lane 2). The same held true if AMS was added after incubation with [14C]NEM (lane 3). If, however, membrane integrity was first compromised with 1% Triton X-100, pretreatment with AMS prevented subsequent radiolabeling by [14C]NEM (lane 4). In contrast, if the detergent-lysed cells were first challenged with [14C]NEM and only then treated with AMS, radiolabeling was still realized (lane 5). These results thus show the cysteine residues of Alg5-B and Mpg1-D to be localized within the confines of the cell.

The AglD portion of CBD-AglD, in contrast, contains two cysteine residues (Cys-61 and Cys-176). However, since both are predicted to reside in an N-terminal segment of the protein assigned to the cytosol by the subcellular localization algorithms consulted above, the protein was considered as if it contained a single cysteine residue. With this in mind, H. volcanii cells transformed to express CBD-AglD were challenged with [14C]NEM either before or after AMS treatment. A pattern of radiolabeling similar to that observed with CBD-Alg5-B or CBD-Mpg1-D was also obtained with CBD-AglD. Specifically, preincubation with AMS did not prevent subsequent [14C]NEM labeling (Fig. 3B, CBD-AglD panel, lane 2). If the cells were, however, first solubilized with 1% Triton X-100, AMS succeeded in blocking any ensuing binding of [14C]NEM and radiolabeling of the protein (lane 4). These observations therefore point to the pair of cysteine residues in AglD, and hence the N terminus, as being sequestered within the confines of the cytoplasm.

To experimentally confirm the assumption made above, namely, that both AglD cysteines indeed present the same topological orientation, site-directed mutagenesis was employed to selectively transform either Cys-61 or Cys-176 into alanines. H. volcanii cells transformed to express CBD-AglD(C61A) or CBD-AglD(C176A) were challenged with AMS and [14C]NEM, as described above. As with the native AglD protein, preliminary incubation with AMS could not prevent later binding by [14C]NEM to either CBD-AglD(C61A) or CBD-AglD(C176A) unless the cells were first solubilized with 1% Triton X-100 [Fig. 3B, CBD-AglD(C61A) and CBD-AglD(C176A) panels, lanes 2 and 4, respectively]. These observations thus confirm the assignment of the AglD N terminus to the cytosol.

Like AglD, the AglJ component of CBD-AglJ also includes two cysteine residues (Cys-8 and Cys-181). The various topology prediction programs listed above assign both AglJ cysteine residues to an N-terminal segment of the protein exposed either to the cytosol or to the exterior. Given that AglD could be considered a single-cysteine-containing protein, the same initial assumption was made in the case of AglJ, presenting cysteine residues at positions in the polypeptide chain comparable to those present in AglD. This supposition was justified upon cysteine modification, which revealed that neither AglJ cysteine residue was available for AMS-mediated protection against [14C]NEM binding unless membrane integrity was first compromised by detergent treatment (Fig. 3B, CBD-AglJ panel, lanes 2 and 4, respectively). As such, these experiments assign both AglJ cysteines and the N terminus of the protein to the cytoplasm.

The distribution of cysteine residues in the Dpm1-A portion of CBD-Dpm1-A presents a more complicated picture. Here, three of the four cysteines of the protein (Cys-62, Cys-131, and Cys-211), along with the N terminus, are assigned to the cytoplasm by the various topology prediction algorithms consulted, with the first transmembrane domain predicted to begin at residue 279. In contrast, Dpm1-A Cys-362 is predicted to lie within a membrane-spanning domain of the protein. With these predictions in mind, H. volcanii cells were transformed to express CBD-Dpm1-A(C362A) and incubated with AMS and [14C]NEM. As reflected in the CBD-Dpm1-A(C362A) panel of Fig. 3B, pretreatment with AMS was unable to prevent subsequent radiolabeling (lane 2) unless cell integrity was first compromised by detergent treatment (lane 4). These results, therefore, point to the N-terminally linked CBD domain as well as at least the first 211 residues, including Cys-62, Cys-131, and Cys-211, as being cytoplasmic in orientation. The observation that AMS likewise failed to prevent [14C]NEM labeling of CBD-Dpm1-A(C62S,C131,362A), a version of CBD-Dpm1-A that includes a Dpm1-A mutant containing only a single cysteine residue (i.e., C211), offers additional support for the topology described above [Fig. 3B, CBD-Dpm1-A(C62S,C131,362A)].

Conclusions. Although protein N glycosylation occurs across evolution, much remains unknown of the archaeal version of this universal posttranslational modification. Indeed, many basic aspects of archaeal N glycosylation have yet to be examined, including the subcellular localization of the process. Accordingly, the present research has addressed the topology of proteins known to be involved in N glycosylation in H. volcanii, as well as other proteins suspected but not yet verified as participating in this pathway. Specifically, AglD has been experimentally demonstrated to participate in the addition of the fifth monosaccharide to the pentasaccharide decorating the H. volcanii S-layer glycoprotein (2), while AglJ is involved in adding either the first or second monosaccharide to the pentaglycan (Abu-Qarn et al., unpublished data). Gene deletion, together with mass spectrometry, has also provided preliminary evidence for the involvement of Alg5-B in assembly of this pentasaccharide (Abu-Qarn et al., unpublished data). In contrast, the involvement of Dpm1-A and Mpg1-D in H. volcanii N glycosylation remains to be confirmed.

In the present study, three independent approaches were employed to define the topologies of the H. volcanii proteins being considered. Topologies were predicted using five different predictive algorithms, by assessing protease accessibility, and by a cysteine modification protocol. As summarized in Table 2 and Fig. 4, all three approaches assign both Alg5-B and Mpg1-D to the cytosol. In the case of Dpm1-A, all three approaches assign the N terminus to the cytoplasm, including the putative glycosyltransferase 2 superfamily domain (residues 47 to 210) of the protein. In addition, the results of the bioinformatics and protease accessibility studies describe Dpm1-A as an integral membrane protein, predicted to expose two small loops (residues 293 to 306 and 375 to 389) to the exterior. The latter loop is likely susceptible to the actions of proteinase K, releasing a 34- to 38-residue fragment, as reflected in the transformation of the 61-kDa chimera to an approximately 56-kDa species following challenge with the protease. All three approaches also agree that AglJ corresponds to a membrane protein with an internally oriented N terminus. Indeed, AglJ is thought to contain two transmembrane domains, with only a small 14-amino-acid stretch (residues 251 to 265) being exposed to the exterior. As such, the bulk of the protein is sequestered within the cytoplasm. As proposed based on bioinformatics predictions (2), AglD was shown to present a cytoplasmically oriented N terminus that includes the putative catalytic domain. Considering that AglD is predicted to span the membrane six times (residues 260 to 281, 294 to 312, 377 to 401, 426 to 450, 495 to 519, and 555 to 579) (2), it appears that a third external loop (situated between residues 519 to 555) contains the proteinase K-accessible site, given the loss of a 69- to 105-residue fragment, as reflected in the transformation of the 85-kDa CBD-AglD chimera to a faster-migrating species upon protease treatment. Thus, given the cytoplasmic assignment of the putative active sites of AglJ and AglE (3) and AglD, which are participants in adding the first or second (Abu-Qarn et al., unpublished data), the fourth (3), and the final (2) saccharide subunits, respectively, to the pentasaccharide decorating the H. volcanii S-layer glycoprotein, it appears that pentasaccharide assembly occurs within the cytoplasm.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Topology of H. volcanii N-glycosylation proteins


Figure 4
View larger version (26K):
[in this window]
[in a new window]

  		FIG. 4. Schematic depiction of the topologies of the H. volcanii proteins considered in this study. The consensus topologies of CBD-AglD, -AglJ, -Alg5-B, -Dpm1-A, and -Mpg1-D, as determined by bioinformatics predictions, protease accessibility, and cysteine modification approaches, are shown. The numbers shown for each chimera correspond to the amino acid residue of the H. volcanii component; the N terminus (N) is also shown. Cysteine residues discussed in the text are designated with a C. The open curled bracket framing the intracellular domains of AglD, AglJ, and Dpm1-A in each chimera depicts the position of the putative catalytic domain of each H. volcanii protein. The CBD portion of the chimera, drawn with a thin line, is also indicated in each diagram by an open curled bracket. The arrowheads in the CBD-AglD and CBD-Dpm1-A diagrams point to external loops apparently susceptible to the actions of proteinase K. Drawings are not to scale.

Apart from evidence assigning the activity of the archaeal oligosaccharide transferase to the external surface of the cell (14, 26), relatively little was known about the topology of the steps involved in N glycosylation in Archaea prior to this study. The findings reported here thus suggest that the archaeal N-glycosylation process relies on the assembly of oligosaccharides on lipid carriers facing the cytoplasm, as occurs in bacterial N glycosylation and during the first phase of eukaryal glycan assembly in the ER (8, 23). As such, the basic mechanisms of N glycosylation appear to be conserved in Bacteria and Archaea, becoming more sophisticated in Eukarya. Nonetheless, aspects of the N-glycosylation process may be unique to Archaea. Homology-based searches have thus far failed to predict candidates for the archaeal "flippase" responsible for translocating the lipid-linked oligosaccharide across the plasma membrane prior to oligosaccharide transferase-mediated modification of the target protein, pointing to this N-glycosylation pathway component as being unique to Archaea. Such uniqueness may be related to the unique chemistry of the archaeal plasma membrane.


arrow
ACKNOWLEDGMENTS
 
Support for this work came from the Israel Science Foundation (grant 30/07) and the U.S. Air Force Office for Scientific Research (grant FA9550-07-10057).


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Dept. of Life Sciences, Ben Gurion University, P.O. Box 653, Beersheva 84105, Israel. Phone: 972 8646 1343. Fax: 972 8647 9175. E-mail: jeichler{at}bgu.ac.il Back

{triangledown} Published ahead of print on 17 October 2008. Back


arrow
REFERENCES
 
    1
  1. Abu-Qarn, M., and J. Eichler. 2006. Protein N-glycosylation in Archaea: defining Haloferax volcanii genes involved in S-layer glycoprotein glycosylation. Mol. Microbiol. 61:511-525.[CrossRef][Medline]
    2. 2
  2. Abu-Qarn, M., S. Yurist-Doutsch, A. Giordano, A. Trauner, H. R. Morris, P. Hitchen, O. Medalia, A. Dell, and J. Eichler. 2007. Haloferax volcanii AglB and AglD are involved in N-glycosylation of the S-layer glycoprotein and proper assembly of the surface layer. J. Mol. Biol. 374:1224-1236.[Medline]
    3. 3
  3. Abu-Qarn, M., A. Giordano, F. Battaglia, A. Trauner, P. Hitchen, H. R. Morris, A. Dell, and J. Eichler. 2008. Identification of AglE, a second glycosyltransferase involved in N-glycosylation of the Haloferax volcanii S-layer glycoprotein. J. Bacteriol. 190:3140-3146.[Abstract/Free Full Text]
    4. 4
  4. Burda, P., and M. Aebi. 1999. The dolichol pathway of N-linked glycosylation. Biochim. Biophys. Acta 1426:239-257.[Medline]
    5. 5
  5. Chaban, B., S. Voisin, J. Kelly, S. M. Logan, and K. F. Jarrell. 2006. Identification of genes involved in the biosynthesis and attachment of Methanococcus voltae N-linked glycans: insight into N-linked glycosylation pathways in Archaea. Mol. Microbiol. 61:259-268.[CrossRef][Medline]
    6. 6
  6. Reference deleted.
    7. 7
  7. Eichler, J., and M. W. Adams. 2005. Posttranslational protein modification in archaea. Microbiol. Mol. Biol. Rev. 69:393-425.[Abstract/Free Full Text]
    8. 8
  8. Helenius, A., and M. Aebi. 2004. Roles of N-linked glycans in the endoplasmic reticulum. Annu. Rev. Biochem. 73:1019-1049.[CrossRef][Medline]
    9. 9
  9. Igura, M., N. Maita, J. Kamishikiryo, M. Yamada, T. Obita, K. Maenaka, and D. Kohda. 2008. Structure-guided identification of a new catalytic motif of oligosaccharyltransferase. EMBO J. 27:234-243.[CrossRef][Medline]
    10. 10
  10. Irihimovitch, V., and J. Eichler. 2003. Post-translational secretion of fusion proteins in the halophilic archaea Haloferax volcanii. J. Biol. Chem. 278:12881-12887.[Abstract/Free Full Text]
    11. 11
  11. Kohda, D., M. Yamada, M. Igura, J. Kamishikiryo, and K. Maenaka. 2007. New oligosaccharyltransferase assay method. Glycobiology 17:1175-1182.[Abstract/Free Full Text]
    12. 12
  12. Konrad, Z., and J. Eichler. 2002. Lipid modification of proteins in Archaea: attachment of a mevalonic acid-based lipid moiety to the S-layer glycoprotein of Haloferax volcanii following protein translocation. Biochem. J. 366:959-964.[Medline]
    13. 13
  13. Lechner, J., and F. Wieland. 1989. Structure and biosynthesis of prokaryotic glycoproteins. Annu. Rev. Biochem. 58:173-194.[CrossRef][Medline]
    14. 14
  14. Lechner, J., F. Wieland, and M. Sumper. 1985. Biosynthesis of sulfated saccharides N-glycosidically linked to the protein via glucose. Purification and identification of sulfated dolichyl monophosphoryl tetrasaccharides from halobacteria. J. Biol. Chem. 260:860-866.[Abstract/Free Full Text]
    15. 15
  15. Mescher, M. F., and J. L. Strominger. 1976. Purification and characterization of a prokaryotic glucoprotein from the cell envelope of Halobacterium salinarium. J. Biol. Chem. 251:2005-2014.[Abstract/Free Full Text]
    16. 16
  16. Mevarech, M., and R. Werczberger. 1985. Genetic transfer in Halobacterium volcanii. J. Bacteriol. 162:461-462.[Abstract/Free Full Text]
    17. 17
  17. Mizanur, R. M., C. J. Zea, and N. L. Pohl. 2004. Unusually broad substrate tolerance of a heat-stable archaeal sugar nucleotidyltransferase for the synthesis of sugar nucleotides. J. Am. Chem. Soc. 126:15993-15998.[CrossRef][Medline]
    18. 18
  18. Namboori, S. C., and D. E. Graham. 2008. Acetamido sugar biosynthesis in the euryarchaea. J. Bacteriol. 190:2987-2996.[Abstract/Free Full Text]
    19. 19
  19. Ortenberg, R., and M. Mevarech. 2000. Evidence for post-translational membrane insertion of the integral membrane protein bacterioopsin expressed in the heterologous halophilic archaeon Haloferax volcanii. J. Biol. Chem. 275:22839-22846.[Abstract/Free Full Text]
    20. 20
  20. Reuter, C. J., S. J. Kaczowka, and J. A. Maupin-Furlow. 2004. Differential regulation of the PanA and PanB proteasome-activating nucleotidase and 20S proteasomal proteins of the haloarchaeon Haloferax volcanii. J. Bacteriol. 186:7763-7772.[Abstract/Free Full Text]
    21. 21
  21. Shams-Eldin, H., B. Chaban, S. Niehus, R. T. Schwarz, and K. F. Jarrell. 2008. Identification of the archaeal alg7 gene homolog (N-acetylglucosamine-1-phosphate transferase) of the N-linked glycosylation system by cross-domain complementation in yeast. J. Bacteriol. 190:2217-2220.[Abstract/Free Full Text]
    22. 22
  22. Sumper, M., E. Berg, R. Mengele, and I. Strobel. 1990. Primary structure and glycosylation of the S-layer protein of Haloferax volcanii. J. Bacteriol. 172:7111-7118.[Abstract/Free Full Text]
    23. 23
  23. Szymanski, C. M., and B. W. Wren. 2005. Protein glycosylation in bacterial mucosal pathogens. Nat. Rev. Microbiol. 3:225-237.[CrossRef][Medline]
    24. 24
  24. Tozik, I., Q. Huang, C. Zwieb, and J. Eichler. 2002. Reconstitution of the signal recognition particle of the halophilic archaeon Haloferax volcanii. Nucleic Acids Res. 30:4166-4175.[Abstract/Free Full Text]
    25. 25
  25. Weerapana, E., and B. Imperiali. 2006. Asparagine-linked protein glycosylation: from eukaryotic to prokaryotic systems. Glycobiology 16:91R-101R.[Abstract/Free Full Text]
    26. 26
  26. Wieland, F., J. Lechner, G. Bernhardt, and M. Sumper. 1981. Sulphation of a repetitive saccharide in halobacterial cell wall glycoprotein: occurrence of a sulphated lipid-linked precursor. FEBS Lett. 132:319-323.[CrossRef]
    27. 27
  27. Yurist-Doutsch, S., B. Chaban, D. VanDyke, K. F. Jarrell, and J. Eichler. 2008. Sweet to the extreme: protein glycosylation in Archaea. Mol. Microbiol. 68:1079-1084.[CrossRef][Medline]
    28. 28
  28. Yurist-Doutsch, S., M. Abu-Qarn, F. Battaglia, H. R. Morris, P. G. Hitchen, A. Dell, and J. Eichler. 2008. aglF, aglG and aglI, novel members of a gene cluster involved in the N-glycosylation of the Haloferax volcanii S-layer glycoprotein. Mol. Microbiol. 69:1234-1245.[CrossRef][Medline]

Journal of Bacteriology, December 2008, p. 8045-8052, Vol. 190, No. 24
0021-9193/08/$08.00+0     doi:10.1128/JB.01200-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Plavner, N.
Right arrow Articles by Eichler, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Plavner, N.
Right arrow Articles by Eichler, J.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»