SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Proteins in all three domains of life can experience N-glycosylation. The steps involved in the archaeal version of this post-translational modification remain largely unknown. Hence, as the next step in ongoing efforts to identify components of the N-glycosylation pathway of the halophilic archaeon Haloferax volcanii, the involvement of three additional gene products in the biosynthesis of the pentasaccharide decorating the S-layer glycoprotein was demonstrated. The genes encoding AglF, AglI and AglG are found immediately upstream of the gene encoding the archaeal oligosaccharide transferase, AglB. Evidence showing that AglF and AglI are involved in the addition of the hexuronic acid found at position three of the pentasaccharide is provided, while AglG is shown to contribute to the addition of the hexuronic acid found at position two. Given their proximities in the H. volcanii genome, the transcription profiles of aglF, aglI, aglG and aglB were considered. While only aglF and aglI share a common promoter, transcription of the four genes is co-ordinated, as revealed by determining transcript levels in H. volcanii cells raised in different growth conditions. Such changes in N-glycosylation gene transcription levels offer additional support for the adaptive role of this post-translational modification in H. volcanii.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Post-translational protein modifications are responsible for much of the variety and diversity found within the proteome of any organism. Of the various modifications a protein can experience, glycosylation is one of the most prevalent, occurring either on Asn residues, in the case of N-glycosylation, or on amino acids presenting a functional hydroxyl group, such as Ser or Thr, in the case of O-glycosylation (Spiro, 2002). Both N- and O-glycosylation transpire in all three domains of life, i.e. Eukarya, Bacteria and Archaea (Spiro, 2002; Messner, 2004; Eichler and Adams, 2005), although current understanding of each version of these processes is not consistent. In particular, the archaeal N-glycosylation pathway is not as well defined as the parallel eukaryal and bacterial processes (Yurist-Doutsch et al., 2008).

In Haloferax volcanii, sequences homologous to eukaryal and bacterial N-glycosylation genes have been detected and several have been experimentally verified as participating in the N-glycosylation of a reporter protein (Abu-Qarn and Eichler, 2006; Abu-Qarn et al., 2007; 2008). AglD plays a role in the addition of the final subunit of a pentasaccharide decorating the H. volcanii S-layer glycoprotein (Abu-Qarn et al., 2007), while AglE participates in the addition of the fourth subunit to the same pentasaccharide (Abu-Qarn et al., 2008). Finally, AglB acts to transfer the pentasaccharide to at least two of the seven putative N-glycosylation sites of the S-layer glycoprotein, i.e. Asn-13 and Asn-83 (Abu-Qarn et al., 2007).

Examination of those ORFs found upstream of aglB (HVO_1530;http://archaea.ucsc.edu/cgi-bin/hgGateway?db=haloVolc1; Schneider et al., 2006) revealed two sequences previously listed as possible participants in the H. volcanii protein glycosylation process (Abu-Qarn and Eichler, 2006). HVO_1528 corresponds to the H. volcanii homologue of C. jejuni pglI, the product of which adds a glucose branch to the undecaprenolpyrophosphate-linked polysaccharide structure that is ultimately transferred to polypeptide targets in this bacterium (Linton et al., 2005). HVO_1527 corresponds to mpg1-B (Abu-Qarn and Eichler, 2006), one of the five H. volcanii homologues of eukaryal mpg1, encoding the GTP:mannose-1-phosphate guanyltransferase responsible for catalysing the last step in GDP-mannose production (Kruszewska et al., 1998). In the eukaryal N-glycosylation pathway, this nucleotide-activated mannose is transferred to a dolichol phosphate carrier facing the cytoplasmic side of the ER membrane. The charged carrier is then re-orientated to face the ER lumen and the mannose subunit is transferred to the growing polysaccharide that is ultimately attached to select Asn residues of nascent polypeptides being translated and translocated into the ER lumen (Helenius and Aebi, 2004).

The current annotation of the H. volcanii genome lists HVO_1529 as encoding a homologue of ExoM, a β1–4 glucosyltransferase thought to be involved in the biosynthesis of a bacterial exopolysaccharide succinoglycan (Glucksmann et al., 1993). However, given the documented difficulty in assigning function to sugar-binding proteins (Coutinho et al., 2003), this annotation may be incorrect. Indeed, blast searches reveal HVO_1529 to be homologous to sequences simply referred to as glycosyltransferases, including N-acetylgalactosamine transferases, involved in mucin-type protein O-glycosylation in Eukarya (Marth, 1996; Ten Hagen et al., 2003).

Given their physical proximity to H. volcanii genes implicated in protein glycosylation and homology to genes putatively implicated in protein glycosylation elsewhere, the involvement of HVO_1527, HVO_1528 and HVO_1529 in H. volcanii protein N-glycosylation was further considered. The results obtained confirm the participation of the three genes found upstream of aglB, namely HVO_1527 (now renamed aglF), HVO_1528 (now renamed aglI) and HVO_1529 (now renamed aglG), in H. volcanii S-layer glycoprotein N-glycosylation.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Deletion of HVO_1527, 1528 or 1529 does not compromise H. volcanii survival

As a first step in assessing the putative involvement of HVO_1527, 1528 and 1529 in H. volcanii protein glycosylation, each gene was deleted according to the protocol developed by Allers et al. (2004) and successfully used by numerous laboratories for the study of a variety of genes (cf. Soppa et al., 2008). In this approach, the sequence under consideration is replaced by the tryptophan synthase-encoding H. volcanii trpA gene (HVO_0789), introduced into the genome of the uracil and tryptophan auxotrophic H. volcanii strain WR536 by the pyrE-containing plasmid pTA131 and plating onto casamino acids lacking uracil and tryptophan.

PCR amplification was performed to follow genomic integration of the introduced plasmids as well as the subsequent expulsion of the plasmid together with HVO_1527, 1528 or 1529. However, given the comparable sizes of these genes and the trpA sequence (732, 888, 942 and 834 nucleotides, respectively), each gene was followed by dual PCR amplifications using forward primers raised against internal sequences within HVO_1527, 1528, 1529 or trpA and a reverse primer directed against a sequence within the flanking region downstream to HVO_1527, 1528 or 1529, as appropriate. As revealed in Fig. 1B (left panels), whereas those primer pairs directed against internal and downstream flanking regions of HVO_1527, 1528 or 1529 yielded PCR amplification products in the background strain (right pair of lanes in each panel; 1467, 1422 and 1251 bp, respectively), only those primer pairs directed against an internal sequence of trpA and the flanking regions downstream to HVO_1527, 1528 or 1529 yielded PCR amplification products in the deletion strain (left pair of lanes in each panel; 1353, 1368 and 1359 bp, respectively). These results thus point to respective replacement of HVO_1527, 1528 and 1529 by trpA. Deletion of each gene was further confirmed when PCR amplification was performed using genomic DNA from the deletion strains as template and primers directed against the HVO_1527, 1528 or 1529 coding regions (732, 888 and 942 bp, respectively; Fig. 1B, right panels).

image

Figure 1. HVO_1527, 1528 and 1529 are not essential for H. volcanii survival. A. Schematic representation showing the orientations of HVO_1527, 1528, 1529 and 1530 as well as proven (HVO_1530) annotations. B. Left panels: PCR amplification was performed using a forward primer directed at the HVO_1527, 1528 or 1529 3′ flanking regions and a reverse primer directed at a sequence within the HVO_1527, 1528 or 1529 coding regions (primer pair a) or a sequence within the trpA sequence (primer pair b), together with genomic DNA from cells of the WR536 background strain (bkgnd) or from cells that had replaced the HVO_1527, 1528 or 1529 gene (deletion; top, middle and bottom panels, respectively), as template. Right panel: PCR amplification was performed using primers directed against the HVO_1527, 1528 or 1529 coding regions, together with genomic DNA from cells of the WR536 background strain (bkgnd) or the HVO_1527, 1528 or 1529-deleted strains (deletion; top, middle and bottom panels, respectively). C. RT-PCR was performed using primers directed at HVO_1527 (top row of panels), HVO_1528 (middle row of panels) or HVO_1529 (bottom row of panels) and cDNA (left lane of each panel), RNA (middle lane of each panel) from HVO_1527, 1528 or 1529-deleted strains (left, middle and right columns of panels, respectively) as template. In the right lane of each panel, no nucleic acid template was added to the reaction (blank). The identities of PCR products were confirmed by sequencing.

Download figure to PowerPoint

The absence of HVO_1527, 1528 or 1529 in the respective deletion strains was next ascertained at the RNA level by RT-PCR, performed as described previously (Abu-Qarn and Eichler, 2006). In these experiments, RNA or cDNA generated from the RNA of each deletion strain or no nucleic acids (blank) served as template for PCR amplifications, together with primers directed against the coding region of HVO_1527, 1528 or 1529. As reflected in Fig. 1C, no PCR products were obtained when cDNA from any of the deletion strains served as template in reactions involving primers directed against the deleted gene in question. By contrast, PCR products were readily obtained when the same reactions were repeated using primers directed against either of the other two sequences. For example, when PCR amplification was performed using cDNA obtained from cells deleted of HVO_1527 (Fig. 1C, left panels, left lanes), PCR products were obtained when primers to the HVO_1528 or 1529 coding regions were employed (middle and bottom panels, respectively) but not when primers to the HVO_1527 coding region were included in the reaction (top panel). Similarly, no PCR products appeared when RNA served as template (middle lanes of each panel) or when no nucleic acids were present (blank, right lanes of each panel). These results thus reflect the deletion of HVO_1527, 1528 and 1529 at the RNA level and, moreover, reveal that the absence of HVO_1527, 1528 or 1529 does not compromise the transcription of the other two genes.

H. volcanii cells lacking HVO_1527, 1528 or 1529 present a modified S-layer

Having determined that HVO_1527, 1528 and 1529 are not essential for H. volcanii viability, the participation of the gene products in protein glycosylation was considered by examining the S-layer glycoprotein, a well-characterized archaeal reporter of this post-translational modification (Sumper et al., 1990; Mengele and Sumper, 1992, Eichler, 2000), in cells deleted of each gene. As reflected in Fig. 2A, when the S-layer glycoprotein from the H. volcanii WR536 strain background and the same strain lacking HVO_1527, 1528 or 1529 were compared by SDS-PAGE and Coomassie staining, the faster migration of the protein from the mutant cells was evident. To confirm that such enhanced migration of the S-layer glycoprotein on SDS-PAGE was due to the absence of the individual genes and not the outcome of a general perturbation of the genome in the region of HVO_1527, 1528 and 1529, each deletion strain was transformed to express a plasmid-based copy of the absent gene, engineered to include a cellulose-binding domain (CBD) tag, with expression being confirmed by immunoblot using anti-CBD antibodies (Fig. 2B). In the case of HVO_1527 and 1529, such complementation restored the original SDS-PAGE behaviour of the S-layer glycoprotein (Fig. 2C). The failure of plasmid-encoded CBD-HVO_1528 to restore S-layer glycoprotein migration in SDS-PAGE to that of the native protein may be due to several causes, including steric interference by the fused CBD tag, introduced for purposes of detection. Nonetheless, the observation that HVO_1527 and HVO_1529 mRNA is detected in the HVO_1528 deletion strain (Fig. 1C) argues that effects resulting from the absence of HVO_1528 (such as modified S-layer glycoprotein apparent molecular weight) are due to the missing gene product rather than arising in a non-specific, unrelated manner because of genome disruption.

image

Figure 2. The absence of HVO_1527, 1528 or 1529 affects S-layer glycoprotein migration on SDS-PAGE. A. Equivalent aliquots of H. volcanii WR356 cells (bkgnd), or the same cells lacking HVO_1527HVO_1527; top panel), HVO_1528HVO_1528; middle panel) or HVO_1529HVO_1529; bottom panel) were separated by 5% SDS-PAGE and Coomassie blue-stained. The position of the S-layer glycoprotein is shown. B. The expression of CBD-tagged HVO_1527 (top panel), HVO_1528 (middle panel) and HVO_1529 (top panel) in the complemented deletion strains, as confirmed by immunoblot after separation on 15% SDS-PAGE using anti-CBD antibodies. C. Equivalent aliquots of H. volcanii WR356 cells lacking HVO_1527HVO_1527), HVO_1529HVO_1529) or cells of the deletion strain transformed with a plasmid encoding a CBD-tagged version of the deleted gene (ΔHVO_1527/CBD-HVO_1527;ΔHVO_1529/CBD-HVO_1529) were separated by 5% SDS-PAGE and Coomassie blue-stained. The position of the S-layer glycoprotein is shown.

Download figure to PowerPoint

To determine whether the enhanced migration of the S-layer glycoprotein in the HVO_1527-, 1528- or 1529-lacking cells was indicative of modifications that affected the integrity of the S-layer surrounding H. volcanii, thought to be composed solely of the S-layer glycoprotein (Sumper et al., 1990), WR536 background cells, as well as cells of the same strain deleted of HVO_1527, 1528 or 1529, were challenged with proteinase K for up to 3 h. The proportion of non-digested S-layer glycoprotein remaining at increasing intervals from the onset of proteolysis was then considered. Such analysis revealed the S-layer glycoprotein in the background strain (as well as in cells deleted of HVO_A0586, a seemingly unrelated putative nucleoside diphosphate sugar pyrophosphorylase) to be less susceptible to proteolytic digestion than its counterparts in the HVO_1527-, 1528- or 1529-deleted strains (Fig. 3). Moreover, complementation of HVO_1527- or HVO_1529-deleted cells to express a CBD-tagged version of the encoded protein restored the protease resistance of the S-layer to that of the background strain. It thus appears that the source of the enhanced HVO_1527-, 1528- and 1529-deleted strain-derived S-layer glycoprotein SDS-PAGE migration also compromises the proper assembly of the protein shell surrounding H. volcanii cells in these strains.

image

Figure 3. The S-layer surrounding H. volcanii cells is protease-sensitive in cells lacking HVO_1527, 1528 or 1529. Background strain WR536 (top panel), and HVO_1527-, 1528- or 1529-lacking cells of the same strain (second, third and fourth panels, respectively) were challenged with 1 mg ml−1 proteinase K at 42°C. In the fifth and sixth panels, cells lacking HVO_1527 or HVO_1529, respectively, transformed to express a CBD-tagged version of the deleted gene, were similarly challenged. Aliquots were removed immediately prior to incubation with proteinase K and at 15–30 min intervals following addition of the protease for up to 3 h and examined by 7.5% SDS-PAGE. In a control experiment, H. volcanii cells deleted of a seemingly non-related gene (HVO_A0586) were similarly challenged (bottom panel).

Download figure to PowerPoint

HVO_1527, 1528 and 1529 participate in the assembly of the pentasaccharide decorating the H. volcanii S-layer glycoprotein

As homology-based predictions assign the deduced products of HVO_1527, 1528 and 1529 roles in N-glycosylation, experiments were next performed to directly test these predictions. Indeed, defective N-glycosylation could explain the observations described above. Accordingly, SDS-PAGE gel pieces containing the S-layer glycoprotein from the H. volcanii background strain, as well as from cells lacking HVO_1527, 1528 or 1529, were subjected to in-gel tryptic digestion. The obtained peptides were separated by liquid chromatography and MS/MS was employed to reveal peptide sequences. The six peptides generated in this manner included the N-terminal 1ERGNLDADSESFNK14 peptide (1581 m/z), encompassing the glycosylated Asn-13 residue (Sumper et al., 1990; Abu-Qarn et al., 2007). Previous MALDI TOF mass mapping of the nanoLC-purified tryptic digest, complemented by MS/MS analyses using MALDI TOF/TOF and electrospray Q-TOF instrumentation, had shown this S-layer glycoprotein-derived peptide to be modified by a novel pentasaccharide (Abu-Qarn et al., 2007). In agreement with this earlier study, the Asn-13-containing peptide isolated from the WR536 background strain was now shown to be decorated by the same pentasaccharide moiety (m/z 2447), composed of a hexose (162 Da), followed by two 176 Da residues (hexuronic acids), one 190 Da residue (likely either dimethylated hexose or the methyl ester of hexuronic acid) and an additional hexose residue at the end of the glycan chain. In addition, the same peptide modified by precursor mono- (m/z 1743.7), di- (m/z 1919.7), tri- (m/z 2095.8) and tetrasaccharides (m/z 2285.5) were also observed (Fig. 4, top left panel, bkgnd).

image

Figure 4. The products of H. volcanii HVO_1527, 1528 and 1529 are involved in the biogenesis of the pentasaccharide decorating S-layer glycoprotein Asn residues. The MALDI-TOF spectra of the Asn-13-containing tryptic peptide derived from the S-layer glycoprotein of the WR536 background cells (upper left panel) and cells from the HVO_1527- [ΔHVO_1527 (aglF); upper right panel], HVO_1528- [ΔHVO_1528 (aglI); lower left panel] or HVO_1529- [ΔHVO_1529 (aglG); lower right panel] deleted strains are shown. The components of the peptide-associated glycan are shown as an inset in the upper left panel, while the sugar subunits decorating the peptide peaks detected are indicated on the MALDI-TOF spectra.

Download figure to PowerPoint

When the same S-layer glycoprotein peptide was derived from H. volcanii cells deleted of HVO_1527, a very different profile was obtained. In this case, the monosaccharide-bearing species as well as a lesser amount of the disaccharide-bearing peptide were observed [Fig. 4, top right panel, ΔHVO_1527 (aglF)]. In the case of cells deleted of HVO_1528, an identical pattern was obtained [Fig. 4, bottom left panel, ΔHVO_1528 (aglI)]. Due to the weak nature of the signal at m/z 1919.7 in the ΔHVO_1527 (aglF) and ΔHVO_1528 (aglI) profiles, relative to the m/z 1743.8 signal, two further analyses were performed (data not shown). Standard deviations of ±2.2 and ±0.9 were obtained for the relative intensities between these signals in the ΔHVO_1527 (aglF) and ΔHVO_1528 (aglI) samples, respectively. When the N-terminal S-layer glycoprotein tryptic peptide from cells lacking HVO_1529 was examined as above, only the monosaccharide-decorated species was detected [Fig. 4, bottom right panel, ΔHVO_1529 (aglG)]. The products of HVO_1527 and 1528 are therefore involved in the addition of the distal hexuronic acid of the pentasaccharide decorating Asn-13, while the product of HVO_1529 is involved in addition of the proximal hexuronic acid of the same oligosaccharide.

Thus, given the involvement of HVO_1527, 1528 and 1529 in H. volcanii S-layer glycoprotein N-glycosylation, as revealed by analysis of SDS-PAGE migration, susceptibility to proteolysis and mass spectrometry of this reporter, HVO_1527, 1528 and 1529 are now renamed aglF, aglI and aglG, according to the nomenclature proposed by Chaban et al. (2006) for genes involved in archaeal N-glycosylation.

The transcription of aglB, aglF, aglG and aglI is regulated in a co-ordinated manner

Given their physical proximity in the genome, as well as the common involvement of their products in N-glycosylation, efforts next focused on whether the transcription of aglB, aglF, aglG and aglI is co-ordinated. Towards this end, the transcription profile of these genes was initially considered by investigating their upstream regions for the presence of promoters.

As aglF and aglI lie adjacent to each other on the H. volcanii genome, assume the same orientation and are apparently separated by only 51 nucleotides, the possibility that the two genes are co-transcribed was tested. Accordingly, RT-PCR was performed using cDNA derived from RNA extracted from cells grown to mid-exponential phase, together with a forward primer directed against the 5′ end of aglF and a reverse primer directed against the 3′ end of aglI. As reflected in Fig. 5A, a single PCR product was obtained, confirmed by sequencing to contain both aglF and aglI. To further demonstrate that the transcription of aglF and aglI is under the control of a common promoter, the DNA sequence separating HVO_1526 and 1527, i.e. that region lying upstream of the predicted start site of aglF, was introduced into plasmid pJAM1020 (Reuter and Maupin-Furlow, 2004), encoding for GFP, in place of the Halobacterium cutirubrum rRNA P2 promoter originally present in the plasmid. Preliminary control experiments confirmed that in the absence of the native promoter, no GFP expression could be detected (Fig. 5B). When the modified plasmid containing the 180 bp region upstream of the predicted aglF start site in place of the original promoter was used to transform H. volcanii cells, the expression of GFP could be clearly seen. By contrast, far less GFP expression was achieved when the original plasmid pJAM1020 promoter region was replaced with the 51 bp sequence separating aglF and aglI. Thus, although the region upstream of aglI is capable of directing protein expression to a limited extent, the augmented level of protein expression directed by the stronger aglF promoter offers additional support for the concept that aglF and aglI, the products of which jointly participate in addition of a hexuronic subunit to the S-layer glycoprotein pentasaccharide, are co-transcribed under the control of a single promoter.

image

Figure 5. Functional characterization of the promoter regions of H. volcanii aglB, aglF, algG and aglI. A. RT-PCR reveals the co-transcription of aglF and aglI. PCR amplification was performed using a forward primer directed against the start of the coding region of aglF and a reverse primer against the end of the coding region of aglI together with cDNA (lane 1), RNA (lane 2) or DNA (lane 3) from H. volcanii strain WR536 background cells as template, or no nucleic acid template (lane 4). B. Upper panels: H. volcanii strain WR536 cells were transformed to express GFP, as directed by plasmid pJAM-1020 in which the promoter region had been removed, in which the native promoter was present, or when the region upstream to aglF or aglI replaced the native promoter of the plasmid. Lower panel: The 118 bp region separating aglG and aglB served as promoter (aglB lane). In lane aglG, the same region, this time in the reverse orientation, served as the promoter in plasmid pJAM-1020. GFP expression was visualized by immunoblotting using anti-GFP antibodies.

Download figure to PowerPoint

Unlike aglF and aglI, which assume the same orientation in the H. volcanii genome, aglG and aglB are oriented in opposite directions, with the direction of aglG being inverted. To determine whether the 118 bp DNA sequence separating the predicted start sites of aglG and aglB drives the transcription of either or both genes, the original plasmid pJAM1020 promoter was replaced by the 118 bp region, introduced in either orientation. As also reflected in Fig. 5B, the transcription of GFP was driven by this H. volcanii sequence, regardless of its orientation. The level of GFP detected was, however, somewhat higher when the 118 bp region was inserted in the forward direction, pointing to the promoter being stronger in driving the transcription of aglB than of algG.

While aglB, aglF, aglG and aglI do not form an operon, given the differential orientation of aglG, the possibility remains that their transcription is somehow linked. To begin testing this hypothesis, the relative levels of aglB, aglF, aglG and aglI transcription were qualitatively assessed by comparing the levels of GFP in cells transformed to transcribe the encoding gene under the control of the aglB, aglF, aglI or aglG promoters, grown to mid-exponential phase. As reflected in Fig. 5, GFP expression driven through the aglB promoter exceeded that expression directed through the aglF, aglI or aglG promoters. To quantify these observations, real-time RT-PCR was performed to assess the relative amounts of AglB, AglF, AglG and AglI mRNA in H. volcanii cells grown to mid-exponential phase in complete medium, using primer pairs that bind with equivalent efficiencies (as determined in preliminary experiments involving the drawing of standard curves describing primer pair binding to serial dilutions of known quantities of cDNA). Based on the results of three experiments, each involving triplicate samples, it could be concluded that H. volcanii cells grown to mid-exponential phase in complete medium contain threefold less AglF mRNA, and fivefold less AglG and AglI mRNA than AglB mRNA (with all differences being significant to P < 0.01) (Fig. 6A). At present, it is not clear why different levels of AglF and AglI mRNA were detected, if the two genes are co-transcribed. Thus, cells grown to mid-exponential phase contain different amounts of aglB, aglF, aglG and aglI mRNA.

image

Figure 6. The transcription of aglB, aglF, algG and aglI is co-ordinated. A. Real-time RT-PCR was employed to assess the relative amounts of aglB, aglF, algG and aglI mRNA in H. volcanii strain WR536 cells grown to mid-exponential phase in rich medium. Values shown represent the average of three experiments ± standard deviation, expressed relative to the level of aglB mRNA, taken as 1. Differences from the level of aglB RNA marked with the double asterisk are statistically significant to P < 0.01, as determined by Student's t-test. B. Real-time RT-PCR was employed to assess the fold increase in aglB, aglF, algG and aglI mRNA in H. volcanii strain WR536 cells grown to stationary phase (stat.), subjected to heat shock, or raised in low or high salt-containing medium, relative to those levels detected in cells grown to mid-exponential phase in rich medium. The values shown represent the average of 3–5 experiments. Bars marked with the double asterisk are statistically distinct to a significance of P < 0.01, while those marked with single asterisks are statistically distinct to a significance of P < 0.05, as determined by Student's t-test.

Download figure to PowerPoint

To determine whether changes in the transcription profile of aglB, aglF, aglG and aglI take place in a co-ordinated manner, real-time PCR was employed to quantify AglB, AglF, AglG and AglI mRNA levels in cells grown to stationary phase, challenged with heat shock (i.e. 65°C for 45 min) or raised to mid-exponential phase in low salt- or high salt-containing medium (i.e. 1.75 or 4.8 M NaCl, respectively). In these experiments, 16S rRNA was considered as a housekeeping marker to allow for normalization of mRNA levels from cells in each growth condition. The AglB, AglF, AglG and AglI mRNA levels measured in these various growth conditions were, in turn, expressed in terms of fold increase relative to those values obtained from cells grown to mid-exponential phase in complete medium. Initially, AglB, AglF, AglG and AglI mRNA levels in cells either grown to stationary phase or exposed to heat shock conditions are considered. In cells grown to stationary phase, AglB, AglF, AglG and AglI mRNA levels were substantially reduced, relative to those levels realized during mid-exponential growth (Fig. 6B). AglB, AglF, AglG and AglI mRNA levels were also depressed upon transfer to heat shock conditions, relative to the situation realized in cells grown to mid-exponential phase. In both growth conditions, the decrease in AglB mRNA levels was 10- to 20-fold greater than the observed reduction in AglF, AglG or AglI mRNA levels.

A very different picture was obtained when real-time RT-PCR was performed with cells grown in the presence of reduced or elevated salt concentrations. In the case of cells grown in 1.75 M NaCl-containing medium, statistically significant (P < 0.01) increases in AglF, AglG and AglI mRNA levels were detected, while in cells grown in 4.8 M NaCl-containing medium, statistically significant (P <  0.01) increases in AglF and AglI mRNA levels were noted, in both cases relative to those levels measured in cells grown to mid-exponential phase in medium containing 3.5 M NaCl. The observed increases in transcription were higher in the case of cells grown in low-salt conditions. In both low- and high-salt growth conditions, the increase in AglB mRNA levels was not statistically significant.

Thus, real-time RT-PCR reveals that changes in the transcription of aglB, aglF, aglG and aglI transpire in a co-ordinated manner in the face of different growth conditions, although the nature of such changes depends on the conditions experienced, possibly reflecting an adaptive role of N-glycosylation in H. volcanii.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The list of H. volcanii genes whose products participate in the N-glycosylation of a reporter glycoprotein, the S-layer glycoprotein, is growing (Abu-Qarn and Eichler, 2006; Abu-Qarn et al., 2007; Abu-Qarn et al., 2008). The present study confirms the involvement of three additional gene products in the N-glycosylation pathway, namely AglF, AglG and AglI, showing that all three proteins serve roles in the biogenesis of the pentasaccharide decorating at least two of the modified sequons of the H. volcanii S-layer glycoprotein (Abu-Qarn et al., 2007). Specifically, AglF and AglI are involved in the addition of the hexuronic acid found at position three of the pentasaccharide, while AglG contributes to the addition of the hexuronic acid found at position two. These findings, together with the earlier identification of AglE and AglD as, respectively, participating in the addition of the 190 Da and hexose species found at positions four and five of the pentasaccharide (Abu-Qarn et al., 2007; Abu-Qarn et al., 2008), as well as the oligosaccharide transferase, AglB (Abu-Qarn et al., 2007), are leading to the delineation of the N-glycosylation pathway in H. volcanii (Fig. 7 and Yurist-Doutsch et al., 2008).

image

Figure 7. Schematic depiction of the H. volcanii N-glycosylation pathway, as described to date. Based on the findings presented in this and earlier reports (Abu-Qarn and Eichler, 2006; Abu-Qarn et al., 2007; 2008), AglB, AglD, AglE, AglF, AglG and AglI have been shown to participate in the series of reactions ultimately leading to N-glycosylation of the H. volcanii S-layer glycoprotein. Apart from AglB, all of the other N-glycosylation pathway components appear to act on the cytoplasmic face of the plasma membrane. After reorientation of the lipid-linked pentasaccharide to face the cell exterior by an as yet unidentified agent, AglB transfers the glycan moiety to select sequons in the protein target. While AglB also transfers precursor polysaccharides to the S-layer glycoprotein, it remains to be determined whether the same ‘flippase’ translocates the lipid-linked pentasaccharide precursors across the plasma membrane. The legend describes the components of the S-layer glycoprotein-bound pentasaccharide.

Download figure to PowerPoint

As is the case with AglD and AglE, the specific roles played by AglF, AglG and AglI remain undefined. Homology-based analysis reveals that AglF, first identified through its similarity to eukaryal Mpg1 (Abu-Qarn and Eichler, 2006), contains a COG1210 UDP-glucose pyrophosphorylase domain, involved in UDP-glucose generation. If AglF indeed serves such a role, then the observation that aglF deletion prevented addition of the second but not the first hexuronic acid to the S-layer glycoprotein pentasaccharide suggests these two monosaccharides to be non-identical. This assumption awaits a complete chemical description of the pentasaccharide. By contrast, both AglI, originally detected due to its homology to C. jejuni PglI (Abu-Qarn and Eichler, 2006), and AglG, notpreviously implicated in H. volcanii N-glycosylation, contain Pfam00535 glycosyltransferase 2 domains. These enzymes may thus be responsible for, respectively, adding the third and second sugar subunits to a dolichol-linked monosaccharide as part of the assembly of the S-layer glycoprotein-linked pentasaccharide. Still, like AglD (Abu-Qarn et al., 2007) and AglE (Abu-Qarn et al., 2008), accurate description of AglF, AglG and AglI function awaits the development of in vitro H. volcanii N-glycosylation assays.

It is, however, clear that the perturbed N-glycosylation of the S-layer glycosylation resulting from deletion of aglF, aglG or aglI affects the behaviour of the H. volcanii S-layer, as revealed by the enhanced protease sensitivity of the S-layer glycoprotein in the deletion strains. This observation lends support to the previously proposed hypothesis that a properly glycosylated S-layer is important for H. volcanii survival (Abu-Qarn et al., 2007). Nonetheless, the perturbation to the S-layer that transpires in the absence of AglF, AglG or AglI (or indeed AglE; Abu-Qarn et al., 2008) seems less significant than what occurs either in the absence of AglD, involved in addition of the final hexose of the N-linked pentasaccharide, or AglB, the H. volcanii oligosaccharide transferase (Abu-Qarn et al., 2007). While the absence of AglD led to the appearance of an S-layer with modified architecture, and deletion of aglB resulted in enhanced S-layer glycoprotein release from the cell, no such effects were detected in the absence of AglF, AglG or AglI (not shown). Moreover, the absence of AglF, AglG or AglI had little effect on the ability of H. volcanii cells to grow in increasingly saline medium (not shown), in contrast to what was observed in the aglD and aglB deletion strains (Abu-Qarn et al., 2007).

In addition to assessing the contributions of AglF, AglG and AglI to N-glycosylation in H. volcanii, the present study also addressed questions related to the transcription of genes comprising the N-glycosylation gene island that includes aglF, aglI, aglG and aglB. By allowing the regions lying upstream of these genes to direct the expression of GFP, as well as through real-time RT-PCR, relations between the various genes were elucidated. Such analyses revealed aglF and aglI to be co-transcribed and showed that the same region directs the expression of both aglG and aglB, albeit in different orientations, and at different efficiencies. Nonetheless, the co-ordinated behaviour of the four genes could be demonstrated. In response to stationary phase growth or heat shock, H. volcanii cells present less mRNA directing the synthesis of those enzymes involved in the assembly of the pentasaccharide decorating the S-layer glycoprotein, i.e. aglF, aglI and aglG, than in cells grown to mid-exponential phase. Cells grown to stationary phase or challenged with heat shock also drastically reduce the amount of AglB mRNA, directing the biosynthesis of the enzyme catalysing the final step of N-glycosylation. When H. volcanii cells are, however, grown in medium containing lowered or elevated salt concentrations, the opposite holds true. Here, AglB mRNA levels are maintained relatively constant while transcription of the neighbouring AglF-, AglG- and AglI-encoding genes is augmented. Thus, while transcription of AglB, AglF, AglG and AglI mRNA transpires in a co-ordinated manner, the levels of these N-glycosylation island genes can be either jointly augmented or diminished, depending on the growth conditions. These observations are in line with earlier studies reporting differential transcription of the members of homologous H. volcanii gene families putatively involved in N-glycosylation as a function of growth conditions (Abu-Qarn and Eichler, 2006). Together, these findings support the hypothesis that in H. volcanii, N-glycosylation may be modulated in response to the stage or conditions of growth.

The gene island encompassing HVO_1527, 1528, 1529 and 1530, respectively, encoding AglF, AglI, AglG and AglB, lays downstream of AglE, shown to be encoded by a DNA sequence lying between the wrongly delineated HVO_1523 and HVO-1524 sequences (Abu-Qarn et al., 2008). The intervening sequences, i.e. HVO_1525 and 1526, are currently annotated as encoding a putative membrane protein and an insertion element respectively. Future efforts will address whether the products of these genes also play a role in N-glycosylation. The same holds true for HVO_1531 and 1532, annotated as encoding a UDP-glucose dehydrogenase and a predicted membrane protein, respectively.

In conclusion, while the present study has expanded our understanding of protein N-glycosylation in Archaea, much remains unknown of the archaeal version of this post-translational modification (Yurist-Doutsch et al., 2008). Continued efforts at deciphering the N-glycosylation in H. volcanii as well as in other species, such as Methanococcus voltae (Chaban et al., 2006; Shams-Eldin et al., 2008) and Pyrococcus furiosus (Igura et al. 2008), will help provide a more complete picture of this protein processing event across evolution.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Growth conditions

Haloferax volcanii WR536 (H53; Allers et al. 2004; Table 1), obtained from Moshe Mevarech (Tel Aviv University), was grown in complete medium containing 3.4 M NaCl, 0.15 M MgSO47H20, 1 mM MnCl2, 4 mM KCl, 3 mM CaCl2, 0.3% (w/v) yeast extract, 0.5% (w/v) tryptone, 50 mM Tris-HCl, pH 7.2, at 40°C (Mevarech and Werczberger, 1985). For low-salt growth conditions, 1.75 M NaCl was included in the growth medium, for high-salt growth conditions, 4.8 M NaCl was included in the growth medium, whereas for heat shock, cultures grown in complete medium were transferred to a 65°C environment for 45 min. In casamino acids medium, the yeast extract and tryptone were replaced by casamino acids (Difco, Detroit MI) at a final concentration of 0.5% (w/v). Escherichia coli were grown in Luria–Bertani medium.

Table 1.  Strains and plasmids used in this study.
 DescriptionReferences
Strains
 WR536 (H53)H. volcanii DS70 background, ΔpyrE2, ΔtrpAAllers et al. (2004)
 WR536ΔaglFH. volcanii WR536 pTA131aglF pop-in/pop-out; trpA+, ΔpyrE2, ΔaglFThis study
 WR536ΔaglGH. volcanii WR536 pTA131aglG pop-in/pop-out; trpA+, ΔpyrE2, ΔaglGThis study
 WR536ΔaglIH. volcanii WR536 pTA131aglI pop-in/pop-out; trpA+, ΔpyrE2, ΔaglIThis study
Plasmids
 pTA131pBluescriptII with pGB70 BamHI-XbaI pyrE2-containing fragment (Bitan-Banin et al., 2003) under control of Halobacterium salinarum ferrodoxin promoter (Pfeifer et al., 1993).Allers et al. (2004)
 pTA131aglFpTA131 containing XhoI-HindIII and BamHI-XbaI flanking regions of aglFThis study
 pTA131aglGpTA131 containing XhoI-HindIII and BamHI-XbaI flanking regions of aglGThis study
 pTA131aglIpTA131 containing XhoI-HindIII and BamHI-XbaI flanking regions of aglIThis study
 pJAM10200.74 kb BamHI-SacI fragment of pSMRSGFP blunt end ligated with a 9.94 kb NdeI-BlpI fragment of pJAM202, an H. volcanii-E. coli shuttle expression plasmid with psmB-his6 geneReuter and Maupin-Furlow (2004)
 pJAM1020nopropJAM1020 lacking its promoter (BamHI-XbaI cut), blunt-end ligationThis study
 pJAMaglFpropJAM1020 with aglF promoter (BamHI-XbaI cut) replacing native promoterThis study
 pJAMaglGpropJAM1020 with aglG promoter (BamHI-XbaI cut) replacing native promoterThis study
 pJAMaglIpropJAM1020 with aglI promoter (BamHI-XbaI cut) replacing native promoterThis study
 pJAMaglBpropJAM1020 with aglB promoter (BamHI-XbaI cut) replacing native promoterThis study
 pWL-CBDpWL-Nov vector containing the PrR16 promoter fused to the cbd gene encoding for the C. thermocellum cellulosome CBDIrihimovitch and Eichler (2003)
 pWL-CBD-HVO_1527pWL-CBD containing NdeI-KpnI-flanked HVO_1527This study
 pWL-CBD-HVO_1528pWL-CBD containing NdeI-KpnI-flanked HVO_1528This study
 pWL-CBD-HVO_1529pWL-CBD containing NdeI-KpnI-flanked HVO_1529This study

Gene deletion and complementation

Deletion of H. volcanii HVO_1527,1528 and 1529 was achieved as previously described (Abu-Qarn and Eichler, 2006). The primers used to amplify regions of approximately 500 bp in length flanking the coding sequences of each gene are listed in Table 2. XhoI and HindIII sites were introduced in the 1527-, 1528-, 1529-for 5′ up and -rev 5′ up sequences, respectively, while BamHI and XbaI sites were introduced in the corresponding 3′ down and rev 3′ down sequences respectively. For complementation, HVO_1527, 1528, 1529 were PCR amplified from H. volcanii strain WR536 genomic DNA using primers designed to introduce NdeI and KpnI restriction sites (Table 2) at the 5′-and 3′ ends of the coding region, respectively, and ligated into the pGEM-T Easy vector (Promega). The HVO_1527, 1528, 1529 genes were then excised upon digestion with NdeI and KpnI and inserted into the pWL-CBD vector (Irihimovitch and Eichler, 2003), also digested with the same restriction enzymes, resulting in DNA encoding the Clostridium thermocellum cellulosome CBD fused to the 5′ end of HVO_1527, 1528 or 1529 respectively.

Table 2.  Primers used in this study.
Primer nameForward primersReverse primers
  1. Genomic DNA sequences are in capitals.

Flanking region primers
 aglF 5′ upgggctcgagCGTCATTACGAACCCATACTcccaagcttAGTAAGAGAGTCATCGAGGC
 aglF 3′ downgggggatccACATCTAATCACGTGTGCTTccctctagaTGTGCGTCAAACCTTGCTGG
 aglG 5′ upgggctcgagCCAAAAGCGACTTGGCTACGcccaagcttAAGTCGGAGTTACCGAGGAG
 aglG-3′ downgggggatccTTGGCATTTCAGCGGGTGTTccctctagaCACAGACCGCCTTTCCCATA
 aglI 5′ upgggggctaccGCTGATGCTTGGCGACAACATcccctcgagTGAAATCAGGTTTACTCCCAC
 aglI 3′ downgggggatccCCACGAGGTTCGGCGTCAACAccctctagaCGTCGGGTGTGACGAACGTG
Open reading frame primers
 aglFTAAAGGAACCCGTCTTCGACGTCGTTGTTCTGCTTCGTCA
 aglGCTCGATGGAACGGTACGAGTTTCGTCTTCTCCACGAGGTT
 aglIATGGCTGATTCTCCGTTTCCTCAGCGGGTGTTCCCGCGAACG
 trpAcccgaattcTTATGTGCGTTCCGGATGCG 
Complementation primers
 CBD-aglFgggcatATGCAAGCTGTTGTCCTCGCCcccggtaccCTACTCGGTCGCCTGTGTCGTTTCC
 CBD-aglGgggcatATGAAAGTCTCCGTCGTGGTCcccggtaccCTAATTATTCGTCTTCTCCACG
 CBD-aglIgggcatATGGCTGATTCTCCGTTTCCTTGcccggtaccTCAGCGGGTGTTCCCGCGAACG
Plasmid pJAM1020 promoter primers
 aglF promotergggtctagaATAACCGCAGGACACCAACCCcccggatccTAGTAAGAGAGTCATCGAGGC
 aglG promotergggtctagaTTGTGACCAACAACCGCCAAGcccggatccTAAGTCGGAGTTACCGAGGAG
 aglI promoterctagaGACATCTAATCACGTGTGCTTTTTATTA GTGGGAGTAAACCTGATTTCAAggatccTTGAAATCAGGTTTACTCCCACTAA TAAAAAGCACACGTGATTAGATGTCt
 aglB promotergggtctagaTAAGTCGGAGTTACCGAGGAGcccggatccTTGTGACCAACAACCGCCAAG
Real-time RT-PCR primers
 aglF RTGTGAGGCAATCGACCTTCTCGGTCTTCTGGGTAGCCGATA
 aglG RTGAAAGTCTCCGTCGTGGTCTGTCTGTGCGAGGACACTCTC
 aglI RTACATACCCGACGGAGAGAGTGAGTGTGACGTTCTCGTGCT
 aglB RTAACCGGATGGAGTACTACGGAGGACGGTAATCCAGTGACC
 16S rRNA RTCGGGTTGTGAGAGCAAGAGGGTCGAGAAAAGCGAGGAC

Real-time RT-PCR

Real-time RT-PCR was performed essentially as described (Yurist et al., 2007) using SYBR Green PCR 2× Master Mix (Applied Biosystems, Foster City, CA). Primers (listed in Table 2) were designed using Primer Express 2.0 software (PerkinElmer Life Sciences) and employed at a final concentration of 400 nM. For PCR amplification, 20 μl reactions were subjected to 40 reaction cycles in an ABI Prism 7300 light cycler (Applied Biosystems). Primer efficiency was ascertained by drawing a standard curve based on fivefold serial dilutions of cDNA when using primers for HVO_1527,1528, 1529 and 1530 and 10-fold serial dilutions when using primers for 16S rRNA. For expression analysis of HVO_1527,1528, 1529 and 1530 in H. volcanii WR536 cells, 100 ng of cDNA were used in 20 μl PCR amplifications. For measuring the levels of 16S rRNA housekeeping gene, 0.16 ng of cDNA were used in a 20 μl reaction. Relative quantification of mRNA levels was calculated using the standard 2−ΔΔct formula.

Mass spectrometry

Mass spectrometry was performed essentially as described elsewhere (Abu-Qarn et al., 2008).

Immunoblotting

For immunoblotting, proteins were electrotransferred from SDS-PAGE gels to nitrocellulose membranes (0.45 μm, Schleicher and Schuell, Dassel, Germany) and incubated with anti-GFP (Roche) or anti-CBD (a gift from Ed Bayer, Weizmann Institute of Science, Rehovot, Israel) antibodies, each at a 1:1000 dilution. HRP-conjugated goat anti-mouse (1:2500; KPL, Gaithersburg, USA) or anti-rabbit (1:4000, Bio-Rad) antibodies, serving as secondary antibodies for binding of the anti-GFP or anti-CBD sera respectively. Detection of antibody binding was achieved using ECL Western blotting detection reagent (Amersham Biosciences, UK).

Accession numbers

The sequence of H. volcanii aglF, aglI and aglG have been deposited into the EMBL/GenBank/DDBJ databases and assigned accession number AM991128, AM991129 and AM991130, respectively.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

J.E. is supported by grants from the Israel Science Foundation (grant 30/07) and the US Air Force Office for Scientific Research (grant FA9550-07–10057). The Imperial College work was supported by the Biotechnology and Biological Sciences Research Council (grants B19088, SF19107 and BBC5196701). S.Y.D. is the recipient of a Negev-Faran Associates Scholarship.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References