N-Glycosylation Is Important for Proper Haloferax volcanii S-Layer Stability and Function

ABSTRACT

N-Glycosylation, the covalent linkage of glycans to select Asn residues of target proteins, is an almost universal posttranslational modification in archaea. However, whereas roles for N-glycosylation have been defined in eukarya and bacteria, the function of archaeal N-glycosylation remains unclear. Here, the impact of perturbed N-glycosylation on the structure and physiology of the haloarchaeon Haloferax volcanii was considered. Cryo-electron microscopy was used to examine right-side-out membrane vesicles prepared from cells of a parent strain and from strains lacking genes encoding glycosyltransferases involved in assembling the N-linked pentasaccharide decorating the surface layer (S-layer) glycoprotein, the sole component of the S-layer surrounding H. volcanii cells. Whereas a regularly repeating S-layer covered the entire surface of vesicles prepared from parent strain cells, vesicles from the mutant cells were only partially covered. To determine whether such N-glycosylation-related effects on S-layer assembly also affected cell function, the secretion of a reporter protein was addressed in the parent and N-glycosylation mutant strains. Compromised S-layer glycoprotein N-glycosylation resulted in impaired transfer of the reporter past the S-layer and into the growth medium. Finally, an assessment of S-layer glycoprotein susceptibility to added proteases in the mutants revealed that in cells lacking AglD, which is involved in adding the final pentasaccharide sugar, a distinct S-layer glycoprotein conformation was assumed in which the N-terminal region was readily degraded. Perturbed N-glycosylation thus affects S-layer glycoprotein folding. These findings suggest that H. volcanii could adapt to changes in its surroundings by modulating N-glycosylation so as to affect S-layer architecture and function.

IMPORTANCE Long held to be a process unique to eukaryotes, it is now accepted that bacteria and archaea also perform N-glycosylation, namely, the covalent attachment of sugars to select asparagine residues of target proteins. Yet, while information on the importance of N-glycosylation in eukaryotes and bacteria is available, the role of this posttranslational modification in archaea remains unclear. Here, insight into the purpose of archaeal N-glycosylation was gained by addressing the surface layer (S-layer) surrounding cells of the halophilic species Haloferax volcanii. Relying on mutant strains defective in N-glycosylation, such efforts revealed that compromised N-glycosylation affected S-layer integrity and the transfer of a secreted reporter protein across the S-layer into the growth medium, as well as the conformation of the S-layer glycoprotein, the sole component of the S-layer. Thus, by modifying N-glycosylation, H. volcanii cells can change how they interact with their surroundings.

INTRODUCTION

It is now clear that the protein content of a given cell exceeds what is encoded by the genome, with posttranslational modifications being a major source of proteomic expansion. Of the various processing events that a protein can undergo, N-glycosylation, namely, the covalent attachment of glycans to selected Asn residues of a target protein, is one of the most prevalent and likely the most complex. Indeed, it has been reported that over 50% of all eukaryal proteins are N-glycosylated (1). In eukarya, N-glycosylation has been shown to be important in a plethora of biological processes spanning from protein folding and quality control to a range of recognition events (2–7). Long thought to be restricted to eukarya, it has since become evident that archaea and bacteria also perform this posttranslational modification. In bacteria, where N-glycosylation is seemingly restricted to the deltaproteobacteria and epsilonproteobacteria (8), such protein processing is important for interactions with host cells, as well as other aspects of pathogenicity (9, 10). On the other hand, the purpose of N-glycosylation in archaea remains poorly defined, despite that fact that such posttranslational protein modification is thought to be an almost-universal trait of this domain (11).

Efforts to understand the functional significance of archaeal N-glycosylation have been largely limited by challenges in cultivating the vast majority of identified strains in the laboratory, as well as the few cultivatable strains for which genetic tools are available. As such, current insight into the importance of N-glycosylation in archaea is largely based on studies of two proteins, namely, surface layer (S-layer) glycoproteins and archaellins (12). The S-layer glycoprotein is the sole or one of a limited number of proteins comprising the S-layer surrounding many archaeal cells (13), while archaellins are the basic building blocks of the archaellum, the archaeal counterpart of the bacterial flagellum (14). N-Glycosylation of S-layer glycoproteins has been implicated in maintenance of cell shape, adaptation to environmental changes, and S-layer stability (15–17), while N-glycosylation of archaellins has been shown to be important for the appearance of archaella and for cell motility (18–20). Still, it remains unclear how N-glycosylation contributes to these traits.

With the growing understanding of pathways of N-glycosylation in several archaea, including the methanogens Methanococcus voltae and Methanococcus maripaludis, the thermoacidophile Sulfolobus acidocaldarius, and the halophile Haloferax volcanii, more-detailed insight into the role of the archaeal version of this universal posttranslational modification is possible (for a recent review, see reference 21). In each organism, Agl (archaeal glycosylation) proteins are responsible for the assembly of N-linked glycans on dolichol-based lipid carriers, from where they are delivered to target protein Asn residues. In the H. volcanii Agl pathway, the first four sugars of an N-linked pentasaccharide are sequentially added to a common dolichol phosphate (DolP) lipid carrier by the glycosyltransferases AglJ, AglG, AglI, and AglE (17, 22–24). The lipid-linked tetrasaccharide is subsequently transferred to the target protein. AglD adds the terminal sugar of the pentasaccharide to its own DolP carrier, from where it is delivered to the protein-bound tetrasaccharide (22, 25). In addition to these glycosyltransferases, other proteins involved in the pathway, such as those responsible for modification of the individual sugars, transfer of the glycan to the protein, and other steps, have also been identified (26–30).

Through the use of H. volcanii strains from which genes encoding components of the Agl pathway have been deleted, it is possible to generate cells that contain S-layer glycoproteins presenting truncated N-linked glycans. In the present report, structural, biochemical, and physiological studies addressing such mutants provided insight into the relationship between H. volcanii S-layer glycoprotein N-glycosylation and S-layer architecture, as well as cellular function.

RESULTS

S-layer integrity is compromised in membrane vesicles prepared from cells lacking Agl glycosyltransferases.

Earlier electron microscopy studies revealed that perturbation of the N-glycosylation pathway affected S-layer symmetry in H. volcanii cell envelope preparations (25). It would appear that S-layer glycoprotein N-glycosylation is important for H. volcanii S-layer structure. To better understand how N-glycosylation contributes to S-layer architecture, cryo-transmission electron microscopy (cryo-TEM) was used to consider the S-layer in preparations obtained from cells of the parent strain and from mutant strains lacking the genes encoding AglJ, AglG, AglI, AglE, or AglD, glycosyltransferases sequentially involved in assembly of the N-linked pentasaccharide decorating the S-layer glycoprotein (Fig. 1) (17, 23–25, 31). In each case, both the plasma membrane and the surrounding S-layer were clearly visible. However, because no obvious differences stood out in a comparison of the S-layer of parent strain cells (Fig. 2A) with that of any of the mutant strain cells (not shown), right-side-out (RSO) membrane vesicles were instead prepared from parent and ΔaglD, ΔaglE, ΔaglG, ΔaglI, and ΔaglJ mutant strain cells and examined by cryo-TEM so as to allow for more-focused examination of the H. volcanii cell surface.

FIG 1.

Schematic depiction of the N-linked pentasaccharide decorating the H. volcanii S-layer glycoprotein and the glycosyltransferases responsible for the addition of each sugar. The monosaccharides are depicted according to the symbol nomenclature for glycans (46). Blue circle, glucose; blue/white diamond, glucuronic acid; yellow/white diamond, galacturonic acid; green circle, mannose; Me, methyl group.

FIG 2.

The S-layer of RSO membrane vesicles prepared from H. volcanii agl pathway mutants is compromised. (A) Electron micrographs of the cell surface of an H. volcanii parent strain cell. The arrowheads denote the S-layer, while the arrows denote the plasma membrane. The bars represent 100 nm (top) or 50 nm (bottom). (B) An electron micrograph of representative membrane vesicle derived from H. volcanii parent strain cells. The scale bar represents 100 nm. (C) Representative electron micrographs of membrane vesicles prepared from H. volcanii ΔaglJ, ΔaglG, ΔaglI, ΔaglE, and ΔaglD strain cells. The scale bars represent 100 nm. (D) The percentages of intact S-layer in membrane vesicles prepared from H. volcanii ΔaglJ, ΔaglG, ΔaglI, ΔaglE, and ΔaglD strain cells, relative to parent strain-derived membrane vesicles, were calculated. In each column, the average of two repeats of the experiment (horizontal line in each column) and the distribution of values collected from the two repeats are presented.

Whereas H. volcanii cells examined by cryo-TEM assumed a variety of shapes and sizes, sometimes reaching lengths of several micrometers, the vesicle preparations were more consistent, appearing as spherical entities 200 to 400 nm in diameter (Fig. 2B and C). In the vesicle samples prepared from the parent strain, both the plasma membrane and the concentric S-layer were clearly visible (Fig. 2B and D). The S-layer appeared as a “half-zipper,” with regularly repeating and evenly spaced troughs and peaks, the length of each trough and peak pair being 13 nm and the height of such pairs being 6 nm. The S-layer was found some 22 nm beyond the plasma membrane, on average. These values are in good agreement with those reported in other studies of the H. volcanii cell surface (31) and similar to what was seen with whole-cell preparations in our study (Fig. 2A). In addition, a regularly repeating electron-dense moiety that appeared as a series of short lines of equal length under the peaks of the S-layer structure was noted.

When similar preparations obtained from the various agl deletion strains considered in this study were examined, S-layers of differing degrees of intactness were observed (Fig. 2C). In some cases, the S-layer was almost completely absent, whereas in other instances, most of the S-layer remained intact (Fig. 2C and D). For example, in RSO membrane vesicles prepared from ΔaglJ or ΔaglE mutant cells, just over 50% of the S-layer found to be associated with parent strain vesicles was observed. At the same time, vesicles prepared from ΔaglD mutant cells presented almost 80% of the intact S-layer found to be associated with parent strain vesicles. On average, coverage of the plasma membrane by an intact S-layer was between 20 and 50% lower in the different deletion strains than in the same preparation from the parent strain (Fig. 2D and Table 1). It thus appears that proper N-glycosylation of the S-layer glycoprotein is important for S-layer structural integrity. Finally, where segments of intact S-layer were detected in the deletion strain-derived vesicles, the distance between the S-layer and plasma membrane, the width of the S-layer, and the regularity of the pattern associated with the S-layer were as seen in parent strain-derived vesicles.

TABLE 1.

S-layer intactness in RSO membrane vesicles prepared from H. volcanii parent and Agl pathway mutant strains

Straina	Intact S-layer (%)b		No. of samplesc	Significance of difference (P value) from parent straind
	Mean	SEM
Parent	96.9 (96.6, 98.5)	0.7 (0.8, 1.5)	27	0
ΔaglJ mutant	53.4 (58.7, 50.8)	3.4 (4.1, 4.7)	45	<0.0001
ΔaglG mutant	57.0 (56.5, 57.4)	2.4 (3.7, 3.4)	51	<0.0001
ΔaglI mutant	71.7 (75.6, 67.0)	2.5 (2.7, 4.3)	68	<0.0001
ΔaglE mutant	51.5 (47.5, 54.7)	4.7 (9.6, 3.8)	41	<0.0001
ΔaglD mutant	79.4 (78.6, 80.1)	2.0 (3.3, 2.5)	57	<0.0001

^aThe deletion strains are listed in the order in which the sugar normally processed by the absent glycosyltransferase is added to the N-linked pentasaccharide.

^bValues in parentheses in the mean and standard error of the mean (SEM) columns represent the mean and SEM percentages of each repeat of the experiment.

^cTotal number of fields examined in two biological repeats of the experiment.

^dCalculated according to Student's t test.

Perturbed S-layer glycoprotein N-glycosylation hinders protein secretion.

Having shown that S-layer glycoprotein N-glycosylation contributes to S-layer structure, the possibility that proper S-layer glycoprotein N-glycosylation also impacts the function of this extracellular structure was addressed. As in other S-layer-bearing archaea, the H. volcanii S-layer presents a barrier that secreted proteins must traverse following their translocation across the plasma membrane. With this in mind, the impact of compromised S-layer glycoprotein N-glycosylation on H. volcanii protein secretion was considered.

In silico studies have predicted that in H. volcanii, as in other haloarchaea, a large number of secreted proteins rely on the twin-arginine translocation (Tat) pathway, rather than the generally more widely used secretory (Sec) pathway, for translocation across the plasma membrane (32, 33). Accordingly, H. volcanii parent and ΔaglD, ΔaglE, ΔaglG, ΔaglI, and ΔaglJ mutant strain cells were transformed to express green fluorescent protein (GFP) bearing the N-terminal Natronococcus sp. strain Ah36 α-amylase signal peptide (here called TatSP-GFP) (34). Earlier studies reported the ability of transformed H. volcanii to secrete this amylase via the Tat pathway (32). In the present study, the ability of the translocated GFP to traverse the S-layer and reach the growth medium when N-glycosylation was compromised was considered.

Initially, the secretion of TatSP-GFP by cells transformed with the plasmid generated here was confirmed by microscopic inspection. Analysis of parent strain cells and cells of the same strain transformed to express TatSP-GFP were examined by bright-field and fluorescence microscopy. Fluorescence was only observed with the transformed cells, with both the cells and the medium containing more fluorescence than was seen with the nontransformed cultures (Fig. 3A). Transformation of parent and ΔaglD, ΔaglE, ΔaglG, ΔaglI, and ΔaglJ mutant strain cells to express TatSP-GFP had no major impact on the protein content of the transformed cells, as reflected by the similar Coomassie-stained total protein extracts obtained from all cultures (Fig. 3B), nor on the growth of these strains (not shown). The expression of TatSP-GFP and its secretion into the growth medium were confirmed by immunoblotting using antibodies to GFP. Whereas no bands were detected in the growth medium of parent or ΔaglD, ΔaglE, ΔaglG, ΔaglI, or ΔaglJ mutant strain cells (Fig. 3C, left), the antibodies recognized a single band in the media of the same cultures transformed to express TatSP-GFP (Fig. 3C, right). Although GFP was detected in the growth medium of each transformed strain, cells lacking AglD seemed to secrete less of this reporter protein than the other strains. To assess the extent of TatSP-GFP secretion into the growth medium, the fluorescence of cell-free media obtained from cultures containing equivalent amounts of GFP-secreting cells of the parent strain and of strains lacking AglD, AglE, AglG, AglI, or AglJ was measured. The results of a representative of four similar repeats of such analysis revealed a major peak at 509 nm, corresponding to the peak of emission for GFP, in medium derived from transformed parent strain cells (Fig. 3D). When the media of cultures containing ΔaglD, ΔaglE, ΔaglG, ΔaglI, or ΔaglJ mutant cells were similarly examined, peaks at 509 nm were also detected, although they were less pronounced than those seen with the parent strain. Ever-decreasing peak heights, reflecting diminished GFP secretion, were seen in cells lacking aglJ, aglG, aglI, and aglE, in that order. Secretion of GFP from ΔaglD mutant cells was especially compromised, reflected by a 509-nm fluorescence peak similar in intensity to that of a negative control in which medium alone was considered. To quantitate the impact of deletion of Agl pathway genes on TatSP-GFP secretion, the fluorescence measured at 509 nm in cells of each deletion strain was expressed as a percentage of the same fluorescence obtained with parent strain cells. The averages of such values obtained in four repeats of the experiment are presented in Fig. 3E.

FIG 3.

Perturbation of N-glycosylation affects protein secretion. (A) Cells of the parent strain (top images), in some cases transformed to express TatSP-GFP (bottom images), were examined by bright-field microscopy (left images) and fluorescence microscopy (right images) at ×400 magnification. (B) The protein contents of aliquots of cultures of parent and ΔaglJ, ΔaglG, ΔaglI, ΔaglE, and ΔaglD mutant strain cells transformed to express TatSP-GFP were separated on 7.5% SDS-PAGE gels and Coomassie stained. The arrow on the right denotes the S-layer glycoprotein (SLG). The positions of molecular mass markers (in kilodaltons) are indicated on the left. (C) Aliquots of growth media of parent and ΔaglJ, ΔaglG, ΔaglI, ΔaglE, and ΔaglD mutant strain cells (−TatSP-GFP; left), in some cases transformed to express TatSP-GFP (+TatSP-GFP; right), were probed with anti-GFP antibodies. The positions of molecular mass markers (in kilodaltons) are indicated on the left. (D) The fluorescence of aliquots (800 μl) of the cell-free growth media of parent, ΔaglJ, ΔaglG, ΔaglI, ΔaglE, and ΔaglD mutant strain cells was examined in a fluorometer. A representative of four repeats shows the fluorescence recorded in the region of 509 nm, corresponding to the emission peak for GFP. The lines shown represent a smoothing of data points collected every 0.5 nm. The legend lists the color assigned to the medium of each strain. (E) The fluorescence detected at 509 nm in the medium of each deletion strain is provided relative to that detected in the medium of the parent strain of each repeat, which was considered 100%. The values given correspond to the mean ± standard error of 4 repeats.

N-Glycosylation changes S-layer glycoprotein conformation.

The molecular basis for the impact of perturbed S-layer glycoprotein N-glycosylation on S-layer structure and function was next considered. A series of previous reports addressing the contribution of N-glycosylation to H. volcanii S-layer stability showed the ability of added protease to more rapidly degrade the S-layer of cells lacking the genes encoding select components of the Agl pathway, relative to parent strain cells (17, 24, 25, 28, 30). Such efforts did not, however, systematically compare the effects of deleting genes encoding the different glycosyltransferases of the Agl pathway on the S-layer glycoprotein. Hence, to gain insight into how perturbed S-layer glycoprotein N-glycosylation led to compromised S-layer behavior, as reported above, aliquots of exponential-phase cultures containing equivalent amounts of parent strain or ΔaglJ, ΔaglG, ΔaglI, ΔaglE, or ΔaglD mutant cells were removed prior to and 15, 30, and 45 min after the addition of proteinase K. Following SDS-PAGE and Coomassie staining, the S-layer glycoprotein was considered in each case. Throughout the 45-min window of proteolysis, levels of the S-layer glycoprotein in the parent strain remained largely consistent (Fig. 4A). The same was true of the S-layer glycoprotein in the ΔaglJ and ΔaglI mutant strains, whereas some digestion occurred in the ΔaglG and ΔaglE mutant strains by the 45-min point. In no case were any readily discernible degradation products observed. In contrast, limited proteolysis of the S-layer glycoprotein from ΔaglD mutant cells was seen as early as after 15 min of incubation with proteinase K. Glycostaining of the proteolytic fragment generated in the ΔaglD mutant cells provided evidence that it originated from the S-layer glycoprotein (Fig. 4B). To assess whether the limited proteolysis of the S-layer glycoprotein from ΔaglD mutant cells had occurred at the N- or C-terminal end of the protein, immunoblotting was performed using antibodies raised against the 13 N-terminal residues of the protein (35). Whereas labeling of the full-length protein was readily seen, the proteolytic fragment was not recognized, pointing to the loss of a stretch of residues beginning at the N-terminal end of the protein (Fig. 4C).

FIG 4.

The S-layer glycoprotein modified by an N-linked glycan lacking the final sugar of the pentasaccharide assumes a unique conformation. (A) Aliquots of parent and ΔaglJ, ΔaglG, ΔaglI, ΔaglE, and ΔaglD mutant strain cells were removed prior to and 15, 30, and 45 min after the addition of 1 mg/ml proteinase K, separated by SDS-PAGE, and Coomassie stained. The region of the S-layer glycoprotein is shown in each panel. In the ΔaglD mutant sample (bottom right image), a unique proteolytic fragment was detected (arrowhead). (B) Aliquots of parent and ΔaglD mutant strain cells treated as in panel A were glycostained with periodic acid-Schiff's reagent. (C) Aliquots of parent and ΔaglD mutant strain cells treated as in panel A were probed by immunoblotting using antibodies against the N-terminal 13 residues of the S-layer glycoprotein. At the left of each panel, the positions of molecular mass markers (in kilodaltons) are indicated. Where present on the right side of the panels, the arrowhead indicates the position of the S-layer glycoprotein-derived proteolytic fragment.

As such, it would appear that modification of the S-layer glycoprotein by only the first four sugars of the N-linked pentasaccharide resulted in a more protease-susceptible conformation than that assumed by the same protein modified by the complete pentasaccharide or by N-linked glycans containing three or fewer sugars.

DISCUSSION

In eukarya, the essential process of N-glycosylation affects the vast majority of proteins that enter the secretory pathway (1), where proper N-glycosylation is important for correct protein folding and oligomerization. At the same time, quality control systems can recognize aberrant proteins in the secretory pathway on the basis of their N-glycosylation profile, targeting them for either rescue or degradation (5). Later, N-glycosylation of eukaryal proteins helps determine their ultimate subcellular localization and may also contribute to cellular recognition or related interactions (2, 7). In archaea, where N-glycosylation seems to be an almost universal posttranslational modification (11, 36), the roles played by such protein processing remain to be defined. What is clear is that many of the roles assumed by N-glycosylation in eukarya are not relevant for archaea. For instance, N-glycosylation takes place on the outer surface of the archaeal cell, beyond the reach of the site of protein folding in the cytosol. Moreover, these cells do not contain subcellular compartments. As such, it is reasonable to assume that in archaea, N-glycosylation serves roles specific to this domain of life. In the present study, such roles were considered in the haloarchaeon H. volcanii.

In H. volcanii, the S-layer surrounding the cell is composed of a single component, the S-layer glycoprotein (37). Moreover, the pathway responsible for the assembly and attachment of the pentasaccharide (comprising mannose-1,2-[methyl-O-4-]glucuronic acid-β-1,4-galacturonic acid-α-1,4-glucuronic acid-β1,4-glucose-β1-Asn [38]) known to modify the S-layer glycoprotein at four of the seven putative N-glycosylation sites in the protein has been defined (39). As such, the H. volcanii S-layer glycoprotein represents an ideal system to address the functional role(s) of N-glycosylation. In the present study, perturbed N-glycosylation was shown to compromise S-layer integrity, reflected in the loss of an intact S-layer surrounding RSO membrane vesicles from cells lacking genes encoding glycosyltransferases involved in N-linked pentasaccharide biogenesis. Where fragments of the S-layer were maintained in such vesicles, proper assembly seemingly had occurred. Indeed, when the S-layers of intact cells from the parent and deletion strains were considered, no apparent effect of compromised S-layer glycoprotein N-glycosylation was apparent (not shown). As such, it would appear that S-layer glycoprotein N-glycosylation is not as important for proper S-layer assembly as it is for robustness of this structure in the face of physical challenges, including those encountered when preparing RSO membrane vesicles. Of course, it remains possible that higher-resolution microscopic approaches will reveal effects of perturbed S-layer glycoprotein N-glycosylation on S-layer architecture not detectable at the resolution considered in the present study. It should be noted that the importance of N-glycosylation for proper archaeal S-layer assembly was also proposed in earlier studies on Halobacterium salinarum (15) and Sulfolobus acidocaldarius (40).

In addition to the effect on S-layer stability, compromised S-layer glycoprotein N-glycosylation was also shown to affect S-layer function. Specifically, the role of the S-layer as a barrier encountered by proteins translocated across the plasma membrane en route to the growth medium, and, by analogy, access of proteins and other molecules in the extracellular milieu to the plasma membrane, was considered in response to perturbed S-layer glycoprotein N-glycosylation. By following the secretion of a reporter protein into the growth medium, the importance of proper N-glycosylation was demonstrated. It was shown that as the precursor of the N-linked pentasaccharide decorating the S-layer glycoprotein lengthened from one to four sugars, there was an increasingly detrimental impact on GFP secretion, with the most pronounced effect being noted in cells lacking AglD, namely, where only the first four sugars of the N-linked pentasaccharide are added. The impact of aglD deletion on S-layer function could be related to the change in conformation of the S-layer glycoprotein in cells lacking this glycosyltransferase, revealed by the distinct protease sensitivity of the protein in ΔaglD mutant cells. At this point, it is not clear why the absence or presence of the final N-linked glycan sugar, mannose, should have such a profound impact on the S-layer glycoprotein. One possibility is that the fifth sugar somehow masks charges associated with other sugars in the N-linked glycan. Regardless, the N-glycosylation-related change in S-layer glycoprotein conformation could affect the hexameric assembly of S-layer glycoproteins that correspond to the repeating complex that comprises the H. volcanii S-layer (31), either by modulating the assembly of the hexameric unit or in interactions between individual hexamers. Either scenario could affect the size of S-layer pores through which proteins and other molecules traverse this structure. Finally, it should be noted that the augmented sensitivity of the ΔaglD mutant cells to proteinase K, relative to parent strain cells, seen here is in contrast to the enhanced resistance of the mutant cells to trypsin reported earlier (25). These differences could be related to the substrate preferences of each protease and the effects of differential N-glycosylation on access to cleavage sites.

In the present study, the contribution of S-layer glycoprotein N-glycosylation to H. volcanii S-layer structure and function was addressed. In agreement with earlier efforts addressing the impact of compromised N-glycosylation on the functions of other H. volcanii glycoproteins (18, 41), perturbed S-layer glycoprotein N-glycosylation was shown to have physiological implications. Still, heterogeneity in S-layer glycoprotein N-glycosylation may be responsible for more-subtle effects than suggested here (38). Future efforts will be directed at delineating conditions that modulate such heterogeneity so as to better understand how H. volcanii can respond to the changing demands of its environment through the rapid and reversible usage of posttranslational protein modification.

MATERIALS AND METHODS

Strains and growth conditions.

H. volcanii WR536 (H53) strain cells and the same strain from which aglD, aglE, aglG, aglI, or aglJ was deleted were grown in complete 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) (42) at 42°C. Preparation of the various deletion strains was described previously (17, 23, 24, 43). Novobiocin (1 μg/ml) was added to the growth medium of plasmid-transformed cells.

Plasmid construction.

To generate H. volcanii cells expressing GFP, cells were transformed with plasmid pJAM-1020I, a plasmid based on pJAM1020 (44), obtained from J. Maupin-Furlow, University of Florida. To create plasmid pJAM-1020I, NdeI and BglII cloning sites were introduced into pJAM-1020 between the GFP gene promoter and coding sequences, using the primers listed in Table 2. To generate H. volcanii cells capable of secreting GFP synthesized with a Tat pathway signal peptide (TatSP), DNA from the TatSP-encoding region of Natronococcus sp. strain Ah36 α-amylase (34) found in plasmid pAMY-RR, obtained from Mechthild Pohlschroder, University of Pennsylvania, was PCR amplified using primers (Table 2) designed to introduce NdeI and BglII cloning sites. The amplified product was ligated into pJAM-1020I predigested with the same restriction enzymes to generate plasmid pJAM-TatSP-1020I. The H. volcanii parent strain and ΔaglD, ΔaglE, ΔaglG, ΔaglI, and ΔaglJ mutant cells were transformed with plasmid pJAM-TatSP-1020I.

TABLE 2.

Primers used in this study

Primer	Sequence	Purpose
1020-NdeI	TTCCATGGCCAACACTTGTCACTACTTTCACTTATGGTGTTCAATGCTTTTCAAGATACCCAGATCACATGAAG	Eliminate NdeI site in plasmid pJAM-1020
1020-StuI	CTGAACAGGCCTAGGATAGCCC	Reverse primer
1020-NdeI-BglII	AACGGATCCTAGAAATAATTTTGTTAACTTTAAGAAGGAGATATACACAAGGAGATATAACATATGGGCAGATCTATGAGTAAAGGAGAAGAACTTTTC	Introduce NdeI and BglII sites into plasmid pJAM-1020 to generate plasmid pJAM-1020I
TatSP-NdeI	GGACATATGCGACGGAATCACAGCC	Introduce NdeI site into TatSP sequence
TatSP-BglII	CTCAGATCTGGCGGCCGCACTGGTCG	Introduce BglII site into TatSP sequence

Cryo-TEM of RSO membrane vesicles.

RSO membrane vesicles were prepared from H. volcanii parent strain and ΔaglD, ΔaglE, ΔaglG, ΔaglI, and ΔaglJ mutant cells, essentially as described previously (31). Cultures of each strain were grown to mid-exponential phase (optical density at 550 nm [OD₅₅₀], 1.8), and 1-ml aliquots were removed. After harvesting of the cells (11,500 × g for 20 s in a microcentrifuge), the supernatant was removed, and the cells were incubated for 1 min in 1 ml of vesicle buffer (2.14 M NaCl, 250 mM MgCl₂). The cells were again harvested, the supernatant was discarded, and the pellet was resuspended in 500 μl of vesicle buffer and frozen in liquid nitrogen. The frozen sample was thawed at room temperature, and 5 μl of DNase (1 mg/ml) was added. After 1 h of incubation at 37°C, the sample was centrifuged (11,500 × g for 15 s in a microcentrifuge), and the supernatant was transferred to a new microcentrifuge tube. Following centrifugation (11,500 × g for 7 min in a microcentrifuge), the supernatant was removed, and the pellet was resuspended in 750 μl of vesicle buffer and centrifuged again. The supernatant was removed, and the pellet was resuspended in 200 μl of 10 mM CaCl₂. Aliquots (2 μl) were applied onto nickel grids, which were then plunged into liquid ethane using a Leica EM GP apparatus and manually transferred into liquid nitrogen, where they were stored until viewed. The preparations were examined in a 120 kV FEI Tecnai T12 transmission electron microscope using a charge-coupled-device (CCD) camera, and the data were analyzed using ImageJ software (NIH). The degree of S-layer intactness in the RSOs was calculated as the length of S-layer defined as a percentage of the circumference of the S-layer-based edge of the vesicle. The data reported were collected from at least two biological repeats, and tens of images were analyzed for each.

Proteolytic digestion of the S-layer.

H. volcanii cells (1 ml) of the parent and the aglD, aglE, aglG, aglI, and aglJ deletion strains were grown to an OD₅₅₀ of 1.0 and challenged with proteinase K (1 mg/ml, final concentration) at 37°C. Aliquots (100 μl) were removed immediately prior to incubation with proteinase K (considered the t = 0 point) and at 15-min intervals following the addition of the protease for up to 45 min. Following separation on 7.5% SDS-PAGE gels, the proteins were Coomassie stained, glycostained with periodic acid-Schiff's reagent (45), or transferred to nitrocellulose membranes (0.45-μm pore size; Schleicher & Schuell, Dassel, Germany) and probed with antibodies raised to a peptide corresponding to the 13 N-terminal residues of the H. volcanii S-layer glycoprotein (35).

Fluorescence microscopy.

A 5-μl drop of parent or ΔaglD, ΔaglE, ΔaglG, ΔaglI, or ΔaglJ mutant strain culture on a microscope slide covered with a cover slide was examined in an Olympus FV1000 confocal microscope using a 60× 1.35-numerical aperture (NA) lens, with excitation by laser at 488 nm.

Measurement of fluorescence in the growth medium.

The growth media of 20-ml cultures (OD₅₅₀, 2.0) of H. volcanii cells of the parent and ΔaglD, ΔaglE, ΔaglG, ΔaglI, and ΔaglJ mutant strains transformed to express TatSP-GFP was collected by centrifugation (13,000 × g, 3 min). The supernatant was removed and passed through a 0.22-μm-pore-size filter. The fluorescence of an aliquot (800 μl) was analyzed using a PerkinElmer LS55 fluorescence spectrometer with excitation set at 480 nm. Measurements were taken from 480 to 600 nm in 0.5-nm steps.

Immunoblotting.

The protein contents of aliquots (100 μl) of 20-ml cultures (OD₅₅₀, 2.0) of H. volcanii cells of the parent and ΔaglD, ΔaglE, ΔaglG, ΔaglI, and ΔaglJ mutant strains, either untransformed or transformed to express TatSP-GFP, were precipitated with 15% trichloroacetic acid (TCA) and left on ice for 30 min. Alternatively, the growth medium of each of the same 20-ml cultures was collected by centrifugation (13,000 × g, 3 min) and passed the through a 0.22-μm-pore-size filter. The protein content of aliquots (100 μl) of the medium was precipitated with 15% TCA and left on ice for 30 min. Following centrifugation in a microcentrifuge (10,600 × g, 15 min, 4°C), the supernatants of the TCA-treated samples were removed, and the pellets were washed with ice-cold acetone. After a second round of centrifugation in a microcentrifuge (10,600 × g, 15 min, 4°C), the supernatants were discarded and the pellets were air-dried. SDS-PAGE sample buffer was added, and proteins were separated on 15% SDS-PAGE gels. The separated proteins were electrotransferred to nitrocellulose membranes (0.45-μm pore size; Schleicher & Schuell, Dassel, Germany) and probed with horseradish peroxidase-conjugated goat anti-GFP antibodies (1:500 dilution; Rockland Immunochemicals, Limerick, PA). Antibody binding was detected using an enhanced chemiluminescence kit (Advansta, Menlo Park, CA).

Statistics.

Data were analyzed using GraphPad QuickCalcs software, available online at http://www.graphpad.com/quickcalcs/index.cfm, or KaleidaGraph version 4.1 (Synergy Software, Reading, PA).