|Home | About | Journals | Submit | Contact Us | Français|
Cryptococcus neoformans is an encapsulated pathogenic fungus. The cryptococcal capsule is composed of polysaccharides and is necessary for virulence. It has been previously reported that glucuronoxylomannan (GXM), the major capsular component, is synthesized in cytoplasmic compartments and transported to the extracellular space in vesicles, but knowledge on the organelles involved in polysaccharide synthesis and traffic is extremely limited. In this paper we report the GXM distribution in C. neoformans cells sectioned by cryoultramicrotomy and visualized by transmission electron microscopy (TEM) and polysaccharide immunogold staining. Cryosections of fungal cells showed high preservation of intracellular organelles and cell wall structure. Incubation of cryosections with an antibody to GXM revealed that cytoplasmic structures associated to vesicular compartments and reticular membranes are in close proximity to the polysaccharide. GXM was generally found in association with the membrane of intracellular compartments and within different layers of the cell wall. Analysis of extracellular fractions from cryptococcal supernatants by transmission electron microscopy in combination with serologic, chromatographic and spectroscopic methods revealed fractions containing GXM and lipids. These results indicate an intimate association of GXM and lipids in both intracellular and extracellular spaces consistent with polysaccharide synthesis and transport in membrane-associated structures.
The most prominent feature of the fungal pathogen Cryptococcus neoformans is the presence of a capsule surrounding the yeast cell body. This structure is composed mainly of polysaccharides, including glucuronoxylomannan (GXM) and galactoxylomannan (GalXM) (Zaragoza et al., 2009). The cryptococcal capsule is crucial for fungal cell survival in different situations, including environmental interactions with predators (Steenbergen et al., 2001) or evasion of the immune system during human and animal infections (Monari et al., 2008; Vecchiarelli, 2007). GXM is considered the most important virulence factor of C. neoformans (McClelland et al., 2005) and, for this reason, it is also the most well studied structural component.
GXM biosynthesis involves several enzymatic steps, starting with sugar polymerization using nucleotide-activated sugar monomers as precursors (Wills et al., 2001; Zaragoza et al., 2009). In contrast to cell wall polysaccharides like chitin and glucans, which are presumably synthesized at plasma membrane level, GXM epitopes are found at intracellular sites (Feldmesser et al., 2001; Garcia-Rivera et al., 2004; Yoneda and Doering, 2006), that presumably reflect the sites of initial assembly. Mutation in the gene coding for a Rab/GTPase ortologue of the Saccharomyces cerevisiae Sec4p caused accumulation of cytoplasmic GXM-containing vesicles (Yoneda and Doering, 2006), raising the possibility that newly synthesized polysaccharides might be exported to the cell surface by vesicular traffic. Since Sec4p is involved in post-Golgi secretion events, it was deduced that GXM was at least partially synthesized at the Golgi apparatus (Yoneda and Doering, 2006). GXM-containing vesicles are released to the extracellular milieu (Rodrigues et al., 2008; Rodrigues et al., 2007) in a process that apparently involves the SEC6 gene (Panepinto et al., 2009). Little is known about the biogenesis of extracellular GXM-containing vesicles. Although different studies suggest a role of genes related to conventional post-Golgi secretion in vesicle release (Panepinto et al., 2009; Yoneda and Doering, 2006), proteomic analysis of isolated C. neoformans vesicles reveled similarities with mammalian exosomes (Rodrigues et al., 2008).
Our knowledge on the steps of GXM biosynthesis and cellular traffic is limited by the paucity of information on its intracellular distribution. Previous ultrastructural analysis of C. neoformans cells by immunoelectron microscopy (IEM) revealed that antibodies raised against GXM recognize intracellular sites (Feldmesser et al., 2001; Garcia-Rivera et al., 2004; Yoneda and Doering, 2006; Yoneda and Doering, 2009). Conventional IEM procedures, however, include several steps of dehydration and embedding with different resins, which markedly affects preservation of the cell envelope and intracellular structures and organelles, and generates structural artifacts. Such subcellular modifications usually include loss of cytoplasmic content, poor preservation of membranous structures and cell wall shrinkage (Yamaguchi et al., 2005).
Cryoimmunogold labeling has been successfully used in different studies of the C. neoformans biology (Rodrigues et al., 2008; Rodrigues et al., 2007; Rodrigues et al., 2000). Using this technique, in which cells are promptly fixed, frozen and sectionated, the artifacts that can plague conventional IEM are reduced or avoided, and the result is superior preservation of subcellular structures, such as the nucleus, vacuoles and mitochondria. In addition, cryosectioning has been suggested to be the electron microscopy technique that is best suited for preserving cellular epitopes (Slot and Geuze, 2007).
In a previous study by our group, we used a cryoultramicrotomy protocol to study the cellular distribution of a glycolipid antigen in C. neoformans (Rodrigues et al., 2000). Well preserved cells were obtained, which allowed us to examine in detail different cellular structures, including the cell wall and membranous structures. In the present work, that method was used in association with vesicle isolation and lipid analysis for the study of the distribution of GXM in C. neoformans. Our data revealed that the polysaccharide is associated with lipid matrices and previously unknown subcellular and extracellular structures, indicating that GXM traffic in C. neoformans involves still unclear cellular events that may include unconventional steps of secretion.
C. neoformans strain H99 was grown in minimal medium composed of dextrose (15 mM), MgSO4 (10 mM), KH2PO4 (29.4 mM),glycine (13 mM), and thiamine-HCl (3 µM). Fungal cells were cultivated with shaking in Erlenmeyer flasks (volume) at 30°C for 72 h. Cells were collected by centrifugation and then processed for IEM.
Electron microscopy procedures were based on the studies by Tokuyasu (Tokuyasu, 1973) and Rodrigues and colleagues (Rodrigues et al., 2000). The cells were fixed in 0.1 M sodium cacodylate buffer (pH 7.2) containing 4% paraformaldehyde, 0.2% glutaraldehyde, and 1% picric acid for 60 minutes at room temperature. The cells were then washed three times in PBS, infiltrated in gelatin, cut into cubes of 1mm and infiltrated in 25% polyvinylpyrrolidone (PVP) and 2.3 M sucrose overnight in a cold room. The polyvinylpyrrolidone embedded blocks were then mounted on cryoultramicrotome stubs and flash frozen by immersion in liquid nitrogen. After trimming, ultrathin cryosections were obtained at a temperature of −130°C using an Ultracut UCT cryo-ultramicrotome (Reichert). Flat ribbons of sections were shifted from the knife edge with an eyelash and picked up in a wire loop filled with a drop of 1% (w/v) methyl cellulose, 2.3 M sucrose in PBS buffer. Sections were thawed on the pickup droplet and transferred, sections downwards, to Formvar carbon-coated nickel grids.
For immunolabelling, grids were passed over a series of droplets of washing and were blocked in PBS-bovine serum albumin 1%. After blocking the cryosections were incubated overnight in the presence of 1 µg/ml of the murine monoclonal antibody (mAb) 18B7, washed again and incubated for one hour with 15nm (particle size) gold-labeled anti-mouse IgG. In control systems, the cryosections were incubated with PBS instead of the Mab 18B7, followed by the gold-labeled anti- mouse IgG. After a final washing series in PBS and distilled water, the sections were left for 15 min on 2% uranyl acetate (aqueous) and transferred to 0.75% methyl cellulose droplets. After 30 seconds, the grids were looped out, the excess viscous solution was drained away and the sections were allowed to dry. After drying, specimens were observed in a JEOL 1200EX transmission electron microscope operating at 80kV.
Transmission electron microscopy was used to analyze the morphology and the integrity of extracellular vesicles isolated from culture supernatants. Pellets obtained after centrifugation of cell-free supernatants at 100,000 g (see Vesicle isolation and fractionation below) were fixed with 2% glutaraldehyde in 0.1 M cacodylate at room temperature for 2 h, and then incubated overnight in 4% formaldehyde, 1% glutaraldehyde, and 0.1% PBS. The samples were incubated for 90 min in 2% osmium, serially dehydrated in ethanol, and embedded in Spurrs epoxy resin. Thin sections were obtained on a Reichart Ultracut UCT and stained with 0.5% uranyl acetate and 0.5% lead citrate. Samples were observed in a JEOL 1200EX transmission electron microscope operating at 80 kV.
Isolation of extracellular vesicles was based on a protocol described by Rodrigues et al. (Rodrigues et al., 2007). Cell-free culture supernatants were obtained by sequential centrifugation at 5,000 and 15,000 g (15 min, 4°C) for removal of cells and debris. These supernatants were concentrated approximately 20-fold using an Amicon ultrafiltration system (cutoff 100 kDa). The concentrate was again sequentially centrifuged at 4,000 and 15,000 g (15 min, 4°C) and the remaining supernatant was then ultracentrifuged at 100,000 g for 1 h at 4°C. The supernatant was discarded, and the pellet was washed by five sequential suspension and centrifugation steps, each consisting of 100,000 g for 1 h at 4°C with 0.1 M Tris-buffered saline. To remove extravesicular GXM contamination, vesicles were subjected to passage through a column packed with cyanogen bromide-activated Sepharose coupled to a monoclonal antibody to GXM, as described previously (Rodrigues et al., 2007). Ultracentrifugation pellets were finally fractionated in Optiprep gradients, as previously described for the analysis of virions and mammalian exosomes (Cantin et al., 2008). Optiprep gradients were prepared in PBS as 11 steps of 300 µl in 1,2% increments ranging from 6 to 18%. Vesicles were layered onto the top of the gradient and ultracentrifuged at 250,000 g for 75 minutes. Eleven fractions of the gradient were collected from top to bottom. Each fraction was then analyzed for the presence of sterols and GXM.
The pellets obtained from centrifugation of cell supernatants at 100,000 g were first suspended in methanol, and then two volumes of chloroform were added. The mixture was vigorously vortexed and centrifuged to separate immediately formed polysaccharide-containing precipitates. The supernatant was removed, dried with a nitrogen stream, and partitioned according to Folch et al. (Folch et al., 1957). The lower phase, containing neutral lipids, was recovered and dried by vacuum centrifugation. The dry residue was then prepared for phospholipid analysis by combined liquid chromatography-tandem mass spectrometry (LC/MS/MS) or sterol analysis by thin-layer chromatography (TLC). For sterol analysis, the lipid extract was loaded into TLC silica plates (Si 60F254s; LiChrospher, Germany) and separated using a solvent system containing hexane: ether: acetic acid (80:40:2, vol:vol:vol) solvent. The plate was sprayed with a solution of 50 mg ferric chloride (FeCl3) in a mixture of 90 ml water, 5 ml acetic acid, and 5 ml sulfuric acid. Sterol spots were identified by the appearance of a purple color after heating the plates at 100°C for 3 to 5 min. For phospholipid identification, the dry residue was suspended in 90% methanol. After centrifugation, aliquots of the supernatant (typically 100 µl) were injected onto a reverse phase HPLC column (Supleco Ascentis® Express, C18, 150 × 2.1 mm) equilibrated in buffer A (methanol/water, 95/5, v/v, containing 1 mM ammonium acetate), and eluted (100 µl/m) with an increasing concentration of buffer B (chloroform/water, 500/0.2, v/v, containing 1 mM ammonium acetate; min/%B, 0/0, 5/0, 55/100). The effluent from the column was split, with a portion (about 70%) directed to a fraction collector (1 min fractions) and the remainder (about 30%) passed to an Ionspray® source connected to a triple quadrupole mass spectrometer (PE Sciex API III+) operating in the precursor (parent) scan mode in which Q1 was scanned (0.3 Da step, 6.7 secs/scan, orifice 50 volts), the collision chamber was flooded with argon gas (collision gas thickness setting at 120), and Q3 was set to transmit ions at m/z 184.1 Da.
To confirm assignments made from the LC/MS/MS data, and to further define the composition of the phosphocholine lipids in the sample, positive and negative ion mass spectra were recorded off-line by direct injection of aliquots of fractions collected during LC/MS/MS. Fractions 24–29 were pooled, and aliquots (20 µl) were injected into a stream of methanol/chloroform (4/1, v/v; 50 µl/m) entering the Ionspray® source of the mass spectrometer, before and following the addition of formic acid (1 µl, positive ion mode) and triethylamine (1 µl, negative ion mode). These experiments were done with the mass spectrometer operating in either the normal (Q1 scan range m/z 200–1500, orifice 70 volts, 4.6 secs/scan), or the tandem mass spectrometric modes (precursor and fragment scan modes). In particular, negative ion fragment scans at high orifice voltage (120) and relatively high collision gas density (CGT setting 150), were used to assign phosphocholine fatty acid substituents after the addition of triethylamine. These experiments were done using methods originally described by Jensen et al (Jensen et al., 1986). Putative parent ions for these experiments were assigned from the results of the LC/MS/MS parent ion data, and confirmed by the results from the direct injection analyses. The parent ions were assigned as 16 Da lighter than that determined experimentally, with the loss of 15 Da predicted to result from source fragmentation at high orifice voltage (loss of one nitrogen methyl group), and 1 Da loss resulting from abstraction of the proton attached to the phosphorous oxygen under basic conditions. Data was analyzed with instrument-supplied software (MacSpec version 3.3), and optimal conditions for the experiments were established using authentic di-palmitoylphosphatidyl choline, heptadecanoyl-lyso-phosphatidyl choline and di-heptanoylphosphatidyl choline standards.
A 96-well polystyrene plate was coated with the capture IgM MAb 2D10 (10 µg/ml) for 1 h and then blocked with 1% bovine serum albumin for. The plates were then incubated with GXM fractions from vesicle samples, obtained by precipitation with a mixture of chloroform and methanol (see item Lipid analysis). After 3 washing cycles with TBS 0, 05% Tween 20, the plates were incubated with an IgG1 to GXM (MAb 18B7, 2 µg/ml) for 1 h at room temperature. Positive reactions were developed by incubation with an alkaline phosphatase-conjugated anti-IgG1 followed by reaction with p-nitrophenyl phosphate disodium hexahydrate for 30 min. Absorbance values were measured at 405 nm in a microplate reader.
Electron microscopy procedures for the study of the C. neoformans biology are complex because this fungal species is surrounded by a polysaccharide capsule, which is attached to a thick cell wall. The combination of a large capsule and thick cell wall can slow fixation process resulting in cellular degradation. Hence, obtaining high quality micrographs of cryptococcal cells can take some trial and error to identify the best conditions for fixation. In this context, the commonly used protocols for immunogold analysis of C. neoformans usually include particular steps, such as partial digestion of the cell wall (Yoneda and Doering, 2006), to ensure a more efficient process of embedding and consequent preservation of intracellular compartments. Unfortunately, these procedures are often accompanied by loss of surface components and, consequently, produce micrographs that convey limited information on the cryptococcal cellular anatomy (Yamaguchi et al., 2005). The advantages of using cryosectioning techniques have already been demonstrated in the model yeast Saccharomyces cerevisiae, where protocols based on the method of Tokuyasu (Tokuyasu, 1973) resulted in high resolution images with well preserved cellular structures and efficient immunogold labeling (Griffith et al., 2008). The general procedures to prepare biological samples according to this method include fixation, infiltration/cryoprotection and rapid freezing in liquid nitrogen.
In the current study cryoultramicrotomy allowed the generation of well-preserved C. neoformans sections (Figure 1). Intracellular compartments such as the nucleus, vacuoles, mitochondria, cell wall and capsule were clearly distinguishable. Even in the absence of a post-fixation step with OsO4, lipid bilayers were well preserved and easily discernible, as evidenced in nuclear compartments and mitochondrial lamellas (Figure 1). In addition to the better known organelles, other subcellular structures were evident, including reticular membrane clusters close to the plasma membrane (Figure 1C, black arrows). These compartments showed morphological similarities to structures derived from the endoplasmic reticulum, although their peripheral position inside the cell suggested early endosomes (Gould and Lippincott-Schwartz, 2009). A membranous compartment that resembled the endoplasmic reticulum was also visible in the proximity of nucleus (Figure 1C). Finally, vesicular structures manifesting electron dense contents were also apparent (Figure 1B). These structures had similarities to intermediate/late endosomal compartments called acidosomes (Allen et al., 1993) and also to mammalian melanosomes at different maturation stages (Raposo and Marks, 2007). These compartments, which were consistently observed in our preparations, are shown in more detail in Figure 2. Similar intracellular structures that contained GXM were recently described by Yoneda and Doering using a secretion-deficient mutant of C. neoformans (Yoneda and Doering, 2009).However, the electron dense compartments observed in our model do not display such intense GXM labeling (Figure 2), as described in that study for C. neoformans cells. In our model, some of the intracellular compartments also contained internal vesicles (Figure 1B and C) that show similarities to the so-called multivesicular bodies (MVBs) found in mammalian cells (Gruenberg and Stenmark, 2004). The existence of MVBs in C. neoformans was previously suggested using different techniques (Rodrigues et al., 2008; Takeo et al., 1973), although differences in morphology and electron density were noticed by comparing the currently described results with those from previous studies.
Intracellular sites manifesting immunogold staining with mAb to GXM were usually found in association with membranous structures (Figure 3). Quantification of immunogold labeling in intracellular compartments revealed that approximately 70% of the antibody-binding sites were membrane-associated. On the basis of this reactivity we tentatively assigned these areas as sites of GXM synthesis/transport. Membrane compartments resembling vesicles carrying GXM were repeatedly observed (Figure 3), as well as structures resembling mammalian MVBs (Figure 3C) and peripheral reticular clusters resembling endosomes (Figure 3E). Since GXM staining was often found in close association with the lipid bilayer, we hypothesize that polysaccharide-binding proteins could be inserted in these lipid bilayers functioning as transient GXM anchors, which still remains to be investigated. Considering that the size of the complex gold-labeled antibody to GXM may in some instances exceed the diameter of those structures, we cannot rule out the possibility that they are located in neighboring intracellular regions, although previous data indeed demonstrate that GXM associated with secretory vesicles (Casadevall et al., 2009; Feldmesser et al., 2001; Garcia-Rivera et al., 2004; Rodrigues et al., 2007; Yoneda and Doering, 2006).
The intracellular, electron dense compartments described by Yoneda and Doering (Yoneda and Doering, 2009) were called sav1 bodies, since they were most prominently observed in C. neoformans cells with defective expression of sav1p, a putative small GTPase involved in post-Golgi secretion. The compartments observed in our model are apparently distinct from the sav1 bodies, since they did not manifest significant reactivity with an antibody to GXM, despite morphological similarities (Figure 2, Figure 3 A and D). The presence of GXM in vacuoles (Figures 3 F) could be related to secretion mechanisms initiated by endocytosis or to mechanisms of GXM recycling. In fact, the GXM catabolic process is poorly understood and so far hydrolytic enzymes able to cleave GXM have not been reported in C. neoformans (Bose et al., 2003; Zaragoza et al., 2009).
Conventional procedures for TEM usually generate artificially compacted cell wall structures, probably because of dehydration steps. In our model, as well as in a previous study, cryoultramicrotomy procedures allowed the observation of structures that illustrate the thickness and complexity of the cell wall. As suggested in previous studies with fungal cells (Soragni et al., 2001), clear distinctions between regions of different electron densities were observed (Figure 1 B, C and 3 F), which may be related to the description of secondary cell walls during stationary growth of C. neoformans (Farkas et al., 2009). Most of the cellular sites recognized by the mAb to GXM were found in association with the cell wall or within the capsular region (Figure 1 C), corroborating with previous observations (Yoneda and Doering, 2006).
GXM has been proposed to be exported to the extracellular space in secretory vesicles, suggesting an association of lipids and polysaccharides in C. neoformans (Rodrigues et al., 2007). This notion is supported by our current observations in the cryptococcal cytoplasm. Glucosylceramide and sterols were previously characterized as membrane components of the C. neoformans vesicles (Rodrigues et al., 2007), but other lipid components of GXM-containing vesicles are still unknown. To evaluate the relationship of lipids with GXM export, we first tested whether extracellular vesicle samples contained GXM by ELISA. Vesicles obtained from encapsulated cells reacted with the antibody to GXM, while vesicles from an acapsular mutant manifested a much lower reactivity (data not shown). These results were in agreement with those described by Rodrigues and co-workers (Rodrigues et al., 2007) and support the hypothesis that the major capsular polysaccharide of C. neoformans associates with lipids for extracellular export. Since phospholipids are major membrane constituents in most cells, but also in extracellular vesicles of the fungal pathogen H. capsulatum (Albuquerque et al., 2008), we evaluated whether the crude fraction of GXM-containing vesicles contained this class of molecules.
To ensure that lipid characterization did not include membrane fractions from non-vesicular structures, the integrity and morphology of 100,000g fractions isolated from culture supernatants was analyzed by TEM. Vesicles in these fractions showed typical bilayered membranes that generally formed round compartments in the size range of 50 to 300 nm (Figure 4A), which is in agreement with previous reports on the morphological characterization of fungal extracellular vesicles (Rodrigues et al., 2008; Einsenman et al., accepted for publication). These fractions were then used for phospholipid extraction.
Cryptococcal phospholipids include phosphatidylinositol, lysophosphatidyl ethanolamine, cardiolipin, glycophospholipids, lysophosphatidyl choline, phosphatidic acid, phosphatidyl ethanolamine and phosphatidyl choline (PC) (Rawat et al., 1984). PC was the major phospholipid component of C. neoformans cellular membranes (Rawat et al., 1984), so this molecule was selected as a potential candidate component of cryptococcal vesicles. Samples were subjected to precursor ion scanning of the m/z 184.1 fragment. This ion corresponds to phosphorylcholine, and therefore represents a molecular marker for the presence of PC (Rawat et al., 1984). Control samples consisted of LCMS analysis of the solvent alone, which showed no significant peaks. From the vesicle samples, three strong peaks were observed eluting between 24–29 minutes (Fig. 7A). These peaks revealed parent masses of 782.7, 784.6 and 786.8 Da (Fig. 7B - D). LCMS analysis of two different samples produced essentially identical results (data not shown). Fractions eluting between 24–29 minutes were pooled and examined by direct flow-injection MS and MS/MS. Fragmentation of the peak at m/z 782.7 generated a strong signal at m/z 184 (data not shown), confirming the occurrence of PC. To analyze fatty acid components, negative MS/MS analysis was performed after addition of triethylamine. Analysis of the peaks at m/z 782.7 and 784.6 revealed the presence of related fragments at m/z 279.2 (Fig. 7E) and 281.0 (not shown), consistent with the presence of, respectively, fatty acids corresponding to C18:2 and C18:1. The peak at m/z 786.8 did not yield a significant response probably because of limitations in amount of sample (data not shown).
In our study, intracellular GXM-containing vesicles were repeatedly observed (Figure 3), as well as several unlabeled vesicular structures (Figure 1C and and2).2). In this context, we evaluated whether the distribution of GXM in extracellular vesicles could reflect the results observed in cellular structures. Fractionation of vesicle samples by gradient centrifugation followed by lipid and polysaccharide analysis revealed a previously unknown compositional profile of extracellular fungal vesicles (Figure 5). Sterol analysis was used to track membrane-containing fractions, since ergosterol has been characterized as an easily detectable component of vesicular membranes (Rodrigues et al., 2007). Phospholipid analysis in gradient fractions was not pursued due to poor sensitivity for lipid detection in chromatographic methods.
Ergosterol was detected in all fractions obtained after ultracentrifugation. GXM-containing fractions of different densities (fractions 4–6 and 10), as well as sterol-containing fractions that showed virtually no reactivity with an antibody to GXM (fractions 1–2 and 7–9), were observed. In some cases, the amounts of GXM and sterols in gradient fractions were directly correlated (e.g. fractions 4–6), but fractions with lower ergosterol content and high GXM density were also observed (e.g. fraction 10). These results are consistent with previously noted heterogeneity in vesicle size and composition (Albuquerque et al., 2008; Rodrigues et al., 2008) and indicate that some vesicles are preferentially associated with polysaccharide transport while others contain no GXM, as noted in prior immunogold labeling studies. The observation of extracellular vesicle fractions with such different properties indicate that the cellular events required for vesicle biogenesis are multiple and complex, as previously suggested by morphological analysis of cryptococcal extracellular vesicles (Rodrigues et al., 2008). Our results add to the evolving theme an intimate and intriguing association between lipid and polysaccharide components in C. neoformans capsule (Nicola et al., 2009; Sebolai et al., 2007; Sebolai et al., 2008).
In this work we report the use of two methodological approaches for studying the process of polysaccharide synthesis and transport: 1) a cryosectioning protocol for C. neoformans, which allowed a distinguished resolution of the fungal cytoplasmic content and surface structures, and 2) a technique for fractionating vesicles that separates them into GXM-and non-GXM containing groups. The cryosectioning EM method allowed us to discern in more detail the distribution of GXM inside the cell. The vesicle fractionation approach should allow finer resolution studies of vesicle composition and content by biochemical and mass spectrometry analysis. The results of these studies revealed new details for the mechanisms required for polysaccharide traffic and raise new questions about this complex cellular process. Of particular interest was the association of GXM with membrane structures suggesting that this polysaccharide is synthesized in close proximity to lipids, possibly by anchored enzymes. The possibility of combining cryoultramicrotomy with other less invasive methods such as high pressure freezing followed by freeze substitution and electron tomography could potentially generate new information for the study of still unknown cell biology attributes of cell-wall containing pathogens.
A.C. was supported by NIH Grants AI33774, HL59842, AI33142, and AI52733. M.L.R., L.N. and K.M. were supported by grants from the Brazilian agencies CNPq and FAPERJ. We thank Leslie Gunther and the Analytical Imaging Facility team at Albert Einstein College of Medicine for their assistance with the electron microscopy experiments.