High Pressure Freezing/Freeze Substitution Fixation Improves the Ultrastructural Assessment of Wolbachia Endosymbiont – Filarial Nematode Host Interaction

Background Wolbachia α-proteobacteria are essential for growth, reproduction and survival for many filarial nematode parasites of medical and veterinary importance. Endobacteria were discovered in filarial parasites by transmission electron microscopy in the 1970’s using chemically fixed specimens. Despite improvements of fixation and electron microscopy techniques during the last decades, methods to study the Wolbachia/filaria interaction on the ultrastructural level remained unchanged and the mechanisms for exchange of materials and for motility of endobacteria are not known. Methodology/Principal Finding We used high pressure freezing/freeze substitution to improve fixation of Brugia malayi and its endosymbiont, and this led to improved visualization of different morphological forms of Wolbachia. The three concentric, bilayer membranes that surround the endobacterial cytoplasm were well preserved. Vesicles with identical membrane structures were identified close to the endobacteria, and multiple bacteria were sometimes enclosed within a single outer membrane. Immunogold electron microscopy using a monoclonal antibody directed against Wolbachia surface protein-1 labeled the membranes that enclose Wolbachia and Wolbachia-associated vesicles. High densities of Wolbachia were observed in the lateral chords of L4 larvae, immature, and mature adult worms. Extracellular Wolbachia were sometimes present in the pseudocoelomic cavity near the developing female reproductive organs. Wolbachia-associated actin tails were not observed. Wolbachia motility may be explained by their residence within vacuoles, as they may co-opt the host cell’s secretory pathway to move within and between cells. Conclusions/Significance High pressure freezing/freeze substitution significantly improved the preservation of filarial tissues for electron microscopy to reveal membranes and sub cellular structures that could be crucial for exchange of materials between Wolbachia and its host.


Background
Most medically and economically important filarial parasites in the subfamilies Onchocercinae and Dirofilariinae depend on Wolbachia endosymbionts for development, reproduction and survival. Intracellular bacteria were first observed in filarial worms in early ultrastructural studies of Dirofilaria immitis, Brugia pahangi, Brugia malayi and Onchocerca volvulus [1][2][3]. Subsequent studies provided additional morphological information on these bacteria [4][5][6][7]. However, it was not until the late 1990's that these endobacteria were shown to belong to the genus Wolbachia, and this led to research that explored the bacteria as a novel drug target for filariasis [8][9][10][11].
Morphological studies by our group on the dynamics of Wolbachia distribution during the life cycle of B. malayi showed that Wolbachia densities are relatively low in microfilariae and vector stage larvae. Bacterial densities increase exponentially in the lateral chords of developing worms within the vertebrate host, and they later invade the growing ovaries of immature female worms [12]. Other authors have confirmed this infection process across tissue membranes. It now appears well established that Wolbachia are not only passively transferred to daughter cells in the course of cell division, but they also are able to cross cell membranes [13][14][15]. Ultrastructural studies with improved morphological preservation of fine structure of the bacteria and nematode tissues may provide new insight regarding Wolbachia movement and bacteriahost cell interactions. Initial ultrastructural studies of Wolbachia in filarial nematodes described endobacterial morphology extensively and reported that the bacterial cytoplasm was surrounded by three membranes [1,2]. Subsequent examinations of filarial parasites show varying levels of morphological quality, but studies of Wolbachia ultrastructure were often hampered by suboptimal preservation of membranes and organelles [12,[16][17][18][19].
Despite technical advances in electron microscopy in the last two decades, protocols for detecting Wolbachia and their interaction with filarial host tissues have changed very little since the bacteria were discovered in the 1970's. High pressure freezing/ freeze substitution fixation (HPF/FS) is a relatively new fixation method for transmission electron microscopy (TEM) that is especially useful for studies of cell and organelle membranes [20]. Comparison of chemical fixation and HPF/FS in platyhelminthes demonstrated the superiority of HPF/FS for the detection of membranes and organelles such as the ciliary rootlet system [21]. This fixation method also preserves antigenic properties of proteins, making it an excellent choice for immunogold electron microscopy [22].
Therefore, the objective of the present study was to use HPF/FS for TEM and immunogold TEM to better assess the relationship of Wolbachia to membranes and host tissues. This revealed many new and exciting features and suggested new hypotheses regarding Wolbachia-filaria interactions.

Comparison of Chemical and HPF/FS Fixation
Ultrastructural studies on various single cell as well as on multicellular organisms indicate that HPF/FS is superior to conventional, chemical fixation for preservation of membranes and subcellular structures [20,21]. For example, studies using HPF/FS improved the understanding of the ultrastructure and functional anatomy of the model nematode Caenorhabditis elegans [22,23]. Therefore we postulated that HPF/FS would improve the preservation of filarial nematodes and their endosymbiont for ultrastructural analysis. While membranes can be detected using chemical fixation and Epon embedding, their relation to Wolbachia is often difficult to recognize (Fig. 1A, C). Glycogen, which is often seen in the vicinity of Wolbachia in the lateral chord can be easily differentiated by HPF/FS, but is difficult to recognize in chemically fixed specimens (Fig. 1A, B). Undoubtedly using aldehydes as fixatives resulted in good morphological preservation in former studies [24,25]. However, it is known that LR White The same region as in A in a specimen fixed by HPF/FS. Note the improved preservation of membranes (arrowheads) and subcellular structures. C Cluster of Wolbachia within the lateral chord of a chemically fixed specimen. D Similar cluster as in C in a HPF/FS fixed specimen. Note the pleomorphy of the endobacteria (e.g. small electron dense bacteria with large nucleoid, arrow, next to large bacteria) and the small vesicles in the same area (arrowheads). W, Wolbachia; rER, rough endoplasmic reticulum; g, glycogen. Scale bar corresponds to 500 nm. doi:10.1371/journal.pone.0086383.g001 embedded material results in ''halo'' artifacts surrounding Wolbachia and it can be a problem to distinguishing Wolbachia from the host cytoplasm in Epon embedded material [26]. With HPF/FS fixed specimens none of these problems were encountered.

Morphology of Wolbachia
Wolbachia endobacteria were highly pleomorphic in all developmental stages of the parasite that we studied ( Fig. 2A-F, Table 1). Wolbachia distribution and morphology are stage and gender dependent, and vary by location in the worm (tissue and body region). For this study we decided to focus on developing females, because we wanted to evaluate Wolbachia morphology in embryonic stages. Pleomorphy was observed with regard to size, shape, and electron density of the cytoplasm. We observed striking variability in the appearance of the nucleoid, vesicle formation, the number of membranes surrounding the bacteria, the space between membranes, vacuole morphology, and in patterns of Wolbachia aggregation. Relatively small (250-500 mm), electron dense, coccoid Wolbachia with thick membranes were typically seen in developing embryos in the uterus of 8-12 weeks old females (Fig. 2 F). Similar but less electron dense endobacteria were present in the lateral chords of 5-6 week old, immature females ( Fig. 2D). In areas with high Wolbachia densities, clustered Wolbachia sometimes were present in a single vacuole (Fig. 2B). Interestingly, Wolbachia within a vacuole were not always morphologically identical, and this suggests that the cells may be in different stages of the cell cycle, perhaps due to asynchronous cell division within vacuoles. An early ultrastructural study described different morphological forms of endobacteria in B. malayi [1]. The author suggested that the endobacteria reside in vacuoles formed by the host's membranes and that they are randomly distributed throughout the length of the lateral chord. While our results clearly show that intracellular Wolbachia are surrounded by three bilayer membranes and that the space between the membranes varies in different Wolbachia forms (compare Fig. 2E and 2F), they do not support Kozek's suggestion that they are randomly distributed. Indeed, Wolbachia exhibit different distribution patterns in the lateral chords during different stages of worm development [12].
Lipopolysaccharides (LPS) present on the outer leaflets of the outer membrane of Gram negative bacteria are stained by heavy metals, and this can be observed by transmission electron microscopy. Previous studies showed that HPF/FS fixation is   especially suitable for preserving this asymmetric bilayer membrane structure [27]. Although Wolbachia do not possess LPS [28], our results show strong staining of the outer and the inner leaflets of the bilayer membrane structure (Fig. 2G). This symmetric ultrastructural staining pattern is similar to the one recently reported for the rickettsial endosymbiont 'MIDORIKO' of green algae using HPF/FS fixation [29]. We did not observe hostderived actin tails in Wolbachia that are involved in intracellular mobility in some other a-proteobacteria such as Rickettsia rickettsii [30,31]. For Drosophila it was reported that Wolbachia localization is dependent on microtubules, but not on the actin-based cytoskeleton [32].The absence of actin tails in Wolbachia of B. malayi suggests that this also applies to filarial Wolbachia.

Wolbachia Form and Secrete Vesicles
Some morphological forms of Wolbachia in the lateral chord and in early embryos of B. malayi synthesize and release membrane vesicles (Table 1). Large vesicles with diameters of 80-160 nm are released by electron dense coccoid Wolbachia in morula stage embryos in the uterus of 12 week old females (Fig. 3). These vesicles have a low electron density and often have two membranes when they are released into the cytoplasm (Fig. 3D, E). Vesicles with a single bilayer membrane were frequently observed within an outer vacuole membrane together with Wolbachia (Fig. 3B, C, F). This suggests that these vesicles are produced by the bacteria, especially in areas devoid of organelles such as the Golgi apparatus. In addition to single vesicles we saw polymorphic structures, where a number of smaller vesicles are enclosed by a large outer membrane ( Fig. 4A-C, Figs. S1, S2). While the formation of membrane vesicles has not been specifically studied in filarial Wolbachia, it is well known that Gram-negative bacteria release outer membrane vesicles. Vesicle formation is often associated with detoxification and acquisition of nutrients [33] or with virulence in pathogenic bacteria [34,35]. Although isolation of membrane vesicles is challenging, vesicles of the intracellular pathogen Chlamydia trachomatis have been isolated from host cells [36]. However, different subpopulations of vesicles in the vicinity of C. trachomatis have been described and some of them are derived from host cells. So far the detailed content of vesicles is known only for very few bacteria species [37].

Immunogold Labeling of Wolbachia and Wolbachiaderived Vesicles
Although the localization of many vesicles suggested that Wolbachia are synthesizing and releasing the vesicles, it is possible that some of these vesicles in close proximity to Wolbachia were formed by host organelles and taken up by the endobacteria. Therefore, we used a monoclonal antibody (mab) directed against Wolbachia surface protein 1 (WSP-1) to assess the origin of these vesicles (Fig. 4). As a technical control an antibody against double stranded (ds) DNA was used that showed specific labeling of dsDNA in host's nuclei and in Wolbachia cells while resulting in very little background (Fig. S3) The experiments using mab WSP-1 showed that Wolbachia membranes, including a large number of vesicles in host cell cytoplasm in the vicinity of Wolbachia, were labeled by mab WSP-1. The surface protein is localized at the outer, vacuole membrane which had previously thought to be formed by the host [1]. However, it is not clear how Wolbachia transfer WSP-1 into a host-derived membrane. Gold particles were seen only in association with membranes, but membranes of vesicles present in tissues devoid of Wolbachia were not labeled. Immunostaining affected the integrity of the membranes slightly, and the morphological preservation was not as good as in the regular TEM. However, no 'free' or non-membrane bound WSP-1 was detected, and this result differs from results recently reported by other authors [17]. Even membranes of multivesicular structures in the vicinity of Wolbachia were WSP-1 positive (Fig. 4A-D). A recent ultrastructural study using chemically fixed specimens suggested that WSP-1 (wBm0432) is co-localized with aldolase and interacts with the host's glycolytic pathway [26]. A bilayer membrane expected to be present in traditional bound vacuoles was not detected and the authors hypothesized that their specimens were not preserved well enough to demonstrate membrane binding of the antibody. Our results indicate that vesicle formation by Wolbachia is common and that WSP-1 is membrane-bound and localized to membranes forming the Wolbachia vacuole and Wolbachia derived vesicles. Pathogen-containing intracellular compartments sometimes interact with secretory pathways of the host cell [38]. Therefore, we studied the association of Wolbachia with host organelles such as the endoplasmic reticulum and Golgi apparatus. HPF/FS resulted in excellent preservation of these subcellular structures. Single or small clusters of Wolbachia within the lateral chord were frequently detected in the vicinity of the Golgi apparatus, the endoplasmic reticulum, and inside or outside of phagolysosomes (Fig. 5, Figs. S1, S2). Single (Fig. 5B) or dividing (Fig. 5C) Wolbachia and vesicles were observed on both the cis and the trans sides of the Golgi apparatus. For comparison, it has been reported that Chlamydia disrupts the Golgi apparatus to ensure reproduction, while some insect Wolbachia reside within Golgi-related vesicles [39,40]. A Drosophila Wolbachia strain has been reported to be associated with the endoplasmic reticulum [41]. We often observed small numbers of Wolbachia near nucleus-associated endoplasmic reticulum in the lateral chord (Fig. 5D). However, large clusters of Wolbachia, which are often seen in the lateral chord of L4 and young adult worms during the exponential growth of the endobacteria population, were rarely seen in direct contact with the endoplasmic reticulum or the Golgi apparatus. Based on these observations, it is possible that Wolbachia use the secretory pathway of the host cell including vesicle trafficking and membrane fusion for intracellular and syncytial motility.

Wolbachia in the Lateral Chord and in the Pseudocoelom
Large numbers of Wolbachia were found in the lateral chord in the posterior end of the worm (near the ovaries and the uterus) in young, immature female B. malayi (Fig. 6). Extracellular Wolbachia were sometimes present in the pseudocoelomic cavity (Fig. 6A). This observation is consistent with those recently reported (compare to Fig. 8 in [12]). HPF/FS fixation showed that these extracellular Wolbachia have a thick cell envelope comprised of two membranes (Fig. 6B). The utrastructural images did not reveal a mode of motility for these extracellular Wolbachia; it is possible that passive motility based on changes in pressure and flow of fluid in the pseudocoelomic cavity enable bacteria to invade the reproductive system of the worm host. It is also possible that these extracellular Wolbachia forms induce endocytosis by host cells in a manner similar to elementary bodies of Chlamydia.
Wolbachia have their highest density and greatest morphological diversity in the lateral chords of female L4, in immature adult females, and in mature females adjacent to the distal portion of the reproductive organs, especially ovary, oviduct and seminal receptacle. In some parts of the lateral chord, the tissue is highly vacuolized and consists almost exclusively of Wolbachia either as single cells or as clusters (Fig.7). These dense accumulations of Wolbachia are often associated with high densities of glycogen granules. Physical proximity of Wolbachia and glycogen suggest a functional association. Transcriptomic and molecular biological studies of filarial Wolbachia have suggested a mitochondrion-like function [26,28,42]. However, direct consumption of glucose or glycogen by Wolbachia seems unlikely, because its genome lacks genes coding for two enzymes required for glycolysis [28].

Conclusions
HPF/FS fixation dramatically improved tissue preservation for enhanced ultrastructural examination of Wolbachia and filarial nematodes. Our study provides more detailed information regarding the pleomorphy of Wolbachia in different tissues, their association with vesicles, and their movement within cells and across tissues. Membrane and membrane vesicles appear to be critically important morphological structures involved in the Wolbachia/filarial nematode interactions. Vesicles may shuttle metabolic products between bacteria and worm tissues, this could help explain the interdependence of and B. malayi.

Parasite Material
The protocol to maintain B. malayi in Mongolian gerbils for morphological studies was approved by the Animal Studies Committee at Washington University School of Medicine. Male gerbils were infected i.p. with 200 B. malayi infective larvae and sacrificed 2, 3, 4, 5, 6, 8 and 12 weeks post infection. Parasites were obtained in intraperitoneal lavage using PBS cooled to 4uC and older stages were separated by sex. After external examination worms were immediately submitted for HPF/FS. For each time point at least 5 larvae or 3 female worms were examined.

Standard Processing for TEM
For conventional ultrastructural analysis of chemically fixed specimens, live worms were fixed in 2% paraformaldehyde/2.5% glutaraldehyde (Polysciences Inc., Warrington, PA) in 100 mM phosphate buffer, pH 7.2 for 1 hr at room temperature. Samples were washed in phosphate buffer and postfixed in 1% osmium tetroxide (Polysciences Inc.) for 1 hr. Samples were then rinsed extensively in distilled water prior to en bloc staining with 1% aqueous uranyl acetate (Ted Pella Inc., Redding, CA) for 1 hr. Following several rinses in distilled water, samples were dehydrat- High-pressure Freezing and Freeze Substitution (HPF/FS) Live worms were placed in specimen planchettes containing 20% bovine serum albumin as a cryoprotectant. Planchettes were then high-pressure frozen in a Leica EM PACT2 high-pressure freezer (Leica Microsystems) at 2180uC, 2,100 bar and maintained under liquid nitrogen. For optimal ultrastructural analysis, samples were transferred to freeze substitution medium (acetone containing 2% osmium tetroxide, 0.1% uranyl acetate, and 5% distilled water) under liquid nitrogen and placed in the Leica AFS, automatic freeze substitution system (Leica Microsystems) precooled to 2130uC. For freeze substitution, the samples were brought to 290uC over 1 hr, remaining at 290uC for 10 hr, and subsequently warmed to 220uC over a period of 18 hr. Samples were placed at 4uC for 30 min prior to washing with anhydrous acetone at room temperature. Samples were infiltrated and embedded in Eponate 12 resin, and sectioned and stained as described above.

Immunogold TEM
For immunolabeling of high-pressure frozen samples, the freeze substitution medium consisted of acetone containing 0.2% glutaraldehyde, 0.1% uranyl acetate, and 5% distilled water. Following freeze substitution in the AFS unit as described above, samples were infiltrated in LR Gold resin (Ted Pella) at 220uC, polymerized under UV light, and subsequently sectioned and stained. A mab against WSP-1 (Wbm0432) was used. The antibody was kindly provided by Patrick Lammie (CDC, Atlanta) and extensively evaluated by immunohistochemistry previously [12]. A monoclonal antibody directed against ds DNA (clone AC-3-10, Novus Biologicals, CO) was diluted 1:20 and used as a positive control according to the instructions of the manufacturer. showing also Golgi cisternae proximal to Wolbachia and endoplasmic reticulum. C Another part of panel A showing Wolbachia near single (arrowhead) and multiple (double arrowheads) vesicles. D Magnification of a portion of panel A demonstrating a Wolbachia enclosed by multiple membranes with numerous small vesicles (arrowhead) surrounded by the same outer membrane. This structure may be similar to those in panel C, but from a different perspective. W, Wolbachia; nu, nucleus; sER, smooth endoplasmic reticulum; rER, rough endoplasmic reticulum; Go, Golgi; Scale bar corresponds in A-C to 500 nm and in D to 200 nm.