Introduction

The presence of an endosymbiont of the Bacteroidetes phylum and its role in the feminization of Brevipalpus phoenicis (Geijskes) (Acari: Tenuipalpidae) were first described by Weeks et al. (2001). Since then, this bacterium, now classified as Cardinium, has been found in many arthropods (Weeks et al. 2003; Zchori-Fein et al. 2004), including three other Brevipalpus species, B. obovatus Donnadieu, B. lewisi McGregor and B. californicus (Banks) (Weeks et al. 2003; Novelli et al. 2005; Chigira and Miura 2005). The presence of the symbiotic bacterium has been determined through PCR using primers that amplify the 16S rDNA gene from all known Eubacteria or using sequence specific primers for Cardinium (Weeks et al. 2001; Weeks et al. 2003; Zchori-Fein et al. 2004; Chigira and Miura 2005; Novelli et al. 2005).

Natural populations of Brevipalpus are made up mostly of haploid females which reproduce by thelytokous parthenogenesis. Occasionally, males are present, and although they mate with females, fertilization is probably not effected (Pijnacker et al. 1981). Antibiotic treatment provided strong evidence that Cardinium is responsible for the feminization; untreated females produce almost exclusively females while antibiotic treated females produce significantly more males (Weeks et al. 2001; Chigira and Miura 2005; Groot and Breeuwer 2006). However, aposymbiotic populations of B. phoenicis were discovered that are thelytokous and halpoid. It has been suggested that in those mites, genes responsible for feminization of unfertilized haploid eggs may have been transferred from the symbiont to the host genome (Groot and Breeuwer 2006). Although parthenogenetic reproduction may be an evolutionary disadvantage by limiting the amount of genetic variation, some species of Brevipalpus seem well adapted for the environmental variation. Specifically, B. phoenicis, B. californicus and B. obovatus have been reported from nearly 1,000 different plant species belonging to 139 families (Childers et al. 2003). In addition, transplantation experiments of B. phoenicis to alternative host plant species led Groot et al. (2005) suggest that the model of frozen niche variation, which postulates that different clones are specialized to different niches, best explains the clonal adaptation of these mites. Brevipalpus mites are also currently attracting interest because of their role as vector of several plant viruses. These viruses characteristically induce localized lesions in infected plants (Kitajima et al. 2003). Among them citrus leprosis is economically the most important viral disease of citrus industry in the Americas. In the Brazilian state of São Paulo alone, expenditures for chemical control of B. phoenicis, the vector of the causal virus, have been estimated to be up to US$ 80 million/year (Rodrigues et al. 2002, 2003).

PCR has been efficient to detect Cardinium in Brevipalpus mites (Weeks et al. 2001; Chigira and Miura 2005; Novelli et al. 2005; Groot 2006) but it does not provide information regarding the distribution of the bacterium within the host organism. On the other hand Cardinium has a unique structural feature among bacteria, the presence of microtubule-like material (ML) (Zchori-Fein et al. 2004; Bigliardi et al. 2006), which permits its discrimination from other bacterial types in sections.

To explore such a peculiarity, an electron microscopic study was undertaken in sections of three Brevipalpus species (B. californicus, B. obovatus and B. phoenicis) to confirm the presence of Cardinium in different instars and observe its distribution among different organs. Here we report the results of such studies which confirm the presence and absence of Cardinium bacteria as determined by PCR-based techniques. In addition we describe the distribution of the bacteria among distinct organs and tissues of the host, as well as in eggs. Finally, we provide details on the typical ML of these symbionts.

Materials and methods

Mite populations

The populations of Brevipalpus species studied are listed in Table 1. Most of them were collected in Brazil, but samples from The Netherlands, Costa Rica, Colombia and Japan are included. Some samples came from populations collected in Brazil and The Netherlands, maintained on bean (Phaseolus vulgaris L.) at the University of Amsterdam. Presence or absence of Cardinium in these populations was previously assessed by PCR (Novelli et al. 2005, 2006; Groot 2006).

Table 1 List of the Brevipalpus mite populations studied and the results of electron microscopic examination
Full size table

Electron microscopy

Adult mites of all populations (Table 1) were fixed in a solution of 3% glutaraldehyde in 0.1 M cacodylate buffer pH 7.2, with 1.8% (w/v) sucrose (R. Dallai. Personal communication), post fixed in 1% osmium tetroxide, dehydrated in acetone and embedded in Spurr’s low viscosity epoxy medium. To enable penetration of the fixative and infiltration of the embedding medium, a small portion of either the anterior part of the propodosoma or the posterior end of the hysterosoma was cut off with a razor blade under a binocular. Sections were made in a Leica Ultracut UC2 microtome equipped with a diamond knife. Thin sections were collected on copper grids, stained with 3% uranyl acetate and Reynold’s lead citrate (Maunsbach and Afzelius 1999) and examined in a Zeiss EM 900 transmission electron microscope. To orient the localization of the organs, semi-thins sections (1–2 μm) were mounted on glass slides and stained with basic 1% toluidine blue. All the chemicals used were purchased from Electron Microscope Sciences (Hartfield, PA, USA). Males, larvae, protonymphs, deutonymphs and aposymbiotic adults (Table 1) of some populations were also included in these studies. At least six mites per population and instars were examined.

Results and discussion

Because of the cuticle, tissue preservation and cutting properties of the processed mites were poor if entire mites were used. The opening made by the cutting off mite’s extremities resulted in good fixation and infiltration of the resin and in good quality sections.

Cardinium was the only bacterium (Figs. 1 and 2) found in Brevipalpus mites, confirming data of Groot (2006). In sections the bacterial cell profile varied from oval/elliptical to sausage-shaped. The largest cells were 4–5 μm long and 0.5–1 μm wide. Cardinium cells were enclosed by the plasma membrane and a thin, Gram negative type cell wall and contained small ribosomes and fibrillar DNA as other prokaryonts. The most characteristic feature of Cardinium, and unique among bacteria, is the presence of a bundle of thin tubular structure, generally arranged perpendicularly to the lateral part of the bacterial body (Fig. 2). These tubules were associated with a plate-like structure as originally described by Zchori-Fein et al. (2004) and detailed by Bigliardi et al. (2006) and referred to as microtubule-like structures (ML). In cross section these ML had a uniform diameter of ca. 12–15 nm, and were up to 100 nm long as seen in longitudinal sections. Counts made in cross sections revealed up to 50 of MLs/cell. The ultrastructural features of the Cardinium found in Brevipalpus were essentially similar to the detailed description made by Bigliardi et al. (2006) in the vine leafhopper Scaphoideus titanus Ball except for the diameter of the MLs which was rather uniform in Brevipalpus. The function of the ML is still unknown. Bigliardi et al. (2006) suggested that they may serve as cytoskeleton, have a role in cell transport or in communication between cell membrane and cell wall. These ML seem to be different from the recently described filament bundles in Caulobacter crescentus using electron cryotomography (Briegel et al. 2006) that are thinner and have varied location in the bacterial cell.

Fig. 1
figure 1

Transmission electron micrograph of sections of three Brevipalpus mite species (B. phoenicis, B. obovatus and B. californicus) indicating the presence of Cardinium symbionts in different organs of adult females. (A) Symbionts in the epidermis of B. californicus collected from orchid. (B) Cardinium can be seen within the eye complex of B. phoenicis infesting Clerodendron thomsonae. (C) Numerous bacteria are present in the ovary, within the oocytes in B. phoenicis infesting Passiflora edulis f. flavicarpa. (D) A single bacterium in an egg within the ovary of B. obovatus collected from Solanum violaefolium. (E) Cardinium in the epithelial cell of branched digestive tube (cecum) of a B. phoenicis infesting Citrus sinensis. (F) Symbiont in a nerve cell surrounding the nervous ganglion in B. obovatus collected from Cestrum nocturnum. Cut–cuticle; Vm–vitelline membrane. Arrows point to the bacterial cells

Full size image
Fig. 2
figure 2

Details of the microtubule-like structures (ML) in the cells of Cardinium from adult Brevipalpus mites. (A and B) Longitudinal sections of the ML arranged perpendicularly onto the cell membrane. A layer is visible adjacent to the insertion of the ML on the cell membrane (arrow), respectively from B. phoenicis infesting Clerodendron x speciosum and B. obovatus colonizing S. violaefolium. (C and D) Cross section of bundles of the ML, showing their tubular nature and the ordered array, respectively in B. phoenicis collected from C. sinensis and C. thomsonae

Full size image

Groot (2006) mentions at least two different types of Cardinium in Brevipalpus based on molecular data. One of them is present in B. phoenicis and B. obovatus and the other, in populations of B. californicus. No morphological difference in the morphology of the Cardinium was observed in the different specimens studied here.

Cardinium cells were detected in every part of the Brevipalpus body including the legs and palps. However, even without critical morphometric analysis, the examination of a large number of sections of different organs of the three Brevipalpus species indicated that the number and distribution of Cardinium was highly variable between individuals including different instars, and in the inner organs. The symbionts were located in most wide array of tissues (Alberti and Coons 1999), including epidermis, muscle, eye, fat body, digestive system epithelium, nervous cells (ganglia and innervations), trachaeal cells, prosomal gland, ovarium and developing eggs (Fig. 1). This is in contrast to S. titanus where the presence of symbionts was restricted to the salivary gland and reproductive system (Bigliardi et al. 2006). Cardinium cells were absent in males, confirming PCR assays by Weeks et al. (2001). The symbionts appeared immersed directly in the cytoplasm of the infected cells (Figs. 1 and 2). This is a different situation in comparison with Wolbachia symbionts which consistently are located within a vacuole, separated from the cytoplasm by a membrane (Louis and Nigro 1989).

Electron microscopy revealed that Cardinium bacteria were present in all three Brevipalpus species studied. However, in three populations of B. phoenicis no symbiont could be observed confirming PCR data (Groot 2006; Novelli et al. 2006). One of these populations was collected from Wisteria floribunda DC found in Piracicaba, SP, Brazil. Another was collected from Citrus sinensis (L.) Osbeck at Vila Fão, RS, Brazil. The third population was collected from Strongylodon macrobotrys A. Grey in The Netherlands. The latter population though morphologically is B. phoenicis type 2, based on the idiosomal reticular ornamentation (Mesa Cobo 2005), is more related to B. obovatus based on mitochondrial DNA markers and might actually be a fourth species (Groot 2006). No similar analysis was made with other two aposymbiotic populations.

Because Cardinium is widespread within the mite body (Fig. 1) it is likely that most of the mite cells, including ovary, become infected during early embryogenesis, though uninfected cells may appear and later be infected through cell-to-cell movement of the bacterium. The appearance of the rare males in natural population could be due to the production of some bacteria-free egg cells. This situation is enhanced with tetracycline treatment which kills Cardinium in females and significantly increases the number of males among the offsprings (Weeks et al. 2001). This led to the notion that infection by Cardinium results in the feminization of the haploid Brevipalpus. The finding of female-only, aposymbiotic populations seem enigmatic but Groot and Breeuwer (2006) suggest as an explanation that the gene(s) of the symbiont responsible for the feminization may have been transferred to the mite genome. What the effect would be of a reinfection of such population, a possibility that exists because evidence of horizontal transmission of Cardinium was presented (Groot 2006), has to be assessed. Whether or not Cardinium has a role in the transmission of the viruses by Brevipalpus mites is under investigation.

One advantage of the visualization of the Cardinium in situ was to determine that the symbiont is widespread within the body of the examined Brevipalpus species, a situation that is difficult to conclude through PCR based technique or by light microscopy. On the other hand, this fact raises some questions regarding the role of the symbiont in the biology of the mite host. If the symbiont is involved only in feminization or related function, it would be expected to be restricted to the host reproductive system as is commonly observed for Wolbachia symbionts (Stouthamer et al. 1999) and also noticed for Cardinium in the cicadelid S. titanus (Bigliardi et al. 2006). But being present everywhere in the mite body, the possibility of some additional role should be considered, for it is likely that maintaining these bacteria is of high metabolic cost to the host. A critical comparison of the physiology and behavior of normal and aposymbiotic populations may provide some answers. For instance, Weeks and Stouthamer (2004) showed that infection with Cardinium had a positive effect on the fecundity of the predatory mite Metaseiulus occidentalis.