Wolbachia pipientis grows in Saccharomyces cerevisiae evoking early death of the host and deregulation of mitochondrial metabolism.

Abstract Wolbachia sp. has colonized over 70% of insect species, successfully manipulating host fertility, protein expression, lifespan, and metabolism. Understanding and engineering the biochemistry and physiology of Wolbachia holds great promise for insect vector‐borne disease eradication. Wolbachia is cultured in cell lines, which have long duplication times and are difficult to manipulate and study. The yeast strain Saccharomyces cerevisiae W303 was used successfully as an artificial host for Wolbachia wAlbB. As compared to controls, infected yeast lost viability early, probably as a result of an abnormally high mitochondrial oxidative phosphorylation activity observed at late stages of growth. No respiratory chain proteins from Wolbachia were detected, while several Wolbachia F1F0‐ATPase subunits were revealed. After 5 days outside the cell, Wolbachia remained fully infective against insect cells.

In contrast, eliminating Wolbachia with tetracycline in filaria, increases respiratory-chain gene expression in the host and causes an early death. This result, lead to the hypothesis that at least in filariae Wolbachia contributes as an energy generator for the host (Strübing, Lucius, Hoerauf, & Pfarr, 2010;Darby et al., 2012Darby et al., , 2014. Culturing obligate intracellular bacteria is a challenge. Insect cells support Wolbachia growth, but culture times are long and cells are difficult to manipulate. Alternative systems such as mammalian blood have proven helpful to grow intracellular organisms such as Sodalis (Dale & Maudlin, 1999). However, Wolbachia did not seem to grow in blood and this was not pursued further (Result not-shown; see Methods). In contrast, Saccharomyces cerevisiae did support the growth of Wolbachia strain wAlbB.
As it can be extensively manipulated, S. cerevisiae is widely used as a model organism in biochemistry and molecular biology. In S. cerevisiae, it is possible to study processes such as the Crabtree effect observed in tumor cells (Diaz-Ruiz, Rigoulet, & Devin, 2011) and to model cell death in response to anoxia or ischemia/reperfusion (Stella, Burgos, Chapela, & Gamondi, 2011). In addition, it is used as a host to study DNA and RNA viral replication (Alves-Rodrigues, Galão, Meyerhans, & Díez, 2006), to identify and characterize bacterial effectors and toxins (Siggers & Lesser, 2008) and to analyze the function of heterologously expressed proteins such as the Yarrowia lipolytica and the mammalian brown-fat mitochondrial uncoupling proteins (UCPs) (Guerrero-Castillo et al., 2011). Thus, when it was observed that Wolbachia grew in S. cerevisiae, the system was characterized and the effects of Wolbachia infection on its host were analyzed.
Growing Wolbachia in insect cell cultures or in live hosts presents difficulties that have precluded detailed biochemistry and physiology studies Khoo, Venard, Fu, Mercer, & Dobson, 2013). Here, we used the S. cerevisiae strain W303 as an alternative host for Wolbachia wAlbB and analyzed the host/endosymbiont system. Infected yeasts died earlier than controls. This probably resulted from an abnormally high mitochondrial oxidative phosphorylation activity observed at late stages of growth. Understanding Wolbachia and host-Wolbachia interactions holds great promise for medical, parasitological, and biotechnological applications.
Prior to use, heat-inactivated fetal bovine serum (FBS; 30 min at 56°C) was added to a final concentration of 10% (Shih et al., 1998).
The insect cell line was grown on True Line TR 4003 140 mm sterile petri dishes at 27°C in a 5% CO 2 atmosphere (ESCO CelCulture CO 2 incubator or in Corning culture flasks, Shanghai, China). Subcultures were performed in a 1:10 split at 90% confluence. A sample from this cell line was treated with tetracycline to eliminate Wolbachia infection (Aa23Tet) (Dobson, Marsland, Veneti, Bourtzis, & O'Neill, 2002).

| Cell viability assays
Viability of Aa23 cell line, Wolbachia or yeast was assessed using the BacLight live-dead staining kit (Molecular Probes, Carlsbad, CA).
Ten microliters of cell suspension were stained according to the manufacturer suggested protocol and viewed in an epifluorescence NIKON microscope.

| Failed attempts to grow Wolbachia ex-vivo and a serendipitous finding
The original idea was to find a system where Wolbachia would grow ex-vivo. To do this, diverse protocols used for other endosymbionts such as Sodalis and Coxiella were followed (Dale & Maudlin, 1999;Omsland et al., 2009Omsland et al., , 2013. It was found that some components did improve survival in isolated Wolbachia, even if we never observed substantial growth. Some of these agents were: (1) Trehalose and other compatible solutes such as mannitol, glycerol and sucrose, known to stabilize pollen (Crowe, Reid, & Crowe, 1996;Leslie, Israeli, Lighthart, Crowe, & Crowe, 1995) and isolated proteins (Sampedro & Uribe, 2004) (2) Actin, which supports binding and movements of some endosymbionts in vivo. (3) Catalase which deactivates hydrogen peroxide (Dale & Maudlin, 1999) and (4) Blood from large mammals, which has been used to grow Sodalis (Dale & Maudlin, 1999) and increases Wolbachia titers (Amuzu, Simmons, & McGraw, 2015;McMeniman, Hughes, & O'Neill, 2011). Human blood was also effective.
First, we tried growing Wolbachia using sheep blood. However, it was easily contaminated at the sites of extraction, so cultures had to be discarded often. On one occasion we obtained positive wsp gene amplification from a yeast colony grown in one of the agar plates. Out of curiosity, we studied the host, which turned out to be S. cerevisiae. From this accidental finding we decided to test a known strain of S. cerevisiae as an alternative host. We learned that, in order to support growth of Wolbachia, yeast culture media needed to be supplemented with blood, which eventually was substituted with ammonium ferric citrate with excellent results and none of the contamination problems. Neither compatible solutes, nor catalase nor actin enhanced growth. The second addition needed was bovine fetal serum, which was present in all original growth media but not in yeast culture media. FBS was titrated and we ended up using 1%.

| Wolbachia wAlbB infection of the Saccharomyces cerevisiae W303 yeast strain (wScW303)
A first yeast infection was performed following a modified cell line infection protocol (Dobson et al., 2002). All procedures were performed under sterile conditions. The Aa23 cell line (containing Wolbachia) was grown in Corning cell culture flasks (225 cm 2 ) as described in (Shih et al., 1998). After 20 days of culture, cells were scrapped and concentrated by centrifugation at 3,000g for 5 min.
For homogenization, ~1*10 7 cells were resuspended in 10 ml Eagles medium and vortexed for 10 min with (50% v/v) 3 mm sterile borosilicate glass beads (Rasgon, Gamston, & Ren, 2006). The homogenate was centrifuged at 3,000g for 10 min to remove unbroken cells. The supernatant was passed through a 2.7 μm syringe filter and the fil-  grown in a liquid YPD culture for 3 hr, harvested and centrifuged at 3,000g for 3 min. Culturing yeast in low oxygen environments prevents thickening of the cell wall (Aguilar-Uscanga & Francois, 2003;Smith et al., 2000;Avrahami-Moyal, Braun, & Engelberg, 2012). To induce contact between bacteria and yeast both bacteria (the whole 2 ml sample) and yeast (60 mg ww) were mixed and centrifuged at 2,500g for 1 hr at 20°C (Dobson et al., 2002). Bacteria-infected yeast were plated (all 2 ml) on a Petri dish containing MM supplemented with 1 mmol L −1 ammonium ferric citrate plus 25% v/v outdated human packaged erythrocytes and 2% agar (MM Fe-blood) and incubated at 27°C in a 5% CO 2 chamber (ESCO, Cell Culture CO 2 incubator, Singapore) for 14 days (Dale & Maudlin, 1999). Infection was confirmed by FISH and PCR. Infected yeast was transferred to a fresh agar plate every month for up to 6 months, then yeast was discarded and a new sample was used. Some aliquots were added with 40% glycerol, frozen and stored at −80°C, these samples have remained infective for nearly 10 months.
To transfer Wolbachia from yeast to yeast, slight modifications to the protocol were made: An aliquot of 100 μl of yeast taken from a glycerol-frozen sample or a loophole of infected yeast cells was diluted in 2 ml YPD Fe 20% FBS and plated in YPD Fe-blood agar plates, which were grown in 5% CO 2 . After 14 days, all cells grown in a Petri dish were collected and washed by centrifugation at 3,000g for 3 min at 20°C with sterile water and the pellet was suspended in 10 ml MM. The suspension was vortexed for 10 min in the presence of 0.425-0.600 mm sterile borosilicate glass beads (60% v/v) to disrupt yeast cells (note that beads were smaller than those used for insect cell lines). Disrupted yeasts were centrifuged at 3,000g for 10 min and the supernatant was centrifuged again 3,000g for 10 min. The washed supernatant was filtered through different 0.8-0.65-0.45 μm syringe filters. Again, we used filters with smaller pores than those used for cell lines due to the small size of yeast cells. The last filtrate was centrifuged at 16,500g for 10 min. The pellet (~60 mg ww) was suspended in 2 ml MM Fe FBS and used to infect yeast from 3-h cultures as described above. The yeast-bacterium mixture was plated in a YPD Fe agar plate and incubated at 27°C with 5% CO 2 for at least 7 days. Infection was evaluated using FISH and PCR.

| Culture and maintenance of wAlbB-infected Saccharomyces cerevisiae W303
Infected S. cerevisiae strains were kept in YPD plus 1 mmol L −1 ammonium ferric citrate agar plates. When transferring to liquid medium, a loophole from the desired strain was suspended in 100 ml of sterile YPDS and incubated at 28°C, 130 rpm for 48 hr. Precultures were decanted in one liter YPDS and incubated at the same conditions for up to 14 days. When transferring from solid to solid media, a loophole of yeast was suspended in 1 ml YPD supplemented with 1 mmol L −1 ammonium ferric citrate plus 20% FBS and plated on YPD agar. A cell passage every 2-3 weeks was performed in order to maintain the infection. When it was desired to eliminate Wolbachia from yeast, tetracycline 30 μg/ml was added five consecutive times to the medium as passages were performed (Dobson et al., 2002).

| Wolbachia wAlbB infection of the C6C36 Aedes albopictus cell line
To determine whether Wolbachia cells retained its infective ability after all treatments, Wolbachia were isolated from S. cerevisiae grown in liquid YPD Fe 1% FBS and they were tested for infection against a C6C36 insect cell line.

| Z-cut images for cell reconstruction
Fourteen day old infected and noninfected yeast samples were visualized with a Olympus-FV1000 or FV-3000 microscopes. Z-cut images were reconstructed using Imaris 7.2.1 and Image J software.

| Antibodies
Primary antibodies: Mouse monoclonal Anti-Wolbachia Surface Protein NR-31029 was from BEI Resources, NIAID, NIH. Mouse monoclonal Anti-VDAC was from Abcam. Secondary antibody: HRP coupled Anti-mouse antibody from Jackson ImmunoResearch (West Grove, PA). ). PVDF membranes were stripped as indicated by abcam protocol using a mild-stripping buffer, blocked with 5% Blotto nonfat dry milk in TBS-T and reprobed with a different antibody as indicated.

| Transmission electron microscopy of wScW303
Infection was assessed by transmission electron microscopy (TEM) following a protocol from (Sun et al., 2015). Briefly, 500 μl of cells were harvested from 100 ml cultures of infected and uninfected Saccharomyces cerevisiae cultures form the first unintentional infection (wSc) at 10 days and wScW303 of fourteen days. Yeast and Wolbachia samples were washed twice in distilled water at 740 g for 5 min for yeast and 23400 g for 10 min for bacteria in an Eppendorff Centrifuge 5415C. Samples were fixed in 2% KMnO 4 at 4°C overnight. Next day, samples were washed for 15 min with deionized water six times and dehydrated with sequential 10-minute washes with 50%, 70%, 80%, 90% ethanol and three washes with 100% ethanol. Samples were washed with ethanolpropanone (1: 1) for 8 min, then with anhydrous propanone for 5 min, then with propanone-EPON 821 (3: 1) for 1 hr and left in propanone-EPON 821 (1 : 3) overnight. Next day, samples were concentrated and resuspended in propanone-EPON 821 (1: 1) for 1 hr. Then, samples were concentrated again and left in resin for 24 hr. Then they were incubated for 12 hr at 37°C and then further incubated for 36 hr at 60°C. Resins were cut into 70 nm slices on an ultra-microtome (Ultracut Reicheit-jung) and observed in a JEOL JEM-1200 EXII electron microscope. Data were processed using Gatan Digital Micrograph Software.

| Mitochondrial (or Mitochondria/Wolbachia mixture) isolation
Yeast were centrifuged at 3,000g for 5 min, washed twice in water and resuspended in MES-mannitol buffer (5 mmol L −1 MES, 0.6 mol L −1 mannitol, 0.1% BSA pH 6.8 adjusted with triethanolamine). Yeast were disrupted using a Bead Beater cell homogenizer (Biospec Products, OK, USA, final volume 50 ml) with 0.425-0.6 mm glass beads during three 20 s pulses separated by 40 s resting periods in ice (Uribe, Rangel, Espínola, & Aguirre, 1990). The homogenate was differentially centrifuged to isolate mitochondria similar to described in (Peña, Piña, Escamilla, & Piña, 1977). Briefly, cells were centrifuged at 1,100g for 5 min. The supernatant was centrifuged at 9,798g for 10 min and the pellet was resuspended in MES-mannitol buffer and centrifuged at 3,000g for 5 min. Finally, the supernatant was centrifuged at 17,500g for 10 min. The resulting pellet was resuspended in minimal volume and protein concentration was measured by Biuret (Gornal, 1957) using a Beckman Coulter spectrophotometer at 540 nm.

| At the expense of its own viability, the artificial host Saccharomyces cerevisiae W303 supports growth of Wolbachia wAlbB
To study Wolbachia (wAlbB) large biomass yields plus a host that is easy to manipulate are needed. After testing different alternatives  Table S1.
3.2 | Proliferation of Wolbachia in ScW303 was further confirmed using different independent methods as follows (Figure 2)

| PCR of the Wolbachia outer surface protein gene (wsp)
Both the Aa23 cell line (Figure 2a)

| Western Blot analysis detected Wolbachia wsp in S. cerevisiae
In the Aa23 cell line, a ~37 kDa protein corresponding to the Wolbachia Surface protein (wsp) was revealed with anti wsp antibodies (Bei resources, NIH, MD) (Figure 2c, Aa23). This band disappeared after growth in the presence of tetracycline (Aa23 Tet). VDAC (Voltage dependent anionic channel) protein was used as a loading control. In non-infected yeast wsp was not detected, (Figure 2c, ScW303), while in infected yeast the wsp western blot signal was first detected at day 3 and increased gradually up to day 10, remaining stable until day 14 (Figure 2c, wScW303). (For images of original Western Blots, see Figure S1a). When tetracycline was added to the medium, the wsp signal decreased, disappearing by day 10 ( Figure S1b, wScW303Tet).

| Normal growth and early death were observed in infected S. cerevisiae
During the first 12 days of culture, growth curves of infected wScW303 were similar to the controls (Figure 3a). Then, beginning at day 14, wScW303 absorbance decreased. Cell wall degradation ( Figure S2) and viability staining (Figure 3b) confirmed that wScW303 viability was rapidly lost during the late stages of the stationary phase, from 14 to 18 days of culture.
In addition, during growth the transcriptional activity of both S. cerevisiae 18S rRNA and the Wolbachia wsp were tested.
Transcription was high in S. cerevisiae from the first day, decreased at day fourteen and became negligible at days 16 and 18 (Figure 3c).
In contrast, transcription of the wsp from Wolbachia became detectable only after 3 days, increased exponentially until day 10 and remained constant until day 14. Then, at days 16 and 18, transcription decreased abruptly (Figure 3c). Transcription data in the Wolbachia/S. cerevisiae system indicated that Wolbachia activity grew later than S. cerevisiae, reaching a maximum at 10 days. Later, beginning at 14 days both transcription activities decreased abruptly in parallel with the death of the host.

Tridimensional reconstructions of z-cuts performed in a wScW303
sample show the intracellular location of different bacteria (movies S1 and S2). In the periphery of movies S1 and S2, few independent bacterial labels were detected, which we speculate, may come from bacteria inside heavily deteriorated host cells whose cell wall was not stained by Calcofluor (movies S1 and S2).

| Wolbachia-infected yeast retained high mitochondrial oxidative phosphorylation activity for abnormally long periods
A possible mechanism for the early death of infected yeast was explored in our infected ScW303/wAlbB system. This system exhibited an abnormal preservation of mitochondria ( Figure 5), so it was logical to explore aerobic metabolic activity. The relationship between Wolbachia and aerobic metabolism in the host is a matter of controversy. Some authors have proposed that these endo-cellular organisms possess an aerobic metabolism that contributes to overall activity (Strübing et al., 2010) Kurtz, 2014;Foster et al., 2005;Heddi et al., 1999;Strübing et al., 2010). Thus, we decided to evaluate oxidative phosphorylation activities in our system, which preserved mitochondrial structure beyond the stationary phase ( Figure 5).
When isolation of Wolbachia was attempted, it was found that the bacterium and mitochondria migrated together Uribe et al., 1985). Thus, it was decided to characterize oxidative phosphorylation activity in the mitochondria/Wolbachia mixture and then determine the contribution of each entity using different bioenergetics techniques. The rate of oxygen consumption was measured using ethanol as a substrate (Table 1). We isolated the mitochondria/Wolbachia fraction from either one-day cultures where there are very few Wolbachia cells or from 14-day cultures, where Wolbachia numbers were high (Table 1). In one-day cultures from ScW303 and wScW303 respiratory activities were very similar. However, at 14 days the rates of oxygen consumption and respiratory controls (RC) were widely different as follows: In noninfected yeast, both the rate of oxygen consumption and respiratory control decreased at the expense of state 3 inhibition, while in contrast, wScW303 retained high rates of oxygen consumption plus high respiratory controls, i.e. in 14-day old Wolbachia-infected yeast exhibited high oxidative phosphorylation activity, consistent with the presence of mitochondria observed by TEM in the infected cells (Table 1).

| In the presence of Wolbachia the activity of different mitochondrial respiratory complexes was preserved
In the isolated mitochondria/Wolbachia mixture, we tested specific substrates for each respiratory chain complex/enzyme (Table S3). In one-day cultures the rates of oxygen consumption were similar in infected and noninfected S. cerevisiae (Figure 6). In aged mitochondria from noninfected yeast, external NADH dehy- Complex IV and NADH-dependent oxygen consumption rates were still decreased as compared to mitochondria from one-day cultures ( Figure 6). Other respiratory substrates, namely glutamate and glutamine, which are used by Rickettsia (Winkler & Turco, 1988) where assayed and they did not support oxygen consumption. The respiratory activities measured indicate that the mitochondria/Wolbachia fractions from the infected and noninfected yeast consume the same substrates and are inhibited by the same respiratory chain inhibitors.
F I G U R E 5 Electron microscopy images of infected and noninfected Saccharomyces cerevisiae at different times of incubation. Transmission electron microscopy images confirm the intracellular location of Wolbachia. 10-day images were taken with uninfected (a) Sc and infected (b-c) wSc; 14 day-old images were taken with (d) ScW303 and (e-g) wScW303. Images show the presence of bacteria-like bodies (*) that are not present in uninfected yeast and mitochondria (m) whose cristae can be easily identified Reaction mixture: 0.6 mol L −1 mannitol, 5 mmol L −1 MES, pH 6.8, 4 mmol L −1 Pi, 10 mmol L −1 KCl. As substrate, 5 mmol L −1 ethanol. For state III, 1 mmol L −1 ADP.
TA B L E 1 Oxygen consumption rates of mitochondrial fractions from 1 and 14 day-old cultures of Wolbachia-infected (wScW303) and noninfected (ScW303) Saccharomyces cerevisiae cells | 11 of 16 URIBE-ALVAREZ Et AL.

| Under the experimental conditions tested, infected wScW303 oxygen consumption activity was mitochondrial
The experiments above suggested that either Wolbachia has the exact same electron transport chain as mitochondria or Wolbachia respiratory proteins may be damaged when the mitochondria/Wolbachia fraction is isolated and exposed to oxygen.
To explore this possibility further, we measured in-gel activities in the mitochondria/bacterium fraction. As eukaryote and prokaryote respiratory complexes I, II, III, and IV have different molecular masses the contribution from each organism to a given activity would be easily detected by native gel electrophoresis. The in-gel activities for each complex from infected and noninfected yeast from 1 and 14 day-old cultures were analyzed and, in all cases, activities were detected only at MWs corresponding to the mitochondrial enzymes (Table S3, Figure 7) suggesting that in the artificial ScW303/wAlbB system and under the specific conditions of growth reported here, Wolbachia did not express any functional respiratory chain proteins. The above results suggest that mitochondria were responsible for all the observed oxygen consumption activity. Still, one NADH dehydrogenase (Table S4) was weakly expressed making it impossible to conclude on whether different Wolbachia strains may be aerobic or not.

| F 1 F 0 -ATPase subunits from Wolbachia were detected in wScW303
In the in gel ATPase activity from the mitochondria/Wolbachia isolate no differential bands were observed. This was expected as the proposed MWs are similar for of both ATPases: 543 kDa for the eukaryote S. cerevisiae and 530 kDa for prokaryotes Escherichia coli and Paracoccus denitrificans (Bakhtiari, Lai-Zhang, Yao, & Mueller, 1999;Jonckheere, Smeitink, & Rodenburg, 2012; Morales-Rios, Montgomery, Leslie, & Walker, 2015;Robinson et al., 2013;Schagger, 2002). However, the ATPase activity band ( Figure 7A1, Table S4) sequence exhibited a mixture of yeast and Wolbachia ATPase proteins. BN and hrCN-PAGE results indicate that if Wolbachia expresses any electron transport chain proteins (still a possibility), under our experimental conditions their concentration was negligible when compared to the mitochondrial proteins and to its own F 1 F 0 -ATPase.

| Wolbachia remains infective against insect cell lines
After being cultured in a yeast cell, the question arose on whether Wolbachia remained viable and infective when isolated. To test this, we extracted Wolbachia from wScW303, incubated it in isolation for 5 days and then infected a C6C36 insect cell line which was previously reported to support bacterial infection .
Aged Wolbachia infection was successful as assessed by specific staining using FISH (Figure 8 Movie S3).

| D ISCUSS I ON
Growing obligate endosymbionts in cell lines yields low biomass at high costs Khoo et al., 2013 Since Wolbachia is an alpha-proteobacterium closely related to mitochondria, it seemed likely that the aerobic metabolic machinery of Wolbachia might mimic, enhance, or supplement the respiratory activity from the host. (Strübing et al., 2010). However, under our conditions, Wolbachia respiratory chain proteins were not detectable, instead, we found an increase in host mitochondrial activity. Another obligate endosymbiont, the Sytophilus oryzae Principal Endosymbiont (SOPE), has also been reported to increase the mitochondrial activity in the host, probably by providing nutrients such as riboflavin (Heddi, Lefebvre, & Nardon, 1993;Heddi et al., 1999). Several authors suggest that Wolbachia provides riboflavin or heme groups to their arthropod and nematode hosts (Brownlie et al., 2009;Darby et al., 2012;Foster et al., 2005;Wu et al., 2009). This may vary with strains as Wolbachia from Brugia malayi (wBm) contains complete sets of riboflavin, heme and nucleotide biosynthesis genes the filarial host lacks (Darby et al., 2012;Foster et al., 2005;Klasson et al., 2008;Wu et al., 2004). In return, the host provides amino acid, proteins and a safe, stable environment (Brownlie et al., 2009;Darby et al., 2012;Foster et al., 2005;Wu et al., 2009 (Mavingui et al., 2012) shows that some respiratory complex subunits are missing, e.g. nuoC and nuoD for a functional complex I (Sazanov, 2015); yet other Wolbachia sequenced genomes contain all the genes necessary for a functional electron transport chain (Klasson et al., 2008), so maybe under different growth conditions, hosts and Wolbachia strains, bacterial respiratory proteins may be detected. It is suggested that other Wolbachia strains should be tested in order to determine whether some consume oxygen.
In our hands, Wolbachia infection resulted in activation of mitochondria beyond the stationary growth phase. It may be speculated that such activation constitutes an advantage for Wolbachia either due to quenching of oxygen in the cytoplasm (Rosas-Lemus et al., 2016) or because Wolbachia needs high ATP that an active mitochondria provides (Potter, Badder, Hoade, Johnston, & Morten, 2016). It has already been suggested by experiments using paraquat that Wolbachia sensitivity to free radicals is higher than that of the host (Fallon et al., 2013) and it cannot survive outside a host cell unless it is kept in a 5% CO 2 atmosphere (Rasgon et al., 2006). Also, high agitation speeds, which would increase oxygen concentrations, lead to loss of the Wolbachia infection (Result not shown). Thus, it is possible that Wolbachia enters the cytoplasm to hide from high atmospheric oxygen and then it optimizes cell metabolism to both, use host metabolites and find low cytoplasmic oxygen concentrations. Avoidance, i.e. hiding from oxygen, is a common behavior in oxyconformers (Rosas-Lemus et al., 2016). In air, oxygen saturation concentration is ~21% (200 μmol L −1 ) while intracellular oxygen concentration ranges between 13.2% and 14% (126-133 μmol L −1 ) for rhabdomyosarcoma (RD) cells (Potter et al., 2016) or HEK293T cells (Abcam, 2016). When cells are exposed to lower ambient oxygen, intracellular oxygen concentration is also decreased: HEK293T cells exposed to 6% oxygen (50 μmol L −1 ) have an intracellular oxygen concentration below 2% (19 μmol L −1 ) (Abcam, 2016); RD cells exposed to 10% or 5% ambient oxygen reduce their intracellular oxygen concentration to 5.4% and 2.1% respectively (Potter et al., 2016). In addition, there is an intracellular oxygen gradient in the area surrounding the mitochondria in rat heart and hepatocytes, where oxygen concentration ranges between 3 (Gnaiger, 2003) and 6 μmol L −1 (Jones & Kennedy, 1982;Tamura, Oshino, Chance, & Silver, 1978). Thus, a mitochondrion-containing host such as cell lines and yeast would probably provide the endosymbiont with a microaerobic environment. The mechanism for the increase in host mitochondrial activity needs to be defined. In

CO N FLI C T O F I NTE R E S T
None declared.