|Home | About | Journals | Submit | Contact Us | Français|
We used Wolbachia pipientis strain wAlbB from Aedes albopictus Aa23 cells to infect clonal Ae. albopictus TK-6 cells, which are resistant to 5-bromodeoxyuridine. Infected TK-6 cells were cultured in medium containing 5-bromodeoxyuridine to select against Aa23 cells that might have persisted in the inoculum. Infected TK-6 lines retained the Wolbachia infection for 5 mo, indicating that their metabolic processes support Wolbachia growth and multiplication. To investigate early events after Wolbachia infection, we labeled infected cells with 35S[methionine/cysteine]. Patterns of labeled proteins on sodium dodecyl sulfate gels were similar in control and infected cells, with the exception of a 29-kDa protein. Tandem mass spectrometry revealed that the 29-kDa band included α and β subunits of the 26S proteasome. Independent confirmation of the up-regulation of the proteasome was established by probing Western blots with a monoclonal antibody to the proteasome-associated co-factor, ubiquitin. Wolbachia’s loss of metabolic pathways for the synthesis of most amino acids and retention of pathways for their uptake and metabolism suggest that proteasome activation provides a mechanism whereby controlled degradation of intracellular host proteins would increase availability of amino acids to support establishment and maintenance of the Wolbachia infection.
Genome analysis reveals that Wolbachia and other members of the Rickettsiales differ from free-living bacteria in their loss of most of the pathways for the biosynthesis of amino acids while retaining pathways for the uptake of amino acids as organic nutrients. Moreover, Wolbachia’s ability to metabolize amino acids suggests that this obligate intracellular microbe uses amino acids sequestered from its host as an energy source (Wu et al. 2004; Hotopp et al. 2006). These predictions provide a conceptual basis for the experimental investigation of metabolic interactions between Wolbachia and its host in cultured insect cell lines that maintain the Wolbachia infection.
The Aedes albopictus cell line known as Aa23 was derived from Wolbachia-infected Ae. albopictus mosquito embryos and is persistently infected with Wolbachia pipientis strain wAlbB (O’Neill et al. 1997). Aa23 cells have been a source of infectious inoculum that has been transferred to other insect cell lines (Dobson et al. 2002; Rasgon et al. 2006). Conversely, Aa23 cells can be made Wolbachia-free by treatment with tetracycline. Cells thus cured, designated Aa23T, can be reinfected with diverse strains of Wolbachia, including isolates from the fruit fly, Drosophila simulans, the mosquito, Culex pipiens, and the moth, Cadra cautella (Dobson et al. 2002). In aggregate, these in vitro studies indicate that Wolbachia can become established and persist over several passages in diverse insect cell lines regardless of their prior exposure to Wolbachia. Likewise, an individual cell line can productively support diverse strains of Wolbachia.
Most insect cell lines consist of a heterogeneous population of cell types. The Aa23 cell line contains at least two cell types (O’Neill et al. 1997) and has a relatively long population doubling time even in medium containing 20% fetal bovine serum (Fallon 2008; Fallon and Hellestad 2008). We began this study by investigating whether wAlbB from Aa23 cells could be stably transferred to a clonal population of TK-6 cells, which have a deficiency in thymidine kinase activity (Mazzacano and Fallon 1992, 1995). TK-6 cells were derived from Ae. albopictus C7-10 cells (Fallon 1997) after mutagenesis and selection for resistance to the thymidine analog, 5-bromodeoxyuridine (BrdU).
When previously uninfected (naïve) TK-6 cells were inoculated with wAlbB from Aa23 cells and labeled with 35S[methionine/cysteine], the predominant difference in labeled proteins was increased synthesis of an approximately 29-kDa protein. We analyzed this band using mass spectrometry, which revealed peptides corresponding to α and β subunits of the 26S proteasome, a large 2.4-MDa proteolytic complex responsible for the degradation of most intracellular proteins (DeMartino and Gillette 2007).
To verify that Wolbachia infection was accompanied by the up-regulation of the 26S proteasome, we took advantage of the 100% amino acid sequence identity between mosquito and bovine ubiquitin to probe Western blots with a commercially available antibody. Using either spot blots, or conventional Western blots, we showed increased protein ubiquitination in infected cells. These results support the hypothesis that Wolbachia growth and proliferation are accompanied by increased degradation of host cell proteins by the ubiquitin–proteasome pathway.
Cells were maintained in Eagle’s medium supplemented with nonessential amino acids, vitamins, glutamine, and penicillin/streptomycin as detailed by Shih et al. (1998). For TK-6 cells and infected derivatives, heat-inactivated fetal bovine serum was added to a final concentration of 5%; for Aa23 cells, the serum concentration was 20%. These formulations are called E-5 and E-20, respectively. Adaptation of Aa23 cells (O’Neill et al. 1997) to Eagle’s medium has been described previously (Fallon 2008). Cells were maintained at 28–30°C under a 5% CO2 atmosphere. For infection (see below), TK-6 cells were plated in advance and grown to confluence in 60-mm plates containing 4 ml of E-5 medium. Cells were typically labeled with 35S[methionine/cysteine], 25 μCi/ml (Tran [35S] Label; ~1,100 Ci/mmol; MP Biomedicals, Solon, OH). At appropriate time points, cells were resuspended in the culture medium, harvested by centrifugation, washed in phosphate-buffered saline (Dulbecco and Vogt 1954), and pellets were resuspended by sonication in 0.1% sodium dodecyl sulfate (SDS). Radioactivity was determined by scintillation counting, and samples containing equal amounts of label were reconstituted in SDS sample buffer and boiled prior to electrophoresis. For some experiments, samples were prepared in RIPA buffer (PBS containing 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) containing HALT protease inhibitor complex (Thermo Scientific, Waltham, MA), but no differences in labeling patterns were observed.
Stock cultures of Aa23 cells were diluted tenfold in two 180-cm2 flasks, each containing 40 ml of E-20 medium. Initially, a robust patchy monolayer forms; after 2 wk, when cells formed attached aggregates (Fallon 2008), an additional 40 ml of E-20 was added to each flask. Cell growth was monitored microscopically over the course of 2 to 4 wk, during which the patchy monolayer from the initial seeding, plus additional cells that grow after refeeding, converted to aggregates of floating cells. Aggregates and attached cells were pooled, harvested by centrifugation, resuspended in 16 ml (0.1 volume relative to starting material) of E-5 medium, and distributed in 1 ml aliquots to sterile tubes for sonication (5 s in a Kontes GE 70.1 ultrasonic processor, at an amplitude setting of 40) under sterile conditions. Sonicated material was pooled, filtered in 5-ml portions through 2.7-μm syringe filters, and added to TK-6 recipient cells (0.3 ml of inoculum per 60-mm plate containing 4 ml of E-5 medium). A slide was made for Giemsa staining, and the infected status of the inoculum was verified by light microscopy. For control treatments, the inoculum was boiled for 5 min and cooled on ice prior to addition to the cells.
Three infected lines, designated TW-180, 280, and 380, were obtained from infected TK-6 cells after two rounds of selection in E-5 medium containing BrdU at a final concentration of 80 μg/ml. The BrdU selection was initiated 7 d after infection and continued for 4 wk, after which the cells were subcultured by tenfold dilution into fresh E-5 medium at 10-d intervals.
A ~29-kDa Coomassie blue-stained band (Fig. 2) that directly correlated to the mass as well as increased expression observed in the methionine-labeled sample was excised and subjected to tryptic digestion (Kinter and Sherman 2000). Peptides were rehydrated in water/acetonitrile (ACN)/formic acid (FA) 95:5:0.1 and loaded using a Paradigm AS1 autosampler system (Michrom Bioresources, Inc., Auburn, CA). Each sample was subjected to Paradigm Platinum Peptide Nanotrap (Michrom Bioresources, Inc.) pre-column (0.15×50 mm, 400-μl volume) followed by an analytical capillary column (75 μm×12 cm) packed with C18 resin (5 μm, 200 Å Magic C18AG, Michrom Bioresources, Inc.) at a flow rate of 250 nl/min. Peptides were fractionated on a 60 min (10–40% ACN) gradient on a flow MS4 flow splitter (Michrom Bioresources, Inc.). Mass spectrometry (MS) was pre-formed on a LTQ (Thermo Electron Corp., San Jose, CA). Ionized peptides eluting from the capillary column were subjected to an ionizing voltage (2.0 kV) and selected for MS/MS using a data-dependent procedure alternating between an MS scan followed by four MS/MS scans for the four most abundant precursor ions.
Sequest (ThermoFinnigan, San Jose, CA, version 27, rev. 13) and X! Tandem (www.thegpm.org; version 2007.01.01.1) were used to search the Aedes_NCBI_011508_CTM database, 37,114 entries. The search was repeated with a combined Aedes, Wolbachia reference sequence database (rs_wolbachia_aedes_v200808_cRAP database, 23,725 entries) to determine if any Wolbachia proteins were present in the excised band. Sequest and X! Tandem were searched with a fragment ion mass tolerance of 1.00 Da and a parent ion tolerance of 1.00 Da. Oxidation of methionine was specified as a variable modification. Scaffold (version Scaffold-01_07_00, Proteome Software Inc., Portland, OR) was used to validate MS/MS-based peptide and protein identifications with filters set to a protein identification probability, 99.9%; minimum number of peptides, 3; and minimum identification probability, 95%. By these criteria, 13 proteins were identified, including the trypsin precursor, mosquito actin (42 kDa), and bovine serum albumin (66 kDa), which were excluded from further consideration.
Samples were electrophoresed on 12% polyacrylamide SDS gels as described by Laemmli (1970). Gels were stained in Coomassie brilliant blue, photographed, dried, and exposed to X-ray film for autoradiography or transferred directly to Whatman PRO-TRAN nitrocellulose (0.45-μm pore size; PerkinElmer Life and Analytical Sciences, Boston, MA) for Western blots. Filters were blocked by incubation in 5% nonfat dry milk containing 0.1% Tween-20 at room temperature for 1 h. Primary antibody [mouse monoclonal antibody to full-length bovine ubiquitin (Ub-P4D1) from Santa Cruz Biotechnology, Inc., Santa Cruz, CA, catalog # sc-8017] was added at a 1:1,000 dilution, and incubation was continued overnight at 4°C, with gentle shaking. Filters were washed in 20 mM Tris–HCl, pH 7.5, containing 0.5 M NaCl and 0.1% Tween-20 for 15 min and three times for 5 min. Filters were placed in 5% milk containing 0.1% Tween 20 and secondary antibody (goat anti-mouse IgG, coupled to horse radish peroxidase; sc-2005) at a 1:1,000 dilution for 1 to 3 h at room temperature, washed as described above, and developed using SuperSignal West Pico detection system (Thermo Scientific) and Kodak Biomax X-OMAT AR film, typically with a 5-min exposure.
In contrast to Aa23 cells, which have a doubling time of 45–70 h in medium containing 20% serum (Fallon and Hellestad 2008), TK-6 cells have a doubling time of 30–32 h in medium containing 5% serum (Mazzacano et al. 1992). We reasoned that their more rapid growth and reduced serum requirement might facilitate investigation of metabolic events accompanying Wolbachia infection. Moreover, passage of infected cells in medium containing BrdU provided a rigorous means of eliminating any Aa23 cells that persisted through preparation of the inoculum.
We evaluated proliferation of Wolbachia from Aa23 cells in three lines (designated TW-180, TW-280, and TW-380) derived from infected TK-6 cells by assaying polymerase chain reaction amplifiable Wolbachia DNA in the cell cultures. Two primer sets, designed to amplify Wolbachia DNA and host cell DNA targets of similar size, were used in a single reaction as described previously (Fallon 2008). As was the case with Aa23 cells, Wolbachia DNA persisted in the culture medium of infected TW-280 cells (Fig. 1, passages 2 through 12), while cellular DNA degraded during prolonged culture and was barely detectable in the early samples (Fig. 1, passages 2 through 7). By passage 12, Wolbachia DNA was barely detectable, and levels declined as passage continued, while the band representing cellular DNA remained robust. Lines TW-180 and TW-280, which were established earlier and were therefore at a higher passage number than the TW-380 line, spontaneously lost wAlbB after approximately 5 mo. Because the three TW lines were comparable in their cytological and growth properties, we discontinued the experiment before the TW-380 cells were subcultured to the passage number at which wAlbB was lost from TW-180 and TW-280 cells. The wAlbB infection is considerably more stable in Aa23 cells, which, despite variability in infection levels (O’Neill et al. 1997; Fallon 2008), has remained stable in our hands for over 2 yr. Overall, the Aa23 line has maintained Wolbachia for more than a decade, during which time it has been distributed to a number of laboratories and maintained in a variety of culture media.
Long-term persistence of wAlbB in Aa23 cells over years, as opposed to months for the TK-6 cells, may reflect subtle physiological adaptations to Wolbachia in the infected embryonic tissues from which the Aa23 line was derived. Likewise, eventual loss of infection in TW-180 (not shown) and TW-280 (Fig. 1) cells suggested that the metabolic accommodations between wAlbB and the naïve host cell are less well-adjusted, resulting in eventual loss of the Wolbachia infection. We reasoned that this instability might facilitate detection of changes in protein synthesis during establishment of wAlbB during the first week after inoculation.
To investigate intracellular changes early after infection, we labeled infected cells with 35S[methionine/cysteine] and harvested cells at various times to examine changes in protein synthesis on SDS polyacrylamide gels. Against a background of largely identical protein labeling, a single ~29-kDa band showed substantially increased radioactivity in infected cells, relative to cells treated with boiled inoculum (Fig. 2, left panel, note the open arrow). This band was also detectable by Coomassie blue staining, and a corresponding band was absent in the control lane (Fig. 2, right panel). In independent experiments (not shown), we verified that TK-6 cells treated with boiled inoculum and untreated TK-6 cells expressed identical patterns of labeled proteins. The up-regulated 29-kDa band was detected in more than six independent experiments initiated with independent preparations of live inocula from Aa23 cells, suggesting that the response was invariant.
When the 29-kDa stained band was recovered from SDS gels and processed for mass spectrometry, MS/MS data analyzed at the highest levels of stringency in the Scaffold protein analysis software revealed ten peptides that originated from mosquito proteins with an approximate mass of 29 kDa. The highest ranking non-ribosomal protein candidate was the β subunit of the 26S proteasome (accession no. EAT45312), which had an identity probability of 100% for each of two preparations, with five unique peptides and 20% (56/280 amino acids) protein coverage, indicated by boxed residues in Fig. 3. The β subunit has a calculated molecular mass of 30.4 kDa. The α subunit (accession no. ABF18446) was also present among the top ten candidates, with probabilities of 100% and 94% for the two samples, 15% coverage in one sample (Fig. 3), and 4% coverage in the other. Reduced coverage of the 26.9-kDa α subunit is consistent with its lower mass, which would reduce its representation in the excised band. These bands were of particular interest because they are components of a multi-protein complex, the 26S proteasome, whose activity potentially increases the availability of amino acids to Wolbachia via controlled degradation of host cell proteins.
We anticipated that both Wolbachia and Aedes proteins would be present in the band recovered after SDS gel electrophoresis. However, when search parameters were expanded to include the reference sequence database of Wobachia, no proteins with a high confidence level were identified, which suggests that either the excised band contained no Wobachia protein or that levels of Wobachia protein contained in the excised slice were below the detection range of the mass spectrometer.
Because proteasome activity is responsible for normal turnover of cytoplasmic proteins during cellular homeostasis, we sought independent experimental evidence for increased proteasome activity in infected cells. Proteins destined for degradation are targeted to the 26S proteasome by covalent attachment of the 76 amino acid co-factor, ubiquitin. Amino acid sequence identity between mosquito and bovine ubiquitin allowed use of a monoclonal antibody to bovine ubiquitin to probe Western blots of extracts from infected and uninfected mosquito cells. In Fig. 4, the left panel shows that a large number of proteins in infected cells show increased labeling with ubiquitin (lanes 2, 3, and 4), relative to uninfected controls (lane 1). Note that 35S [methionine/cysteine]-labeled bands in Fig. 4, right panel, exposed from the filter used for the Western blot verifies that lanes 1 through 4 contain comparable levels of transferred protein. Comparison of the Western blot with the labeled bands also reveals that distribution of ubiquitinated proteins differs from that of labeled bands, suggesting that proteins are not uniformly destined for proteolytic degradation.
The present study provides a novel insight into a mechanism by which Wolbachia interacts with host cellular machinery in a naïve cell line. Continual turnover of proteins by regulated proteolysis is an important component of overall cellular homeostasis, growth, and metabolism. Proteins destined for degradation by the 26S proteasome are tagged by covalent attachment of the 76 amino acid cofactor, ubiquitin. Within the 24-MDa complex, the core 20S proteasome houses a proteolytic chamber in which the α and β subunits mediate catalytic degradation of ubiquitin-tagged proteins, releasing peptides of three to 20 residues. These peptides are further hydrolyzed to amino acids by cytoplasmic enzymes, while ubiquitin itself is recycled (for reviews, see Lee and Goldberg 1998; Lecker et al. 2006; DeMartino and Gillette 2007).
While this work was in progress, Brennan et al. (2008) reported use of a proteomics approach based on two-dimensional polyacrylamide gels to identify differential up-regulation of host antioxidant proteins in Aa23 cells compared with Aa23 cells cured after seven passages in the presence of rifampicin. In their analysis, proteins with pI values ranging from 5 to 8 were separated on a second dimension gel with discrimination between 47 and 12 kDa. Although the α subunit of the proteasome has a pI of 4.72, and would be undetectable on their gels, the β subunit has a calculated pI of 6.34, assuming absence of phosphorylation. Brennan et al. (2008) do not describe an up-regulated protein that matches these parameters. Likewise, in studies related to those reported here, we have not detected the 29-kDa band in tetracycline-cured Aa23 cells reinfected with Wolbachia, while the band was present when naïve C7-10 cells were infected with Wolbachia. We suggest that the absence of the 29-kDa band in reinfected Aa23 cells may reflect unique reciprocal adaptations between Wolbachia and its host cells that evolved during establishment of the infected Aa23 cell line.
The overall similarity in protein labeling in Wolbachia-infected cells compared with uninfected control cells and the absence of Wolbachia proteins in the excised 29-kDa band support the hypothesis that the metabolic needs of Wolbachia impose minimal disruption of host cell metabolism once the infection becomes established. It remains to be determined whether activation of intracellular protein degradation via the proteasome remains detectable as the naïve cell becomes adapted to Wolbachia and whether this activity imposes a subtle metabolic cost that leads to the eventual loss of Wolbachia from TW-280 cells (Fig. 1). We anticipate that the identification of ubiquitin-tagged proteins in infected TK-6 cells, and the evaluation of their roles in cell metabolism, will lead to an improved understanding of this widespread bacterium and its diverse effects on insect reproduction.
This work was funded by NIH grant AI 070913 and by the University of Minnesota Agricultural Experiment Station, St. Paul, MN. Protein analysis was done at the University of Minnesota Mass Spectrometry Consortium for the Life Sciences & Proteome Analysis Core Facilities.
Ann M. Fallon, Department of Entomology, University of Minnesota, 1980 Folwell Ave., St. Paul, MN 55108, USA, e-mail: ude.nmu@200ollaf.
Bruce A. Witthuhn, Center for Mass Spectrometry and Proteomics, University of Minnesota, St. Paul, MN, USA.