Wolbachia spp. are maternally inherited obligate intracellular bacteria belonging to the α-
Proteobacteria. They infect a broad range of insect species, a number of noninsect arthropods such as isopods and mites, and most species of filarial nematodes (
3,
34,
37,
38). In arthropods they have been implicated in several host reproductive modifications, including cytoplasmic incompatibility in various insect species (
16), parthenogenesis in wasps (
33), feminization in isopods (
29), and virulence in
Drosophila melanogaster(
25). Within the Nematoda, it appears that
Wolbachia spp. are required for fertility and normal development of the filarial worms they infect (
22,
34).
In preparation for complete genome sequencing, we have determined the genome sizes of a number of Wolbachia strains using pulsed-field gel electrophoresis (PFGE) and have developed a method to rapidly purify Wolbachia chromosomal DNA in quantities sufficient for library construction.
MATERIALS AND METHODS
Wolbachia strains.
The seven
Wolbachia strains used in this study are listed in Table
1.
Drosophila simulansRiverside previously treated with tetracycline (DSRT) was used as a
Wolbachia-free control insect strain.
Wolbachia purification fromDrosophila.
Drosophila organisms were reared on standard corn flour-sugar-yeast medium at 25°C. Young adults were harvested for extraction of
Wolbachia, except for
Drosophila melanogaster w1118, which harbors
wMelPop. The infection density in this strain rises dramatically with age (
25). Newly emerged adults of this strain were transferred to standard egg-laying bottles for aging. Twenty-day-old flies were harvested for
wMelPop purification.
All purification methods published to date have been unable to separate
Wolbachia from
Drosophila mitochondria. In this report, the methods used to purify mitochondria from
Drosophila (
28,
35) were modified to prepare DNA from
Wolbachia in quantities that could be visualized by ethidium bromide staining of agarose gels. Around 5 ml of adult flies (about 1,000) was collected and then homogenized in buffer as previously described (
9), except without Lubrol (90 mM KCl, 55 mM CaCl
2, 15 mM MgSO
4, 30 mM NaCl, 250 mM sucrose) using a Dounce tissue grinder (Wheaton, Millville, N.J.). The homogenate was filtered through a 95-μm-pore-size nylon mesh. The filtrate was centrifuged at 200 × g
maxfor 25 min at 4°C to pellet
Drosophila nuclei. The supernatant was then centrifuged at 4,100 × g
max for 5 min at 4°C to pellet
Wolbachia. The pellet was resuspended at 56°C in a mixture consisting of 1 volume of Tris-EDTA (TE) plus 1 volume of 2% molten GPG low-melting-point agarose (American Bioanalytical, Natick, Mass.), and the resuspension was loaded into a plug module (Bio-Rad, Hercules, Calif.). Plugs were treated with 40 μg of DNase I (Roche, Basel, Switzerland)/ml in DNase I reaction buffer (10 mM Tris-HCl [pH 8.0], 1 mM MgCl
2) for 40 min at room temperature (RT) (25°C). After DNase I treatment, the plugs were incubated overnight at 56°C in the lysis buffer (
2) (100 mM EDTA [pH 8.0], 10 mM Tris-HCl [pH 8.0], 1% [wt/vol]
N-lauroylsarcosine sodium salt [Sigma, St. Louis, Mo.], 200 μg of proteinase K [Roche]/ml). The plugs were stored in this lysis buffer at 4°C before restriction digestion or electrophoresis.
Wolbachia purification from nematodes.
Adult female nematodes were selected for purification of Wolbachianot only because they are larger than certain other life cycle stages (e.g., microfilariae) that also have high densities of endosymbionts but also because they were more readily and uniformly homogenized using the Dounce tissue grinder. Typically, 5 mature adult females ofDirofilaria immitis or approximately 275 mature females of the smaller Brugia malayi were used for extractions. The purification procedure for Wolbachia from nematodes was essentially the same as the one for Wolbachia from insects, but with the modifications discussed below.
Live worms, supplied by TRS Laboratories (Athens, Ga.), were placed in a petri dish on ice and chopped into small pieces using a sterile razor blade. The homogenization buffer used was physiological saline (0.85% NaCl) supplemented with 0.001% Nonidet P-40 detergent (Sigma). Inclusion of this very low concentration of detergent in the homogenization buffer was found to decrease the amount of
Wolbachia pelleting with the worm tissue fragments without causing any noticeable increase in degradation of DNA. The filtrate was passed through two layers of cheesecloth (Veratec, Walpole, Mass.). The first centrifugation was carried out at 350 × g
max for 25 min at 4°C to pellet nematode nuclei. The supernatant was then centrifuged at 4,100 × g
max for 5 min at 4°C to pellet
Wolbachia. This final pellet was resuspended in an equal volume of saline without detergent and 2 volumes of molten 2% SeaPlaque (low-melting-point) agarose (FMC, Rockland, Maine) in 0.5× Tris-borate-EDTA (TBE) to give a final concentration of 1% agarose. The sample was allowed to set in 75-μl plug molds (Bio-Rad). Following DNase I treatment, the plugs were transferred to proteinase K lysis buffer (
6) (0.5 M EDTA [pH 8.0], 1% lauroylsarcosine, sodium salt supplemented with 2 mg of proteinase K [Gibco BRL, Gaithersburg, Md.]/ml) and incubated at 55°C for 48 h. The proteinase K was diffused out of the plugs by performing a minimum of six washes, each of 30 min, in TE at RT. The plugs were stored short term in TE at 4°C.
Optimization of DNase I treatment.
After formation of plugs, DNase I was used to digest any fragmented DNA produced during homogenization. Several concentrations of DNase I (10, 20, 30, 40, 50, and 60 μg/ml) in combination with a time course (0, 5, 10, 15, 22, 28, 35, 40, or 50 min) were used at RT to determine the best conditions for digestion. The limited amounts of nematode sample precluded optimization of DNase I treatment as was carried out for theWolbachia from Drosophila. The conditions determined optimal for Drosophila (40 μg of DNase I/ml for 40 min at RT) were applied.
Plug preparation from cell line culture.
Aedes albopictus cell line Aa23 containing
Wolbachia strain
wAlbB was used (
27). Cells were maintained in a 25-mm
2 flask at 25°C in 5 ml of medium (45% Mitsuhashi and Maramorosch insect medium [Sigma], 45% Schneider medium [Sigma], and 10% heat-inactivated fetal bovine serum). The cells were harvested and washed in phosphate-buffered saline twice. The end pellet was then used to make agarose plugs as described above and directly placed into the lysis buffer and treated at 56°C overnight.
Restriction digestion of Wolbachia genome DNA.
The Drosophila Wolbachia plugs were washed twice with TE buffer, treated twice with TE buffer supplemented with 40 μg of phenylmethylsulfonyl fluoride (Life Technologies, Rockville, Md.)/ml to inactivate the proteinase K, and then washed twice with TE buffer again. All washes were done for 30 min each at RT. For the nematodeWolbachia plugs, the proteinase K was diffused out of the plugs before they were stored in TE buffer.
The Drosophila Wolbachia plugs were equilibrated with restriction endonuclease buffer for 1 h at RT and transferred to fresh buffer for endonuclease digestion overnight. Four restriction enzymes were used: AscI (GG∧CGCGCC),ApaI (GGGCC∧C),FseI (GGCCGG∧CC), and SmaI (CCC∧GGG) (New England Biolabs, Beverly, Mass.). The nematodeWolbachia plugs were equilibrated on ice for 2 h in restriction enzyme buffer containing 50% of the final number of enzyme units. The remaining 50% of the enzyme was then added, and the endonuclease digestion continued for 3 h at the appropriate reaction temperature.
PFGE.
Contour-clamped homogeneous electric field (CHEF) (
10) gels were run to separate DNA fragments that included at least one fragment with a size greater than 50 kb, using either a CHEF Mapper XA (Bio-Rad) or a CHEF-DR II (Bio-Rad). For resolution of DNA fragments of less than 50 kb, field inversion gel electrophoresis (
7) gels were used with only the CHEF Mapper XA. All of the electrophoresis was carried out at 14°C using 0.5× TBE as the running buffer. Plugs prepared from filarial nematode samples were equilibrated in 0.5× TBE prior to electrophoresis, while
Drosophila Wolbachia plugs underwent no treatment before electrophoresis. The migration profiles were determined using CHEF Mapper XA interactive software, version 1.2 (Bio-Rad). Fragment lengths and the presence of multiple fragments were determined using Gel-doc and Quantity One one-dimensional analysis software (Bio-Rad).
Southern hybridization.
The
Wolbachia surface protein (
wsp) gene fragment from
wRi was amplified by PCR using total DNA from DSRT flies as the template and
wsp-specific primers 81F (5′-TGGTCCAATAAGTGATGAAGAAAC-3′) and 691R (5′-AAAAATTAAACGCTACTCCA-3′) as a probe (
9). The nematode
ftsZ gene fragment was amplified with total DNA from
B. malayi as the template and with
ftsZ-specific primers
ftsZ1F (5′-GTTGTCGCAAATACCGATGC-3′) and
ftsZ1R (5′-CTTAAGTAAGCTGGTATATC-3′) as a probe (
39). The mitochondrial 12S rRNA gene fragment was amplified with primer pair 12SAI (5′-AAACTAGGATTAGATACCCTATTAT-3′) and 12SBI (5′-AAGAGCGACGGGCGATGTGT-3′) (
32). The amplification conditions were the same as those previously described (
9). Total DNA of
Drosophila was extracted using the Holmes-Bonner method (
18). PCR products were gel purified with either β-agarase (New England Biolabs) or Qiagen (Valencia, Calif.) gel extraction kits. These probes were radioactively labeled using either the Random Primed DNA labeling kit (Roche) or the NEBlot kit (New England Biolabs) according to the manufacturer's instructions.
A cocktail of seven probes derived from
wBma and known from preliminary mapping to be well dispersed around the
Wolbachia genome was also prepared. These were a 16S rRNA fragment, a 23S rRNA fragment, HSP-60 (GroEL homologue), a DNA mismatch repair protein homologue, the DNA polymerase III γ subunit, the RNA polymerase β subunit, and serine hydroxymethyltransferase. These
Wolbachia sequences had been identified among the expressed sequence tags reported from
B. malayi as part of the Filarial Genome Project (see
http://neb.com/fgn/filgen1.html ). The sequences were amplified from the appropriate phage stocks representing these cDNA clones using T3 and T7 primers (New England Biolabs). The PCR products were precipitated to remove excess primers, and nucleotides and then labeled by hot PCR (
15) with the same primers but using a nucleotide mixture that contained dATP at 1/10 the concentration of the other deoxynucleoside triphosphates but that included [
32P]dATP. The labeled PCR products were purified using QIAquick PCR purification columns (Qiagen) according to the manufacturer's instructions and pooled to make a
Wolbachia probe cocktail.
After gel electrophoresis, Southern transfer was done with a VacuGene XL vacuum blotting system (Amersham Pharmacia Biotech, Uppsala, Sweden), and the filters were hybridized at either 60 or 65°C and washed under high-stringency conditions.
DISCUSSION
Previous studies of
Wolbachia have focused heavily on ultrastructure, reproductive phenotypes, and phylogeny. In the past, the difficulty in culturing and purifying the bacteria has hindered the progress of genetic and biochemical studies. A new protocol based on the purification of
Drosophila mitochondria proved to be suitable for purification of
Drosophila Wolbachia. Three modifications were key: (i) a change to the composition of the homogenization buffer, (ii) incorporation of a DNase I digestion step to obtain purer
Wolbachia DNA, and (iii) the use of
Drosophila adults as starting material. The homogenization buffer had previously been shown to be effective in separating
Wolbachia from host materials (
9). The detergent Lubrol was removed from the original recipe, however, since its presence increased degradation of DNA (data not shown). A digestion step with DNase I was added to remove sheared DNA, generated during homogenization. Addition of the DNase I step also appeared to remove contaminating host mitochondrial DNA from the preparation. Levels of purification of
Drosophila Wolbachia from both adults and embryos were compared, and the latter generated very poor results (data not shown).
Bacterial chromosomes demonstrate different forms by PFGE studies; not all bacteria have circular genomes (
5,
13). In circular forms, DNA with a large size is not expected to migrate into pulsed-field gels (
31). As such, the fragments resolved on PFGE gels without restriction digestion (Fig.
1a and
3) are likely to be the result of nicking during homogenization. This is also consistent with the observation that most of the
Wolbachia DNA was retained in the loading wells (Fig.
1a and
3). Furthermore,
FseI was determined to be a single cutter for the
wMelPop strain. If the genome of
wMelPop were linear, then
FseI digestion should have resulted in two fragments (complete digestion) or three fragments (partial digestion). However, digestion resulted in a single fragment. Similarly, digestion with
AscI produced two fragments. Comparing restriction patterns of
wMelPop from single digestion with
ApaI or
SmaI to those from double digestions with
ApaI and
AscI or
SmaI and
AscI clearly showed that
AscI cut the chromosome in two places. These data strongly suggest that the
Wolbachia chromosome is circular.
Studies using either PFGE or whole-genome sequencing have revealed a diversity of bacterial genome sizes, ranging from as low as 0.58 Mb to as high as 9.5 Mb. For all characterized bacterial genomes, the sizes of free-living species are generally larger than the sizes of intracellular species. Within the α-
Proteobacteria, reported genome sizes of the free-living species are typically above 3.0 Mb: 3.8 Mb for
Rhodobacter capsulatus (
14), 3.8 to 4.0 Mb for
Caulobacter crescentus (
11,
12), 3.4 Mb for
Rhizobium meliloti (
19), and 8.7 Mb for
Bradyrhizobium japonicum (
21). The strictly obligate species, on the other hand, typically have genome sizes below 2.0 Mb: 1.6 Mb for
Bartonella bacilliformis (
20), 1.1 Mb for
Rickettsia prowazekii, and 0.9 to 1.5 Mb for
Ehrlichia spp. (
30). Consistent with these previous results we have also demonstrated reduced genome sizes for
Wolbachia: 0.95 and 1.1 Mb for the
Wolbachiainfecting nematodes and 1.4 to 1.6 Mb for the different A group
Wolbachia strains infecting
Drosophila.
At the present time four major monophyletic clades of
Wolbachia are recognized and referred to as
Wolbachia groups A, B, C, and D. The A and B groups are found in a range of arthropods and crustaceans. The C and D groups are restricted to filarial nematodes. Infections with A and B group
Wolbachia strains are associated with various parasitic traits that indicate a conflict between their own vertical transmission and the normal reproduction of their host. In these cases
Wolbachia has evolved various mechanisms to increase its vertical transmission including cytoplasmic incompatibility, parthenogenesis, and feminization phenotypes in the hosts they infect (
24). In addition phylogenetic and experimental studies indicate that these infections are capable of moving horizontally among hosts, albeit as presumably rare events (
26). In contrast C and D group
Wolbachia strains infecting nematodes appear to be more like classical mutualists, being required for normal reproduction and development of their hosts, presumably through the supply of metabolic products required by the worm (
4,
22). In addition the phylogeny of these
Wolbachia strains mirrors that of the host worms, indicating a long period of concordant evolution between host and symbiont (
3). At the present time, the lack of a suitable outgroup has prevented resolution of the evolutionary relationships among the four
Wolbachia clades.
The large difference between the genome sizes of representatives from these different groups is intriguing. Nematode
Wolbachiastrains have a genome 30% smaller than those of the A group counterparts. The reduction in the genome sizes of these strains is consistent with the reduced genome sizes reported for other mutualistic symbionts (
1,
36).
ACKNOWLEDGMENTS
We thank Tetsuhiko Sasaki, Henk Braig, and Melinda Pettigrew for technical assistance, Serap Aksoy for providing a CHEF-DR II apparatus, Liangbiao Zheng for providing a gel documentation system, and Elizabeth McGraw for suggestions on the drafts.
This work was supported by grants from the National Institutes of Health (AI40620 and AI47409), the McKnight Foundation, New England Biolabs, and the UNDP/World Bank/WHO program for Research and Training in Tropical Diseases.