Determination of Wolbachia Genome Size by Pulsed-Field Gel Electrophoresis

Abstract

Genome sizes of six different Wolbachia strains from insect and nematode hosts have been determined by pulsed-field gel electrophoresis of purified DNA both before and after digestion with rare-cutting restriction endonucleases. Enzymes SmaI, ApaI, AscI, and FseI cleaved the studied Wolbachia strains at a small number of sites and were used for the determination of the genome sizes of wMelPop, wMel, and wMelCS (each 1.36 Mb), wRi (1.66 Mb), wBma (1.1 Mb), and wDim (0.95 Mb). The Wolbachia genomes studied were all much smaller than the genomes of free-living bacteria such as Escherichia coli (4.7 Mb), as is typical for obligate intracellular bacteria. There was considerable genome size variability among Wolbachia strains, especially between the more parasitic A group Wolbachia infections of insects and the mutualistic C and D group infections of nematodes. The studies described here found no evidence for extrachromosomal plasmid DNA in any of the strains examined. They also indicated that the Wolbachia genome is circular.

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).

Most research attention to date has been focused on the phenomenology of Wolbachia-host interactions. Little is known about the molecular genetics of Wolbachia. For example, there has been little characterization of Wolbachia genes other than a few loci that have been cloned and used mainly for phylogenetic purposes. This is largely due to the fastidious nature of Wolbachia and the difficulty in obtaining large amounts of pure material for laboratory studies.

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 simulans Riverside previously treated with tetracycline (DSRT) was used as a Wolbachia-free control insect strain.

TABLE 1.

Wolbachia strains studied

Strain	Host	Phenotype
wRi	D. simulans Riverside	mod+ resc+
wMelCS	D. melanogaster Canton-S	mod− resc+
wMel	D. melanogaster yw67c23	mod+ resc+
wMelPop	D. melanogaster w1118	Virulence
wDim	D. immitis	Mutualism
wBma	B. malayi	Mutualism
wAlbB	A. albopictus	mod+ resc+

Wolbachia purification from Drosophila.

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 CaCl₂, 15 mM MgSO₄, 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_max for 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 MgCl₂) 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 Wolbachia not 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 of Dirofilaria 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 the Wolbachia 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-mm² 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 nematode Wolbachia 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 nematode Wolbachia 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.

RESULTS

Wolbachia purification.

The methods used led to the successful enrichment of Drosophila Wolbachia. When run on a PFGE gel, Wolbachia genomic DNA could be resolved with ethidium bromide staining (Fig. 1a, lane 2, b, lanes 2 and 3, and c, lane 2). This band was not visible in preparations from uninfected Drosophila. Moreover, Southern hybridization using a wsp gene fragment as the probe confirmed that the visible band represented Wolbachia genomic DNA (Fig. 1a, lane 3, b, lanes 4 and 5, and c, lane 3). This procedure also showed that much of the uncut Wolbachia DNA still remained in the loading wells (Fig. 1). Southern hybridization using labeled total DNA from Wolbachia-free strain DSRT as a probe produced a faint background smear, which indicated the presence of trace amounts of degraded Drosophila DNA but which did not hybridize to any distinct fragments. No extrachromosomal DNA was detected on the gel, indicating an absence of plasmids.

FIG. 1.

Composite ethidium bromide-stained gel and corresponding autoradiograph of Southern blot probed with a wRi wsp gene fragment. (a) Lane 1, yeast chromosomal size marker; lane 2, undigested wMelPop genome fragment; lane 3, Southern blot. (b) Lane 1, yeast chromosomal size marker; lane 2, undigested wMelCS genome fragment; lane 3, undigested wMel genome fragment; lane 4, Southern blot (undigested wMelCS genome fragment); lane 5, Southern blot (undigested wMel genome fragment). (c) Lane 1, yeast chromosomal size marker; lane 2, undigested wRi genome fragment; lane 3, Southern blot.

Optimal DNase I reaction conditions were found to be 40 μg of DNase I/ml for 40 min of treatment at RT. A time course study with DNase I (40 μg/ml) showed that the background DNA smear on gels gradually decreased and finally disappeared with the increasing reaction time (data not shown). Southern hybridization with a mitochondrial DNA 12S rRNA gene fragment as the probe showed that the mitochondrial genome migrated at approximately 500 kb (owing to its circular conformation) and was cleared with increasing DNase I treatment (Fig. 2). Excessive DNase I treatment was found to digest Wolbachia DNA completely as well. The final conditions used were found to remove mitochondrial DNA while still maintaining a good yield of Wolbachia DNA.

FIG. 2.

Autoradiograph of Southern blot probed with mitochondrial 12S rRNA gene fragment. Shown is a DNase I (40 μg/ml) reaction time course at RT (25°C). Lane 1, 15 min; lane 2, 22 min; lane 3, 28 min; lane 4, 35 min; lane 5, 40 min. Arrow, Drosophila mitochondrial genome fragment.

The methods described above also led to the successful enrichment of Wolbachia from the filarial nematodes. Staining pulsed-field gels with ethidium bromide showed bands of around 1,100 and 950 kb for extracts made from B. malayi and D. immitis, respectively (Fig. 3). There was a sufficient amount of Wolbachia DNA in the extracts made from D. immitis worms for hybridization with the ftsZ gene fragment probe, confirming that the 950-kb band represented Wolbachia genomic DNA. For the extracts made from the smaller B. malayi worms it was usually necessary to use the Wolbachia probe cocktail to demonstrate that the 1,100-kb band was the Wolbachia genome (Fig. 3). The majority of uncut Wolbachia was found to remain in the well of the gel, paralleling the situation observed for extracts from Drosophila.

FIG. 3.

Pulsed-field gel sizing of Wolbachia genomes from D. immitis (a) and B. malayi (b). Lanes 1, ethidium bromide staining of the uncut genomes; lanes 2, corresponding autoradiographs, which were probed with the ftsZ gene fragment (a) and with the Wolbachia probe cocktail (b); lanes 3, autoradiographs of the genomes after digestion with ApaI and probing with the Wolbachia probe cocktail. Sizes of selected DNA standards (yeast chromosome and MidRange II pulsed-field gel markers; New England Biolabs) are indicated.

Restriction digestion of Wolbachia genomic DNA.

The sequences of several genes previously cloned from Wolbachia strains suggest that the genomes of these bacteria are A+T rich (8, 9, 17, 23). Therefore, to find rare-cutting restriction enzymes, we utilized those with six- or eight-base GC recognition sites initially using wMelPop and wRi genomic DNA.

Out of a total of 15 restriction enzymes screened, ApaI, SmaI, AscI, and FseI produced small numbers of fragments that facilitated the calculation of the genome size for wMelPop. Among them, ApaI and SmaI cut wMelPop into multiple fragments, AscI cut this chromosome into two fragments, and FseI cut only once (Fig. 4a). The size of the genome of wMelPop was calculated by digestion with multiple enzymes to be 1.36 Mb (Table 2). Digestions of the closely related wMel and wMelCS strains with AscI indicated genome sizes equal to that for wMelPop (Fig. 4b; Table 2).

FIG. 4.

CHEF gels of digested genomes of arthropod Wolbachia strains wMelPop (a), wMelCS and wMel (b), and wRi (c and d). (a) Lane 1, Saccharomyces cerevisiae chromosomal size marker; lane 2, lambda ladder; lanes 3 to 6, digested wMelPop genome. (b) Lane 1, S. cerevisiae chromosomal size marker; lane 2, lambda ladder; lane 3, digested wMelCS genome; lane 4, digested wMel genome. (c) Lane 1, lambda ladder; lane 2, S. cerevisiae chromosomal size marker; lanes 3 to 6, digested wRi genome. (d) Lane 1, S. cerevisiae chromosomal size marker; lane 2, lambda ladder; lanes 3 and 4, digested wRi. Each lane is labeled with the enzyme(s) used.

TABLE 2.

Sizes of DNA fragments produced by digestions of Wolbachia genomes with selected restriction enzymes

Strain	Sizes (kb) of DNA fragmentsa produced by digestions with restriction enzyme(s)
	FseI	AscI	ApaI	ApaI + AscI	SmaI	SmaI + AscI
wMelPop	1,380	986, 376 (1,362)	343, 230,b 154, 146, 47, 45, 42, 36, 33, 33, 18 (1,357)		389, 245, 190, 173, 129, 125, 48, 29, 18, 6, 6 (1,358)
wMelCS			986, 376 (1,362)
wMel			986, 376 (1,362)
wRi			363, 200, 191,b 155, 128, 119, 100, 58, 47, 36, 33, 20, 18 (1,659)	363, 200, 191, 155, 128, 119,b 100, 72, 58, 47, 36, 33, 20, 18 (1,659)	381, 287, 243, 226, 130, 121, 79, 57, 45, 33, 28, 20 (1,653)	381, 287, 243, 206, 130, 121, 79, 57, 45, 33, 28, 23, 20 (1,650)
wDim			420, 290, 225, 60 (995)
wBma			380, 272, 250, 215 (1,117)

^aTotal numbers of kilobases are in parentheses. 

^bTwo fragments comigrated. 

The size of the genome of wRi was obtained via restriction enzyme digestion with ApaI, SmaI, and AscI (Fig. 4c and d). The size estimated from ApaI digestion and ApaI and AscI double digestion was 1.66 Mb (Table 2), while that from SmaI digestion and SmaI and AscI double digestion was 1.65 Mb (Table 2). Thus the genome size of wRi was designated 1.66 Mb, the mean of 1.65 and 1.66 Mb.

A panel of restriction enzymes having GC-rich recognition sequences were also tested for their ability to cut the nematode Wolbachia DNA into only one or a few linear fragments. Several enzymes were found not to cut the genomes of wBma or wDim, while others cut the genome into multiple fragments. The optimal enzyme for wBma was found to be ApaI, whose cut produced only four large fragments (Fig. 3), the sum of which was approximately 1,100 kb (Table 2), confirming the genome size determined from analysis of uncut DNA. Similarly, the Wolbachia genome from D. immitis extracts was cut into four bands totaling 950 kb by enzyme ApaI, again confirming the genome size by an independent method (Fig. 3; Table 2). For the Wolbachia genomes enriched from nematode tissues it was necessary to hybridize Southern blots of digested DNA with the probe cocktail in order to identify the individual fragments.

Conformation of the Wolbachia chromosome.

To determine if the Wolbachia chromosome was linear or circular, plugs containing intact mosquito cells which harbored Wolbachia wAlbB were prepared (27). The plugs of intact Aa23 (wAlbB) were digested with AscI and FseI, respectively. Then, PFGE was performed for both digested and undigested plugs. After PFGE, the gel was Southern blotted. Hybridization with the wRi wsp gene fragment clearly showed that uncut and FseI-digested Wolbachia DNAs were retained in the loading well, while AscI-digested Wolbachia DNA migrated into the gel (Fig. 5). These results suggest that the Wolbachia chromosome is circular.

FIG. 5.

Autoradiographs of Southern blot of wAlbB probed with a wRi wsp gene fragment. Lanes 1 to 3, plugs of Aa23 cells (wAlbB) digested with restriction enzymes AscI (lane 1), FseI (lane 2), and AscI and FseI (lane 3); lane 4, uncut DNA. Arrow, Wolbachia genome fragment that migrated into the gel after AscI digestion.

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 Wolbachia infecting 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 Wolbachia strains 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).