Solenopsis fire ants are 5-10 times more abundant where they occur in the United States than they are in their South American homelands (Porter et al. 1992; Porter et al. 1997). Escape from numerous natural enemies left behind in South America (Williams et al. 2003) is a likely explanation for this difference. Natural enemies include almost 2 dozen species of phorid flies (Porter & Pesquero 2001; Brown et al. 2003), a parasitic ant (Calcaterra et al. 1999), several nematodes (Williams et al. 2003), a virus (Valles et al. 2004b; Valles et al. 2007), and two species of microsporidian pathogens: Thelohania solenopsae Knell, Allen & Hazard and Vairimorpha invictae Jouvenaz & Ellis.

The first microsporidian, T. solenopsae, was described by Knell et al. (1977) from Brazil. Briano et al. (1995a, 1995b, 1995c, 1996) published extensively on its distribution, abundance, and impacts at sites in Argentina. This pathogen was first detected in the United States by Williams et al. (1998) who found it widespread and apparently specific to imported fire ants.

The second microsporidian, V. invictae, was described by Jouvenaz & Ellis (1986). This pathogen occurs in both monogyne and polygyne fire ant colonies (Valles & Briano 2004) and appears to be more common in the red fire ant S. invicta than in the black fire ant S. richteri (Briano & Williams 2002). Its occurrence in the field is associated with smaller colonies and declining fire ant populations (Briano 2005). Recent laboratory studies have shown that this pathogen can be transmitted among fire ant colonies by the transfer of larvae, pupae, or even dead workers and that infected colonies grew 85% slower than uninfected control colonies (Oi et al. 2005).

Terrestrial microsporidian pathogens, as a group, are normally limited to one host or a group of related hosts (Solter & Maddox 1998; Briano et al. 2002; Keeling & Slamovits 2004). Narrow host ranges also apparently apply to Vairimorpha (Solter et al. 2000); however, some species like Vairimorpha necatrix appear to have relatively broad physiological host ranges in the laboratory when lepidopteran larvae challenged with high spore doses. Nevertheless, actual ecological host ranges in the field are likely less broad (Maddox et al. 1981).

Vairimorpha invictae is a promising self-sustaining or classical biocontrol agent for fire ants because of (1) documented impacts on fire ant colonies, (2) its prevalence on red fire ants, and (3) its absence from the United States. Furthermore, V. invictae appears to be specific to fire ants. Briano et al. (2002) reported that none of the 9 genera of non-Solenopsis fire ants collected at 167 baits from >25 sites or from 50 colonies at >20 sites were infected with V. invictae. This absence of observed infections occurred despite the fact that V. invictae was found in 61 of 535 fire ant colonies collected at 18 of 90 sites.

The objective of the current study was to further examine the host range of V. invictae by investigating its presence in arthropods and other species of ants that co-occur in close proximity with V. invictae infected fire ant colonies in South America.

Materials and Methods

We found 3 populations of mostly polygyne S. invicta infected with V. invictae in Apr 2004, 5–20 km east of Corrientes city, Argentina. The first site was a campground and boat launch area along the Paraná River (S 27°22.601', W 58°40.891'). The second site was by a bridge over a small stream near a family farm (S 27°24.283', W 58°41.127'). The third site was along Route 12 at km marker 1034 just outside the Corrientes Airport (S 27°27.414', W 58°45.743'). Two additional sites with V. invictae infected colonies were found near San Javier, Santa Fe Province about 350 km south of Corrientes. The first site was near Estancia Liriolay (S 30°30.951W 60°01.052). The second site was near an intersection with a clay road marked by white tires on the roadside (S 30°31.615, W 60°00.730). The proportion of polygyne colonies at these 2 sites was 33%, based on PCR tests done for a recent study (Briano et al. 2006).

In order to examine the host range of V. invictae at the five study sites, we used small vials (12 by 85 mm) baited with pieces of hotdog to collect other species of co-occurring ants. Sugar water baits were also used at the San Javier sites. At each site, 20-50 baits were set out at 5-m intervals in several transects. Some hand collecting was also used for ants not attracted to baits. At the Corrientes sites, we also used sweep nets to collect co-occurring arthropods. A supplemental sample of 79 non-ant Hymenoptera collected in Jan 2005. Ants and other arthropods were stored submerged in 95% ethanol to preserve the DNA (King & Porter 2004).

Infected fire ant colonies were detected by macerating about 30 workers with an electric tissue grinder. A drop of aqueous extract was then examined under a phase-contrast microscope for the presence of spores (Briano 2005). At baits, we macerated 15-20 fire ant workers for microscopic examination if that many were available. These sampling rates should have been sufficient on average to detect V. invictae in samples with worker infection rates as low as 3-7%.

PCR was used to detect V. invictae in non-Solenopsis ants and other kinds of arthropods (Valles et al. 2004a). Briefly, PCR was carried out with primer pairs specific for the 16S rRNA gene of V. invictae (p90, 5'CACGAAGGAGGATAACCACGGT and p93, CGCAATCAGTCTGTGAATCTCTTCA) by using the hot start method in a PTC 100 thermal cycler (MJ Research, Waltham, MA) under the following optimized temperature regime: 1 cycle at 94°C for 2 min, then 35 cycles at 94°C for 15 s, 55°C for 15 s, and 68°C for 45 s, followed by a final elongation step of 5 min at 68°C. The reaction was conducted in a 25 μL volume containing 2 mM MgCl₂, 200 μM dNTP mix, 1 unit of Platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA), 0.4 μM of each primer, and 0.5 μL of the genomic DNA preparation (10 to 50 ng). PCR products were separated on 1.2% agarose gel and visualized by ethidium bromide staining. For all experiments, positive and negative controls were run alongside treatments. In situations in which verification of V. invictae positive non-Solenopsis insects was required, the amplicon was cloned and sequenced. Amplicons were gel-purified, ligated into the pCR-4 vector and transformed into TOP10 competent cells (Invitrogen, Carlesbad, CA). Insert-positive clones were identified and DNA sequences of inserts were elucidated by the University of Florida, Interdisciplinary Center for Biotechnology Research (Gainesville, FL).

At the Corrientes sites, samples of ants from several baits were pooled, generally by genus and site, into 18 samples. Most pooled samples consisted of ants from 2-5 baits and contained 20-50 individuals depending on their size and availability. The arthropod samples were similarly pooled, mostly by order or family, into 33 samples. At the San Javier sites, ant samples from non-fire ant genera were similarly pooled into 16 samples with each sample containing 20-40 workers from 2-3 baits.

PCR was also conducted to verify that the test could detect the V. invictae target gene in other ant genera (Linepithema, Dorymyrmex, Crematogaster, Paratrechina). This experiment was done by spiking 30-70 non-Solenopsis workers with 5 Solenopsis workers from a V. invictae-positive colony prior to maceration for DNA extraction. Similar tests were not done with non-ant arthropods, because DNA extraction techniques are quite reliable, sample sizes were limited, and it would have considerably increased the number of samples needing testing. We also used PCR to test 3 samples of infected fire ant workers from San Javier that were initially detected by microscopic inspection. Based on the experience from a previous study conducted in Argentina (Valles & Briano 2004), V. invictae is always detected with PCR tests if it was previously detected by microscopic examination.

Results

As expected (Valles & Briano 2004), the PCR technique identified V. invictae in fire ant workers in which spores had been detected by light microscopy (n = 3). Similarly the PCR technique detected V. invictae in samples of Paratrechina, Crematogaster, and Linepithema ants spiked with 5 infected fire ants, thereby showing that ants in these genera do not appear to have chemistry that would mask detection of V. invictae DNA. Vairimorpha invictae was not detected in the first sample with Dorymyrmex ants, perhaps because none of the 5 fire ant workers used to spike the sample were infected (Briano & Williams 2002) or perhaps because of a faulty extraction. A retest with these ants, however, produced a positive result.

Infected fire ants were detected in both fire ant mounds and among foraging workers collected at baits (Table 1). At the Corrientes sites, the percentage of positive baits was lower than the percentage of positive mounds when we included baits with ≤4 workers. However, when we only compared baits with ≥5 workers, the percentages were similar. This comparison is more appropriate because only 10-60% of workers in a mound are infected (Briano & Williams 2002) and mound samples always contained many workers (15-30) while bait samples, with only a few workers, may have produced false negatives relative to their colony of origin. However, at the clay road site near San Javier (Table 1), the percent of infected baits was still about half that of the infected mounds.

All of the non-ant arthropod samples from the Corrientes area (Table 2) were negative for V. invictae (Table 1). One sample from the campground area containing 2 solitary bees and 2 sphecid wasps produced a faint positive band, but sequencing confirmed that the amplicon was not from the V. invictae 16S rRNA gene.

All of at least 19 species of ants from 12 non-Solenopsis genera including 947 individuals from 107 samples collected near Corrientes and San Javier (Table 3) were also negative for V. invictae (Table 1).

When we resampled ant colonies at the 3 Corrientes sites 20 months after the initial sampling (Dec 2006) and tested them for V. invictae infections, we found that V. invictae had disappeared from the bridge site (0/8) and only one weakly positive sample was found each at the campground and airport sites (1/12; 1/19), a considerable decline from what we found in 2004 (Table 1). This drop in frequency of V. invictae infections was not unexpected because similar drops in infection rates have been reported previously, often in conjunction with lower nest densities (Briano 2005; Briano et al. 2006).

Discussion

Overall, these data indicate that V. invictae is specific to Solenopsis fire ants because this pathogen was not found in any of the sympatric ants or co-occurring arthropods. Briano et al. (2002) failed to find V. invictae in 217 collections of non-Solenopsis ants from 9 genera in Argentina; although, it appears that only 20% of the sites may have had V. invictae present in co-occurring fire ants. While V. invictae has not been found in other ant genera, it has been reported from 3 species of Solenopsis fire ants (S. invicta, S. richteri, S. macdonaghi; Briano et al. 2002), all in the saevissima complex of South American fire ants (Trager 1991). It likely occurs in other South American Solenopsis species as well. The known geographic range of this pathogen extends from Mato Grosso, Brazil (Jouvenaz & Ellis 1986) south through Paraguay and part of Bolivia (Briano et al. 2006) to Buenos Aires Province, Argentina (Briano et al. 2002). Extensive searches in the United States for T. solenopsae and other pathogens (Jouvenaz et al. 1977; Williams et al. 1998) have failed to detect V. invictae infections.

We currently do not know whether V. invictae has the ability to infect native Solenopsis fire ants in the United States (e.g.; S. geminata, S. xyloni) which belong to a different complex of fire ants (geminata) or the more distantly related group of very small Solenopsis thief ants. Laboratory tests in quarantine are currently being conducted by D. H. Oi in an effort to answer the question of whether our native North American fire ants and other species of ants are susceptible to V. invictae infections.

Low levels of infection in native fire ants would probably be tolerable (Porter 2000) because the native fire ants can be pests in their own right, especially the tropical fire ant S. geminata which has become a world-wide exotic pest. However, infections in other ant genera or in other families of insects would be a complicating factor which would likely require further tests and impact evaluations across a range of ant genera or arthropod families.