Genotypes and phenotypes in a Wolbachia-ant symbiosis
- Published
- Accepted
- Received
- Academic Editor
- Joseph Gillespie
- Subject Areas
- Entomology, Evolutionary Studies
- Keywords
- Resistance, Development, Virulence, Caste determination, Social insect, Endosymbiosis
- Copyright
- © 2024 Frost et al.
- Licence
- This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. For attribution, the original author(s), title, publication source (PeerJ) and either DOI or URL of the article must be cited.
- Cite this article
- 2024. Genotypes and phenotypes in a Wolbachia-ant symbiosis. PeerJ 12:e17781 https://doi.org/10.7717/peerj.17781
Abstract
The fitness effects of overt parasites, and host resistance to them, are well documented. Most symbionts, however, are more covert and their interactions with their hosts are less well understood. Wolbachia, an intracellular symbiont of insects, is particularly interesting because it is thought to be unaffected by the host immune response and to have fitness effects mostly focussed on sex ratio manipulation. Here, we use quantitative PCR to investigate whether host genotype affects Wolbachia infection density in the leaf-cutting ant Acromyrmex echinatior, and whether Wolbachia infection density may affect host morphology or caste determination. We found significant differences between host colonies in the density of Wolbachia infections, and also smaller intracolonial differences in infection density between host patrilines. However, the density of Wolbachia infections did not appear to affect the morphology of adult queens or likelihood of ants developing as queens. The results suggest that both host genotype and environment influence the host-Wolbachia relationship, but that Wolbachia infections carry little or no physiological effect on the development of larvae in this system.
Introduction
The intracellular bacterium Wolbachia is thought to be one of the most abundant endosymbionts on the planet and has been estimated to infect 40–66% of all arthropod species (Hilgenboecker et al., 2008; Zug & Hammerstein, 2012). Wolbachia transmits vertically in the maternal line only, resulting in it evolving mechanisms of distorting host sex ratio to increase the proportion or relative fitness of infected female hosts and thereby its own transmission (Werren, 1997). However, Wolbachia can also have more classical fitness effects, such as the decrease in longevity seen with the Wolbachia strain wMelPop in Drosophila melanogaster and the enhanced survival seen in the same host when infected with Wolbachia strain wDm (Carrington et al., 2009; Fry, Palmer & Rand, 2004). Switching between these phenotypes can occur rapidly, with Wolbachia in newly infected populations of Drosophila simulans having been shown to evolve from parasitic to mutualistic phenotypes in only 20 years (Weeks et al., 2007). This suggests that the relationship between Wolbachia and the biology, and therefore genetic make-up, of its host is important in the effects that Wolbachia induces.
The phenotypic effects induced by Wolbachia are likely to have a significant impact on the way in which the host immune system interacts with it. While, it was thought that Wolbachia can evade detection by the host immune system (Siozios et al., 2008), genetic backcrossing experiments, microinjection experiments, and cycling within natural populations suggest that host physiology affects the success and nature of Wolbachia infections (Bordenstein, Uy & Werren, 2003; Evans et al., 2009; Frost et al., 2010; McGraw et al., 2002; Van Borm et al., 2001; Wenseleers, Sundstrom & Billen, 2002). Despite this, research into host genetic effects on Wolbachia infection is still limited (Bordenstein, Uy & Werren, 2003; Mouton et al., 2007; Reynolds, Thomson & Hoffman, 2003).
For host resistance traits to spread within a population, resistant individuals must have a selective advantage over non-resistant hosts. One of the most obvious morphological indicators of fitness is size, with larger individuals generally having higher survival and fecundity (Kingsolver & Huey, 2008). Morphological effects of symbionts are well known (Miura et al., 2006), and may also be important in Wolbachia infections. If infections cause energetic stress to the host then this may result in larvae developing into smaller adults or adults storing less energy reserves in their fatbody (Arrese & Soulages, 2010; Emlen & Allen, 2003). In support of this, larger mosquitoes have been found to be less likely to pass Wolbachia to their offspring, possibly due to these larger females having lower density infections (Kittayapong et al., 2002).
Social insect colonies are made up of reproductive queens and sterile workers, with workers, like males, being an evolutionary dead end for the vertical transmission of Wolbachia, potentially selecting for phenotypic effects which increase the representation of Wolbachia in queens (termed gynes sensu stricto prior to mating). Wolbachia infections are common in the ants in particular (Ramalho & Moreau, 2020; Russell et al., 2012; Wenseleers et al., 1998), and in at least one ant species appear to cause sex ratio distortion (Pontieri et al., 2017), but what other fitness effects Wolbachia has is unknown. It has been shown in several social insects that patrilines (groups of sisters within a colony that share the same father) may differ in their propensity to develop into different castes, but the reasons for this are largely unknown (Chéron et al., 2011; Holman et al., 2011; Hughes & Boomsma, 2007; Hughes & Boomsma, 2008; Jaffé et al., 2007; Rheindt, Strehl & Gadau, 2005). Symbionts can affect the energy available to hosts for growth and development, so could affect the availability of resources needed to develop as a queen. It is therefore possible that resistance of specific patrilines to Wolbachia infection could be the reason for this queen skew, i.e. more resistant individuals would have a lower burden of Wolbachia infection, leading to less energetic stress and in turn more nutritional resources to develop as queens. If Wolbachia exert detrimental effects on their host, resistant patrilines would be expected to have lower density infections and a higher propensity to develop into queens, as these would be overall fitter individuals. The opposite would be the case if Wolbachia has beneficial or no effects, as well as if the host has evolved to tolerate Wolbachia infections.
Acromyrmex leaf-cutting ants have high prevalence Wolbachia infections, with infections that appear to cycle within a population over time, suggesting that Wolbachia in this group may be subject to significant host regulation (Andersen et al., 2012; Frost et al., 2010; Van Borm et al., 2001). Individual Acromyrmex ants may be infected by a single Wolbachia strain or multiple coinfecting strains, and Wolbachia cells are commonly present in both reproductive and nonreproductive tissues (Andersen et al., 2012; Frost et al., 2014; Sapountzis et al., 2015; Zhukova et al., 2017). Queens of these ants mate with multiple males (Sumner et al., 2004), which facilitates genotypic comparisons because the offspring within a colony share the same environmental conditions, maternal cues and maternal genotype (on average), differing only in their paternal genotype (patriline). This makes Acromyrmex ideal candidates for investigation of host effects on Wolbachia infection. Here we investigate whether hosts vary genotypically in resistance to Wolbachia infection by comparing the Wolbachia infection density in queens between patrilines within colonies of A. echinatior. We also determine whether any variation in Wolbachia density correlates with morphological measures of host fitness or with the propensity of individuals in different patrilines to develop into reproductive queens rather than sterile workers.
Methods
Sample collection
We used five monogynous colonies of Acromyrmex echinatior (Ae48, Ae357, Ae153, Ae088 and Ae07P4) that were collected in Gamboa, Panama, between 1996 and 2008 (Ae48 in 1996, Ae153 in 2001, Ae07P4 in 2007, and Ae357 and Ae088 in 2008) under permit from the Autoridad Nacional del Ambiente (ANAM), and maintained in the laboratory at 27 ± 2 °C and 80 ± 10% RH on a diet of privet leaves and rice. A total of 96 queens (gynes sensu stricto), 96 small workers (<1.2 mm head width) and 96 large workers (1.8–2.4 mm head width) were collected over a week from each colony to allow the queen-worker propensity of each patriline to be determined. All individuals were sampled from the surface of the fungus garden and were of similar age based upon their cuticular colouration (Armitage & Boomsma, 2010). This avoided the propensity estimates being confounded by any temporal changes in sperm use, something which in any case appears to be random in leaf-cutting ants (Holman et al., 2011; Stürup et al., 2014).
Patriline determination
A single leg per individual was incubated at 56 °C overnight in 100 µl of 5% Chelex 100 (BioRad, Hercules, CA, USA) suspended in 10 µM Tris buffer with 4 µl of Proteinase K (5 µg/ml), and then boiled for 15 min. After spinning down, the DNA extract (supernatant) was removed and used for subsequent PCRs using five microsatellite loci: Ech1390, Ech3385, Ech4126, Ech4225 and Atco15 (see Table 1 for conditions; Helmkampf, Gadau & Feldhaar, 2008; Ortius-Lechner, Gertsch & Boomsma, 2000). An ABI 3130xl genetic analyser with Genemapper® v3.7 software was used to score microsatellite sizes in relation to GeneScanTM LIZ-500 size standard (ABI). Individuals were then manually assigned to patrilines based upon their multilocus genotypes, so that queen-worker skew could be measured for each individual patriline. Queen-worker skew was defined as the corrected proportion of females within that patriline that were queens as opposed to workers (the number of queens was corrected by multiplying by two to account for twice the number of workers being collected compared to queens), with <0.5 being worker skewed and >0.5 being queen skewed. Patrilines for which less than seven queens had been sampled were excluded from further analysis.
| Primer | Ech1390 | Ech3385 | Ech4126 | Ech4225 | Atco15 |
|---|---|---|---|---|---|
| Taq polymerase (Promega) (units/µl) | 0.04 | 0.04 | 0.04 | 0.04 | 0.03 |
| Taq buffer | 1x | 1x | 1x | 1x | 1x |
| MgCl2 (mM) | 1.62 | 0.67 | 0.67 | 3.38 | 2.00 |
| Total dNTPS (µM) | 270 | 100 | 100 | 340 | 200 |
| Primer F (µM) | 0.4 | 0.5 | 0.5 | 0.5 | 0.2 |
| Primer R (µM) | 0.4 | 0.5 | 0.5 | 0.5 | 0.2 |
| Template DNA (µl) | 1.1 | 1.1 | 1.1 | 1.1 | 2 |
| Annealing temperature (°C) | 48 | 50 | 57 | 52 | 55 |
Wolbachia infection densities
For each queen (only), the density of Wolbachia infection was determined using a sample of the ovaries and midgut combined. The tissue samples therefore included both somatic and sexual tissues, as it has been shown that both can harbour substantial densities of Wolbachia in these ants (Andersen et al., 2012). Tissues were extracted in the same manner as for genotyping, with the addition of the DNA extract (supernatant) being run through a Onestep-96 PCR inhibitor removal kit (Zymo Research, Tustin, CA, USA). Wolbachia was quantified using the comparative Ct method, which corrects for differences in the quantity of host tissue in a sample by combining Wolbachia and host DNA quantities into a single value (Schmittgen & Livak, 2008). Host and Wolbachia primers and probes were designed using the ABI custom design service based on widely aligned sequences of the relevant Wolbachia CoxA gene, designed to detect the multiple Wolbachia strains that may infect Acromrymex (Andersen et al., 2012; Van Borm et al., 2001), and host 18S rRNA gene (Genbank accession numbers HM211030 and EF012826, respectively). The Wolbachia assay included Walter forward primer TTGTATCAACTTTTTCTCACAGACCTGTA, Walter reverse primer AAAAAATGCAGCAGCGT and Walter probe (FAM-NFQ) ACCATAAAGCCAAATACTG. The host assay included Brenda forward primer GCATTCGTATTGCGACGTTAGAG, Brenda reverse primer CATTTTTGGCAAATGCTTTCGCTTC and Brenda probe (FAM-NFQ) ACGATCCAAGAATTTC. 10 µl qPCRs using the Taqman® Universal Master Mix II system were used to determine infection status. Each qPCR contained 1 µl of DNA template, 1× Master mix II, 2.25 mM of each Primer and 0.625 mM of Taqman® probe. The qPCR cycle used was 50 °C for 2 min, 95 °C for 10 min, 40 cycles of 95 °C for 15 s and 60 °C for 1 min. All qPCRs were run in triplicate (average Ct used for analysis), with negative controls and positive reference samples included in each run. Multiple calibration curves were run for both assays. Calibration curves from samples containing ovary and gut tissue had efficiencies of 91–99% for the Wolbachia assay and 95–99% for the host assay over 10,000 fold changes for both assays. Additional calibrations were run using whole ants (which had been previously bead-beaten) to extend the calibration curves and these gave similar efficiencies over a greater range (Ct values between 16-37 for the host and 23-36.5 for the Wolbachia assay). Any samples which gave Ct values outside of this range were removed from analysis. Wolbachia infection density was calculated as the relative quantity (RQ) of Wolbachia normalized against the 18S rRNA control gene.
Morphometric data collection
Morphometric data for queens from four of the five colonies (Ae48, Ae07P4, Ae088 and Ae357) were collected as part of a previous study (Mitchell, Frost & Hughes, 2012). Briefly, six measurements of body size were made for each queen: (1) head width (the maximum head width across the eyes); (2) forewing length (from the first vein intersection to the base of the wing); (3) thorax length; (4) weight (after drying at 70 °C for 5 days). Body parts were scanned (Epson Scan V300 Photo; Epson UK Ltd, Hemel Hempstead, UK) with a resolution of 9600 pixels, and measured using IMAGEJ 1.42q (Rasband 1997–2011) calibrated with a scanned 0.1-mm graticule. Measurement error was estimated as the average coefficient of variation (CV) for each character based on a random subset of 10 individuals measured three times, using Haldane’s correction for small sample size (Haldane, 1955), and was on average <1%.
Statistical analysis
All analyses were carried out in R version 2.8.1 (R Development Core Team, 2005). The effect of colony on the Wolbachia infection density of queens was investigated using a generalised linear model (GLM) with quassi-Poisson error structure and the anova function. The effect of patriline on Wolbachia infection density was tested with a linear mixed effects model with patriline nested within colony and both specified as random factors, using the lmer function and lme4 package, with a quasi-Poisson error structure and p values computed using the likelihood ratio test method (Bates et al., 2015). The relationships between Wolbachia infection density and the various morphological variables on Wolbachia infection density were also explored with linear mixed effects models using lmer, while the relationship between queen-skew and Wolbachia infection density was tested using the lme and ANOVA functions (Pinheiro et al., 2019). Colony and the interaction between colony and the morphological variable were also tested for in these models. Model choice was determined by examination of residuals.
Results
The density of Wolbachia infections in queens (which may be infections by single or multiple strains) differed significantly both between the ant colonies and between patrilines within colonies (respectively: ANOVA (GLM), f360, 4 = 21.2, p < 0.001; LRT (LMER), f24, 20 = 45.3, p = 0.001). The between colony differences were larger than those seen between patrilines within colonies, with colonies Ae48 and Ae153 having infection densities that were 2–4 times as high as colonies Ae357, Ae07P4 and Ae088 (Fig. 1). Colonies with higher density Wolbachia infections tended to have greater variation between patrilines, with the largest difference between patrilines being over two-fold (between Patrilines 1 and 5 in colony Ae088; Fig. 1). The more highly infected colonies also appeared to have greater variation within patrilines than the less infected colonies (Fig. 1).
Figure 1: Wolbachia infection densities across patrilines.
Mean ± SE Wolbachia infection density (RQ) for patrilines within five colonies of Acromyrmex echinatior. The number of individuals examined for each patriline is shown above the columns. Wolbachia infection density refers to the density of Wolbachia normalized against the 18S rRNA host control gene.
There was no significant relationship between Wolbachia infection density of queens and any of the morphological variables (LRT (LMER), head width, f7, 1 = 0.036, p = 0.849; wing length, f7, 1 = 5.58, p = 0.233; weight, f10, 4 = 5.796, p = 0.215), except for thorax length for which there was a significant interaction between the effects of colony and the density of Wolbachia infections on the length of the thorax (LRT (LMER), f10, 4 = 18.7, p < 0.001). Colony Ae357 had a negative relationship between Wolbachia infection density and thorax length, while the remaining three colonies had relatively flat relationships (Fig. 2). There was no significant relationship between the infection density and queen-worker skew of patrilines (LME, f15, 1 = 0.0002, p = 0.989, Fig. 3).
Figure 2: The relationships between Wolbachia infection density (RQ) and body size of Acromyrmex echinatior queens.
Data are for (A) head width, (B) thorax length, (C) wing length and (D) weight for queens from four colonies (colony denoted by different markers: Ae48, red circle and trendline; Ae07P4, blue circle and trendline; Ae088, orange circle and trendline; Ae357, white square and black trendline). Wolbachia infection density refers to the density of Wolbachia normalized against the 18S rRNA host control gene.
Figure 3: The relationship across leaf-cutting ant patrilines between the density of Wolbachia infections (RQ) and the propensity to develop into queens or workers.
Caste propensity is queen-worker skew, from 0 being entirely worker-biased to 1 being entirely queen-biased, and 0.5 showing no skew towards either caste. Data are presented for five Acromyrmex echinatior colonies indicated with different markers (Ae48, red circle; Ae153, blue triangle; Ae07P4, blue circle; Ae088, orange circle; Ae357, white square). Wolbachia infection density refers to the density of Wolbachia normalized against the 18S rRNA host control gene.
Discussion
There was significant variation among the queens analysed in the density of their Wolbachia infections. The greatest differences were between colonies, but there were also statistically significant differences in the infection density of queens between patrilines within colonies. There was a negative relationship between Wolbachia infection density and thorax length in one colony, but Wolbachia density did not show any significant relationship with morphology in the other colonies, nor the propensity of individuals in a patriline to develop into queens rather than workers. Infections may have involved single or multiple strains because our protocol did not aim to distinguish these.
The largest variation in the Wolbachia density of queens was found between colonies, with two colonies (Ae48 and Ae153) having higher density infections than the other three colonies. There is greater genetic variation at this level than within colonies, with individuals across colonies having different maternal, as well as paternal, genes. In addition, environmental effects, epigenetics, the overall health of the colony and differences in the strain composition of Wolbachia may also have produced colony-level differences in Wolbachia infections. These colony-level differences may therefore be similar to the way populations in other hosts vary in their Wolbachia infections due to differences in environmental conditions, genetic composition and symbiont strains (Bordenstein, Uy & Werren, 2003; Fry, Palmer & Rand, 2004; Reynolds, Thomson & Hoffman, 2003). It is interesting to note in this regard that the two colonies with the highest infection densities were the oldest and had been maintained in the laboratory the longest. Laboratory colonies of Acromyrmex have previously been found to have higher titres of Wolbachia than field colonies (Andersen et al., 2012), and Wolbachia can show rapid evolution in response to environmental condition (Weeks et al., 2007). Our results raise the intriguing possibility that this may be cumulative, with the longer a colony is kept in the laboratory under relatively favourable conditions, the greater the infection density. Temperature can have significant effects on Wolbachia infection dynamics (Mouton et al., 2007; Pintureau & Bolland, 2001; Reynolds, Thomson & Hoffman, 2003), and it may be that other environmental factors have similar effects. Wolbachia in Acromyrmex ants is present in the gut and other non-reproductive, as well as reproductive, tissues (Andersen et al., 2012; Frost et al., 2014; Sapountzis et al., 2015; Zhukova et al., 2017), so may show more dynamic infections than endosymbionts that are restricted solely to the ovaries of hosts.
We also found that some patrilines within colonies differed in the density of Wolbachia infection of queens. As patrilines within colonies differ only in their paternal genotypes and sperm use in Acromyrmex is random (Stürup et al., 2014), this suggests that host genotype can have a significant effect on this endosymbiont. There are three potential explanations for the genotypic variation in infection density. First, that the fitness effects of Wolbachia have caused hosts to evolve resistance mechanisms that directly target Wolbachia either in general or in a strain-specific way. Second, that Wolbachia infection is affected indirectly by immune responses targeted at other threats (Siozios et al., 2008). Third that genotypic differences in some other trait affect Wolbachia infection density as an epiphenomenon. The second explanation seems most likely because silkworm cells in vitro do not alter gene expression in direct response to Wolbachia infection (Nakamura et al., 2011). In addition, the genotypic variation in Wolbachia infection seen here is much lower than the genotypic differences found in resistance to more virulent parasites (Baer & Schmid-Hempel, 2003; Hughes & Boomsma, 2004; Invernizzi, Peñagaricano & Tomasco, 2009). It is important to note that within-colony variation may result from differences in distribution of maternal factors as well as differences between patrilines, but we did not investigate this here.
There were no relationships between Wolbachia infection density of queens and any of the morphological variables, with the sole exception of a negative correlation with thorax length in Colony Ae357. The relative lack of Wolbachia effects on morphology implies that Wolbachia infection density may not be associated with a significant energetic burden or benefit in these hosts, although further experiments would be needed to confirm this by investigating other measures of energetic impact on the host such as metabolic rate (Evans et al., 2009). Comparatively benign infections of many parasites can become more virulent in times of stress (Brown, Schmid-Hempel & Schmid-Hempel, 2003; Comiskey, Lowrie & Wesson, 1999), and it may be that studies looking at other parts of the colony cycle may show a different host-symbiont relationship.
Caste determination in leaf-cutting ants and other social insects is thought to occur through a threshold mechanism whereby environmental cues, such as nutrition, determine the morphological caste into which an individual will develop (Wheeler, 1991). It has been shown in several social insects, including A. echinatior, that genotype may affect the propensity of individuals to develop into different castes (Chéron et al., 2011; Holman et al., 2011; Hughes & Boomsma, 2007; Hughes & Boomsma, 2008; Jaffé et al., 2007; Rheindt, Strehl & Gadau, 2005). If Wolbachia had an overall positive or negative burden on these insects it may have been expected that genotypic differences in the density of Wolbachia infections may correlate with fitness parameters of the host, but we found no evidence to suggest that this is the case. However, it must be noted that colonies varied greatly in their apparent relationships across patrilines between queen-worker skew and the density of Wolbachia infection, and the relatively few patrilines studied meant that the analysis had limited power. An analysis which included more colonies and patrilines, with sufficient sample size to provide high statistical power, would be needed to give a more conclusive result.
Conclusions
Although Wolbachia is often considered simply as a maternally transmitted, sex ratio-distorting endosymbiont, there is now abundant evidence that its effects and infection dynamics can be more complex than this. The large differences in infection density between colonies, and smaller differences between patrilines, that we observed here in leaf-cutting ants support this. Endosymbiont infections are unlikely to ever be completely cost-free, and although we found little evidence of any physiological costs or benefits here, further investigation of the physiological impacts of infection would be warranted.