SEARCH

SEARCH BY CITATION

Keywords:

  • condition dependence;
  • genetic diversity;
  • parasite load;
  • parent–offspring heterozygosity

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Positive correlations between heterozygosity and fitness traits are frequently observed, and it has been hypothesized, but rarely tested experimentally, that parasites play a key role in mediating the heterozygosity–fitness association. We evaluated this hypothesis in a wild great tit (Parus major) population by testing the prediction that the heterozygosity–fitness association would appear in broods experimentally infested with a common ectoparasite, but not in parasite-free broods. We simultaneously assessed the effects of parental and offspring heterozygosity on nestling growth and found that body mass of nestlings close to independence, which is a strong predictor of post-fledging survival, increased significantly with nestling levels of heterozygosity in experimentally infested nests, but not in parasite-free nests. Heterozygosity level of the fathers also showed a significant positive correlation with offspring body mass under an experimental parasite load, whereas there was no correlation with the mothers’ level of heterozygosity. Thus, our results indicate a key role for parasites as mediators of the heterozygosity–fitness correlations.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Heterozygosity may be linked to Darwinian fitness either because increased allelic diversity lowers the risk of deleterious recessive alleles to occur in homozygous state or because heterozygotes benefit from overdominance (Keller & Waller, 2002). Individual heterozygosity is usually assessed using neutral genetic markers, such as microsatellites. Heterozygosity–fitness correlations (HFCs) require heterozygosity at neutral markers to reflect heterozygosity levels at loci under selection (Szulkin et al., 2010). Two hypotheses have been put forward for such an association: first, the general effect hypothesis holds that inbreeding effects cause associations between neutral markers and genome-wide heterozygosity resulting in HFCs (Hansson & Westerberg, 2002). Second, the local effect hypothesis suggests that linkage disequilibrium between neutral markers and selected genes leads to the observed correlation (David, 1998; Hansson & Westerberg, 2002). Many studies reported associations between heterozygosity and important life history traits such as number or quality of offspring (Foerster et al., 2003; Seddon et al., 2004; Slate et al., 2004; Bensch et al., 2006), survival (Coltman et al., 1998; Coulson et al., 1998; Hansson et al., 2001, 2004; Da Silva et al., 2009) and immunocompetence (Reid et al., 2007; Fossoy et al., 2009). However, such correlations are generally weak and high variation in correlational strength occurs among populations and years (Coltman & Slate, 2003), likely due to condition-dependent effects of heterozygosity on fitness (Balloux et al., 2004; Kempenaers, 2007; Chapman et al., 2009). Environmental factors such as thermal stress, limited food availability and harsh weather conditions have been found to strengthen the association between heterozygosity and fitness (Scott & Koehn, 1990; Lesbarreres et al., 2005; Da Silva et al., 2006; Marr et al., 2006). In wood frogs, for example, larval survival was associated with heterozygosity in the wild, but not in the laboratory, and the difference has been attributed to abiotic factors being harsher in the wild but also to increased predation risk, competition or disease prevalence (Halverson et al., 2006).

As parasites can impose strong selection and fitness costs by reducing growth, weight, breeding success and survival of hosts (Milinski & Bakker, 1990; Moller, 1990; Lehmann, 1993; Richner, 1998), they are likely mediators for the strength of the HFC. Facing parasite pressure, genome-wide effects of heterozygosity may not only provide general fitness benefits, for example by enhancing overall condition, but also increase variability at loci involved in pathogen recognition and defence (Balloux et al., 2004). Both of these mechanisms can cause the often observed correlation between parasite resistance and individual heterozygosity (e.g. Acevedo-Whitehouse et al., 2003, 2005; Hawley et al., 2005). Consequently, the observed variability of HFCs might be generated by variation of parasite prevalence and intensity, which would predict HFCs to be stronger in the presence of parasites but weak in their absence (Coltman et al., 1999). Experimental evidence from free-living populations, however, is scarce.

In this study, we investigated the effects of experimental parasite infestation on HFCs in a socially monogamous bird, the great tit (Parus major). Given that the effects of many pathogens are dosage-dependent (e.g. Poulin, 2011), we experimentally infested great tit nests with increasing numbers of hen fleas (Ceratophyllus gallinae), thus manipulating the infestation intensity faced by breeding parents and their nestlings. Hen fleas are commonly occurring nest-based ectoparasites (Harper et al., 1992; Heeb et al., 1996), known to reduce nestling growth and number (Richner et al., 1993). We assessed first the association between nestling heterozygosity and growth in the presence and absence of an ectoparasite burden and second the relationship between parental heterozygosity and nestling growth to evaluate whether heterozygous parents cope better with stressful breeding conditions and thus can buffer the negative impact of parasite infestation on offspring.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Experimental procedures

The experiment was carried out in spring 2009 in a great tit population breeding in the Spilwald (47°56′ N, 7°18′ E), a forest near Bern, Switzerland. For experimental parasite infestation, we used the hen flea (Ceratophyllus gallinae), a naturally occurring ectoparasite of many hole-nesting bird species with highest intensity and prevalence in great and blue tits (Tripet & Richner, 1997). Adult fleas live in the nesting material and suck blood from adult and nestling great tits. Flea larvae develop in the nesting material, where they feed on detritus and undigested blood. In great tit nests, two flea generations per breeding cycle can usually be observed (Harper et al., 1992). Nestling great tits suffer from reduced growth and survival when ectoparasite load is high in their nest (Richner et al., 1993). Before the start of the breeding season, we emptied and brushed all nest boxes thoroughly to remove the parasites from the previous breeding season. Nest boxes were regularly visited to determine clutch size and start of incubation. On the second day of incubation, we heat-treated the nesting material for 3 min using a microwave (Richner et al., 1993) to kill ectoparasites and brushed the nest boxes thoroughly to remove all remaining ectoparasites. Each nest box was randomly assigned to one of four infestation treatments with 100, 50 or 10 fleas added plus a control without added fleas. Fleas used for infestation were collected from nesting material of the previous breeding season, which had been stored in a climatic chamber at 5 °C, and added to the nest using a sex ratio of three females per two males.

The subsequent procedures were identical for all treatments: 10 days after infestation, nests were regularly checked for hatching (referred as day 1) of young. Newly hatched nestlings were weighed to the nearest 0.01 g using an electronic balance and individually marked by partially removing tuft feathers from their heads, shoulders and backs. Nestlings were ringed with standard aluminium rings 9 days post-hatch, and blood samples were collected. Blood samples were stored in ethanol 96% until further analyses. 16 days post-hatch, we measured nestling body mass (± 0.1 g), metatarsus (± 0.1 mm) and the length of the third primary feather (± 0.5 mm). Parents were captured on day 14 post-hatch with a spring trap inside the nest box. Body weight, tarsus length and the length of the third primary feather were measured, and the birds were sexed according to the presence or absence of a brood patch. Blood samples were collected and stored in ethanol 96%.

Genetic and paternity analyses

DNA was extracted from blood samples using magnetic beads. Forty-seven autosomal microsatellite markers were amplified by polymerase chain reaction using Qiagen Multiplex PCR kit (Qiagen AG, Hombrechtikon, Switzerland) as described in the study of Saladin & Richner (2012). We used a subset of 11 microsatellite loci (PmaC25, PmaCAn1, PmaD105, PmaD22, PmaGAn27, PmaGAn30, PmaTAGAn71, PmaTAGAn86, PmaTGAn33, PmaTGAn42 and PmaTGAn45) for paternity analyses (Saladin et al., 2003). Nestlings were considered as extra-pair offspring if their genotype mismatched their putative father’s at two or more loci. We used Cervus 3.0 (Kalinowski et al., 2007) for parentage assignment and Fstat (version 2.9.3; Goudet, 1995) to calculate deviation from Hardy–Weinberg equilibrium and linkage disequilibrium.

Homozygosity by loci (HL) was used as an estimate of individual genetic diversity. HL takes into account the allelic variability of each locus and thus improves heterozygosity estimate by weighting the contribution of each locus to the homozygosity index, giving more weight to more informative loci (Aparicio et al., 2006). HL values range from 0 (all loci heterozygous) to 1 (complete homozygosity). We computed a heterozygosity parameter (HetHL) as 1-HL. HL was calculated using Rhh, an extension package for R (Alho et al., 2010).

Statistical analyses

Linear mixed-effect models with restricted maximum likelihood were used to evaluate the effects of parental and nestling heterozygosity levels on nestling body weight, tarsus and feather length. To account for the nonindependence among siblings, we included nest identity as a random factor. Explanatory variables in the initial models were male, female and chick heterozygosity and their interaction terms with parasite treatment, brood size, nestling sex and status (extra-pair or within-pair nestling). Treatment, sex and status were included as factors. For the analysis of nestling feather and tarsus length, we additionally fitted the mean of maternal and paternal feather or tarsus length of the genetic parents into the model to account for heritability of these traits.

We used within-group centring (Model 2; Van de Pol & Wright, 2009) to assess whether associations between nestling fitness and nestling heterozygosity resulted from within-nest or between-nest effects. We calculated standardized effect size (r statistics for nonindependent data; Nakagawa & Cuthill, 2007) for the effect of heterozygosity on nestling growth when significant effects of either parental or nestling heterozygosity were found.

For all models, we started with a full model and followed a backward, stepwise elimination procedure based on the Akaike information criterion and maximum likelihood.

The association between parental heterozygosity and offspring heterozygosity was assessed using another linear mixed-effect model with male and female heterozygosity as explanatory variables. Only nestlings with both genetic parents known were included in this analysis. To assess the relationship between female and male heterozygosity, we used two-tailed Pearson correlations. All analyses were performed using the statistical software R, version 2.11.1 (R Development Core Team, 2010).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Genetic diversity

We genotyped 55 females, 54 males and 499 nestlings from 60 nests across a panel of 47 microsatellite loci with a minimal number of 46 loci typed per individual. After Bonferroni correction, none of the loci deviated from Hardy–Weinberg equilibrium, and there was no evidence for linkage disequilibrium between any pair of microsatellite markers used.

Female heterozygosity and male heterozygosity were not significantly correlated (= −0.19, = 0.17, = 51), and neither parental heterozygosity was significantly associated with offspring heterozygosity (maternal: F1,48 = 0.001, = 0.98; Paternal: F1,48 = 2.847, =0.10).

Influence of infestation treatment on HFC

As post hoc tests of linear mixed-effect models revealed no differences between any of the three infestation treatments (10, 50, 100 fleas added), we pooled the data of these infestation treatments for further analysis. Number of nests in the control treatment was somewhat small (NBox = 15) but contained a reasonable high number of nestlings (NNestling = 112). Given that also in this group, data distribution was relatively homogeneous, and results from the mixed model appear robust despite the smaller sample.

Uninfected nestlings were heavier (F1,47 = 12.42, = 0.001, Table 1) than infested nestlings. The deleterious effect of fleas depended on the heterozygosity level of both nestlings and fathers, as indicated by the significant interaction terms between treatment and heterozygosity (Table 1). Chick heterozygosity was positively associated with nestling body mass in infested nests (F1,306 = 5.26, = 0.023, = 0.1), but not in control nests (F1,98 = 2.45, = 0.121, = −0.02, Fig. 1). This association resulted from a within-nest effect, as revealed by the within-group centring (Table 2). Similarly, heterozygosity of the social father was associated with nestling body mass in infested nests but was uncorrelated in parasite-free control nests (infested nests: F1,37 = 17.0, = 0.0002, = 0.17, control nests: F1,10 = 1.61, = 0.234, = −0.08, Fig. 2). Maternal heterozygosity was not associated with nestling mass for both treatments (infested nests: F1,39 = 0.01, = 0.942, control nests: F1,10 = 0.09, = 0.766). Nestling feather and tarsus length did not show a correlation with chick or parental heterozygosity for both treatments, although there was a tendency for a negative correlation between male heterozygosity and feather length (Table 1). The treatment did significantly affect feather length but not tarsus length, and both were associated with mean parental feather and tarsus length, respectively (Table 1).

Table 1.   Linear mixed-effect models testing for an influence of parental and nestling heterozygosity, treatment and other nongenetic terms on nestling size, 16 days after hatching.
EffectVariablesEstimate ± SEFd.f.P
  1. SE, standard error.

  2. Nest identity was included as a random factor in all models to account for nonindependence between siblings.

  3. F and P values of nonsignificant terms are those just before dropping from the model.

  4. *Relative to control group.

  5. †Relative to female nestlings.

  6. ‡Relative to within-pair nestlings.

Nestling body massIntercept22.22 ± 3.86
Chick HetHL−2.13 ± 1.482.081, 4050.15
Female HetHL2.28 ± 2.221.061, 440.31
Male HetHL−4.70 ± 4.840.941, 470.38
Treatment*−15.51 ± 4.4012.421, 470.001
Nestling sex†0.85 ± 0.07137.311, 405< 0.001
Nestling status‡0.02 ± 0.160.011, 3860.92
Brood size−0.07 ± 0.090.491, 430.49
Chick HetHL × treatment*3.96 ± 1.675.631, 4050.018
Male HetHL × treatment*16.19 ± 5.518.631, 470.005
Nestling feather lengthIntercept14.03 ± 10.08 ─
Chick HetHL−0.71 ± 1.630.191, 3940.66
Female HetHL−0.04 ± 3.560.011, 440.91
Male HetHL−7.14 ± 3.683.771, 460.059
Treatment*−1.07 ± 0.455.511, 480.023
Mean parental feather length0.40 ± 0.175.231, 480.027
Nestling sex†0.37 ± 0.184.221, 4020.04
Nestling status‡−0.01 ± 0.380.141, 3920.71
Brood size0.20 ± 0.151.991, 450.17
Chick HetHL × treatment*−1.22 ± 3.980.091, 3910.76
Male HetHL × treatment*−12.69 ± 8.612.171, 430.15
Nestling tarsus lengthIntercept11.98 ± 2.27 ─ ─
Chick HetHL−0.44 ± 0.371.451, 3990.23
Female HetHL0.88 ± 0.831.711, 450.19
Male HetHL0.85 ± 0.900.91, 460.35
Treatment*−0.01 ± 0.110.001, 440.96
Mean parental tarsus length0.38 ± 0.1211.031, 490.002
Nestling sex†0.50 ± 0.04155.321, 400< 0.001
Nestling status‡−0.07 ± 0.090.681, 3900.41
Brood size0.03 ± 0.030.771, 480.38
Chick HetHL × treatment*1.11 ± 0.911.491, 3890.22
Male HetHL × treatment*2.62 ± 2.171.451, 430.24
image

Figure 1.  Nestling body mass (mean mass per nest box) in relation to nestling heterozygosity and parasite treatment. Intercept and slope of the line are those obtained by the linear mixed-effect model. Control: open circles and dashed line. Flea treatment: black dots and solid line.

Download figure to PowerPoint

Table 2.   Linear mixed-effect model based on within-group centring to assess the within-nest and between-nest relationship between nestling weight and nestling heterozygosity while controlling for the influence of treatment, nestling sex and paternal heterozygosity.
 Estimate ± SEFd.f.P
  1. Nest identity was included as a random factor in all models to account for nonindependence between siblings.

  2. *Relative to control group.

  3. †Relative to female nestlings.

Intercept20.65 ± 8.08
Male HetHL−4.28 ± 5.190.681, 450.42
Treatment*−17.90 ± 8.834.101, 450.05
Nestling sex†0.85 ± 0.07137.871, 405< 0.001
Within-nest−2.19 ± 1.502.121, 4050.15
Between-nest−0.47 ± 7.660.011, 450.95
Male HetHL × treatment*16.31 ± 5.837.821, 450.007
Within-nest × treatment*3.78 ± 1.704.941, 4050.03
Between-nest × treatment*7.20 ± 8.450.721, 450.40
image

Figure 2.  Nestling body mass (mean mass per nest box) in relation to social male heterozygosity and parasite treatment. Intercept and slope of the line are those obtained by the linear mixed-effect model. Control: open circles and dashed line. Flea treatment: black dots and solid line.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Studies on HFCs report wide variation among years or populations, and it has been suggested that this inconsistency arises due to environmental conditions (Coltman & Slate, 2003; Balloux et al., 2004). Abiotic factors have been shown to influence the strength of HFCs (Lesbarreres et al., 2005; Halverson et al., 2006; Marr et al., 2006), but the role of parasites as mediators of HFCs has been experimentally investigated in one single study only (Coltman et al., 1999). In Soay sheep, heterozygosity predicted winter survival among animals infected with gastrointestinal parasites. In contrast, among individuals treated with an anthelmintic therapy, survival was independent of heterozygosity levels (Coltman et al., 1999).

Here, we found that experimentally induced flea infestation did significantly affect the correlation between heterozygosity and fitness in nestling great tits. We found positive effects of offspring heterozygosity on nestling body mass in infested nests but no significant relationship in parasite-free control nests. This association could be attributed to a within-nest effect, meaning that in treated nests, heterozygous nestlings were heavier than their siblings.

Heterozygous individuals have been suggested to be less susceptible to parasites (Coltman et al., 1999; Acevedo-Whitehouse et al., 2005; Hawley et al., 2005). Thus, our results suggest that heterozygous nestlings benefit either from an increased parasite resistance or from parasite tolerance due to their higher genetic diversity and are therefore able to invest more into development and growth in spite of the parasite load. Alternatively, heterozygous individuals might be of higher vigour in general and could therefore cope better with the negative impact of parasite infestation. Both would lead to the observed HFC in infested nests. Heterozygosity-dependent parasite resistance and/or tolerance may also modulate HFCs via other life history traits and thus more generally explain the observed variation in HFCs, given that parasite abundance often varies among populations and years (e.g. Krasnov & Lareschi, 2010; Gomez-Flores et al., 2011; Jachowski et al., 2011).

We found effects of flea prevalence but not infestation intensity on the strength of the HFCs. The lack of intensity effects might be explained by density-dependent processes within the flea population, as reproductive rate of hen fleas decreases when the density of the founder flea population is high (Tripet & Richner, 1999).

Regarding parental heterozygosity, we found that males with higher heterozygosity levels had heavier offspring when breeding in flea infested nests. In control nests, paternal heterozygosity did not significantly influence nestling body mass. As nestling body mass at fledging is a strong predictor of post-fledging survival in great tits and other species (Tinbergen & Boerlijst, 1990), this might result in a higher reproductive success for heterozygous males. The effect of male heterozygosity may be explained by an increased parental care. Males have been shown to increase food provisioning rates by more than 50% if nests are infested by ectoparasites (Christe et al., 1996), and thus, the strong effect of male heterozygosity on nestling weight may appear if more heterozygous males encounter lower costs when intensifying feeding effort.

Results of male and nestling heterozygosity may be interpreted with caution, however, as the contribution of parental and offspring heterozygosity to offspring growth was not experimentally separated in the present study, for example by cross-fostering nestlings. However, parental heterozygosity was not correlated with parental condition (data not shown) and offspring heterozygosity was not significantly correlated with parental heterozygosity. Additionally, male heterozygosity was significantly associated with mean body mass of both within- and extra-pair offspring (within-pair: = 0.35, = 0.01, = 51; extra-pair: = 0.62, = 0.006, = 18).

In conclusion, the study provides experimental evidence that in a natural great tit population, the occurrence of HFCs depends on parasite pressure, providing strong evidence for parasites as mediators of HFC.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We thank Fabrice Helfenstein and Michael Coslovsky for advice on statistical questions, discussions and support and two anonymous reviewers for constructive comments. The experiments were conducted under a license of the Ethical Committee of the Agricultural Office of the Canton Bern, Switzerland. The work was funded by Swiss National Science Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References