Fig 1.
Using a captive experimental approach, we assayed house finches from seven populations (grey circles) spanning the temporal invasion gradient of the bacterial pathogen, Mycoplasma gallisepticum (MG).
From oldest to most recent, these are: Virginia (VA), Alabama (AL), Iowa (IA), Washington (WA), California (CA), Arizona (AZ), and Hawaii (HI). This space-for-time approach leverages the well-documented spread of MG eastward from VA beginning in the early 1990s. After reaching the northwestern US around 2004 (e.g., WA), MG spread along the western coast (CA) and into some populations in the desert southwest within the last 0–10 years (e.g., AZ). MG has not been detected in finches in HI [16,19–27]. Use of line drawing maps created using the maps package in R is allowed under a general public license: https://cran.r-project.org/web/licenses/GPL-2.
Table 1.
Experimental design for the longitudinal infection studies (Experiment 1 and 2), where birds from each population were randomly assigned to one of three inoculation treatments.
We collected pre-inoculation samples from all birds, which served as controls. Sample sizes are not even across dose / isolate treatments due to the ethical and logistical limitations of working with wild vertebrates from seven geographically separated populations.
Table 2.
Summary of treatments for the gene expression study (Experiment 3).
In this study, we used a subset of four populations representative of the extremes of MG endemism and tolerance (two “Less-tolerant”, two “More-tolerant”). Harderian glands were collected from all individuals three days post-inoculation with M. gallisepticum. *The sample from one bird from HI showed a near zero read-count after sequencing and was excluded from subsequent analyses.
Fig 2.
House finch populations (n = 2–3 populations per endemism category) in which the bacterial pathogen, Mycoplasma gallisepticum (MG) has been endemic longer (20–25 years, lighter symbols and shading, dotted line) were more tolerant of experimental infection than house finch populations with little or no history of MG endemism with both (a) an evolutionarily basal, lower-virulence MG isolate (n = 141 birds from 7 populations) and (b) an evolutionarily derived, higher-virulence isolate (n = 61 birds from 6 populations). We quantified tolerance as the ability to maintain health (y-axis, lower maximum conjunctivitis scores) despite increasing pathogen loads (x-axis; log10(maximum pathogen load + 1)). The most pronounced difference we detected in resistance, measured as lower log10(maximum pathogen load + 1), was within the low-dose, low-virulence treatment, between populations with 0–10 years of pathogen endemism (dark triangles) and those with 10–20 years of pathogen endemism (medium-shaded triangles). For the low-virulence isolate (a), we used data collected from the same individuals both pre-infection and at post-infection peaks in pathogen load and conjunctivitis scores following either a low- or high-dose inoculation (7.5x102 and 7.5x106 CCU/mL, respectively). For high-virulence infections (b), we performed analyses similarly, but used only the 7.5x102 CCU/mL inoculation dose (see main text). Lines show predictions from generalized linear mixed effects models with shading representing bootstrapped 95% confidence intervals. Data have been jittered in the graph to clearly visualize all points.
Fig 3.
The percentage of individuals with asymptomatic infections differed between populations when birds were experimentally infected with a lower-virulence MG isolate (a), a difference that was most pronounced between populations with 10–20 years and 0–10 years of MG endemism. When birds were experimentally infected with a higher virulence isolate (b), there were no strong differences in the percentage of asymptomatic infections based on years of MG endemism. Sample sizes are included above bars, expressed as the number of asymptomatic infections/all successful infections. Only one experimentally inoculated animal showed no pathogen load following inoculation and was removed from the analysis. Error bars show 95% confidence interval estimates derived using Wilson scores in the ‘prevalence’ package in R.
Fig 4.
In house finch Harderian glands, we identified ˜90% fewer differentially expressed genes when comparing non-infected controls to MG-infected birds from populations that are more tolerant to MG (light circle) and populations that are less tolerant to MG (dark circle).
There were only 50 genes differentially expressed between non-infected controls and infected individuals in all populations.
Fig 5.
Birds from less-tolerant populations (dark bars) had differentially expressed genes for a higher number of immune system Gene Ontology Biological Process terms than did birds from more-tolerant populations (light bars).
These terms are significantly enriched for birds from the less-tolerant populations but not for birds from more-tolerant populations. Genes that were upregulated in infected birds compared to non-infected birds are represented to the right of zero and genes that were downregulated in infected birds compared to non-infected birds are represented to the left of zero. Data are from 39 individuals from four populations, each with five infected and five uninfected animals (except Hawaii, as data from one uninfected animal was excluded due to low reads, see main text).