Priority effect theory, a foundational concept from community ecology, states that the order and timing of species arrival during species assembly can affect species composition. Although this theory has been applied to co-infecting parasite species, it has almost always been with a single time lag between co-infecting parasites. Thus, how the timing of parasite species arrival affects co-infections and disease remains poorly understood. To address this gap in the literature, we exposed post-metamorphic Cuban tree frogs (Osteopilus septentrionalis) to Ranavirus, the fungus Batrachochytrium dendrobatidis (Bd), a nematode Aplectana hamatospicula, or pairs of these parasites either simultaneously or sequentially at a range of time lags and quantified load of the secondary parasite and host growth, survival and parasite tolerance. Prior exposure to Bd or A. hamatospicula significantly increased viral loads relative to hosts singly infected with Ranavirus, whereas A. hamatospicula loads in hosts were higher when co-exposed to Bd than when co-exposed to Ranavirus. There was a significant positive relationship between time since Ranavirus infection and Bd load, and prior exposure to A. hamatospicula decreased Bd loads compared to simultaneous co-infection with these parasites. Infections with Bd and Ranavirus either singly or in co-infections decreased host growth and survival. This research reveals that time lags between co-infections can affect parasite loads, in line with priority effects theory. As co-infections in the field are unlikely to be simultaneous, an understanding of when co-infections are impacted by time lags between parasite exposures may play a major role in controlling problematic co-infections.

Amphibian responses to parasites

To measure the growth of amphibian hosts over the duration of the experiment, 24 h after secondary parasite exposure, treefrogs were weighed and snout-vent length (SVL) was measured, and these measurements were taken weekly for four weeks. Survival and disease signs were assessed twice daily. At the time of death, we weighed deceased individuals, measured SVL, and assessed infection for the secondary parasite to which the individual was exposed. Disease signs were determined to be internal hemorrhaging or erythema near the feet for Ranavirus and discoloration, sloughing, or thickening (the latter of which was only noted during dissections) of the skin for Bd (Cunningham et al. 1996, Kilpatrick et al. 2010). For Bd- and Ranavirus-exposed individuals, we swabbed the skin and conducted qPCR (see below).  For A. hamatospicula exposed individuals, we conducted dissections of the gut. Twenty-eight days after exposure to the secondary parasite, all surviving frogs were euthanized and dissected. Host growth rate was calculated as the final mass or snout-vent length (SVL) minus initial mass or SVL divided by weeks alive (g wk^-1).

Assessing parasite loads

To assess A. hamatospicula infection, treefrog feces were checked weekly for larval worms and amphibian hosts were dissected at time of morality and at the end of the experiment to quantify adult A. hamatospicula infection load in the gastrointestinal tract. To assess Bd and/or viral load, amphibians were swabbed 2, 4, 8, 16, and 24 days after exposure to the second parasite and at the time of mortality. Individuals were swabbed 5 times from hip to toe on both rear legs for Bd and around the cloaca and mouth to test for Ranaviral infection. Swabs were placed in 2-ml sterile microcentrifuge tubes and stored at -80 °C until processing.

DNA was extracted (Qiagen DNEasy Blood & Tissue Kit) from 2-, 4, 8, and 24-day swabs where Ranavirus was the second infecting parasite and from day 4-, 8-, 16-, and 24-day swabs where Bd was the second infecting parasite. Day 2 swabs were not extracted for Bd because Bd DNA from the original inoculum can be detected for up to 2 days after dosing, making it impossible to differentiate infection from exposure at this time (McMahon et al. 2014). Bd and Ranavirus swabs were analyzed using quantitative polymerase chain reaction (qPCR) to quantify parasite load (Boyle et al. 2004, Picco et al. 2007; see supplement methods for more details).

Assessing amphibian immune response

We conducted a separate experiment that included 48 additional Cuban tree frogs to quantify amphibian antibody responses to single and simultaneous co-infections with the parasites that we tested above. Four individuals were tested for each of the three single infection treatments, for a total of twelve singly infected individuals. Eight individuals were tested for each of the three, pairwise, co-infection treatments and half of the individuals in each pairwise co-infection treatment were tested for one parasite and the other half were tested for the other. Finally, twelve individuals were tested as controls, with four individuals tested to compare with each parasite (Appendix S1: Table S4). To measure the immune response for each parasite, frogs were euthanized as described above and all available blood (approximately 100-300uL depending on the frog size) was extracted directly from frog hearts with microcapillary tubes 11, 19, and 28 days after exposure to Ranavirus, A. hamatospicula, and Bd, respectively. These times were chosen to reflect when antibody immune responses to each parasite should be detectable (Gantress et al. 2003, Ramsey et al. 2010, Knutie et al. 2017). While we acknowledge that immunity is a complex, multifactorial response, we chose to focus on IgY antibodies because IgY is a metric of acquired immunity, is the most common immunoglobulin, and is used to combat the parasites we tested (Gantress et al. 2003, Maniero et al. 2006, Knutie et al. 2017, Grogan et al. 2018).  General IgY measurements cannot differentiate between Th1 and Th2 immune responses (Bretscher 2014, Menon et al. 2018). However, assuming co-infected individuals would mount higher IgY levels than singly infected individuals to combat both pathogens, no significant increase in IgY levels for co-infected individuals could suggest that a Th1/Th2 tradeoff is limiting the hosts’ ability to mount a significantly higher immune response. 

Blood plasma was used in enzyme-linked immunosorbent assays (ELISA) to detect IgY antibody presence. Individual frog serum was first diluted at 1:100 in carbonate coating buffer (0.05 M, pH 9.60) and then added to ninety-six well plates with 100 μL/well. Each individual sample was run in triplicate. After serum was added, plates were incubated at 4 °C overnight.  Plates were washed five times with 300 μL/well of a Tris-buffered saline wash solution. The plates were then filled with 200 μL/well of bovine serum albumin (BSA) blocking buffer and incubated for 2 h at room temperature on an orbital table. Plates were washed again five times, loaded with 100 μL/well of primary detection antibody (Goat-αAlligator-IgY, diluted 1:1000; Bethyl), placed on an orbital table for 1 h, washed again (five times), loaded with 100 μL/well of secondary detection antibody (Rabbit-αGoat-IgG, diluted 1:5000; Bethyl) for 1 h, and then washed for the final time (five times). Plates were then loaded with 100 μL/well of tetramethylbenzidine (TMB: Bethyl Laboratories) and incubated for 30 mins before the reaction was stopped with 100 μL/well of stop solution (Bethyl Laboratories). Finally, optical density (OD) of the wells was measured with a spectrophotometer (BioTek, PowerWave HT, 450-nanometer filter).  All samples were run in triplicate.

There are two provided data sheets with the Priority Effects Ecology spreadsheet. The first sheet (All Swab Days) was used to analyze parasite loads over time since co-infection. The metadata for this spreadsheet is included and titled Metadata-All Swab Days. The second sheet (Focal Swab Days) was used to analyze how the time lags between infections altered disease progression. The metadata for this spreadsheet is also included and titled Metadata-Focal Swab Days.

There is an additional excel file for the separate immune experiment titled, Priority Effects ELISAs Ecology. Metadata is included.