Assessing rates of parasite coinfection and spatiotemporal strain variation via metabarcoding: Insights for the conservation of European turtle doves Streptopelia turtur

Abstract

Understanding the frequency, spatiotemporal dynamics and impacts of parasite coinfections is fundamental to developing control measures and predicting disease impacts. The European turtle dove (Streptopelia turtur) is one of Europe's most threatened bird species. High prevalence of infection by the protozoan parasite Trichomonas gallinae has previously been identified, but the role of this and other coinfecting parasites in turtle dove declines remains unclear. Using a high‐throughput sequencing approach, we identified seven strains of T. gallinae, including two novel strains, from ITS1/5.8S/ITS2 ribosomal sequences in turtle doves on breeding and wintering grounds, with further intrastrain variation and four novel subtypes revealed by the iron‐hydrogenase gene. High spatiotemporal turnover was observed in T. gallinae strain composition, and infection was prevalent in all populations (89%–100%). Coinfection by multiple Trichomonas strains was rarer than expected (1% observed compared to 38.6% expected), suggesting either within‐host competition, or high mortality of coinfected individuals. In contrast, coinfection by multiple haemosporidians was common (43%), as was coinfection by haemosporidians and T. gallinae (90%), with positive associations between strains of T. gallinae and Leucocytozoon suggesting a mechanism such as parasite‐induced immune modulation. We found no evidence for negative associations between coinfections and host body condition. We suggest that longitudinal studies involving the recapture and investigation of infection status of individuals over their lifespan are crucial to understand the epidemiology of coinfections in natural populations.

Successful storage of Trichomonas gallinae on Whatman FTA cards following culture

“…identified; consistent with additional studies from the same site (Thomas 2017;Young et al unpubl. data) and reported in more detail elsewhere (Thomas et al 2022). Overall, the samples stored in ethanol gave longer sequence reads than the samples stored on FTA cards (Mann Whitney-U, W = 111, p < 0.001; Fig.…”

Evidence for strain-specific virulence ofTrichomonas gallinaein African columbiformes

“…We isolated the C6 sub-type from 2 laughing doves (A/N ON936878): this sub-type has previously been reported from a booted eagle Hieraaetus pennatus in Spain (A/N KP099941). As previously reported (Thomas et al ., 2022 ), we identified sub-type C8-TD (A/N MT418242) from 1 turtle dove. All 3 novel sub-types were identified from laughing doves, with C9-LD (A/N ON93689), C10-LD (A/N ON936881) and C12-LD (A/N ON936875) isolated from 5, 1 and 1 individual, respectively; C9-LD (A/N ON936880) was also isolated from 1 vinaceous dove.…”

“…We obtained good quality sequence from 119 birds (81%; 13 black-billed wood doves; 23 laughing doves; 41 Namaqua doves; 41 turtle doves and 1 vinaceous dove; Fig. 1 ); prevalence and sequence data for turtle doves has been reported elsewhere (Thomas et al ., 2022 ), but are summarized below for comparison. The most common strain present, GEO, was identical to one previously isolated from Australasian columbids [GenBank accession number (hereafter A/N) JQ755287], and was found in 6 black-billed wood doves, 9 laughing doves, 41 Namaqua doves and 25 turtle doves.…”

“…Positive PCR products were either purified using Wizard SV Gel & PCR Clean-Up System (Promega, Southampton, UK) and sent for bidirectional Sanger sequencing by Beckman Coulter Genomics (Takeley, Essex, UK) ( n = 52), or prepared for sequencing on an Illumina MiSeq ( n = 101) as part of a wider project (Thomas, 2017 ; Thomas et al ., 2022 ). Details of preparation for Illumina sequencing are published in detail elsewhere (Thomas et al ., 2022 ), and summarized in the Supplementary material.…”

“…Sequences returned from Sanger sequencing were manually assessed for sequencing errors, trimmed and aligned in BioEdit (Hall, 2005 ). Analysis of MiSeq sequences is described in detail elsewhere (Thomas et al ., 2022 ), and available in the Supplementary material. Following processing, all sequences were queried using the NCBI-BLAST algorithm (Altschul et al ., 1997 ) to assign strain identity.…”