INTRODUCTION
Parasite distribution shows distinct heterogeneity in wild animal populations (Poulin, Reference Poulin2007). In different species, which are hosts for a given type of parasites (e.g. helminths, mites, ticks), the proportion of infected individuals (i.e. prevalence), richness and diversity of the parasite community may range from very similar up to very dissimilar (Clayton and Moore, Reference Clayton and Moore1997; Schmid-Hempel, Reference Schmid-Hempel2011). Such pattern of variation in parasitic infections may be primarily associated with host life-history traits, host ecology and the parasite characteristics. Since parasites may constitute the important selection drivers in animal populations (Schmid-Hempel, Reference Schmid-Hempel2011), exploring the patterns of variation in parasitic infections among host species is important for understanding many of the biological processes, e.g. mate choice (Hamilton and Zuk, Reference Hamilton and Zuk1982).
Birds act as hosts for a highly diverse and geographically widely distributed vector-transmitted haematozoan parasites from the genera Plasmodium and Haemoproteus (phylum: Apicomplexa, order: Haemosporida) (Valkiūnas, Reference Valkiūnas2005). The majority of bird species have been confirmed to harbour these parasites (Atkinson and Van Riper, Reference Atkinson, Van Riper, Loye and Zuk1991), although the prevalence and their diversity substantially differ among the host species (Scheuerlein and Ricklefs, Reference Scheuerlein and Ricklefs2004). In some species, none or only a small fraction of individuals in the population are being infected (Yohannes et al. Reference Yohannes, Križanauskienė, Valcu, Bensch and Kempenaers2009; Krams et al. Reference Krams, Suraka, Rattiste, Āboliņš-Ābols, Krama, Rantala, Mierauskas, Cīrule and Saks2012), while in the others all or nearly all birds in the population carry an infection (Van Rooyen et al. Reference Van Rooyen, Lalubin, Glaizot and Christe2013). Despite the great number of studies focusing on the prevalence and diversity of haemosporidian parasites, surprisingly little is known about patterns of differences in these two parameters among bird species. This is because in most studies data originate either from different years, different time of sampling within a year or a breeding cycle, different study sites or is based on a small sample size, all of which may affect the estimates of prevalence and parasite diversity (Jovani and Tella, Reference Jovani and Tella2006; Arriero and Møller, Reference Arriero and Møller2008; Svoboda et al. Reference Svoboda, Marthinsen, Turčoková, Lifjeld and Johnsen2009; Zamora-Vilchis et al. Reference Zamora-Vilchis, Williams and Johnson2012). Among- and within-year variability in the prevalence and the diversity of haemosporidian communities is common in avian hosts (Garvin and Greiner, Reference Garvin and Greiner2003; Bensch et al. Reference Bensch, Waldenström, Jonzén, Westerdahl, Hansson, Sejberg and Hasselquist2007; Cosgrove et al. Reference Cosgrove, Wood, Day and Sheldon2008). In temperate regions the prevalence peaks in the spring, which coincides with the relapse of chronic infections (Valkiūnas, Reference Valkiūnas2005; Cosgrove et al. Reference Cosgrove, Wood, Day and Sheldon2008). Commonly observed differences in the prevalence and parasite diversity among geographically distant sites (Pagenkopp et al. Reference Pagenkopp, Klicka, Durrant, Garvin and Fleischer2008; Szöllősi et al. Reference Szöllősi, Cichoń, Eens, Hasselquist, Kempenaers, Merino, Nilsson, Rosivall, Rytkönen, Török, Wood and Garamszegi2011) are probably primarily driven by different local communities of vectors and parasites as well as environmental conditions which may affect the activity of vectors and the development of parasites (Martínez-de la Puente et al. Reference Martínez-de la Puente, Merino, Lobato, Rivero-de Aguilar, del Cerro, Ruiz-de-Castañeda and Moreno2009). The number of screened host individuals and the number of infected individuals affect the estimates of the prevalence and parasite richness respectively, as at low sample sizes the accuracy of the prevalence estimates is low and richness maybe underestimated (Jovani and Tella, Reference Jovani and Tella2006; Jenkins and Owens, Reference Jenkins and Owens2011). Because of these factors, the most reliable estimates of the prevalence and diversity of the parasite community in different species should be derived from large sample sizes collected at the same location and the same year. There are only few studies presenting data on the prevalence and diversity of haemosporidian parasites in closely related bird species, which fulfil these requirements (Shurulinkov and Chakarov, Reference Shurulinkov and Chakarov2006; Wiersch et al. Reference Wiersch, Lubjuhn, Maier and Kampen2007; Kulma et al. Reference Kulma, Low, Bensch and Qvarnström2013; Scordato and Kardish, Reference Scordato and Kardish2014). However, they mostly consider migratory species, in which differential prevalence and parasite diversity at breeding grounds may at least partly result from differences in vector and parasite exposure at stopover and wintering sites. For example, long distance migrants – Collared (Ficedula albicollis) and Pied Flycatchers (Ficedula hypoleuca) – sampled at the breeding site on the island of Öland (Sweden) – have been shown to differ in the prevalence of infection with haemosporidian parasites by 50% (Kulma et al. Reference Kulma, Low, Bensch and Qvarnström2013). Interestingly, the Pied Flycatcher – the more prevalently infected species – had a less diverse parasite community.
Here, we compare the prevalence and diversity of haemosporidian parasites from the genera Plasmodium and Haemoproteus in two closely related non-migratory passerines – the Great Tit (Parus major) and the Blue Tit (Cyanistes caeruleus) based on molecular screening of blood samples. While data on these two parameters of haemosporidian infections are available from several populations of each species (e.g. Wood et al. Reference Wood, Cosgrove, Wilkin, Knowles, Day and Sheldon2007; Stjernman et al. Reference Stjernman, Råberg and Nilsson2008; Szöllősi et al. Reference Szöllősi, Cichoń, Eens, Hasselquist, Kempenaers, Merino, Nilsson, Rosivall, Rytkönen, Török, Wood and Garamszegi2011; Ferrer et al. Reference Ferrer, García-Navas, Sanz and Ortego2012; Van Rooyen et al. Reference Van Rooyen, Lalubin, Glaizot and Christe2013), only a single study looked into interspecific infection patterns with these parasites in sympatrically breeding populations (Lachish et al. Reference Lachish, Knowles, Alves, Sepil, Davies, Lee, Wood and Sheldon2012). However, this study did not examine the composition of the parasite communities. Here, we compare sympatric populations of Great and Blue Tits, sampled in the same years and the same point of the annual cycle, for the prevalence of infection with Plasmodium and Haemoproteus parasites and richness and the diversity of parasite community at location characterized by a high infection frequency. Great and Blue Tits share an array of characteristics in their breeding biology and ecology. Both species are small and short-living insectivorous hole-nesters. Only females build the nest and incubate the eggs, while both parents feed the young. However, Blue Tits invest more in a single breeding attempt per unit of body mass since they lay larger clutches than Great Tits (Cramp, Reference Cramp1985). We predicted that Great and Blue Tits either (i) do not differ in the prevalence given their similar biology and ecology and close genetic relatedness, (ii) Blue Tits are more prevalently infected than Great Tits because they may exhibit stronger reproductive effort-mediated immunosuppression (Knowles et al. Reference Knowles, Nakagawa and Sheldon2009) as a consequence of their higher investment in reproduction, or (iii) Blue Tits are less prevalently infected than Great Tits because of their predicted lower exposure to vectors resulting from the presence in their nests of plants with insect repelling properties (Petit et al. Reference Petit, Hossaert-McKey, Perret, Blondel and Lambrechts2002). We had no clear prediction about the diversity of the parasite community, although given close genetic relatedness of the two host species we expected that they would harbour similar parasite community (Ricklefs and Fallon, Reference Ricklefs and Fallon2002; Davies and Pedersen, Reference Davies and Pedersen2008).
MATERIALS AND METHODS
Data were collected as part of two projects focusing on fitness consequences of infection with malaria parasites in Great and Blue Tits (Podmokła et al. Reference Podmokła, Dubiec, Drobniak, Arct, Gustafsson and Cichoń2014a ; unpublished results). Birds were sampled during two breeding seasons (2011–2012) in nest-box breeding populations in Southern Gotland, Sweden (57°03′N, 18°17′E). The study site consists of over ten large and several small wood plots (primarily deciduous) separated by arable areas (Fig. 1).
The map of study plots monitored for box-breeding Great and Blue Tits in 2011 and 2012 on the island of Gotland (Sweden). Plots with individuals used for inter-species comparisons of the prevalence and the diversity of haemosporidian infections are depicted in dark grey.
From the middle of April nest-boxes were regularly inspected to determine the first egg-laying date, hatching date (day = 0) and the parameters of reproductive success. Adult Blue Tits were sampled for blood once over the nesting period, during the nestling stage (modal nestlings’ age at catching of adult birds – 14 days), while Great Tits – twice: during the nest-building stage and the nestling stage (modal nestling age at catching of adult birds – 14 days). Birds were caught with traps installed inside the box or with mist nets located in the vicinity of the nest box. The subset of breeding birds of both species was subject to an additional treatment. In the case of Blue Tits, some pairs had their brood size increased by three nestlings on day 2 post-hatching (for a detailed description of the treatment see Podmokła et al. Reference Podmokła, Dubiec, Drobniak, Arct, Gustafsson and Cichoń2014a ), while in Great Tits some females were injected during the nest-building stage with either a physiological salt or an anti-malarial drug – primaquine. Because the experimental increase of brood size is expected to elevate reproductive effort, which in turn may induce immunosuppression and increase blood parasitaemia (Knowles et al. Reference Knowles, Nakagawa and Sheldon2009), and the primaquine may potentially eradicate malaria parasites and influence the probability of developing a new infection (Marzal et al. Reference Marzal, de Lope, Navarro and Møller2005), birds subject to these treatments (21 Blue Tits and 50 Great Tits) were excluded from the analyses. Moreover, because the probability of acquiring infection with malaria parasites may differ at a local spatial scale (e.g. Wood et al. Reference Wood, Cosgrove, Wilkin, Knowles, Day and Sheldon2007), only birds breeding in plots where individuals of both species were sampled (seven wood plots in 2011 and eight wood plots in 2012) were considered in the analyses (Fig. 1). The size of these plots ranges from ca. 0·7 to 19·8 ha and the mean yearly breeding density in these plots in 2011–2012 was 1·8 and 0·6 pairs ha−1 for Great and Blue Tits, respectively.
Birds were ringed, sexed based on the presence of a brood patch and aged as yearlings or at least 2-year-old based on ringing records and plumage characteristics (Svensson, Reference Svensson1992). Blood samples were obtained by venipuncture from the wing vein using non-heparinized capillaries and stored in 96% ethanol in ambient temperature. Genomic DNA was extracted from the blood using either Chelex (Bio-Rad, Munich, Germany) in the case of Blue Tits (Walsh et al. Reference Walsh, Metzger and Higuchi1991) or an ammonium acetate method in the case of Great Tits. The repeatability of PCR results based on chelex and ammonium acetate DNA isolation methods was 97·8% (n = 45 Blue Tit samples from 2015) with a slightly lower detection of infection with the chelex method (40 positive samples vs 41 positive samples based on ammonium acetate method). The presence of haemosporidian parasites (genus Haemoproteus and Plasmodium) was assessed by amplifying 478 bp long fragment of the mitochondrial cytochrome b gene using nested polymerase chain reaction (Waldenström et al. Reference Waldenström, Bensch, Hasselquist and Östman2004). This method is very sensitive and allows one to detect the infection with sensitivity of one infected cell per 10 000 erythrocytes. PCR conditions followed the protocol of Cosgrove et al. (Reference Cosgrove, Wood, Day and Sheldon2008) and PCR products were processed as described in Podmokła et al. (Reference Podmokła, Dubiec, Drobniak, Arct, Gustafsson and Cichoń2014a , Reference Podmokła, Dubiec, Drobniak, Arct, Gustafsson and Cichoń b ). In short, PCR products were run on 2% agarose gel and PCR products of all positive samples were purified and then sequenced uni-directionally (except for novel lineages, which were sequenced bi-directionally) from the 5′ end with the primer HaemF with an automated ABI 3130 DNA analyser (Applied Biosystems). Sequences were edited, aligned and compared with the MalAvi database (Bensch et al. Reference Bensch, Hellgren and Pérez-Tris2009) using BioEdit software (Hall, Reference Hall1999). When individuals harboured multiple infections, i.e. infections caused by two or more lineages simultaneously (indicated by double peaks in the chromatogram), parasite lineages were in most cases assigned following the visual comparison of sequences with the pool of lineages known to occur at the study site. The reliability of this method was confirmed by cloning of PCR products in ten individuals with multiple infections (see Podmokła et al. Reference Podmokła, Dubiec, Drobniak, Arct, Gustafsson and Cichoń2014b for more details). When unambiguous identification of the lineages was not possible, PCR products were cloned. In the case of unique, not previously described lineages (the lineage was considered unique when a difference of at least one nucleotide was present, Bensch et al. Reference Bensch, Hellgren and Pérez-Tris2009), the PCR and sequencing of the sample yielding such sequence were repeated to exclude the possibility of an error. Novel lineages were named following recommendation of Bensch et al. (Reference Bensch, Hellgren and Pérez-Tris2009) and deposited in GenBank (accession no. KU695262- KU695264).
Statistical analyses
To avoid potential differences between host species associated with timing of sampling within the breeding cycle, only samples collected during the nestling period were considered in this study. In total, 124 blood samples from Blue Tits and 363 samples from Great Tits were considered. In the case of six Blue Tits and 38 Great Tits, the samples were collected in both breeding seasons. One positive Blue Tit sample was excluded from the statistical analyses because the amplified product was shorter than the targeted fragment of the parasite gene. Moreover, some samples were excluded from the analyses of the infection type (single vs multiple infections) and the parasite diversity because parasite lineages could not be identified either because of the poor quality of the PCR product or when cloning of the PCR product of multiple infections did not yield all unique lineages. The final data set for the analyses of infection prevalence included 64 female/59 male and 84 yearling/39 older Blue Tit samples and 160 female/203 male and 187 yearling/176 older Great Tit samples. Mean estimates of prevalence were calculated using the infection rates recorded in each season and were based on the dataset which included only one record (selected randomly) per individual to omit the issue of non-independence of data in individuals sampled twice.
Data on infection prevalence were analysed using restricted maximum-likelihood method (REML) implemented in the R package ASReml-R (Butler Reference Butler2009). Since the data was expressed as binary contrasts (presence/absence of infection) we have used generalized linear mixed models with a logit link function. All models were run for 50 optimizing iterations, but in all cases convergence to the maximum-likelihood estimate was achieved within no more than ten iterations.
We have analysed four types of models. In the first three models the response variable described the prevalence of parasitic infection in each individual (three different response variables: presence/absence of haemosporidian infection, presence/absence of Plasmodium infection and presence/absence of Haemoproteus infection). Prevalence at the parasite genus level was based on individuals harbouring single and multiple infections. Consequently, birds which were co-infected with Plasmodium and Haemoproteus were scored as positive for both parasite types. In the fourth model, run on the dataset containing only infected individuals, the response variable described the presence/absence of the multiple infections. Each model included a random categorical effect of individual identity (to account for repeated inclusion of the same individuals in the dataset) and a set of fixed effects. The first three models included: host species, year of sampling, individual age and sex, and all second-order interactions, which contained the term species with the remaining fixed terms. The fourth model (the prevalence of multiple infections) included only species, year of sampling and their interaction because of the very low frequency of multiple infections in Blue Tits (see the Results). Non-significant interactions were removed sequentially, starting with those with the highest P-values. Fixed effects were tested using conditional Wald tests (equivalent to conventional F test) with Satterthwaite method of approximating the number of degrees of freedom.
Richness is expressed as the total number of detected lineages and the lineage diversity as the Shannon index, which combines information on species richness and relative abundance (Shannon and Weaver, Reference Shannon and Weaver1962). The index and its 95% confidence intervals (CIs) were calculated using the software EstimateS v. 9.1.0 (Colwell, Reference Colwell2013). CIs were estimated based on 1000 randomizations. The diversity was considered to differ between compared sets of samples, if CIs did not overlap the mean of the other group. Richness and lineage diversity were calculated based on the dataset containing only one record per individual sampled twice.
RESULTS
Prevalence of infection with haemosporidian parasites
Both host species were frequently infected with haemosporidian parasites with, on average, over 70% of individuals in each species infected during the breeding period with either one or two of the surveyed parasite genera. In general, Great Tits were more frequently infected than Blue Tits in terms of the overall prevalence as well as the prevalence of Haemoproteus infections, while there was no difference between the host species in the Plasmodium prevalence (mean overall prevalence: Great Tits – 97·1%, Blue Tits – 71·2%; mean Plasmodium prevalence: Great Tits – 71·4%, Blue Tits – 62·7%; mean Haemoproteus prevalence: Great Tits – 38·4%, Blue Tits – 9·8%; Table 1, Fig. 2). The prevalence of infections (overall as well as the prevalence at the parasite genus level) differed between the seasons (Table 1). Overall and Plasmodium prevalence decreased from 2011 to 2012, while Haemoproteus prevalence increased during this period (Fig. 2).
The overall prevalence and the prevalence of infections with Plasmodium and Haemoproteus in Great and Blue Tits sampled during the late nestling period on Gotland (Sweden) in two breeding seasons. Data is presented separately for each season. Individuals co-infected with Plasmodium and Haemoproteus were scored as positive for both parasite genera. The number of screened individuals: Great Tits – 118 in 2011 and 207 in 2012, Blue Tits – 41 in 2011 and 76 in 2012.
The association between the probability of infection with haemosporidian parasites in Blue and Great Tits on Gotland (Sweden) with species, year, age (yearlings or older birds) and sex as fixed effects
Data for overall hamosporidian infections and infections with either Plasmodium or Haemoproteus parasites are presented separately. Full models contained all fixed effects and two-way interactions which included the factor ‘species’. Non-significant interactions were removed sequentially, starting with those with the highest P-values. Each model included a random categorical effect of individual identity to account for repeated inclusion of individuals sampled in both breeding seasons.
In both host species sex of the individual did not explain the variation in the prevalence (both overall and at the parasite genus level); however, the probability of being infected was associated with host age (Table 1). Specifically, the overall prevalence and the prevalence of infection with Plasmodium were in both species more common among older birds (mean overall prevalence: Great Tits: 95·3% in yearlings vs 99·0% in older birds, Blue Tits: 65·6% in yearlings vs 83·3% in older birds; mean Plasmodium prevalence: Great Tits: 64·2% in yearlings vs 79·0% in older birds, Blue Tits: 59·8% in yearlings vs 68·8% in older birds), while in the case of infection with Haemoproteus age-related changes in the prevalence differed between the host species. In Blue Tits mean prevalence was over twice higher in older birds than in yearlings, while in Great Tits older birds were slightly less prevalently infected than yearlings (Great Tits: 41·6% in yearlings vs 35·5% in older birds, Blue Tits: 6·8% in yearlings vs 16·6% in older birds).
Great and Blue Tits differed in the frequency of multiple infections, however, the proportion of individuals which harboured such infections did not differ between years (species: F 1, 432·0 = 13·95, P < 0·001, year: F 1, 432·0 = 1·821, P = 0·178). In the subset of infected individuals, multiple infections were found on average in 26·2% of Great Tits and 3·2% of Blue Tits. The most common were double infections – 83·4% (mean prevalence) in Great Tits and all multiple infections in Blue Tits. Great Tits also harboured triple (15·3%) and quadruple infections (1·4%).
Richness and diversity of the parasite community
In total, 11 parasite lineages were detected in the set of Great and Blue Tit samples used in the present study (Fig. 3, Appendix 1). Among eight lineages, which have been already described, five belong to the Plasmodium genus (BT7 and TURDUS1 representing morphospecies P. circumflexum, GRW11 and SGS1 representing P. relictum and SW2 representing P. homonucleophilus) and three to the Haemoproteus genus (PARUS1, PHSIB1 and WW2 representing H. majoris). Out of 11 detected lineages Great and Blue Tits shared five lineages (BT7, PARUS1, PHSIB1, SGS1 and TURDUS1), while six lineages, including three novel ones (GRW11, SW2, WW2, PARUS65, PARUS66 and PARUS67) were found only in Great Tits. Based on the comparison of novel lineages with the pool of lineages deposited in MalAvi database, PARUS65 most closely matches lineages from the Haemoproteus genus (the closest genetic similarity with PARUS1), while PARUS66 and PARUS67 most closely match lineages from Plasmodium genus (the closest genetic similarity with SW2 and TURDUS1, respectively).
The composition and the prevalence of Plasmodium and Haemoproteus lineages detected in sympatric populations of Great and Blue Tits on Gotland (Sweden) in two breeding seasons. Data is presented separately for 2011 and 2012.
The diversity of the parasite community was higher in Great Tits than Blue Tits (Shannon index and 95% CI, 2011: Great Tits 1·64, 1·56–1·72; Blue Tits 0·61, 0·37–0·84; 2012: Great Tits 1·90, 1·82–1·98; Blue Tits: 1·12, 0·94–1·30). Although both Tit species were most commonly infected with Plasmodium lineage TURDUS1, the distribution of infections with other lineages co-occurring in both species differed (Fig. 3, Appendix 1). For example, Haemoproteus lineage PHSIB1 showed a moderate infection rate in Great Tits, while in Blue Tits it was the least prevalent lineage among all detected in this species. In both host species, some lineages were found only sporadically with the mean annual prevalence <2% (Great Tits: GRW11, PARUS65, PARUS66, PARUS 67, Blue Tits: PHSIB1).
DISCUSSION
Based on screening a large sample of individuals from sympatric populations we show that Blue Tits and Great Tits – two closely related passerines – differ in the prevalence and diversity of infections with haemosporidian parasites. Great Tits not only are more prevalently infected with these vector-transmitted parasites, but also harbour a higher proportion of multiple infections and have a more diverse parasite community than Blue Tits. The same pattern of interspecific difference in haemosporidian prevalence in sympatric populations of these two host species has been found in Southern UK (Lachish et al. Reference Lachish, Knowles, Alves, Sepil, Davies, Lee, Wood and Sheldon2012). Since both study sites differ in an array of parasite- and host-associated characteristics (e.g. prevalence, composition of the parasite assemblage, the frequency of multiple infections, ratio of Blue Tit/Great Tit population density), a higher frequency of infections in Great Tits seems to be a consistent pattern.
In general, closely related host species from sympatric populations are expected to show similarity in the susceptibility to parasitic infections and the parasite community because of the common evolutionary background (Ricklefs and Fallon, Reference Ricklefs and Fallon2002; Davies and Pedersen, Reference Davies and Pedersen2008) and exposure to the same vectors in the habitat they occupy. However, currently available data, including this study, seems to challenge this prediction. Not only closely related migratory species breeding sympatrically differ in the prevalence and composition of the haemposporidian community (Shurulinkov and Chakarov, Reference Shurulinkov and Chakarov2006; Kulma et al. Reference Kulma, Low, Bensch and Qvarnström2013; Scordato and Kardish, Reference Scordato and Kardish2014), which may be largely attributed to contracting infections at stopover and wintering sites, but importantly such differences occur in closely related non-migratory species (Lee et al. Reference Lee, Martin, Hasselquist, Ricklefs and Wikelski2006; Wiersch et al. Reference Wiersch, Lubjuhn, Maier and Kampen2007, Lachish et al. Reference Lachish, Knowles, Alves, Sepil, Davies, Lee, Wood and Sheldon2012, but see Jenkins and Owens, Reference Jenkins and Owens2011 for no difference in infection rates of Leucocytozoon, the sister genus to Plasmodium and Haemoproteus, in Great and Blue Tits).
Several factors, including immunological and behavioural characteristics, may contribute to differences between Great and Blue Tits in the prevalence and diversity of the haemosporidian parasite community. Direct comparison of immune function activity of two study species is currently lacking, so it is not possible to make any association between the two parameters reflecting the susceptibility to haemosporidian parasites and immunocompetence. However, in two other closely related passerines – House (Passer domesticus) and Tree Sparrows (Passer montanus) – immune defences have been shown to differ and the species with higher antibody responsiveness – the House Sparrow – has been also found to be less prevalently infected with haemosporidians from the genera Plasmodium and Haemoproteus (Lee et al. Reference Lee, Martin, Hasselquist, Ricklefs and Wikelski2006). In general, based on studies in mammals, it may be expected that species with stronger antibody responsiveness should be better able to control and clear chronic haemosporidian infections (Taylor-Robinson, Reference Taylor-Robinson1995). Regardless of the potential differences in the constitutive and adaptive immunity, immune reaction to infection may differ between these two species if immune function is differently modulated by reproductive effort. Blue Tits invest more in a single breeding event than Great Tits, because they lay more eggs and consequently rear more nestlings per unit of their body mass. As a consequence more pronounced suppression of immune function resulting from a trade-off between immunity and reproductive effort (Knowles et al. Reference Knowles, Nakagawa and Sheldon2009) should be expected in this species. Moreover, in the study area adult Great Tits seem to be more prone than Blue Tits to infestation/infection with other parasitic and pathogenic agents including ticks and pox viruses (own observation). Infestation/infection with other parasites and pathogens may suppress the immune function (e.g. ticks’ saliva contains molecules which activate an anti-inflammatory TH2 response; Andrade et al. Reference Andrade, Texeira, Barral and Barral-Netto2005) resulting in a higher probability of developing haemosporidian infection if the parasite gets transmitted by the vector.
The majority of 1-year old females of both host species are infected already at the nest building stage (Great Tits) or late incubation stage (Blue Tits, unpublished results). Since a prepatent period in Plasmodium and Haemoproteus parasites lasts from a few days to several weeks (Valkiūnas, Reference Valkiūnas2005) and some vectors (e.g. mosquitoes) become active at the earliest at the incubation stage of both Tit species, such infections are most probably contracted by juveniles during the previous year. Such pattern of acquiring infection may indicate that other factors, not associated with reproduction-mediated reallocation of resources, play a role in this process. One factor, which may substantially contribute to differences between species, is the exposure to vectors transmitting haemosporidian parasites. Such differences may be especially present during the nestling stage. Blue Tits rear larger broods, which should attract more blood-sucking dipterans, because more nestlings should produce more cues, which are used by ornithophilic arthropods to locate the host (Russell and Hunter, Reference Russell and Hunter2005; Allan et al. Reference Allan, Bernier and Kline2006). However, on the other hand, Blue Tits incorporate in their nests plants producing volatile compounds (Petit et al. Reference Petit, Hossaert-McKey, Perret, Blondel and Lambrechts2002; review in Dubiec et al. Reference Dubiec, Góźdź and Mazgajski2013), which may act as repellent against parasite-transmitting insects (Lafuma et al. Reference Lafuma, Lambrechts and Raymond2001; Krams et al. Reference Krams, Suraka, Rantala, Sepp, Mierauskas, Vrublevska and Krama2013, but see Tomás et al. Reference Tomás, Merino, Martínez-de la Puente, Moreno, Morales, Lobato, Rivero-de Aguilar and del Cerro2012). Krams et al. (Reference Krams, Suraka, Rantala, Sepp, Mierauskas, Vrublevska and Krama2013) showed that 1-month old Great Tit fledglings from nest boxes treated with the insect repellent (citronella oil) had much lower prevalence and intensity of infection with haemosporidian parasites than fledglings from untreated nest boxes. In the study area some Blue Tit nests contain green plant material (own observation), which may indicate that nestlings of this species are exposed to fewer parasite-transmitting dipterans than Great Tit nestlings. Other potential behaviourally mediated mechanism of different exposure to vectors includes exploitation of foraging microhabitats which vary in vector abundance.
The probability of infection with haemosporidian parasites may also be associated with density of either conspecific or heterospecific hosts. Generally, some theoretical models suggest that the probability of infection with vector-transmitted parasites follows the dynamics along the continuum between pure density-dependent to pure frequency-dependent transmission (density and frequency refer in the models to infected individuals, Antonovics et al. Reference Antonovics, Iwasa and Hassell1995). However, in birds, the probability of vector-transmitted haemosporidians in some host-parasite systems has been shown to be well explained by the overall local host density (Ortego and Cordero, Reference Ortego and Cordero2010; Lachish et al. Reference Lachish, Knowles, Alves, Sepil, Davies, Lee, Wood and Sheldon2012; Isaksson et al. Reference Isaksson, Sepil, Baramidze and Sheldon2013). Lachish et al. (Reference Lachish, Knowles, Alves, Sepil, Davies, Lee, Wood and Sheldon2012) showed that in sympatric populations of Blue and Great Tits, infected primarily with two Plasmodium morphospecies – P. circumflexum and P. relictum, the probability of belonging to P. circumflexum cluster for nest-boxes occupied by Great Tits increased with local density of Great Tits and Blue Tits, while for boxes occupied by Blue Tits – with density of Great Tits, but not conspecifics. We may not exclude that at our study site certain density-dependent processes of acquiring and/or developing of infection also occur contributing to a higher infection frequency in Great Tits.
In both species age but not sex of the host was associated with variation in infection prevalence. Higher frequency of infection with haemosporidian parasites in birds at least 2-year-old than in yearlings observed in the current study (overall and Plasmodium infections in Great and Blue Tits and Haemoproteus infections in Blue Tits) is commonly found in birds (Kulma et al. Reference Kulma, Low, Bensch and Qvarnström2014; Marzal et al. Reference Marzal, Balbontín, Reviriego, García-Longoria, Relinque, Hermosell, Magallanes, López-Calderón, de Lope and Møller2016, but see Zylberberg et al. Reference Zylberberg, Derryberry, Breuner, MacDougall-Shackleton, Cornelius and Hahn2015 for the lack of age-related patterns) and is being attributed to the increased probability of the vector encounter with age. Interestingly, in the case of Haemoproteus in Great Tits older birds were slightly less prevalently infected than yearlings. Possible mechanisms behind such a pattern include a competitive exclusion of Haemoproteus lineages by some Plasmodium lineages (Beadell et al. Reference Beadell, Gering, Austin, Dumbacher, Peirce, Pratt, Atkinson and Fleischer2004) and differential mortality rates in Haemoproteus-infected Great and Blue Tits.
We found a much higher frequency of multiple infections in Great than Blue Tits. Similarly to the general prevalence of infections, the interspecific differences in the frequency of multiple infections in these two host species may be explained by aforementioned factors including immunological and behavioural characteristics and difference in the density. It has to be noted, that the actual rates of multiple infections in both species are possibly markedly higher than reported because PCR protocols underestimate this parameter (Bernotienė et al. Reference Bernotienė, Palinauskas, Iezhova, Murauskaitė and Valkiūnas2016). Limited detection of parasite lineages simultaneously infecting the host may be associated with preferential amplification of lineages with higher parasitaemia (Pérez-Tris and Bensch, Reference Pérez-Tris and Bensch2005; Valkiūnas et al. Reference Valkiūnas, Bensch, Iezhova, Križanauskienė, Hellgren and Bolshakov2006).
Prevalence differed between years, which is a pattern commonly found in studies of haemosporidian infections in birds (Bensch et al. Reference Bensch, Waldenström, Jonzén, Westerdahl, Hansson, Sejberg and Hasselquist2007). Changes in the prevalence between years are most probably associated with fluctuations in the population size and the activity of vectors, which are known to be affected by temperature and rainfall (Martínez-de la Puente et al. Reference Martínez-de la Puente, Merino, Lobato, Rivero-de Aguilar, del Cerro, Ruiz-de-Castañeda and Moreno2009). Apart from its effect on vectors, temperature may also affect rates of parasite development within vectors (Hoshen and Morse, Reference Hoshen and Morse2004). All these factors may in turn affect the transmission frequency of haemosporidian parasites. Since at the parasite genus level only Plasmodium prevalence decreased, while Haemoproteus prevalence increased, most probably weather conditions negatively affected only the population of mosquitoes, but not dipterans transmitting Haemoproteus.
Based on the subset of samples used in this study, Great Tits harbour richer and more diverse parasite community than Blue Tits. This pattern holds even after taking into account four lineages (GRW11, PADOM02, SW2 and WW2, all occurring at very low frequencies), which have been found in the study population of Blue Tits in a set of nearly 1400 samples collected in years 2008–2014. It has to be noted, however, that Great and Blue tits are not necessarily the competent hosts for all detected lineages, especially those occurring sporadically, because haemosporidians may sometimes replicate in non-competent hosts without forming the infective stages, the gametocytes (Valkiūnas et al. Reference Valkiūnas, Iezhova, Loiseau and Sehgal2009). To rule out such a possibility, assessing the presence of gametocytes in blood smears is necessary.
In order to better understand the patterns of haemosporidian infection rates and the variation in the parasite community composition between Blue and Great Tits as well as in other avian hosts, studies at different sites within hosts’ distribution range and in different habitats are required.
ACKNOWLEDGEMENTS
We thank two anonymous reviewers for valuable comments on an earlier version of the manuscript. We also thank Giulia Casasole, Ewa Poślińska, Joanna Sudyka and Javier Lázaro Tapia for assistance with the fieldwork, Adam Krupski for assistance with molecular analyses and Kevin Fletcher for improving the manuscript's English. The study conforms to the legal requirements of Sweden.
FINANCIAL SUPPORT
Financial support was provided by the Polish National Science Centre (A.D., grant number N N303 818340), (E.P., grant number 2011/03/N/NZ8/02106); and the Polish Ministry of Science and Higher Education (M.C., grant number N N304 409838), (S.M.D., grant number N N304 061140). The long-term nest box study was supported by The Swedish Research Council (to L.G.).
Appendix 1
The parasite assemblage and the number of birds infected with Plasmodium and Haemoproteus lineages in Great and Blue Tits on Gotland (Sweden) in years 2011–2012. Dataset includes only one record (selected randomly) per individual for birds which were sampled in both years (six Blue Tits and 38 Great Tits). Number of individuals harbouring a given lineage was calculated based on the set of birds with single and multiple infections. N in brackets denotes the number of individuals screened molecularly for the presence of haemosporidian infections.
