Avian malaria in a remnant population of red-fronted parakeets on Little Barrier Island, New Zealand

Abstract

Avian malaria is recognised as a potential threatening factor for endangered New Zealand birds; nevertheless, analyses of its prevalence are few and often retrospective, following outbreaks in managed species. We conducted an opportunistic polymerase chain reaction (PCR)-based survey for Plasmodium on a remnant population of red-fronted parakeet (Cyanoramphus novaezelandiae) on Little Barrier Island alongside an analysis of haematology profiles as a first assessment of the effects of this parasite on parakeets. We sampled 22 parakeets and detected Plasmodium relictum DNA in nine samples (prevalence 40.9%; 95% CI = 20.49–61.51%). One successfully amplified sequence corresponded to P. relictum haplotype GRW4. Lymphocyte and heterophil to lymphocyte counts did not differ between PCR-positive and PCR-negative parakeets. However, it is unclear which state of the infection cycle the parakeets experienced during our sampling scheme. From a management perspective, our results indicate that translocation of parakeets from Little Barrier Island to sites where introduced reservoirs of P. relictum occur is a sound management option given the current exposure to this microorganism at the source site.

Keywords:

• pathogen
• avian malaria
• prevalence
• wildlife
• parakeet
• health
• New Zealand

Introduction

Avian malaria is a haemoparasitic infectious disease caused by various lineages of protozoans of the genus Plasmodium (Fallon et al. 2003; Schrenzel et al. 2003; Beadell et al. 2004) whose intermediate hosts are mosquito species (Diptera: Culicidae) (Piovesan-Alves et al. 2005). The disease occurs worldwide and affects numerous bird species causing anaemia, dyspnoea, weakness and in extreme cases progressive anorexia and death (Grenier & Ritchie 1994). Avian malaria is considered the driving factor behind various bird extinctions in Hawaii following the spread of the invasive mosquito Culex quinquefasciatus (van Riper et al. 1986). Avian malaria has also caused local population declines of extant species such as the amakihi (Hemignathus virens) (Foster et al. 2007). Consequently, avian malaria is considered a disease of concern for the management of endangered bird species in captivity as well as in the wild (Ricklefs & Fallon 2002; Atkinson et al. 2005; Belo et al. 2009).

The recent introduction of C. quinquefasciatus to avian endemic-rich archipelagos such as Galápagos Islands (Whiteman et al. 2005; Bataille et al. 2009) and New Zealand (Laird 1995; Tompkins & Gleeson 2006) has raised concerns about the potential effects of a novel disease on immunologically naïve populations and species (Miller et al. 2001; Parker et al. 2006). In addition, New Zealand has one of the most extinction-prone avifaunas in the world (Sekercioglu et al. 2004) and there is potential for synergistic impacts of introduced pathogens on these endangered and genetically depauperate remnant populations (Tompkins et al. 2006; Hale & Briskie 2007).

Furthermore, in New Zealand, in addition to C. quinquefasciatus, one native mosquito species Culex pervigilans has been recently identified as a potential vector of Plasmodium relictum (Tompkins & Gleeson 2006; Massey et al. 2007) indicating the likelihood of a complex infection network involving native and introduced vectors and hosts (Tompkins & Gleeson 2006; Massey et al. 2007). Tompkins & Gleeson (2006) reported that widespread introduced passerines in New Zealand such as house sparrow (Passer domesticus), song thrush (Turdus philomelos) and blackbird (Turdus merula) could act as reservoirs for malarial pathogens. Given the low host specificity of Plasmodium spp. (Beadell et al. 2004), it is likely that the range of avian reservoirs in New Zealand also includes common native species; however, this topic has received little attention.

Mortalities of endangered New Zealand birds related to Plasmodium infection have been documented, including captive yellowheads (Mohoua ochrocephala) (Derraik et al. 2008), New Zealand dotterel (Charadrius obscurus) (Reed 1997) and kea (Nestor notabilis) (Bennett et al. 1993). Plasmodium has also been detected in free-living saddlebacks (Philesturnus carunculatus carunculatus) (Hale 2008). Yet, although the potential threat of avian malaria for endangered New Zealand birds is recognised, the number of published field surveys to identify prevalence and diversity of Plasmodium among native birds in the archipelago are limited.

One additional complication to manage the risk of avian malaria and other diseases on threatened New Zealand species is the widespread practice of translocation as a conservation tool (Armstrong & McLean 1995). During translocations, the density and composition of species on managed sites is altered by the removal of individuals from a source population and subsequent release at a new site (Ar mstrong et al. 1999; Taylor et al. 2005). During this process, microorganisms can be introduced to previously unexposed populations or translocated individuals can be exposed to a new suite of pathogens at the release site (Viggeers et al. 1993; Cunningham 1996). Increasingly, captive-bred and wild individuals are being released to new sites, often outside their historical distribution (McHalick 1999). Such translocations provide a good opportunity to gather baseline information about the range of microorganisms present at source and release locations (Ortiz-Catedral et al. 2009b, 2009c).

We conducted a polymerase chain reaction (PCR)-based survey for Plasmodium on a remnant population of red-fronted parakeet (Cyanoramphus novaezelandiae) on Little Barrier Island. This site has been identified as a potential source for translocations of this parrot species. The red-fronted parakeet is restricted to predator-free and predator-managed areas of New Zealand (Higgins 1999). Formerly common and widespread, the species declined throughout the archipelago as a result of predation by introduced mammals and hunting (Higgins 1999). Red-crowned parakeets have a generalist diet and inhabit regenerating forest patches as well as forest remnants (Greene 1998; Nixon 1982). Remnant populations and populations established via translocation have been observed to use diverse nesting sites (Greene 2003; Ortiz-Catedral & Brunton 2009). Rapid recoveries in population numbers of the species following the eradication of introduced mammals (Graham & Veitch 2002; Ortiz-Catedral et al. 2009a) and their current distribution suggests that predation pressure rather than habitat specialisation is a key determinant for population establishment and persistence. Given their broad diet, low nest specialisation and susceptibility to predation, they are considered an excellent target species to be translocated from a natural remnant population to predator-free areas undergoing community-led habitat restoration in the Auckland region (Ritchie 2002; Hawley 2005).

For our survey on malaria, we targeted Plasmodium because of the aforementioned mortalities of native New Zealand birds suffering Plasmodium infection and the comparatively low reported occurrence of Haemoproteous in New Zealand (Ishtiaq et al. 2006). Our aims were to determine the prevalence of Plasmodium within a sample of red-fronted parakeets, to determine the strain present in parakeets and to provide a preliminary assessment of the effects of Plasmodium infection using leucocyte counts from blood smears. Such information is an opportune contribution to the growing research on Plasmodium in New Zealand and the diversity of native hosts susceptible to infection by this microorganism.

Methods

Little Barrier Island (36°12′S, 175°04′E) lies in the Hauraki Gulf approximately 80 km north of Auckland, and it is New Zealand's oldest wildlife reserve, established in 1894 (Cometti 1986). The island is approximately 30 km² in area, and its vegetation types include regenerating coastal and kauri (Agathis australis) forests (Hamilton 1961). Most of the island is rugged terrain but an accessible area, Te Maraeroa Flat (Dodd 2007), holds numerous breeding pairs of red-fronted parakeets (Dawe 1979; Greene 2003) and feeding flocks of parakeets are common.

We captured red-fronted parakeets between 6 and 8 May 2008 using mist-nets placed around known feeding grounds in the vicinity of Te Maraeroa Flat. All captured individuals were banded with a uniquely numbered metal band and each parakeet was weighed and measured (folded wing length; tarsus and tail length). Blood samples were obtained by venipuncture of the brachial vein (Campbell 1994) with a 27-gauge sterile needle (Terumo®, Tokyo Japan). Approximately 70 µl of blood were collected using a heparinised capillary tube. Once the approximate volume was obtained, the ends of the capillary tube were sealed with plasticine and kept at 4°C until PCR analysis 2 days after sampling. DNA extraction and PCR amplification of Plasmodium followed Tompkins & Gleeson (2006). A recent review has shown problems associated with malaria tests done using blood samples stored in buffers (Freed & Cann 2006). Thus we limited our mist-netting sampling to only 3 days to ensure delivery of refrigerated blood samples within 2 days of collection and thereby maximising the sensitivity of PCR detection. To determine the species and strain of Plasmodium, we attempted to sequence PCR-positive samples following the methodology described by Massey et al. (2007). Only one sample was successfully sequenced, the rest were damaged during storage and could not be sequenced. We amplified a 355-bp segment from the cytochrome b gene, using primers L15368 (5′-AAA AAT ACC CTT TCC AAA TCT-3′) and H15730 (5′-CAT CCA ATC CAT AAT AAA GCA-3′) developed by Fallon et al. (2003). Samples were then sequenced (forward and reverse) to confirm strain identity. We did not assess prevalence via microscopy on blood slides given the greater precision of PCR to detect haemoparasites (Ricklefs & Sheldon 2007; Waldenströn et al. 2004).

For haematological counts, one drop of blood collected with a non-heparinised capillary tube was dropped on a microscope slide and smeared to a thin layer following the ‘push-slide’ method (Walberg 2001). Non-heparinised capillary tubes were chosen because heparin is reported to affect the quality of staining for blood cell counts (Walberg 2001). Slides were then air-dried, fixed in 100% methanol (Bennett 1970) and stored in a slide container until staining 2–4 days after collection. Slides were stained with May Grunwald–Giemsa followed with a phosphate buffer/rinse (Robertson & Maxwell 1990; Technecult Laboratories Ltd., Napier, New Zealand). White blood cell counts were completed by haematologists at Gribbles Veterinary Pathology (Auckland, New Zealand) following the methodology described in Parker et al. (2006). We used lymphocyte counts and heterophil/lymphocyte ratio as indicators of parasite induced stress and immunocompetence, these blood cell counts represent an inexpensive technique widely used in wild bird studies (Fokidis et al. 2008; Salvante 2006). Because of the known proliferation of lymphocytes in response to infection (Martin et al. 1994), and the decrease in heterophil/lymphocyte ratio during stress in birds (Gross & Siegel 1983), we expected PCR-positive individuals to present a higher lymphocyte count and a decreased heterophil/lymphocyte ratio than PCR-negative individuals.

We determined sample size and confidence intervals for prevalence using Episcope© 2.0 (www.clive.ed.ac.uk/winepiscope). A minimum sample size of n=14 is required to detect Plasmodium at a prevalence of 20%, at the 95% confidence interval (CI). We assumed a 90% sensitivity of the PCR test and 95% specificity (Beadell et al. 2004) and a population size of 6000 (Ortiz-Catedral et al. 2009c). Prevalence values of avian malaria across New Zealand vary from 5% to 53% among introduced passerines near urban areas (Tompkins & Gleeson 2006). For our calculation of sample size, we assumed a 20% prevalence value, intermediate within this range. We considered an intermediate value would apply for this island population given that, although isolated from urban areas, it is within 20 km of highly modified urban and suburban areas where introduced passerines are common. Furthermore, known avian reservoirs of Plasmodium such as thrushes, house sparrows and blackbirds are common on Little Barrier Island (Ortiz-Catedral, pers. obs.). All statistical analyses were completed in StatView© Version 5.0.1. Normality of the dataset was tested using the Kolmogorov–Smirnov test. Finally, we compared the total leucocyte count between Plasmodium PCR-positive and PCR-negative individuals using an independent Student t-test.

Results

In total, eight males and 14 females (n=22) were captured, measured and tested for Plasmodium via PCR. We detected Plasmodium DNA in nine samples (four males, five females; 40.9% prevalence, 95% CI 20.49–61.51%). We successfully sequenced one PCR product with a 100% match to P. relictum haplotype GRW4 (Genbank accession number AY099041.1). PCR-positive vs PCR-negative individuals did not differ significantly in leucocyte abundance in blood smears (PCR-positive leucocytes = 4.95±1.10×10⁹/l, n=9; PCR-negative leucocytes = 5.65±0.72, n=14; t-test_leucocytes=0.56, P=0.58). Likewise, the number of lymphocytes and heterophils as well as their ratio did not vary significantly between PCR-positive vs PCR-negative individuals (PCR-positive lymphocytes=3.98± 1.07×10⁹/l; PCR-negative lymphocytes=4.46 ±0.59×10⁹/l; t-test_lymphocytes=0.42, P=0.67; PCR-positive heterophils=0.57±0.15×10⁹/l; PCR-negative heterophils=0.39±0.09×10⁹/l; t-test_heterophils=−1.07, P=0.30; PCR-positive heterophil/lymphocyte ratio=0.18±0.04×10⁹ /l; PCR-negative heterophil/lymphocyte ratio =0.09±0.02×10⁹/l; t-test_{heterophil/lymphocyte ratio}=−2.01, P=0.62).

Discussion

Infection by P. relictum was common in our sample of parakeets from Te Maraeroa flats; however, given the habitat complexity of Little Barrier Island, it is likely that prevalence varies according to features such as micro-climate, proximity to streams, forest structure and composition, and other factors that might determine the distribution of vectors. In the Auckland region, mosquito larvae of several species inhabit pools formed in leaf axils of kahakaha (Collospermum hastatum) and fallen leaves of nikau (Rhopalostylis sapida) (Derraik 2009). These plants are common on Little Barrier Island in addition to wetlands noticed close to our sampling sites, which combined are likely to provide suitable habitat for mosquito species. However, we did not assess Plasmodium prevalence in relation to spatial distribution of potential mosquito habitat.

Despite the high prevalence of P. relictum during our sampling period, we found no indication of acute malarial infection in our sample of parakeets when analysing blood cell counts. Experimental infection with malaria causes increase in white blood cells in chickens (Gallus gallus) (Silveira et al. 2009) and a significant decrease in locomotory activities and body condition of apapane (Himatione sanguinea) (Yorinks & Atkinson 2000). It is possible that acutely infected parakeets were not detected because our sampling methodology was biased towards mobile and active parakeets, whereas weak and stationary individuals might be unlikely to fly into mist-nets. Alternatively, the prevalence and effects of malaria infection might have a seasonal component not detected in our snapshot sampling.

The lymphocyte counts and heterophil to lymphocyte ratios were similar between PCR-positive and PCR-negative parakeets indicating a non-discernible effect of Plasmodium infection as assessed by leucocyte counts. Increased lymphocyte counts related to malarial infection have been documented in various bird species including thrushes (Turdus spp.) (Ricklefs & Sheldon 2007) and great tits (Parus major) (Ots & Horak 1998). An increase in lymphocytes is related to activation of the immune system (Campbell 1994). However, given that our study focused on free-living parakeets, nothing is known about the onset and stage of the Plasmodium infection at the time of sampling.

Besides malaria, beak and feather disease virus (BFDV) has been detected in our study population (Ortiz-Catedral et al. 2009c) but such virus was not detected in our study parakeets. Likewise, Campylobacter, Salmonella and Yersinia were not found in the sample group (Ortiz-Catedral et al. 2009b). Thus a confounding effect of other infections influencing the leukocyte counts of non-PCR-positive parakeets seems unlikely. Nonetheless, we acknowledge that microorganisms not tested for in this study might also be present in our study population and affect leucocyte counts.

From our results, it is not possible to determine whether parakeets pass through an acute malarial infection phase and subsequently recover with resistance to re-infection as it has been documented for the amakihi (Atkinson et al. 2001). The generalist P. relictum strain found on Little Barrier Island is found in a broad range of avian hosts in New Zealand (Tompkins & Gleeson 2006; Sturrock & Tompkins 2008) but its effects among species of conservation concern remain largely unknown. Recent research in managed populations of New Zealand birds has shown that population bottlenecks resulting from translocation (Hale & Briskie 2007) or severe population declines related to predation and habitat modification (Tompkins et al. 2006) can lead to a reduced immune response. Thus, although no discernible negative effects were detected in a remnant population of red-fronted parakeets, it is likely that the strain found in this study may cause common symptoms associated with malaria among bottlenecked populations of the same bird species. Finally, to detect discernible health effects of malaria infection on the Little Barrier Island population, further sampling to include the temporal variability of the infection and its effects is recommended in further surveys of this and other naturally occurring pathogens of parakeets.

From a management perspective, we conclude that the use of Little Barrier Island as a source of parakeets for translocation to nearby areas (where potential reservoirs of malaria exist) is unlikely to increase the risk of exposure of the species to a generalist strain of Plasmodium found throughout New Zealand (Tompkins & Gleeson 2006). However, we recommend further research into the effects of Plasmodium infection in managed populations of this species. Lastly, the diversity of Plasmodium on the Little Barrier Island population might include haplotypes other than GRW4 not detected in our sampling. GRW4 has been detected in other bird species around the world, such as great reed warbler (Acrocephalus arundinaceus) in Israel and Kenya (Bensch et al. 2000) and house sparrows (Passer domesticus) in California (Schrenzel et al. 2003). As mentioned before, further sampling on Little Barrier Island and elsewhere in New Zealand including other seasons and hosts might reveal a greater diversity of Plasmodium and the range of species it affects.

Acknowledgements

This research was completed with the logistical and financial support of the New Zealand Department of Conservation, Institute of Natural Sciences (Massey University), Motuihe Island Trust, Landcare Research and National Council of Science from Mexico (CONACYT). We thank numerous volunteers who made the collection of samples possible and to Dan Tompkins (Landcare Research) for helpful comments to earlier drafts of the manuscript. Thanks also to Karen Cooper, Lab Manager of Gribbles Pathology for comments and advice on sample preparation in the field. This research was conducted under full approval of the New Zealand Department of Conservation (permits AK-15300-RES, AK-20666-FAU and AK-22857-FAU) and Massey University Animal Ethics Committee (protocols MUAEC 07/138 and 08/24).