Maselloetal.Parasites&Vectors (2018) 11:357 https://doi.org/10.1186/s13071-018-2940-3 RESEARCH Open Access Can the intake of antiparasitic secondary metabolites explain the low prevalence of hemoparasites among wild Psittaciformes? Juan F. Masello1* , Javier Martínez2, Luciano Calderón1, Michael Wink3, Petra Quillfeldt1, Virginia Sanz4, Jörn Theuerkauf5, Luis Ortiz-Catedral6, Igor Berkunsky7, Dianne Brunton6, José A. Díaz-Luque8,9, Mark E. Hauber10, Valeria Ojeda11, Antoine Barnaud12, Laura Casalins11, Bethany Jackson13,14, Alfredo Mijares15, Romel Rosales15, Gláucia Seixas16, Patricia Serafini17, Adriana Silva-Iturriza15, Elenise Sipinski18, Rodrigo A. Vásquez19, Peter Widmann20, Indira Widmann20 and Santiago Merino21 Abstract Background: Parasites can exert selection pressure ontheirhosts through effects onsurvival, on reproductive success, on sexually selected ornament, withimportantecological and evolutionary consequences, such as changes inpopulation viability. Consequently, hemoparasites have become the focus of recent avian studies. Infection varies significantly among taxa. Various factors might explain the differences ininfection among taxa, including habitat, climate, host density, thepresence of vectors, life historyand immune defence. Feedingbehaviour can also be relevant both through increased exposure to vectors and consumption of secondary metabolites withpreventative or therapeutic effects that can reduce parasite load. However, thelatter has been little investigated.Psittaciformes (parrots and cockatoos) are a goodmodel to investigate these topics, as they are known to use biological control againstectoparasitesandtofeedontoxicfood.Weinvestigatedthepresenceofavianmalariaparasites(Plasmodium), intracellularhaemosporidians(Haemoproteus,Leucocytozoon),unicellularflagellateprotozoans(Trypanosoma)and microfilariaein19PsittaciformesspeciesfromarangeofhabitatsintheIndo-Malayan,AustralasianandNeotropical regions.WegatheredadditionaldataonhemoparasitesinwildPsittaciformesfromtheliterature.Weconsidered factorsthatmaycontrolthepresenceofhemoparasitesinthePsittaciformes,compilinginformationondiet,habitat, andclimate.Furthermore,weinvestigatedtheroleofdietinprovidingantiparasiticsecondarymetabolitesthatcould beusedasself-medicationtoreduceparasiteload. Results:Wefoundhemoparasitesinonlytwoof19speciessampled.Amongthem,allspeciesthatconsumeatleast onefooditemknownforitssecondarymetaboliteswithantimalarial,trypanocidalorgeneralantiparasiticproperties, werefreefromhemoparasites.Incontrast,theinfectedparrotsdonotconsumefooditemswithantimalarialoreven generalantiparasiticproperties.Wefoundthatthetwoinfectedspeciesinthisstudyconsumedomnivorousdiets. Whenwecombinedourdatawithdatafromstudiespreviouslyinvestigatingbloodparasitesinwildparrots,the positiverelationshipbetweenomnivorousdietsandhemoparasiteinfestationwasconfirmed.Individualsfromopen habitatswerelessinfectedthanthosefromforests. (Continuedonnextpage) *Correspondence:[email protected] 1DepartmentofAnimalEcologyandSystematics,Justus-LiebigUniversität Gießen,Heinrich-Buff-Ring26,D-35392Gießen,Germany Fulllistofauthorinformationisavailableattheendofthearticle ©TheAuthor(s).2018OpenAccessThisarticleisdistributedunderthetermsoftheCreativeCommonsAttribution4.0 InternationalLicense(http://creativecommons.org/licenses/by/4.0/),whichpermitsunrestricteduse,distribution,and reproductioninanymedium,providedyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinkto theCreativeCommonslicense,andindicateifchangesweremade.TheCreativeCommonsPublicDomainDedicationwaiver (http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle,unlessotherwisestated. Maselloetal.Parasites&Vectors (2018) 11:357 Page2of15 (Continuedfrompreviouspage) Conclusions:Theconsumptionoffooditemsknownfortheirsecondarymetaboliteswithantimalarial,trypanocidalor generalantiparasiticproperties,aswellasthehigherproportionofinfectedspeciesamongomnivorousparrots,could explainthelowprevalenceofhemoparasitesreportedinmanyvertebrates. Keywords:Antiparasiticmetabolites,Bloodparasites,Cacatuidae,Haemoparasites,Herbivorous,Omnivorous,Plant secondarymetabolites,Psittacidae,Self-medication Background descriptions see [24]), and by eliciting chronic infections Parasites can exert ecological and evolutionary pressures inwildbirds[1–3,24].Yet,arelapseoftheparasiteinfec- on their hosts [1–4]. This pressure can affect the host in tion usually takes place during the breeding season of the various ways, including decreased body condition, avianhost, facilitatingthe infection ofthe vectorsandthe reproductive success, and survival, as well as physio- transfer of infection to the offspring [24]. Delayed logical castration [5–8]. Ultimately, parasites affect the reproduction, reduced clutch sizes, reduced parental fitness of the host by promoting the evolution of behav- working capacity while feeding nestlings, and in- ioural, physiological or immunological anti-parasite creased predation risk, are some of the reported ef- defences [1, 9–11]. The co-evolutionary dynamics of host fects of blood parasites on the host, particularly defence mechanisms and parasite counter-adaptations are during situations of stress that deteriorate the condi- influencedbymanyfactors,includingenvironmentalcon- tion of an individual [1, 2, 7, 10, 26]. Although some ditions, genetic background, immune defence investment, blood parasites have been classified as non-pathogenic, lifehistorytraits,behaviour,hostageandsex[12–16]. they certainly remove resources from the host that could Due to their ecological and evolutionary importance, be otherwise be used for growth, maintenance or avian malaria parasites (Plasmodium) and related hae- reproduction, reducing thus reproductive success and ul- mosporidians (Haemoproteus and Leucocytozoon), the timatelyfitness[1]. unicellular parasitic flagellate protozoa Trypanosoma, The degree of blood parasite infection reported varies and the early stage in the life-cycle of some parasitic greatly among birds, including taxa with high prevalence nematodesknownasmicrofilaria,havebecomethefocus (e.g. songbirds), but also taxa with low prevalence (e.g. of a number of avian studies and has resulted in the some seabirds) or even an absence of parasites (e.g. establishment of an open access database dedicated to storks, waders, nightjars, some seabirds, sandgrouse, haemosporidians (MalAvi) [3, 16–23]. Avian haemospo- parrots, swifts) [24, 30–44]. Various factors may explain ridians are single-celled, intracellular parasites in which this disparity. Songbirds have been more often sampled the fertilisation, formation of zygotes, and asexual spor- than some other bird groups that have not been re- ogony take place in a blood-sucking dipteran vector cordedharbouringbloodparasites[19,24–44].Secondly, whilesexualgametogonyandasexualmerogonyoccur in microscopy may underestimate the prevalence in cases the avian host [24]. Plasmodium haemosporidians are of very low infection, poor quality of the blood smears, usually transmitted by dipteran vectors belonging to the or if the observers are not properly trained (e.g. [40, 41], Culicidae, while Haemoproteus are transmitted by and references therein). Various abiotic and biotic fac- dipterans of the Ceratopogonidae and Hippoboscidae, tors might also explain the large taxonomic variability of and Leucocytozoon through species belonging to the infection prevalence e.g. nesting habitat, nest character- Simuliidae [24]. Avian Trypanosoma are transmitted by istics, host density, temperature, elevation, topography, biting midges (Ceratopogonidae) [25], while in the case wateravailabilityandvectorabundance[41,45–47].Sev- of microfilariae the common intermediate hosts are eral of these factors have been shown to influence the blood-sucking insects of the Simuliidae, Ceratopogo- hemoparasite-host interactions by affecting parasite nidae, Tabanidae and Culicidae [26]. Although all these prevalence in the insect vectors and the probability of blood parasites are widely distributed, they are restricted transmission to the avian host [20, 47–49]. Birds from to the distribution of both their avian hosts and their certain habitats, such as tundra, arid, island and/or mar- vectors [24, 27, 28]. Plasmodium appears to be more ine environments, have been reported as having a lower cosmopolitan than the other related haemosporidians. prevalence than birds from other habitats [19, 32, 43, Haemoproteus appears to be absent from some oceanic 50–53]. Another possible reason for the variation in regions, while Leucocytozoon appears to be less abundant abundance of blood parasites includes the variable pres- inthe Neotropical and Australian regions [19, 24, 28, 29]. ence and density ofthe vectors [50, 54, 55],and the abil- Blood parasites are distinguished by having at least one ity by the host immune system to resist or control developmental stage in the hostbloodstream (for detailed infection [24]. Moreover, a blood parasite will reach its Maselloetal.Parasites&Vectors (2018) 11:357 Page3of15 host only during a specific life-stage when the avianhost domed nests within cavities [75, 76]. Most species tend has to be susceptible, the necessary vectors need to be to be gregarious forming loose to very dense colonies present and competent, and the environmental condi- [77–79]. The wide range of habitats used, the gregarious tions need to be appropriate for the transmission [47]. behaviour and the nesting characteristics of the Food availability appears to play a key role by affecting Psittaciformes could favour contact with the vectors and the condition of the birds, which in turn can affect their hence the transmission of parasites. In fact, some hae- susceptibility to parasites [56–58]. Feeding behaviour mosporidians like Haemoproteus (Parahaemoproteus) can likewise play an important role [12, 20]. A handai, H. (P.) psittaci, and H. (P.) homohandai have wide-ranging study on wild Neotropical birds found that been originally described from captive Psittaciformes species with omnivorous diets had a higher prevalence [24, 80, 81]. Moreover, blood parasites appear to be ofPlasmodium,whereasinsectivoreshad ahigherpreva- common among captive parrots particularly in zoos lence of microfilariae [20]. Moreover, some animals con- ([24, 80–82], most records in [83–89]). However, sume food containing secondary metabolites with when considering only wild Psittaciformes, 66% of preventative or therapeutic effects that can be used as parrotpopulationsstudiedsofarreportedanapparentab- self-medication to reduce parasite load, against microbes sence of bloodparasites(44 of 67populations; Additional or that can even serve as antioxidants [59–63]. These file 1: Table S1, and references therein). A plausible ex- secondary metabolites often interact with proteins, bio- planation for this difference is that the stress associated membranes or nucleic acids of the parasites, disrupting withcaptivitymayincreasetheimmunologicalsusceptibil- their bioactivities and thus acting as effective antipara- ityofindividualsorreducetheircapacitytoavoidthevec- sitic medication [63]. Non-human primates, ruminants, tors commonly present in zoos, thus increasing parasite wolves (Canis lupus), cougars (Puma concolor) and load [78–92]. Another explanation is that the absence of domestic dogs (C. l. familiaris) ingest plants with anti- hemoparasitescouldberelatedtoastronginnateimmun- parasitic properties but with little or no nutritional value ity in some Psittaciformes [42, 93]. Also, the monk [59–61, 64–66], wood ants (Formica paralugubris) use parakeet and the red-fronted parakeet (Cyanoramphus resin to inhibit the growth ofmicroorganisms [67], some novaezelandiae) have been shown to bring fresh green passerines use lime rind against lice [68] and fresh plant leavestothenest,abehaviourthathasbeeninterpretedas material to repel parasites or mask the chemical cues a way to actively deter ectoparasites [94, 95] acting as that parasites use to find the host [69], while great bus- blood parasite vectors. Many parrots in the wild feed on tards (Otis tarda) have been shown to consume blister toxicfruits,seedsorflowerbuds[96–98],whereasincap- beetles (Meloidae) that contain secondary metabolites tivity they obtain food that does not contain toxins [99]. with antimicrobial and pathogen-limiting activity [62]. Thus, it could also be possible that parrots use secondary Consequently, there is not a unique explanation for metabolites present in the diet as self-medication to re- the complex variety of hemoparasite-host interactions, duce parasite load [61–69, 95]. Alternatively, differences and comparative multivariate analyses suggest that in the diet, habitats or environmental factors like climate phylogenetic, ecological, behavioural, climatic, and could also determine the large interspecific variation in life-history traits determine the large variation ob- blood parasite prevalence reported for Psittaciformes served in hemoparasite prevalence [3, 7, 12, 16, 19, (Additional file 1: Table S1 and references therein). Until 20, 43, 47, 48, 70–72]. Lastly, it is important men- now,noneofthesepotentialexplanations hasbeeninves- tioning that a reduction or the absence of parasite in- tigated in detail and information on blood parasite infec- fections might influence the evolutionary trajectories of tion in wild Psittaciformes remains patchy, including birdspecies.Identifyingtheunderlyingdriversofvariation mostly occasional data collected during general surveys inpathogenprevalencehasimportantramificationsinthe (Additional file 1: Table S1 and references therein), and fieldsofevolutionaryecologyanddiseaseecology.Accord- fewindividualspeciesinvestigatedindetail[100–102]. ingly,theavoidanceofinfectionsmaypositivelyaffecthost Given the wide range of habitats, climates and diets traits, such as reproduction and survival, allowing species that Psittaciformes exploit, as well as their previously re- not subject to pathogen pressure to locally outcompete ported antiparasitic behaviour, this group of birds has a other species or to become successful invaders in intro- great potential as a model to investigate environmental ducedcommunities[73,74]. and behavioural factors. This includes diet selection and ThePsittaciformes (parrots and cockatoos) avianorder self-medication, which may determine the interspecific is distributed from the tropics to sub-Antarctic regions, variation in blood parasite prevalence in vertebrates. We in a wide range of habitats extending from tundra to therefore studied the presence of hemoparasites of the rainforest [75]. Psittaciformes are mostly cavity nesters, genera Haemoproteus, Plasmodium, Leucocytozoon and with only monk parakeets (Myiopsitta monachus) build- Trypanosoma, as well as microfilariae across populations ing twig nests and some Agapornis species building of wild Psittaciformes. We sampled 19 Psittaciformes Maselloetal.Parasites&Vectors (2018) 11:357 Page4of15 species from 25 localities covering an extensive range of Maidstone, UK), in lysis buffer (100 mM Tris pH 8, 10 habitats and climate types in the Indo-Malayan, Austra- mM NaCl, 100 mM EDTA, 2% SDS), or in ethanol 70% lasian and Neotropical zoogeographical regions. We (see details per species in Table 1). Liver, kidney and considered extrinsic and intrinsic factors that may con- heart samples were sampled under sterile conditions trol the presence of hemoparasites in Psittaciformes, (tools andcabinets)andpreservedin70%ethanol. compilinginformation onfood itemsconsumed, habitats We only included samples obtained from wild Psittaci- used and climate. We also investigated the potential role formes. We excluded samples of parrots and cockatoos of the food consumed in providing antiparasitic second- that could have been in contact with (i) poultry; (ii) pets; ary metabolites that could be used as self-medication to (iii) zoos; (iv) aviculture facilities; (v) wildlife hospitals; reduce parasite load. Furthermore, we searched the lit- (vi) rehabilitation facilities; (vii) reintroduction program erature for wild parrots previously investigated for facilities; or (viii) the staff of any of the previously men- hemoparasites, and additionally compiled information tioned facilities. One exception were the parakeets sam- on the corresponding diets, habitats and climates. We pled at Tiritiri Matangi Island (New Zealand), which hypothesised that parrots consuming antiparasitic sec- were descendants from birds translocated from captivity ondary metabolites have a lower prevalence of to the island c.30 years ago [111]. Co-workers who are hemoparasites. veterinarypractitioners and also work with captivebirds, strictly refrained from contact with the above-mentioned Methods facilities in the weeks before sampling. We considered Ownsamples these precautions mandatory, since parasites are com- Between 1999 and 2014, we obtained blood (n = 329), mon in captive birds, probably as an effect of captivity liver (n = 23), heart (n = 1) and kidney (n = 1) samples [86, 89–91]. We took special precautions to avoid both belonging to 19 Psittaciformes species from 25 localities. the contamination of samples and spreading of disease Maps showing the position of all localities sampled in into wildpopulations duringsampling. this study are provided in Additional file 2: Figures Genomic DNA from samples stored in FTA cards was S1–S5. Adults (n = 213), nestlings (n = 112) and juve- extracted according to Martínez et al. [112]. The DNA niles (n = 4) were captured in their nests. The age of the solution was purified using the commercial kit NZYGel- adult parrots sampled was unknown. We obtained sam- pure (NZYTech, Lisbon, Portugal). By means of PCR, ples from nestlings during the pre-fledging (pf) period; we amplified a part of the cytochrome b gene or the 18S this means the nesting period shortly before the young ribosomal RNA gene using previously published primers leave the nest and long after the minimal prepatent [113]. Sequences of the primers, size of the amplicons, period reported for several blood parasites [103–105]. andPCRconditions areshown inTable 2.PCRreactions The prepatent period for blood parasites i.e. the period consisted of 10 μl reaction volumes containing between between infection and presence of infective forms in 20 and 100 ng of template DNA, 0.25 μM of each pri- blood, varies between 5 and 14 days [24, 103]. This mer and SYBR® Select Master Mix (Applied Biosystems, might be the reason for the usually low blood parasite Foster City, CA, USA). The reactions were cycled using counts in nestlings [106, 107]. Nevertheless, blood para- a StepOnePlus Real-Time PCR System (Applied Biosys- sites were found in 20% of 13-day-old nestlings of the tems). The diagnosis was performed by visualizing the pied flycatcher (Ficedula hypoleuca) [104] and in 67% of melting curve of the amplicons. After screening, positive 15-day old blue tits (Cyanistes caeruleus) [108]. Juveniles samples were amplified again to obtain larger amplicons were sampled during their first year of life (i.e. that facilitate the identification of haplotypes [113]. PCR post-fledging).Forthisreason,theage categoriesconsid- reactions contained between 20 and 100 ng of the DNA ered are ‘pre-fledging’, ‘juvenile’ and ‘adult’. Liver, heart template, Supreme NZYTaq 2× Green Master Mix and kidney samples were obtained from birds found (NZYTech)and250nMofeachprimer (Palu F/Palu R). dead in the vicinity of parrot nests, colonies or roosting Using aVeriti thermal cycler (Applied Biosystems), reac- places. For ethical reasons, we did not accept samples tions were run using the following conditions: 95 °C for from huntedorlethallysampled birds forthis study. 10 min (polymerase activation), 40 cycles at 95 °C for 30 We collected blood samples via puncture of the cuta- s, annealing at 56 °C for 30 s, 72 °C for 30 s, and a final neous ulnar vein immediately after capture. In order to extension at 72 °C for 5 min. All amplicons were recov- minimize stress, we sampled individuals only once and ered from agarose gels and subjected to direct sequen- kept handling time to a minimum, usually below 5–10 cing using an ABI 3730 XL automated sequencer min, before returning them to their nests. Like in previ- (Applied Biosystems). DNA extraction and PCR set up ous studies, blood sampling had no detectable adverse were always performed in different laminar flow cabi- effects [42, 93, 109, 110]. We stored blood samples on nets. We never detected amplicons in negative controls FTA classic cards (Whatman International Ltd., added in each PCR batch. A positive control for each Maselloetal.Parasites&Vectors (2018) 11:357 Page5of15 erences 0] 1,142] –3145] 3,146] 1] 2] 2] 2] 2] 2] 7] 7] 7] 7] Ref [14 [14 [14 [14 [11 [10 [14 [14 [14 [14 [14 [14 [14 [14 AntiparasiticeSMindiet? Yes No No No No No No Yes Yes alregions Hemoparasitepresenceandprevalence − − − − − − − Genus:Plasmodium;lineage:LIN4(BELL02);identity100%;prevalence18% − Genus:Plasmodium;lineage:LIN4(BELL02);identity100%;prevalence5%;lineage:CN73;identity99.7%;prevalence5% − − − − − − − − − − c pi o dNeotr dTissue FTA FTA FTA FTA FTA FTA FTA lys lys lys FTA heart kidney FTA FTA FTA FTA lys FTA FTA n a n a ce alasi Ag pf ad pf ad ad ad ad ad pf ad pf ad pf ad juv ad pf ad ad pf str u A 6 0 4 4 2 n 1 2 1 5 2 2 3 3 2 4 1 1 1 6 1 2 1 1 5 2 n, a y a do-Mal bClimate Am Af Af Af Af fCfb gCfa Cfb Cfb Af Af Af Af Am Am Am Am n I e h t atesof abitat MF Sa,PP Gr,NTP Sa,PP Sa,PP Ma CP S S S S clim Ha M, Sa Rf, Su Su CS EP BF, CK Rf, Rf, Rf, Rf, PFI PFI PFI PFI d n a 8) esfromdifferenthabitats Locality(localitynumber) RasaI.,Palawan,Philippines(1) Ouégoa,NewCaledonia(2) ParcdesGrandesFougères,NewCaledonia(3) Motorpool,Nouméa,NewCaledonia(4) ValléedesColons,Nouméa,NewCaledonia(5) MangereI.,ChathamIs.,NewZealand(6) RaoulI.,NewZealand(7) TiritiriMatangiI.,NewZealand( LittleBarrierI.,NewZealand(9) ParcdesGrandesFougères,NewCaledonia(3) ParcdesGrandesFougères,NewCaledonia(3) ParcProvincialdelaRivièreBleue,NewCaledonia(10) Gossana,Ouvéa,NewCaledonia(11) Trinidad,Bolivia(12) Sachojere,Bolivia(13) Beni,Bolivia(14) Trinidad,Bolivia(12) m or cif a Table1Hemoparasitesin19Psitt Cacatuidae Indo-Malayan PhilippinecockatooCacatuahaematuropygia Psittacidae Australasian NewCaledonianrainbowlorikeetTrichoglossushaematodusdeplanchii ’ForbesparakeetCyanoramphusforbesi Red-frontedparakeetCyanoramphusnovaezelandiae NewCaledonianparakeetCyanoramphussaisseti HornedparakeetEunymphicuscornutus OuvéaparakeetEunymphicusuvaeensis Neotropical BlueandyellowmacawAraararauna Blue-throatedmacawAraglaucogularis Maselloetal.Parasites&Vectors (2018) 11:357 Page6of15 AntiparasiticReferenceseSMindiet? Yes[148] No[147] Yes[149] Yes[150] Yes[151] [152] No[46] [153] Yes[147] Yes[147] Yes[154] [154] Yes[155] [148] [150] CTScactusandthornscrub,DXFmaquis,MFmonsoonforest,MoPSPatagoniansteppes,Rf utdryseason,hotsummer;Cfb, Table1Hemoparasitesin19PsittaciformesfromdifferenthabitatsandclimatesoftheIndo-Malayan,AustralasianandNeotropicalregions(Continued) abcdLocality(localitynumber)HabitatClimatenAgeTissueHemoparasitepresenceandprevalence −Blue-crownedconureThectocercusChaco,Argentina(15)DXFCfa1adlysacuticaudatus−4pflys −White-eyedconurePsittacaraSachojere,Bolivia(13)PFISAm2pfFTAleucophtalmus −f9adFTABrown-throatedconureEupsittulaIslaMargarita,Venezuela(16)CTSAwpertinax −NandayconureAratinganendayPrincipeNegro,Pantanal,Gr,Sa,SS,RiAw11pfFTABrazil(17) −BurrowingparrotCyanoliseuspatagonusElCóndor,Patagonia,MoBSk32adethArgentina(18)−14adliver −Comallo,Patagonia,PSCsb5adliverArgentina(19) AustralparakeetEnicognathusNavarino,Chile(20)BFNET2adFTAGenus:Leucocytozoon;ferrugineusprevalence100% Bariloche,Patagonia,BFNCsb3adFTAGenus:Leucocytozoon;Argentina(21)prevalence100% −1pfFTA 3adliverGenus:Leucocytozoon;prevalence33% −1pfliver −Blue-wingedparrotletForpusxanthopterygiusTrinidad,Bolivia(12)PFISAm9adFTA −Yellow-chevronedparakeetBrotogerisTrinidad,Bolivia(12)PFISAm4adFTAchiriri −Red-tailedAmazonAmazonabrasilensisIlhaRasa,Guaraqueçaba,LAF,MCfa29pfFTABrazil(22) −IlhadasGamelas,LAF,MCfa4pfFTAGuaraqueçaba,Brazil(23) −Blue-frontedAmazonAmazonaaestivaJujuy,Argentina(24)DXFCwa6adlys −Chaco,Argentina(15)DXFCfa13adlys −Brasil(25)Gr,Sa,SS,RiAw17pfFTAPantanal, aHabitats:BFremnantsofbroadleafforest,BFNbroadleafforestsdominatedbyNothofagusspp.,CKcoastalandkauri(Agathisaustralis)forests,CScoastalscrub,CPcoconutplantations,deciduousxerophyteforest,EPendemicpohutukawa(Metrosideroskermadecensis)sub-tropicalmoistforest,Grgrasslandwithsparsetrees,LAFlowlandAtlanticforest,Mmangrove,MaMontexerophyteforest,NTPnativetreesplantedunderare-vegetationprogramme,PFISpalmdominatedforestislandssurroundedbyregularlyfloodedsavannah,PPpineplantations,rainforest,Ririparianforest,Sasavannah,SSscrubsavannahs,Susub-urbanbThediversityofclimatesfollowing[143,156],exceptwhereindicated:Af,tropicalrainforest;Am,tropicalmonsoon;Aw,tropicalsavannah;BSk,arid,steppe,cold;Cfa,temperate,withotemperate,withoutdryseason,warmsummer;Csb,temperate,drysummer,warmsummer;Cwa,temperate,drywinter,hotsummer;ET,polar,tundracAge:ageofindividualsatthetimeofsampling;ad,adult;pf,pre-fledgling;juv,juveniledTissue:tissueusedtoobtainDNA;FTA,bloodinFTAcards;lys,bloodinlysisbuffer;eth,bloodinethanol70%eFooditemsknownfortheirsecondarymetabolites(SM)withantimalarial,trypanocidalorgeneralantiparasiticpropertiesffollowing[143]andWorldClimdatabase(http://www.worldclim.org)[145],using-(http://www.diva-gis.org/)DIVAGISgfollowing[143]anddatafromtheNewZealandNationalClimateDatabase(http://cliflo.niwa.co.nz)Abbreviation:n,samplesize Maselloetal.Parasites&Vectors (2018) 11:357 Page7of15 Table2PrimersusedforPCRdetectionofhemoparasitesinwildNeotropical,Indo-MalayanandAustralasianPsittaciformes Gene Primername Primersequence5’→3’ Size(bp) Annealing Parasite Cytochromeb Palu-Fq CAAGGTAGCTCTAATCCTTTAGG 201 54°C/30s Haemoproteus;Plasmodium Palu-R DGGAACAATATGTARAGGAGT Cytochromeb L180 GAGAACTATGGAGTGGATGG 221 60°C/30s Leucocytozoon Leunew1-R CCCAGAAACTCATTTGWCC 18SrRNA Try-F GGAGAGGGAGCCTGAGAAATA 121 60°C/30s Trypanosoma Try-R ATGCACTAGGCACCGTCG 18SrRNA NF110 GCTAATACATGCACCAAAGCTCC 119 60°C/30s microfilaria NR228 CAAGACCATGCGATCAGC pair of primers was routinely used. All analyses were using the five-letter species code to indicate the bird carried out at the Departamento de Biomedicina y Bio- host, and submitted the results/sequences to GenBank tecnología, Área de Parasitología, Facultad de Farmacia, andMalAvidatabases. Universidad de Alcalá, Alcalá de Henares, Spain, with The main food items consumed by the 19 Psittaci- the exception of the brown-throated conure (Eupsittula formes species that we sampled in this study is provided pertinax) samples. As we were not able to export these in Additional file 3: Table S2. Most of this information samples from Venezuela, some co-authors (AM, RR and was gathered by the co-authors during ongoing research VS) analysed the 9 samples of the brown-throated projects on the species considered, while some of the in- conure at the Instituto Venezolano de Investigaciones formation originates from recent literature cited in Científicas. In this case, Haemoproteus, Plasmodium and Additionalfile 3:Table S2.Informationonthesecondary Leucocytozoonparasiteswerealso screened bythe nested metabolites contained in the items consumed by the PCR method but the partial amplification of the cyto- sampled parrots was obtained from the references also chrome b (here 471 bp) gene was carried out using dif- provided in Additional file 3: Table S2. The secondary ferent primers. For the first PCR round, we used the metaboliteswere classifiedaccordingtotheirmainactiv- primer pair Haem NFI (5'-CAT ATA TTA AGA GAA ity in (i) antimalarial, trypanocidal or general antipara- ITA TGG AG-3') and Haem NR3 (5'-ATA GAA AGA sitic properties; (ii) anthelmintic; (iii) antimicrobial; and TAA GAA ATA CCATTC-3') [114]. We then used 2 μl (iv) antioxidant (Additional file 3: Table S2). Detailed of the first PCR reaction mixture as the template for a information on the habitats where we obtained our second-round PCR, in which Haem R2 (5'-GCA TTA samples is provided in Table 1. Information on the cli- TCT GGA TGT GAT AAT GGI-3') was paired with mate associated with these habitats was obtained from Haem F (5'-ATG GTG CTT TCG ATA TAT GCA thereferencesprovidedinTable1. TG-3'), and Haem R2L (5'-CAT TAT CTG GAT GAG ATA ATG GIG C-3') with Haem FL (5'-ATG GTG TTT Additionaldatafromtheliterature TAG ATA CTT ACA TT-3') [114]. In order to detect We also searched the literature for wild parrots previ- possible positive samples that were not detected by the ously investigated for hemoparasites, and additionally firsts primersset, we amplified an additional cytochrome collected information on the corresponding diets, habi- b gene fragment also using a nested PCR [115]. The tats and climates. For the literature search, we used the outer reaction was carried out with the primers DW2 library of the Working Group Psittaciformes of the (5'-TAA TGC CTA GAC GTA TTC CTG ATT ATC International Ornithologists’ Union. At the time of the CAG-3') and DW4 (5'-TGT TTG CTT GGG AGC TGT literature search, this comprehensive library, which is AAT CAT AAT GTG-3’) [116]. We used a 2 μl aliquot updated monthly, comprised > 3600 papers (from 1817 of this product as a template for a nested reaction with to present). The literature in the library was reviewed primers DW1 (5’-TCA ACA ATG ACT TTA TTT manually by JFM in search for any previously published GG-3') and DW6 (5'-GGG AGC TGT AAT CAT AAT paper containing information on blood parasites affect- GTG-3') [116]. Additionally, we carried out a nested ing wild parrots. The search terms used to assist this PCR using 2 μl aliquot of the first reaction with the search arelistedinAdditionalfile1:Table S1. primers DW1 and HaemR (5'-CAT ATC CTA AAG We found 24studiesandreviewspreviouslyinvestigat- GAT TAGAGC TACCTTGTAA-3'). ing blood parasites in wild parrots belonging to 52 spe- We identified hemoparasite lineages using the Basic cies (covering 67 different populations; Additional file 1: Local Alignment SearchTool (BLAST) in GenBank and Table S1). For these parrot species, we summarized in the MalAvi databases [17]. We named new lineages Additional file 1: Table S1 the following information: (i) Maselloetal.Parasites&Vectors (2018) 11:357 Page8of15 number ofindividuals investigated(adultsandnestlings); omnivorous, and the screening methods as PCR-based (ii) number of individuals infected, including the total or not (Additional file 4). We assessed the combined number, and the subtotals discriminated in Haemopro- dataset (N = 520) for the effects of diet, habitat, cli- teus, Plasmodium, Leucocytozoon, Trypanosoma and mate and screening method (as factors) on the pres- microfilaria; (iii) diets; (iv) habitats used; (v) climates; ence of parasites in the studied individuals using (vi) screening methods used for parasite detection and Generalized Linear Mixed-Effects Models with bino- (vii) pertinent references. Yet, detailed information on mial error distribution and model selection based on specific food items consumed by a number of those par- Akaike information criterion (AIC) in the R program- rot species is scarce (i.e. include a few plant items men- ming language (script and full dataset provided in tioned in the literature or vague observations like in e.g. Additional file 4; packages MuMIn and lme4) [118]. Aratinga euops ‘they feed on fruits, seed, berries nuts To account for inherent variation among species, we andprobably blossomsandleaf buds’,[75])orevencom- also included the species to which each individual pletely unknown (e.g. Pionopsitta pulchra, [75, 117]). subject belonged as a random intercept in the Gener- For this reason, following González et al. [20], we classi- alized Linear Mixed-Effects Models (script provided fied the diets as (i) herbivorous; (ii) omnivorous-carrion, in Additional file 4). In the dredge function in for species including carrion in their diets; (iii) MuMIn, all possible candidate models (i.e. subsets of omnivorous-insectivorous, for species including insects the global model) were tested using each unique lin- in their diets; (iv) ‘herbivorous?’-scarce information; or ear combination of covariates. The best models are (v) unknown (Additional file 1: Table S1). The shortage then selected based on ΔAIC scores less than or of information on specific food items consumed by a equal to two. In order to evaluate the predictive number of these parrot species prevented us from being power of our diet models and its balance between able to investigate the secondary metabolites that could sensitivity and specificity, we ran a 10-fold cross be present in their diets (Additional file 1: Table S1), as validation, fitting the model to training sets and pre- we did for the parrot species sampled in this study dicting for with-held test sets (see script provided in (Additionalfile3:TableS2). Additional file 4). The 10-fold cross validation pro- vides a mean area under the receiver operating Comparativeanalyses characteristic curve (mean AUC). Odds ratios were To facilitate comparisons, we included in Additional file calculated to provide a measure of how the probabil- 1: Table S1 the data corresponding to the 19 Psittaci- ity of infection is predicted to change, for instance formes speciesfrom the 25localities thatweinvestigated when a species has an omnivorous diet compared to (marked ‘this study’ in the References column) and pro- a species without an omnivorous diet (for calculation duced a combined dataset. Then, we transformed the in- see Additional file 4). Additionally, to allow an ad- fection information of the combined dataset into equate evaluation of our results, we simulated the presence/absence data for each individual investigated probability that the parasites will actually be detected (see combined dataset provided in Additional file 4), as given the sample size and an expected true prevalence 37% of previously published data on hemoparasites in of 0.08213 based on prevalence data previously re- wild parrots lacked information on prevalence. We ex- ported in wild Psittaciformes (Additional file 1: Table cluded studies from statistical testing for which data S1). The simulation script is provided in Additional were only partially available or uncertain (i.e. marked file 4 and the probabilities of detection for all species ‘NA’, ‘herbivorous?-scarce information’ or ‘unknown’ in and sample sizes are provided in Additional file 1: Additional file 1: Table S1). We also excluded studies Table S1. from statistical testing for which only data on nestlings existed. This was done to avoid cases in which infection Results could have been not detected because the nestlings were Hemoparasites were present in adult birds of only two too young at the time of sampling thus not giving the of the 19 species sampled for this study (Table 1). In the parasites time to develop. Although we were certain that red-fronted parakeet, hemoparasites were detected in this was not the case for the nestlings sampled in our samples from Raoul Island and Little Barrier Island, but study, and in order to be conservative, we likewise ex- not on Tiritiri Matangi Island (New Zealand; Additional cluded our data on nestlings from statistical testing. file 2: Figure S4; Table 1). The red-fronted parakeets Giventhelargenumberof habitatandclimate categories from Raoul Island were infected with hemoparasites in the combined dataset, we transformed the data into fromthegenusPlasmodiumcorrespondingtothehaplo- binomial categories to allow adequate statistical compar- type LIN4 (BELL02, new lineage name in MalAvi isons.Habitatswere classifiedasforestornon-forest,cli- database;identity100%;prevalence18%;GenBankacces- mates as tropical or non-tropical, diets as herbivorous or sion number MH238461). The birds from Little Barrier Maselloetal.Parasites&Vectors (2018) 11:357 Page9of15 Island were infected with two Plasmodium haplotypes, cockatoo (Cacatua haematuropygia), blue and yellow one that corresponded to LIN4 [102] (identity 100%, macaw(Araararauna),blue-throatedmacaw(Araglauco- prevalence 5%), and another one which differed in a sin- gularis), blue-crowned conure (Thectocercus acuticauda- gle nucleotide, considering a 312 bp length sequence, tus),brown-throatedconure(Eupsittulapertinax),nanday from the haplotype GRW06 [119] (thereafter haplotype conure (Aratinga nenday), burrowing parrot (Cyanoliseus CN73, GenBank accession number MH238460, identity patagonus), blue-winged parrotlet (Forpus xanthoptery- 99.7%,prevalence5%). gius), yellow-chevroned parakeet (Brotogeris chiriri), All adult austral parakeets (Enicognathus ferrugineus), red-tailed Amazon (Amazona brasilensis), and blue- both from Navarino Island and Bariloche, Patagonia fronted Amazon (Amazona aestiva) (Table 1, and (separated by more than 1500 km; populations 20 and Additionalfile3: TableS2).Noneofthespeciesfound in- 21 in Additional file 2: Figure S5), were infected with fected (red-fronted parakeets, austral parakeets) con- hemoparasites from the genus Leucocytozoon belonging sumed any food item with antimalarial or general to an un-described haplotype (Merino et al. description antiparasitic properties (Table 1, and Additional file 3: in prep.). This new Leucocytozoon haplotype differs 3 to Table S2). However, red-fronted parakeets consumed a 4% in its sequence with respect to several haplotypes food item known for the presence of secondary previously described in birds from the genus Atrix (e.g. metabolites with anthelmintic activity (Additional file 3: GHOW93-00-55, NSPOWORRO15, BAOW5909, SPO Table S2). In addition, all 19 studied parrot species con- W7; Merino et al. unpublished data). Compared to the sumed items that contained secondary metabolites known blood samples, only a third of the austral parakeet liver for their antimicrobial activity, and 17 species (89%) in- samples contained the same Leucocytozoon haplotype volved items with antioxidant properties (Additional file 3: (Table1). TableS2). Toalesserextent,someparrotsconsumedfood All other 17 species of Psittaciformes sampled in this itemscontainingsecondarymetabolitesrecognizedfortheir study, covering a wide range of habitats and climates in e.g.antiinflammatory,anticarcinogenic,analgesic,expector- the Indo-Malayan, Australasian and Neotropical zoogeo- ant,orantipyreticeffects(Additionalfile3:TableS2). graphical regions, were not infected by any of the tested When the data corresponding to the Psittaciformes hemoparasites (Haemoproteus, Plasmodium, Leucocyto- sampled in this study were combined with the data from zoon, Trypanosoma and microfilariae; Table 1). No 24 studies previously investigating blood parasites in pre-fledging nestling or juvenile was infected, including wild parrots (Additional file 4), we found that 7 of 58 those from the red-fronted parakeet and austral herbivorous wild parrots (12%) were reported to be in- parakeet. fected with Haemoproteus, Plasmodium or microfiliaria. The main food items consumed by the 19 parrot spe- In contrast, 7 of 10 omnivorous wild parrots (70%) were cies that we sampled are provided in Additional file 3: reported to be infected with Leucocytozoon, Haemopro- Table S2. Most parrot diets were herbivorous, except for teus or Plasmodium (Additional file 1: Table S1). Model the two species in which blood parasites were detected: selection revealed that the four best models included the red-fronted parakeet and the austral parakeet which diet, whereas the second best included diet and habitat, also consume items of animal origin (Additional file 3: the third best included diet and screening method, Table S2). In addition to fruits, flowers and unripe seeds, and the fourth best included diet and climate (Additional red-fronted parakeets consume invertebrates and marine file 4). The model-averaged coefficients (7.5, SE = 3.7; molluscs and occasionally scavenged animal carrion, P = 0.04) and the odds ratio (OR = 1882.6) suggested a including birds (omnivorous-carrion diet). Austral strongpositiverelationshipbetweenomnivorousdietsand parakeetsincludeintheir dietlarvaeofHomoptera, Lepi- hemoparasite infestation), but a weaker relationship with doptera and Hymenoptera (Aditrochus fagicolus) (omni- foresthabitat(OR=0.5),non-tropicalclimate(OR=0.8), vorous-insectivorous diet; Additional file 3: Table S2). and non-PCR-based screening method (OR = 0.8). The Thus, in our samples, only omnivorous species were in- 10-fold cross-validation showedfor our diet model a high fectedwithhemoparasites. predictive power and a good balance between sensitivity Fifteen of the 19 parrot species that we sampled (79%) andspecificity(meanAUC=0.9). regularly consume food items that include secondary metabolites known for their antiparasitic activity, in- Discussion cluding antimalarials, antifungals, leishmanicidals, try- Two clear patterns emerged from our results. First, our panocidals, anthelmintics, insecticides and mosquitocidals results appear to support the hypothesis that parrots (Additional file 3: Table S2). Of the 15 species, those that engaging into self-medication are free from hemopara- consume food with antimalarial or general antiparasitic sites. All the studied parrots that regularly consume at properties were free from Haemoproteus, Plasmodium, least one food item known for its secondary metabolites Leucocytozoon,Trypanosoma and microfilariae: Philippine with antimalarial, trypanocidal or general antiparasitic Maselloetal.Parasites&Vectors (2018) 11:357 Page10of15 properties were indeed free from Haemoproteus, Plas- predators and parasites [122]. Also, parrots have been modium, Leucocytozoon, Trypanosoma and microfilariae shown to use plants for prophylactic reasons. Previous (Table 1). Moreover, no food items with antimalarial or studies on monk parakeets and red-fronted parakeets even general antiparasitic properties were consumed by described the use of plants known for their insecti- red-fronted parakeets, which were infected with Plasmo- cidal activity, which protect their nests from parasites dium haplotypes, or by austral parakeets, found infected [94, 95, 123, 124]. Furthermore, red-fronted parakeets with Leucocytozoon. These results suggest that chew leaves of these plants, mix the chopped material self-medication in parrots may be more widespread than withpreenoilandspreadthemixtureontheirfeathersto previously thought. Nevertheless, some species that do repel insects [95]. Thus, our results add further evidence not regularly incorporate antiparasitic metabolites in infavourofprophylacticanti-parasiteself-medicationasa their diets were also free from blood parasites [e.g. New wider than previously thought behaviour in both verte- Caledonian rainbow lorikeet (Trichoglossus haematodus brates and invertebrates, and increase our understanding deplanchii), Forbes’ parakeet (Cyanoramphus forbesi), of why certain food items are taken regardless being poi- New Caledonian parakeet (Cyanoramphus saisseti), sonous or with little nutritional value [61–63, 97–99]. horned parakeet (Eunymphicus cornutus), Ouvéa Even more, anti-parasite self-medicationcan be relatedto parakeet (Eunymphicus uvaeensis) and white-eyed con- humanfoodconsumptionandhealth.Forinstance,thein- ure (Psittacara leucophtalmus)]. In the case of the creaseindiseaseinhoneybees(Apismellifera)asaconse- white-eyed conure, available data on their diet are lim- quence of agricultural practices that interfere with the ited (Additional file 3: Table S2) and thus, we may have ability of bees to self-medicate [61]. Research on animal missed food items that could provide this species with self-medication has the potential to trigger the discovery antiparasitic protection. In the case of the Forbes’ para- ofnewsecondarymetabolitescontainedinthefood items keet from Mangere Island, the lack of necessary vectors consumed by animals. Some of these secondary metabo- previously reported [120] could explain the results. Add- lites may have pharmacological properties and therefore, itionally, small sample sizes could have played a role in couldcontributetohumanhealthcare.[63].However,itis at least three of these species (Table 1). Possible sam- important to mention that to fully test prophylactic pling bias should be considered when interpreting our anti-parasiteself-medicationinparrots,furtherwork,par- findings and should steer future work to assess how ticularly experimental research, should be conducted. It these patterns hold following additional surveys, particu- wouldalsobenecessarytoinvestigatespecieslivinginthe larly among poorly sampled parrot species such as the same or close regions/habitats to those infected in order New Caledonian parakeet, horned parakeet, Ouvéa para- totestiftheyarefreeofparasiteswheningestingantipara- keet and orange-fronted parakeet (Aratinga canicularis). siticsubstances,i.e.thattheyareexposedtothesamevec- Nevertheless, results from previous studies lend support tors but remain uninfected. Alternatively, infection by to our primary conclusions by demonstrating that some blood parasites in these species could be lethal, making animalsdeliberatelyconsumefooditemswithprophylac- thediscoveryofinfectedbirdsunlikely. tic or therapeutic activity against pathogens or parasites The second interesting pattern that we observed is [59–68]. These include a wide range of species such as that the two infected parrot species (red-fronted and wood ants (Formica paralugubris) engaging in social austral parakeets) regularly consume omnivorous diets. prophylaxis, therapeutic self-medication in the woolly This contrasts with the non-infected species, which are bear caterpillar (Grammia incorrupta), or prophylactic all herbivorous (Additional files 1: Table S1 and self-medication in the Ethiopian baboon (Papio anubis Additional file 3: Table S2). Furthermore, when we com- [61]). In birds, great bustards increase their mating suc- bined our data with data from previous studies, the cess and thus fitness by consuming blister beetles strong positive relationship between omnivorous diets (Meloidae), which are known to contain secondary me- and hemoparasite infestation was confirmed. Our results tabolites with antimicrobial and pathogen-limiting activ- are in fully agreement with a previous wide-ranging ity ([62] but see [121]). Some bird species have been study including 246 species of wild Neotropical birds, reported to be toxic, including the European quail which found that species with omnivorous diets had (Coturnix coturnix), several New Guinea species from higher prevalence of Plasmodium, whereas insectivores the genus Pitohui, the North-American ruffed grouse had higher prevalence of microfilariae than birds with a (Bonasa umbellus), the spur-winged goose (Plectropterus different feeding behaviour [20]. More recently, Naqvi et gambensis) from Benin, and some Australian bronzew- al. [125] found that the prevalences of Haemoproteus, ings from the genus Phaps [122]. The toxicity is pro- Plasmodium, and Leucocytozoon in chickens were higher duced by batrachotoxin, cantharidin, andromedotoxin in free-ranging individuals that also feed on carrion. and alkaloids contained in the skin and feathers, which Therefore, the consumption of carrion and its associated have been suggested to protect the birds against scavenging behaviour by red-fronted parakeets or the
Description: