mn 08 February 2022 BIOLOGY LETTERS royalsocietypublishing.org/joural/rsbl Research 8 ® Gite this artic: Kaiser SW, Grenles Ml, Shine R. 2021 Wildes modify the paraste loads of invasive cane toads. Bil Lett. 17: mno4n. tps/doorg/101098/sb 20210470 Received: 31 August 2021 Accepted: 30 November 2021 Subject Areas: clog, heath and disease and epidemialogy Keywords: Bufo marinus, habitat change, bost-paase, invasive species ‘Author for correspondence: chard Shine emai: rick shine@m ed au econ supplementary material i avaiable online at hips. do org/0.6084m9. figshare 5744237 THE ROYAL SOCIETY PUBLISHING CO a Wildfires modify the parasite loads of invasive cane toads Shannon W. Kaiser, Matthew J. Greenlees and Richard Shine Deparment of Bologal Scenes, Macquane Unersty, NSW 2108, saa (© 00 001505530; NG, 000-001 505-30; HS, 0000 G01 7529-5657 ‘The frequency and severity of wildfires are increasing due to anthropogenic ‘modifications to habitats and to climate. Postfire landscapes may advantage invasive species via multiple mechanisms, including changes to host parasite interactions. We surveyed the incidence of endoparasitic lungworms (Riatdias pseudosphaerocephala) in invasive cane toads (Rhinella marina) in near-coastal sites of eastern Australia, a year after extensive fires in this ‘ogion, Both the prevalence of infection and number of worms in infected toads increased with toad body size in unburned areas. By contrast, parasite Toad decreased with toad body size in burned areas. By killing moisture- dependent free-living lungwworm larvae, the intense fires may have liberated adult cane toads from a parasite that can substantially reduce the viability of its host, Smaller toads, which are restricted to moist environments, did not receive this benefit from fires. 1. Introduction Over recent decades, changes to temperature and rainfall patteras are causing more extrome events sich as bushfire, flood and drought {1,2}. Those phenom. ena can disrupt natural systems in many ways, some of them obvious (such as, direct mortality and habitat destruction) but others less apparent (such as the disruption of important interspecific interactions). For example, the open land- scapes produced by fire can facilitate predation upon individuals of prey species that survived the fie [3], and herbivores may migrate to post-burn areas to exploit the newly available nutrient-rich vegetation [4]. Many invasive species are major beneficiaries of the habitat changes, able to thrive in disturbed habitats (5. (One important but under-studied impact of intense fies involves changes to parasite load, Ifa parasitic species includes a free-living stage in the life cycle, high temperatures and ground-cover elimination by fine may substantially rexiuce parasite abundance and hence reduce deleterious effects of parasites on their hosts (eg. [6)). Thus, for example, wildfires have been reported to recluce ectoparasitic mite infestation rates in lizards [7] and chytrid fungus loads of boreal toads [8]. The impacts of fires on parasite loads of invasive species do not appear to have been studied, however. In the summer of 2019-20, a prolonged drought followed by atypically high summer temperatures resulted in massive fires across near-coastal regions of castem Australia [9]. The fies burnt almost 100.000 km? of vegetation [9] in an area that was already impacted by habitat destruction, disease, drought and invasive species [10]. At fist sight, we might expect an intense fire to be deleterious to an invasive amphibian such as the cane toad (Rhinella marin), by killing toads and climinating the moist refuges required by this water- dependent species [11]. However, cane toads actively prefer relatively open habitats for foraging [12] and thus can thrive in postfire landscapes (13) {© 201 The Authors Pubsed by the Royal ocety under the ts ofthe Ceatne Commons Atibton License Hntpfoeatecammonscgfcensesy/01, whi permits anestided use, proved the og author and source ae cited.wary 2022 Downloaded from https://royalsocietypublishing.org/ on 08 Febri Plausibly, intense fire also may reduce the incidence of para~ sitic hangworms that are widespread in Australian cane toads [14] and can enforce major reductions in host viability [15]. To test this idea, we compared parasite loads in cane toads from adjacent sites that differed in whether or not they had ‘been burned during the 2019-2020 bushfires. 2. Material and methods (a) Study species Cane toads (Rhinello marina; previously Bufo marinus) ane large exceptionally, to greater than 1 kg) bufoniel anurans native to northeastern South America but translocated to many countries ‘worldwide in the 1930s for pest contro in commercial agriculture {12J. Adult cane toads are primarily nocturnal (but see [16, fora ging at night for insect prey and retreating to moist shelter sites by ‘lay [17], Small tonds are restricted to the margins of natal ponds ‘until they are large enough to withstand desiccating conditions {181, but adult toads are highly dispersive and can use arid habitats if moist diurnal retreat-sites are available (eg. [19D. ‘The nematode hungworm Rhshdies pseudosphaerocephala is abundant and widespread in cane toads across Australia as ‘well as in the species’ native minge, having been brought to Aus tralia when the toads were introduced to that continent in 1935 [20]. Adult worms are hermaphroditic and live in the lining, of the host lung, producing larvae that are defaecated by the host [21]. Infective larvae are free-living in moist soil and enter the bodies of new hosts by crawling through areas of thin epidermis, such as around the eyes and cloaca (22]. Experimental manip lation of parasite loads in free-ranging toads has revealed strongly negative effects on Tungworms on host survival and behaviour [15]. Surveys in topical Australia. showed that prevalence and intensity of lungworm infections in cane toads increase with toad body size and vary seasonally [141 (b) Methods ‘We objaines 512 cane toads that had been collected by comm nity environmental groups (Border Ranges Richmond Valley LandCare Network (BRRVLN) and Clarence Valley Landcare) and another 62 animal i the course of ou own fsidwork. All toads were humanely euthanized by cooling then freezing [231 ‘We asked the LandCare groups to mcond the locaton and date of collection, and whether the area was bumed or usburned by the 2019-2000 fies based on maps provided by the Rural Fire Service, and checked by onsite inspections. All toads came from northeastem New South Wales (NSW), within 75km of the city of Casino (ee electronic supplementary materials for locations). Collections were made in November 220 to Apri 2021, TI to 16 months since intense wildfires aifected this region in December 2019 and January 2020. ‘We thawed 572 toads and recorded their body length using ‘Vernier calipers (snout urosyle length, =SUL) and sex, then dis sectec! them, To score parasite Toads, we made a midvental incision, stretched the let Tung out of the body with forceps and counted lungworms (gure 1, From those data we quanti fied parasite load in terms of prevalence (proportion of toads that contained visible Iangworms) and abundance (number of Jungworms per infected toad, omitting nominfected hosts). (0 ‘Statistical analysis Data analysis was conducted in Rv. 0 24 using the pakages Ime 25 for miveeflets model ‘me’ 1 for ineat moves a tiiyer’ [271 for data manipulation, We analysed data on Foresite prevalence ting a mined mel with 9 binary dbs Eaton an gt ink neti, For data on sbundanee, we used Figure 1. A diseded cane toad (nels erin), shoving ungworms (Rhabdas psewdosphaeoephala) in the left lung of the tad. Photgaph by Matthew Greenies ‘9 mixed model with negative binomial link distribution and log link. In both cases, we included fire condition (burned versus unburned) as a factor, and tend body size (UL) as a covariate, plus the interaction between these two variables. Population (ampling location) was included as a random factor. Preliminary analyses showed no significant differences in parasite load between sexes, nor any interaction between sex ancl body 0 we report only analyses on the combined dataset Because initial analyses identified a strong, interaction between fie condition and body size in diving parasite load, wwe arbitrarily divided toads into juvenile versus adult animals, (Ge. greater or less than 70 mm SUL, as toads below this size were difficult to sex). We then repeated the above analyses ‘with fire condition as the factor (plus population as a random Variable), to see if fire condition significantly affected parasite loads in either juvenile toads or in their adult conspecifis. ‘To explore possible shifts through time, we added ‘number of months since fire’ as an additional covariate in the analyses above, 3. Results (a) Sample sizes and mean body sizes of toads We obtained data on parasite lads of 572 tons (71 aus, 101 juveniles): 279 from five burned sites ard 293 fom five unburned sites, OF those 572 toads, 35% contained tang worms in the let Tung range 0 to 88 worms per toad Mean boy size of tonds id not fer significantly betwen burned and unbumed ste (= -083, di.=8, p=038), (b) Effects of body size and fire condition on parasite load ‘Analyses of parasite prevalence and abundance yielded simi- lar patierss and conchisions (figure 240). Fansite loads tended to be higher in unburned sites, with a strong eect of toad body size. In unburned sites, larger toads were more likely to be infected and contained more lungworms By contrast, larger tonds in bummed sites were les likely to be infected and contained fewer worms (figure 2a). AS a ocvolz07 22k 227 ow esyneuno( xo buysyqndczposieto.Downloaded from https://royalsocietypublishing.org/ on 08 February 2022 @ 10 z o mean number of worms pet 4 50 67108090100 snout-urostyle category (mm) @ bummed / unburned 110 120 130, 140150, 160170180. Figure 2. The relationship between a cane toas body sie (snout-ursye length) and its parate loadin burned and unburned sites; (a) shows data for preva- lence (proportion of toads infected with lungworms) whereas (6) shows data for abundance (numberof lungworms per toad, omitting data for non-infected individual). To ctf pattems inthe data, the gure shows data (means and associated standard eos) foreach 5 mm SUL ize as of toads. However, tatisial analyses in this paper use body size as a continuous rather than categrcal variable result, the interaction term between toad body size and fire condition was significant both for parasite prevalence (Z=287, d4.=567, p<0.005) and abundance (Z=2.47, df. =187, p< 0.014). “The data in figure 2 suggest that parasite loads of small toads were unaffected by fire condition, whereas langer toads benefited considerably. When we divided toads into juveniles and adults (ie. greater or less than 70 mm SUL), and repeated the analyses with fire condition as the only factor (plus population as a random variable), fre condition did not affect parasite loads in juveniles (I=0.225, d.f.=6, 83) but did so in adults (1=-2.75, df.=8, p=0.025). In analyses with the additional covariate ‘number of ‘months since fire’, parasite loads decreased over time but did so equally in bumed and unburned sites, consistent with a seasonal effect rather than an impact of fire per se. With time since fire included as a covariate, fire condition continued to significantly affect parasite loads (prevalence, p<00001; abundance, p <0.012). 4, Discussion Intense wildfires affected the parasite loads of cane toads that were collected a year postfire, but that effect depencled upon the body size of the host. The parasite loads of small toads were unaffected by fire, whereas those of large toads were Tower in burned areas (figure 22,5). A toad’s sex did not affect its parasite load, consistent with earlier studies [14,22]. The pattern for parasite load to increase with toad. body size, as seen in our unburned sites, mirrors the results of an earlier study in tropical Australia [14], whereas a Droader-scale sampling study reported that parasite loads were highest in toads of intermediate body sizes [28]. The effect of fire on the allometry of parasite load—the main result of our analyses above—suggests that this allometry ‘may depend upon local habitats. ‘Why should a toad’s body size affect its parasite load? Several reasons may explain why parasite load might be ‘expected to increase in larger toads (as observed in a previous ocvolz07 2k 227 jo esyneuno(Gvobuysyqndczposieto.Downloaded from https://royalsocietypublishing.org/ on 08 February 2022 study [14] and in unburned sites in our own work). For example, larger toads are likely to be older, thus, have had a longer period of exposure to parasites, andl may have less effective immune responses [29]; they may also be more vul- nerable simply because they offer a larger target for uptake of parasite larvae [14] ‘The fre dependency of ths size effect in our data suggests an additional hypothesis. Larger toads are more resistant to esiceation and thus tend to use drier micohabitats (eg. 117). Such an ontogenetic shift in habitat use might exacerbate exposure to parasites if toads in dry available shelters, increasing opportunities for paras mission (as may explain seasonal variation in parasite loads of tropical toads [14)). Alternatively, in an area with relatively low densities of toads (and thus, little sharing of refuge sites), drier areas may curtail the longevity of fe-living infective larvae [22] and thus reduce the infection rate. The later situation may apply to postfire sites in our own resus. Our study does not identify the proximate mechanism by which a previous fire (a year earlier) reduced parasite loads in adult cane toads. One possibility is that infective larvae in les away from water were killed by fire. Another is that the open sun-exposed landscape postfire provided few places for larvae to survive, except around the margins of ponds, Experimental studies have confirmed that lungworm larvae are unable to survive in dry soil [30]. Yet another is that the toads currently found in the broader landscape (is. avvay from water) within burned areas are recent immigrants from nearby sites. If lungworm infection reduces dispersal rates of toads [15] (but see [31D, such immigrants might be less heavily infected than most of their conspecifics, Future ‘work could test these possibilities by following the time course of toad colonization of postfire habitats at a finer tem- poral scale, and by directly measuring, rates of survival of Iungworm larvae, and of infection of toads, in the refuge sites available in unburned versus burned locations. More es cluster in the few trans: References 1. Dale WH et of 2001 imate change and fost 6, itrbances, Boscence $1, 73-734. (de10.164) 6 3es(00nastor23cCArOR.0«02) 2 Speer M, Lee L Hartigan J Manama 5, 2021 hagas in fequecy and laxaton fess low presure gots afecing outa Auta, inate 9, 119, (4 103390/89030045) 3. McGregor H, Legge S, Joes ME, Johnson 201 Landsape management offre and grazing regimes alters the fel habitat uisaton by 8, fel as, Plas ONE 9, 105097. 10.1711 jouma pone 105097) 44 Davies H, Ranges Tes MAN, Stok D, Miler . Gillespie GR, Murphy BP. 2021 Vaan in fel 08013) Cat densty between two lage ajcent nisin 8 Astral. Pc Coser, Bol 0.107 C2088) Brown Jk, Smith el). 2000 ian ren a7.) URS-GTR42. 2. Ogden, UT: US. Deparment of ice, Forest Service, Rokyounain Reseach Hosack BR, Lowe WH, Wae J, Com FS. 2013, Disease ina dyamic cape: host behaver and Plzat L KelhearG Shine R, 2013 Seasonal wilde reduce amphibian cyt ifecion BL (nsers 157, 23-299, 0 O16 icon 2012. Word et al 2020 rpc of 2018-2020 mega fires on Ausra fowa habitat. Nat Ea 4 1321-1326. doi10.1034155-000-1251-1) 15. Fiery PB, Sine R Bown GP, 2018 The oss of comes: ees fhe on fra. Genteh Rep. 10. Keaey Sta. 208 Te thes to Ausalas impel species anéimplkatons for 2 natn conservation response ac. Conse. Bol. 25, Stat, 231-24, (ovo. 107YPC18029 ‘generally, given that parasitism can massively influence the Viability of their hosts, we need additional research on the ways in which anthropogenically driven shifts in phenomena such as droughts, floods and bushfires modify important interspecific interactions including predation, competition ‘and parasitism. Invasive species tend to thrive in highly ‘modified habitats, for multiple reasons (eg, (3,12), and our study suggests that abiotic shifts caused by intense fire may contribute to that success, by releasing invaders from the negative effects of parasites. Fibs AM procedures were conducted sith approval fom the ‘Macquarie ‘University Animal Ethics Committee (AEC reerence no, 2019/040-2). All procedures involving animals were carvied out in accordance with relevant guidelines and regulations Gnclacing, ARRIVE guidelines) Dato acesibity. AIR code fs nchded as supplementary files. Data om collection dates and locations, toad body sizes and parasite loads are available from Dryad_hitps://doi.org/105061/ dryad.rwdbr9b. ‘The data are provided in the electronic supplementary material [32] Autos! onbutons, SK: canceptualization, formal analysis, investi- ‘gation, methodology, writing — original draft, writing the review land editing: Mj.G.: conceptualization, data curation, investigation, ‘methodology, supervision, writing the review and editing: RSS com. ceptualization, data curation, formal analysis, funding, acquisition, methodology, project administration, supervision, writing. the ‘original draft, writing the review and editing. All authors: gave final approval for publication and agreed ta be held accountable for the work performed therein, Cnpeting interes. We declare we have no competing interests Funding. The work was funded by the Australian Research Council (grant no, LPITDIONIS, to RS). ‘dnonledgemenis We thank the Border Ranges Richmond Valley LandCare Network (BRRVEN) and Clarence Valley Lanare for pro- viding toad carcasses for dissection, Grafton Primary Industries (NSW DPIE) for Inbortory facilities, Russell Jago for advice om field sites, Greg Brown for statistical advice and Melanie Elphick for help with figures. Scat JD, ng DM, Talley JL, Wer JR, Fulendort 11, Schwa L, Ald RA. 1996 Desicaton and 50, Debs DM, 2015 Drought inns cont of pars ls of cate on pastures managed with path gig, Range Fc. Manas, 68, 290-297. (0.10160 201503001) 12, Shine R. 2018 Cne too wos. ety, CR: ‘Avare Ru Balle J Sats X, aus. 2021 Fre reduces parasite lad ina Meiteranean Hard. 3, Hannah D, Smith G. 1985 Prescribed es in nave Pr. R Sec. B 288, 20211230 4:10.08 sheet we in a tpi amphibian cmparing toads with ysl made, fund, Fo 10, 183-20, (4:10.2307 2388645) Unierty of Calta rss. forest rested in hgh amphibian and reptile abundance but not pees nes in Austra. Meo, Quen Mus. 38, 53-53) yas ofthe lngworm, tabs seuiphoercephaa in weeny colonised cane ‘ood (hina maine) poplin in oil asa Pras 48, 753-761, (10. ‘wisipxa 20105002) rast ifcion: eles of rmaig ogwors on perfomance, gouth and sunivalof fe ‘angng cane tas. Fat Ec. 32, 402-41, (ostorr11365 243512352) 01901207 LL 137 Yow synewno[o-SuysyandhaposetasDownloaded from https://royalsocietypublishing.org/ on 08 February 2022 Petit, Ducter 5, Deir, Ware 6, Stine 200 Dial acy in cane toads (ino ‘maria 's geapialy widespread. Si. Rep. 10, 7B, (ol 0. 1038'41588 020624023) Seebucer ,Afrd RA 2002 Sher miahabitats tein boy tempectre and dehydration rates of a terest amphibian (Bu main). 4 Were 36, 68-75 (4oe102307/1568808) (ila, Piips BL, Sine R. 2009 oes esicaton vik dv the dstrutian of joven cane toads (Blo marinus a wpial Aus? Top. fal, 28, 183-200. (430.1017) 5055740800565) Ward fear G Greeles My, Stine 2016 Tas on loa patil ecagy and hata use of vase Cane tots (hinle maria) in Hawa. Plas ONE M1, 151700. (10.137 oul ponects1720) Duty 5, hie 2008 Og ofthe pasate of an ivating specs, the Astalion cane toad (ufo ‘marin athe lngwoms Asta or Area? Mol Ba 17, 4418-404 (410.1111 41365 2902008039723) 2. a u %. % Boker MA. 1979 The fe-Ivng and parse evelopment of habs 5p. Nenad: Abas) in aghibians. Con J Zo. $7, 161-178. (ok 10.139479-014) Kedar G Webb J, Hagman M, Sine R 2011 Intact between lfc elit laae and ther anuan hos. Heptlagic 67, 378-85. (dos 1.15 MERPETOLIGIA. 0-1-0007.) ‘Shine Ail, Mana A, Sew pst AL, Les A, 2015 “aang then fexing’ a humane ay tol ampbians an reps? Bia. Open 4, 160-163. (ok10.120,61012179) 1 Cae Team. 200 fe Lguoge ond enerent or stil computing. enna, asia: R Foundation fr Sata Computg, Bates 0, Mace Baller 8 Waker S. 205, Fig Siew mined fet models ung Ine. 4 StS. 67, 1-88 (10.1637 067101) Pine J, Bates, Deby , Sattar D,R ore Team. 2020 ne: ea andrea med ees models. padage veson 31-14. See psi! BIN projec 1gpadtage=nime 3h Wid Het. 208 kore othe ene, pen Sore Sof 4,185, (oe 021s 1686) Ban DP 1858 mais of mata infections of habs ae (Neat) in Bufo mars (Amphibia) in Astal,Pastlgy 117, 505-51. (det. 1017 0032086008159) Boley PW, Sayer PW, Blaustein AR, 2019 Host age ates amphibian suscep 1 Bavacecyian Sndabtis, an emerging infectious fngal pathogen. PLoS ONE 14 218. (410137Vjounal po 0227181) Noleear (abe Gazmin E, Sine R 2012 Inadvertent consequences of community ised fl to cnt invasive specs. sere. Lt 8, 340-36, (J10.1119.155 269201200513) Bown Hear Pat Shine R. 2016 Te npc of lungworn pastes oo ts of dps thi aan th ashe cane a Bast 18, W114 (oi 10100751053001549831) Kaier 9H, Genes My, Sine R 2021 ies moi the prate lads of ase cane toads. Fst lqsyjewno(/u0-Buysyqndaaposjetos ovvouzez 2b 027 1g