Mitochondrial-Dna Polymorphism in the Oilbird (Steatornis Caripensis, Steatornithidae) in Venezuela

The Auk 111(3):573-578, 1994

MITOCHONDRIAL-DNA POLYMORPHISM IN THE OILBIRD (STEATORNIS CARIPENSIS, STEATORNITHIDAE) IN VENEZUELA

PATRICIA C. GUTIERREZ 1 BiologyDepartment, State University of New York at Albany,Albany, New York 12222,USA

A13sTRAcT.--Mitochondrial-DNA(mtDNA) polymorphismswere studied in the Oilbird (Steatorniscaripensis). Twelve closely related (P = 0.06 to 0.35%) mtDNA haplotypes were found in Oilbird colonies studied in northeastern and northwestern Venezuela. Eleven of the mtDNA clonesare related to the ancestralone by one or two mutational steps.Female- mediated gene flow is high (Nm > 1) among the coloniesstudied. As a consequenceof the high female-mediatedgene flow, no phylogeographicstructuring among the mtDNA composite haplotypeswas observed.Evidence from mtDNA analysissuggests that Oilbird pop- ulationsin Venezuela have gone through a bottleneck.Results also seemto indicate that the annual postbreedingmigrations of Oilbirds from the GuacharoCave to the cavesin the Mata de Mango area involve mostly the breeding adults, whereas juveniles disperse from the GuacharoCave to the Mata de Mango cave systemfor longer periods.Received 3 August1993, accepted15 November1993.

OILBIRDS(Steatornis caripensis) are cave-dwell- that stay behind are probably too old or weak ing, nocturnal,and gregariousfrugivores. They for the journey. Postbreedingmigrations occur are restricted to the cave habitats of northern in other Venezuelan colonies, in Ecuador and South America and Trinidad, and occur in Ven- Colombia (Bosqueand Ramirez 1988). Return ezuela, Colombia, Guyana, Ecuador, Peru, Bo- of Oilbirds to the Guacharo Cave in March co- livia, northernBrazil, and Panama(Bosque 1986). incides with the onset of the breeding season, Oilbird coloniesin Venezuela are found pri- and most adults return to mate, build, or rebuild marily in the northeastand northwest (Fig. 1), the samenests (Snow 1961,Bosque and Ramirez with the exceptionof two coloniesin the Andes 1988, Roca 1994). and six colonies near the Brazilian border (Bos- Demographiccharacteristics that influencethe que 1986). geneticarchitecture of Oilbirds are:colonies of In Venezuela,Oilbirds breedonce a year (Roca limited size; breeding philopatry to particular 1994). Reproductionwithin a colony is mod- caves;and life-long monogamy (Snow 1961). eratelysynchronized and the reproductivepe- Their distribution in caves that are distant from riod extendsfrom March to September.At the one another suggeststhat the coloniesmay be end of the breeding season,most Oilbirds from genetically isolated from one another. How- the GuacharoCave migratesoutheast to an area ever, postbreeding migrations may facilitate known as Mata de Mango. Mata de Mango is a gene flow among different colonies.Mitochon- limestonemountain range coveredby pristine drial-DNA (mtDNA) analysis can provide a tropicalforest; deep canyons,rivers, waterfalls, measureof the degreeof geneticseparation of caverns, and sinkholes are common in this area. coloniesas a function of isolation.It is expected Oilbirds begin to leave the Guacharo Cave that maximumdivergence occurs between those shortly after nestlingsfledge and there is no coloniesthat have been isolated for the longest bias for age class.Snow (1979) and Roca(1994) time. Conversely,if there is female-mediated indicatedthat postbreedingmigrations are in- gene flow betweencolonies, their geneticsim- complete;between 5 and 27%of the colonyre- ilarity will be proportionalto the rate of female mains at the GuacharoCave during the non- dispersal. I used restriction-fragment-length breedingseason. Roca (1994) postulated that lack polymorphisms(RFLPs) in mtDNA to examine of food triggersmigration and that thosebirds population structure in Venezuelan Oilbirds, and determine whether femalesdisperse from their natal colony,or whether philopatry leads Current address:Casilla, 9813, Santiago,Chile. to genetic structure.

573 574 PAtRICrAC. GUTIERREZ [Auk, Vol. 111

1• 1,2,3,5 n=10 11,12 • • 6,7,8,10,n=15 • •,•n=10 200 km n =11

Fig. 1. Distributionof Oilbird in Venezuela.Asterisk indicates extinct colony, and dot an activecolony. Enlargedarea shows geographic distribution of mtDNAhaplotypes numbered 1-12. Localities indicated: (G) Guacharo Cave; (S) Sucia Cave; (C) Clara Cave; and (A) Los Anadidos Cave.

MATERIALS AND METHODS x g for 10 min to reducenuclear contamination of the mitochondrialpellet, and to 17,369 x g for 40 min Nestlingsthat fell from their nestswere collected to enhancemitochondrial yield. The pelletwas washed from four localities in northeastern and northwestern several times (until no fat was visible in the super- Venezuela. In the northeast(State of Monagas),sam- natant). pleswere collectedfrom GuacharoCave (colony size Harsherconditions were usedduring lysis to rid ca. 10,000 individuals, n = 47), Sucia Cave (ca. 1,000 the mtDNA of lipidsand proteins.Mitochondria were individuals, n = 26), and Clara Cave (ca. 1,500 indi- resuspendedin 5 ml of TE (10 mM Tris-HC1,1 mM viduals, n = 17). Suciaand Clara cavesare about 5 km EDTA, pH 8.0), lysed with 0.30 ml of 20% sodium apart and locatedin the Mata de Mango area, and dodecylsulfate (SDS), and incubatedat 65øC,for 30 thesetwo cavesare approximately45 km from Gua- min. The lysatewas cooledon ice and 0.175 ml/ml charo Cave. In northwestern Venezuela (State of Yar- of lysate,of a high salt solution(60 ml 5 M K acetate, acuy), nestlingswere collectedfrom Los Anadidos 11.5 ml glacialacetic acid, 28.5 ml H20) was added. Cave (ca. 1,200 individuals, n = 10), which is approx- After 2 h on ice, centrifugationat 34,858 x g for 5 imately 500 km from the northeasterngroup of caves min precipitatedmost of the SDS-proteincomplex. I (Fig. 1). extractedmtDNA in the supernatantat leastfive times The method of Lansman et al. (1981) for the iso- with phenol-chloroform-isoamylalcohol(24:24:1) and lation of mtDNA was modified in several ways be- twice with chloroform; the mtDNA was dialyzed cause it would not render intact mtDNA from Oilbird overnightand precipitated. heart and liver tissues. The concentration of EDTA in Thisextraction-purification protocol did not render the homogenizingbuffer was increasedto 200 mM. an mtDNA preparation completelyfree of nuclear Centrifugationspeeds and time were increasedto 754 contamination.Therefore, mtDNA preparationswere July1994] OilbirdmtDNA Polymorphism 575 digestedwith PstI, which produced two large frag- TABLE1. Compositehaplotypes of 46 Oilbirdsfrom mentsthat were cut out of the agarosegel and purified four colonieswith haplotypelabels (1-12) and fre- (Maniatis et al. 1982).This purified mtDNA was then quenciesin eachcolony. Positions 1 to 6 of com- digestedwith AvaI, AvaII, HindIII, MspI, Sau96I,and positehaplotype contain letters designating different fragment patterns produced by restriction SpeI. Digestions were carried out in 10 •1 reaction enzymesAvaI, Avail, HindIll, MspI, Sau96I,and mixtures containing mtDNA and at least three units SpeI. of enzyme.Digestions were extendedfor 16 h at 37øC to ensure completedigestion, and stoppedby freez- Composite No. ing. Label haplotype Colony birds Restrictionfragments were end-labeledwith •2pus- 1 CCCCCC Guacharo 6 ing rapid end-labelling of DNA with the Klenow Sucia 5 fragmentof E. coliDNA polymeraseI (Maniatiset al. Clara 3 1982).Vertical polyacrylamide gel electrophoresiswas Anadidos 9 usedto separatemtDNA fragmentsof a 100 to 1,000 2 CCCCFC Guacharo 1 base pairs (4% gels), and fragments of 80 to 500 bp 3 CCCCAC Guacharo 1 (5% gels; Maniatis et al. 1982). Autoradiographywas 4 CCCBCC Clara 1 usedto visualize the fragments(Maniatis et al. 1982). 5 CIgCCCC Guacharo 1 Presence or absenceof restriction fragments was 6 CCCECC Guacharo 1 7 CCCCDC Guacharo 1 usedto estimatenucleotide sequence divergence be- Sucia 5 tweenmtDNA haplotypesusing the indexof the rel- Clara 4 ative genetic similarity (P) of Nei and Li (1979). A 8 CCCBDC Guacharo 1 G-testof significancefor the differencesin the hap- Clara 1 lotype frequencieswas used(Sokal and Rohlf 1969). 9 CCCCGC Anadidos 1 The mtDNA haplotypeswere clusteredusing the 10 CDCCIgC Guacharo 1 unweighted pair-group method with arithmetic av- 11 CDCCCC Guacharo 1 erages(UPGMA) of Sneathand Sokal(1973). The frag- Sucia 1 ment presence/ absence matrix was subjected to a phy- Clara 1 12 CIgCCBC Guacharo 1 logeneticanalysis using parsimony (PAUP; Swofford and Olsen 1990). Gene flow between the different colonies was estimatedby threedifferent methods. The private-allele the size reported for many birds (Avise and approach of Slatkin (1985) and Slatkin and Barton Zink 1988,Shields and Helm-Bychowski1988). (1989) usesthose alleles that are "rare" (i.e. unique The seven restriction endonucleasesused pro- haplotypes).The genetic-identityapproach of Taka- duced between 65 and 69 mtDNA restriction hataand Palumbi (1985)takes into accountthe genetic similarityof alleleswithin and betweendemes. This fragmentsin eachindividual assayed.In total, approach uses an estimator of Wright's (1951) Fst- 61 fragmentswere identical and 13 were poly- statistic(Gst) to estimateNm and is basedon identity morphic.This represents355 bp of information probabilities,the probability that two sampledho- or 2.21%of the mtDNA genomeanalyzed per mologousDNA segmentswill be identical. individual. A 10-bp size polymorphism was ob- The cladistic method of Slatkin and Maddison (1989) served in the restriction-fragmentprofiles of is basedon the minimum number of migration events Sau96I.The two size classeswere always scored necessaryto explain the observedgeographic distri- as the samefragment. bution of mtDNA haplotypes.The method provides an estimate of Nm, where N is the size of each local Every individual was assigneda composite population and m is the migration rate. When the haplotypeof six letters,each corresponding to phylogenyof mtDNA haplotypesis known, the geo- the restriction-fragmentpattern observedfor graphic location of each population is treated as a each restriction endonuclease. The common multistatecharacter with one statefor eachgeograph- haplotype was designatedby C, and variants ic location.A parsimonycriterion is usedon a tree of were designatedby A, B, D, E, F, and G. The the phylogeny to determine the minimum number observed frequencies and geographic distri- of migration events consistentwith the tree. Slatkin butionsof the 12 haplotypesobserved are shown and Maddison (1989) provided confidenceintervals in Table 1. A G-test with a Williams' correction for Nm. for small sample size (Sokal and Rohlf 1969) indicatedthat haplotypefrequencies among dif- RESULTS ferent colonieswere significantly different (P < 0.05). Eleven of the mtDNA haplotypes are The sizeof the mtDNA genomeof the Oilbird related,by one or two mutationalsteps, to the is 16,250 _+SE of 150 basepairs (bp), typical of ancestralone (haplotype1). It is consideredto 576 PATRICIAC. GUTIERREZ [Auk,Vol. 111

TAllI•E2. Nrn values (lower left) and Gsrvalues (up- up to 96.3 km 2, with the maximum distancebe- per right) estimatedby Takahataand Palumbi (1985) tween feeding localitiesof about 150 km. Dur- method. ing their postbreeding dispersal they are ca- Gua- Ana- pable of flying 240 km in one night in search charo Sucia Clara didos for food (Roca 1994). Therefore, the limited

Guacharo -- 0.129 0.133 0.296 mtDNA phylogeographicpopulation structure Sucia 3.37 -- 0.0076 0.333 is probably attributable to a high level of dis- Clara 3.25 131.0 -- 0.329 persal and gene flow. Anadidos 1.18 2.50 2.53 -- In a number of other avian species,a similar lackof phylogeographicdifferentiation has been attributed to female dispersal and gene flow. be ancestralbecause it is present in 50% of the Tegelstrom(1987) investigatedmtDNA varia- birds sampled, it is found in all colonies, and tion in the GreatTit (Parusmajor) in threeneigh- it is related to all others (Avise et al. 1987). boring localities in Sweden, and found no ob- Values of P among Oilbird haplotypesranged viousspatial structuring of mtDNA clones.Ball from 0.0006 to 0.0035 (œ= 0.0018), indicating et al. (1988) conducteda continentwide survey that haplotypesare closely related. of Red-wingedBlackbirds (Agelaius phoeniceus) The geographicdistribution of the composite and found very little population structure.Te- mtDNA haplotypes(Fig. 1) showsthat only 4 gelstromet al. (1990)studied the Pied Flycatch- of 12 haplotypes are widespread. The PAUP er (Ficedulahypoleuca) in Sweden,and found no parsimony analysis generated 16 most-parsi- obvious geographic structuring in the very monious trees and the consensus tree showed closely related mtDNA clones. a polychotomy (uncertain resolution) of hap- In Oilbirds, gene flow was inversely related lotypes2 through 9. The UPGMA phenogram to distance,the exceptionwas the estimatesbe- showed no obvious geographicclustering of tween Sucia and Anadidos, and Clara and An- mtDNA haplotypes. adidos. This is the result of Sucia, Clara, and Gene flow (Nm) was estimatedby three dif- Anadidos having fewer different haplotypes ferent methods. The private-allele approach than Guacharoand Anadidos(Fig. 1) and, there- gave an estimatefor Nm of 0.76. The genetic- fore, appearingmore similar than Guacharoand identity approachgave Nm valuesabove 1 (Ta- Anadidos. ble 2). The cladistic method estimated Nm to be My mtDNA analysishas shown that the Oil- 1.81 and 1.23 based on UPGMA and PAUP trees, bird colonies studied in Venezuela do not rep- respectively.The respective95% confidence in- resentisolated populations. Rather, they are dy- tervals were 0.17 and 30.69. namicallyinterconnected by migrationof adults that have been shown to be philopatric (Snow DISCUSSION 1961, Roca 1994) and by dispersalof juveniles from the GuacharoCave to the Mata de Mango These P values (0.06 to 0.35%) for Oilbirds are area.Nesting spaceis very limited in the caves similar to those obtained for other conspecific where Oilbirds are found (pers.obs., Snow 1961, comparisonsof avian mtDNA (Kesslerand Av- Roca 1994). When adults return to breed, they ise 1985, Ovenden et al. 1987, Shields and Wil- tend to occupythe samenest. A few young and son 1987,Tegelstrom 1987, Ball et al. 1988,Hare inexperienced birds end up without a nest and, and Shields 1992). In avian specieswhere fe- therefore, without a mate remain as floaters in malescan disperseover large areas,closely re- the colonyand cannotbreed. I believethat most lated mtDNA haplotypescan be found in geo- of the juveniles remain in the Mata de Mango graphicallydistant localities,and considerably area for some time, a few may return to the divergent ones can be present in the same lo- Guacharo Cave to breed, and others breed in cality (Tegelstrom1987). Thus, mtDNA haplo- the cavesof the Mata de Mango. typeswill not show geographicstructure. Birds The Oilbird is probably a relatively old spe- are highly mobile; therefore, phylogeographic cies,given that the family Steatornithidaemay differentiation in mtDNA haplotypes is ex- have arisen about 50 mya near the baseof the pected to be minor (Ball et al. 1988). Oilbirds caprimulgiformradiation, possibly derived from are highly mobile;during the breeding season the Coraciiformes (Olson 1987). Oilbirds ap- their home ranges can expand over an area of parently originated and differentiated in North July1994] OilbirdmtDNA Polymorphism 577

America,and had the samediet asmodern rep- completelyto extinctionof lineagesas a result resentatives.The present-dayspecies must be of more recent ecologicaldisturbances or the regarded as a relic in South America (Olson appearanceof new mutations, but could also 1987). reflectthe fact that samplesizes were relatively Ancestralrepresentatives of today'sSteatornis small. A more exhaustivesurvey would prob- may have colonized South America after the ably show that rare compositemtDNA haplo- Central Americanland bridge was formed and types are more widely distributed,confirming when the Pleistoceneclimatic changesforced the lack of geographicstructuring due to gene them to migratesouth. Oilbirds were probably flow observedin this survey. establishedin South America during the last glacial periods. Warm and humid conditions ACKNOWLEDGMENTS duringinterglacial periods may have caused the This researchwas supported by grantsfrom Wild- expansionof low montaneforests on which oil- life Conservation International, a division of the New birds depended.This, in turn, may have facil- York ZoologicalSociety, the InternationalCouncil for itated migration and colonization of new hab- Bird Preservation,Sigma Xi, StateUniversity of New itats with subsequent establishment of new York at Albany Benevolent Award, and the Foun- colonies.Less favorable conditions may have dation for the Defense of Nature (FUDENA, Vene- caused the reduction of some forest habitats, zuela). In Venezuela, the Institute of National Parks and the extinctionof somecolonies, and may (INPARQUES), the Ministry of the Environment havereduced gene flow amongthose surviving. (MARNR) and Profauna were fundamental in the de- These conditionscould have produceda bot- velopment of this work. I am grateful to R. Roca,T. Valdez, E. Hall, K. Scully, R. Gutierrez, and L. Aular tleneck (by founder effectsand local extinction for helping collectsamples. M. J. Braun, R. M. Zink of lineages)in the population,whereby genetic and G. Sattier provided useful comments on the variability would have been reduced. As cli- manuscript. matic conditionsbecame favorable again, subpopulations that were subdivided becauseof LITERATURE CITED ecologicalconditions could have begun to re- coverfrom this bottleneck.New mutationsap- AVISE, J. C., J. ARNOLD, R. g. BALL,E. BERMINGHAM, T. LAMS, J. E. NEIGEL, C. A. R•s, AND N. C. peared,and with the reestablishmentof gene SAUNDERS.1987. Intraspecificphylogeography: flow, they couldhave begun to spreadthrough The mitochondrialDNA bridge between popu- the subpopulations. lation geneticsand systematics.Annu. Rev. Ecol. A reductionin Oilbird geneticdiversity due Syst.18:489-522. to a bottleneckeffect is also suggestedfrom AWSE,J. C., ANDR. M. Zn•IK. 1988. Molecular genetic morphologicaland nuclear-genomedata. Len- divergencebetween avian sibling species:King tin o (1990)assayed 37 proteinloci from Oilbirds and Clapper rails, Long-billed and Short-billed collected from five colonies in Venezuela and dowitchers,Boat-tailed and Great-tailedgrackles, determinedthat all lociwere monomorphic and and Tufted and Black-crested titmice. Auk 105: fixedfor the sameallele. He alsofound no sig- 516-528. nificant difference in 49 morphologicalchar- BALL,R. M., JR.,S. FREEMAN,F. C. JAM•S,E. BERMING- HAM,AND J. C. AVISE. 1988. Phylogeographic actersstudied. The reducedgenetic divergence population structureof the Red-winged Black- and low frequenciesof most mtDNA variants birdsassessed by mitochondrialDNA. Proc.Natl. suggeststhat these mtDNA mutationsmay have Acad. Sci. USA 85:1558-1562. arisen in the populationin the recent past. BOSQUE,C. 1986. Actualizaci6n de la distribuci6n However, this homogeneity could also be a del Guftcharo(Steatornis caripensis) en Venezuela. productof geneflow. Due to the greatdispersal Bol. Soc.Venezolana Espel.22:1-10. capabilitiesof Oilbirds and to the observedrates BOSQUE,C., AND R. RAMIREZ.1988. Post-breeding of female mediated gene flow, these mtDNA migration of Oilbirds. Wilson Bull. 100:675-677. haplotypeshave spreadthrough the popula- HARE, M.P., AND G. F. SHIELDS. 1992. Mitochondrial tion. DNA variation in the polytypic Alaska Song Sparrow. Auk 109:126-132. The presenceof uniquemtDNA haplotypes KESSLER,L. G., ANDJ. C. ARISE. 1985. A comparative in somecolonies and not in others(i.e. haplo- description of mitochondrial DNA differentia- type 9 in Anadidos;haplotype 4 in Clara;hap- tion in selectedavian and other vertebrategen- lotype8 in Claraand Guacharo;haplotypes 2, era. Mol. Biol. Evol. 2:109-125. 3, 5, 6, 10, 12 in Guacharo)may not be due LANS•AN, R. A., R. O. Sr•ADE,J. F. Sr•APIRA,AND J. C. 578 PATRICIAC. GUTIERREZ [Auk,Vol. 111

AVISE. 1981. The use of restriction endonu- SLATKIN,M. 1985. Rare alleles as indicatorsof gene cleasesto measuremitochondrial DNA sequence flow. Evolution 39:53-65. relatedness in natural populations. Ill. Tech- SLATKIN,M., ANDN.H. BARTON.1989. A comparison niques and potential applications. J. Mol. Evol. of three indirect methodsfor estimatingaverage 17:214-226. levels of gene flow. Evolution 43:1349-1368. LENTnqO,M. 1990. Flujo g6nicoy variabilidadmor- SLATKIN, M., AND W. P. MADDISON. 1989. A cladistic fo16gicaen el Guficharo(Steatornis caripensis). M.S. measureof gene flow inferred from the phylog- thesis, Univ. Sim6n Bolivar, Caracas, Venezuela. enies of alleles. Genetics 123:603-613. MANIATIS, T., E. F. FRITSCH,AND J. SAMBROOK.1982. SNEATH, P. H. A., AND R. R. SOKAL. 1973. Numerical Molecular cloning. A laboratory manual. Cold taxonomy. W. H. Freeman, San Francisco. Spring Harbor Laboratory. SNOW,B.K. 1979. The oilbirds of Los Tayos.Wilson Bull. 91:457-461. NEI, M., AND W-H. LI. 1979. Mathematical model for studying geneticvariation in terms of restriction SNOW,D.W. 1961. The natural history of the Oil- endonucleases. Proc. Natl. Acad. Sci. USA 76:5269- bird, Steatorniscaripensis, in Trinidad. W.l. Part 1, 5273. General behaviorand breeding habits.Zoologica 46:26-48. OLSON,S.L. 1987. An early EoceneOilbird from the SOKAL,R. R., ANDJ. J. ROHLF. 1969. Biometry. W. H. Green River formation of Wyoming (Caprimul- Freeman and Co., San Francisco. giformes: Steatornithidae). Docum. Lab. Geol. SWOFFORD,D. L., ANDG. J. OLSEN. 1990. Phylogeny Lyon 99:57-69. reconstruction.Pages 411-501 in Molecular sys- Ox ENDEN,J. R., A. G. MACK•NLAY,AND R. H. CROZIER. tematics (D. M. Hillis and C. Moritz, Eds.). Sin- 1987. Systematicsand mitochondrial genome auer Associates, Sunderland, Massachusetts. evolution of Australian rosellas(Aves: Platycer- TAKAHATA, N., AND S. R. PALUMEI. 1985. Extranu- cidae). Mol. Biol. Evol. 4:526-543. clear differentiation and gene flow in the finite ROCA,R. 1994. Oilbirds of Venezuela: Ecology and island model. Genetics 109:441-457. conservation.Nuttall Ornithological Club, No. TEGELSTROM,H. 1987. Genetic variability in mito- 24 (R. A. Paynter, Ed.). Cambridge, Massachu- chondrial DNA in a regional population of the setts. Great Tit (Parusmajor). Biochem. Genet. 25:95- SHIELDS,G. F., ANDA. C. WILSON. 1987. Subspecies 110. of the Canada Goose (Branta canaden$is)have dis- TEGELSTROM,H., H. P. GELTER,AND M. JAAROLA.1990. tinct mitochondrial DNAs. Evolution 41:662-666. Variation in Pied Flycatcher(Ficedula hypoleuca) SHIELDS, G. F., AND K. M. HELM-BYCHOWSKI. 1988. mitochondrial DNA. Auk 107:730-736. Mitochondrial DNA of birds. Curr. Ornithol. WmGHT,S. 1951. The genetical structure of popu- 5:273-295. lations. Ann. Eugen. 15:323-354.