Genetic Characterization of an Invasive Boa Constrictor Population on the Caribbean Island of Aruba Author(S): Lauretta M

Genetic Characterization of an Invasive Boa constrictor Population on the Caribbean Island of Aruba Author(s): Lauretta M. Bushar, R. Graham Reynolds, Sharese Tucker, LaCoya C. Pace, William I. Lutterschmidt, R. Andrew Odum, and Howard K. Reinert Source: Journal of Herpetology, 49(4):602-610. Published By: The Society for the Study of Amphibians and Reptiles DOI: http://dx.doi.org/10.1670/14-059 URL: http://www.bioone.org/doi/full/10.1670/14-059

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. Journal of Herpetology, Vol. 49, No. 4, 602–610, 2015 Copyright 2015 Society for the Study of Amphibians and Reptiles

Genetic Characterization of an Invasive Boa constrictor Population on the Caribbean Island of Aruba

1,2 3 1 1 4 LAURETTA M. BUSHAR, R. GRAHAM REYNOLDS, SHARESE TUCKER, LACOYA C. PACE, WILLIAM I. LUTTERSCHMIDT, 5 6 R. ANDREW ODUM, AND HOWARD K. REINERT

1Biology Department, Arcadia University, 450 South Easton Road, Glenside, Pennsylvania 19038 USA 3Department of Organismic and Evolutionary Biology & Museum of Comparative Zoology, Harvard University, 26 Oxford Street, Cambridge, Massachusetts 02138 USA 4Department of Biological Sciences and the Texas Research Institute for Environmental Studies, Sam Houston State University, 1806 Avenue J, Huntsville, Texas 77340 USA 5Department of Herpetology, Toledo Zoological Society, PO Box 140130, Toledo, Ohio 43614 USA 6Department of Biology, The College of New Jersey, 2000 Pennington Road, Ewing, New Jersey 08628 USA

ABSTRACT.—Boa constrictor was first documented on the Caribbean island of Aruba in 1999. Despite intensive efforts to eradicate the snake from the island, B. constrictor has established a stable, reproductively successful population on Aruba. We generated mitochondrial sequence and multilocus microsatellite data for individuals from this population to characterize the origins and means of introduction to the island. Phylogenetic analyses and measures of genetic diversity for this population were compared with those for invasive B. constrictor imperator from Cozumel and B. constrictor constrictor from Puerto Rico. Cozumel populations of B. c. imperator had significantly higher numbers of alleles and significantly higher values for FIS than the Puerto Rico and Aruba populations. Observed, expected, and Nei’s unbiased heterozygosities, as well as effective number of alleles, were not significantly different. The effective population sizes from Aruba and Puerto Rico were generally lower than those for either of the Cozumel populations; however, there were broad confidence intervals associated with published estimates. We conclude that the present B. constrictor population on Aruba most likely resulted from the release or escape of a small number of unrelated captive snakes originating from South America (B. c. constrictor) and are phylogenetically distinct from introduced boas on Puerto Rico and Cozumel. This study adds to the growing body of evidence suggesting the ease with which a small number of relatively slow-maturing B. constrictor can quickly invade, become established, and avoid eradication efforts in a new location with suitable habitat.

Invasive species can alter habitats, compete with native The release of several related neonate B. constrictor constrictor wildlife for limited resources, reduce biodiversity, and lead to around 1992 is believed to be responsible for the population on extirpation of native species (Pysˇek and Richardson, 2010). Puerto Rico, and B. c. constrictor on Puerto Rico share a Many reptile species have become naturalized and subsequently mitochondrial DNA haplogroup commonly found in breeding invasive outside of their native ranges, occasionally leading to collections and zoo specimens (Reynolds et al., 2013). significant ecological damage to invaded areas (Kraus, 2009). A Aruba is a small continental island approximately 33 km long now classic example, the Brown Treesnake (Boiga irregularia), and 8 km wide (~175 km2). The first five wild B. constrictor were was responsible for the extinction of at least 18 different species documented on the island in 1999, and in 2000 the government of birds, lizards, and bats within 50 yr of its introduction to of Aruba established a program whereby any B. constrictor Guam (Savidge, 1991; Mackessy et al., 2006). The Burmese found was collected, recorded, and euthanized. Despite these Python, Python bivittatus, was first observed in southern Florida efforts to eradicate the snake from the island, a reproductively approximately 30 yr ago and is implicated in the decline of successful population had been established by 2005 (Quick et several species of mammals in Everglades National Park al., 2005). All size classes have been reported, from neonates to (Dorcas et al., 2011). adults approaching 3 m total length, as well as gravid females A particularly interesting situation in invasion biology is (Quick et al., 2005). Ongoing research is determining the when species that are threatened in their native range become ecological impacts of this invasion, and initial dietary data invasive elsewhere. For example, several mainland and insular suggest that B. constrictor snakes on the island are euryphagous populations of Boa constrictor are considered threatened or (Quick et al., 2005). There is concern that B. constrictor could endangered (e.g., Chiaraviglio et al., 2003; Boback, 2005; directly compete with the endemic Aruba Island Rattlesnake Henderson and Powell, 2009), primarily as a result of (Crotalus unicolor), impact populations of native wildlife, and persecution, collection for the pet trade (Dodd, 1986), and prey upon domestic animals (Quick et al., 2005; van Buurt, hunting for their skins. Nevertheless, invasive populations of B. 2011). constrictor have recently become established in southern Florida Four important lines of inquiry can yield meaningful (Snow et al., 2007), on the Mexican island of Cozumel information in forensic invasive species genetics: species or (Martıńez-Morales and Cuaroń, 1999), and on the Caribbean lineage identification, geographic origins of the invasive islands of Puerto Rico (Reynolds et al., 2013) and Aruba (Quick individuals, assessment of evidence for propagule pressure or et al., 2005). The Cozumel population of B. constrictor imperator multiple introductions, and possible means by which the species is ecologically damaging (Romero-Na´jera et al., 2007) and is arrived in the invaded range (Fitzpatrick et al., 2012). believed to have become established island wide following the Furthermore, comparison among separately introduced popu- release of between 2 and 30 individuals in 1971 (Martıńez- lations of invasive species might yield important population Morales and Cuaroń, 1999; Va´zquez-Domıńguez et al., 2012). genetic information germane to control efforts and the study of invasion biology. 2Corresponding Author. E-mail: [email protected] Boa constrictor is comprised of multiple highly divergent DOI: 10.1670/14-059 lineages (up to seven million years divergent; Sua´rez-Atilano et INVASIVE BOA CONSTRICTOR ON ARUBA 603 al., 2014), and identifying lineages present in an invasive data from studies on the invasive B. constrictor populations on population might shed light on origins and introduction history the islands of Cozumel and Puerto Rico. (Va´zquez-Domıńguez et al., 2012; Reynolds et al., 2013). Boas on Cozumel belong to the Central American subspecies B. c. MATERIALS AND METHODS imperator, and established boas in Puerto Rico are South American B. c. constrictor. Some authors (e.g., Henderson and Sample Collection.—The first B. constrictor was documented on Powell, 2009) recognize West Indian populations of B. constrictor Aruba in 1999 (Quick et al., 2005), and the administration of as separate species (Boa nebulosa and Boa orophias), and recent Parke Nacional Arikok in conjunction with the Aruba Veterinaire work has suggested that B. c. constrictor and B. c. imperator Dienst established a program whereby any B. constrictor found should be elevated, with the potential for recognizing a third was collected, recorded, and euthanized. Prior to euthanasia, we species (Boa sigma) in Mexico (Reynolds et al., 2014; Sua´rez- obtained blood from 22 B. constrictor collected in 2007 (Fig. 1). We Atilano et al., 2014). These species are not yet formally excised tissue samples (tail tips) from 24 B. constrictor that were described; hence we continue to refer to the subspecies, but collected and euthanized in 2014 and stored at -208Cuntil we note that there are important ecological, evolutionary, and processing. Blood samples were stored at -20 or -808C, and tail conservation implications for different lineages of B. constrictor. tips were stored at room temperature in 95% ethanol. The geographic source of B. constrictor on Aruba is not Mitochondrial DNA analyses.—We extracted and purified known; however, these snakes may have arrived on Aruba by genomic DNA from tail tips using the Promega Wizardt SV kit several possible modes. The original founders might have been (Promega, Inc., Madison, WI) and we used polymerase chain released or escaped captivity, arrived as stowaways in shipping reaction (PCR) to amplify a fragment of the mitochondrial cargo, or rafted naturally from mainland South America. Aruba genome (cytochrome B [CTYB]) with the use of primers and is only 27 km north of the Paraguana´ Peninsula of Estado conditions described in Burbrink et al. (2000). We purified and Falcoń, Venezuela, and B. constrictor are known from Paraguana´ sequenced products on an automated sequencer (ABI 3730XL) at (Quick et al., 2005) as well as coastal areas and river drainages Massachusetts General Hospital DNA Core Facility, Cambridge, of Estado Falcoń (http://www.vertnet.org accessed 9/1/2014). MA. We then assembled contigs and manually verified ambig- Propagule pressure can greatly contribute to the success of an uous base calls using GENEIOUS 7.1.7 (Biomatters, Auckland, New invasion (Simberloff, 2009), and there are several alternative Zealand). We deposited representative sequences in GenBank (though not mutually exclusive) hypotheses regarding the (KP116263–KP116286). number of snakes that were originally introduced to Aruba. To examine the Aruba boas in a phylogenetic context, we Multiple snakes may have been introduced to the island downloaded all 259 available B. constrictor (sensu lato; Reynolds through one of the aforementioned routes or, given the et al., 2014; Sua´rez-Atilano et al., 2014) CYTB sequences from potentially large litter size produced by B. constrictor (Bertona GenBank (August 29, 2014) with the use of a custom R script. and Chiaraviglio, 2003; Boback, 2005), a single gravid female This is more than double the number of sequences available in may have produced the current population. Furthermore previous studies of invasive B. constrictor populations (Reynolds parthenogenesis has been reported for boids, including B. c. et al., 2013), largely owing to extensive study of Central imperator (Booth et al., 2011); hence a single snake might not American populations (Sua´rez-Atilano et al., 2014). We note necessarily have been gravid at the time of its introduction. that some sequences are of unclear origin owing to samples Quick et al. (2005) hypothesized that the there were multiple originally obtained from the pet trade (Hynkova´ et al., 2009; separate invasions or releases of captive snakes across Aruba, Reynolds et al., 2013). We then aligned these sequences with owing to the widely separated geographic locations of the first representative haplotypes from Aruba boas with the use of the five snakes found in 1999. van Buurt (2011) posited the CLUSTALW 2.1 (Larkin et al., 2007) algorithm implemented in population to be at least partially derived from released captive GENEIOUS, with default parameters. Sequence of Corallus animals, as some individuals had color patterns rarely found in annulatus (KC750011) and Corallus ruschenbergerii (HM348838) wild B. constrictor. Boas have been kept as pets in Aruba, and were used as outgroups. We excluded sequences shorter than anecdotal accounts suggest that at least two boas were 700 base pairs (bp), trimmed ends of the alignment, and verified accidentally released (Quick et al., 2005). coding regions by translating the matrix to protein sequences. Understanding the source of an introduced population and We analyzed the mtDNA data set with the use of a Bayesian whether the invasion is the result of a single or multiple method implemented in BEAST v1.8 (Drummond et al., 2012). We introductions is important for control and management of selected the best-fit model of molecular evolution (TIM+I+G) invasive species (Ascunce et al., 2011). Population genetic and using Bayesian information criterion (BIC) in JMODELTEST2 phylogenetic methods are widely used in invasive species (Guindon and Gascuel, 2003; Darriba et al., 2012). We ran the biology, and though some important inadequacies exist (Fitzpat- Markov chain Monte Carlo (MCMC) algorithm for 100 million rick et al., 2012), these methods can impart useful information generations with the use of a Yule speciation prior and an (Sakai et al., 2001). Invasive B. constrictor have been the subject uncorrelated lognormal relaxed (UCLN) molecular clock model. of two recent genetic investigations (Va´zquez-Domıńguez et al., We repeated the analyses three times with different starting 2012; Reynolds et al., 2013), and these and other studies on numbers, sampling every 10,000 generations and discarding the invasive snakes have provided insight into the origin and first 25% of generations as burn in. We assured adequate mixing approximate number of population founders (Gautschi et al., of the chains by calculating the effective sample size (ESS) 2002; Guicking et al., 2006). In the present study, we use values for each model parameter, with ESS values >200 mitochondrial sequence data and species-specific microsatellite indicating adequate sampling of the posterior distribution. We loci to test 1) phylogenetic affinities, 2) geographic origins, 3) assessed convergence of the independent runs by a comparison relative number of founders, and 4) introduction vectors for the of likelihood scores and model parameter estimates in TRACER invasive B. constrictor population on Aruba specifically. We v1.5 (Rambaut et al., 2013). We combined the results from the further examine our results in a comparative framework with three analyses with the use of LOGCOMBINER and generated a 604 L. M. BUSHAR ET AL.

FIG. 1. Capture locations on Aruba of Boa constrictor sampled for analysis of mtDNA (triangles) and microsatellite DNA (circles). maximum clade credibility (MCC) tree with TREEANOTATOR.We We used MICRO-CHECKER version 2.2.3 (van Oosterhout et deposited the final MCC tree in TreeBase (http://purl.org/ al., 2004) to check for the presence of null alleles, calculating the phylo/treebase/phylows/study/TB2:S16656). null allele frequency according to Brookfield (1996). We Analysis of Microsatellites.—We purified genomic DNA from determined the effective number of alleles and Nei’s unbiased blood samples as described in Bushar et al. (1998) and used PCR expected heterozygosity with the use of GENALEX (Peakall and to amplify five microsatellite loci specific for B. constrictor (Bci14, Smouse, 2006, 2012). To estimate the effective population size Bci15, Bci18, Bci21, and Bci23; primers and PCR conditions as (Ne) and 95% confidence intervals (assuming random mating) described in Booth et al., 2011), with the use of Life Technologies from linkage disequilibrium we used a parametric method (Grand Island, New York) core reagents. We ordered one primer implemented in LDNe 1.31 (Waples, 2006; Waples and Do, in each set labeled with D4 dye (Life Technologies) and we 2008). The lowest allele frequency used was 0.05. We used ML- resolved PCR products on a CEQ8000 genetic analysis system Relate (Kalinowski et al., 2006) to conduct maximum likelihood (AB SCIEX, Framingham, Massachusetts). We called alleles from tests to estimate pairwise relatedness between all pairs of the resulting electropherograms with at least two researchers individuals with the assumption of a null allele for locus Bci23. independently analyzing the electropherograms. ML-Relate also calculates the most likely relationship between We used the program CONVERT (Glaubitz, 2004) to convert pairs of individuals (unrelated, sibling, half-sibling, or parent- data from an EXCEL spreadsheet into the format required for offspring). GENEPOP on the web version 4.2 (Raymond and Rousset, 1995; We compared our results to similar data generated for Rousset, 2008), which we then used to calculate observed and invasive B. constrictor populations on Cozumel (Va´zquez- expected heterozygosity. We assessed significant deviations Domıńguez et al., 2012) and Puerto Rico (Reynolds et al., from Hardy–Weinberg equilibrium with the use of a Hardy– 2013). We separated the Cozumel population into the two Weinberg exact test (Haldane, 1954; Weir, 1990; Guo and subgroups (Cozumel 1 and Cozumel 2) identified by Va´zquez- Thompson, 1992). We evaluated linkage disequilibrium with a Domıńguez et al. (2012). Although we used the same five log likelihood ratio statistic (G-test) in GENEPOP. We estimated microsatellite loci as Reynolds et al. (2013), only four of our loci probability (P) values for these tests with the use of a Markov (Bci14, 15, 18, and 21) were in common with those used by chain method (Guo and Thompson, 1992) with a dememoriza- Va´zquez-Domıńguez et al. (2012). As a result, we included only tion number of 1,000, a batch number of 100, and 1,000 the results for loci Bci14, 15, 18, and 21 in our comparison of iterations per batch. We calculated FIS according to Weir and genetic diversity statistics among the four populations. We used Cockerham (1984). one-way, Model 1 ANOVAs (Sokal and Rohlf, 2012) to compare INVASIVE BOA CONSTRICTOR ON ARUBA 605

FIG. 2. Annual number of Boa constrictor collected and euthanized on Aruba since 1999.

total number of alleles (na), effective number of alleles (ne), ing sites, whereas haplotype 3 is 1.35% and 1.25% divergent observed heterozygosity (Ho), expected heterozygosity (He), from haplotypes 1 and 2, respectively (Tamura–Nei corrected unbiased expected heterozygosity (HNei), and FIS values. sequence divergence). These data indicate that at least three Significant ANOVAs were followed by Tukey’s a posteriori unrelated maternal lineages are present on the island. Phyloge- comparisons of group means for unequal sample sizes (T’- netic analysis of a maximum of 1,062 bp from each of 198 method; Sokal and Rohlf, 2012). We performed all statistical haplotypes (from 249 original retained CYTB sequences) analyses with the use of STATISTICA 12, version 12 (StatSoft, revealed that boas introduced to Aruba correspond to the Inc., 2012; www.statsoft.com). We calculated means, standard South American B. constrictor lineage, rather than Central errors, and 95% confidence intervals from the raw data for American (B. c. imperator) or Argentinean (Boa constrictor Cozumel and Puerto Rico if not provided by Va´zquez- occidentalis) lineages (Fig. 3). Furthermore, these boas are highly Domıńguez et al. (2012) or Reynolds et al. (2013). divergentfromboasintroducedtoPuertoRico(Fig.3). Haplotypes 1 and 2 are sister lineages, phylogenetically related RESULTS to animals originating from Guyana and Suriname, whereas Annual captures of B. constrictor on Aruba rose rapidly from haplotype 3 represents a lineage sister to animals originating their first appearance in 1999, peaked at 741 snakes in 2008, and from Amazonian Brazil and Peru (Hynkova´ et al., 2009; Sua´rez- began to show a decline in recent years (Fig. 2). From 1999 to Atilano et al., 2014). 2013 capture rate averaged 278 snakes per year (Fig. 2). None of the 22 B. constrictor individuals genotyped in this Currently, B. constrictor is found island wide (Fig. 1) and the study were homozygous across the five microsatellite loci, and encounter rate has dropped and might be stabilizing (Fig. 2). only two were homozygous at four of the five loci. Although We obtained a total of 1,093 bp (near-complete coding DNA there was no evidence for significant heterozygote excess, we sequences [cds]) from 24 mtDNA sequences of Aruba B. did find evidence for significant heterozygote deficiency (P < constrictor, which collapse into three distinct haplotypes. We 0.001) at locus Bci23 with an estimated null allele frequency of found haplotype 1 in 18 individuals, haplotype 2 in 5 0.24. We found significant linkage disequilibrium (P = 0.025) individuals, and haplotype 3 in 1 individual. Haplotypes 1 between loci Bci14 and Bci23. This could indicate linkage; and 2 are minimally divergent, separated by only six segregat- however, other studies found no evidence for linkage disequi- 606 L. M. BUSHAR ET AL.

librium between these two loci (Green, 2011; Va´zquez-Domıń- guez et al., 2012). Therefore, we believe this might be a result of the small number of founding individuals in the Aruba population. We calculated standard measures of genetic diversity for the Aruba population of B. constrictor (Table 1) and compared these with published values for the introduced populations of B. c. imperator on Cozumel and B. c. constrictor on Puerto Rico (Table 2). Both the Cozumel 1 and Cozumel 2 populations had significantly higher average number of alleles than the Puerto Rico and Aruba populations (9.7 and 8.0, respectively, versus 4.0 and 4.2, respectively; Table 2). Likewise, both Cozumel 1 and Cozumel 2 populations also had a significantly higher FIS than the Puerto Rico and Aruba populations (0.23 and 0.21, versus 0.12 and 0.01, respectively; Table 2). There were no significant differences in average effective number of alleles (ne), average observed heterozygosity (Ho), or average expected heterozygosity (He) among the four groups (Table 2). Unbiased expected heterozygosity (HNei) was not calculated for the Puerto Rico population, but there were no significant differences among the Aruba and two Cozumel populations for this measure (Table 2). The effective population sizes (Ne) for both Aruba and Puerto Rico B. constrictor were lower than for either of the Cozumel populations (15 and 16, respectively, versus 256 and 749, respectively; Table 2); however, the broad confidence intervals associated with all estimates of Ne preclude drawing any meaningful conclusion regarding this measure. As discussed in Reynolds et al. (2013), using linkage disequilibrium to estimate Ne is problematic in populations originating from a biological invasion; as such, populations generally do not meet the assumption of population genetic equilibrium. As a result, the Ne calculated does not generally provide an accurate estimate of the effective population or census population size. Mean pairwise relatedness on Aruba was 0.14 (SE = 0.009; 95% CI = 0.12–0.16), with 72.3% of the pairs classified as unrelated, 10.8% as half siblings, 10.4% as parent-offspring, and 6.5% as full siblings. We emphasize that for many pairs, however, the 95% confidence interval included ‡1other possible relationships. Only 10% of the pairs were assigned to a single category (8.7% were unrelated, 1.3% were full siblings, and none were half siblings or parent offspring); 25.2% of the pairs were assigned to two categories, 40% to three categories, and 24.8% to all four categories.

DISCUSSION Despite intensive eradication efforts set in place within a year after the ﬁrst snakes were discovered on Aruba, B. constrictor have established a robust, reproductively successful population on the island. With the use of multilocus genetic data, we wished to ascertain the geographic origins, possible means of

known geographic origins for the tips in the tree, while the abbreviations ‘‘PAC’’ and ‘‘GYCA’’ designate lineages identified in Sua´rez-Atilano et al. (2014). Red triangles show locations in the phylogeny of introduced populations of B. constrictor, except for the FIG. 3. Bayesian maximum clade credibility tree for the mtDNA Cozumel, Mexico population, which likely belongs to the ‘‘GYCA’’ locus CYTB from combined runs in the program BEAST. This tree lineage of B. c. imperator (no sequence data available). Note that there are represents all available haplotype sequence data for Boa constrictor to three Aruba haplotypes. Haplotype 3 (labeled on corresponding edge) is date, including 198 haplotypes generated from 249 sequences. Posterior represented by a single individual and corresponds to a divergent probabilities >0.95 are shown as black dots at nodes. Three subspecies lineage affiliated with samples from Amazonia. The other two of Boa constrictor are shown shaded as B. c. occidentalis (light blue), B. c. haplotypes are similar and are affiliated with samples from north of constrictor (blue), and B. c. imperator (pink). Gray triangles represent the Guiana Shield. INVASIVE BOA CONSTRICTOR ON ARUBA 607

TABLE 1. Comparison of genetic diversity statistics of total number of alleles (na), effective number of alleles (ne), observed heterozygosity (Ho), expected heterozygosity (He), unbiased heterozygosity (HNei), and FIS for insular, invasive populations of Boa constrictor constrictor and Boa constrictor imperator. Signiﬁcant FIS value is indicated by an asterisk. Puerto Rico data from Reynolds et al. (2013), Cozumel data from Va´zquez-Domı´nguez et al. (2012). N/A, data not available.

Population Locus na ne Ho He HNei FIS n Aruba Bci14 6 4.46 0.68 0.78 0.79 0.14 22 Bci15 4 2.32 0.59 0.57 0.58 -0.02 22 Bci18 4 3.45 0.73 0.71 0.73 0.00 22 Bci21 2 1.82 0.50 0.45 0.46 -0.09 22 Bci23 2 1.60 0.05 0.38 0.38 0.88* 22 Mean 3.6 2.73 0.51 0.58 0.59 0.18 22 SE 0.75 0.538 0.122 0.075 0.078 0.178 22 Puerto Rico Bci14 4 3.02 0.63 0.67 N/A 0.08 32 Bci15 6 3.28 0.81 0.69 N/A 0.15 32 Bci18 5 2.67 0.59 0.63 N/A 0.07 32 Bci21 2 1.99 0.41 0.50 N/A 0.20 32 Bci23 3 2.14 0.41 0.53 N/A 0.25 32 Cozumel 1 Bci14 11 7.00 0.90 0.86 0.87 0.11 30 Bci15 9 3.31 0.40 0.70 0.71 0.29 30 Bci18 8 4.24 0.53 0.76 0.78 0.33 30 Bci21 11 6.50 0.70 0.85 0.86 0.18 30 Cozumel 2 Bci14 11 4.57 0.63 0.78 0.79 0.11 46 Bci15 4 2.07 0.46 0.52 0.52 0.27 46 Bci18 7 4.58 0.50 0.78 0.79 0.32 46 Bci21 10 4.63 0.67 0.78 0.79 0.14 46 introduction, relative number of founding individuals, and maternal lineage. Haplotype 3 belongs to a divergent lineage genetic diversity of this population. sister to animals originating from Amazonia (Hynkova´ et al., Phylogenetic analyses of our mtDNA data set point to three 2009; Sua´rez-Atilano et al., 2014). Third, none of the animals in meaningful conclusions. First, we found that all boas on Aruba Aruba are phylogenetically similar to animals introduced are South American B. c. constrictor (B. constrictor sensu stricto). elsewhere (Cozumel or Puerto Rico). Second, it seems likely this population results from the release Possible vectors of arrival to Aruba include intentional or of snakes obtained from mainland South America. Haplotypes 1 accidental release of captive animals, passive arrival in shipping and 2 appear to have originated from populations north of the cargo, or rafting from Venezuela. These scenarios are difficult to Guiana Highlands and east of the Cordillera Oriental. We are test precisely and are not necessarily mutually exclusive; unable to obtain samples from Venezuelan populations because however, our data allow for some plausible explanations for of restrictive sample export laws, and we note that we do not the origins of these animals. Given the disparate lineages and have complete sampling across the range of B. constrictor. genetic diversity represented in the Aruba population, the Nonetheless, we find these haplotypes cluster with lineages arrival of animals likely did not occur naturally or via passive originating from Guyana and Suriname (fide Hynkova´ et al., dispersal in shipping cargo from mainland South America. 2009). A single sequence from GenBank (AM236348) is highly Although we cannot rule out some animals resulting from these similar to haplotype 2 and is instructive of the caution that must vectors (owing to apparent relatedness to adjacent mainland be used when mining GenBank for phylogenetic analyses. This populations), the presence of haplotype 3 suggests that at least sequence is trimmed from a whole mtDNA genome of an one animal from the pet trade has become naturalized on the animal labeled as a Central American ‘‘B. c. imperator,’’ likely island. That a pet trade animal is included in the Aruba owing to gross phenotype (Douglas et al., 2006); however, it is population, and that introductions from captive animals clear this individual belongs to a South American B. c. constrictor resulted in invasive populations elsewhere (e.g., Va´zquez-

TABLE 2. Comparison of effective population size (Ne) and mean genetic diversity statistics of average number of alleles (na), effective number of alleles (ne), observed heterozygosity (Ho), expected heterozygosity (He), unbiased heterozygosity (HNei), and FIS for insular, invasive populations of Boa constrictor constrictor and Boa constrictor imperator (95% confidence interval [CI] in parentheses). These data are from loci Bci14, 15, 18, and 21, the four loci common to all three studies. Values with different superscripts indicate significant differences based upon ANOVA and Tukey’s honestly ns significant difference a posteriori tests; P > 0.05; *P < 0.05; **P < 0.01; N/A, data not available. Means, 95% CIs, and Ne statistics for Cozumel and Puerto Rico from published data (Va´zquez-Domıńguez et al., 2012; Reynolds et al., 2013).

Population Ne na ne Ho He HNei FIS Aruba 15 4.0a 3.01 0.62 0.63 0.64 0.01a n = 22 (3–247) (1.4–6.6) (1.13–4.89) (0.46–0.79) (0.39–0.86) (0.40–0.88) (-0.14 to 0.16) Puerto Rico 16 4.2a 2.74 0.61 0.62 N/A 0.12a n = 32 (5–58) (1.5–7.0) (1.85–3.63) (0.35–0.87) (0.49–0.76) (0.03–0.22) Cozumel 1 256 9.7b 5.26 0.63 0.79 0.80 0.23b n = 30 (88–520) (7.4–12.1) (2.44–8.08) (0.29–0.98) (0.67–0.91) (0.69–0.92) (0.07–0.39) Cozumel 2 749 8.0b 3.96 0.56 0.71 0.72 0.21b n = 46 (140–1,847) (3.0–13.0) (1.95–5.97) (0.40–0.72) (0.51–0.92) (0.51–0.94) (0.05–0.37) ns ns ns ns ANOVA N/A F(3,12) = 7.21** F(3,12) = 3.23 F(3,12) = 0.15 F(3,12) = 2.02 F(2,9) = 1.77 F(3,12) = 4.84* 608 L. M. BUSHAR ET AL.

Domıńguez et al., 2012; Reynolds et al., 2013), lends strong constrictor is likely around 5 yr (Greene, 1983), we would not support to the hypothesis that the invasion resulted from necessarily expect to see genetic effects of inbreeding by the year released or escaped pets. A small number of captive B. 2007, only 8 yr after the first B. constrictor was reported on the constrictor were maintained on the island in the early 1990s (P. island; future studies might assess this, however, if the Barendsen, E. R. de Cuba, C. Raz, pers. comm.). Importantly, at population cannot be eradicated and persists. least two of these snakes, a large adult female and an adult Although B. constrictor might have historically occurred on snake of unknown sex, are thought to have escaped or been Aruba (Quick et al., 2005) and subsequently went extinct, the liberated from outdoor enclosures prior to 1995 (Quick et al., newly arrived boas clearly represent a novel invasion. A 2005; E. R. de Cuba, C. Raz, pers. comm.). growing body of evidence suggests a very small number of We are able to reject the scenario of the present invasion introduced B. constrictor can successfully establish viable resulting from a single gravid female, as we found three populations in new geographic locations. As a result of its high divergent maternal haplotypes and six alleles at locus Bci14. fecundity, broad habitat requirements, and prevalence in the pet Multiple maternal haplotypes could result from the introduction trade, B. constrictor was determined to pose the highest risk as a of a single female and at least two unrelated males; however, we potentially invasive species among 23 species of large constrict- found that at least two haplotypes are represented by multiple ing snakes evaluated by Reed (2005). The B. constrictor individuals of different age classes, indicating a greater population on the island of Aruba was established cryptically, likelihood that multiple females were introduced (as opposed increased at a rapid rate, and has evaded eradication efforts to multiple males introduced from a single maternal line). (Quick et al., 2005) This should serve as a cautionary note for Furthermore, if a single female boa had colonized Aruba, the conservationists wherever habitat conditions exist that may be presence of more than four alleles at a single locus would suitable for the survival of B. constrictor. require multiple paternity. Although breeding aggregations of B. c. occidentalis have been observed (Bertona and Chiaraviglio, Acknowledgments.—This work is dedicated to our friend, the 2003), multiple paternity has not been reported for this species late T. J. M. Wools, for his inspiration, support, and commitment (reviewed in Rivas and Burghardt, 2005). Parthenogenesis has to the conservation of Aruba’s wildlife. We thank A. Curiel, R. been reported for B. c. imperator (Booth et al., 2011); however, it Croes, D. Marquez, I. Zaandam, and all the rangers and staff of is unlikely that parthenogenesis was involved in the establish- Arikok National Park who assisted with this project, especially ment of the current population, because no individual in our S. Franken and A. Gomes. We thank M. E. Avalos, M. A. Miller, sample was homozygous at all five loci and three loci had more B. J. B. Reinert, L. M. Y. Reinert, Q. W. Reinert, and E. D. Wilson than two alleles. for their assistance in sample collection. We are indebted to P. Based upon our genetic data, the most likely scenario for the Barendsen and the staff of the Aruba Veterinaire Dienst for their founding of the population of B. constrictor on Aruba is that a interest and support of this project. E. R. de Cuba, DVM, of the small number of founding individuals (likely including at least Veterinaire Klinieken Aruba, and C. Raz provided important three unrelated reproductive females) were introduced to the historical information regarding captive boas on Aruba. Our island. These factors lead to this conclusion: 1) three different work on Aruba was conducted with the permission of the mtDNA haplotypes were identified in the population; 2) the Directie Landbouw, Veeteelt, Visseru, en Markthallen (LVV), total number of alleles was low, suggesting a small number of and we thank T. Damian (Director), P. Portier, and F. Franken for founders (similar to Puerto Rico), but not low enough to result their consistent help in so many ways in making our efforts from a single female; 3) most measures of genetic variation successful. Funding and support for this project was provided (effective number of alleles, observed and expected heterozy- by the American Association of Zoological Parks and Aquaria, gosity) were moderate to low and similar to that of the Puerto Valero Refining Company–Aruba, MetaCorp (particularly R. Rican population; 4) the N estimates for Aruba and Puerto Rico e Timmer), the University of Massachusetts Boston, the Museum were on the lower end of values compared to Cozumel, and in of Comparative Zoology at Harvard University, the Toledo the case of biological invasions, Ne provides an approximation Zoological Society, Arcadia University, Sam Houston State of the effective number of breeders, Nb, in the parental generation (Waples, 2006; Reynolds et al., 2013); 5) the mean University, and The College of New Jersey. Treatment of all pairwise relatedness (0.14; SE = 0.009) was similar to that animals followed the Guidelines for Use of Live Amphibians reported by Reynolds et al. (2013) for the invasive Puerto Rican and Reptiles in Field Research (compiled by the American population (0.16, SE = 0.007, n = 32) and in both cases Society of Ichthyologists and Herpetologists, The Herpetolo- approached half-sibship (r = 0.25). In contrast, relatedness (r)on gists’ League, and the Society for the Study of Amphibians and Cozumel was reportedly somewhat lower, at least for the Reptiles; http://www.asih.org/publications) and were ap- Cozumel 1 population (Cozumel 1: -0.01, SE = 0.034, n = 30; proved by the Institutional Animal Care and Use Committee Cozumel 2: 0.12, SE = 0.034, n = 46; Va´zquez-Domıńguez et al., (IACUC) of The College of New Jersey (Protocol 0708-002HR1). 2012). Taken together, these findings suggest the number of All samples were imported to the United States with approvals founding individuals was at least three, but likely far fewer than from the Aruba Veterinaire Dienst and the United States Fish the number responsible for the Cozumel invasion (n 30). The and Wildlife Service. first occurrences of B. constrictor on Aruba were not localized, but distributed across the island (Quick et al., 2005), suggesting that either introductions occurred at multiple sites on Aruba or LITERATURE CITED the population remained cryptic for an extended period of time. The latter scenario is unlikely given the level of development ASCUNCE, M. S., C. C. YANG,J.OAKEY,L.CALCATERRA,W.J.WU,C.J.SHIH, J. GOUDET,K.G.ROSS, AND D. SHOEMAKER. 2011. Global invasion and human population density throughout most of the island. history of the fire ant Solenopsis invicta. Science 331:1066–1068. We found no evidence for inbreeding in the Aruba population BERTONA, M., AND M. CHIARAVIGLIO. 2003. Reproductive biology, mating of B. constrictor. Because the average generation time for B. aggregations, and sexual dimorphism of the Argentine Boa INVASIVE BOA CONSTRICTOR ON ARUBA 609

Constrictor (Boa constrictor occidentalis). Journal of Herpetology 37: LARKIN,M.A.,G.BLACKSHIELDS,N.P.BROWN,R.CHENNA,P.A. 510–516. MCGETTIGAN,H.MCWILLIAM,F.VALENTIN,I.M.WALLACE,A.WILM, BOBACK, S. M. 2005. Natural history and conservation of island boas (Boa R. LOPEZ,J.D.THOMPSON,T.J.GIBSON, AND D. G. HIGGINS. 2007. Clustal constrictor) in Belize. Copeia 2005:879–884. W and Clustal X version 2.0. Bioinformatics 23:2947–2948. BOOTH, W., D. H. JOHNSON,S.MOORE,C.SCHAL, AND E. L. VARGO. 2011. MACKESSY, S. P., M. N. SIXBERRY,W.H.HEYBORNE, AND T. FRITTS. 2006. Evidence for viable, non-clonal but fatherless Boa constrictors. Biology Venom of the Brown Treesnake, Boiga irregularis: ontogenetic shifts Letters 7:253–256. and taxa-specific toxicity. Toxicon 47:537–548. BROOKFIELD, J. F. Y. 1996. A simple new method for estimating null allele MARTIŃEZ-MORALES, M. A., AND A. D. CUAROŃ. 1999. Boa constrictor,an frequency from heterozygote deficiency. Molecular Ecology 5:453– introduced predator threatening the endemic fauna on Cozumel 455. Island, Mexico. Biodiversity and Conservation 8:957–963. BURBRINK, F. T., R. LAWSON, AND J. B. SLOWINSKI. 2000. Mitochondrial DNA PEAKALL, R., AND P. E. SMOUSE. 2006. GENALEX 6: genetic analysis in phylogeography of the polytypic North American Rat Snake (Elaphe Excel. Population genetic software for teaching and research. obsoleta): a critique of the subspecies concept. Evolution 54:2107– Molecular Ecology Notes 6:288–295. 2118. PEAKALL, R., AND P. E. SMOUSE. 2012. GenAlEx 6.5: genetic analysis in BUSHAR, L. M., H. K. REINERT, AND L. GELBERT. 1998. Genetic variation and Excel. Population genetic software for teaching and research—an gene flow within and between local populations of the Timber update. Bioinformatics 28:2537–2539. Rattlesnake, Crotalus horridus. Copeia 1998:411–422. PYSˇEK,P.,AND D. M. RICHARDSON. 2010. Invasive species, environmental CHIARAVIGLIO,M.,M.BERTONA,M.SIRONI, AND S. LUCINO. 2003. change and management, and health. Annual Review of Environ- Intrapopulation variation in life history traits of Boa constrictor mental Resources 35:25–55. occidentalis in Argentina. Amphibia Reptilia 24:65–74. QUICK, J. S., H. K. REINERT,E.R.DECUBA, AND R. A. ODUM. 2005. Recent DARRIBA, D., G. L. TABOADA,R.DOALLO, AND D. POSADA. 2012. jModelTest occurrence and dietary habits of Boa constrictor on Aruba, Dutch 2: more models, new heuristics and parallel computing. Nature West Indies. Journal of Herpetology 39:304–307. Methods 9:772. RAMBAUT, A., M. A. SUCHARD,D.XIE, AND A. J. DRUMMOND. 2013. Tracer DODD, C. K., JR. 1986. Importation of live snakes and snake products into v1.5. Available from: http://beast.bio.ed.ac.uk/Tracer. the United States, 1977–1983. Herpetological Review 17:76–79. RAYMOND, M., AND F. ROUSSET. 1995. GENEPOP Version 1.2: population DORCAS, M. E., J. D. WILLSON,R.N.REED,R.W.SNOW,M.R.ROCHFORD,M. genetics software for exact tests and ecumenicism. Journal of A. MILLER,W.E.MESHAKA JR., P. T. ANDREADIS,F.J.MAZZOTTI,C.M. Heredity 86:248–249. ROMAGOSA, AND K. M. HART. 2011. Severe mammal declines coincide REED, R. N. 2005. An ecological risk assessment of nonnative boas and with proliferation of invasive Burmese Pythons in Everglades pythons as potentially invasive species in the United States. Risk National Park. Proceedings of the National Academy of Science Analysis 25:753–766. 109:2418–2422. REYNOLDS, R. G., A. R. PUENTE-ROLOŃ,R.N.REED, AND L. J. REVELL. 2013. DOUGLAS, D. A., A. JANKE, AND U. ARNASON. 2006. A mitogenomic study Genetic analysis of a novel invasion of Puerto Rico by an exotic on the phylogenetic position of snakes. Zoologica Scripta 35:545–558. constricting snake. Biological Invasions 15:953–959. DRUMMOND, A. J., M. A. SUCHARD,D.XIE, AND A. RAMBAUT. 2012. Bayesian REYNOLDS, R. G., M. L. NIEMILLER, AND L. J. REVELL. 2014. Toward a Tree-of- phylogenetics with BEAUTi and the BEAST 1.7. Molecular Biology Life for the boas and pythons: multilocus species-level phylogeny and Evolution 29:1969–1973. with unprecedented taxon sampling. Molecular Phylogenetics and FITZPATRICK, B. M., J. A. FORDYCE,M.L.NIEMILLER, AND R. G. REYNOLDS. Evolution 71:201–213. 2012. What can DNA tell us about biological invasions? Biological RIVAS, J. A., AND G. M. BURGHARDT. 2005. Snake mating systems, behavior, Invasions 14:245–253. and evolution: the revisionary implications of recent findings. GAUTSCHI, B., A. WIDMER,J.JOSHI, AND J. C. KOELLA. 2002. Increased Journal of Comparative Psychology 119:447–454. frequency of scale anomalies and loss of genetic variation in serially ROMERO-NA´ JERA,A.D.CUAROŃ, AND C. GONZA´ LES-BACA. 2007. Distribu- bottlenecked populations of the Dice Snake, Natrix tessellata. tion, abundance, and habitat use of introduced Boa constrictor Conservation Genetics 3:235–245. threatening the native biota of Cozumel Island, Mexico. Biodiversity GLAUBITZ, J. 2004. Convert: a user-friendly program to reformat diploid and Conservation 16:1183–1195. genotypic data for commonly used population genetic software ROUSSET, F. 2008. Genepop’007: a complete reimplementation of the packages. Molecular Ecology Notes 4:309–310. Genepop software for Windows and Linux. Molecular Ecology GREEN, S. 2011. Ecology, conservation and commercial exploitation of the Resources 8:103–106. ‘Hog Island’ Boa constrictor in the Cayos Cochinos, Honduras. SAKAI, A. K., F. W. ALLENDORF,J.S.HOLT,D.M.LODGE,J.MOLOFSKY,K.A. Unpubl. Ph.D. diss., University of Kent, United Kingdom. WITH,S.BAUGHMAN,R.J.CABIN,J.E.COHEN,N.C.ELLSTRAND,D.E. GREENE, H. W. 1983. Boa constrictor (boa, be´quer, boa constrictor). Pp. MCCAULEY,P.O’NEIL,I.M.PARKER,J.N.THOMPSON, AND S. G. WELLER. 380–382 in D. H. Janzen (ed.), Costa Rican Natural History. 2001. The population biology of invasive species. Annual Review of University of Chicago Press, USA. Ecology and Systematics 32:305–332. GUICKING, D., R. A. GRIFFITHS,R.D.MOORE,U.JOGER, AND M. WINK. 2006. SAVIDGE, J. A. 1991. Population characteristics of the introduced Brown Introduced alien or persecuted native? Resolving the origin of the Treesnakes (Boiga irregularis) on Guam. Biotropica 23:294–300. viperine snake (Natrix maura) on Mallorca. Biodiversity and SIMBERLOFF, D. 2009. The role of propagule pressure in biological Conservation 15:3045–3054. invasions. Annual Review of Ecology, Evolution, and Systematics GUINDON, S., AND O. GASCUEL. 2003. A simple, fast and accurate method 40:81–102. to estimate large phylogenies by maximum-likelihood. Systematic SNOW, R. W., K. L. KRYSKO,K.M.ENGE,L.OBERHOFER,A.WARREN-BRADLEY, Biology 52:696–704. AND L. WILKINS. 2007. Introduced populations of Boa constrictor GUO,S.W.,AND E. A. THOMPSON. 1992. Performing the exact test of (Boidae) and Python molurus bivittatus (Pythonidae) in southern Hardy-Weinberg proportion for multiple alleles. Biometrics 48:361– Florida. Pp. 416–438 in R. W. Henderson and R. Powell (eds.), 372. Biology of the Boas and Pythons. Eagle Mountain Publishing, USA. HALDANE, J. B. S. 1954. An exact test for randomness of mating. Journal SOKAL, R. R., AND F. J. ROHLF. 2012. Biometry, 4th ed. W. H. Freeman, of Genetics 52:631–635. USA. HENDERSON,R.W.,AND R. POWELL. 2009. Natural History of West Indian SUA´ REZ-ATILANO, M., F. T. BURBRINK, AND E. VA´ ZQUEZ-DOMIŃGUEZ. 2014. Amphibians and Reptiles. University Press of Florida, USA. Phylogeographic structure within Boa constrictor imperator across the HYNKOVA´ , I., Z. STAROSTOVA, AND D. FRYNTA. 2009. Mitochondrial DNA lowlands and mountains of Central America and Mexico. Journal of variation reveals recent evolutionary history of main Boa constrictor Biogeography, doi: 10.1111/jbi.12372. clades. Zoological Science 26:623–631. VAN BUURT, G. 2011. Conservation of amphibians and reptiles in Aruba, KALINOWSKI, S. T., A. P. WAGNER, AND M. L. TAPER. 2006. ML-Relate: a Curacao and Bonaire. Pp. 145–159 in A. Hailey et al. (eds.), computer program for maximum likelihood estimation of related- Conservation of Caribbean Island Herpetofauna. Volume 1: Conser- ness and relationship. Molecular Ecology Notes 6:576–579. vation Biology and the Wider Caribbean. Brill, Netherlands. KRAUS, F. 2009. Alien Reptiles and Amphibians: A Scientific Compen- VAN OOSTERHOUT, C., W. F. HUTCHINSON,D.P.M.WILLS, AND P. S HIPLEY. dium and Analysis. Springer, Netherlands. 2004. Micro-Checker: software for identifying and correcting 610 L. M. BUSHAR ET AL.

genotyping errors in microsatellite data. Molecular Ecology Notes 4: WAPLES, R. S., AND C. DO. 2008. LDNe: a program for estimating effective 535–538. population size from data on linkage disequilibrium. Molecular VA´ ZQUEZ-DOMI´NGUEZ, E., M. SUA´ REZ-ATILANO,W.BOOTH,C.GONZA´ LEZ- Ecology Resources 8:753–756. O´ BACA, AND A. D. CUAR N. 2012. Genetic evidence of a recent WEIR, B. S. 1990. Genetic Data Analysis. Methods for Discrete Population successful colonization of introduced species on islands: Boa Genetic Data. Sinauer Associates, USA. constrictor imperator on Cozumel Island. Biological Invasions 14: WEIR, B. S., AND C. C. COCKERHAM. 1984. Estimating F-statistics for the 2101–2116. analysis of population structure. Evolution 38:1358–1370. WAPLES, R. S. 2006. A bias correction for estimates of effective population size based on linkage disequilibrium at unlinked gene loci. Conservation Genetics 7:167–184. Accepted: 16 November 2014.