Testing for hybridisation of the Critically Endangered Iguana delicatissima on Anguilla to inform conservation efforts

The Caribbean Island of Anguilla in the north-eastern Lesser Antilles is home to one of the last populations of the Critically Endangered Lesser Antillean iguana Iguana delicatissima. This population is highly threatened primarily because of hybridisation with non-native Iguana iguana. This study assesses the degree of hybridisation between Anguilla’s Iguana species firstly using morphological characteristics and then genetic analysis to validate the genetic integrity of morphologically identified I. delicatissima. We also examined the genetic diversity of Anguilla’s I. delicatissima population, and that of a population on the nearby island of Îlet Fourchue, St Barthélemy. Forty-five iguanas were captured in Anguilla and 10 in St Barthélemy, and sequences from 3 nuclear and 1 mtDNA genes were obtained for each. Of the 45 iguanas captured in Anguilla, 22 were morphologically identified as I. delicatissima, 12 as I. iguana and the remainder were identified as hybrids. Morphological assignments were all confirmed by genetic analyses except for one I. iguana and one hybrid individual. These two individuals appeared likely to have originated following ancestral hybridisation events several generations ago. A significant paucity of genetic diversity was found within Anguillan and St Barthélemy I. delicatissima populations, with a single haplotype being identified for each of the three nuclear genes and the mtDNA sequence. This study highlights the urgency for immediate action to conserve Anguilla’s remnant I. delicatissima population. Protection from hybridisation will require translocation to I. iguana-free offshore cays, with supplementary individuals being sourced from neighbouring islands to enhance the genetic diversity of the population.


Introduction
West Indian iguanas are considered the most endangered lizards in the world (Alberts et al. 2004). Two genera of the Iguanidae, Cyclura and Iguana, are native to the West Indies: the genus Cyclura is restricted to the Bahamas and Greater Antilles, while the genus Iguana is native to the Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s1059 2-020-01258 -6) contains supplementary material, which is available to authorized users.

3
Lesser Antilles and parts of Central and South America. Two species of Iguana are currently recognised: the Critically Endangered Lesser Antillean iguana Iguana delicatissima (van den Burg et al. 2018), the green or common iguana Iguana iguana (classified by IUCN as Least Concern Bock 2018). The latter includes several native subspecies, including Iguana iguana insularis (first described and evaluated as Vulnerable by Breuil et al. 2019) and Iguana iguana sanctaluciae (described and evaluated as Critically Endangered, Breuil et al. 2019).
Historically, I. delicatissima is believed to have existed throughout the northern Lesser Antilles, from Anguilla to Martinique, having been recorded from Anguilla, Saint-Martin/Sint Maarten, St Barthélemy, St Eustatius, St Kitts and Nevis, Antigua and Barbuda, Guadeloupe, Dominica, and Martinique (Fig. 1). However, its current range is much reduced and the conservation status of I. delicatissima has recently been elevated on the IUCN Red List to Critically Endangered due to its restricted and severely fragmented distribution. These factors, along with reports from 2017 of the first sightings of I. iguana on the island of Dominica, which was considered one of the most secure strongholds of I. delicatissima, have highlighted the extreme vulnerability of this species (van den Burg et al. 2018).
Iguana delicatissima is threatened across its range by several factors including habitat loss and fragmentation, particularly in recent years by coastal development for tourism (Debrot et al. 2013) and extensive over-browsing by livestock that causes a shift in plant species composition and habitat structure (Breuil 2002). Predation by introduced invasive species (rats, cats and mongooses) is also a major threat (Barun et al. 2011;Medina et al. 2011). For example, on islands where Asian mongoose Herpestes auropunctatus have been introduced I. delicatissima is either extinct or highly threatened (Breuil 2002). However, the most widely reported threat facing I. delicatissima is competition and hybridisation with I. iguana (Vuillaume et al. 2015). In recent years I. iguana has been expanding its range northwards through the Lesser Antilles and northern Caribbean (Brisbane 2018;Lopez-Torres et al. 2010;van den Burg et al. 2019), with invasions of this species being assisted by tropical storms (Censky et al. 1998), hurricanes, ship traffic between islands (which have inadequate biosecurity measures) and the pet trade (van den Burg et al. 2018). Hybridisation between I. delicatissima and I. iguana has been confirmed through both molecular and morphometric analyses from Guadeloupe and all other main islands across their range (Day and Thorpe 1996;Vuillaume et al. 2015;van den Burg et al. 2018van den Burg et al. , 2019Brisbane 2018). Post-invasion displacement of I. delicatissima is rapid and population extirpations due primarily to the presence of I. iguana have been recorded from several islands in the French West Indies (van den Burg et al. 2018;Breuil 2013).
Based on the range of threats, recent predictions suggest that within one generation, five of the eight remaining populations of I. delicatissima plus four of the currently invaded/ hybridised locations will be extirpated (van den Burg et al. 2018). In addition, only 13% of the species' current estimated area of occupancy is predicted to remain three generations from now (van den Burg et al. 2018). Eradicating introduced and established I. iguana populations is probably unrealistic in most cases, both logistically and financially (Radford 2017). As such, recent conservation efforts have focused on protecting populations in sites that have not yet been colonised by I. iguana. These efforts include translocation from mainland sites to more easily protected offshore cays with no human or I. iguana populations. This conservation strategy has been successful in the French Overseas Territory of St Barthélemy where, in 2011, 14 individuals were translocated from the main island to augment the Îlet Fourchue population (reported as a few lone individuals) (Questel 2019 Day and Subin (1997) indicated that the I. delicatissima population had a restricted range on Anguilla, occurring in small pockets along the northern coast (Fig. 2). The study reported that Anguilla's I. delicatissima population was below the environmental carrying capacity and that intervention would be required to prevent the local extinction of the species (see also Gerber 1999). Adding further pressure to I. delicatissima on Anguilla, an unknown number of I. iguana (and/or I. iguana × I. delicatissima hybrids) unexpectedly appeared on mainland Anguilla following Hurricane Luis in September 1995, having been dispersed on flotsam (Censky et al. 1998). Questionnaire surveys of residents of Anguilla, conducted in January 2015, and subsequent captures and sightings of I. iguana, have revealed a rapid colonisation of I. iguana across Anguilla's mainland over the last 23 years, with this species now widespread over much of the island (Mukhida personal observation). With the increase and spread of the alien population of I. iguana it is plausible that any new offspring from the remnant declining population of I. delicatissima on Anguilla's mainland represent hybrids from recent matings with the invasive I. iguana.
Unique morphological characteristics have been defined for both I. delicatissima and I. iguana (Breuil 2013) and the use of several diverse characters is known to be advantageous when identifying hybrids within iguana populations (Vuillaume et al. 2015). While it has been alluded to (Martin et al. Vuillaume et al. 2015) it is unknown whether, when using Breuil (2013) detailed morphological descriptions for identification, "cryptic hybrids" may also be present in the population that bear the diagnostic characteristics of only one species. This study assesses the degree of hybridisation between Anguilla's Iguana species, firstly using morphological characteristics, and then using genetic analysis to test the genetic integrity of morphologically-identified I. delicatissima. This work was conducted with the aim of informing conservation strategies such as translocating I. delicatissima to I. iguana-free offshore cays. With this purpose in mind, we also analysed the genetic integrity of a growing population of previously translocated I. delicatissima individuals captured on the I. iguana-free island of Îlet Fourchue one of St Barthélemy uninhabited offshore cays, and the closest neighbouring I. delicatissima population to Anguilla.

Field methods
Between May 2015 and May 2018, comprehensive surveys for I. delicatissima were conducted along the northern coast of Anguilla (reflecting previously reported known distributions of this species (Gerber 1999) (Fig. 2). Surveyors attempted to capture all iguanas encountered using a noose or net or by hand. Captured animals were weighed, Fig. 2 The most northerly distributed populations of Iguana delicatissima located on Anguilla and St Barthélemy morphological measurements and photographs were taken, and those individuals deemed to be I. delicatissima or those exhibiting any of the 12 defined I. delicatissima morphological characteristics (see Breuil 2013) were temporarily housed in a specifically designed captive facility. The presence of I. delicatissima or I. iguana states for each of the 12 morphological characters was recorded for each iguana using a checklist ("Appendix"). In addition to surveying the previously reported sites for I. delicatissima, field surveys were undertaken ad hoc in a range of sites across the island, and whenever a member of the public reported an iguana, field staff would visit to identify and attempt capture. In addition to the surveys undertaken on Anguilla, during a 2-day visit in June 2018, 10 iguanas morphologically identified as pure I. delicatissima were captured using a stick with a noose or directly by hand on Îlet Fourchue (17°57'23.59"N, 62°54'5.70"W), St Barthélemy ( Fig. 2).
For all iguanas captured, the end of the tail was sanitised with 70% alcohol, and 1-2 cm was removed from the tip with scissors. This tissue was immediately scored with a sanitised blade and cut into smaller segments before being stored in a 3 ml cryovial containing DNAgard® (Sigma Aldrich, UK). Samples were shipped to the University of Liverpool, UK for genetic analysis with permission from the Government of Anguilla. Samples were analysed to test for the presence of I. iguana haplotypes circulating in this population and to compare the genetic diversity of populations in Anguilla and St Barthélemy.

Gene sequencing and analyses
Genomic DNA was extracted from approximately 10 mg iguana tail segments using the DNeasy Blood and Tissue Kit (Qiagen, UK) according to the manufacturer's instructions. PCR was used to amplify partial sequences as previously described for 3 nuclear genes and 1 mitochondrial gene: 572 bp oocyte maturation factor Mos (C-mos) (Noonan and Chippindale 2006), 498 bp Neurotrophin-3 segment (NT3) (NT3-F3 and NT3-R4: Noonan and Chippindale 2006), 569 bp polymerase alpha catalytic subunit (PAC) (PACs1 and PACs2: Pasachnik et al. 2009) and 801 bp NADH dehydrogenase subunit 4 (ND4) (a maternally inherited marker used to elucidate the direction of crosses Martin et al. 2015). PCR products were verified by gel electrophoresis and amplicons purified using Qiagen PCR purification kits (Qiagen, UK). DNA sequencing of the purified products was performed by GATC Biotech (GATC Biotech, Germany). Unique sequences were deposited in GenBank.
Nucleotide sequences from all specimens and available published Iguanidae sequences were aligned for all loci using MEGA7 (Kumar et al. 2016). Accession numbers for all sequences retrieved from GenBank are shown next to taxonomic names on the phylogenetic tree. Sequence identities were compared using Geneious 11.0.5 (http://www. genei ous.com, Kearse et al. 2012). DNA substitution models were selected based on likelihoods, using the Akaike corrected information criterion in MEGA7. Maximum likelihood phylogenetic trees were generated using these models in MEGA7 (Kumar et al. 2016) with node support based on 10,000 bootstrap replicates (Felsenstein 1985). Further, haplotype networks were constructed using PopART version 1.7 (Leigh and Bryant 2015) using the TCS method (Clement et al. 2002), in order to visualise instances of hybridisation. Finally, the ability for the genetic markers employed in this study to categorise species and hybrids within a population, was ascertained by comparing our data to species specific loci identified from currently published sequences.
A Bayesian analysis of population structure (BAPS) v.6.0 (Corander et al. , 2008Cheng et al. 2011) was employed to identify existing genetic groups among the individuals sampled for each of the four genes utilised and in order to identify putative hybrids. The haplotypes for 55 unclassified [Anguilla (n = 45) and St Barthélemy (n = 10)] and two reference individuals were analysed with no priori assumption. An alignment of all published sequences of each Iguana species was used to identify all species specific sites in order to generate a consensus for each reference individual. Our test population was assumed to be comprised of two species, therefore the number of clusters was set to two (K = 2). An admixture analysis was performed after the mixture analysis to calculate the genetic mixture of each individual from the genetic groups found (Corander et al. , 2008. The analysis was run with 1000 iterations.

Sampling and identification
On Anguilla, approximately 600 h of survey effort over the previously reported range of I. delicatissima (Fig. 2) resulted in the capture of 26 individuals (a further 8 individuals with morphological characteristics of I. delicatissima were sighted but evaded capture). Of those captured, 22 individuals exhibited all morphological characteristics of I. delicatissima, three exhibited morphological characteristics of both I. delicatissima and I. iguana, and 1 appeared morphologically to be I. iguana (Table 1). In addition, 12 individuals that were morphologically diagnosed as I. iguana, and 7 that exhibited at least 1 characteristic of both I. delicatissima and I. iguana (Table 1) were caught from a range of locations distributed widely across the main island of Anguilla (Fig. 3). Based on the 12 morphological features (Table 1), 9 of the 10 morphologically identified hybrids appeared most similar to I. iguana and 1 most similar to I. delicatissima.
Only 1 of the 12 morphological characteristics that were diagnostic for I. iguana was present in all hybrids: the presence of black bands on the tail (Table 1). In contrast, no single diagnostic feature of I. delicatissima was consistently observed in all of the hybrids, although presence of a bumpy head (but not necessarily the presence of occipital bumps) was a common I. delicatissima feature observed in these individuals and recorded in 9 of the 10 morphologically identified hybrids (Table 1, "Appendix").

DNA analyses
BAPS demonstrated a clear pattern of grouping all tested individuals into two distinct genetic clusters with high probability (p < 0.0001 for all). Animals morphologically assigned as I. delicatissima clustered together, as did all individuals designated I. iguana. Clustering differences between genes were only evident within individuals classified as hybrids (Fig. 4).

C-mos
All sequences that originated from the 22 individuals with I. delicatissima morphologies were identical and represent a currently unreported haplotype IdANG (MK605403). The haplotype showed 99.7% similarity to that previously reported for Anguilla, and neighbouring islands of Dominica and St Eustatius (HM352538; Stephen et al. 2013). Hybrid D1 was notable for solid bands on the tail, a feature commonly reserved for I. iguana, provided identical (homozygous) sequences clustering along with the previously reported Anguillian I. delicatissima ( Fig. 5a; Table 2; Supplementary Fig. 1a). All individuals morphologically diagnosed as I. iguana were most similar (99.7%) to previously identified I. iguana C-mos sequences from individuals from Belize, Brazil, Curacao, El Salvador, Honduras, Mexico, St Lucia and Suriname (Stephen et al. 2013). Three individuals that were diagnosed as hybrids (GR9, GR11, GR19) were heterozygous at two interspecific sites located across the partial sequences, demonstrating the presence of haplotypes of both I. iguana and I. delicatissima (Table 2). The remaining  Table 2; Supplementary Fig. 1a). Fig. 4 Bayesian analysis of population structure (BAPS) clustering for each of the partial nuclear genes, a C-mos, b NT3, c PAC, and the mitochondrial gene, d ND4. Analyses were conducted on 57 Iguana individuals (55 unclassified and 2 reference) under the assumption of 2 genetic clusters (K = 2) present within the population (I iguana and I. delicatissima). Vertical bars represent the genotypes that have been apportioned to cluster 1 (I. delicatissima) and cluster 2 (I. iguana), red and green respectively. Hybrid order-D1, GR6, GR9, GR10, GR11, GR15, GR16, GR17, GR18, GR19, GR22

NT3
All 22 morphologically assigned I. delicatissima, along with hybrid D1 were 100% identical across the partial NT3 gene, and identical to previously reported haplotypes (HM352528 and MF191675; Stephen et al. 2012 and van den Burg et al. 2018 respectively), isolated from individuals on Anguilla, Dominica and St Eustatius. None of the sequences obtained from these individuals showed any polymorphic sites when aligned together (Table 2). All sequences from morphologically identified I. iguana as well as eight of the hybrids clustered with previously reported I. iguana sequences from Mexico, Caribbean, as well as Central and South America (Stephen et al. 2012), with evidence suggesting the presence of at least three I. iguana NT3 haplotypes on Anguilla, CA1, SA1 and ANG reported here (HM352529, HM352530 and MK605401, respectively) ( Fig. 5b; Supplementary Fig. 1b). Six of these individuals including hybrid GR6 and GR18 were heterozygous for two I. iguana NT3 haplotypes (SA1 and Ang) ( Table 2). GR9 and GR11 were the only two hybrids identified genetically, clustering with I. delicatissima as a result of heterozygosity at three interspecific sites across the partial regions (  interspecific sites and was thus identical to previously reported I. iguana haplotype IihapSA7 (JN811116) and clustered along with other I. iguana sequences rather than I. delicatissima ( Fig. 5c; Table 2; Supplementary Fig. 1c). The absence of I. delicatissima sequence suggests this individual was not a first generation (F1) hybrid. All I. iguana, along with hybrids that possessed primarily I. iguana features, clustered among other I. iguana PAC sequences ( Fig. 5c; Supplementary Fig. 1c) (Stephen et al. 2012). However, the presence of several haplotypes within the I. iguana population on Anguilla was notable relative to other genes.

ND4
The ND4 gene was identical for all 22 morphologically elected I. delicatissima and hybrid D1, showing that they carried I. delicatissima mtDNA ( Fig. 5d; Table 2; Supplementary Fig. 2). The samples from Anguilla were identical to previously reported I. delicatissima haplotype 1 and A (AF217783 and KJ561221; Malone et al. 2000 andMartin et al. 2015 (Table 2). This suggests that the direction of cross is predominantly male I. iguana mating with I. delicatissima females (Table 2).

Genetic identity of St Barthélemy iguanas
All 10 I. delicatissima individuals from St Barthélemy were found to have identical haplotypes to those circulating in the Anguillan I. delicatissima population for the three nuclear loci. Two ND4 haplotypes were identified with three individuals being identical to the Anguillan I. delicatissima haplotype, and seven identical to haplotype E previously reported from St Barthélemy (Martin et al. 2015). No indication of hybridisation was reported in the iguana's sampled from St Barthélemy.

Discussion
This present study evaluated the genetic integrity of the I. delicatissima population on Anguilla. We demonstrated that the remnant population on the Anguilla mainland has already hybridised with the introduced I. iguana, with evidence of hybridisation in 12% of individuals captured in the last remaining pocket of habitat containing I. delicatissima. In addition, there is evidence of hybrid matings occurring more than one generation ago. Our work further highlights that hybridisation with I. iguana is a major threat facing I. delicatissima populations and is of even wider significance in light of recent reports of I. iguana hybridizing with other native and endangered Iguana and Cyclura populations in the region (Moss et al 2018).

Morphological and genetic comparison
Two mismatches between morphological and genetic identity were identified. One individual (GR6) was morphologically identified as a hybrid due to one I. delicatissima feature (presence of rounded scales and a row of oval scales between the labials and the sub-labials) but clustered with I. iguana in all four genes analysed and could not be identified genetically as a hybrid. In contrast, GR10 was morphologically identified as I. iguana but genetically was found to have I. delicatissima mtDNA. However, this individual was a subadult so may not have yet developed all of the distinguishing morphological features. These two exceptions, which constitute 7.7% of the sampled animals, indicate that the ancestral hybridisation event(s) occurred several generations ago, which could explain discrepancies in morphological identification, and highlights why morphological identification should not be relied upon on its own, particularly if several generations have passed since the invasion of I. iguana and in the identification of juveniles. Whilst several hybrids were identified on Anguilla through genetic analysis, none demonstrated evidence of hybridisation in all of the three nuclear loci. Two hybrids were identified at two of these loci, one by two genes, and seven in just one gene (Table 2). This discordance among the genes analysed suggests that none of the 10 hybrids genetically identified are first generation hybrids and directly points to more distant hybridisation in their ancestry, and in fact the hybridisation event could have even taken place prior to its ancestors arriving on Anguilla (Censky et al. 1998). It is also worth noting that ancestral polymorphisms and incomplete lineage sorting can disrupt phylogenetic inference and may contribute to differences observed (Maddison 1997;Steiner et al. 2012). In the future, next generation technologies such as RADseq could be utilised to amplify thousands of loci from across the genome to further explore and provide insights into the relatedness and ancestry of individuals.

Genetic diversity of Anguillan iguana populations
It is apparent from the genetic markers used in this study that there is a significant paucity of genetic diversity within the Anguillan I. delicatissima population with only one haplotype being identified in each of the four loci. This is perhaps not surprising as other studies have demonstrated a lack of unique haplotypes between geographically isolated populations (Day and Thorpe 1996;Martin et al. 2015;Stephen et al. 2012), and has also been reported in the populations on St Eustatius, Dominica and the French West Indies (Judson et al. 2018;van den Burg et al. 2016;Vuillaume et al. 2015). Other genetic markers such as microsatellites could be informative in understanding the structure and genetic diversity of a population, as well as estimating effective populations sizes (Valette et al. 2013;van den Burg et al. 2016), but this was not an objective of the current study.
In contrast to the low genetic diversity observed in Anguilla's I. delicatissima population, the I. iguana population exhibits more variation and indicates a larger founder population (Crawford and Whitney 2010). This is not surprising as I. iguana are reported to have arrived into Anguilla through several routes, such as being transported on hurricane debris (Censky et al. 1998) and by boat traffic from the neighbouring island of Saint Martin/Sint Maarten which lies only 12 km from Anguilla (Mukhida personal observation).

Conservation implications
The genetic identification of hybrid iguanas, and the discovery of both hybrids and I. iguana in the remaining range of I. delicatissima on the Anguilla mainland highlights the urgency of taking immediate conservation action to ensure the continued survival of I. delicatissima on Anguilla, and in maintaining this species range across the Lesser Antilles. Since 2016 any individual identified both morphologically and genetically as I. delicatissima has been translocated to one of Anguilla's uninhabited and I. iguana-free offshore cays: Prickly Pear East (18° 15′ 60.00″ N, − 63° 10′ 60.00″ W) (Fig. 2), which had been previously identified as a suitable site for reintroduction (Day and Subin 1997;Mukhida and Soanes 2016). Population monitoring and biosecurity (to ensure I. iguana do not reach the island) will continue to monitor the progress of this translocation and, following a regional I. delicatissima conservation planning meeting in March 2018 (Mukhida et al. 2018), discussions have been taking place regarding the translocation of further individuals from sites that support larger populations.
According to Franklin's (1980) somewhat controversial 50/500 rule, an effective population size (defined as a measure of the genetic behaviour of a population relative to that of an 'ideal' population) of under 50 individuals is at high risk of extinction. Other authors have suggested this number should be higher (Reed and Bryant 2000). While we cannot be sure how many individuals would provide a sustainable population size on Prickly Pear East, it is clear that the founders of this newly reintroduced population will need to be supplemented with additional bloodlines to ensure long-term survival (IUCN/ SSC 2013). The analysis of samples from St Barthélemy has highlighted the growing population on Îlet Fourchue as one such potential donor site. The genetic analysis of 10 individuals from this population revealed an almost identical genetic composition to those on Anguilla, and no evidence of hybridisation was found in the 10 individuals sampled. If translocation from Îlet Fourchue is deemed appropriate under international translocation guidelines (IUCN/SSC 2013) further genetic analysis will be undertaken of all individuals due to be translocated to avoid inadvertently introducing hybrids. In addition, before any translocations are undertaken under the auspices of 'genetic rescue' (Bell et al. 2019) it is necessary to assess the genomic heterozygosity of the donor population and to examine the history of population (i.e. have historic bottlenecks been recorded) . The prevalence and the risk of disease should also be thoroughly assessed (Hellebuyck et al. 2017).
This study has further highlighted the urgency of action to preserve Anguilla's remaining I. delicatissima population on offshore islands. Without intervention, the rapid spread of I. iguana on Anguilla's mainland and the level of hybridisation already reported here will rapidly lead to the extirpation of Anguilla's mainland I. delicatissima population.