Conservation Genetics of Two Critically Endangered Island Dwarf Carnivores

Conservation Genetics of Island Dwarf Carnivores

Introduction

• Genetic diversity is crucial for conservation biology and understanding evolutionary processes.
• Islands are ideal natural systems for studying population genetics and evolution due to:
  • Geographic isolation.
  • Small area.
  • Low population sizes.
  • Unique dispersal and colonization processes.
• Oceanic islands harbor unique biota and high endemism, leading to genetic differentiation after establishment from mainland migrants.
• Island populations often face detrimental genetic processes:
  • Bottlenecks.
  • Drift.
  • Inbreeding.
• These processes make island species susceptible to extinction compared to mainland counterparts.
• Human activities (habitat transformation, exotic species introduction) increase extinction risk.
• Mammals are the most endangered insular fauna, with carnivores particularly vulnerable due to their ecological traits.
• Some insular mammals have low genetic diversity, compromising long-term persistence.
• Cozumel Island harbors endemic taxa, including two dwarf carnivore procyonids:
  • Pygmy raccoon (Procyon pygmaeus).
  • Dwarf coati (Nasua nelsoni).
• The pygmy raccoon is the smallest Procyon species with a declining population (323-955 individuals).
  • National protection status: "En Peligro de Extinción" (Threatened with Extinction).
  • IUCN Red List: Critically Endangered.
  • Scarce genetic information reveals low mitochondrial genetic variability.
• Coatis on Cozumel Island date back to ancient Mayan civilization and are considered taxonomically distinct from mainland counterparts (Nasua narica).
  • Dwarf coati was described based on morphometrical study with a small sample size (n=6).
  • Population reductions have occurred, with an estimated population of 150±95 individuals.
  • IUCN: Least Concern (due to subspecies-level status).
  • Threatened by Mexican laws.
• High conservation concerns for Cozumel biota:
  • Exotic species introduction.
  • Habitat fragmentation.
  • Natural resources overexploitation.
• These factors increase the extinction risk of both endemic procyonids.
• The study aims to estimate genetic diversity and structure of the pygmy raccoon and dwarf coati populations, encompassing a wider temporal and spatial scale with comprehensive genetic analyses.
• Molecular markers used:
  • Mitochondrial cytochrome b gene.
  • Bi-parental nuclear microsatellites loci.
• Predictions:
  • Low genetic diversity.
  • Incipient to moderate genetic differentiation.
  • Reduced effective population sizes.
  • Signals of genetic bottlenecks.
• The study reviews the genetic relationship of the two species and their continental counterparts to evaluate if endemic species exhibit unique lineages.
• Specific conservation measures are discussed for their long-term survival.

Methods

• Study area: Cozumel Island (ca. 486 km2), an oceanic island of coral-line origin, never connected to the mainland.
  • Located 17.5 km off the Yucatán peninsula.
  • Separated by the Cozumel Channel (400 m deep).
  • Dominant vegetation: semi-evergreen tropical forest, deciduous tropical forest, and mangroves.
• Samples were obtained from four intensive fieldwork seasons:
  • February–March 2003 (P. pygmaeus).
  • May–August 2006 (N. nelsoni).
  • June–July 2012 (insular and mainland individuals).
  • March–July 2014 (N. nelsoni).
• Tomahawk live traps were used, baited with sardine, banana, and pineapple; traps set at sunset and checked daily.
• Trapped animals were immobilized with Zoletil 100® at a dosage of $10 mg/kg$, sexed, assigned to an age category, and released at the sampling site.
• Tissue samples (ear punches and/or tailsnips) were collected for genetic analyses and stored in 96% ethanol.
• All procedures followed the American Society of Mammalogists guidelines and were conducted with appropriate sampling permits.
• Mainland samples from Mexican states (Morelos, Jalisco, Colima, and Quintana Roo) were included for comparative genetic diversity analysis.
• A total of 150 samples were analyzed:
  • 45 insular P. pygmaeus.
  • 46 N. nelsoni.
  • 31 mainland P. lotor.
  • 28 N. narica.
• Two individuals captured on Cozumel were suspected to be P. lotor.
  • Distinguishing features: larger size, robust complexion, red hues on tail and neck fur.
• DNA extraction: Zymo Research Genomic DNA Tissue MiniPrep kit.
• DNA quantity and quality were assessed with 1% agarose gels stained with ethidium bromide.
• Population genetic analyses for P. pygmaeus were done with microsatellite loci developed for P. lotor.
  • 12 fluorescently labeled primers were tested; four loci were successfully amplified by multiplex reactions.
• Regarding N. nelsoni, 12 fluorescently labeled microsatellite primers developed for N. narica were tested.
• Microsatellite products were run on an ABI Prism 3730xl and 3100 Genetic Analyzer, with ROX-500 as internal size standard, and allele size determined with GeneMarker v2.2.0 software.
• Negative controls were included in all runs, and multiple samples were sized at least twice to assure reproducibility.
• For the mitochondrial evaluation, the cytochrome b (cyt b) gene was amplified using specific primers.
• The cyt b was divided into two fragments (ca. 750 bp each) with an overlap of nearly 300 bp, based on primer pairs.
• Polymerase chain reactions (PCR) were carried out in a $20 µL$ volume with specific reagents and thermocycling conditions.
• PCR products were run on an ABI Prism 3730xl Analyzer.

Genetic Analyses

Nuclear Microsatellite Loci

• Genetic structure for Procyon pygmaeus and Nasua nelsoni was defined using the Bayesian clustering method Structure v2.3.4.
• Values of $K=1–4$ were tested with the admixture and correlated allele frequencies models.
• 10 independent runs were performed for each K value, based on 500,000 Markov chain Monte Carlo iterations and a burn-in of 150,000.
• The best number of genetic clusters (K) was selected using the rate of change in the log probability of data between successive K values.
• Subsequent genetic estimates were performed for each genetic cluster obtained with Structure and for the entire island population.
• Genetic diversity parameters were estimated using R v.3.3.2:
  • Deviations from Hardy–Weinberg equilibrium (HWE) for each locus were estimated with pegas, using an exact test based on 10,000 Monte Carlo permutations.
  • A measure of correlation (rbarD) was calculated for testing overall linkage disequilibrium using poppr.
  • Null alleles were estimated with Brookfield's estimator in PopGenReport.
  • Bonferroni corrections were applied where necessary.
  • The number of observed alleles (Ao), observed (Hobs) and expected heterozygosities (Hexp) were estimated with adegenet.
• To evaluate the degree of genetic differentiation between genetic clusters within the island, FST was estimated using pegas.
• The distribution of genetic variation between and within genetic clusters was analyzed using an analysis of molecular variance (AMOVA) based on FST and a significance non-parametric test with 30,000 permutations.
• The number of migrants (Nm) was evaluated based on the allelic differences between genetic clusters (FST) using Arlequin v3.0.
• Nei’s genetic distance (DN) was estimated with GenAlex v6.5.
• Relatedness among individuals was evaluated using Ml-Relate.
• Genetic signatures of recent bottlenecks were assessed using Bottleneck v5.1.26, comparing observed and expected heterozygosity values estimated under three mutational models:
  • Infinite allele (IAM).
  • Stepwise mutation (SMM).
  • Two-phase (TPM).
• Models were run with 10,000 replicates, and significance was calculated with a Wilcoxon test.
• Additionally, a graphical method based on the distribution of allele frequencies was used.
• The Garza–Williamson index (M) was calculated using MPVal, and the critical value (Mc) was obtained using CRITICAL_M.
• The effective population size (Ne) was estimated with Neestimator v.2, using a Jackknife with LD values between pairs of loci and a random mating system, considering a critical value (Pcrit) of 0.05.
• Microsatellite genetic structure and genetic diversity estimates were also performed for P. lotor and N. narica populations.
• For the Structure analyses, $K=1–6$ were tested with the same parameters described above, based on the mainland and island shared loci.
• The rarefied allelic richness (Ar) was estimated with PopGenReport.
• The degree of genetic differentiation (FST) between insular and mainland populations was assessed using diveRsity.
• A Discriminant Analysis of Principal Components (DAPC) was performed with adegenet.

Cytochrome b sequences

• Fragments of each sequence were assembled with BioEdit and aligned using CLUSTAL W, with P. lotor and N. narica as alignment references.
• Genetic diversity parameters, including the number of haplotypes (Nh), haplotype diversity (h), nucleotide diversity ($\pi$), average number of nucleotide differences (K), and number of polymorphic sites (S), were estimated.
• Neutrality tests were estimated based on Tajima’s D, D and F statistics, and Fu’s Fs, using DNAsp v5.0.
• Phylogenetic relationships between insular and mainland populations were estimated using maximum likelihood (ML) methods.
• jModeltest 2 v0.1.1 was used to select the best-fit model of evolution based on the Akaike Information Criteria (AIC).
• ML analyses were implemented using PhyML v3.1, with the nearest-neighbor interchange algorithm and clade support assessed by an approximate likelihood-ratio test.
• Resulting trees were edited in FigTree v1.4.0.
• Demographic analyses were performed only for the insular populations, consisting on the goodness of fit test between the observed and expected distributions of pairwise differences among sequences.
• The effective population size (Ne) was estimated based on the formula: $Ne = θ / (2μT)$, where μ is the gene mutation rate and T is the generation length in years.

Results

Contemporary population genetics of Cozumel endemic carnivores

Procyon pygmaeus

• Successfully amplified 12 microsatellite loci for 45 P. pygmaeus (excluded two individuals identified as P. lotor).
• Results are described for the entire population (Cozumel) and two genetic clusters, COZ1 and COZ2.
  • Moderate differentiation (FST = 0.105) and some migration (Nm = 4.19).
  • PLOT-09 locus was monomorphic in COZ1 and showed null alleles signal, so it was excluded from subsequent analyses.
• Low and significant linkage disequilibrium was detected.
• PL03-71 locus exhibited signals of null alleles.
• Diversity indices:
  • COZ1: 28 alleles.
  • COZ2: 35 alleles.
  • Number of alleles per locus = 2–5.
  • Moderate values for Cozumel: Ho = 0.485, HE = 0.541.
  • COZ2 had higher values: Ho = 0.576, HE = 0.539.
  • COZ1 had lower values: Ho = 0.452, HE = 0.465.
• Genetic variation resided mainly within individuals (87.4%; p<0.05).
• Genetic variation between genetic clusters and among individuals within genetic clusters was 10.6% (p
• Relatedness results:
  • Highest proportion of unrelated individuals for Cozumel (73.3%).
  • Followed by half-siblings (12.4%), parent/offspring (9.8%), and siblings (4.4%).
  • COZ1 and COZ2 had similar proportions, but COZ2 had a higher percentage of unrelated individuals (86.4%).
• Evidence of a recent genetic bottleneck under all models was observed for COZ1 (p<0.01) with all but the SMM model for Cozumel, and none for COZ2.
• Allele frequencies followed a shifted-mode distribution only in COZ1, suggesting a bottleneck.
• Results from Garza–Williamson tests showed M values significantly lower than the respective Mc, indicating a historical population decline.
• Effective population size for Cozumel was Ne = 21.3.

Procyon lotor

• The same 12 microsatellite loci were amplified.
• After correcting for multiple comparisons, only the Jalisco population had a low proportion of loci out of HWE (0.083).
• Moderate and significant linkage disequilibrium was detected.
• Null alleles were identified for four loci (PLM13, PLM14, PLOT-04, PLOM15).
• Mainland genetic diversity levels showed 5–9 alleles per locus and a varied number of alleles per region.
• Rarefied allelic richness (Ar) ranged from 1.30 for Colima to 1.73 for Jalisco, while Cozumel exhibited an intermediate value (1.51).
• Moderate to high observed (Ho = 0.58–0.75) and expected heterozygosity (HE = 0.30–0.72).
• Genetic differentiation between Cozumel and continental populations was significant for all pairwise comparisons.
• Mainland and island genetic clusters were clearly identified with Structure, followed by a genetic subdivision within the mainland group in concordance with the sampling localities.
• DAPC results showed a clear separation between P. pygmaeus from Cozumel and mainland P. lotor individuals.

Nasua nelsoni

• Three of the 12 microsatellites loci tested were not successfully amplified, while Nasua1251 was monomorphic.
• One genetic group was considered for subsequent analyses.
• After Bonferroni correction, only Nasua931 showed significant departures from HWE.
• Null alleles signature was detected for Nasua931.
• Low and non-significant linkage disequilibrium was detected globally.
• Total of 41 alleles for Cozumel (3–7 alleles per locus).
• Mean heterozygosities: Ho = 0.478 and HE = 0.535.
• No evidence of a recent genetic bottleneck under any model was observed.
• M value for Garza–Williamson test was higher than the critical value, suggesting no sign of historical genetic bottleneck.
• High proportion of unrelated individuals was found, followed by half-siblings, parent/offspring, and siblings.
• Effective population size estimated was Ne = 127.3.
• In contrast with the Cozumel population, locus Nasua1251 was polymorphic on the continental populations.
• No evidence of departures from HWE was found after correcting for multiple comparisons, while non-significant linkage disequilibrium was detected.
• Two loci (Nasua1050 and Nasua833) showed evidence of null alleles.
• Total of 41 alleles (2-8 alleles per locus), rarefied allelic richness for Cozumel was Ar=3.38.
• Clear genetic differentiation between Cozumel and continental individuals was evidenced by Structure and DAPC.

Mitochondrial relationships and demography

• A cyt b fragment was amplified from 69 Procyon and 73 Nasua individuals.
• Procyon pygmaeus exhibited one unique haplotype and hence null cyt b diversity.
• Individuals Pyg3 and Pyg7 showed the haplotype belonging to Quintana Roo, thereby they were analyzed with the P. lotor sequences.
• Three haplotypes were found for the P. lotor mainland individuals.
• Nasua nelsoni exhibited six haplotypes, with high levels of genetic diversity and significant departures from neutrality tests.
• Mainland N. narica showed lower genetic diversity per locality and no significant departures from neutrality.
• Phylogenetic ML analyses showed a topology in agreement with the geographic distribution of Procyon populations, in which one clade includes exclusively P. pygmaeus individuals from Cozumel with a high level of support.
• Pyg3 and Pyg7 individuals exhibit the Quintana Roo haplotype thus were grouped within this clade.
• The mismatch distribution for the insular population showed significant values.
• No estimation of population size could be done due to the low number of haplotypes.
• Regarding Nasua populations, ML analyses showed three main clades: Clades I and II grouped mainland individuals, separating Morelos and Jalisco localities, respectively, while three groups were identified within Clade III: Groups 1 and 3 include individuals only from Cozumel and Group 2 from Cozumel and Quintana Roo.
• No significant mismatch distribution was found for the island population.
• The historical effective population size estimated was Ne = 8955.

Discussion

• The study presents a comprehensive analysis of genetic diversity and divergence at historical and contemporary scales of the populations of two endemic dwarf carnivore procyonids from Cozumel Island.
• Island-continent genetic pattern: species inhabiting islands harbor lower genetic diversity values when compared with their continental counterparts.
• Pygmy raccoon Procyon pygmaeus and dwarf coati Nasua nelsoni exhibited moderate levels of nuclear genetic diversity, similar to other mammals from the island.
• Procyon lotor have higher variability levels (Ho = 0.589–0.910; HE = 0.542–0.940).
• The pygmy raccoon shows null mitochondrial diversity, while the mainland P. lotor exhibits a range of cyt b haplotype diversity.
• We find four cyt b haplotypes for N. nelsoni (h=0.674; n=45), within the range observed for N. narica suggesting moderate historical diversity in N. nelsoni that has been maintained.
• The genetic structuring of P. pygmaeus in two distinct clusters is mostly concordant with its distribution on Cozumel, where family groups are present on the north and the south, but not across the island.
• The proportion of genetic variability resides within individuals (87.4%), but also significantly so between genetic clusters (10.6%).
• The extremely low effective population size (Ne = 21.4) and evidence of historical and contemporary genetic bottleneck signals support a trend of decreasing population size.
• Relatedness also differs between P. pygmaeus clusters, COZ1 having a higher percentage of related individuals.
• COZ1 encompasses individuals predominantly from two sites, the area near the dock to Isla de la Pasión and Club de Golf Cozumel Country Club, where individuals have access to a food supply directly from people and indirectly from garbage.
• COZ2 population is from Punta Sur, a protected area where human presence and activity have been restricted for decades, thus raccoons face less human influence.
• Human direct and indirect interactions with pygmy raccoons on the north have undoubtedly contributed to the decreased dispersal of its populations, enhancing the genetic diversity and differentiation patterns observed.
• Nasua nelsoni did not show a clear nuclear genetic differentiation (Structure depicted one genetic cluster, DAPC two) but, the mitochondrial phylogenetic relationships revealed three clearly distinct groups, where one haplotype is found exclusively in individuals from the south (Punta Sur).
• Our findings show that N. nelsoni has historically maintained a level of differentiation defined by distinct mitochondrial haplotypes, whereas its dispersal and behavior features likely have prevented a more conspicuous contemporary genetic differentiation.
• A high percentage of unrelated individuals (78%) was documented, which is uncommon given that coati groups primarily consist of highly related individuals.
• Although, N. nelsoni did not exhibit bottleneck signals, recent bottleneck effects may cause a drop in alleles before a drop in levels of heterozygosity.
• The phylogenetic analyses were not intended to describe the species’ phylogeny but to obtain an approximation of their endemism.
• Mitochondrial relationships among P. pygmaeus, P. cancrivorus and P. lotor haplotypes are concordant with previous phylogenies.
• We found one unique haplotype in P. pygmaeus, divergent from populations of Mexico and the USA, strengthening its endemic status.
• Nuclear genetic structuring among the mainland and the island further supported the distinctiveness of the Cozumel populations, identifying the island as a unique genetic cluster for both Nasua and Procyon.

Conservation of the Cozumel dwarf carnivores

• Nasua nelsoni and Procyon pygmaeus face habitat transformation and fragmentation, introduction of exotic species, and anthropogenic interactions on Cozumel Island.
• Conservation plans and actions for these two endemic species and their habitats should consider the genetic information provided herein.
• Distinct taxonomic position is considered and that Cozumel carnivores are managed as Evolutionary Significant Units.
• Our findings significantly support a change in the IUCN classification for Nasua nelsoni.
• Previous studies have estimated small population sizes for the two species, while our demographic results demonstrate they are facing significant population declines.
• Effective population size numbers are extremely low for both species, while P. pygmaeus also exhibits historical and contemporary bottleneck signals.
• Another crucial conservation concern is the limiting dispersal of the pygmy raccoon on the north and their increased interactions with humans, which can have significant negative effects.
• There is a potential increase of health problems and disease transmission in non-naturally densely populated animal groups, as these carnivores.
• This problem can be exacerbated with the introduction of exotic species, like the feral dogs and cats that are present in significant numbers on Cozumel.
• The anthropogenic food sources and human interactions occurring with the northern P. pygmaeus population should be significantly controlled and reduced.
• Ensuring the connectivity between the northern and southern pygmy raccoon populations should be a key target, via corridors harboring the native vegetation.
• Prevent further habitat loss and fragmentation.
• Control programs of exotic and feral species must continue to be developed and implemented on Cozumel.
• We found a phylogenetic group (Fig. 3) that included an haplotype present both in Cozumel and Quintana Roo; this could likely be the result of those N. narica individuals (and their progeny) introduced to the island.
• Moreover we identified two Procyon individuals captured on 2012 that exhibit a haplotype exclusively found in P. lotor from Quintana Roo, thus evidencing the introduction of this mainland species onto the island.
• It is urgent to carefully assess the presence and control the movement and introduction of these non-native species.
• Only with a comprehensive management and conservation program can we expect to genuinely reduce the threats the pygmy racoon and the dwarf coati are facing, and hope for the long-term viability of these endemic dwarf carnivores on Cozumel Island.