Whole genome survey of big cats (Genus: Panthera) identifies novel microsatellites of utility in conservation genetic study

Big cats (Genus: Panthera) are among the most threatened mammal groups of the world, owing to hunting, habitat loss, and illegal transnational trade. Conservation genetic studies and effective curbs on poaching are important for the conservation of these charismatic apex predators. A limited number of microsatellite markers exists for Panthera species and researchers often cross-amplify domestic cat microsatellites to study these species. We conducted data mining of seven Panthera genome sequences to discover microsatellites for conservation genetic studies of four threatened big cat species. A total of 32 polymorphic microsatellite loci were identified in silico and tested with 152 big cats, and were found polymorphic in most of the tested species. We propose a set of 12 novel microsatellite markers for use in conservation genetics and wildlife forensic investigations of big cat species. Cumulatively, these markers have a high discriminatory power of one in a million for unrelated individuals and one in a thousand for siblings. Similar PCR conditions of these markers increase the prospects of achieving efficient multiplex PCR assays. This study is a pioneering attempt to synthesise genome wide microsatellite markers for big cats.


Introduction
The genus Panthera includes five hyper carnivorous apex predator species that are typically referred to as big cats [1][2][3] . These are the tiger (Panthera tigris), leopard (Panthera pardus), lion (Panthera leo), snow leopard (Panthera uncia), and jaguar (Panthera onca). Big cats have great conservation value. They play a significant role in ensuring proper ecosystem function through top-down regulation 4 . Being charismatic, big cats helps in mobilising mass audiences and funds for the conservation cause 5 . Also, big cats are species of great cultural and historical significance with references found in artwork, folk tales, and old sayings throughout their distribution 6,7 . Nevertheless, a rampant decline in their wild populations has been observed in the recent past, mainly due to excessive hunting (including the prey species) 3 and overexploitation of habitat resources. From 1970 onward, several measures have been undertaken globally to fight the cause of this falloff. These include a (1) hunting and trade ban, (2) periodic population census, (3) regional and international cooperation to initiate activities for habitat restoration and reintroductions, and (4) community sensitisation campaigns to mitigate conflict with humans 8-12 . However, the success of such measures has been limited as these species continue to be listed among the IUCN (International Union for Conservation of Nature) endangered species [13][14][15][16][17] .
Illegal Wildlife Trade (IWT) of big cats is a highly lucrative and unlawful transnational commercial activity that is worth millions of dollars annually 18  accounting for more than 1,700 individuals 24 . Poaching-driven regional tiger extinctions have occurred in India, Cambodia, Vietnam, Thailand, Korea, and other Asian countries in the past two decades [25][26][27][28] . Regulations (i.e. national laws, international treaties, and conventions) have failed to curb the illegal trade of big cats as this illicit trade is a complex, fast-evolving and a heterogeneous transnational issue involving multiple trading 4 partners/middlemen. Traded articles mostly lack morphological features to ascertain the species, reducing the ability to track their origins reliably.
Incremental adoption of genetic tools and techniques for wildlife conservation and management have been observed globally in the past 25 years mainly due to the development of the robust protocols for DNA extraction and PCR (Polymerase chain reaction) [29][30][31][32] . DNA tools are now increasingly employed for establishing species-level identity 33,34 , resolving taxonomic ambiguities 6,35,36 , wildlife conflict mitigation 37,38 , and more recently, establishing the source of origin 39-41 . Microsatellites or short tandem repeats (STR) are neutral, codominantly inherited, widely distributed, hypervariable, short repetitive nuclear DNA units that have been regarded as the best candidate to develop a genetic signature of the individual (DNA fingerprint), population, and subspecies. Multiplex STR systems to undertake geographic assignments of confiscations have been proposed for tigers, leopards, elephants, rhinos and many other endangered species 39,41-45 . However, except for rhinos and elephants, microsatellite-based applications have failed to achieve global consensus in wildlife offense investigation. Efficient and simple protocols with established utilities in wildlife forensics across the range and species of rhinos and elephants have convinced wildlife managers and law enforcement agencies to adopt DNA methods for seizure investigations.
Tiger, leopard, lion, and snow leopard are the four most commercially exploited (by poaching and illegal trade) Panthera species. Their conservation demands stringent law enforcement.
Here, we report the development of novel microsatellite markers for genus Panthera by mining the genome sequences of four (tiger, leopard, lion, and snow leopard) most exploited big cat species. This study is a part of an ongoing India-Korea-Russia collaborative initiative to develop and test microsatellite based multiplex PCR panels of the pantherine species for genetic identification of the whole genus Panthera. 5

Abundance and distribution of STR in genomes of big cat species
We analysed the whole genome sequences of seven big cat individuals and found a total of 80,474,871 variant sites. These include SNVs (single nucleotide variants), indels, and microsatellites. Potential target variants were mined within these variant sites following the protocols described in the materials and methods section. Some of these variants were consistently polymorphic across all genomes, whereas some had limited polymorphism. Due to a large number of potential target variant candidates, we selected only those that were at least polymorphic in 5 of the 7 big cat genomes. Altogether, there were 8,947 such potential target variants. Of these, 6,283 were found to be located on unique sites in the genome (unique target variant, UTV). We found 2,614 UTVs in all seven genomes, and these were finally processed for microsatellite screening using the program MSDB 46 .

Development of microsatellite markers for genus Panthera
Program batch primer 3 was used to design PCR primers 47 . About 4% of the UTVs were found suitable for primer design (i.e. sufficient flanking sequences and not single-copy sequences). These include 176 dinucleotides, 39 trinucleotides, 45 tetranucleotides, 11 pentanucleotides, and 3 hexanucleotides. The designed primer pairs for these loci were further screened based on GC content and the presence of secondary structures. Finally, primer pairs for 41 loci were shortlisted for oligonucleotide synthesis. PCR was subsequently attempted with the synthesised primer pairs with four DNA samples, one each of the tiger, leopard, lion, and snow leopard. Thirty-two microsatellite loci (Table 1) showed clear amplification in the expected size range and were considered further. The forward primers of these loci were fluorescently labelled with one of the four dyes -6FAM, VIC, NED, and PET. These labelled microsatellites were then used to genotype samples of tiger, leopard, lion, snow leopard, and lynx.

Microsatellite polymorphism evaluation
The fluorescently labelled microsatellites were used to genotype 99 big cat individuals and 7 lynxes. Loci Pan3A2 and Pan8A1 failed to produce scorable profiles in lynx samples and thus, were assigned zero allelic value (Table 2). Overall, all loci were found to be polymorphic (4 to 18 alleles/locus), but some showed no variations within species -Pan2D1 in tiger and lynx; Pan1A2 and Pan8A1 in lion; Pan3A2, Pan3D2, and Pan2C1 in snow leopard; and Pan1A2, Pan2A1, Pan2C1, Pan6C2, and Pan9C2 in lynx ( Table 2). The species wise microsatellite characteristics and polymorphism are as follows: Tiger (Pantheratigris): We genotyped 41 tiger individuals of wild and captive origin. They were collected from India (n = 8), Russia (n = 4), and South Korea (n = 29, zoo individuals). 7 We pooled the samples from different locations to make a single tiger population to study polymorphism of the markers. Though not appropriate, this was the best possible way as (1) there were not enough Russian tiger samples to perform genetic analysis independently, (2) the ancestry of most of the zoo tigers was presumed to be hybrid (Bengal tiger and Amur tiger) 48 , (3) Amur tiger and Bengal tiger are two ecotypes of a same subspecies 49 .
The number of alleles per locus ranged from 1 to 11 (mean: 5.6) with a mean expected heterozygosity of 0.5 (0.00 -0.84). Twenty-six of 32 loci deviated significantly from HWE (Hardy-Weinberg Equilibrium) after Bonferroni correction (adjusted p-value < 0.002, Table   S2), and null alleles were detected in 24 loci (threshold limit of 10%, Table S2). Deviation from HWE was expected due to Wahlund effect. Overall, the markers were found to be polymorphic (except Pan2D1) with a mean polymorphic information content (PIC) of 0.45.  Table S2). According to studbook records, all leopards sampled from Korean zoos belong to Indochinese subspecies (Panthera pardus delacouri). Thus, the probable HWE deviation may have resulted due to higher average relatedness or hybrid ancestry. Null alleles (≥10%) were detected in 19, 18, and 8 loci in leopards sampled from Russia, Korean zoos, and India (Table   S2). However, there were inconsistencies in their occurrence in three tested populations.
Thus, there is high probability of discovery of additional alleles in these developed markers, 8 if tested with a greater number of samples. Fourteen of the 32 markers were found suitable for conservation genetic studies with PIC ≥ 0.5 (Table 2). Lion (Pantheraleo): A total of 18 captive African lions from Korean zoos were genotyped.
Out of 32 loci, 2 were monomorphic and 30 were polymorphic loci, with the number of alleles ranging from 1 to 8 (mean = 3.2). The mean expected heterozygosity was 0.4 (0.00 -0.84) for lions. We did not observe any significant deviation from HWE after Bonferroni correction (adjusted p-value < 0.002) in any loci (Table S2). Null alleles were detected in 9 loci (≥10%, Table S2). The mean polymorphic information content was estimated to 0.35, with 8 loci having PIC > 0.5 (Table 2).
Snow leopard (Pantherauncia): Snow leopards (n = 8) were sampled from the wild (Mongolia) and zoo (Korea). All these samples were considered as a single population during genetic analysis as there were not enough samples from the wild or captivity to be considered as distinct populations. Moreover, Korean zoos sourced snow leopards from Mongolia.
There was no sign of HWE deviation in tested loci after Bonferroni correction (adjusted pvalue < 0.002, Table S2). Only 7 loci had null alleles above the threshold of 10% (Table S2).
Twelve markers had PIC ≥ 0.5 ( Table 2). 9 This study aims to propose a universal microsatellite marker system capable of undertaking individual identification and geographic assignments of big cat seizures. We understand that the loci with higher expected heterozygosity (He) are more useful for individual identification. Similarly, loci with PIC values higher than 0.5 are considered informative enough for estimating genetic diversity. In our study, the locus wise heterozygosity and PIC varied across the species. We selected fifteen microsatellite loci based on the comparative marker's PIC, heterozygosity, and allele diversity ( Table 3). These loci showed no signs of linkage disequilibrium (LD) with big cats' wild populations. The average PIC of 15 markers was 0.48, 0.50, 0.54, and 0.56 for the snow leopard, lion, tiger, and leopard, respectively. The cumulative power of discrimination among unrelated individuals (P ID ) was found to be 5.2X10 -10 , 7.9X10 -10 , 3.0X10 -11 ,and 5.2X10 -12 for lion, snow leopard, tiger, and leopard, respectively, using the recommended panel of 15 microsatellites. Similarly, the cumulative power of discrimination among siblings (P ID sib) was found to be 1.5X10 -4 , 7.8X10 -5 , 3.3X10 -5 , and 2.5X10 -5 for the snow leopard, lion, tiger, and leopard, respectively.

Discussion
Even with the development of more sophisticated and elaborate markers such as SNPs, microsatellites are still considered the best tool to study conservation genetics due to their codominant inheritance pattern and hypervariability. There are two kinds of microsatellitesspecies-specific and heterologous. The former is developed for a species of interest, while the latter is screened from a pool of STR loci that were previously described for other species.
Geneticists have used both species-specific and heterologous microsatellites to study the genetic diversity and population structures of big cats 31,32,50-52 . However, the use of heterologous markers is more prevalent due to the availability of a limited number of speciesspecific STRs. Mishra et al. (2014) compared the polymorphism of species-specific vs. cross-10 specific markers in Bengal tiger and concluded the former's superiority over the latter 53 .
Moreover, the chances of genotyping errors due to mispriming, false alleles, and null alleles are lesser with species-specific STRs. In this study, the genome sequences of seven big cat individuals belonging to four species were analysed rapidly to identify and develop thirty-two polymorphic loci. The procedure of microsatellite development involved four steps: (1) mapping of big cat genomes on the assembled reference genome of the domestic cat to develop a multiple sample construct, (2) screening of the unique variant sites from the multiple sample construct, (3) scanning of unique variants to identify the polymorphic STR loci with conserved flanking regions, and (4) designing of PCR primers for these loci and evaluation of polymorphism with the collected samples ( Figure 1). Since the whole process involved comparative genome analysis and selection of universally located STRs with conserved flanking regions, the developed microsatellite markers were regarded as speciesspecific for all the four target big cat species. This makes our study a pioneering attempt to develop microsatellite markers for a genus. The autosomal location of each marker was assigned based on the karyotype of the domestic cat as its karyotype is reported to be similar to that of Panthera species. The microsatellite markers were named according to the genus Panthera (Pan) and autosome location (A1, A2, D1, etc., Table 1). For example, Pan10C2, Pan14C2, Pan15C2, and Pan16C2 are markers located on chromosome C2 in all Panthera species. Microsatellites were found to be located on six of the eighteen autosomal chromosomes, thereby ensuring at least 33% genome coverage.
We developed fluorescently labelled primer pairs for 32 novel microsatellite loci. Their polymorphism potential was evaluated with the DNA samples of four big cat species and lynx. All markers amplified successfully and produced scorable profiles with tiger, lion, leopard, and snow leopard. However, the profiles of Pan3A2 and Pan8A1 were un-scorable with lynx samples. The faulty genotyping profiles could have resulted due to non-target 11 primer annealing. The increased phylogenetic distance between the source (big cats: genus Panthera) and target (Eurasian lynx: genus Lynx) species greatly reduces transferability of markers 54 .
All markers were found polymorphic in leopards. Pan2D2 in tiger, Pan1A2, and Pan8A1 in lion and Pan3A2, Pan3D2, and Pan8A1 in snow leopard were monomorphic. Mean allelic diversity was found highest for tigers followed by leopard, lion, and lynx ( Table 2). The evidence of null alleles in several locus suggests that more alleles may be discovered. We also reported significant deviation from HWE in several loci in tiger, leopard, and snow leopard (Table S2) (Table 3). Cumulatively, these markers have a high discriminatory power of one in a million for unrelated individuals and one in a thousand for siblings (Table 3).
Such a high degree of discriminatory power also makes this panel suitable for population genetic studies. In the wild, more than two big cat species often inhabit the same region or country simultaneously (e.g., tiger, leopard, lion, and snow leopard in India; lion and leopard in Africa; tiger, leopard, and snow leopard in Russia). The universal marker system for all the big cat species will reduce the necessary reagent cost and technical burden of researchers 13 working on different big cat species in a laboratory or a network of laboratories. This will also promote data exchange and cooperative research. The similar range of annealing temperatures of primers (Table 1) for the markers in this study will be useful for developing a multiplex PCR system. Besides, since the markers are developed by mining the polymorphic STR loci with conserved flanking regions using the assembled genomic sequence of the domestic cat as the reference sequence, most of the markers have the potential to be applied to a variety of other endangered cat species. The potential is exemplified by the use of lynx in this study; 30 out of 32 markers were successfully amplified using lynx samples. Hence, the proposed microsatellite panel is of great utility in establishing DNA fingerprints, population signatures, and wildlife forensics.

Sample collection and DNA preparation
We collected biological samples of tiger, leopard, lion, and snow leopard from nature reserves, zoos, and sample repositories of India, Mongolia, Russia, and South Korea (Table   S1). These include blood, muscle, faeces, shed hair, and DNA extracts. In our study, we also included DNA extracts of seven Mongolian Eurasian lynx (Lynx lynx) to assess the utility of the developed markers over other cat species (Table S1). This experimental work was conducted with permission by the Conservation Genome Resource Bank for Korean Wildlife (CGRB)that provided the biological samples of wild cats for this study. All samples were legally and ethically collected and wherever applicable, the necessary permissions and permits were obtained from competent authorities. The procedures involving animal samples followed the guidelines by Seoul National University Institutional Animal Care and Use Committee (SNU IACUC).The species identity of each of the sourced samples were 14 reverified using conservation genetic tools i.e., amplifying either species-specific primers 55,56 or by sequence analysis of Cyt b gene using universal primers 57 .
The commercial column-based DNA extraction kits were employed to extract DNA following the recommended protocols. The whole process was carried out in a sterile environment of a dedicated laboratory to avoid any chance of contamination. Further, a positive and a negative control per experimental setup were included. Post extraction, DNA was resolved on 0.8% agarose gel to assess quality and quantity. Finally, the DNA was preserved at -20°Cfor long term storage.

Microsatellite development for genus Panthera
In our study, we analysed previously published genome sequences of seven big cat individuals 58,59 . These include three tigers, two lions, a leopard, and a snow leopard.
Additionally, we downloaded the assembled genome of domestic cat, Felcat6.2 60 , that served as areference. The whole process has been schematised in Figure 1. Non-labelled primer pairs were synthesised for loci qualifying the primer designing and selection criteria. These primers were subsequently tested for PCR amplification with one sample each, of tiger, leopard, lion, and snow leopard. Gradient PCR (annealing temperature, 16 T a -52°C to 62°C, reaction volume -10 µL and primer concentration -5 pm each) was performed independently for each primer pair. Primer pairs producing a single product band of expected size during PCR amplification were shortlisted for fluorescent dye labelling (forward primers) with one of four fluorescent dyes (6-FAM, VIC, NED, or PET, Invitrogen, South Korea) to perform fragment analysis using Applied Biosystems 3130 Genetic Analyser. During primer dye-labelling, due consideration was given to avoid dye range and product size overlap.

Microsatellite polymorphism evaluation
Fluorescently labelled microsatellites were tested for their polymorphism potential in an independent PCR assay with 106 samples of big cats and lynx. In a reaction, the total volume