This item is only available as the following downloads:
READING THE SIGNS: TECHNIQUES OF CONSERVATION GENETICS APPLIED TO BOBCATS ( Lynx rufus) AND PUMAS (Puma concolor) Asiatic Goldencat ( Catopuma temmenckii) photo from Zoologie.de BY ELDRIDGE WISELY Baccalaureate Examination Wednesday, April 29 t h 2009 From 1 3PM in the Caples Carriage House
READING THE SIGNS: TECHNIQUES OF CONSERVATION GENETICS APPLIED TO BOBCATS ( Lynx rufus) AND PUMAS (Puma concolor) BY ELDRIDGE WISELY A Thesis Submitted to the Division of Natural Sciences New College of Florida in partial fulfillmen t of the requirements for the degree Bachelor of Arts Under the sponsorship of Dr. Sandra Gilchrist Sarasota, Florida April, 2009
ii Table of Contents page Personal Histories and Photos of the Bobcats and Pumas Species Determination of Bobcat and Puma Scats by DNA Analysis: A Pilot Study.................................................. ..................................................................21 31 Sample S3 Sample S13 Sample S14 Sample S2 Sample S8a Bobcats Sample S10 Sample S16 Sample S17 Sample S18 Rai
iv Preface This thesis project builds upon ideas that I have been working on for several years here at New Col lege. It began with an interest and aptitude for science and a special place in my heart for felines. When I picked my Area of Concentration at New College I had a love for Archaeology and Cellular Biology and a desire to work in the field of Conservatio n Biology, so I picked Biology. I did an Independent Study Project under the sponsorship of Dr. Bill Tiffany in South Africa working at a game reserve and analyzing different Big Cat Conservation schemes throughout the Eastern Cape. As I read more and in vestigated more for this project, I found that obtaining population census information was one of the large basic challenges for any conservation program focusing on large felids. I investigated and tried several methods to tackle this problem through tut orials, ISPs and internships sponsored by Dr. Al Beulig. After my successful project in South Africa, Dr. Tiffany introduced me to Sharon Matola, an alumna who is the founder and director of the Belize Zoo. Sharon offered to mentor me and helped me to c ontact Omar Figueroa, a PhD candidate at the University of Florida and the first Belizean jaguar researcher. Omar took me under his wing and showed me the ropes of Big Cat Conservation Biology by letting me volunteer with him in the field. Omar Figueroa is an amazing person and a role model to me as a field biologist. All of these people have defined my journey and made this dream and the path I am now on a reality. Finally, after trying my hand at camera trapping, tracking, and trapping and radio collar ing, I became very interested in non invasive DNA techniques including those discussed in this thesis. Around the same time that I decided I wanted to work with feline
v DNA for my thesis, I visited Big Cat Rescue in Tampa and appreciated their mission and operation so much that I decided I really wanted to collaborate with them on this project. I originally wanted to study the DNA of exotic cats like tigers, leopards and fishing cats. When I requested scats from them, I was surprised to find out that Big Cat Rescue could not provide scats from CITES listed endangered species because the scats may contain fur from the animal grooming, and therefore it would be considered trading in endangered species fur! Thankfully, I was able to adapt the protocols for tiger scats that I intended to emulate for use with puma and bobcat scats which are not endangered according to CITES. Pumas and bobcats provide their own challenges, and surprisingly, are not as commonly studied as I first expected. During this project I found that very little is known about bobcat genetics, and that further projects on this species such as sequencing of their mitochondrial genome still need to be done. These species also provided a local interest in my thesis as both of these species are found in Florida and are threatened habitat fragmentation and development like much other Florida wildlife.
vi Dedication and Acknowledgements First I would like to thank my Mom and Dad for encouraging me along the way and giving me me freedom to cho ose my own path in life. My parents are my most trusted advisors and best friends, they have given me everything I have and have stood behind me in all of my decisions that have led me to be who I am. The accomplishment of this milestone is theirs as wel l as mine because they have been there for every step along the way. I would like to thank my grandparents as well for emphasizing the importance of an education throughout my life and my family for letting me know that they are proud of me. My academic a dvisors and professors deserve much gratitude for their forbearance and their for their time and effort to help me achieve this goal. Dr. Gilchrist, my thesis sponsor, has provided guidance while still making me discover what I am capable of on my own wit h no help or interference from anyone else. I want to thank Dr. Beulig for the conversations about Central America and field research, and for the tutorials leading up to this thesis project. With his help, the idea for this thesis was developed and the foundation for this project was laid. Dr. Meg Lowman was the first person in academia that I told that I wanted to one day be a field biologist working with big cats. Her reaction reinforced my desire to pursue this seemingly unattainable goal. Dr. Amy Clore was an excellent academic advisor and gave me the laboratory background that I needed to make this project succeed. She is an excellent professor and I was glad to have been in her classes. Last, but not least, Dr. Bill Tiffany was the first profes sor to give me a chance with a big cat related project, and helped me make connections in the field. He did everything possible to help me reach my goal and gave
vii me a leg up into the world of big cat research and conservation. His interest in the lives a nd goals of his students is a credit to his profession. I want to thank Bethany Highsmith, Amy McDavid, and all of my other friends I especially want to thank Dominic Amaral f or his constant support and encouragement along the way. He is my biggest fan and bluntest critic. His faith in my ability to achieve my goals has been a driving force for me to do more and aim higher than I could have alone. His photography skills made my gel images beautiful, and his pipetting skills made long hours in the lab much more fun and productive. Finally I want to thank my laboratory assistants, Katrina Bang and John Correa for their help and companionship in the lab and for going out of thei r way to help with my project, despite their busy schedules. They always made time to help me and never complained about the long hours, the obsessive attention to detail, the unglamorous tasks, or even my music. This thesis is dedicated to the one who go t me interested in felines in the first patterns of their lives, and who instilled in my the desire to understand the subtleties of the signs and signals with which a nimals communicate. I dedicate this thesis to the memory of my cat, Tinkerbell, for teaching me the most valuable lessons I will ever learn.
viii List of Figures page Figure 1: Phylogenetic tree of the Felidae color coded by zoogeographical r Figure 2: Inferred intercontinental migrations associated with phylogenetic relationships and sea Figure 3: Map of the location of Big Cat Rescue in Tampa Florida Figure 4: Photo of Bobby B Figur Figure 17: Clearer image of G el 2 of mitochondrial cytochrome b PCR products Figure 18: Sequence of mitochondrial cytochrome b PCR product from sample S3 Figure 19: The 6 top BLAST matches to the sequence obtained from sample S3 Su Figure 20: Chromatogram and sequence of sample S13
ix Figure 21: Alignment of the sequence from sample S13 Tobi with 8 closely Figure 22: Sequence of mitochondrial cytochrome b PCR product from sample S14 Figure 23: Alignment of the sequence from sample S14 Tobi with 8 closely Figure 24: Chromatogram and sequenc e of sample S2 Figure 25: Alignment of the sequence from sample S2 Aspen with 14 closely Figure 26: Sequence of mitochondrial cytochrome b PCR product from sample S8a Figure 27: Alignment of mitochondrial cytochrome b genes from the five most closely matched species to sample S8a Figure 28: Chromatogram and sequence of sample S10 Figure 29: Alignm ent of the sequence from sample S10 Bobby Blue Rose Figure 30: Sequence of mitochondrial cytochrome b PCR product from sample S16 Figure 31: Alig nment of the sequence from sample S16 Angie with Figure 32: Sequence and chromatogram obtained from sample S17 Figure 33: Sequence of mitochondrial cytochrome b PCR product from sample S18 Figure 34: Alignment of the sequence from sample S18 Raindance with
x List of Tables page Table 1: Selected studies that used mol ecular methods to determine 30 Table 3: The 20 most similar sequences to sample S3 Table 4: Th e top 8 most closely matched sequences to sample S13 Table 5: The top 20 most closely matched sequences to sample S14 Table 6: The top 20 most closely matched sequences to sample S2 Table 7: Top 18 most similar sequences to sample S8a Table 8: The 19 most closely matched sequences to sample S10 Table 9: The 20 most closely matched sequences to sample S16 Table 10: The 20 most closely matche d sequences to sample S18
xi READING THE SIGNS: TECHNIQUES OF CONSERVATION GENETICS APPLIED TO BOBCATS ( Lynx rufus) AND PUMAS ( Puma concolor) Eldridge Wisely New College of Florida 2009 ABSTRACT Species determination of the origin of scats is a useful technique in the field of carnivore conservation and monitoring. Because most felids are cryptic and dangerous to study, methods of non invasive monitoring are commonly used. Better, more standard ized methods of species determination of scats could vastly improve conservation efforts for these animals. This pilot study examines the new universal primer and sequencing technique by applying it to scat samples from five pumas and five bobcats. An id entification success rate of 80% was obtained for pumas and 75% for bobcats. Nucleotide BLAST alignments were created for each successfully sequenced sample, and the resulting sequences were ranked in terms of their similarity to the target sequences. Ph ylogenetic relationships of the mitochondrial cytochrome b gene are explored in relation to the sequencing results. Other techniques such as AFLP, species specific primers, and microsatellite analysis are also qualitatively examined. Finally, methodologi cal recommendations for further research are presented to help standardize molecular scatology techniques. Dr. Sandra Gilchrist, PhD Division of Natural Sciences
1 Phylogenetics of the Felidae The relationships among the members of the Felidae family are complex. These cats are far ranging and share territories with other felid species (Janczewski et al., 1995). As a result, many subspecies may arise and closely r elated species may interbreed on occasion. The current trend of habitat fragmentation combined with hunting may cause and genetic drift. The Florida panther is a n excellent example of a population that became isolated on a peninsula far from most of its kind and became a subspecies with diagnostic features. The Florida panther usually has a direction change called a cowlick in the fur on its back and a kinked tai l. These features are a result of genetic drift and make the phylogeny of the Felidae family more difficult to sort by any taxonomic measure. To study the phylogeny o f the cats of the world, scientists have studied skull morphologies, overall physical features, behavior patterns, immune system karyotypes, and even vocalizations (Janczewski et al., 1995; Parson et al., 2000). Recently, with the advances in PCR technolo gy, large scale genetic studies comparing many different species have become much more practical. Two such studies are presented here. The first study created a highly detailed phylogenetic tree of the family Felidae using autosomal, X linked, Y linked, and mitochondrial DNA from all known extant cat species and 16 fossils of felid ancestors. The second study assessed felid phylogeny using only mitochondrial cytochrome b genes and mitochondrial 12S RNA genes. The second study is less extensive than the first, but is extremely salient to the interpretation
2 of mitochondrial cytochrome b sequence alignment data used in new non invasive species monitoring techniques. In 2006, Warren Johnson and colleagues created a highly detailed phylogenetic tree of the fa mily Felidae using autosomal, X linked, Y linked, and mitochondrial DNA from all known extant cat species and 16 fossils of felid ancestors. They found that y linked DNA was the most informative, with 47% shared derived sites compared to 38% for autosomal DNA, and 14% for X linked DNA. Mitochondrial DNA performed the worst at detecting relationships between distantly related species. The mitochondrial DNA had only 24 variable sites that were suitable for large scale lineage groupings, compared to 42 suit able variable sites in X linked DNA, 118 suitable variable sites in Y linked DNA, and 123 suitable variable sites in autosomal DNA. Johnson and colleagues used 39 genes from the 37 extant felid species, 16 fossil felids, and 7 non felid carnivores to creat e their phylogenetic analysis of the Felidae family. From these data, they calculated the most probable phylogenetic relationships between and among these species, and were able to estimate divergence times of the eight major branches and even reconstruct a possible migration scenario to explain the
Figure 1: Phylogenetic tree of the Felidae color coded by zoogeographical regions. Figure and caption from Johnson, W.E., Eizirik, E., Pecon n Felidae: A genetic assessment. Science 311:73 3
4 From these data, the first radiation of the modern felids was the divergence of the Panthera lineage including the ancestor to modern tigers, lions, leopards, jaguars, and also including the ancestor to the modern clouded leopard just recently discovered in Borneo. Next, many rapid divergence events took place resulting in the emergence of the Asian bay cat lineage, the caracal and serval lineage, and the ocelot lineage. The lynx lineage diverged about 8.0 million years ago, and the puma lineage soon after, between 7.5 and 7.2 million years ago. This radiation took place just after a theoretical migrat ion event concurrent with lowered sea levels across the Bering land bridge. The earliest known Lynx species fossils were found in Florida and were dated to the late Miocene, approximately 5.3 million years ago. The first bobcat appeared in fossil records approximately 2.5 million years ago. Modern pumas can be found in fossil records as far back as 1.8 million years ago. Finally, the leopard cat and domestic cat lineages are the most recent additions to the Felidae family, emerging approximately 6.7 and 6.2 million years ago respectively. The work of Johnson and colleagues supports the theory that feline evolution happened very quickly and produced 8 distinct branches. Twenty one of the thirty six divergence events elucidated by Johnson and colleagues o ccurred in less than 1.0 million years (Johnson et al., 2006).
5 Figure 2: Inferred intercontinental migrations associated with phylogenetic relationships and sea level changes. Figure and caption from Johnson, W.E., Eizirik, E., Pecon Slattery, J., Mu Miocene Radiation of Modern Felidae: A genetic assessment. Science 311:73
6 The second study discussed here was performed by Janczewski and colleagues in 1995. It specifically addresses mitochondrial DNA evolution throughout the history of the Felidae family. The more recent study by Johnson and colleagues, discussed above, was partially built on the work done by Janczewski and others. A closer examination of mitochondrial DNA evolutio nary similarities is a necessary basis for any study that compares sequences of mitochondrial DNA among species. Mitochondrial DNA is a natural choice for studying population level ecology and phylogeny of closely related species because of its lack of rec ombination and its high rate of base substitution. Mitochondrial DNA is a circular molecule that is found in the mitochondria of all animal cells. It does not participate in sexual reproduction and is therefore only passed from females to their offspring through the cytoplasm of the egg cell. Recombination does not occur in mitochondrial DNA as it does in nuclear DNA and therefore variation rates are more easily calculated. There are four main types of sequence changes that affect mitochondrial DNA: rea rrangements, additions, deletions, mitochondrial DNA is 9.5 times faster than that of felid autosomal DNA, 12 times faster than that of X linked DNA, and 6 times f aster than that of Y linked DNA based on the mean pairwise genetic distances between felid species for each type of DNA (2006). The regular rate of change in mitochondrial DNA allows researchers to study relationships between species, and their approximat e times of divergence can be calculated when estimates are grounded in dated fossil evidence. By analyzing expected levels of variation in mitochondrial DNA and nuclear DNA, scientists can even detect bottleneck events in the past.
7 Janczewski and her c olleagues (1995) compared the mitochondrial cytochrome b genes and mitochondrial 12S RNA genes of 17 felid and 5 non felid species to elucidate the phylogenetic relationships of the family Felidae. They isolated DNA from tissue samples of 75 individual ca ts, then amplified a 289 base long region of the mitochondrial cytochrome b gene and sequenced them. By calculating the difference between each species, they were able to reconstruct the phylogenetic tree of the Felidae family and estimate the times of di vergence of each clade. Transposition, the movement of a piece of DNA from one position to another in the genome, can be responsible for mutations if it moves into a coding sequence or gene, or can simply change the length of a non coding sequence. Whe n a change in DNA 1994). The average rate of transposition among all animals is between 10 2 to 10 4 events per generation. Because of this low rate of transposit ion, different populations can be (1995) also studied the different positions of their amplified region of mitochondrial cytochrome b looking for the possible transversions and transitions that phyletic groups of felines share. mitochondrial genes including substitution, transposition and transition data. They found that the lynx clade (including bobcats and three geographically separated species of lynx) is one of the most closely interrelated clades. The next most closely related mitochondrial gene matches to the lynx clade are the marbled cat, followed by the serval, caracal, African goldencat, and the Asi atic goldencat. The genetic data suggested that the
8 puma mitochondrial DNA is most similar to that of the cheetah, followed next by the entire branch encompassing the Asiatic golden cat, caracal, serval, marbled cat and the lynx and bobcat clade (Janczews ki, et al 1995). Because of the high number of subspecies of pumas and leopards, Janczewski and colleagues (1995) inferred a high amount of genetic variation below species level in the Felidae family in general. This could mean that felines are natural ly well protected from genetic disease and able to adapt quickly when their populations are able to intermingle. Janczewski concluded that because mainly third base transitions were found in their cytochrome b analysis and third base transitions are so rap id, that mitochondrial cytochrome b is most informative in phylogenetic studies between more closely related species that may have diverged within the last few million years (1995). The current study uses 472 bases of the mitochondrial cytochrome b gene r egion to differentiate closely related sympatric felids based on amplified sequences.
9 Personal Histories of the Cats that Contributed to this Thesis The cats examined in this study are all housed at Big Cat Rescue in Tampa, Florida. This facility is l ocated at Expressway and Gunn Highway, immediately adjacent to Westfield Shopping Town. Figure 3: Map of the location of Big Cat Rescue in Tampa Florida. The red star marks the location of the san ctuary. Map courtesy of MapQuest, 2009. Big Cat Rescue is the largest accredited non profit big cat sanctuary in the United States, and houses the most diverse collection of exotic cats in the world. Over 150 cats
10 representing 16 species and subspecies live at Big Cat Rescue. This sanctuary takes in cats from all over the United States and provides them with a place to live out their lives. Every year hundreds of big cats must be turned away. Because of the growing numbers of abandoned and abused bi g cats that cannot be rescued, the sanctuary has begun an extensive preventative education program about the underlying causes of big cat suffering. Big Cat Rescue is active in encouraging the public to speak out against Over time, the values of Big Cat Rescue have evolved to recognize that captive breeding programs for big cats do not result in the continued viability of these animals in the wild, and therefore they do not breed or support efforts to breed big cats for lives in captivity (Big Cat Rescue, 2008).
11 Bobby Blue Rose Figure 4: Photo of Bobby Blue Rose. Photo courtesy of Big Cat Rescue Bobby Blue Rose is a female bobcat that may have some lynx ancestr y due to her origins at a fur farm. She was born on May 1, 1992 at a fur farm in Minnesota that raised mink, foxes, bobcats, Canadian lynx, and Siberian lynx. Bobby Blue Rose became one of the first cats at Big Cat Rescue on May 27, 1993 along with 55 ot her bobcat and lynx kittens. It is unlikely that she has been spayed because Big Cat Rescue believes in minimal human contact and performs invasive procedures only in emergencies. She lives alone in her enclosure. She is very friendly toward humans, and never learned wilderness survival skills as a kitten; therefore, she will never be releasable. Her favorite toy is a cardboard tube stuffed with mice and chicks.
12 Angie Figure 5: Photo of Angie. Photo courtesy of Big Cat Rescue Angie is a former pet that was brought to Big Cat Rescue by her owner along with two pet pumas on January 9, 1998. She was five years old at the time. Her housing situation was supposed to be temporary while her home cage was refurbished, but after two years, she was given h er own habitat at the sanctuary and she has lived there ever since. Angie has the characteristic dark coloring and spots of a southern bobcat.
13 Cherokee Figure 6: Photo of Cherokee. Photo courtesy of Big Cat Rescue Cherokee was purchased at an aucti on at the age of six weeks. She was destined for taxidermy or breeding stock at a fur farm. Cherokee was born on July 13, 1994 and came to Big Cat Rescue straight from the auction block on September 13, 1994. Like most bobcats, she prefers to live in a solitary enclosure. She was not handled as a kitten destined for the pet trade would have been; therefore, she is very elusive and prefers to avoid humans. Still, she was not taught to survive in the wild and can never be released. Her favorite activity is playing with cardboard cat toys stuffed with spices.
14 Breezy Figure 7: Photo of Breezy. Photo courtesy of Big Cat Rescue Breezy was born at Big Cat Rescue on April 14, 1996. She was hand raised. When the founder of Big Cat Rescue realized that t here are no conservation breeding programs that contribute to wild populations, the cats were housed separately or neutered as appropriate to prevent any more births. Breezy is very timid around humans and likes a very lush enclosure that provides plent y of opportunity for hiding. She lives by the lakeshore at Big Cat Rescue and enjoys watching ducks, peacocks, and swans pass by. She can often be glimpsed while sleeping in the large oak tree in her enclosure. Breezy has the light fur color of a northern bobcat or lynx, but the fine bone structure and small stature of a southern bobcat, suggesting her genetically mixed lineage.
15 Raindance Figure 8: Photo of Raindance. This photo was taken by Daphne Butters, and is used courtesy of Big Cat Rescue. Ra indance was a member of the first group of 56 bobcats and lynx that began Big Cat Rescue. She was born on May 8, 1993, and she came to the sanctuary only 19 days later. She had to be bottle fed by humans because her mother was still in a breeding cage at a fur farm in Minnesota. Some time later, Big Cat Rescue purchased the adults and paid off the owner of the operation to agree never to breed cats for slaughter again. Raindance and most of her immediate family now live at the sanctuary. She lives alon e in her enclosure as is her preference. She has mixed physical features from southern bobcats, northern bobcats and lynx as well.
16 Aspen Echo Figure 9: Photo of Aspen Echo. Photo courtesy of Big Cat Rescue Aspen is a female puma that has the reddis h fur, kinked tail, and mid back cowlick that are characteristics of the Florida panthers. She was born on August 30, 1997 in captivity for the pet trade. Not much is known about her life before she came to Big Cat Rescue, so it is unknown if she is spay ed. Aspen is a very energetic puma, always entertaining herself by chasing small animals that find their way into her enclosure and stalking passersby from behind her large palmettos.
17 Enya Figure 10: Photo of Enya. Photo courtesy of Big Cat Rescue Enya was born on December 9, 1997 into the pet trade. When she was only 10 days old, she and her mother were featured in a BBC documentary about the Florida Panther. She was hand raised and trained to walk on a leash from a very young age, but when she was 1 year old, she became too wild to handle. She was then sent to live out the rest of her years at Big Cat Rescue. Just recently, Enya was relocated to a new enclosure with a large oak tree in it. This year, she has been practicing her tree climbing skills for the first time.
18 Hallelujah Figure 11: Photo of Hallelujah. Photo courtesy of Big Cat Rescue Hallelujah, known as Hal to his keepers, was purchased at an auction on September 20, 1993. He was only 4 days old at the time. He lives alone a nd is not neutered, as evidenced by the attention he pays to his neighbor Tobi. Hal greets all of the humans that come near his enclosure, and spends his days serenading Tobi with a cacophony of chirps and mews. He is particularly large for a puma which suggests that he is more closely related to the large western pumas instead of the eastern or Florida breeds. Unlike most pumas, Hal loves water! His favorite activity is chasing water from
19 T obi Figure 12: Photo of Tobi. Photo courtesy of Big Cat Rescue unknown. She was born on March 1, 1991 and lived in Oneida, Florida until August 18, 1995, when she was sent to live at Big Cat Rescue. She is not spayed. Tobi is a very active cat. She enjoys waiting until a person passes her enclosure and then leaping out from behind her palmetto bushes to surprise them from behind.
20 Sugar Figure 13: Photo of Sugar. Photo courtesy of Big Cat Rescue Sugar was born in Oneida, Florida on March 1, 1991. She and her life partner Shadow were then sold to a man who lived in South Florida. Sugar and Shadow lived with their owner in a trailer for two years, until he became ill and could no longer care for them. Sugar and Shadow lived in the same enclosure at Big Cat Rescue together for 12 years. Sugar was not spayed, but Shadow was neutered. She enjoyed lounging on a large pile of dirt that she scraped together in their habit at, presumably to better survey her territory. Suddenly on October 2, 2008, Shadow began having seizures and died of a brain tumor. Sugar watched the door to their enclosure where she last saw him leave until she eventually lost her will to live and died on November 2, 2008, exactly a month after Shadow.
21 Species Determination of Bobcat and Puma Scats by DNA Analysis: A Pilot Study Introduction: The importance of species determination of DNA samples from non invasively sampled sources cannot be overstate d. Species determination is the basis for all other genetic studies of wildlife such as sex determination, individual fingerprinting, and evolutionary relatedness studies. Before any of these problems can be addressed, the species of the DNA source must be identified unequivocally. Species identification is of utmost importance when studying scats for other purposes as well. When estimating abundance and density of populations by indirect measures, the method hinges on the ability to distinguish the di agnostic character of the species of interest with surety (Ruiz Gonzles et al., 2008). Scientists studying diet from scats must first identify the source species of the scats. For instance, fecal stress hormone studies aim to determine the stress level of a population by measuring corticosteroid levels present in scats (Bonier, 2004). Before stress hormone levels can be measured, the species of origin of the scats must be ascertained. When studying kinship or phylogeny the species of the animal in ques tion is a basic underlying assumption of any analysis confirming the origin of scats are necessary for many types of inquiry in wildlife science. It is difficult to iden tify scats of sympatric carnivores by morphological features and content alone. Typically, scats must be identified by associated signs such as tracks and scrapes (Murie and Elbroch, 2005). Occasionally trained trackers misidentify the species from which carnivore scats originated (Davison et al., 2002). In a 2008 study of the molecular scatology of sympatric carnivores of the Iberian peninsula, Fernandes and
22 colleagues found that even the experienced naturalists that they worked with misidentified felin e scats as fox or mongoose scats. Scats from both bobcats and pumas can easily be confused with those of coyotes and domestic dogs (Murie and Elbroch, 2005). Feline Scat Morphology: Felid scats can usually be differentiated between species by their size (Murie and Elbroch, 2005). However, juvenile felids of larger species may be confused with adults of smaller species by this rough measure. Bobcat (Lynx rufus) scats vary from 1.1 2.5 cm in diameter and can be from 7.6 22.9 cm long. The scats of Canadi an lynx (Lynx canadensis) are usually between 1.3 and 2.4 cm in diameter, and from 7.6 25.4 cm in diameter. Puma scats can be from 1.9 to 4.1 cm in diameter and from 16.5 to 43.2 cm long (Elbroch, 2003). The size ranges of these scats are very much over lapping and based on rough empirical observation. Because the sizes of scats are not always diagnostic of species of origin, more accurate methods are necessary (Farrell et. al., 2000). Felids are never the sole carnivore species in their ranges, just us ually the dominant one (Janczewski et al., 1995). The presence of other sympatric carnivores in field conditions immensely complicates species determination of scats. Felid scats are easily confused with those of other carnivores, particularly with those of canines and mustelids (Murie and Elbroch, 2005). When relying on morphological features to identify scats in the field, one must take into account the possible presence of sympatric carnivores, the age of the scats, and the conditions in the study are a.
23 The morphology of scats varies greatly with climate and prey base. Both pumas and bobcats have segmented scats composed entirely of animal remains, fur, and very rarely, traces of grass. They never contain masticated fruits as canine scats sometime s do. In arid conditions, the segmented nature of feline scats may cause them to resemble pellets (Murie and Elbroch, 2005). photo by Dominic Amaral Figure 14: Photographs of sampled scats. On the left is a scat from T obi, a puma. On the right there is a scat from Raindance, a bobcat. These are typical comparative sizes for the sampled scats, but size and shape of scats may vary greatly. Therefore, molecular methods for species identification or verification are extr emely useful. In milder conditions, the average feline scat is extremely twisted on the inside but appears smooth on the outside. Segmentation, blunt ends, and a smooth surface are characteristics of scats that are usually used to differentiate scats of felines from those of canines. With pumas, however, the smooth outer layer is not always present, which can lead to erroneous classification as canine scat (Elbroch, 2003). As the age of the scat increases, so too does the chance of confusing feline sca t for canine scat as the smooth outer surface may be washed away or otherwise removed (Godbois et al., 2005). Most felines sometimes create scrapes and mounds of debris accompanying their scats, bobcats and pumas included. On occasion, scats may actuall y be placed in scrapes, which can be useful in species determination (Murie and Elbroch, 2005). According to
24 the accomplished North American tracker, Sue Morse, the best way to differentiate puma scats from bobcat scats is to look at the width of the scra pe if present. Puma scrapes are generally between 50 and 90 cm long, while bobcat scrapes can range from 15 to 50 cm long and from 7.6 19 cm wide, not including mounded debris (Elbroch, 2003). The range of scrape measurements from bobcats and pumas do no t overlap as much as the scat measurements do, which can aid in differentiating these two species when scrapes are present. However, coyotes and domestic dogs also sometimes scrape to accompany their scats, though never in an attempt to cover them. The f inal word on bobcat scats in the Peterson Field Guide entitled Animal Tracks is that whenever coyotes are present in the area where a putative bobcat scat has been found, the origin of the scat is suspect (Murie and Elbroch, 2005). Molecular Identificat ion Methods: Before molecular screening methods were developed, bile characteristics and blood heme analysis were used to narrow the possible list of species from which a forensic wildlife sample may have originated (Verma and Singh, 2003; Khorozyan et al. 2007). Today, the mitochondrial cytochrome b gene has become the most popular molecular region for species determination, and the number of species with mitochondrial cytochrome b sequences represented in the GenBank database reflects and facilitates th is methodological trend (Parson et al., 2000). Genetic means of species identification from non invasive sources is particularly important for the conservation and management of carnivores due to their generally elusive nature (Fernandes, 2008). Molecula r methods, once standardized, may prove to be the best tool yet for species identification of scats and hair samples of wildlife.
25 Several molecular methods exist that can be useful for determining the species of a DNA sample. These include enzymatic restr iction digestion of PCR products, Random Fragment Length Polymorphism (RFLP), Amplified Fragment Length Polymorphism (AFLP), species specific PCR, sequence analysis of whole mitochondrial genomes, microsatellite analysis, and sequence analysis of conserved genes or regions of DNA Srisamoot et al., 2007; Pandey et al., 2007). Each method has its own advantages and disadvantages that must be considered in the context of the goals, budget, and ecological factors of the project. RFLP and AFLP are recently developed methods that combine enzyme digestion, PCR (in the case of AFLP), and gel electrophoresis of total DNA extracts (Powell et al., 1995). These methods do not r equire prior knowledge of target sequences, but rather rely on many small comparative differences in fragment lengths created by enzyme digestion due to polymorphism among the target sequences (Srisamoot et al., 2007). However, RFLP and AFLP are most info rmative for questions of relatedness between species or subspecies rather than between individuals. These two methods are very expensive and are best for high quality DNA extracts (Mueller and Wolfenbarger, 1999). Microsatellites are better suited to qu estions of individual identity and relatedness of individuals or haplotype groups (Powell et al., 1995). Microsatellites do require foreknowledge of the desired target sequence and specific primers for each microsatellite locus of interest. Development o f microsatellite analysis methods for each species in question is extremely time consuming and expensive (Adams et al., 2003). Microsatellite
26 loci are found in genomic DNA, which is better suited to high quality invasively sampled DNA sources (Murphy et a l., 2000). The three most common methods used for degraded DNA samples are enzymatic restriction, species specific PCR, and universal PCR combined with sequencing. All of these methods rely on PCR amplification of a small section of mitochondrial DNA to then be further analyzed. For enzymatic restriction, a universal, order specific, or family specific primer is used to amplify a small (100 500bp) length of mitochondrial DNA; then the PCR product is digested with a combination of 2 or 3 enzymes to creat e different sized bands based on species specific polymorphism within the PCR product area. Knowledge of the expected cut patterns for each species is required for this method (Adams et al., 2003; Mills et al., 2000). Species specific PCR is based on the ability of specially designed PCR primers to amplify the DNA of only one species of all of the animals that live in a certain geographic area (Bhagavatula and Singh, 2006). For this method, multiple positive and negative controls and repeats of procedure are required for each sample in order to reduce error when target DNA may be degraded (Pirez and Fernandes, 2003). This method is sometimes preferred over universal primers for scat studies because it will not amplify prey DNA (Haag, 2009). Species speci fic primers may appear to be less expensive than sequencing until the necessity for replication and multiple extractions is considered. Finally, universal vertebrate PCR in combination with sequence analysis has been used in the field of wildlife forensics for identification of species of many types of DNA. Often, forensic samples are degraded or are found in small quantity, similar to the case with scat DNA (Pandey et al., 2007; Verma and Singh, 2003; Parson et al., 2000).
27 Universal primers may amplify p rey DNA in addition to carnivore DNA, however, the ability of these primers to distinguish between distantly related species with similar scat morphologies may outweigh the possibility of prey DNA sequencing. The current study aims to address this issue. See table 1 for a summary of recent molecular studies incorporating species identification.
Authors Year Focus Species DNA Source ID Method Gene or Locus Amplification success Species ID Success Adams et al. 2003 hybridization monitoring red wolf, coyote blood, scat canid specific PCR, enzyme restriction and sequencing mt cytochrome b gene for enzyme restriction and mt control region for sequencing 83% for restriction enzymes, and 86% for sequencing 92% for restriction enzymes, and 97% for sequencing Bhagavatula and Singh 2006 conservation monitoring, method improvement Bengal tiger c aptive scat and field scat carnivore specific PCR & species specific PCR & microsatellites mt cytochrome b gene 83% 70% Davison et al. 2002 population monitoring, method improvement pine marten scat carnivore specific PCR and sequencing mt contr ol region 53% 90% Ernest et al. 2000 conservation monitoring and individual ID puma, bobcat field scat microsatellites 12 Felis catus microsatellite loci (Menotti Raymond and O'Brien, 1995) 47% 87% Farrell et al. 2000 conservation monit oring and dietary research Jaguar, puma, ocelot, crab eating fox field scat carnivore specific PCR and sequencing mt cytochrome b gene 59% 100% Haag et al. 2009 conservation monitoring, method improvement jaguar and puma captive scat and field sca t carnivore specific PCR and sequencing ATP6/ATP8 genes 55% 100% 28
Authors Year Focus Species DNA Source ID Method Gene or Locus Amplification success Species ID Success Kitano et al. 2007 universal primer development for forensics 16 vertebrates and 2 invertebrates blood, tissue universal vertebrate PCR and sequencing mt 12S RNA and mt16S RNA genes 1 00% for vertebrates, 0% for inverts 100% Mills, Pilgrim, Schwartz, McKelvey 2000 conservation monitoring, method improvement Canadian lynx, bobcat, puma, domestic cat hair enzyme restriction mt D loop & 16S RNA gene 82% 100% Murphy et al. 2003 m ethod improvement brown bears captive scat undisclosed genus or species specific PCR mt DNA undisclosed region 88% 100% Pandey et al. 2007 conservation monitoring, method improvement, phylogeny leopard captive scat and field scat universal vert ebrate PCR and sequencing 12S rRNA gene 53% for captive scat, 31% for field scat 100% for captive, 57% for field scat Parson et al. 2000 universal primer development for forensics 44 vertebrates tissue, feathers, hairs, bristles universal vertebrate PCR and sequencing mt cytochrome b gene 100% mammals =100% Paxinos et al. 1997 conservation monitoring foxes, coyote, wolves tissue, scat canid specific PCR and enzyme restriction mt cytochrome b gene 100% 100% Pires and Fernandes 2003 cons ervation monitoring Iberian lynx field scat carnivore specific PCR & species specific PCR ATP8 gene 91% 2.10% 29
Authors Year Focus Species DNA Source ID Method Gene or Locus Amplification success Species ID Success Ruiz Gonzlez et al 2007 conservation monitoring, method improvement pine marten and stone marten tissue, hair, scat RFLP mt D loop reg ion 88% 100% Tamada et al. 2005 hybridization monitoring and phylogeny domestic cat and Tsushima leopard cat blood, tissue mustelid mt cytochrome b primers, domestic cat mt control region primers, and sequencing mt cytochrome b gene and mt cont rol region 100% 100% Uphyrkina et al. 2002 method improvement for conservation Amur leopard tissue leopard specific PCR and sequencing mt NADH 5 and mt control region 100% 100% Verma and Singh 2003 universal primer development for forensic s Indian vertebrates tissue, feathers, hairs, bristles universal vertebrate PCR and sequencing mt cytochrome b gene 100% 100% Table 1: Selected studies that used molecular methods to determine the species of origin of wildlife samples. The advantage s and ersal al. The Species ID Success column refers to the percentage of successfully (verified or statistically significant for the study) iden tified samples as a subset of the successfully amplified samples. 30
31 Methodological Examples: In 2006, Bhagavatula and Singh did a pilot study to assess the feasibility of a large scale tiger monitoring proj ect using non invasively sampled DNA from scats. They first designed their methodology on captive tiger blood samples and scat samples, then tested it on scats collected in protected areas in India. Tiger specific primers were designed for the purpose of quickly and efficiently screening scats presumed to be tiger in origin. To design these primers, Bhagavatula and Singh aligned the mitochondrial cytochrome b regions from tigers from blood samples and from GenBank with those of sympatric carnivores and p rey species including humans. Then tiger specific nucleotides these and the primers were tested using computer programs and in silico on captive tiger blood DNA sampl es. The possibility of false negatives resulting from interspecif ic tigers, 10 leopards, and several prey species. One hundred percent of the extractions performed on ca positive amplifications, but only 2 of 8 putative tiger scats collected from the wild amplified with their tiger specific primers. Recently, a study of the carnivore community in Spain and Po rtugal by Fernandes and colleagues used the same principles to identify which of a possible 15 species that various field collected scats came from (2008). They first amplified whole mitochondrial cytochrome b sequences from 15 Iberian carnivore species. They then made species specific primers for these animals, excluding the brown bear for its obvious scat morphology, and Iberian lynx because primers for it had already been developed. Their
32 species specific primers amplified a region less than 250 bases long, and therefore these primers were expected to be very useful in degraded DNA samples from scats. For each sample, they had to amplify it with 15 different species specific primer pairs and run 17 lanes of agarose gel including a ladder and a control. In the end, 28 (72%) of their 39 field collected scats were successfully amplified and identified. In 2003, Verma and Singh developed universal mitochondrial cytochrome b primers that amplify a conserved region shared among at least 222 animal species including humans. Universal mitochondrial cytochrome b primers make the process of screening DNA samples for species of origin faster and more efficient than designing primers unique to each possible species. These universal primers were developed for t he purpose of investigating wildlife crimes such as poaching. The only carnivore species used to develop this set of primers was Canis familiaris Other animals included rodents, ruminants, wild pigs, elephants, the Indian river dolphin, dugong, chimpanz ee, birds, and highly conserved regions of the gene sequence. The target sequence of these universal mitochondrial cytochrome b primers is 472 bases long. Because t hey amplify a small region of DNA, they are not as sensitive to DNA degradation as those primers that amplify longer target lengths. After being amplified with the universal primers, the PCR product must then be sequenced and compared to a genetic databas e for identification of species. This protocol was followed for the identification of bobcat and puma scats in the current study, and the effectiveness of this method was evaluated on North American feline species.
33 Methods: Ten scat samples were colle cted from the Big Cat Rescue sanctuary in Tampa, Florida on October 18, 2008. Of the ten cats sampled, five were bobcats and five were pumas. The sampled pumas included subspecies such as the Western cougar, the Eastern puma and one suspected Florida pan ther. Among the bobcats, there were three suspected bobcat/lynx hybrids and two putative southern bobcats. On the collection day, the temperature was between 66 F and 85 F with an average temperature of 76 F with rain. Scats were purportedly less than 24 hours old at the time of collection and placement in individually labeled plastic bags. Volunteers at Big Cat Rescue were instructed to collect scats from solitary animals only, place scats in new plastic bags, and label each with the name of the anim al and the date of collection. The scats were then stored in a freezer at Big Cat Rescue. Three days after collection, the scats were taken to New College of Florida where they were promptly stored in a freezer set to 80 C. To prepare for each extractio n, a glass plate was cleaned with abrasive bleach detergent, ethanol, and finally RNAse Away, as were two chiseling and scraping steel tools. Scats were removed from the freezer and individually chiseled to remove the outer layer of fecal material from th e scat before thawing could occur. Approximately 1 gram of scat was used per extraction. DNA was extracted from the scats using the Qiagen DNA Stool Kit (Qiagen Inc.). For samples that yielded less than 0.050 g/ l of DNA from the extraction process, a new piece of scat from that individual was selected for a new extraction. Eighteen extractions were performed in total.
34 The extracted DNA was amplified with universal mitochondrial cytochrome b primers developed by Verma and Singh as reported in Molecula r Ecology Notes in 2003. In the PCR reactions, 150 ng of template DNA was combined with 10 l of HotStarTaq Plus Master Mix (Qiagen), 2 l each of 100 M forward and reverse primers, and an appropriate amount of RNAse free water to bring the total reaction volume to 20 l. A drop of mineral oil was added on top of the reaction mix before beginning thermal cycling. The PCR program was the same as that outlined by Verma and Singh (2003), and was as follows: 95 C activation step for 10 minutes, then 35 cycle s of [95 C for 45 seconds, 51 C for 1 minute, and 72 C for 2 minutes], and a final extension step of 72 C for 10 minutes. Samples were stored in the 20 C freezer until gel visualization. Amplified mitochondrial cytochrome b regions were visualized on a 1.5% agarose gel stained with ethidium bromide, run in a 1X TBE buffer. The loading mix consisted of 2 l of PCR product, 2 l of loading dye (Carolina Biological), and 2 l of RNAse free water. Ten samples were loaded into two 1.5% agarose gels with a PCR control and 2 ladders in each gel. Gels were run at 100 V for 2.5 hours, and visualized in a darkened hood using a handheld UV light. Successfully amplified samples were sent to Genewiz Inc. for sequencing. Genewiz sequenced the samples on an ABI seq uencer using the amplification primer mcb 869 developed by Verma and Singh in 2003 as the sequencing primer. The resulting sequences were then compared to known sequences in the GenBank database using the non redundant nucleotide BLAST script in the progr am Geneious Pro 4.5.4 (Drummond, 2008). Following the guidelines of the Southwest Biotechnology and Informatics Center the sample was considered to match the known sequence when the identity of the bases
35 was greater than 80% for at least 80% of the sampl e length and the E value was below 1.0 E 80 (SWBIC, 2000). Results: Of the eighteen extractions performed, seven (39%) produced bands of PCR primers. These samples were: S2 Aspen, S8a Hal, S3 Sugar, S10 Bobby Blue Rose (very faint), S4 Breezy, S17 Cherokee (very faint), and S18 Raindance. See Figures 15 17 for gel images. These were the seven best extractions in terms of yield, all having over 50 ng of DNA per microliter of elution buffer. See Table 2 for extraction yields and purities. Table 2: Extraction yields and purity ratios for samples analyzed. Samples are sorted by species. Yield and purity measurements were taken with an Eppendorf Biophotometer u sing UV cuvettes (Eppendorf). The 260/280 ratio should be between 1.7 and 1.9 for pure DNA (QIAmp DNA Stool Handbook, 2007). Only S3, S13, and S8a have purity readings within that range. DNA yield when using the Qiagen Stool kit is typically 75 300 ng/ l, but may range from 25 500 ng/ l (QIAmp DNA Stool Handbook, 2007).
36 photo by Dominic Amaral Figure 15: Gel 1 of mitochondrial cytochrome b PCR products. Lanes labeled S2 and S8a show bands at approximately 500 bp long. The target sequence for Verma and Lanes C3, control, S12, and S11 showed no bands.
37 photo by Dominic Amaral Figure 16: Gel 2 of mitochondrial cytochrome b PCR produ cts. This gel shows amplification products in lanes S10, S3, and S4. The bands correspond to approximately 500 bp on the ladders. Sample S10 produced only a very faint band, suggesting lower yield of mitochondrial cytochrome b PCR product. Control, S13 and S7 were not visible.
38 photo by Dominic Amaral Figure 17: Clearer image of Gel 2 of mitochondrial cytochrome b PCR products. Ten samples were sent to Genewiz for sequencing, including the seven that produced successful bands and S13 To bi, S14 Tobi, and S16 Angie. The extra three samples were included because of their borderline yield values (51.9 ng/ l, 44.3 ng/ l, and 43.2 ng/ l respectively). Of these ten, nine (90%) were successfully sequenced and produced sufficiently intact seq uences for comparison with the existing sequence database GenBank. Seven (78%) of the nine sequences were correctly identified to the genus level. Four (80%) out of the five successful puma sequences were correctly
39 identified to the species level. The r emaining puma sequence suggested Puma concolor origin without statistical significance. Among the bobcats, none were aligned with Lynx rufus but rather with the two very closely related species Lynx canadensis and Lynx pardinus. Three (75%) of the four bobcat sequences conclusively identified with Lynx origin. Pumas Sample S3 came from Sugar the puma. It was a very high quality sample that yielded 111.3 ng of DNA per microliter of elution buffer and a 260/280 absorbance ratio of 1.84, suggesting that th e DNA was very pure and free of protein and RNA contamination. This sample produced a single band at approximately 500 base pairs when the universal mitochondrial cytochrome b primer PCR products were visualized. The sequence obtained from sample S3 show ed consistently high quality base assignments over 185 bases of its read length with no ambiguities or unassigned bases. The latter part of the sequence was lower quality, but still useful for alignments. See Figure 18 for base assignments and quality. When a BLAST search was performed on this sequence, the top match was a Puma concolor mitochondrial cytochrome b gene matching 94.9% of the sites when ambiguities in the target sequence were counted as mismatches. The E value for this match was 4.42 E 17 8 considerably higher than the E values for the next most closely matched sequence. See Table 3 for sequence similarity data and Figure 19 for a visual representation of the alignment similarities and differences between the closest matches. Basic Local Alignment Search Tool
Figure 18: Sequence of mitochondrial cytochrome b PCR product from sample S3 automatic sequencing machine was not able to unequivocally assign the identity of the base in that position. The blue bar gr aph shows a quality reading for each of the base assignm ents. A total length of 449 base pairs, out of 472 was high enough quality to be sequenced Figure produced using Geneious 4.5.4 developed by Drummond (2008). 40
Table 3: The 20 most similar sequences to sample S3 Sugar. The results of a standard nucle otide BLAST of the sequence obtained from sample S3 are presented here. The E values represent the chance that the match between the test sequence and the databa se sequence could have occurred due to chance (SWBIC, 2000). The sequences were matched bas ed on approximately 390 bases of matched sequence. All of the matched sequences are felid mitochondrial cytochrome b gene sequences, showing that the region amplified is diagnostic of species and not simply a conserved region universal to all mitochondrial DNA. The likelihood that the sample was puma in origin is much higher than the other possible species. 41
Figure 19: The 6 top BLAST matches to the sequence obtained from sample S3 Sugar. From the top to the bottom, the organisms with sequences mos t closely matching the target sequence are: Puma (Puma concolor) Iberian lynx (Lynx pardinus) Asiatic goldencat (Catopuma temmenckii) and Canadian lynx (Lynx canadensis) Highlighted bases are those which differ from the consensus sequence generated f rom combining all of the sequences shown. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 42
43 Sample S13 also came from a female puma, Tobi. This sample had a yield of 51.9 ng of DNA per microliter of buffer and a 260/280 absorbance ratio of 1.74, showing that this sample, though o f marginal quantity, was of good quality and purity. S13 did not show any bands when amplified with universal primers and visualized. However, the sequence produced from sample S13 was just barely of high enough quality to successfully align with the Pum a concolor cytochrome b gene. The many ambiguities in the sequence caused only a 47 base long region to be comparable to the database. The short length of the usable sequence resulted in non significant E values for the species assignments, so the sequen ce of this sample was considered inconclusive but suggestive of puma origin. In addition to the poor quality of species assignment, the second and third closest matches to this sample were from the mitochondrial cytochrome b genes of the Chihuahuan pocket mouse and the South American tent making bat. See Figures 20 and 21 and Table 4 for data.
Figure 20: Chromatogram and sequence of sample S13 e, quality scores are not available for each base call but the full length of the template was able to be sequenced. From visual inspection, this sequence shows a lack of clear peaks, which is indicative of poor quality or contaminated templa te DNA. Figur e produced using Geneious 4.5.4 developed by Drummond (2008). 44
Table 4: The top 8 most closely matched sequences to sample S13 Tobi. The results of a standard nucleotide BLAST of the sequence obtained from sample S13 are presented here. The E values represent the chance that the match between the test sequence and the database sequence could have occurred due to chance (SWBIC, 2000). The closest match is Puma concolor but the next most closely matched sequences are from the Chihuahuan pocket mous e (Chaetodipus eremicus) and the tent making bat (Uroderma bilobatum and U. magnirostrum) The sequence length of the matched area was only 47 base pairs long, which suggests that the most intact part of the sequenced DNA was a short fairly conserved regi on. 45
Figure 21: Alignment of the sequence from sample S13 Tobi with 8 closely matched sequences from 3 species. The bases that disagree with the majority represented are highlighted. From the top, the 3 species shown are puma (Puma concolor) Ch ihuahuan pocket mouse (Chaetodipus eremicus) yellow eared tent making bat (Uroderma bilobatum) and brown tent making bat (Uroderma magnirostrum) uma sequences show the highest degree of similarity with sample S13 Tobi. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 46
47 A second extraction of a different scat from Tobi was performed and labeled sample S1 4. This sample had a lower yield and purity than the previous one. It produced no bands when amplified and visualized in agarose gel with ethidium bromide and a UV light. However, when sequenced, this sample produced a medium to good quality chromatogra m. For 180 bases of the first part of the sequence, the quality was excellent, and the total read length was approximately 390 bases out of a possible 472. See Figure 22. The sequence from sample S14 Tobi aligned conclusively with the Puma concolor cyt ochrome b gene, showing a 94.8% pairwise identity, 93.3% identical sites, and an E value of 1.14E 169 over a sequence length of 389 bases. The next most closely matched sequences were the Iberian lynx (Lynx pardinus) and the Asiatic goldencat (Catopuma te mminckii) These sequences showed markedly lower similarity statistics with the target sequence S14 than the puma sequences did. See Table 5 and Figure 23 for sequence similarity and alignment data.
Figure 22: Sequence of mitochondrial cytochrome b PCR product from sample S14 Tobi. The sequence shows automatic sequencing machine was not able to unequivocally assign the identity of the base in that position. The blue bar gr aph shows a quality reading for each of the base assignments. The chromatogram for this sequence has clear, prono unced peaks. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 48
Table 5: The twenty most closely matched sequences to sample S14 Tobi. The results of a standard nucleotide BLAST of the sequence obtained from sample S14 are presented here The E values represent the chance that the match between the test sequence and the database sequence could have occurred due to chance (SWBIC, 2000). All of the matched sequences are felid mitochondri al cytochrome b gene sequences, showing that the regi on amplified is diagnostic of species and not simply a conserved region universal to all mitochondrial DNA. 49
Figure 23: Alignment of the sequence from sample S14 Tobi with 8 closely matched sequences from 5 species. The bases that disagree wit h the majority represented are highlighted. Most of the highlighted bases found in the sequence of Tobi are also found in the puma sequences directly below it. From the top, the 5 species shown are: Puma (Puma concolor) Iberian lynx (Lynx pardinus) As iatic goldencat (Catopuma temmenckii) Leopard cat (Pronailurus bengalensis) and Canadian lynx (Lynx Canadensis) Figure produced using Geneious 4.5.4 (Drummond, 2008) 50
51 Sample S2 was extracted from scat from Aspen, a putative Florida panther. The yield of this extraction was 105.5 ng/ l and the 260/280 light absorbance ratio was 2.02. An absorbance ratio higher than 1.9 shows some contamination o f DNA by proteins (QIAmp DNA Stool Handbook, 2007). Sample S2 produced a band of approximately 500 base pairs when amplified and visualized in agarose gel made with ethidium bromide. When this sample was sequenced, it produced a medium quality sequence w ith a non continuous read length of 412 bases ( Figure 24 ). A BLAST search using this sequence identified the sample correctly as belonging to a puma with an E value of 2.6 E 86 just exceeding the 1.0 E 80 standard of significance described by The South west Biotechnology and Informatics Center (SWBIC, 2000). The secondary match to this sequence was the tiger (Panthera tigris) followed by the snow leopard (Uncia uncia) These sequences were all from mitochondrial cytochrome b genes of f elids, and the s econdary and tertiary matches had much lower E values than the primary match, therefore the identification of this scat was considered significant and correct ( Table 6 and Figure 25 )
Figure 24: Chromatogram and sequence of sample S2 Aspen. This sequence was obtain be sequenced. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 52
Tabl e 6: The top 20 most closely matched sequences to sample S2 Aspen. The results of a standard nucleotide BLAST of the sequence obtained from sample S2 are presented here. The E values represent the chance that the match between the test seque nce and the database sequence could have occurred due to chance (SWBIC, 2000). All of the matched sequences are felid mitochondrial cytochrome b gene sequences, showing that the region amplified is diagnostic of species and not simply a conserved region uni versal to a ll mitochondrial DNA. 53
Figure 25: Alignment of the sequence from sample S2 Aspen with 14 closely matched sequences from 8 felid species. The bases that disagree with the majority represented are highlighted. From the top, the 8 felid species shown are puma, tiger, snowleopard, leopard, domestic cat, European wild cat, Iberian lynx, and Canadian lynx. These sequences were chosen because of their high similarity score with the target sequence. The puma sequences show the highest degree of similarit y with sample S2 Aspen. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 54
55 The last puma scat analyzed was sample S8a from Hal the male puma. This extraction yielded 66.5 ng of DNA per l of elution buffer and had an excellent 260/280 mitochon drial cytochrome b primer pair, sample S8a produced a good quality band at approximately 500 base pairs in an agarose gel stained with ethidium bromide. The sequence that resulted from this sample was of medium quality, with a 70 base long high quality se ction in the middle of a total of nearly 360 total bases read. The chromatogram shows peaks that are mostly clean but small. See Figure 26. The sequence from Sample S8a Hal aligned in a BLAST search with a Puma concolor cytochrome b gene, showing a 93. 2% pairwise identity, 90.0% identical sites, and an E value of 3.66E 96 over a sequence length of 243 bases. The next most closely matched sequences were the Asiatic goldencat (Catopuma temminckii) and the Canadian lynx (Lynx canadensis), with approximat ely 85% pairwise identity, 83% identical sites, and non significant E values of 4.18 E 70 and 5.10E 69 respectively. See Table 7 and Figure 27 for alignment and similarity data.
Figure 26: Sequence of mitochondrial cytochrome b PCR product from sample S8a automatic sequencing machine was not able to assign the identity of the base in that position unequivocally. The blue bar graph shows a quality reading of the base assignments. Almost 400 base pairs, out of 472 were high enough quality to be sequenced. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 56
Table 7: Top 18 most similar sequences to sample S8a Hal. The results of a standard nucleotide BLAST of the sequence obtained from sample S8a are presented here. The E values represent the chance that the match between th e test sequence and the database sequence could have occurred due to chance (SWBIC, 2000). All of the matched sequences are felid mitochondrial cytochrome b g ene sequences, showing that the region amplified is diagnostic of species and not simply a conserv ed region universal to all mitochondrial DNA. 57
Figure 27: Alignment of mitochondrial cytochrome b genes from the five most closely matched species to sample S8a Hal. The sequences are ordered by their E values and percent similarities to the targ et sequence with a BLAST similar sequence search. From top to bottom the sequences are: Puma (Puma concolor) Asiatic goldencat (Catopuma temmenckii) Canadian lynx (Lynx canadensis) Leopard cat (Prionailurus bengalensis) and Iberian lynx (Lynx pardinus ) The highlighted bases show the variable regions of each sequence that can be used to distinguish each species. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 58
59 Bobcats Sample S10 was extracted from a scat from Bobby Blue Rose. This extraction yielded 102.3 ng/ l of DNA in elution buffer. The 260/280 absorbance ratio measuring purity was 2.02. This purity ratio is close to the desired range of 1.7 1.9, but s uggests some protein contamination. Sample S10 produced an extremely faint band of expected primers and electrophoresed in an agarose gel with ethidium bromide. The sequence obtained from sample S10 showed a fairly high quality sequence of almost 450 bases out of 472. (See Figure 28). The sequence from sample S10 matched most closely with the mitochondrial cytochrome b genes of the Canadian lynx (Lynx canadensis) and the Ibe rian lynx (Lynx pardinus) One Canadian lynx sequence matched the sequence of S10 Bobby Blue Rose with 89.1% pairwise identity, 88.8% identical sites and an E value of 2.19E 148 over a sequence length of 410 bases. The Lynx sequences were all comparably similar to the sequence from S10, and the next closest matches (Felis silvestris and Felis catus) all had E values of 1.06E 139 or less. This indicates that the origin of the sample was correctly identified as a member of the genus Lynx specifically bei ng most closely related to the Canadian lynx ( Table 8 and Figure 29 )
Figure 28: Chromatogram and sequence of sample S10 Bobby Blue Rose. The clear peaks of the chromatogram indicate high quality template DNA, and excellent base calling ability. Nearly 450 bases out of a possible 472 were sequenced with surety. The mitochondrial cytochrome b region of this template DNA was amplified using the universal primers developed by Verma and Singh in Figure produced using Geneious 4.5.4 developed by Drummond (2008). 60
Table 8: The nineteen most closely matched sequences to sample S10 B obby Blue Rose. The results of a standard nucleotide BLAST of the sequence obtained from sample S2 are presented here. The E values represent the chance that the match between t he test sequence and the database sequence could have occurred due to chance ( SWBIC, 2000). All of the matched sequences are felid mitochondrial cytochrome b gene sequences, showing that the region amplified is diagnostic of species and not simply a conser ved region universal to all mitochondrial DNA. This table shows that there is a very high probability that the target sequence is from an animal belonging to the genus Lynx 61
Figure 29: Alignment of the sequence from sample S10 Bobby Blue Rose with 9 closely matched sequences from 7 species. The bases that disagree with the majority represented are highlighted. From the top, the 7 species shown are: Canadian lynx (Lynx Canadensis) Iberian lynx (Lynx pardinus), European wildcat (Felis silvestris), domestic cat (Felis catus), clouded leopard (Neofelis nebulosa), tiger (Pant hera tigris), and the African palm civet (Nandinia binotata) These sequences are the closest matches to the sample S10 from a standard nucleotide BLAST search. The Canadian lynx sequences show the highest degree of similarity with sample S10 Bobby Blue Rose. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 62
63 The next bobcat scat sample was S16 from Angie. S16 yielded 43.2 ng of DNA per l of elution buffer. I t had a rather high 260/280 ratio (2.04). This extraction was considered mediocre in terms of yield and purity. It did not show any band pattern after amplification and electrophoresis. However, when submitted for sequencing, this sample produced the se cond best quality sequence of the current study. The sequence obtained from sample S16 showed a continuous read length of 396 bases with excellent peak clarity in the chromatogram. See Figure 30. The BLAST search of this sequence resulted in a match wit h Lynx canadensis with 90.4% identical sites and 90.4% pairwise identity over the entire 396 base sequence length (E value = 3.35E 151 ). The top eight matches were from Canadian lynx and Iberian lynx. The next best species match was the common housecat ( Felis catus) With sample S16 Angie, the similarity statistics, while all significant, were so close that the sample only showed slightly higher probability of being of the Lynx genus than the Felis genus. See Table 9 and Figure 31 for sequence similari ty and alignment data.
Figure 30: Sequence of mitochondrial cytochrome b PCR product from sample S16 automatic sequencing machine was not able to assign the identity of the base in that position unequivocally. The blue bar graph shows a quality reading for each of the base assignments. The chromatogram for this sequence has very clear, pronounced peaks and excellent quality readings. Figure produced using Ge neious 4.5.4 developed by Drummond (2008). 64
Table 9: The twenty most closely matched sequences to sample S16 Angie. The results of a standard nucleotide BLAST of the sequence obtained from sample S16 are presented here. The E values represent the chan ce that the match between the test sequence and the database sequence could have occurred due to chance (SWBIC, 2000). All of the matched sequences are felid mitochondri al cytochrome b gene sequences, showing that the region amplified is diagnostic of spec ies and not simply a conserved region universal to all mitochondrial DNA. The sequences show only slightly higher probability that the target animal was of the Lynx genus than of the Felis genus. 65
Figure 31: Alignment of the sequence from sample S16 Angie with 9 closely matched sequences from 5 species. The bases that disagree with the majority represented are highlighted. Most of the highlighted bases found in the sequence from Angie are variable among all of the sequences shown, and are indicati ve of species identity. From the top, the 5 species shown are: Canadian lynx (Lynx Canadensis) Iberian lynx (Lynx pardinus) domestic cat (Felis catus) European wildcat (Felis silvestris) and leopard (Panthera pardus) Figure produced using Geneious 4 .5.4 developed by Drummond (2008). 66
67 Sample S17 was extracted from a scat from Cherokee the bobcat. This sample yielded 68.4 ng of DNA per l of elution buffer but had a very low 260/280 ra tio of 1.50. This low ratio shows that sample S17 was impure. After amplification with mitochondrial cytochrome b primers, sample S17 produced a faint band of expected size in an agarose gel stained with ethidium bromide. When sample S17 was sent to Gen ewiz for sequencing, the resulting chromatogram was nearly unreadable with very low quality scores throughout. The background signal continued for 756 bases, far past the 472 bases expected from the PCR product. See Figure 32. A BLAST search performed o n this trace of a sequence resulted in matches with bacterial 16S ribosomal RNA genes. The closest match to the target sequence S17 was a partial sequence of a Streptococcus mutans strain with 84% pairwise identity and 80.9% identical sites over a sequenc e length of 47 bases. The E value of this match was 1.88E 02 which is not significant enough for identification. This sample was considered contaminated and inconclusive of species.
Figure 32: Sequence and chromatogram obtained from sample S17 Cherokee. This chromatogram shows very high background signal and poor peak morphology. The quality scores shown by the blue shaded bar graph are extre mely low. This sample shows a high level of background signal suggesting contamination and poor sequencing priming. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 68
69 The final sample extracted and analyzed was S18 from Raindance. The extraction of her scat yielded 115.0 ng of DNA per l of buffer. The purity measured in the 260/280 absorbance ratio (2.07) was higher than the expected range for DNA, which is 1.7 1.9 (QIAmp DNA Stool Handbook, 2007). Sample S1 8 produced a strong band of expected size after amplification with universal mitochondrial cytochrome b primers and visualization in an agarose gel with ethidium bromide. The sequence resulting from this sample was of excellent quality. The peaks of the chromatogram were clear and exact. This sequence showed near perfect quality scores over the almost 400 bases of continuous read length that it produced. See Figure 33. The BLAST sequence similarity search of S18 identified Raindance as a Canadian lynx. Sample S18 and the Canadian lynx sequences had 90.0% pairwise identity and 89.9% identical sites over 397 bases (E value = 6.09E 148 ). The secondary match for sample S18 was the Iberian lynx, followed by the domestic cat, and the wildcat. The clouded l eopard sequence from GenBank shared 88.3% pairwise identity and 88.2% identical sites with sample S18 over 399 bases of length with an E value of 1.03E 138 While still significantly related, this was the 18 th closest match. See Table 10 for sequence sim ilarity statistics, and Figure 34 for a visual representation of the alignment.
Figure 33: Sequence of mitochondrial cytochrome b PCR product from sample S18 ABI automatic sequencing machine was not able to unequivocally assign the identity of the base in that position. The blue ba r graph shows a quality reading for each of the base assignments. The chr omatogram for this sequence has very clear, pronounced peaks and excellent quality readings. Figure produced using Geneious 4.5.4 developed by Drummond (2008). 70
Table 10: The twenty most closely matched sequences to sample S18 Raindance. The results of a standard nucleotide BLAST of the sequence obtained from sample S18 are presented here. The E values represent the chance that the match between the test sequ ence and the database sequence could have occurred due to chance (SWBIC, 2000). All of the ma tched sequences are felid mitochondrial cytochrome b gene sequences, showing that the region amplified is diagnostic of species and not simply a conserved region uni versal to all mitochondrial DNA. The sequences show only slightly higher probability that the target animal was of the Lynx genus than of the Felis genus. 71
Figure 34: Alignment of the sequence from sample S18 Raindance with 10 closely matched sequences from 8 species. The bases that differ from the consensus sequence are highlighted to sh ow important diagnostic bases for species determination. From the top, the 8 species shown are: Canadian lynx (Lynx Canadensis) Iberian lynx (Lynx pardinus) domestic cat (Felis catus) European wildcat (Felis silvestris) clouded leopard (Neofelis nebul osa) leopard (Panthera pardus) tiger (Panthera tigris) and leopard cat (Prionailurus bengalensis) Figure produced using Geneious 4.5.4 developed by Drummond (2008). 72
73 Conclusion: There was rain on the scat collection day and this may have reduced the yield of DN A from the samples. The Qiagen stool kit typically produces 75 300 ng/ l, but yields between 25 and 500 ng/ l of DNA are within normal range (QIAmp DNA Stook Handbook, 2007). The average yield of the extractions analyzed in this study was 75.5 ng/ l (SD = 32.4). This rather poor and highly variable yield can be attributed to the including rain, heat and humidity. Murphy and colleagues (2003) have suggested that the scats from certain individuals are less likely to yield useful DNA than others. This assessment was not supported by the current study. With multiple extractions from different scats, or different parts of the same scat, the current study was able to produce acceptable yield values for each of the 10 individual cats. To accomplish this, up to 4 extractions were required from one individual. The species determination success rate of the current study is higher than the success rates of comparable stu dies using field collected scats. Bhagavatula and Singh (2006) reported a 100% amplification rate with fresh captive tiger scats extracted on the day of deposition, but report that only 70% of putative tiger scat samples collected from the field amplified with their tiger specific primers. Their amplification success rate was comparable to several other studies of molecular scatology using field samples, which reported amplification success rates between 46% and 66%. Fernandes and colleagues (2008) repor ted that 72% of their field collected scats were successfully amplified by one of their 16 species specific primer pairs and identified to the species level. The current
74 study reports successful species determination of 80% of samples analyzed. The semi controlled environment at Big Cat Rescue approximated field conditions in most ways except the scats were fresher when collected than can be expected with field samples, and the source of the scat was known before analysis. The success rate of this study and other studies of feline scats were comparable all factors considered. The universal primer and sequencing method used in the current study is superior to the species specific primer method in several ways. First, when working with field collected sc at samples, a negative amplification result cannot be considered conclusive evidence that the analyzed sample did not come from the species in question. The strong possibility always exists that the DNA in the extracted section of scat was washed away, or degraded, or contaminated with PCR inhibiting substances or prey or vegetation DNA. Only a positive amplification can be considered conclusive evidence of the origin species of the scat. The universal primer and sequencing method provides more informati on than does the species specific primer method, particularly in cases of negative or ambiguous results. Second, when identifying scats by the species specific primer method, one must go through a process of elimination style screening involving species specific amplifications with unique primers and optimized cycling conditions for each possible species of origin. This process becomes more costly and less efficient as the list of possible species grows. For example, Fernandes and colleagues (2008) ampl ified their scat samples with 16 different species specific primer combinations for each scat sample and had to run 18 lanes of agarose gel for each extraction including ladder and control
75 lanes. The cost in time and money of primers and optimization outw eighs the cost of sequencing, particularly when the list of possible species of origin is large. Third, the species specific primer method relies on visualization of PCR products in an agarose gel for determination of sequencing success or failure. Howeve r, in cases of low yields of PCR product that may result from low quality or quantity of template DNA characteristic of scat samples, positive amplification results can be undetectable by this method. As the current study demonstrates, even amplification products that do not appear on a standard agarose gel stained with ethidium bromide can produce excellent sequences that lead to conclusive species identification. Pumas Four (80%) of the five puma samples were conclusively identified as being Puma concol or in origin by this method. All four of these samples had excellent, statistically significant matches with puma sequences in the GenBank database. The fifth sample, S13 Tobi, matched most closely with two puma sequences, but not to a significant degre e, due to the short usable sequence length. The second and third closest matches to sample S 13 Tobi were sequences from the Chihuahuan pocket mouse and the South American tent making bat. These non felid matches can be simply interpreted to be a resul t of the universal nature of the amplified region combined with the low quality of the sample sequence. An alternative interpretation is that the extraction of sample S13 Tobi was contaminated with non target prey DNA. While it is possible that a mouse and a bat were ingested by Tobi in her enclosure at Big Cat Rescue, it is highly unlikely that the mouse was of the genus Chaetodipus. The range of this genus of pocket mice does not extend from Northwestern
76 Mexico as far as Southwest Florida. These mice are more closely related to pocket gophers found in arid regions than to the common mice found in Florida. The tent making bat sequence match is also suspect because its natural range and that of its close relatives is in the Amazon rainforest of South A merica. Also, the E values of these sequence matches were so large as to suspect that they aligned with the sequence of sample S13 by chance. Close matches to three of the four successfully identified puma sequences were consistently from lynx species, t he Asiatic goldencat, the cheetah, the leopard cat, and domestic and wild Felis species. These sequences support the phylogeny of mitochondrial DNA created in 1995 by Janczewski and colleagues (1995) which placed Puma concolor closest to the cheetah (Acin onyx jubatus) followed by the Asiatic golden cat and then the lynx group. The remaining successfully identified puma sequence was from S2 Aspen. The closest matches to this sequence were the Bengal tiger (Panthera tigris tigris) the Sumatran tiger (Pan thera tigris sumatrae) and the snow leopard (Uncia uncia) followed by domesic and wild Felis species, and the Iberian lynx (Lynx pardinus) Neither the and colleagues plac e the puma near the tiger or snowleopard in terms of genetic relatedness. Because Aspen Echo is a Florida panther, her mitochondrial cytochrome b gene is most likely slightly different from those of other puma subspecies. The fast rate of substitution in the mitochondrial genome means that there is a possibility that a substitution convergent with the snowleopard occurred in the mitochondrial genome of
77 and tertiary BLA ST matches may be due to base substitutions in the Florida panther subspecies that differ from those of other pumas. More directed research would be necessary to ascertain the degree of difference of the Florida panther mitochondrial genome from those of western and eastern puma subspecies. From this data, it cannot be concluded that there is definitely a difference between the mitochondrial cytochrome b genes of Florida panthers and other pumas. First, the secondary and tertiary matches to sample S2 A spen had E values that were slightly higher than the necessary 1.0 E 80 necessary to be considered statistically significant. Second, the scat collection method used in the current study, while fine for interspecific genetic analysis, is not sterile enoug h to preclude traces amounts of other felid species DNA on the collection tools used to remove scats from enclosures. The secondary and tertiary matches of tiger and snow leopard DNA for this sample could be artifact from previous collections of tiger and snow leopard scats from other enclosures earlier on the related matches because domestic cat DNA would not be on the tool. The possibility of a detectable difference i n the sequences obtained from the universal mitochondrial cytochrome b primers used in this study is still very much conjecture, but should not be overlooked. Bobcats Three (75%) of the four bobcat samples analyzed were conclusively identified as being of Lynx genus origin. The top matches for all of these samples were Lynx canadensis complete coding sequences, followed closely by Lynx pardinus isolate sequences. There are many non exclusive potential reasons for this effect.
78 The first possible reason is that there is no complete mitochondrial cytochrome b sequence published in GenBank for Lynx rufus Lynx rufus sequences from 3 different paper s. For comparison, the same search for Panthera tigris mitochondrial cytochrome b genes produces a list of 127 sequences from at least 4 subspecies, including 38 complete sequences of the mitochondrial cytochrome b gene and its nuclear pseudogene counterp art present in tigers. The North American bobcat is woefully underrepresented in genetic sequencing studies. The second possible reason that these three bobcat samples did not most closely match a Lynx rufus sequence is that the bobcats in question are su spected bobcat lynx hybrids that came from fur farms. Fur farms do not keep any kind of parentage records and are known to purposefully interbreed bobcats with Canadian lynx to produce lighter shades of pelts with bobcat like markings. Because the DNA an alyzed is from the mitochondrial genome, if the mother of a bobcat lynx hybrid were a lynx, the hybrid offspring in question would have only the mitochondrial genome of a lynx but the nuclear genome of a hybrid. The third and most compelling potential reas on is that the partial sequences available in GenBank may be from a different section of the mitochondrial cytochrome b seems especially likely, as a Geneious sequence a lignment of the two complete Canadian lynx sequences with a consensus sequence of the three successful bobcat sequences from the current study, the two partial sequences from Iberian lynx, and a consensus sequence of the partial bobcat sequences available in GenBank supports this theory. The alignment
79 1,140 base length. The Lynx rufus consensus sequence aligned with bases 4 through 387 of the complete lynx sequences, and t he consensus sequence of bobcat samples S10, S16, and S18 from this study aligned with bases 414 through 807 of the complete lynx sequence. Data not shown (Drummond, 2008). The last potential reason for the bobcat sequences aligning most closely with Cana dian lynx sequences is that bobcats are so phylogenetically close to lynx that their mitochondrial cytochrome b regions are almost indistinguishable at the region amplified by the universal primers used in the current study. According to both Johnson and Janczewski, the lynx clade is one of the most closely interrelated clades of the Felidae family (Johnson et al., 2006 ; Janczewski et al., 1995). In this case, it may only be possible to resolve the genus of origin of the lynx clade with this method, and f urther manipulations such as restriction enzyme analysis would be necessary to determine the species of lynx (Mills et al., 2000). The remaining bobcat sample, which was unsuccessful, was from S17 Cherokee. Sample S17 produced a faint band of approximat ely 500 base pairs when visualized in an agarose gel stained with ethidium bromide, but did not produce a very useful sequence. On the day that sample S17 was visualized and a portion of it sent to Genewiz for sequencing, both the primary researcher and t he laboratory assistant were sick with symptoms consistent with strep throat caused by streptococcal bacteria. The top match to Sample S17 was a strain of Streptococcus which although statistically insignificant, may suggest contamination of the shipment of sample S17 by breathing streptococcal bacteria into it.
80 The matches to this sequence were all from 16S RNA genes from various bacteria. The 16S RNA gene probably appeared because it is one of the most commonly studied genes for bacterial genomic evo lution studies and therefore many records of it occur in GenBank (Zhang et al., 2002). The 16S RNA gene codes for a structural RNA that forms part of a bacterial ribosome. The cellular function of 16S RNA is unrelated to that of mitochondrial cytochrome b in animals, but the phylogenetic function to scientists is nearly identical. Because the E value of this match with the target sequence was so low and the quality of the target sequence so obviously poor, further inquiry into the reasons behind this par ticular sequence match are unwarranted. Methodological Recommendations for Future Studies When amplifying non invasively sampled DNA, quality and quantity of template DNA become limiting factors (Bhagavatula and Singh, 2006). This effect was evident in th e current study. The best sequences generally came from high yield extractions in the range of 40 120 ng/l. An even more powerful indicator than quantity of DNA was the quality of the extraction. Samples within the purity range of 1.7 2.05 performed be st overall. For example, S16 produced the second best sequence with only 43.2 ng/ l of DNA and a 260/280 ratio of 2.04, whereas sample S17 produced the worst quality sequence with 68.4 ng/ l of DNA but a 260/280 ratio of only 1.50. The quality of the ext raction will affect all downstream applications including sequencing. However, it is not the only factor influencing the success or failure of this technique. Samples S13 and S14 provide good examples. Sample S13 had higher quantity and better quality t han sample S14 when measured with a biophotometer, but the sequence of sample S14 was far superior to that of S13.
81 The extraction quality and quantity of DNA from non invasive sources is highly variable. This depends largely on the part of the scat or ot her degraded DNA source that the sample is taken from. For best results, scat samples should be taken from the outer colorectal epithelial mucous. Also, the outside of scats may contain fewer PCR inhibiting substances (Bhagavatula and Singh (2006). This becomes more complicated and less successful as the age of the scat increases in the field. Often the surface of a scat sample will be covered in leaves and dirt due t defecation. In this case, debris must be removed without removing target DNA. If there has been rain since deposition of the scat, the rain may have washed target DNA off the surface of the scat (Fernandes et al ., 2008). In this case, the sheltered underside of the scat may yield more DNA than the other parts. The current study aimed to maximize the potential target DNA yield by scraping the outsides of a frozen scat until a mass of approximately 1.0 grams of sc at was being of scat as directed in the basic Qiagen stool protocol (QIAmp DNA Stool Kit, 2007). This approach successfully produced 8 extractions out of the 10 anal yzed that were used to determine the genus and species of the target animal. By extracting DNA from a greater volume of scat, one can maximize the chance that a DNA rich portion of scat is included in the extraction. Based on the successful sequencing of samples that did not produce visible bands when visualized in an agarose gel, this study suggests that the visualization step is unnecessary in universal primer and sequencing protocols. To save time and money, the
82 gel visualization step can be omitted wh en using existing universal primers and accompanying optimized thermal cycling protocols. To check the method, gels can be used to verify the presence of at least one sample with a band of expected size and when at least one sample can be visualized, all of the samples from that treatment group should be sequenced as well. Future Directions To improve the universal primer method, the 472 base pair conserved region could be divided into two smaller PCR products with the addition of two more primers anchored on conserved bases between the current primers. In this way, the universal nature of the primers could be preserved while the potential for successful amplification of degraded DNA samples would be vastly increased. Future directions of research should focus on the development of universal primers for smaller regions of the mitochondrial cytochrome b gene in multiplex PCR reactions. Sequencing of the complete mitochondrial genome of bobcats (Lynx rufus) should be a priority for researchers using this me thod on North American carnivores. Bobcats can be found throughout North America and are sympatric to many other carnivore species. Without a complete sequence of at least the mitochondrial cytochrome b gene of this species, erroneous identification of n on invasively sampled DNA is a distinct possibility, particularly in areas where Canadian lynx may be present. Despite the non endangered status of bobcats today, considering the dismal trend of feline carnivore conservation across the board it is never t oo early to develop mechanisms for monitoring and conservation efforts for bobcats.
83 Discussion sequencing method from forensic analysis for scat analysis. These primers were designed for Indian wildlife, but the only carnivore included in the design phase was the domestic confirmed for North American felines. Some previous studies have shied away from the use of universal primers with scat because of the possibility of amplification of non target prey DNA (Haag, 2009). The results of the current study show that non target prey DNA is not a significant problem for scat studies following the current pro tocol. The use of larger volumes of stool in this protocol may counteract the negative effects of prey DNA mixing in the extraction phase. The presence of prey DNA in a few sequences or in DNA extracts taken from the interior of scats may even aid in diet ary studies of sympatric carnivores. The benefits of sequence analysis over species specific PCR product gel visualization far outweigh any potential drawbacks of this technique for conservation oriented scat analysis. The current study highlighted severa l benefits of the universal primer PCR method of species identification. First, far more information is available from sequence analysis than from a present or absent rating of species specific PCR product bands in a gel. When sequences fail to produce a significant match or match most closely to non target species, the matches provide clues as to the failure cause. Second, the technique need only be applied until a usable sequence is obtained. Further repeats and positive controls are unnecessary, maki ng this method cheaper and quicker than species specific primer based studies. The current study showed that a usable
84 sequence could usually be obtained from moderately fresh scats from 1 3 extractions. Third, the current study showed that high quality s equences can be obtained from samples which do not show banding patterns when visualized by gel electrophoresis. This suggests that false negatives may result from gel visualization, which is the final step in species specific PCR protocols. Fourth, phyl ogenetic studies are facilitated by the current technique, and questions of molecular phylogeny can be addressed as a secondary inquiry as a part of an overall molecular scatology study of wild felids. Lastly, the current method has the potential to add n ew sequences to the GenBank database and aid future studies on the target species. This pilot study successfully demonstrated the applicability of the universal primer and sequencing method to analyze scats of North American felids for conservation and man agement goals. It also exemplifies the accessibility of this technique to field biologists and researchers with little background in genetics. If conservation efforts are to be as effective as possible, powerful techniques such as genetic analysis must b e standardized and made accessible to the greater conservation community. This thesis project is a step toward simplification and standardization of molecular analysis of non invasive DNA samples for wildlife conservation and management.
85 References Adams J.R., Kelly, B.T., Waits, L.P. 2003. Using faecal DNA sampling and GIS to monitor hybridization between red wolves ( Canis rufus ) and coyotes ( Canis latrans ). Molecular Ecology. 12: 2175 2186. Bhagavatula, J., Singh, L. 2006. Genotyping faecal samples of Bengal tiger Panthera tigris tigris for population estimation: A pilot study. BMC Genetics. 7:48 Big Cat Rescue. 2008. BigCatRescue.org. Accessed Nov. 10, 2008 Bonier, F., Quigley, H., Austad, S.N. 2004. A technique for non invasively detecting stress response in cougars. Wildlife Society Bulletin. 32: 711 717. Boore, J.L., Macey, J.R., Medina, M. 2005. Sequencing and comparing whole mitochondrial genomes of animals. Methods in Enzymology. 395: 311 341. Davison, A., Birks, J.D.S., Brookes, R.C., Br aithwaite, T.C., Messenger, J.E. 2002. On the origin of faeces: morphological versus molecular methods for surveying rare carnivores from their scats. Journal of Zoology. 257: 141. Drummond AJ, Ashton B, Cheung M, Heled J, Kearse M, Moir R, Stones Havas S, Thierer T, Wilson A. 2008. Geneious v4.0, Available from http://www.geneious.com/ Ernest, H.B., Penedo, M.C.T., May, B.P., Syvanen, M., Boyce, W.M., 2000. Molecular tracking of mountain lions in the Yosemite Va lley region in California: genetic analysis using microsatellites and faecal DNA. Molecular Ecology. 9: 433 441. Farrell, L.E., Roman, J., Sunquist, M.E. 2000. Dietary separation of sympatric carnivores identified by molecular analysis of scats. Molecul ar Ecology. 9: 1583 1590. Fernandes, C., Ginja, C., Pereira, I., Tenreiro, R., Bruford, M., Santos Reis, M., 2008. Species specific mitochondrial DNA markers for identification of non invasive samples from sympatric carnivores in the Iberian Peninsula. Con servation Genetics. 9: 681 690. Godbois, I.A., Conner, L.M., Leopold, B.D., Warren, R.J. 2005. Effect of diet on mass loss of bobcat scat after exposure to field conditions. Wildlife Society Bulletin. 33: 149 153. Haag, T., Santos, A., De Angelo, C., Srbe k Araujo, A., Sana, D., Morato, R., Salzano, F., Eizirik, E. 2009. Development and testing of an optimized method for DNA
86 based identification of jaguar ( Panthera onca ) and puma ( Puma concolor ) faecal samples for use in ecological and genetic studies. Gene tica. of mitochondrial 12S RNA and cytochrome b sequences in the pantherine lineage of Felidae. Molecular Biology and Evolution. 12:4, 690 707. ship, T.L., Hirth, D.H., Tewes, M.E., Kilpatrick, C.W., Grassman, L.I. 2006. Kinship and social struction of bobcats ( Lynx rufus) inferred from microsatellite and radio telemetry data. Journal of Zoology. 269: 494 501. Johnson, W., Eizirik, E., Pecon S lattery, J., Murphy, W.J., Antunes, A., Teeling, E., Assessment. Science. 311:73 Kitano, T., Umetsu, K., Tian, W., Osawa, M., 2007. Two universal primer sets for species identif ication among vertebrates. International Journal of Legal Medicine. 121: 423 427. Mills, L.S., Pilgrim, K.L., Schwartz, M.K., McKelvey, K. 2000. Identifying lynx and other North American felids based on MtDNA analysis. Conservation Genetics. 1: 285 288. M ueller, U.G., Wolfenbarger, L. L. 1999. AFLP genotyping and fingerprinting. TREE. 14 (10): 389 394. Murie, O., Elbroch, M. 2005. Animal Tracks, 3 rd Ed. Peterson Field Guides. Houghton Mifflin. Murphy, M.A., Waits, L.P., Kendall, C. 2003. The influence of diet on faecal DNA amplification and sex identification in brown bears (Ursus arctos). Molecular Ecology. 12: 2261 2265. Murphy, M.A., Waits, L.P., Kendall, C. 2000. Quantitave evaluation of fecal drying methods for brown bear DNA analysis. Wildlife Society Bulletin. 28 (4): 951 957. Review of Genetics. 24: 467 489. Pandey, P.K., Dhotre D.P., Dharne, M.S., Khadse, A.N., Hiremath, U.I., Chaudhari, R.D., Patole, M.S., Sh ouche, Y.S. 2007. Evaluation of mitochondrial 12S rRNA gene in the identification of Panthera pardus fusca (Meyer, 1794) from field collected scat samples in the Western Ghats, Maharashtra, India. Current Science. 92 (8): 1129.
87 Parson, W., Pegoraro, K ., Niedersttter, H., Fger, M. Steinlechner, M., 2000. Species identification by means of the cytochrome b gene. International Journal of Legal Medicine. 114: 23 28. Paxinos, E., McIntosh, C., Ralls, K., Fleischer, R. 1997. A noninvasive method for dist inguishing among canid species: amplification and enzyme restriction of DNA from dung. Molecular Ecology. 6: 483 486. Pires, A.E., Fernandes, M.L. 2003. Last lynxes in Portugal? Molecular approaches in a pre extinction scenario. Conservation Genetics. 4: 525 532 Powell, W., Morgante, M., Andre, C., Hanafey, M., Vogel, J., Tingey, S., Rafalski, A. 1995. The comparison of RFLP, RAPD, AFLP and SSR (microsatellite) markers for germplasm analysis. Molecular Breeding. 2: 225 238. Ruiz Gonzlez, A., Rubines, J., Berdin, O., Gmez Moliner, B. 2008. A non invasive genetic method to identify the sympatric mustelids pine marten ( Martes martes ) and stone marten ( Martes foina ): preliminary distribution survey on the northern Iberian Peninsula. European Journal of Wildlife Research. 54: 253 261. The Southwest Biotechnology and Informatics Center, 2000. BLAST Tutorials. 2009. Srisamoot, N., Chaveerach, A., Nuchadomrong, S., Sattayasai, N., Chaveerach, P., Tanomtong, A., Pinthong, K. 2007. Genetic relationships amo ng wild felidae in Thailand using AFLP markers. Pakistan Journal of Biological Sciences. 10 (16): 2639 2645. Tamada, T., Kurose, N., Masuda, R. 2005. Genetic diversity in domestic cts Felis catus of the Tsushima Islands, based on mitochondrial DNA cytoc hrome b and control region nucleotide sequences. Zoological Science. 22: 627 633. Uphyrkina, O., Miquelle, D., Quigley, H., Driscoll, C., O'Brien, S.J., 2002. Conservation Genetics of the Far Eastern Leopard ( Panthera pardus orientalis ). Journal Heredity. 93: 303 311. Verma, S.K., Singh, L. 2003. Novel universal primers establish identity of an enormous number of animal species for forensic application. Molecular Ecology Notes. 3: 28 31. Zhang, Z., Willson, R., Fox, G. E. 2002. Identification of charac teristic oligonucleotides in the bacterial 16S ribosomal RNA sequence dataset. Bioinformatics. 18:2, 244 250.
88 Appendix I: Extraction Protocol for the QIAmp DNA Stool Kit 1. Turn on the water bath at a power level of approximately 6, and adjust it occasionally so that the water in the wells stays at 70 C. 2. Clean glass plate and scraping utensils with SoftScrub with b leach, 70% EtOH, and RNAse away. Make sure that the plate is dry. Have 2 empty 15ml vials and a box of ice on hand. 3. Remove sample from 80 C freezer using a paper towel to help insulate the scat from body heat. Immediately begin chiseling the outer l ayer of scat off. Scrape the chunks of scat into one of the 15 ml vials until approximately 1 gram has been collected. Discard the used scat and place the sample on ice. 4. Weigh the sample vial against the empty vial and record the mass of scat sampled. 5. Add 10 ml of buffer ASL to each gram of scat. (9ml for 0.9g) Vortex vigorously for 1 min or until the scat is thoroughly homogenized. 6. Pipet 2 ml of lysate into a 2 ml microcentrifuge tube. The tip of the disposable pipet may be cut off in order to suck up viscous samples. 7. Centrifuge for 1 min, 14000 rpm, room temp. 8. Pipet 1.4 ml of supernatant into a new 2 ml tube. 9. Add an InhibitEX tablet directly from the packaging to the tube. Top off the sample until the total volume = 2ml. Vortex immediately unt il tablet is completely suspended. (up to 1 minute). 10. Incubate at room temperature for another minute. 11. Centrifuge for 6 min, 14000 rpm, room temp. 12. Immediately after the centrifuge stops, pipet all of the supernatant into a new 1.5 ml microcentrifuge tube. 13. Centrifuge for 3 minutes, 14000 rpm, room temp. 14. While the centrifuge is running, pipet 25 l of proteinase K into a new 2 ml tube. 15. When the centrifuge stops, pipet 600 l of supernatant into the 2 ml microcentrifuge tube with the proteinase K. Add 600 l of Buffer AL and vortex until thoroughly mixed. 16. Incubate in the water bath at 70 C for 10 minutes.
89 17. After incubation, add 600 l of 99.9% ethanol to the lysate and mix by vortexing. 18. Prepare a spin column in a collection tube. 19. Add 60 0 l of lysate to the spincolumn without moistening the rim. Centrifuge for 1 minute at 14000 rpm, room temp. Repeat 3 times until the lysate has all passed through the column. Discard the filtrate and collection tube every time. 20. Add 500 l of Buffer A W1 to the spin column and centrifuge 1 min, 14000 rpm, room temp. Discard collection tube and flow through. 21. Add 500 l Buffer AW2 to the spin column and centrifuge for 3 min, 14000 rpm, room temp. Discard collection tube and flow through. 22. Centrifug e spin column in a new collection tube for 1 min, 14000 rpm, room temp to get rid of any residual Buffer AW2 (optional step) 23. Add 200 l of Buffer AE into spin column in a clean 1.5 ml tube and incubate for 1 minute before centrifuging for 1 minute, 140 00 rpm, room temp. 24. Transfer eluate to a new 1.5 ml tube and label it if the cap broke off of the previous tube. 25. Store in 20 C freezer.