Amphibians of Northern KwaZulu-Natal : a phylogenetic study

(1)

Amphibians of Northern KwaZulu-Natal:

A Phylogenetic Study

JE Reeder

orcid.org 0000-0003-0525-5857

Dissertation submitted in fulfilment of the requirements for the

degree

Master of Science in Environmental Sciences

at the

North-West University

Supervisor:

Prof LH du Preez

Co-supervisor:

Dr DJD Kruger

Co-supervisor:

Mr EC Netherlands

Graduation May 2019

25995642

(2)

ii

Declaration

I, Jani Elsabe Reeder, declare that this dissertation is my own unaided work, except otherwise acknowledged in the text. This dissertation is being submitted for the

degree of Masters of Science in Environmental Sciences to the North-West University, Potchefstroom Campus. It has not been submitted for any degree or

examination in any other university.

Jani Elsabe Reeder 20 November 2018

(3)

iii

Dedication

This dissertation is dedicated to my loving mother and father, whose love and

support have carried me through my toughest times and greatest

achievements.

“The one process now going on that will take millions of years to correct is

the loss of genetic and species diversity by the destruction of natural habitats.

This is the folly our descendants are least likely to forgive us.”

(4)

iv

Acknowledgements

This project would not have been possible without the guidance and support from numerous people and institutions, and I would like to thank them in no particular order. My deepest appreciation goes out to Prof Louis du Preez, whom without, I would not be where I am today. Prof, your passion inspires me and the rest of your team to be the best that we can be and to never give up on our dreams. You mean so much to so many of your students; words cannot describe your true value.

To my Co-supervisors, Dr. Donnavan Kruger and Edward Netherlands, thank you both for all your hard work and dedication. Ed, whom so patiently took me under his wing, and Donnavan, the person that always made me smile again, thank you, I am truly grateful.

To the rest of the African Amphibian Conservation Research Group (AACRG), thank you all for your support and help with this project. I am truly blessed to have met such wonderful people. I would particularly like to thank my team member and dear friend, Ferdi de Lange, without you, I would never have had this opportunity. Also, Wentzel Pretorius, whom has become like a brother to me, and Fortunate Phaka, whom I can always depend on, thank you for all the comradery and support throughout this journey. To Olena Kudlai, thank you for your assistance, guidance, and friendship, I am truly grateful to have met you during this journey. To Willie Landman, Roman Svitin, Natasha Kruger, Nadine Lepart, Este Mathew, Ruan Gerber, Prof Che Weldon and Oriel Moeti Taioe, thank you for all the assistance and support.

To all of my friends and family: the journey was long and at times difficult, but your support carried me through it, thank you. I especially want to thank my parents, Danielle and Carste Reeder, whom have supported and encouraged me throughout this journey. To my husband, Dennis Quinn, you have been my rock and I am forever thankful for your love and support.

Special thanks goes out to the South African National Biodiversity Institute’s (SANBI) Foundational Biodiversity Information Programme (National Research Foundation Grant Holder Linked Bursary for Grant UID: 98114) for funding this project and to the North-West University (NWU Masters Progress Bursary, and NWU Masters Bursary) for financial assistance towards this degree.

(5)

v

I would also like to thank the Ezemvelo KZN Wildlife team for allowing us to do research in northern KwaZulu-Natal (OP 4092/2016; see Appendix 1). Special thanks to Leonard Muller (Tembe Game Reserve), Catharine Hanekom (District Ecologist North East: Umkhanyakude) for their assistance during field work and to the Ndumo Game Reserve and Bonamanzi Pivate Game Reserve for allowing us to conduct research in the respected reserves. Thank you to the Anim-Care ethics committee of the North West University for providing me ethics clearance, NWU-00006-14-A3.

(6)

vi

Abstract

Northern KwaZulu-Natal (KZN) is a biodiversity hotspot containing the Maputaland centre of endemism and boasts the highest diversity of frog species in South Africa. The tropical to subtropical climate with relatively high rainfall coupled with its unique geography have resulted in a mosaic of habitats and proliferation of frog species in this region. KwaZulu-Natal has become a priority area for conservation due to increasing environmental pressures from population growth and development. Effective conservation and management of biodiversity depend largely on our knowledge of taxonomy. Evolutionary patterns of convergence and parallelism, occurring among relatively closely related taxa, present a great challenge to amphibian taxonomy and diversity research. This not only leads to challenges in identification resulting from similarities between species, but can also completely mask cryptic species. DNA barcoding is a relatively modern tool that uses a short standard gene fragment to identify specimens. Additionally, it can also lead to the discovery of new or cryptic species and provides standardisation for species description. This study aimed to contribute to a better understanding of frog diversity by expanding the DNA barcoding reference library for the frogs of northern KZN and screening for possible cryptic species. Due to a lack in reference sequences of the mitochondrial cytochrome c oxidase subunit I marker, the complementary 16S rDNA marker was also included in this study for phylogenetic analysis. A total of 350 individual frog and tadpole voucher specimens were collected. All voucher specimens, tissue samples, and photographs were processed and accessioned. A subset of 141 specimens was chosen for molecular analysis. The successfully obtained sequences, 224 in total, represent 35 species and two unidentified

Breviceps sp. specimens, accounting for 69% of the known frog species occurring within the

study area. Only the 16S marker was obtained for thee of these species. All sequences and data were uploaded and added to the Biodiversity of Life Database under project code LDPJR. The Barcode Index Number (BIN) system was used to corroborate identifications of sequenced specimens. Identifications were also supported by the acquisition of additional reference sequences (COI and 16S) from BOLD and GenBank and their subsequent analysis through the construction of Neighbour Joining (NJ) trees. Screening for the presence of possible cryptic species was performed through the employment of the Automatic Gap Discovery algorithm, the construction of NJ trees, and the evaluation of genetic divergence between the elucidated groups. The possible presence of cryptic species was found in ten genera. Many other taxa were shown to need further investigation to strengthen morphological characterisation, establish species genetic boundaries, correct distribution estimates and correct identification of reference sequence records. While DNA barcoding alone does not provide enough information to identify and describe new species, it does flag taxa in need of further investigation and provides a rapid identification system, thus aiding in the prioritisation of research efforts and efficiency.

(7)

vii

Opsomming

Noord-KZN is 'n biodiversiteits meka wat die Maputaland-sentrum van endemisme bevat en die hoogste diversiteit van paddaspesies in Suid-Afrika. Die tropiese tot subtropiese klimaat met relatief hoë reënval tesame met unieke geografie het gelei tot 'n mosaïek van habitats en verspreiding van paddaspesies in hierdie streek. KwaZulu-Natal het 'n prioriteitsarea vir bewaring geword as gevolg van toenemende omgewingsdruk van bevolkingsgroei en ontwikkeling. Doeltreffende bewaring en bestuur van biodiversiteit berus grootliks op ons kennis van taksonomie. Evolusionêre patrone van konvergensie en parallelisme, wat voorkom onder relatief nou verwante taksa, bied 'n besondere groot uitdaging vir amfibiese taksonomie en diversiteitsnavorsing. Dit lei nie net tot uitdagings in identifikasie as gevolg van ooreenkomste tussen spesies nie, maar kan ook kriptiese spesies masker. DNS kodering is 'n relatief moderne tegniek wat van 'n kort standaard geen fragment gebruik maak om eksemplare te identifiseer. Daarbenewens kan dit ook lei tot die ontdekking van nuwe of kriptiese spesies en bied standaardisering vir spesiebeskrywing. Hierdie studie het ten doel om by te dra tot 'n beter begrip van padda-diversiteit deur die DNS strepieskode verwysingsbiblioteek vir die paddas van die noord KZN uit te brei en vir moontlike kriptiese spesies te ondersoek. Weens 'n gebrek aan verwysingsvolgorde van die mitochondriale sitochroom-oksidase subeenheid I merker, was die komplementêre 16S rDNS merker ook ingesluit in die studie vir filogenetiese analise. Ons het 'n totaal van 350 individuele paddas en kuikens ondersoek. Alle eksemplare, weefselmonsters en foto's is verwerk en toeganklik. 'N Subset van 141 eksemplare is gekies vir molekulêre analise. Die suksesvol verkrygde sekwense, 224 in totaal, verteenwoordig 35 spesies en twee onbekende Breviceps sp. eksemplare, wat 69% van die bekende paddaspesies wat binne die studiegebied voorkom verteenwoordig. Slegs die 16S-merker is vir drie van hierdie spesies verkry. Alle opeenvolgings en data is opgelaai en bygevoeg aan die ‘Biodiversity of Life Database’ onder projekkode LDPJR. Die Barcode Index Number (BIN) stelsel is gebruik om identifikasies van eksemplare te bevestig. Identifikasies is ook ondersteun deur die verkryging van addisionele verwysingsreekse (COI en 16S) van BOLD en GenBank en hul daaropvolgende analise deur die konstruksie van NJ-bome. Sifting vir die teenwoordigheid van moontlike kriptiese spesies is uitgevoer deur die aanwending van die ‘Automatic Barcode Gap Discovery’ algoritme, die konstruksie van NJ-bome, en die evaluering van genetiese afwykings tussen die verklaarde groepe. Die moontlike teenwoordigheid van kriptiese spesies is in tien genera gevind. Baie ander taksa is in nood van verdure odersoek bevind om morfologiese karakterisering te versterk, genetiese grense van spesies te vestig, korrekte spesie verspreiding te bevestig en korrekte identifikasie van verwysings eksemplare te registreer. Terwyl DNS kodering alleen bevat nie genoeg inligting om nuwe spesies te identifiseer en te beskryf nie, maar kan wel groepe uitwys wat verder ondersoek benodig en vinnige identifiseering voorsien. Dus kan navorsings en bewaring pogings geprioritiseer word.

(8)

viii

3.2. Species accounts ... 27 3.2.1. Arthroleptidae ... 27 3.2.2. Brevicepitidae ... 28 3.2.3 Bufonidae ... 30 3.2.4 Hemisotidae ... 33 3.2.5 Hyperoliidae ... 35 3.2.6 Microhylidae ... 39 3.2.7 Phrynobatrachidae ... 40 3.2.8 Ptychadenidae ... 42 3.2.9 Pipidae ... 46 3.2.10 Pyxicephalidae ... 48 3.2.11 Rhacophoridae ... 52

Chapter 4: Discussion and Conclusion ... 54

4.1. Establishing a DNA barcoding database for northern KZN ... 55

4.1.1. Collection and sequencing of specimens ... 55

4.1.2. Contribution to and utilisation of DNA barcoding reference libraries ... 56

4.1.3. The efficacy of primers ... 58

4.1.4. Using 16S rRNA as a complementary marker for DNA barcoding ... 58

4.2. Species identification using DNA barcoding and the presence of cryptic species in northern KZN ... 59

4.2.1. Species identification using DNA barcoding ... 59

4.2.2. Taxa containing possible cryptic species ... 60

4.2.3. Other taxa requiring further investigation ... 63

4.3. Recommendations for future research and conclusion ... 64

4.3.1 Recommendations for future research ... 64

4.3.2. Conclusion ... 64

References ... 66

Appendix 1: Copy of permit ... 73

Appendix 2: Locality maps ... 77

Appendix 3: BOLD specimen record and BIN page examples ... 85

(10)

x

List of figures:

Figure 1.1.1: Example of subtle morphological differences between two previously cryptic

species.

Figure 1.3.1: Map showing the anthropogenic transformation of KZN from 1994 to 2011. Figure 2.1.1: Study area map showing sampling localities and survey periods

Figure 2.1.2: Examples of sampling sites

Figure 2.2.1: Collecting techniques and field processing

Figure 3.2.1.1: COI Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of six

Arthroleptidae species.

Figure 3.2.1.2: 16S Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of three

Arthroleptidae species.

Figure 3.2.2.1: COI Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of Breviceps

species.

Figure 3.2.2.2: 16S Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of Breviceps

species.

Figure 3.2.3.1: COI Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of Bufonidae

species.

Figure 3.2.3.2: 16S Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of Bufonidae

species.

Figure 3.2.4.1: COI Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of Hemisus

species.

Figure 3.2.4.2: 16S Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of Hemisus

species.

Figure 3.2.5.1: COI Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of

Hyperoliidae species.

Figure 3.2.5.2: 16S Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of

Hyperoliidae species.

Figure 3.2.6: a) COI and b) 16S Neighbor-Joining Phylogenetic Trees. Phylogenetic

analysis of Phrynomantis species.

Phrynobatrachus species.

Ptychadenidae species. 3 8 12 13 15 27 29 28 30 32 34 33 36 31 38 39 41 41 44

(11)

xi

Ptychadenidae species.

Figure 3.2.9.1: COI Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of Xenopus

species.

Figure 3.2.9.2: 16S Neighbor-Joining Phylogenetic Tree. Phylogenetic analysis of Xenopus

species.

Pyxicephalidae species.

Chiromantis species. 47 49 47 46 50 52

(12)

xii

List of tables:

Table 2.3.1: Primers used in this study.

Table 3.1.1: Comprehensive list of species sampled and analysed. Table 3.1.2: List of specimens selected for molecular analysis.

Table 3.2.2.1: Mean divergence (K2P distances, %) for 16S sequences between species/

clades of Breviceps.

Table 3.2.4.1: Mean divergence (K2P distances, %) for 16S sequences between and within

species/ clades of Hemisus.

Table 3.2.5.1: Estimates of evolutionary divergence between 16S sequences of H.

tuberilinguis specimens.

Table 3.2.7.1: Mean divergence (K2P distances, %) of COI sequences between and within

species/ clades of Phrynobatrachus.

Table 3.2.7.2: Mean divergence (K2P distances, %) of 16S sequences between and within

species/ clades of Phrynobatrachus.

species of Ptychadenidae.

species of Pyxicephalidae.

species of Chiromantis. 22 34 20 17 30 37 42 42 45 51 55

(13)

1

Chapter 1: Introduction

(14)

2

1. Introduction

1.1. Amphibian diversity

1.1.1. Introduction to amphibians

Often inconspicuous and overlooked, amphibians play a crucial role in the natural world and contribute significantly to both the academic and research sciences. Modern amphibians are a monophyletic group descending from a common ancestor in the Permian period. Amphibians represent the all-important evolutionary transition of aquatic tetrapods onto land and are the only living vertebrates that have development from water to land in both their ontogeny and phylogeny.

From an ecological perspective, amphibians play an integral role in ecosystem function and health (Duellman, 1999). As predators, both adults and larvae are responsible for controlling insect populations and consequently their associated diseases; as prey, amphibians are an essential food resource for numerous other taxa, including other amphibians. This integral role connects both lower and higher trophic levels and aquatic and terrestrial environments.

Due to their biphasic lifecycle, involving aquatic eggs and larvae that metamorphose into terrestrial or semiaquatic juveniles and adults, amphibians are more vulnerable to environmental change (Stuart et al., 2004). This vulnerability coupled with their highly specific but variable habitat requirements (Indermaur & Schmidt, 2011) makes amphibians good bio-indicators of environmental health (Du Preez & Carutthers, 2017).

Amphibians are facing many challenges leading to global extinctions and population declines. Of the worlds amphibian species, 32% are known to be threatened or extinct (IUCN/SSC, 2008), and at least 43.2% of amphibian species are experiencing some form of population decrease (Stuart et al., 2004). In his study, Alroy, 2015, estimated that since the 1970s approximately 200 amphibian extinctions have occurred and suggested that in the next century hundreds more may be lost. Approximate numbers of threatened or extinct amphibians are likely underestimated, as new species are frequently discovered, and some may have gone extinct prior to discovery.

1.1.2. Describing amphibian species diversity

Effective conservation and management of biodiversity depend largely on our knowledge of taxonomy. Morphological characterisation has been the primary method used for species description and classification throughout history (Adams et al., 2004) and remains the most general criterion used to define amphibian species

(15)

3

(Vences & Wake, 2007). An especially great challenge to amphibian taxonomists is when patterns of convergence and parallelism occur among relatively closely related taxa (Wiens et al., 2003), leading to a lack of morphological diagnostic traits, which not only make identifications difficult, leading to erroneous identifications, but can also completely mask species diversity due to the presence of cryptic species.

Cryptic species are morphologically indistinguishable species blanketed under one species assignment (For example, Breviceps carruthersi and B. passmorei under B. adspersus, Minter et al., 2017, Figure 1.1.1.). Cryptic species tend to occur with more frequency in tropical regions with high species richness. Undescribed species are likely to be of conservation concern as they will be limited in their distribution (Measey, 2011).

The advent of molecular techniques has enabled taxonomists to re-evaluate the systematics of amphibians worldwide, leading to an increase in described species. The molecular revolution has brought about significant advancements in amphibian taxonomy (Measey, 2011), but data from non-molecular sources is by no means obsolete, and the combination of various datasets is preferable (Cannatella, 2007). Morphological and acoustical analyses are deemed required techniques for inclusion in amphibian taxonomy where possible (Measey, 2011). Using both molecular and non-molecular methods requires specialised training, funding, and access to curated collections and reference data, all of which can be difficult to obtain in the face of the taxonomic impediment.

The term ‘taxonomic impediment’ was first used by the International Union of Biological Sciences / Diversitas, to describe the taxonomic crisis characterised by a shortage of taxonomic specialists and curators within several groups and geographic regions and by insufficient funding for such work (Mallet & Willmott, 2003). This problem is observed in South Africa which has contributed the highest number of publications describing new species in all of Africa (Grieneisen et al., 2014), but the majority of species have been described by foreign taxonomists (Hamer, 2013).

South Africa has meagre numbers of professional herpetologists to carry out the work required to ensure conservation of amphibians (Measey, 2011). This lack of professional

Figure 1.1.1.: Example of subtle morphological differences between (A) Breviceps carruthersi, and (B) Breviceps

passmorei, two cryptic species formerly blanketed under B. adspersus. Photos from Minter et al., 2017.

(16)

4

herpetologists may be due to a lack of incentives to pursue taxonomic work and training, such as a lack in positions, funding, low salaries etc., but may also be partly because there has been a general move away from descriptive and revisionary taxonomy within the country towards more phylogenetic analyses (da Silva & Willows-Munro, 2016).

As a means to revitalise traditional taxonomy, and help overcome the constraints imposed by the taxonomic impediment, alternative and complementary methods have been promoted. Molecular taxonomy through DNA barcoding (Hebert et al., 2003), is one of these methods.

1.2. DNA barcoding amphibians

1.2.1. Introduction to DNA barcoding

DNA barcoding, a relatively new molecular technology developed in 2003 by Dr Paul Hebert at the University of Guelph, has helped to rejuvenate taxonomic research (Vernooy

et al., 2010). It is used as a standard tool for the rapid identification of animals and plants

at the species level (Brownlee, 2004). DNA barcoding has received increased acceptance because it is simple, affordable and fast in the discovery and identification of biodiversity (Radulovici et al., 2010; Padial & De La Riva, 2007). For animals, the 650 base-pair fragment of the 5’ end of the mitochondrial cytochrome c oxidase 1 (COI) gene has been designated as the barcode region (Hebert et al., 2003). This region mutates fast enough to distinguish closely related species but slowly enough that individuals within a species have similar barcodes (Hebert et al., 2003), enabling the discrimination of cryptic species and also revealing phylogeographic structures within a species (Hebert et al., 2003).

DNA barcoding involves the extraction, amplification, and sequencing of the COI gene fragment, followed by a comparison with other sequences previously deposited in a database. Specimens are identified by matching the new sequence with sequences of known identity already in a reference database (Hebert et al., 2003). This new sequence can also lead to a novel barcode sequence for a given species (i.e. a new haplotype or geographical variant). Additionally, if a sequence shows divergence from other sequences above a certain threshold, it can suggest the existence of a newly encountered species. The analysis of DNA barcoding data is usually performed through clustering and tree-based methods.

In contrast to other taxa, especially fishes and birds, DNA barcoding of amphibians is in a very early stage (Vences & Wake, 2007). DNA barcoding of amphibians based on COI is not fundamentally different from that of other animal groups and holds the same promise, but specifics to be kept in mind are the old age of many amphibians leading to the potential presence of very deeply diverged mitochondrial lineages. This divergence exacerbates the problem of primer failure even within species or species complexes (Vences et al., 2012), and can lead to the over interpretation of diversity where reference

(17)

5

data are lacking. Thus, it is necessary to have a comprehensive COI reference databases for successful and efficient species identification.

1.2.2. Benefits and limitations of DNA barcoding

One of the beneficial uses of DNA barcoding is in the field of taxonomy, where it has become a standard tool for rapid species identification and assists in discoveries of new species, rejuvenating taxonomic research (Vernooy et al., 2010). This benefit is especially applicable in cases where traditional morphological identifications are not sufficient or subjective, as is the case with many amphibian species. In such instances barcoding helps to: identify cryptic species (morphologically indistinguishable but incapable of interbreeding); uncover phenotypic plasticity (variable phenotypes in different environments); identify multiple life-stages (e.g. amphibian larvae) and; identify damaged or partial specimens (e.g. stomach contents, butchered meat). Barcoding can expedite the taxonomic process by identifying groups of species that require more detailed studies (Goldstein and DeSalle, 2011).

An important application of DNA barcoding in amphibians is the identification of larvae (Vences et al., 2005a) and immature individuals. The morphological identification of tadpoles can be tedious and requires specialised knowledge. In cases where tadpoles are the only life-stage present their identification is crucial for diversity studies.

As much as DNA barcoding has its advantages, it also has some limitations, and the use of DNA Barcoding has been a widely discussed matter. These limitations mainly are: ‘the single-locus approach’; the ‘universal barcode gap’; the cost and redundancy of Sanger sequencing and disconnect between data generators and end-users (da Silva & Willows-Munro, 2016). A problem more specific to amphibians is the high variability of priming sites that hinder the application of universal primers (Vences et al., 2005a).

Because the mitochondrial genome is usually maternally inherited, mtDNA genes may be shared between taxa due to hybridisation or incomplete lineage sorting (da Silva & Willows-Munro, 2016). This sharing can lead to discordance between mitochondrial and nuclear genealogies, with DNA barcoding failing to recognise such taxa as different species, whereas nuclear genes may, resulting in different phylogenies (da Silva & Willows-Munro, 2016). This discordance can have serious conservation implications as a result of different interpretations of species status, rarity and conservation importance. To solve the problem of discordance between barcoding and nuclear phylogenies (especially within amphibians and understudied taxa) Vences et al., (2005a), suggests the use of multiple mitochondrial and nuclear genes as DNA barcoding markers to complement COI. For DNA barcoding to delimit between taxa, a universal barcode gap (i.e. a distinct difference in divergence) needs to be recognised between taxa, which would be defined by interspecific variation exceeding intraspecific variation with species being delineated at a standard threshold. A standard limit of sequence divergence or a ‘universal barcode gap' with which to delimit species has not been established (da Silva & Willows-Munro, 2016). Most barcoding researchers have adopted a 2% divergence threshold as this cut-off value (e.g., Hebert et al., 2003; Ward et al., 2009; Pereira et al., 2013) and is the preliminary

(18)

6

threshold employed by BOLD (Barcode of Life Database; Ratnasingham and Hebert, 2007). For amphibians, Vences et al., 2005a, suggested a divergence threshold of 10% for the COI gene fragment and 5% for the 16S gene fragment for candidate species recognition.

Proponents of DNA barcoding argue that its objective is merely clarification and delimitation of new taxa by highlighting groups that are genetically distinct (Hebert & Gregory, 2005). DNA barcoding does not provide sufficient information to describe a species, but can reveal genetically distinct lineages and gives species-level resolution in 95 to 97% of cases (Hebert et al., 2004, Ward et al., 2005), which ,combined with other sources of data, can contribute to taxonomical clarification (Hebert & Gregory, 2005).

1.2.3. Barcode reference libraries

To strengthen DNA barcoding research, coordinate global efforts and broaden geographic and taxonomic coverage, the International Barcode of Life Project (iBOL; http://www.ibolproject.org) was launched in October 2010. South Africa has committed to assisting in iBOL’s mission, and in 2011 a formal South African node was established on iBOL (da Silva & Willows-Munro, 2016).

The Barcode of Life Data System (BOLD; http://boldsystems.org) is not only the barcode resource library for the iBOL, but also a resource tool that is freely available to the public to assist in all stages of barcoding research, from specimen collection to data submission (Ratnasingham and Hebert, 2007). BOLD requires that photographs, GPS coordinates, and other collection information accompany all sequence data; thus providing a tightly validated barcode library. As of August 2017, BOLD contained barcode records for over 5.5 million specimens and had recognised close to 514 000 animal barcode index numbers (BINs) (boldsystems.org). BINs are sequence clusters that closely approximate species (or operational taxonomic units) recognised through sequence variation in the COI DNA barcode region (Ratnasingham and Hebert, 2013). Of the 2697 public BINs for Anurans (boldsystems.org, 23rd March 2018), only 1608 represent formally described species with the remaining 1089 BINs possibly representing species not yet described; thus requiring further validation through additional molecular markers and taxonomic assessments. Another source of barcoding sequence data is GenBank. GenBank is the largest open-access annotated repository of publicly available genetic sequence data. BOLD shares a tightly integrated data exchange pipeline with NCBI (GenBank) (Boldsystems, 2013).

1.2.4. The use of complementary markers

Even though the COI barcoding gene has been found to identify 94% of amphibian species correctly (Smith et al., 2007), the difficulty with the amplification of COI sequences in amphibians due to their high variability (Vences et al., 2005a) has led to the favoured use of the mitochondrial 16S rRNA gene fragment for molecular species identification. This favourability, in turn, has led to the consequential scarcity of amphibian COI DNA barcodes (Vences et al., 2012), further discouraging researchers from using the COI gene.

(19)

7

The mitochondrial 16S rRNA gene fragment has been suggested as a complement to COI for specific barcoding applications in amphibians (Vences et al., 2005b).

The 16S gene fragment has three advantages over COI for use in amphibians: the existing 16S reference database for amphibians is much more taxon-rich than for COI; truly universal primers exist for 16S; and the phylogenetic signal in the 16S gene is stronger, thus the identification of species to higher clades is more reliable, even with incomplete taxon sampling in the reference database (Vences et al., 2012). Therefore, at present, 16S is more reliable to assign species that are not represented in the COI database to at least the correct major lineages, but as the COI reference database becomes comprehensive, the disadvantages of using COI will disappear (Vences et al., 2012), especially with more robust primers now available for amphibians (Che et al., 2012).

1.3. DNA barcoding the frogs of northern KZN

1.3.1. Frogs of South Africa

The class Amphibia consists of three orders: Anura (frogs) with 7024 species; Urodela (salamanders and newts) with 717 species; and Gymnophiona (caecilians) with 209 species (Amphibiaweb, 2018). The Anurans are the only order of amphibians that occur in South Africa.

South Africa’s landscape, ranging from desert to tropics, is the third most biologically diverse in the world (Mittermeier et al., 1999). Despite having an average annual rainfall well below that of the global average (Cowan, 1995), this diverse landscape has led to the evolution of a rich diversity of frog species adapted to a wide range of ecological niches (Du Preez & Carruthers, 2017). This diversity includes not only a wide variety of shapes and forms, but life histories, habitats, calls, colours and phylogenetic diversity (Measey et

al., 2011). Due to their permeable skins, biphasic life cycle and ectothermic nature, more

frog species are found in warm and moist climates; thus species richness increases from west to east across Southern Africa. Other determinants of distribution patterns are centres of origin and range restriction (see Du Preez & Carruthers, 2017). All of these determinants have led to a high species diversity but low endemicity in the north-east of South Africa, and a low diversity but high endemicity in the south-west of South Africa. In South Africa, frog species distributions have been well documented (Du Preez & Carruthers, 2017; Measey, 2011; Minter et al., 2004; Poynton, 1964). Thus, the likelihood of discovering new morphologically distinct species is low, but the discovery of cryptic species currently blanketed under described species is plausible, as is exemplified by the recent descriptions of eight new species from three genera in the last year (Turner & Channing, 2017;Channing et al., 2017; Minter et al., 2017). According to Measey et al., 2011, there are many known and expected cryptic species still to be described, these species may be threatened and will require further research.

(20)

8

South Africa is ranked fourth regarding the number of threatened amphibian species in the Afrotropical realm (Stuart et al., 2008). Out of the 126 species that have been assessed for conservation status, six species are listed as critically endangered, nine as endangered, one as vulnerable, 12 as near threatened, and five as data deficient (IUCN, 2018). Despite these numbers, South African amphibians appear to be faring relatively well in relation to the global proportions, with proportionately fewer species in the Threatened category or the Data Deficient category. This disproportion may be due to enigmatic declines having a more significant effect at the global level, such as the fungal disease chytridiomycosis (Measey et al., 2011) and does not reflect the seriousness of local threats. Well over 50% of South Africas’s threatened species are being affected by habitat loss caused by agricultural activities, urban development, and biological resource use (Measey et al., 2011).

KwaZulu-Natal forms the central component of the Maputaland-Pondoland-Albany biodiversity hotspot and is internationally recognised for its high levels of species richness and endemism (Di Minin et al., 2013). The tropical to subtropical climate, relatively high rainfall and steep escarpment of KZN have resulted in a mosaic of diverse habitats leading to a proliferation of frog species in this area. KwaZulu-Natal hosts the highest amphibian richness in South Africa (Measey, 2011), with a total of 70 frog species, of which, 10% are endemic to the province (Armstrong, 2001), and account for 42% of the total diversity of frog species occurring in Southern Africa (see Du Preez & Carruthers, 2017).

KwaZulu-Natal also hosts the second highest number of threatened frog species in the country (Measey, 2011) and has become a priority area for conservation due to increasing environmental pressure from population growth and development. An average of 1.2% of the natural landscape in KZN has been transformed per annum since 1994, with the significant drivers of habitat loss being agriculture, timber plantations, the built environment, dams, and mines (Figure 1.3.1., Jewitt et al., 2015). KwaZulu-Natal is the province with the second largest population (19.7%) in South Africa, with approximately 11.4 million people (SSA, 2018).

Figure 1.3.1.: Map showing anthropogenic habitat transformation in KZN from 1994 to 2011. Study area

(21)

9

Northern KZN is a biodiversity hotspot that comprises the Maputaland centre of endemism (Van Wyk, 1994). This area is located at the southernmost extension of the tropical climate, and the northernmost extent of subtropical climate in Africa. The coastal region is recognized as a transition zone between tropical and temperate herpetofauna (Bruton & Haacke, 1975). These factors influence the areas high biodiversity and habitats.

Since the last comprehensive frog survey in northern KZN more than 30 years ago and subsequent publication of “A Review of the Amphibians of Natal” (Lambiris, 1989), northern KwaZulu-Natal underwent a complete transformation, both in terms of the number of people residing there and in terms of habitat alteration. Human-induced landscape change and the clearing of land for monocultures, such as sugar cane plantations, can be seen throughout the area and are major drivers of biodiversity loss.

1.3.2. DNA barcoding the Anurans of SA

Until recently, the vast majority of barcoding had been performed by researchers in developed countries and on particular taxonomic groups, even though most of the Earth's biodiversity is found in the tropical and subtropical regions. This disproportion is primarily due to researchers in developed countries having the resources and capacity to carry out such work (Vernooy et al., 2010). It is thought that as many as 80 000 animal species (>45 000 insects) are left to be discovered or described in South Africa (Hamer, 2013). Of South Africa’s known animal species, only 2.3% are represented in BOLD, with fish being the best represented taxonomic group (approximately 36%), followed by birds (5.4%), and mammals (4.9%) with all other taxonomic groups having less than 2% representation in the database (da Silva & Willows-Munro, 2016). Currently, in BOLD, there are 262 public records of Anurans from Southern Africa (South Africa, Namibia, Botswana, Zimbabwe, Mozambique) forming 99 BINS, while there are 165 described Anuran species in the same area.

After Gauteng, KZN has contributed the highest number of animal samples (11 863) (da Silva & Willows-Munro, 2016), but despite the known Anuran species richness of this area, only 22sequences from KZN representing 21 species were available on BOLD prior to this study. Not only is this concerning from a purely descriptive standpoint, but there is also a high potential that cryptic species exist in the area. With such a diverse complexity of habitats, it is possible that a high number of refugia were available in this area, increasing the potential for speciation over time (Alexander et al., 2004).

Also, the ‘mountain passes are higher in the tropics' (MPHT) hypothesis postulates that reduced climate variability at low latitudes, as is the case in northern KZN, should select for narrower thermal tolerances, lower dispersal and smaller elevational ranges compared with higher latitudes. This higher number of refuges and reduced climate variability could have led to increased species richness in northern KZN, but that species richness increase may have mainly been cryptic because physiological and dispersal traits which isolate populations might not correspond to morphological differences (Gill et al., 2016).

This high possibility of cryptic species being present in the area illustrates the need for DNA barcoding work to be done in northern KZN. Firstly to extend the reference library of

(22)

10

baseline data for future work to be possible, and secondly for the screening of possible cryptic species in the area.

1.4. Research question, aim and objectives

1.4.1. Research question

Since northern KZN is a low-lying area with reduced climate variability the research question of the present study was to determine whether northern KZN may house an even greater amphibian species richness than currently recognised.

1.4.2. Aim

Contribute to a better understanding of frog diversity in northern KZN by conducting field surveys to collect voucher specimens and tissue samples, sequence and analyse DNA barcoding data, and screen for the potential presence of cryptic species.

1.4.2. Objectives

1. Collect specimen vouchers, tissue samples, and photographs to expand the existing collection databases for frog species of northern KZN.

2. Sequence the COI and 16S markers for molecular analysis.

3. Establish a comprehensive dataset of DNA barcoding records for frog species of northern KZN by confirming species identifications through molecular analysis. 4. Analyse sequences to screen for the potential presence of cryptic species.

(23)

11

Chapter 2: Materials and Methods

(24)

12

2. Materials and Methods

2.1. Study area and sampling effort

The present study took place in the north-eastern section of the province of KwaZulu-Natal, South Africa. The study area (Figure 2.1.1) is bordered by Mozambique to the north, Swaziland to the west and the Indian Ocean to the east, and extends south from Jozini to Richard's Bay. The study area incorporates the 332,000-hectare iSimangaliso Wetlands Park World Heritage Site.

Figure 2.1.1: Map showing study area with sampling localities and survey periods. o

(25)

13

In north-eastern KZN the highest average rainfall and temperature occurs during the spring (September to November) and summer (December to March) months. To increase the likelihood of encountering a diversity of frogs sampling trips corresponded closely with these seasons. Sampling trips occurred during the following periods, ranging from four days to one month (Figure 1):

 November 2015 (16th - 25th)

 February 2016 (4th - 11th)

 March 2016 (14th – 17th)

 November/December 2016 (11th Nov – 11th Dec)

Survey efforts were focused at potential breeding sites, identified through the use of aerial imagery, topographic maps and local knowledge (Figure 1). Prior to nocturnal sampling, sites were evaluated diurnally (Figure 2.1.2) for sampling feasibility based on accessibility, safety, hydrology and habitat quality. Sampling sites ranged from water bodies located in

Eucalyptus plantations to those within protected parks (Figure 2). During diurnal surveys,

sampling for tadpoles and Xenopus spp. was done, while nocturnal survey efforts focused on the active sampling of adults. During rainfall events surveying was done by scanning for individuals crossing roads while driving (road cruising) and by listening for frog choruses.

C

Figure 2.1.2.: Examples of sampling sites: A- Mfungeni pan, Tembe. Daytime surveys took place due to the

danger of wildlife at night; B- Mposa offices, Mposa. Daytime surveys took place due to restricted access at night; C- Kosi Bay Lodge pan, Kosi Bay. Both daytime and night-time sampling took place; D- Bonamanzi porcupine pan, Bonamanzi. Example of a site not hydrologically suitable for sampling.

A B

D

C

B

A

(26)

14

2.2. Field Work

2.2.1. Sampling techniques:

Visual and acoustic sampling

Headlamps were used at night to perform visual encounter surveys to detect frogs in and around pools and surrounding vegetation (Figure 2.2.1a). Acoustic surveys were used to home in on calling frogs. Frogs were captured by hand. Upon collection, frogs were assigned a field number and secured individually in labelled plastic bags with a small volume of water and some damp vegetation. Coordinates, time of capture, notes on habitat, and call activity for all species encountered were recorded at each collection site. Frogs were transported in a cooler box back to a field laboratory. Specimens were collected under permit No. 4092/2016 issued by Ezemvelo Wildlife (Appendix 1).

Dip net and bucket trap sampling

Dip net sampling was used to collect tadpoles (Figure 2.2.1b). This method entails intense back-and-forth sweeping through the water at potential breeding pools. Captured tadpoles were placed in plastic sorting trays at the point of capture and when possible at least five representatives of each species (preliminarily field identification) were collected. To prevent specimen degradation, tadpoles were immediately euthanised using buffered 3 % Ethyl 3-aminobenzoate methanesulfonate (MS222) solution following SOP NW-00492-16-A5 for further identification and assignment of field numbers back at a field laboratory. Baited bucket traps (Du Preez & Van Wyk, 2007) were used to collect Xenopus spp. Traps were baited with chicken livers and checked the following day. Captured frogs were field processed using the methods described above.

2.2.2. Field laboratory procedures:

Processing of frogs and tadpoles

After arrival at a field laboratory (Figure 2.2.1c), frog specimens were transferred to plastic containers with a small volume of water, moist soil for fossorial species, and kept cool. Identification of each specimen was established using a field guide (du Preez & Carruthers 2009) and a stereo microscope when necessary. Each specimen’s date of collection, locality, coordinates, sex, and age class was recorded. The mass and snout-to-vent length (SVL) of each frog was measured using a Vernier calliper. Two representative specimens per species were selected for tissue sampling and vouchering; all remaining frogs were released at their point of capture.

Tissue collection and vouchering

Before euthanasia specimens were assigned field tags with the species name, locality, field number, and associated AACRG collection number tags. Specimens were photographed either before or after euthanasia with their collection number tags. Specimens were euthanised using MS222 solution. Muscle tissue from one hind limb was

(27)

15

dissected out and preserved in a cryogenic vial containing 70% molecular grade ethanol with a split sample in RNAlater. Voucher specimens were fixed in 10% neutral buffered formalin (NBF) overnight. Due to their conservation status and permit conditions, only blood samples were obtained from red-listed species, such as Hemisus guttatus prior to their release. Blood samples were collected using an insulin syringe, drawing from the femoral artery or vein, and preserved in cryogenic vials containing 70% molecular grade ethanol. For tadpoles, a tissue sample from one representative of each species was dissected from the tailfin and placed in a cryogenic vial with 70% molecular grade ethanol. Tadpole vouchers were fixed in 10% NBF.

After arrival at the campus laboratory, formalin-fixed specimens were washed with slow running tap water overnight and accessioned into the SAIAB (South African Institute of Aquatic Biodiversity) amphibian collection. Tissues stored in cryogenic boxes were placed in a -80˚C freezer. All voucher specimens, tissue samples, and photographs are housed in the SAIAB amphibian collection at the North-West University; all data are archived in the SAIAB Specify database.

Figure 2.2.1: Collecting techniques and field processing: A- Visual and acoustic sampling; B- Dip net sampling; C- Field

laboratory.

A

B

A C

(28)

16

2.3. Molecular laboratory work and analysis

2.3.1. Amplification and alignment of sequences

To account for geographical variation, and to increase the likelihood of uncovering cryptic species, I aimed to select specimens from multiple localities to reflect the maximum geographic separation allowed by sampling efforts. In cases where species were only collected from one locality, two representative specimens were selected.

Due to the scarcity of COI records for Anurans the sequencing of a portion of the mitochondrial 16S rRNA gene, as well as the standard mitochondrial COI gene, was attempted for each specimen. The inclusion of the 16S gene fragment allowed for the inclusion of the more comprehensive reference data for 16S on GenBank and BOLD, in addition to available COI data.

DNA was extracted from the tissue samples using the rapid DNA extraction method as detailed in the commercial KAPA Express Extract DNA extraction kit (Kapa Biosystems, Cape Town, South Africa). Polymerase chain reactions (PCR) were used to amplify approximately 650 bp fragments of both genes. Four primer sets were used for the amplification of the COI marker and one set for the amplification of 16S (Table 1). Primer sets for the COI gene were used systematically until samples were successfully amplified or until all four primer sets were exhausted. The PCR reaction mixture contained 12.5µl Thermo Scientific DreamTaq PCR mastermix (2x) (2x DreamTaq buffer, 0.4mM of each dNTP, and 4mM MgCl2), 1.25µl of each primer (10 μM), 1µl of template DNA and 9µl of Milli-Q water to adjust the volume to 25µl. PCR reactions were conducted using a ProFlexTM PCR thermal cycler (Applied Biosystems by Life technologies).

For the COI gene, reactions were amplified under the following conditions: initial denaturation at 94°C for 5 minutes, followed by 36 cycles entailing a 94°C denaturation for 30 seconds, annealing at 50°C for 50 seconds, with an end extension of 72°C for 2 minutes, and a final extension at 72°C for 10 minutes. For 16S amplification the following conditions were used: initial denaturation at 95°C for 90 seconds, followed by 34 cycles entailing a 95°C denaturation for 45 seconds, annealing at 51°C for 45 seconds, with an end extension of 72°C for 90 seconds respectively, and a final extension at 72°C for 5 minutes. Gel electrophoresis was used to visualise the DNA. A 1% agarose gel loaded with samples and Gel Red was illuminated using Ultraviolet (UV) light. A BioRad GelDoc Imaging System (BioRad, UK) was used to view the contents.

Products of PCR yielding positive bands were sent to a commercial sequencing company (Inqaba Biotechnical Industries (Pty) Ltd, Pretoria, South Africa) for purification and sequencing in both directions. After receiving the sequences, they were assembled, and chromatogram based contingencies were generated using Genious Ver. 9.1. (http://www.geneious.com, Kearse et al., 2012). The species identification was confirmed

(29)

17 Table 2.3.1: Primers used in this study

Fragment Primer name Primer sequence 5′–3′ References

COI LCO-1490 5′-GGT CAA CAA ATC ATA AAG ATA TTG G-3′ Folmer et al., 1994

HCO-2198 5′-TAA ACT TCA GGG TGA CCA AAA AAT CA-3′

COI Chmf4 5'-TYT CWA CWA AYC AYA AAG AYA TCG G-3′ Che et al., 2012

Chmr4 5'-ACY TCR GGR TGR CCR AAR AAT CA-3′

COI VF1 5′-TTCTCAACCAACCACAAAGACATTGG-3′ Ivanova et al., 2007

VR1 5′-TAGACTTCTGGGTGGCCAAAGAATCA-3′

COI VF1d 5′-TTCTCAACCAACCACAARGAYATYGG- 3′ Ivanova et al., 2007

VR1d 5′-TAGACTTCTGGGTGGCCRAARAAYCA-3′

16S 16SaR-F 5′-CGCCTGTTTAYCAAAAACAT-3′

Kocher et al., 1989

16SbR‐R 5′-CCGGTYTGAACTCAGATCAYGT-3′

to genus level using BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi), and sequences with widely erroneous results were removed from the analysis.

For inclusion in the analysis, additional sequences of the species sampled in this study, all available species occurring in the study area, and additional species belonging to the families in this study, were downloaded from the BOLD and GenBank databases. Sequences were grouped by family and analysed separately. Detailed acquisition and alignment of sequences for each family are given in Appendix 2. Sequences were aligned and trimmed to the length of the shortest sequence using the bioinformatics software programme MEGA7 (Molecular Evolutionary Genetics Analysis version 7.0 for bigger datasets; Kumar, Stecher, and Tamura, 2016).

2.3.2. Sequence analysis

Multiple lines of evidence were used to confirm or dispute the morphological identification of collected specimens and to screen for cryptic species simultaneously; these included BINs delineated in BOLD, the ABGD algorithm, and the construction of Neighbor-joining (NJ) trees. In cases where evidence of cryptic species was detected, genetic distances, the proximity of samples to type localities, and known geographic distributions of the species in question were evaluated.

To statistically detect barcode gaps and identify distinct clusters of DNA sequences two methods were used: the Barcode Index Number (BIN) system implemented by the RESML algorithm on BOLD and the Automatic Barcode Gap Discovery algorithm (ABGD; Puillandre et al., 2012). The ABGD method was used in addition to the BIN system as the RESML algorithm is more sensitive and subject to false positives (Guarnizo et al., 2015). Using the ABGD method also allowed for the analysis of sequences not contained in the BIN database.

The ABGD web interface (http://wwwabi.snv.jussieu.fr/public/abgd/), was used to apply the ABGD method. The COI alignments were processed in ABGD using the complete sequence data and the Kimura two-parameter (K2P) nucleotide substitution model (Kimura, 1980). The K2P model is the standard model of DNA substitution for DNA barcoding studies and performs as well as more complex models in identifying specimens (Collins et al., 2012). The ABGD method uses adjustable parameters. Parameters were

(30)

18

set to result in a variety of partitions. Partitions were selected to maximise the number of resulting groups that matched described species. Groups were then evaluated for merging or splitting of species.

To visualise genetic divergence between species and show clusters of OTUs within them NJ trees (Saitou & Nei, 1987) were constructed using the K2P distance model with 1000 bootstrap replicates in MEGA 7 (Kumar et al., 2016). Identical COI sequence datasets were used for NJ trees and ABGD analyses. The NJ trees were used for clustering of similar sequences, not for estimating evolutionary relationships.

Where evidence of cryptic species was detected, sequences were grouped into OTUs identified by the methods above and their genetic distances analysed. The K2P pairwise genetic distances (p-distances) between sequences, between groups (interspecific), and the mean intraspecific variation within groups (intraspecific) were computed in MEGA 7 using all characters. Maximum and minimum intraspecific distances were calculated from the pairwise genetic distances.

Specimen data, sequence data, trace files (for both gene fragments), primer data, and photographs of specimens are available on BOLD (http://www.boldsystems.org/; Ratnasingham & Hebert, 2007), under project code LDPJR. Sequences have also been deposited in GenBank. All BOLD Process IDs are listed in table 3.1.2. BOLD Process IDs are linked to GenBank accession numbers.

(31)

19

Chapter 3: Results

(32)

20

3. Results

3.1. Molecular data

3.1.1. Voucher collection and sequence recovery

We surveyed 70 localities, collecting and sampling 432 individual frogs and tadpoles, of which 350 were preserved and accessioned. These samples represent 11 taxonomic families, 20 genera, and 35 nominal species and two unidentified Breviceps specimens (Table 3.1.1), accounting for 69% of the known species occurring within the study area (Du Preez & Carruthers, 2017). Maps showing localities of voucher and analysed specimens are given for each species collected in this study in Appendix 2. From the 350 voucher specimens, 141 specimens were chosen for molecular analysis. Sequencing for the COI gene was attempted for 115 specimens, 89 of which were successful (77% amplification success rate), representing 32 recognised species and two unidentified specimens. Table 3.1.1: Comprehensive list of species sampled and analysed

Family Species Reference

Arthroleptidae Arthroleptis wahlbergii Smith, 1849

Leptopelis mossambicus Poynton, 1985

Leptopelis natalensis (Smith, 1849)

Brevicipitidae Breviceps passmorei Minter, Netherlands & Du Preez, 2017

Breviceps sp. n/a

Bufonidae Sclerophrys garmani (Meek, 1897)

Sclerophrys gutturalis (Power, 1927)

Sclerophrys pusilla (Mertens, 1937)

Schismaderma carens (Smith, 1848)

Hemisotidae Hemisus marmoratus (Peters, 1854)

Hemisus guttatus (Rapp, 1842)

Hyperoliidae Afrixalus aureus Pickersgill, 1984

Afrixalus delicatus Pickersgill, 1984

Afrixalus fornasini (Bianconi, 1849)

Hyperolius marmoratus Rapp, 1842

Hyperolius pusillus (Cope, 1862)

Hyperolius tuberilinguis Smith, 1849

Hyperolius argus Peters, 1854

Kassina senegalensis (Dumeril & Bibron, 1841)

Phlyctimantis maculatus (Dumeril, 1853)

Microhylidae Phrynomantis bifasciatus (Smith, 1847)

Phrynobatrachidae Phrynobatrachus natalensis (Smith, 1849)

Phrynobatrachus mababiensis FitzSimons, 1932

Pipidae Xenopus laevis (Daudin, 1802)

Xenopus muelleri (Peters, 1844)

Ptychadenidae Ptychadena anchietae (Bocage, 1867)

Ptychadena mossambica (Peters, 1854)

Ptychadena oxyrhynchus (Smith, 1849)

Ptychadena nilotica (Seetzen, 1855)

Ptychadena porosissima (Steindachner, 1867)

Hildebrandtia ornata (Peters, 1878)

Pyxicephalidae Pyxicephalus edulis Peters, 1854

Tomopterna krugerensis Passmore & Carruthers, 1975

Cacosternum boettgeri (Boulenger, 1882)

Amietia delalandii (Dumeril & Bibron, 1841)

(33)

21

Sequencing of the 16S gene was attempted for all 141 specimens, of which one failed (99.3% amplification success rate), representing 35 recognised species, and two unidentified specimens.

Analyses were performed on one to six individuals per species (mean = 3.4). An average of four 16S rDNA and two COI marker barcode sequences were obtained per species. Nine species were represented by the COI marker of one specimen and four species by the 16S marker of one specimen. The COI marker for three species could not be successfully amplified. These results total 139 analysed specimens (Table 3.1.2), with one or both gene fragments being successfully amplified for 35 species and two unidentified

Breviceps specimens.

Despite using four sets of primers, COI was not successfully amplified for three species:

Leptopelis natalensis, Sclerophrys garmani and Ptychadena porosissima. It should be

noted that all four primer sets were not utilised for P. porosissima, as the COI fragment for a single tadpole specimen, thought to be a P. porosissima had been sequenced, but was later identified as P. anchietae. Due to logistical constraints, only two of the primer sets were utilised for L. natalensis. In total, the VF1-VR1 primer set amplified 17 species, HCO-LCO amplified 11 species, VF1d-VR1d amplified two species and Chmf4-Chmr4 amplified three species. The primer set 16SaR-16SbR was successful in amplifying the 16S marker for all species.

3.1.2. Reference library contribution

In total this study contributed 224 sequences to BOLD and Genbank, comprising 86 COI sequences and 138 16S sequences. Six of the COI sequences contained stop codons and were thus not barcode compliant, all others received BIN numbers. Two 16S sequences lacked successful traces. This study contributed nine new BIN entries to BOLD and four new species entries to Genbank. Examples of a specimen record and BIN record is given in Appendix 2.

3.1.3. BIN delineation and BIN discordance

The barcode compliant COI sequences were delineated into 34 BINs by the RESLM algorithm. Four species were delineated into more than one BIN: Hyperolius marmoratus;

Kassina senegalensis; Breviceps sp.; and Tomopterna krugerensis. One BIN containing

only sequences from this study was deemed as discordant by the BIN discordance analysis. This BIN contained sequences identified initially as one Hyperolius pusillus, one

H. poweri and one Hyperolius sp. After re-examination, the H. poweri and Hyperolius sp.

specimens (which were a juvenile and tadpole respectively) were deemed as misidentifications and reassigned to H. pusillus.

Amphibians of Northern KwaZulu-Natal : a phylogenetic study