Tesis Kroll

Evolutionary patterns of endemic gastropods in ancient Lake
Titicaca: the genus Heleobia (Rissooidea: Cochliopidae)
Oliver Kroll
Rostock, August 2008

Evolutionary patterns of endemic gastropods in ancient Lake
Titicaca: the genus Heleobia (Rissooidea: Cochliopidae)
Diploma Thesis
Submitted 4. August 2008 by
Oliver Kroll
University of Rostock
Institute of Zoology
Supervisors:
Prof. Dr. Wilke, University of Gießen
Prof. Dr. Richter, University of Rostock
Evolutionary patterns of endemic gastropods in ancient Lake Titicaca: the genus
Heleobia (Rissooidea: Cochliopidae)
Table of Contents
Danksagung
Summary
1 Introduction ................................................................................................................ 1
1.1 “Ancient lakes” in evolutionary- and biodiversity science ......................... 1
1.2 Lake Titicaca ............................................................................................... 2
1.2.1. Titicaca research .......................................................................... 2
1.2.2 General information (Origin, morphology and hydrology) ......... 2
1.2.3 Biodiversity .................................................................................. 5
1.2.4 Molluscs ....................................................................................... 6
1.3 The genus Heleobia (Stimpson, 1865) ...................................................... 6
1.4 Heleobia in Lake Titicaca .......................................................................... 8
1.5 Aims of this work ....................................................................................... 10
2 Materials and Methods .............................................................................................. 12
2.1 Sampling design and specimens studied .................................................... 12
2.2 Molecular analysis ...................................................................................... 14
2.2.1 DNA isolation .............................................................................. 14
2.2.2 DNA amplification ...................................................................... 15
2.2.3 DNA purification ......................................................................... 16
2.2.4 DNA sequencing .......................................................................... 16
2.2.5 Sequence saturation ..................................................................... 17
2.3 Phylogenetic analysis of molecular data .................................................... 17
2.3.1 Tree based methods ..................................................................... 17
2.3.2 Network analysis ......................................................................... 19
2.3.3 Molecular clock analysis ............................................................. 20
2.4 Morphology ................................................................................................ 21
2.4.1 Shell parameters .......................................................................... 21
2.4.2 Radula .......................................................................................... 25
3 Results ....................................................................................................................... 26
3.1 Molecular analysis ..................................................................................... 26
3.1.1 DNA isolation and amplification ................................................ 26
3.1.2 DNA purification ......................................................................... 26
3.1.3 DNA sequencing ......................................................................... 26
3.2 Phylogenetic analysis of molecular data .................................................... 27
3.2.1 Tree based methods ..................................................................... 27
3.2.2 Network analysis ......................................................................... 28
3.2.3 Molecular clock analysis ............................................................. 31
3.3 Morphological analyses ............................................................................. 32
3.3.1 Shell parameters .......................................................................... 32
3.3.2 Radula ......................................................................................... 35
4 Discussion ................................................................................................................ 37
4.1 Phylogenetical position to putative Heleobia spp. .................................... 37
4.2 Monophyly of Titicaca Heleobia spp. ....................................................... 39
4.3 Morphological characters in the view of molecular data .......................... 41
4.4 Biogeography ............................................................................................ 44
5 Literature .................................................................................................................. 48
Acknowledgements
Acknowledgements
I want to express my gratitude to:
- Prof. Dr. Wilke (Universität Gießen/ Germany) for supervising this work, especially for
supporting my own ideas
- Prof. Dr. Richter (Universität Rostock/ Germany) for supervising this extern Diploma
tesis, him and his workgroup for helpful comments and critical discussions
- Prof. Dr. Edmundo Moreno (Universidad Puno/ Peru) for his hospitality and the intensive
support of the fieldwork in Peru, without him this work would not have been possible
- Prof. Dr. Julio Pinto (Universidad Mayor de San Andrés, La Paz/ Bolivia) for
coordinating the fieldwork in Bolivia and the kindly permission to use extensive the
institute laboratory
- Blgo. David Aranibar (Reserva Nacional del Lago Titicaca, Puno/ Peru) and all people of
the RNT for hospitality and the intensive support of the fieldwork
- Dr. Christian Albrecht (Universität Gießen/ Germany) for his pragmatic guidance and
constructive comments from the beginning
- Roland Schultheiß (Universität Gießen/ Germany) for helping with the phylogenetic
software and for uncomfortable critics
- Dr. Carlos Molina (Universidad Mayor de San Andrés, La Paz/ Bolivia) for hospitality
- Sylvia Nachtigall (Universität Gießen/ Germany) for introducing me into the laboratory
work
- the workgroup of Prof. Dr. Wilke for answering many quenstions
- Dr. Mateo López (Universität Gießen/ Germany) for his enormous help in the final days
- all biologists and fishermen in Bolivia and Peru that I could not thank here nominel
and especially my mother Carmen Kroll for supporting me during my whole university
time and particular during this tesis.
Summary
Summary
There are more than a dozen so called ancient lakes in the world. With an age of
100.000 years or up to several million years they are much older than most common lakes.
Ancient lake histories are characterized by periods of long stability in abiotic factors,
interrupted by phases of ecological change, which presumably resulted in a high
biodiversity with numerous endemic elements. Many ancient lake taxa, for example
gastropods, are known for their high degree of endemism and a remarkable morphological
diversity. The Lake Titicaca is the only recent South American ancient lake. During its life
history the size, the deep and the climatic situation varied drastic, all caused mainly by
glacial-interglacial influences. Nineteen species of gastropods are described for Lake
Titicaca, from which 16 are considered to be endemic. They belong to the pulmonate
genera Biomphalaria and Gundlachia and to the here investigated caenogastropod genus
Heleobia. Taking into account the actual systematic review, the representatives of
Heleobia occur strongly disjunctive and can be found in the southern part of South
America, as well as in the old world. The type species of the genus Heleobia is to be found
in Lake Titicaca. Important systematic criteria for Heleobia are mainly anatomical
characters like the presence of typical shaped glands at the penis. The remarkable diversity
of shell morphologic characters within the 14 recognized species from Lake Titicaca, some
of them quite uncommon in gastropods (e.g. uncoiled shell, spiraliform operculum), makes
interpretations of phylogenetic relationships within the lake very difficult. This also applies
to the relationships to species outside the lake. The aim of this work was to study the
species of Heleobia spp. present on Lake Titicaca, using a combined approach of
molecular genetics and morphologic methods for the first time. The material studied was
collected during a field trip to Lake Titicaca and its neighbouring lakes Umayo and Arapa,
from April to July 2007. Other specimens from one sample site in South Bolivia and one
site in North Chile completed the collection. The mitochondrial gene cytochrome c oxidase
I (COI) was sequenced for 45 individuals of Heleobia. Compatible topologies were found
using MP, ME, ML, NJ and BI analyses and resulted in important points in equivalent
topologies in all these methods. The results of the analysis showed that: all specimens from
Lake Titicaca cluster in one clade, within the Titicaca clade two haplotypes from Chile
occur, European Heleobia spp. are the next related clade to the Titicaca clade, the Bolivian
haplotype is next related to the Eurasian Heleobia spp., North-American Heleobops spp.
Summary
are next related to the Bolivian haplotype, one Italian Heleobia sp. haplotype clusters not
with the other European Heleobia spp. but within the North-American Heleobops spp.. H.
parchappi and/ or H. australis from eastern South-America are closely related to the
Heleobops spp., and H. davisi from eastern South-America occurs within or next to the
remaining out-groups. Parsimony network analysis of Lake Titicaca Heleobia did not
result in distinct clades for distinct species. In opposite to that haplotypes of different
species can cluster. The specimens from the lakes Umayo and Arapa cluster within the
network, as well as the specimens from Chile. The time of divergence between all
haplotypes within the Titicaca clade was estimated as 0.69 ± 0.27 Ma. Incomplete lineage
sorting (ancestral polymorphism) in the mitochondrial sequences might be responsible for
obscuring the real phylogenetic relationships within the analyses. Therefore future studies
need be performed with nuclear markers (microsatellits and AFLP). The separation
between the Titicaca clade and the Eurasian Heleobia spp. clade was calculated by
molecular clock analyses as 2.13 ± 0.59 Ma. The molecular results on Titicaca mollusks,
performed here for the first time, show that the Heleobia spp. of Lake Titicaca are
phylogenetically closely related to the other Heleobia spp. worldwide. The biogeographical
aspects implicated in these results are interesting, considering the disjunctive distribution
of Heleobia, which is known from Eurasia as well as from America. Surprisingly, species
from Eurasia obtain the sister taxon position within the gene trees regardless of the
analyzing methods. Whether monophyly of the Titicaca Heleobia is present, depends on
the taxonomical status of the Chilean specimens, which cannot be answered with the data
at hand, because of uncertainty in species determination. Furthermore, the discussed
taxonomical status of the Chilean species also implies consequences for the present status
of endemism for the Titicaca Heleobia spp., which would be rejected in the case that the
Chilean species actually belong to one of the Titicaca species. Distribution pattern with
evident connection to the paleohistory have been found in other taxa for the Altiplano
region. Future studies on Heleobia should include specimens from all Altiplano relict
waters and adjoining waters of the Central Andes. Statistical principal component analyses
(PCA) were performed with morphological parameters. Species included in all analyses are
not clearly separable, neither by shell character nor by the marker gene COI.
For the moment it is plausible to say that the Heleobia spp. from Lake Titicaca
might be a recently radiated clade or just a single species with a very high morphological
plasticity.
1 Introduction
1 Introduction
1.1 Ancient lakes in evolutionary- and biodiversity science

There are more than a dozen so called ancient lakes (Brooks 1950) in the world,
each of them much older than most common lakes. According to Gorthner (1994), the term
ancient lake should be changed into long lived lakes, which reflects that their basins have
been constantly filled with water for at least 100,000 years (up to 20 million for Lake
Baikal) (Gorthner 1994, Martens 1997, Albrecht et al. 2006). Their origin is set mostly by
tectonic processes, thus these basins are of considerable profoundness. Ancient lake
geological histories are characterized by periods of long stability in abiotic factors,
interrupted by phases of ecological change, which presumably resulted in a high
biodiversity with numerous endemic elements (Rossiter & Kawanabe 2000, Glaubrecht &
Köhler 2004, Wilson et al. 2004, Albrecht et al. 2006, Seehausen 2006). For several
scientific disciplines like climatology, palaeoecology and evolutionary biology these
waters represent important study sites (Baker 2005, Seehausen 2006). The great age of
these systems allows observations that can help to explain processes that occurred a long
time ago. At the same time, the mechanisms of these processes and their consequences can
be studied in recent examples (Martens et al. 1994). All these reasons make “ancient lakes”
ideal in situ laboratories for studies related with biodiversity and speciation processes
(Mourguiart 2000, Glaubrecht & Köhler 2004).
Immigration and natural extinction, as well as the dynamic process of speciation,
are some of the causal factors that determine the degree of biodiversity. Many ancient lake
taxa, for example gastropods, are known for their high degree of endemism and a
remarkable morphological diversity. However, the gastropod fauna of “ancient lakes”
show various convergent developments, which are still far from being understood (Martens
1997, Albrecht et al. 2006). Some of these lakes were set into focus of research activities
and are consequently well investigated. The Siberian Lake Baikal and the African rift lakes
Malawi and Tanganyika represent waters where intensive research has been done
continuously for al long time (Rossiter & Kawanabe 2000, Schön & Martens 2004). Other
ancient lakes have long been neglected by evolutionary biologists, and the general
knowledge of evolutionary processes in these lakes must be considered as low. The South
American Lake Titicaca constitutes a prima example for such rarely investigated lake
systems (Dejoux 1992a, Mourguiart 2000).
1
1 Introduction
1.2 Lake Titicaca
1.2.1 Titicaca research

Expeditions to Lake Titicaca conducted during the 19th and 20th century gave a first
overview of the biological situation of the lake (Dejoux 1992a, Mourguiart 2000,
Constantini et al. 2004). The most important undertakings and works with naturalistic
backgrounds were done by D’Orbigny (from 1835 to 1847), Castelnau and Weddel (from
1843 to 1847), Agassiz and Garman (year 1876), Crequi-Montfort and Senenchal de la
Grange (year 1904), and Pilsbry (years 1896 and 1924) (for a review see Gilson 1939,
Dejoux 1994). The first large scale ecological research was realized by the British Percy
Sladen Trust Expedition in 1937, which brought the richest results in taxonomical
collection of Titicaca species (Gilson 1939). Since the 1970s, the laboratory of the Instituto
del Mar del Peru (IMARPE) exists at the Peruvian shore line. The Universidad Mayor de
San Andrés of La Paz is responsible for research projects in the Bolivian side of the lake.
In the 1980s the French guided cooperation project ORSTOM investigated several aspects
about the limnology of Lake Titicaca. The results were published as an extensive
compendium (Dejoux 1992a,b). Due to their economical importance, fishes of Lake
Titicaca are the best represented group of organisms in the literature (Lauzanne 1992,
Sterba 1996, Northcote 2000, Lüssen et al. 2003). The last ten years were dominated by
studies with geologic focus like palaeo-limnological and palaeo-climatical processes, the
largest represented by the ICDP Titicaca Drilling Project (Argollo & Mourguiart 2000,
Rowe 2003, Baker 2005, Rigsby et al. 2005, Fritz et al. 2007, Theissen et al. 2008).
1.2.2 General information (Origin, morphology and hydrology)

Lake Titicaca is one of the largest lakes worldwide. During its life history the size,
the deep and the climatic situation varied drastic (i.e. water level change for about 140 m),
all caused mainly by glacial-interglacial influences (Newell 1949, Romero 1973,
Wirrmann 1992, Baker 2005). The Lake Titicaca is the only recent South American
„ancient lake“. Its origin was set by glacial and tectonic procedures during the Pleistocene,
about 2-3 Ma (Lavenú 1992). The lake is situated in the tropical zone of the central Andes
of Peru and Bolivia at an altitude of 3809 m in the northern part of the Altiplano, which
gives the name to a plateau between the Eastern and the Western Cordillera (Newell 1949,
Allmendinger 1997). The Altiplano measures about 2000 km from north to south and 400
km from east to west at its maximum width, and is situated at an altitude between 3700-
2
1 Introduction
4600 m. Its southernmost parts reach the north of Chile and Argentina (Figure 1).
Palaeolimnic processes of the Altiplano wetlands are thought to be of major importance for
influencing speciation patterns within the faunal components of the Lake Titicaca and their
recent biogeography (Gorthner 1994, Schäfer 1997, Lüssen et al. 2003). Therefore, they
will be outlined here in terms of their developmental history. Geological data suggest that
the separation of South America from Africa took place about 118 Ma, and the uplifting
process of the Andes during the last 90 Ma. The creation of the central-andine-high-
mountain-plateau started about 60 Ma and its isolation began about 30 million years later.
During this period it was uplifted about 3000 m.
During cold tempered periods the water level took its maximum, in such a way that
a big part of the Altiplano was dominated by a wetland system much larger than the actual
Lake Titicaca. The oldest of such palaeolake systems is known as Palaeolake Mataro, with
an age of 1.8–1.0 Ma, being the largest known palaeolake system world wide (Figure 1).
Interglacial periods of warming resulted in lowering of the lake surface, leaving relict
waters with an own, more or less isolated development.
There have been five well recognizable paleolimnic periods during the last 1.8 Ma
(Baker 2005). In the southern part of the Altiplano the salty Lake Popoo and the Salares
Uyuni and Coipasa represent the largest examples of such relict waters (Figure 1). Directly
at the periphery of Lake Titicaca two much smaller lakes are located, also representing
relict waters. Both lakes Umayo and Arapa are permanently or seasonally connected
though rivers with their larger neighbour. More than two dozen rivers are to be found
entering Lake Titicaca, but the only drainage is Rio Desaguadero, that leads 200 km
southward into the Lake Popoo. There, temporary overflows to the Salares Uyuni and
Coipasa can occur (Gilson 1939, Dejoux 1992b, Baker 2005). Titicaca’s morphology is
bipartite with a bigger northern basin (Lago Chucuito) and the much smaller one in the
south (Lago Huinaimarca) (Table 1).
3
1 Introduction
Figure 1. Satellite image of the Altiplano Region (delimitated with the yellow line),
indicating with points two sampling sites located at the Laguna Blanca in Bolivia and at
the Rio Loa in Chile. The image was downloaded from the webpage of the NASA
(www.nasa.org).
4
1 Introduction
Table 1. Summary of physical and limnological data (from Dejoux 1994).

Whole lake Northern basin Southern basin
Geographical position 15°13’-16°35’S, 68°33’-70°02’W
Altitude (masl) 3809
Length (km) 178 151 62
Width (km) 58 53 23
Surface area (km2) 8448 7131 1428
Island area (km2) 111 50 61
Shoreline (km) 915 610 305
Max. depth (m) 284 284 41
Mean depth (m) 105 135 10
Volume (109 m3) 895.86 883.5 12.36
Stratification polymictic monomictic
Mean temp. °C 11-14 11.5-15 9.5-17
Extreme temp. °C 8.5-18.5 10.9-17 8.5-18.5
Oxygen (mg/l) 7
pH 8.5
EC (K20, μS/cm) 1500
Transparency (m) 4.5-16 9-10.2*
* Personal data suggest lower values (0.5-6 m).
1.2.3 Biodiversity
About 500 species are described for Lake Titicaca, with some higher taxa
characterized by a high degree of endemic forms. Namely the Mollusca, Amphipoda,
Ostracoda, Pisces and Porifera led to extensive discussions of evolutionary convergence
within “ancient lake” fauna (Dejoux 1994, Martens et al. 1994, Schön & Martens 2004).
The quantity of Titicaca species is relatively low compared to those of other ancient lakes
(see Martens 1997). This is supposed to be caused by the fact that the initial faunal
components were of tropical origin and have been limited during the uplifting processes of
the Altiplano. Only a small quantity of species was able to tolerate and adapt to the
circumstances of a high mountainous life (de Lattin 1967).
5
1 Introduction
1.2.4 Molluscs
Nineteen species of gastropods are described for Lake Titicaca, from which 16 are
considered to be endemic (Dejoux 1992a, Hershler & Thompson 1992). They belong to the
caenogastropod genus Heleobia (Stimpson, 1865) and to the freshwater pulmonate genera
Biomphalaria (Watson, 1954) and Gundlachia (Pfeiffer, 1849). Four species of freshwater
mussels (Sphaeriidae) with the endemic species Musculium titicacaense (Pilsbry, 1924) are
known. The remaining three species do not settle more than a restricted area of the Andes
(Korniushin 2004).
1.3 The genus Heleobia (Stimpson, 1865)

Phylum Mollusca Cuvier, 1795
Class Gastropoda Cuvier, 1797
Clade Caenogastropoda Cox, 1960
Clade Littorinimorpha Golikov & Starobugatov, 1975
Superfamily Rissooidea Gray, 1847
Family Cochliopidae Tryon, 1866
Type locality
D’Orbigny described in “Voyage dans l’Amerique meridionale 1826-35” various
species of gastropods collected during this expedition. One of them, Paludestrina
culminea, was found in Lake Titicaca in the Bay of Achacachi, and later became the type
species of the genus Heleobia (D’Orbigny 1840).
Systematics
Significant and partly conspicuous differences in conchiological characters led
Haas (1955, 1957) to create distinct genera for these Titicaca gastropods. The similarity in
anatomy brought Hubendick (1957) to the postulation of their close systematic
relationship. Later systematic revisions finally merged all into one single genus. Taylor
(1966) placed all microgastropods from Lake Titicaca in the New World genus Littoridina
Eydoux and Souleyet, 1852. Davis et al. (1982), however, separated the now exclusively
neotropical Littoridina and the Heleobia, and furthermore synonymised the European
Semisalsa Radoman, 1974 with Heleobia (Kabat & Hershler 1993). Important systematic
criteria for Heleobia are mainly anatomical characters like the presence of typical shaped
glands at the penis (Figure 2). Heleobia is the more diverse genus within the Cochliopidae
6
1 Introduction
(31 genera in total), represented by 89 species (Hershler & Thompson 1992). This family
has a sister taxon relationship with the Amnicolidae (Tryon, 1863) reconstructed by
anatomical and molecular genetic data (Wilke et al. 2001).
Figure 2. Schematic draw of the penis of Heleobia andicola showing the characteristic
glands present on the penis (Hubendick 1955).
Distribution
Taking into account the actual systematic review (Hershler & Thompson 1992), the
representatives of Heleobia occur strongly disjunctive and can be found in the southern
part of South America, as well as in the old world, where species occur from the North Sea
region down to the Mediterranean region of Africa and eastward to the Caspian Sea region
(Figure 3). They occur in both limnic and marginally brackish habitats. The development is
holobenthic or bentho-pelagic with free swimming veliger larvae within coastal species.
The lack of good fossil data makes it difficult to estimate the palaeontological age of the
Cochliopidae. However, it is hypothesized that the origin is set prior or at the time of the
Pangaea break-up during the early Mesozoic, which is very likely given the neotropical as
well as holarctic distribution of Heleobia (Hershler & Thompson 1992, Wilke et al. 2001).
7
1 Introduction
Figure 3. Distribution of Heleobia spp. (red circles) with type locality Lake Titicaca (red
dot) (distribution map from Hershler & Thomson 1992).
1.4 Heleobia in Lake Titicaca

The cochliopid micrograstropods account for the majority of mollusc species
known from the lake are all endemic to the Titicaca basin, and with one exception,
endemic to the lake itself (H. andicola culminea also occurs in the two neighbor lakes
Umayo and Arapa) (Haas 1955, own observations). Studies on these molluscs started at the
early 19th century, but have been continued only sporadically (D’Orbigny 1835, Bavay
1904, Pilsbry 1924). The richest collection of new species was made by the British Percy
Sladen Trust Expedition in 1937. The taxonomic work by Haas (1955) and the anatomic
complement by Hubendick (1955) on this material are still the most comprehensive works,
including many primary species description. Some single descriptions of new species and
sub-species followed (Haas 1957, Blume 1958). In the 1980s the limnological ORSTOM
Project collected more than 20,000 individuals of Heleobia. Having such an extensive
material on hand, Dejoux (1992a,b) was in the position to note very interesting
morphological details of taxonomical relevance (some unpublished), but without
continuing to force systematic consequences.
The remarkable diversity of morphologic characters within the 14 recognized
species (Haas 1955, Dejoux 1992a,b), some of them quite uncommon in gastropods, makes
interpretations of phylogenetic relationships within the lake very difficult. This also applies
to the relationships to species outside the lake (Figure 4).
8
1 Introduction
Figure 4. Images and drawings of the 14 Heleobia spp. inhabiting Lake Titicaca. The
drawings represent the four described species that have not been found in this study
(drawings from Haas 1955).
This problem becomes more severe due to a high degree of intraspecific

conchiological plasticity. It was Bavay (1904) who recognized first that the type species H.
andicola culminea (former Paludestrina culminea) and H. andicola andicola (former
Paludina andicola) are inseparably connected by means of intermediate forms. For the
same reason, Haas (1955) noted that H. andicola neveui (former Pyrgula neveui) is the
continuation of this chain of forms. Consequently, he established one single species H.
andicola, separated in three morphologically very distinct sub-species, connected by
continuous intermediate forms (see Figure 5).
9
1 Introduction
Figure 5. Set of seven recognizable intermediate morphotypes of Heleobia andicola

representing the Heleobia-andicola-Komplex. At the left H. andicola culminea, in the
middle H. andicola andicola and at the right H. andicola neveui. Drawings taken from
Hass (1955).
Dejoux (1992b) described massive problems of precise species identification for

most Heleobia sp. in Lake Titicaca. The full range of intermediate forms between various
species made decisions, whether an individual belongs to the one or to the other species,
quite subjective and “a matter of chance” (Dejoux 1992b).
1.5 Aims of this work

Rissooidean gastropods are promising candidates for studying processes in
evolution. It is an old group (Paleozoic), very diverse (more than 1,000 recent species) and
has distributional attributes that are interesting for global biogeography, as well as for local
investigations on speciation (Hershler & Ponder 1998). They are known as a group with
various examples of homoplasy, showing similar shell characters in distinct species (i.e.
cryptic speciation; Wilke et al. 2002). On the other hand, disparity, expressed as high
plasticity in shell morphology, is commonly present within the same genus and species (i.e.
Potamopyrgus antipodarum; Glöer 2002).
Up to this moment all studies about Titicaca gastropods are based exclusively on
shell morphology and anatomy. Former workers have recognized the large amount of
10
1 Introduction
problems in interpreting the morphologic and anatomical characters of the Titicaca

Heleobia spp. as appropriate for certain species determination.
Given this background the aim of this work was to study the species of Heleobia
spp. present on Lake Titicaca, using a combined approach of molecular genetics and
morphologic methods for the first time. By means of this combined approach the following
questions were addressed:
a) Which phylogenetic relationships exist between the Heleobia spp. present on Lake
Titicaca and the putative close related taxa outside the lake?
b) Do the Heleobia spp. of Lake Titicaca represent a monophyletic group?
c) Do the molecular analysis results support the current species classifications, all of them
based only on morphological characters?
11
2 Materials and methods
2 Materials and Methods
2.1 Sampling design and specimens studied

The material studied was collected during a field trip to Lake Titicaca and its
neighbouring lakes Umayo and Arapa, in November 2007. Other specimens from one
sample site in South Bolivia and one site in North Chile completed the collection (material
provided by U. Bößneck, Erfurt). Samples close to the shoreline were taken by hand and
sieve, whereas depths up to 3-4 m could be reached by snorkelling. For sampling the
deeper benthos (up to 38 m depth) transects were established, along which a triangular
dredge (25 cm) was launched from a boat to obtain the material from the bottom of the
lake. A total of 29 sampling sites were investigated (Figure 6).
Figure 6. Map of the study area indicating the sampling sites.
Particular attention was paid to cover the lake’s main geographical regions,
including most of the islands, different substrates and type localities of particular species
12
(Figure 7). All collecting sites were described in terms of their ecological settings, like
bathymetric range of occurrence and type of substrate.
All collected material was fixed and preserved in plastic containers (Sarstedt 70 ml)
filled with 60-80% ethanol, and was stored in freezers at -20°C after arriving at the
University of Gießen. To identify the specimens collected, original descriptions and keys
from D’Orbigny (1835-1847), Pilsbry (1896 & 1924), Bavay (1904), Haas (1955, 1957),
Blume (1958) and the taxonomical notes of Dejoux (1992) were used.
Figure 7. Four examples of the main habitat types to be found on the shores of Lake
Titicaca: a) marsh of Totora plants, b) sandy bottoms, c) bottoms dominated by submerse
macrophytes, d) rocky shores.
13
Ten of the 14 Heleobia spp. described for Lake Titicaca were found and could be fully
identified:
Heleobia andicola (D’Orbigny, 1840)
Heleobia aperta (Haas, 1955)
Heleobia berryi (Pilsbry, 1924)
Heleobia forsteri (Blume, 1958)
Heleobia magna (Haas, 1955)
Heleobia mirum (Haas, 1957)
Heleobia otorni (Pilsbry, 1924)
Heleobia parva (Haas, 1955)
Heleobia stiphra (Haas, 1955)
Heleobia umbiculata (Haas, 1955)
Specimens with ambiguous species determination were labeled as Heleobia sp.

Individuals close to species descriptions of Heleobia lacustris (Haas, 1955) and Heleobia
profunda (Haas, 1955) were present in the samples, but they were labeled as Heleobia sp.,
because they exhibited particular morphological characters (e.g. presence or absence of
umbiculus) only described for other species, or because of lacking a crucial autapomorphic
character (see discussion). The specimen from Bolivia could not be determined to species
level. The determination of the two specimens from Chile as H. cf loaensis is tentatively.
2.2 Molecular analysis

The mitochondrial gene cytochrome c oxidase I (COI) was sequenced for 45
individuals of Heleobia. This mitochondrial, fast evolving protein coding COI has been
extensively used in reconstructing phylogenetic relationship in Mollusca and other animal
groups (Folmer et al. 1994, Davis et al. 1998, Wilke et al. 2001, Köhler 2007, Huang et al
2008). Its variability provokes a significant phylogenetical signal from population up to
family level (Remigio et al. 2001, Wilke et al. 2001).
2.2.1 DNA isolation

DNA from ethanol preserved snails was extracted using the protocol of Wilke et al.
(2006). Because of the small sample size, the complete gastropod was taken only after
shooting several digital images. First, all alcohol-preserved specimens were soaked for 8
min in ice-cold exchange buffer to remove any residual alcohol. After that, tissue were
14
transfered into a mix of 200 μl Turner lysis buffer and 3 μl of Proteinase K (20 μg/μl). For
digestion, the soft parts were set an overnight incubation in a water bath of 58 °C. The next
day, 35 μl of 5% CTAB/NaCl and 35 μl of 5 M NaCl were added and afterwards extracted
with 270 μl chloroform. After shaking the solution 2-3 min by hand, it was centrifuged for
5 min at 9000 rpm to separate the phases. The aqueous phase was transferred into a new
tube and 270 μl CTAB precipitation buffer was added and mixed by hand. The
precipitation finished after 45 min at room temperature. After the centrifugation (10 min at
12000 rpm), the supernatant was disposed and the pellet was dissolved in 100 μl of
NaCl/TE and 1 μl RNase (10 mg/ml). Then the incubation was set for 5-10 min (65 °C).
An overnight precipitation (-20 °C) was done using 250 μl ice-cold ethanol (100 %) for
purification. The supernatant was disposed after its centrifugation (15 min at 12000 rpm).
For washing the pellet 300 μl of 70% ice-cold ethanol were used and the sample
centrifuged afterwards for 5 min (12000 rpm). This procedure was passed two times. After
the supernatant was disposed, the pellet was dried at the air for 5−10 min, and was
redissolved in 20 μl ddH2O. The quality and quantity of the fragments were checked by
electrophoresis in a 1% agarose gel in TBE buffer comparing with a λ DNA marker. For
visualizing the DNA, ethidium bromide by ultraviolet light of a Transilluminator TI1
(Whatman Biometra, Göttingen, Germany) was used. The isolated DNA was stored
afterwards in freezers at -20 or -80 °C.
2.2.2 DNA amplification

The fragments of the mitochondrial COI gene were amplified using the polymerase
chain reaction (PCR). The frozen DNA isolation samples were mixed by vortexing. Then
0.5-1.5 μl of DNA solution, 1 μl 10 x ThermoPol reaction buffer, 0.7 μl deoxyribose
nucleotide triphosphates mix (dGTP, dATP, dTTP, dCTP, each 2.5 mM), 0.7 μl forward
primer (10 μM), 0.7 μl reverse primer (10 μM), 6.2 μl ddH2O, 0.1 μl
tetramethylammonium chloride (TMAC, 0.5 M), 0.6 μl bovine serum albumin (BSA, 10
mg/ml) and 0.2 μl Taq Polymerase (all New England Biolabs, Inc, MA, USA) were
pipetted into one tube in order to obtain a PCR mastermix. The tubes were covered with
paraffin oil (Carl Roth GmbH, Germany) during the PCR process to avoid evaporation.
The primers used for DNA amplification are listed in the table 2.
The PCR was proceeded with the Primus 96 advanced thermocycler (PEQOLAB
Biotechnologie GmbH, Erlangen, Germany). The quality and quantity of the PCR product
were checked on a 1 % agarose gel in TBE buffer, using the PhiX 174/Hae III Digest
15
marker (New England Biolabs, Inc, MA, USA). The visualization of DNA was realized
with ethidium bromide and ultraviolet light (Transilluminator TI1, Whatman Biometra,
Göttingen, Germany).
Table 2. List of primers used to amplificate the DNA extracted from Heleobia spp. from
Lake Titicaca and other localities.
Primer Direction Sequence Citation
COF14 F 5’-GGTCAACAAATCA Folmer et al. (1994)
TAAAGATATTGG-3’
COR722B R 5’-TAAACTTCAGGGTG Wilke & Davis (2000)
ACCAAAAAATYA-3’
2.2.3 DNA purification

The purification was carried out using ExoSAP-IT (USB Corporation, Cleveland,
OH, USA). ExoSAP-IT consists of Exonuclease I and Shrimp Alkaline Phosphatase of
hydrolytic character and a buffer and removes the residual dNTPs and primers from PCR
products (USB Corporation). Therefore 1.2 μl of the PCR product was mixed with 0.6 μl
of ExoSAP-IT. The incubation time took 15 min at 37 °C and 15 min at 80 °C (Thermo
Controller PTC-100, MJ Research, Inc., Watertown, MA, USA).
2.2.4 DNA sequencing

DNA sequencing was carried out using the ‘Sanger didesoxy sequencing’ method,
which results in sequence fragments of different length that can be separated using a gel
electrophoresis. All fragments were sequenced in both directions using the Thermo
Sequenase Fluroescent Labeled Primer Cycle Sequencing kit (Amersham Pharmacia
Biotech). Primers with the same sequence as for PCR reaction were used, forward and
reverse primers differently labeled with two fluorescent dyes at 700 nm and 800 nm.
Sequences were determined using the LI-COR (Lincoln, NE, USA) DNA sequencer Long
ReadIR 4200. After performing the cycle sequencing reactions with a Primus 96 advanced
thermocycler (PEQOLAB Biotechnologie Gmbh, Erlangen, Germany) the samples were
mixed with 1.4 μl loading dye and denaturized (3 min, 95°C). The electrophoresis was
performed on polyacrylamide gels using borofloat plates in TBE-buffer. Sequence analysis
was carried out using e-Seq DNA Sequencing and Analysis software Version 2.0 (LI-COR,
16
Lincoln, NE, USA). All automatic readings were checked by eye. Missing data were
signed as N. Forward and reverse sequences were aligned in Bioedit 7.0.5.2 (Hall 1999).
For phylogenetic analyses resulting consensus sequences were used only.
2.2.5 Sequence saturation

In order to test whether the COI data set shows a significant level of saturation, the
test of sequence saturation, as implemented in the software package DAMBE version 5.0.7
(Xia & Xie 2001) was used. The test did not reveal a significant degree of saturation, even
under the very conservative assumption of an extreme asymmetrical tree (Iss = 0.075 <
Iss.cAsym = 0.391).
2.3 Phylogenetic analyses of molecular data
2.3.1 Tree based methods
Out-group selection
Out-groups were chosen to root the phylogenetic trees, paying attention to the out-
group criterion. This recommends the selection of relatives close to the in-group taxa but
without being part of the in-group (Felsenstein 2004). From the same genus Heleobia,
species from South America as well as from Eurasia were integrated in the analysing
dataset. South American species integrated from the eastern part of this continent
(Argentina, Uruguay, and Brazil) were: H. parchappi (Hepa), H. australis (Heau) and H.
davisi (Heda). From the western part of South America (Bolivia, Chile) were integrated: H.
sp. (He8725) and H. sp.(He8726, He8837). From Eurasia, sequences from species of
Semisalsa [S. aponensis (Seap), S. dalmatica (Seda), S. foxiamensis (Sefo), S. scammandri
(Sesc), S. stagnorum (Sest) and S. sp. (Sesp)] were chosen. Note that Semisalsa is not a full
accepted synonym of European Heleobia sp. (see section 1.3). Taxa from other genera
within the same family Cochliopidae were chosen from North America: Cochliopa sp.
(CO_SP), Pyrgophorus platyrachis (Pypl), Spurwinkia salsa (Spsa), Onobops jacksoni
(Onja), Heleobops carrikeri (Heca) and Heleobops docimus (Hedo). A sequence from
Potamopyrgus antipodarum (Poan) was used as outgroup taxon from a closely related
family Tateidae (Thiele, 1925). All out-group sequences (with the exception of the
specimens from Chile and Bolivia) were provided by the working group of Prof. Dr. Wilke
(University of Giessen).
17
Method selection
Different strategies can be used to construct phylogenetic trees from molecular
datasets. Most frequently used statistical programs work with the principle of a) Maximum
Parsimony (MP), b) Maximum Likelihood (ML), c) Bayesian Inference (BI) and d)
Neighbour Joining (NJ) (Nei & Kumar 2000, Felsenstein 2004, Knoop & Müller 2006).
MP, ML and BI are tree based methods using discrete character states of genetic
sequences. NJ is a clustering method based on genetic sequence distances. Minimal
Evolution (ME) is another approach, also based on genetic sequence distances like NJ, but
in opposite to the latter, ME is a tree building method like MP, ML and BI (Felsenstein
2004). The relative performance of different phylogenetic methods depends on numerous
factors such as degree of heterogeneity, sample size and model of sequence evolution (e.g.
Felsenstein 2004; Kolaczkowski & Thornton 2004). Each of these methods owns its
specific problems, advantages and disadvantages. In order to test the robustness of results
regardless of used algorithms, all methods were used to perform the phylogenetic analyses.
Model selection
Prior to the phylogenetic analysis, the Modeltest software version 3.6 (Posada &
Crandall 1998) was used in order to find the optimal model of DNA substitution, based on
the Akaike information criterion. The software performs hierarchical likelihood ratio tests
among 56 possible models of sequence evolution. The model selected was GRT+I+G
(general time reversible model with invariable sites and gamma distribution).
Bayesian inference analyses are based on the posterior probability distribution of

trees using Bayes’ theorem (Felsenstein 2004). The BI was performed using MrBayes
software version 3.1.2 (Roquist & Huelsenbeck 2003). The parameters were explored
using two Markov chain Monte Carlo (mcmc) simulations and sampling in intervals of 10
generations. As model of sequence evolution the GTR+I+G model was chosen. Invariable
sites and Γ distribution with the actual parameters were estimated during the calculation
process automatically by the software. Two independent runs were conducted, generating
one tree per ten sampled generations. The average standard deviation of split frequencies
was calculated after every 1,000 generations, and the analysis was terminated after it
became smaller than 0.01 (after 1,974,000 calculated generations). A 50% majority rule
consensus tree was then calculated from the sampled trees, ignoring the first 50,000 trees
as burnin.
18
Maximum likelihood is a statistical method to find the probability (expressed as

likelihood) of a tree based on the observed data and the assumed model of sequence
evolution (Felsenstein 1981). A ML tree was constructed using the RAxML 7.0.4 online
program. Similar to BI, the GTR+I+G model was chosen, executing 100 rapid bootstrap
inferences and thereafter a thorough heuristic ML search. All free model parameters were
estimated by RAxML. These are the optimal transition and transversion rate ratios, the
base frequencies and the site heterogeneities. The gamma model parameters were
estimated up to an accuracy of 0.1 Log Likelihood units.
Maximum parsimony was used to find a tree with the smallest number of changes
in nucleotide position in DNA sequences (Knoop & Müller 2006). The MP tree was
conducted with the PAUP software version 4.0b10 (Swofford 2002) by performance of
heuristic search with the tree-bisection-reconnection (TBR) branch swapping algorithm.
Subsequently the tree topologies were tested by bootstrapping (100 replicates), resulting in
a consensus tree (50% majority-rule).
Neighbour joining analyses work by clustering based on distances in a matrix
(Felsenstein 2004). The latter is a matrix of all pairwise genetic distances in the nucleotide
sequences (Felsenstein 2004). NJ analyses were performed with the MEGA software
version 4.0 (Kumar et al. 2004), using the Kimura-2-Parameter (K2P) genetic distance
estimates (Kimura 1980). Bootstrapping always comprised 1000 replicates.
Minimal evolution analyses are based on genetic sequence distances. The criterion
of optimum is the minimal main distance between all sequences, respectively the minimal
branch length of the reconstructed tree (Felsenstein 2004, Nei & Kumar 2000).
Phylogenetic analyses were performed in the MEGA software version 4.0.2, using the
Kimura-2-Parameter (K2P) genetic distance estimates (Kimura 1980). Bootstrapping
always comprised 1000 replicates. All trees were visualized using the TREEVIEW
software version 1.6.6.
2.3.2 Network analysis

Traditional (dichotomous) tree building approaches may perform poorly in analyses
within or between closely related species (Posada & Crandall 1998). Therefore the network
approach was chosen. Network analyses are very useful tools if input data are not tree-like
(Moulton 2003). Conflicting information in datasets can be reflected that way. Reticulate
evolution (i.e. hybridisation), rapid evolution, and speciation still in progress are regarding
19
examples (Moulton 2003, Knoop & Müller 2006). Alternative potential evolutionary paths
can be shown in the form of cycles (Bandelt et al. 1999). Network analyses were utilized in
order to demonstrate relationships among the putative closely related Heleobia spp. within
Lake Titicaca, to its neighboring lakes Umayo and Arapa, and to the species collected in
Bolivia and Chile.
Statistical parsimony (SP) networks were performed with the TCS software
version 1.21 (Clement et al. 2000), calculating the most-parsimonious relationships
between two haplotypes, based on DNA pairwise differences. The criterion of parsimony is
fulfilled by minimal numbers of mutational connections between pairs of sequences
(Clement et al. 2000). Haplotypes separated by mutational steps (more than ten) were
excluded by TCS 1.21 and have not been connected with the cladogram.
Median-joining (MJ) network analyses were performed with the Network software
version 4.5.0.0 (Bandelt et al. 1999). MJ reconstructs the shortest least complex network
based on maximum parsimony, inferring ancestral haplotypes and potential haplotypes,
evolutionary branchings and variants (Bandelt et al. 1999).
2.3.3 Molecular clock analyses

According to Zuckerhandl & Pauling (1965), the molecular clock hypothesis bases
on the approach of constant rates in nucleotide substitution in all clades within one
phylogenetic tree. The grade of divergence between two sequences should be therefore
proportional with the time passed since the split of these two sequences. Kimura’s theory
of neutral evolution supported this approach of clockwise DNA sequence evolution
(Kimura 1968).
The Molecular clock hypothesis was tested for the COI dataset by the performance
of two phylogenetic BI analyses with MrBayes version 3.1.2 (Huelsenbeck & Ronquist
2001), including exclusively unique haplotypes. If the mutation rate does not correspond
with divergence time, the molecular clock is not applicable. One topology was constructed
with and one without enforced clock. The BI topologies were compared statistically using
the software R version 2.7.0 (R Development Core Team 2007), to test whether the tree
without enforced clock fits the dataset better than the tree with enforced clock, performing
a hierarchical likelihood ratio test (hLRT). The likelihood values for the COI LRT were –
3760.999 (enforced clock) and –3773.725 (without enforced clock). Likelihood ratio was
25.45, at 69 degrees of freedom. The critical value (p = 0.05) was 89.39. The molecular
20
clock was accepted. Clock calibrations can be carried out for example using major
geological events (e.g. Isthmus of Panama, Mediterranean Salinity Crisis), fixing the
moment of splitting for these two sequences. The universal clock calibration for
“Protostomia” (Albrecht et al. 2006) was applied (3.83 ± 0.45 GTR+I+G/ million years).
2.4 Morphology
2.4.1 Shell parameters

The current taxonomy within the Titicaca Heleobia spp. is based on shell
morphological characters exclusively. Morphologic plasticity, partly in high degrees, is
described for most species of the lake. In some cases this situation was so distinctive, that
the presences of complete intermediate morphs was ascertained (see also Haas 1955,
Dejoux 1992). A principal component analysis (PCA) was performed in order to test
whether the Titicaca Heleobia spp. can be separated based on morphological character
with statistical analysis.
Equipment
For studying morphological characters of interest, digital images were made using a
light microscope Olympus SZX 12, and a digital camera system Colour-View FW. The
images were analysed and saved in the analySIS FIVE software (Soft Imaging System
GmbH, Münster, Germany). Shell dimensions were measured on the screen, using digital
callipers in the Corel PHOTO-PAINT software version 12.
Specimen selection
A total of 80 specimens were included in the analysis, in order to have a
representing range of the different morphotypes collected. Additionally, paratypes from
original descriptions were included, one representing the species, plus two sub-species of
H. andicola, representing the full range of this extreme variable species. Missing data for
the paratypes were measured on original paratype illustrations, using Corel PHOTO-
PAINT.
Character selection
Principal problems in morphological analysis within the Rissooidea are the paucity
of robust characters, their largely unknown phylogenetic relevance, and a potentially high
21
degree of homoplasy. Statistical character analysing is aggravated usually by the relatively

limited suite of characters within most rissoidean taxa (Hershler & Ponder 1998). The
Titicaca Heleobia spp. represent a remarkable exception to that. Restriction in character
selection is maybe the most ambitious challenge for practicable analyses, quite contrary to
the common problems in finding separating characters within other microgastropod taxa
(Hershler & Thompson 1992, Hershler & Ponder 1998, Wilke et al. 2001). Characters and
states proposed by Hershler & Ponder (1998) are basic to the here performed analyses, but
have been amplified with demonstrative characters and states outlined in the original
descriptions for the Titicaca Heleobia spp. (D’Orbigny 1835-1847, Pilsbry 1896, 1924,
Bavay 1904, Haas 1955, 1957, Blume 1958). However, over-splitting was, whenever
possible, avoided.
Character coding:
1. Number of whorls: n
2. Height of shell: mm
3. Wight of shell: mm
4. Height of last whorl: mm
5. Height of aperture: mm
6. Wight of aperture: mm
7. Position of aperture outline on the right in relation to the outline of previous whorls: (0)
out to the right; (1) in line; (2) set to the left
8. Position of aperture outline on the left in relation to the outline of previous whorls: (0)
out to the left; (1) in line; (2) set to the right
9. Whorl outline, dominating in most whorls (>3): (0) strong convex; (1) convex middle;
(2) lightly convex; (3) marginally convex; (4) flat; (5) light concave; (6) concave; (7)
winkled
10. First whorl with character number 7: n
11. Number of keel(s) in maximum: n (see Figure 8)
22
Figure 8. Three specimens of Heleobia spp. to show the recognisability of the character
“Number of keel(s) in maximum” (at the left side 0, in the middle 1 and at the right side 2).
12. Shape of keel(s): (0) not applicable; (1) marginal to medium; (2) strong
13. Outline of the last whorl: (0) strong convex; (1) convex middle; (2) lightly convex; (3)
marginally convex; (4) flat; (5) light concave; (6) concave; (7) winkled
14. Degree of insection of sutur: (0) from marginal to strong; (1) wide, beginning
uncoiling; (2) uncoileled (Figure 9)
Figure 9. Two specimens of Heleobia spp. to show the recognisability of the character
“Degree of insection of sutur” (here: wide, beginning uncoiling).
15. First whorl with character number 12: n

16. Presens of umbiculus: (0) absent; (1) marginal; (2) medium; (3) wide; (4) perforating
(sensu Haas, shell concavely hollowed, found in H. crawfordi); (5) not applicable (if
uncoiled)
23
17. Abundance of pitted structure on the shell surface: (0) absent; (1) present; (2) > 30%
(Figure 10)
Figure 10. One specimen of Heleobia spp. to show the recognisability of the character
“Abundance of pitted structure on the shell surface”.
18. Outer shape of operculum: (0) flat; (1) nucleus peaked; (2) spiraliform (Figure 11)
Figure 11. The three types of operculum shape of Heleobia spp. to show the
recognisability of the character “Outer shape of operculum”. Images made by light
microscope and electron microscope, drawings taken from Haas (1955).
Principal component analyses (PCA)

The PCA is a multivariant analysis used to find correlations between single
variables within a dataset, in order to reduce the number of variables by retaining
characteristics with the strongest contribution to its variance. All statistical analyses have
been performed in the STATISTIKA software version 7 (Amyuni Technologies). The
graphics have been created in the Microsoft EXCEL software version 2003.
24
2.4.2 Radula
Shell shape variation could be an expression of differences in ecological factors. Studies in
radula anatomy can discover potentially trophic specialization (Glaubrecht & Köhler 2004,
Von Rintelen et al 2004). Hubendick (1955) made examinations on all Heleobia spp. from
Lake Titicaca known up to that time (with the exception of H. crawfordi). He did not find
significant differences between all those species. The later described H. mirum (Haas,
1955) has never been studied in anatomical characters. Four individuals (He8678, He8679,
He8680, He8681) of this uncoiled species from two different populations were studied, to
show whether they differ from the other Heleobia spp. in radula morphology. Additionally,
radulae from 16 other individuals (five other nominal species and H. sp) have been studied
to compare the radula morphology of H. mirum with other Heleobia spp. of Lake Titicaca
(H. andicola culminea (He8242, He8486), H. andicola andicola (He8154, He7761), H.
andicola neveui (He7759, He8697); H. aperta (He8163); H. magna (He8247); H. otorni
(He8682); H. umbiculata (He8160, He8161, He8241); not determined specimens (He8158,
He8238, He8239, He8694).
The examinations have been performed with aid of the scanning electron
microscope (SEM). The SEM images were obtained in the Central Biotechnical Unit
(ZBB) at the Justus-Liebig-University in Giessen. Radulae were mounted on stubs, sputter-
coated with gold and then documented using a digital Scanning Electron Mircoscope
Philips XL-20.
25
3 Results
3 Results
3.1 Molecular analysis
3.1.1 DNA isolation and amplification

Due the small body size of the species studied, the DNA quantity was quite low.
The isolation method of Wilke et al. (2006) worked successful for most specimens, after
changing the last step in the protocol (redissolvation of DNA pellet) in primary 40 μl H2O
to then 20 μl. Altogether DNA from 56 specimens was isolated. For the majority of
isolation products the amplification was successful and resulted in PCR products from 48
specimens (see example in Figure 12).
Figure 12. Example of isolation checkgel of the DNA of Heleobia specimens. Next to
each sample lane, DNA preparation numbers are provided.
3.1.2 DNA purification

The ExoSAP-IT (USB Corporation, Cleveland, OH, USA) purification were used
for all PCR products and worked without problems.
3.1.3 DNA sequencing

Forward and reverse strands of 45 COI sequences could be aligned using the
LICOR AlingnIR version 2.0 software. These sequences included haplotypes from nine
species. Sequences from H. berryi have not been obtained. Bioedit version 7.0.5.2 software
(Hall 1999) was used to create the consensus sequences. After cutting the ends of the
sequences the sequences had a length of 638 bp. Six haplotypes were identical (Figure 14).
The sequences comprised 229 variable sites, 166 parsimony informative sites and an
26
3 Results
average nucleotide composition of G=19.9 %, A=23.0 %, T=17.4 % and C=39.6 %. The

uncorrected pairwise distances (p-distances) between two of these 45 COI sequences
constituted in maximum 0.066 (between Hesp8725 from Bolivia and He8360 from Lake
Titicaca), and between haplotypes from the Lake Titicaca basin, in maximum 0.026
(between He8354 and He8360).
3.2 Phylogenetic analyses of molecular data
3.2.1 Tree based methods

Compatible topologies were found using MP, ME, ML, NJ and BI analyses and
resulted in important point in equivalent topologies in all these method. A representative
tree is shown (Figure 13) by the BI topology. An overall result of similarity can be lined
out as follows:
1) All specimens from Lake Titicaca cluster in one clade.
2) Within the Titicaca clade the two haplotypes from Chile occur.
3) European Semisalsa spp. are the next related clade to the Titicaca clade.
4) The Bolivian haplotype is next related to the Eurasian Semisalsa spp..
5) North-American Heleobops spp. are next related to the Bolivian haplotype.
6) One Italian Semisalsa sp. haplotype clusters not with the other Semisalsa spp., but
within the North-American Heleobops spp.
7) H. parchappi and/ or H. australis from eastern South-America are closely related to the
Heleobops spp.
8) H. davisi from eastern South-America occurs within or next to the remaining out-
groups.
The clade for the Heleobia spp. from Lake Titicaca (including H. sp. from Chile) is
supported with 100 % in all methods used, with exception for ML (82 %). The topology
and its supports within this clade vary largely from method to method used. The split
between the Lake Titicaca (inclusive Chile) clade and the European Semisalsa spp. is
supported between 37% and 83%. The split between the latter two clades and the Bolivian
haplotype is supported between 49% and 83%. Most other clades are weakly supported
(<95%). The representing tree for all methods used results from the Bayesian Inference
method (Figure 13).
27
3 Results
Figure 13. Representative tree of the clade Heleobia spp. from Lake Titicaca (including H.
sp. from Chile), using the BI topology. The statistical supports are shown for three
important clades (starting from the top: Bayesian Inference, Minimal Evolution, Neighbour
Joining, Maximum Likelihood, Maximum Parsimony).
28
3 Results
3.2.2 Network analysis

An overall result of the TCS statistical parsimony network combined with
taxonomical information can be lined out as follows (Figure 14):
1) The network excluded the Bolivian haplotype of He8725 for its distance of more than
ten mutation steps.
2) Altogether six cases of identical haplotypes were found.
3) The net does not separate clear clades for nominal species.
4) Individuals of distinct morphology (different species) can obtain identical haplotypes.
5) One nominal species can obtain very different haplotypes.
6) The network suggested the two identical haplotypes 8489/ 7602 (H. andicola andicola/
H. andicola culminea) as having the highest probability of being the ancestral haplotype in
the analysis.
7) H. andicola is ubiquitary present in the network.
Figure 14. Result of the TCS statistical parsimony network combined with taxonomical
information for the clade Heleobia spp. from Lake Titicaca (including H. sp. from Chile).
29
3 Results
Each circle represents one haplotype (= COI sequence of one single specimen). The four
big circles represent cases of identical haplotypes for two specimens. The box in the
middle represents the ancestral haplotype (suggested by the software), in this case two
identical haplotypes. The haplotypes are connected by lines. Each small black dot
represents the distance of one mutation step. The colours are symbolizing the taxonomy.
Results of the TCS statistical parsimony network combined with geographical

information can be lined out as follows (Figure 15):
1) Specimens from different geographical regions can share identical haplotypes.
2) Specimens with different haplotypes can be of the same geographical distribution.
Figure 15. Results of the TCS statistical parsimony network combined with geographical
information for the clade Heleobia spp. from Lake Titicaca (including H. sp. from Chile).
The colours represent the geographical origin of the haplotypes.
30
3 Results
A combination of geographical with taxonomical information with the TCS

network shows that: 1) Specimens with identical species determination and identical
distribution can obtain distinct haplotypes (e.g. H. andicola culminea (He8343, He8344,
He8345) from sample locality TC_24 were collected in exactly the same sample site). 2)
The haplotypes from these lakes do not represent distinct clades within the network. The
specimens from the lakes Umayo and Arapa cluster within the network formed by different
haplotypes from Lake Titicaca.
The analysis performed with Network version 4.5 has shown the same above
mentioned results. Differences occur for example in the presence of additional potential
haplotypes close to central haplotypes He8489/ He7602. Those are suggested as ancestral
by Network 4.5, similar to TCS.
3.2.3 Molecular clock analysis

The divergence time was calculated for two splits within the clock-enforced tree
(Figure 16). The time of divergence between all haplotypes within the Titicaca clade was
estimated as 0.69 ± 0.27 Ma. The separation between the Titicaca clade and the Eurasian
Semisalsa spp. clade was calculated as 2.13 ± 0.59 Ma.
31
3 Results
Figure 16. Divergence time calculated for two splits within the clock-enforced tree for the
clade Heleobia spp. from Lake Titicaca (including H. sp. from Chile). The clock was
calibrated applying the universal molecular clock for “Protostomia” (Albrecht et al. 2006)
3.3 Morphological analyses
3.3.1 Shell characters

The results of the principal component analyses (PCA) shows which characters are
responsible for separating the dataset and can be lined out as follows (Figure 17):
1) Above (green circle): whorl structuring factors (e.g. keel, winkle, flatness).
32
3 Results
2) To the right (red circle): shell dimension.

3) Bellow (black circle): shell surface, umbiculus dimensions.
4) To the left (blue circle): other characters (e.g. operculum structure, sutur insection).
Figure 17. Principal component analyses (PCA) for several species of the genus Heleobia,
showing the four main sets in which the variables were grouped. For explanation of the
abbreviations of the variables see the methods’ chapter.
The figure 18 shows that:

1) The highest abundance is concentrated to the positive side of the first factor, spread
among the positive and negative sectors of the second factor.
2) The red vertical line denotes a gap of abundance occurring between some species.
33
3 Results
showing the distribution of the specimens according to the two principal factors.
The figure 19 shows that:

1) Most species cluster with other species.
2) The determined species differ largely in position.
3) H. andicola is ubiquitous distributed at the right side of the diagram.
4) Two specimens (belonging to H. andicola and H. berryi) are far distanced positioned
(outside the red circle).
5) Specimens with ambiguous determinations cluster between the smaller specimens
(down in middle) and H. andicola (at the right side), both connecting.
7) Specimens of H. mirum occur separated (at the left side).
34
3 Results
8) Specimen without determination cluster with H. mirum.

9) H. otorni takes with its three integrated specimens a close position to H. mirum.
showing the distribution of the species according to the two principal factors. Hest = H.
stiphra, Hema = H. magna, Heca = H. carinifera, Hela = H. lacustris, Hepa = H. parva,
Hecr = H. crawfordi, Hean = H. andicola, Hepr = H. profunda, Heum = H. umbiculata,
Hemi = H. mirum, Hebe = H. berryi, Heap = H. aperta, Heot = H. ortorni, Hefo = H.
forsteri.
3.3.2 Radula
The comparison of radulae from four specimens of Heleobia mirum with five other
species, and with four undetermined specimens, all of them belonging to the same genus,
did not show significant differences (e.g. comparison shown in figure 20). Given the great
35
3 Results
similarity in radula morphology among all members of Cochliopidae and related families,
only the central teeth could be considered as of importance for systematically
interpretations. The shape of the central cusp of the central teeth can be described as
“family like” with one pair of basal cusps (Figure 20).
Among the same individuals of all species, great differences were observed in the
general aspects of the central tooth: those used or “being old” were always more round-
shaped, while the “new” ones were always acute. One example of the variance occurring
within the same individual is shown in figure 20c, where the lower central tooth is pointed,
while the uppermost is rounded. These differences were found in relatively short distances
of no more than ten tooth in between.
Figure 20. Electron microscope images from radulae: a) Heleobia mirum, a_1) frontal,
a_2) central teeth lateral; b) H. andicola.
36
4 Discussion
4 Discussion
4.1 Phylogenetical position to putative Heleobia spp.
All former studies on Heleobia spp. of Lake Titicaca have been performed exclusively
by means of morphological and anatomical characteristics. Their phylogenetical position was
far from being clear, given the unknown phylogenetic significance of anatomical characters
within the Cochliopidae and related families (Hershler & Ponder 1998, Wilke 2001).
Furthermore, the majority of the Titicaca Heleobia spp. exhibit very special shell
morphologies, which make comparisons to other taxa difficult.
The molecular results on Titicaca mollusks, performed here for the first time, show
that the Heleobia spp. of Lake Titicaca are phylogenetically closely related to the other
Heleobia spp. worldwide (Figure 21). Heleobia andicola culminea from Lake Titicaca is the
type species of the genus Heleobia. Hence, the belonging to the genus Heleobia, defined for
all the other species by their phylogenetical position with respect to Heleobia andicola
culminea, has been shown.
The biogeographical aspects implicated in these results are interesting, considering the
disjunctive distribution of Heleobia, which is known from Eurasia as well as from America.
Surprisingly, species from Eurasia obtain the sister taxon position within the gene trees
regardless of the analyzing methods. This may be, however, an artifact due to a lack in
geographical sampling. The Andes in general and the Bolivian and Peruvian parts in
particular are poorly investigated, and additional material from the Altiplano and its
surroundings was obtained from two localities only. Nevertheless, the very young age of the
split between the Titicaca clade and the Eurasian clade obtained from the molecular clock
analyses (Figure 21) excludes that the disjunctive distribution comprising two continents can
be explained as a Gondwanaland relict, as proposed by Hershler & Thompson (1992). The
most recent common ancestor (MRCA) of this both clades is not older than 2.13 ± 0.59 Ma
(Figure 16), which is close to the proposed age of Lake Titicaca itself (Dejoux 1992a,b).
The inclusion of additional species from the Andes in the molecular clock analyses would not
result in a different time of this split. Molecular clock analyses are based on genetic distances
which are constant and not relative. An explanation of the settlement mechanism that caused
this pattern of distribution cannot be given with the data at hand and has to be worked out in
future studies. One possible explanation might be long-distance passive dispersal (e.g. living
snails or eggs attached to or swallowed by migratory birds; see discussion in Glöer 2002; see
Figure 22). Some species of the genus Heleobia are known to settle brackish costal habitats
37
4 Discussion
and having a bentho-pelagic development with free swimming veliger. The marine dispersal
of larvae might be another imaginable way of settlement.
For now, one species sampled in Italy (H. aponensis) shows a closer phylogenetic
relationship to American Heleobia spp., then to species from Europe (Figure 21c). This
indicates intercontinental faunal exchange twice. On the other hand, Heleobia davisi from
Brazil is more closely related to the outgroup taxa than to the other Heleobia spp., which
raises doubts whether this species belongs to the genus Heleobia. The undetermined species
from the southern Altiplano in Bolivia (Figure 21b) is closely related to the clade of Lake
Titicaca Heleobia spp., which can be expected due to their close geographical position. Most
of these clades are not highly supported, but were presented as final topology by all methods
used. A further study with a more extensive dataset of different Heleobia species is necessary
to find systematic positions with better statistical supports in all methods.
Figure 21. Heleobia species from: a) Rio Loa/ Chile, H. cf. loaensis, b) Bolivia, Heleobia sp.,
c) Holland, H. stagnorum
The presence of two specimens from Chile within the Titicaca clade in all gene
trees and network analyses (see Figure 13 and 14) implicates two possible explanations: a)
One possibility is, that the gastropods collected in Chile are the same species as the
gastropods from Lake Titicaca (e.g. the similar shaped H. andicola culminea), but with an
occurrence 600 km south of Lake Titicaca. This distribution pattern might be the result of
long-distance passive dispersal with similar mechanisms as discussed above for
intercontinental exchange. Or it is one example of a species which has a wider distribution
(relict of paleolimnological events (see section 4.4), or just a widespread occurrence of one
38
4 Discussion
species with a recent distribution in the whole central Andean region, between Lake Titicaca
and Rio Loa in North Chile and even broader, but until today this never has been detected).
Figure 22. Egg capsules of Heleobia spp. were found frequently, partly in high abundances
deposited at macrophytes (here Chara spp.).
b) The second scenario to explain the presence of two specimens from Chile within
the Titicaca clade, implicates that the two specimens from Chile belong to one of the 21
Heleobia species described for Chile (e.g. H. loaensis, endemic to the sample site at Rio
Loa; Valdovinos 2006), but they could not be identified due to ambiguity in species
determination. Given the considerable shell plasticity within both the Titicaca Heleobia spp.
as well as in the Chilean Heleobia spp. (Biese 1944, 1947), this explanation cannot be
excluded. Regardless of the process involved in the actual distributional pattern of the Chilean
specimens, molecular clock analyses have shown that this clade is the result of very recent
speciation processes. The divergence time for haplotypes within the Titicaca clade, including
two specimens from Chile, is less than 0.69 ± 0.27 Ma (Figure 16).
4.2 Monophyly of Titicaca Heleobia spp.

Strong differences in shell shaping of the Titicaca Heleobia spp. led the researchers to
separate them in different genera (see section 1.4). Given the special and distinct shell
morphological characters of most Heleobia spp. from Lake Titicaca (e.g. uncoiling, keel) this
decisions are understandable (Hubendick 1955). Later researchers postulated con-generic
39
4 Discussion
relationships that were supported by anatomical studies exclusively (Hershler & Thompson
1992).
The here performed analyses, for the first time also based on molecular methods, have
shown that the Heleobia spp. from Lake Titicaca represent a phylogenetically close related
group. Whether monophyly is present, depends on the taxonomical status of the Chilean
specimens, which cannot be answered with the data at hand, because of uncertainty in species
determination (see section 4.1). Monophyly for the Titicaca Heleobia spp. is only possible if
the specimens collected in Chile belong to the species occurring in Lake Titicaca (e.g. H.
andicola culminea) or if they are basal in the Titicaca clade. If not, the Titicaca Heleobia spp.
do not represent a monophylum, because that clade would not comprise a single common
ancestor and all the descendants of that ancestor (Wägele 2000).
Furthermore, the discussed taxonomical status of the Chilean species also implies
consequences for the present status of endemism for the Titicaca Heleobia spp., which would
be rejected in the case that the Chilean species actually belong to one of the Titicaca species.
Strictly spoken, the occurrence of H. andicola culminea in the two smaller lakes Umayo and
Arapa, close to Lake Titicaca (Figure 6), already contradicts the endemism of the Titicaca
Heleobia spp. Nevertheless, given the permanent connections between these wetlands systems
due to the short distance between them (<25 km), considering them as one limnological
complex is acceptable.
Again, both the status of endemism and monophylism of the Titicaca clade depends
on the taxonomical status of the two Chilean specimens, which is still not clear (Figures 13-
15). All the previous suggestions lead studies to discussions about the presumed species flock
status for the Titicaca Heleobia spp. (Dejoux 1992a,b), since the originally definition of a
species flock, comprising monophyly, endemicity and speciosity (Greenwood 1984), is not
fulfilled for the first two conditions (monophyly and endemicity).
A more recently redefinition of the term “species flock”, applicable to groups
conformed by as few as 3-4 closely related endemic species, with the immediate ancestor not
necessarily endemic to the lake (see Schön & Martens 2004), might be more appropriated for
the case of the Heleobia spp. present at Lake Titicaca. However, prior to any definition, some
points should be considered:
a) not all species described for Lake Titicaca are represented in the dataset (e.g. H.
berryi was found but no DNA could be obtained; H. carinifera and H. crawfordi were not
found),
40
4 Discussion
b) many specimens were of uncertain taxonomical classification (classified as

Heleobia sp.), since they were very close to the original descriptions, but failed to fit in it
partially (e.g. one whorl more, umbiculus present if absent is criteria, etc.; Especially two of
the smaller species H. lacustris and H. profunda with exact congruence to original
descriptions were never found. Some of them were always close to those with intermediate
morphological characters, or presented unique characteristics; see also section 4.3). The same
for other specimens with intermediate morphological characters (e.g. H. mirum and H. otorni,
see section 4.4), that have not been described before (e.g. Figure 23).
c) it is possible that the ancestor of the Titicaca clade had its origins outside the lake
system (e.g. an ancestor coming from Chilean waters (see section 4.4)).
4.3 Morphological characters in the view of molecular data

Species descriptions for Lake Titicaca Heleobia spp. are based exclusively on shell
morphological observation. As shown in the phylogenetical network analyses and the
statistical analyses of shell morphology, the species included in the analyses are not clearly
separable, neither by shell characters nor by the marker gene COI. Specimens with
intermediate morphological characters were found. These results might have several
causes:
a) In Lake Titicaca is one single species present, morphologically extreme plastic.
The low genetically divergences between haplotypes within the Titicaca sequences might
indicate that. The maximal uncorrected p-distance between two Lake Titicaca specimens is
2.6%. The idea of DNA barcoding, proposed first by Herbert (2003), declares a genetic p-
distance of mitochondrial COI-sequences more than 3% as separating two species. But, as
shown in various examples, COI distances are not per se defining taxa and are problematical
for interpretations in taxa that are not well established (Wilke 2003, Köhler 2007, Huang
2008).
b) In Lake Titicaca are distinct species distributed, but due to the young age (0.69 ±
0.27 Mya), the network analyses resulted not in separated clades for specimens congruent to
morphological species description. Incomplete lineage sorting is well known for taxa with
recent speciation, for complicating phylogenetic inferences due to ancestral polymorphism of
mitochondrial sequences (Maddison & Knowles 2006, Wilke et al. 2006). Ancestral
polymorphism can obscure real phylogenitical relationships within analyses by discordance
between species and gene trees. Starting from an ancestral state with variances in sequence
divergence in a population, there is some time necessary for accumulating mitochondrial
41
4 Discussion
sequence mutations with sufficient abundance to overcome the ancestral diversity. Therefore
the split of the Titicaca clade (inclusive the Chilenean specimens) might be even younger than
the molecular clock calculations suggest (Wilke 2004). Therefore future studies need to be
performed with nuclear markers (microsatellits, AFLP).
Given the enormous range of morphological variation within the Titicaca Heleobia
spp., without evident results for decisions about species level, it seems justifiable to outline
general points observed on thousands of specimens. General tendencies in shell character
variation can be stated:
a) complexity: high (e.g. keel baring) vs. low (e.g. whorl flatness)
b) size: elongation vs. miniaturization
c) shape: slender vs. globules
d) uncoiling
e) operculum structuring: outside (flat vs. pitted/ spiraliform) & inside (for some
species described as autapomorphy (pers. observ.: the same character state found in
other species))
f) umbiculus: from absent to extremely wide
g) shell surface: pitted structure, described for some species as taxonomical character
(e.g. H. forsteri pers. observ.) can also be found in other species, microscopy resulted
in the suggestion that this structure is probably an artifact of erosion (as already
mentioned by Blume 1958 and Dejoux 1992)
h) The character states can be present in different degrees and in combination.
i) Specimens were included in the morphological analyses that have never been
described before. Their shell morphologies could be interpreted as intermediate
(linking) forms between already described Heleobia sp. from Lake Titicaca (e.g.
between H. otorni and H. andicola neveui, see Figure 23).
42
4 Discussion
Figure 23. The specimen in the middle (Heleobia sp.) presents a combination of
morphological characters that are described for H. andicola neveui (keel) at the left side and
H. otorni (spiraliform operculum) at the right side.
For giving a more detailed example, the data for one species, the uncoiled H. mirum,
will be outlined here. The specimens within the graphics of the morphological analyses
(Figure 18) separated to the left are Heleobia mirum (and undetermined/ intermediate
specimens). These are separated from all the other species by a gap (shown by the red line).
As seen in figure 17, the responsible factors are the suture character (uncoiling) and the
external operculum structuring (e.g. spiraliform). Interestingly, one of the next species to the
right is Heleobia otorni. It is present here not as the only one species, but in opposite to other
species occurring close to H. mirum, they are present here with all (three) its specimens
included in the analysis. Heleobia otorni and H. mirum are supposed to be phylogenetically
close related, mainly by the similar shaped spiraliform operculum and the rounded shape of
the first (embryonal) whorls in both species (Hass 1957). In the here presented study shells
have been found that can be interpretated as intermediate morphs between H. otorni and H.
mirum, mainly by the presence of the suture character state “beginning of uncoiling” and
whorl outline from “winkled” to “keel” baring (Figure 5, 9). Tendencies for slight winkled
whorls have been reported for some specimens of H. otorni (Haas 1955). Dejoux &
Mourguiart (unpublished) found H. mirum with various degrees of uncoiling, from slightly to
extreme, the same as in the here presented study. Moreover, they found five specimens with
43
4 Discussion
flat instead of spiraliform opercula. The holotype of Heleobia mirum is described with a
rounded aperture which is slightly winkled only (Haas 1957). In the here presented study were
found also specimens with one and two keels (Figure 8). In Figure 23 a specimen presenting a
combination of all character states can be seen. The spiraliform operculum is described for
Heleobia otorni (and the uncoiled H. mirum). The keels are described for the extreme
sculptured variant of the type species, H. andicola neveui (besides H. mirum). Additionally,
the shell surface is covered with the pitted structure (described e.g. for H. forsteri).
Figure 24. Different examples of morphological variation in Heleobia mirum and H. otorni
with different degrees in uncoiling. Note that specimens were found as well with flat (at the
top in the middle) as with spiraliform operculum. Drawings taken from Dejoux & Mourguiart
(unpublished).
The example of combined character states might be an indication for independent

transitions between these character states. Reticulate mechanisms as hybridization should not
be excluded. Seehausen (2004) hypothesized hybridization as forcing factor in evolution. The
results in network analysis conflicting with species descriptions might result from this, but
cannot be proven by this method.
The examinations of radulae in H. mirum and intermediate morphs did not result in
observations of significant differences to other Heleobia spp. Hence, ecophenotypical
adaptation discovered due to distinct radular morphologies indicating trophic adaptation
cannot be assumed.
Other ecological aspects responsible for shell morphological change were discussed
by Gorthner (1992), regarding to ancient lakes. One interesting hypothesis is about separation
effects by macrophyts within microhabitats. For example, lakes with Chara spp. might have
44
4 Discussion
physical barriers formed by the high density of these algae. The extensive presence of Chara
spp. in the littoral of Lake Titicaca might be of interest to study this theory. Additionally,
Gorthner (1992) tried to work out ancient lake intrinsic factors that could potentially be
responsible for the thalassoid gastropod shell shape (marine like shell structures, e.g. rips)
found in ancient lakes. After all his analyses, just the age resulted to be the characteristic that
all ancient lakes have in common.
4.4 Biogeography
The findings of fossil Heleobia mirum shells in Pleistocene horizons in the Southern
Altiplano by Dejoux & Mourguiart (unpublished) leads to some points in recent and
paleolimnological biogeography. The recent lake systems of the Altiplano are relict waters of
paleolakes with extension of about almost the whole Altiplano (<1.8 Ma; Wirrmann 1992).
Recent distribution patterns of endemic species occurring in the now isolated waters may
represent relicts of such paleolakes. Or, as mentioned for many species flocks in ancient lakes,
originated (intralacustrine speciation) herein. Distribution pattern with evident connection to
the paleohistory have been found in other taxa of the Altiplano region. Mourguiart (2000)
found those for ostracoda, with significance in morphological differentiation by ecological
changes (e.g. salinity). Lüssen et al. (2003) investigated fishes of the genus Orestias, by
mitochondrial DNA analysis, and found the occurrence of taxa which are part of the Orestias
Titicaca species flock, in waters situated in the southern Altiplano in Chile. For many species
of Amphibia, Ostracoda and Pisces of the Altiplano, the diversity centers in Lake Titicaca
(Dejoux 1994).
Four endemic Heleobia spp. are described for smaller lakes (relict waters) at the
Altiplano (Haas 1955). In contrast to that, one widespread distributed species, Heleobia
cumingii (D’Orbigny, 1835) (Figure 24) is to be found from the central Peruvian coast down
to central Chile, and even in the Andes (Valdovinos 2006). Interestingly, the presence of this
species is also reported from rivers at the Altiplano (e.g. Rio Ramis, which is entering Lake
Titicaca), and moreover from the small Lake Umayo next to Lake Titicaca, but never from
Lake Titicaca itself (Haas 1955). In opposition to that, Heleobia andicola culminea is the only
Titicaca Heleobia spp. with distribution outside of Lake Titicaca, namely in Lake Umayo and
Lake Arapa (Haas 1955; own observations). These close located lakes with river connections
to Lake Titicaca could be interpreted as a "contact zone" between endemic Lake Titicaca
species and widespread "surrounding" species. Brooks (1950) and Boss (1978) summarized a
pattern found in ancient lake gastropod communities which is “robustness” against
45
4 Discussion
immigration of more widespread (non-endemic) and presumably more generalist taxa. This
eco-insularity was thought to act as a buffer against newer invasions.
Figure 25. Drawings from: a) Heleobia cumingii, b) H. haasi. Drawings taken from Haas
(1955)
Heleobia andicola culminea, the plesiomorphest species in its shell shape of all
Heleobia spp. from Lake Titicaca (by comparison with the outgroups), is the species by far
with the largest abundance and range distribution. The network analyses suggested haplotypes
of Heleobia andicola as ancestral. Haplotypes of H. a. culminea are ubiquitous in the network
and haplotypes of this species identical with other species were found. Heleobia a. culminea
is known as a species with an enormous morphological variation (see Figure 5). Hass (1955),
showed that there is an Heleobia-andicola-complex, while Dejoux and Mouguiart
(unpublished) found that H. otorni and H. mirum also form a sort of species complex (also
found during the present study). Furthermore, during this study some specimens difficult to
classified were found, which might be consider as an intermediate stage between the
Andicola-complex and the complex formed by H. otorni and H. mirum. This would suggest
that all species might form a single complex in Lake Titicaca, and even more, all of them
might be variations (morphotypes) form a single species (H. andicola sensu lato). Outside
Lake Titicaca, just one of all those morphotypes is to be found in the lakes Umayo and Arapa:
Heleobia andicola culminea.
From one of the previously mentioned small relict lakes (Lagunilla Lagunilla), one
endemic species was described [H. haasi (Haas, 1955)], with the same type of spiraliform
operculum as the one exhibited by H. otorni and H. mirum (Figure 25). On the other hand,
Hubendick (1955) observed in the widespread species H. cumingii also tendencies for such a
spiraliform operculum. This pattern of: a) variability of shell morphology inside the Lake
Titicaca, b) one morphological character (spiraliform operculum) typical for Titicaca species
46
4 Discussion
also found inside relict water, and c) this character also found in species with distribution
between this relict water and the "contact zone" between Lake Titicaca species, would suggest
events of hybridization. Such events could be the result of secondary contact after long
isolation periods. Since the widespread distributed species H. cumingii is to be found from the
central Peruvian coast down to central Chile, is not possible to assure what implications that
might have for the biogeographical interpretation of the two specimens found on the Rio Loa
at northern Chile. Future studies should include specimens from all Altiplano reclict waters
(e.g. Lagunilla Lagunilla, Lake Popoo) and waters adjoining the Altiplano.
For the moment it is plausible to say that the Heleobia spp. from Lake Titicaca might
be a recently radiated clade (possible with H. andicola as ancestral) or just a single species
with a very high morphological plasticity (Figure 26).
Figure 26. Several specimens of the Heleobia spp. from Lake Titicaca to show the
morphological variation. The specimens are not organized according to their taxonomical
relationships, but according to their morphological variations.
47
5 Literature
5 Literature
Acosta J (1979) Las Algas Superficiales del Lago Titicaca. Lima – Centro de
Investigaciones Pesceras, pp. 12-16.
Albrecht C, Lohfink D, Schultheiß R (2006) Dramatic decline and loss of mollusc diversity
in long-lived lakes in Greece. Tentacle 14: 11-13.
Allmendinger RW, Jordan TE, Kay SM, Isacks BI (1997). The Evolution of the Altiplano-
Puna plateau of the Central Andes. Annu Rev Earth Planet Sci 25: 139-174.
Anistratenko V, Albrecht Ch, Wilke Th (unpublished). Evolutionary trends in hydrobiid
gastropods (Mollusca : Caenograstropoda) of the Azov-Black Sea Basin :
comparative morphology, palaeozoogeography, ecology and taxonomy. pp. 1-26.
Argollo J, Mourguiart P (2000) Late Quaternary climate history of the Bolivian Altiplano.
Quat Int 72: 37-51.
Baker PA (2005) Holocene hydrological variation at Lake Titicaca, Bolivia/ Peru, and its
relationship to North Atlantic climate variation. J Quat Sci 20: 655-662.
Bandelt HJ, Forster P, Rohl A (1999) Median-joining networks for inferring intraspecific
phylogenies. Mol Biol Evol 16: 37-48.
Bavay A (1904) Mission de Crequi-Montfort et Senechal de la Grange en Amerique du
Sud. Mollusques terrestres et fluviatilis recoltes par le Dr Neveu-Lemaire. Bull Soc
zool France 29: 152-156.
Biese WA (1944) Revisión de los moluscos terrestres y de agua dulce provistos de concha
de Chile. Parte I, Familia Amnicolidae. Bol Mus Hist Nat Chile 22: 169-190.
Biese WA (1947) Revisión de los moluscos terrestres y de agua dulce provistos de concha
de Chile. Parte I, Familia Amnicolidae (continuación). Bol Mus Hist Nat Chile 23:
63-77.
Blume W (1958) Littoridinen aus dem Titicacasee (Mollusca). Opus Zool 25: 1-8 .
Boss KJ (1978) On the evolution of gastropods in ancient lakes, in: Fretter, V., Peake, J.
(Eds.), Pulmonates, Vol. 2a, Systematics, Evolution and Ecology. Academic Press,
London, New York, San Francisco, pp. 385–428.
Brooks JL (1950) Speciation in ancient lakes. Q Rev Biol 25: 131–176.
Cano W (1952) Estudio geografico, historico y sociologico del Lago Titicaca. Edicion
Moreno Buenos Aires, pp. 11-74.
Clement M, Posada D, Crandall KA (2000). TCS: a computer program to estimate gene
genealogies. Mol Ecol 9: 1657-1660.
48
5 Literature
Constantini ML, Sabetta L, Mancinelli G, Rossi L (2004) Spatial variability of the

decomposition rate of Schoenoplectus tatora in polluted area of Lake Titicaca. J Trop
Ecol 20: 1-11.
Davis G (1982) Historical an ecological factors in the evolution, adaptive radiation, and
biogeography of freshwater mollusks. Amer Zool 22: 375-395.
Davis G, Wilke Th, Spolsky Ch, Qiu C, Qiu D, Xia M, Zhang Y, Rosenberg G (1998)
Cytochrome oxidase I – based phylogenetic realtionships among the Pomatiopsidae,
Hydrobiidae, Rissoidae and Truncatellidae (Gastropoda: Caenogastropoda:
Rissoacea). Malacologia 40: 251-266.
Dejoux C (1992) The benthic populations: Distribution and seasonal variations. In: Lake
Titicaca – A Synthesis of Limnological Knowledge (Ed. by C. Dejoux and A. Ilties).
Kluwer Academic Publishers, pp. 383-404.
Dejoux C (1992) The Mollusca. In: Lake Titicaca – A Synthesis of Limnological
Knowledge (Ed. by C. Dejoux and A. Ilties). Kluwer Academic Publishers, pp. 311-
336.
Dejoux C ( 1994) Lake Titicaca. Arch Hyd Beih Ergebn Limnol 44: 35-42.
Dejoux C, Mourguiart (After 1992) P. Ecpomastrum mirum (Mollusque Hydrobiidae) Du
Lac Titicaca, Un Probleme De Taxinomie. Unpublished
de Lattin G (1967) Der Titicacasee. In: Grundriss der Zoogeographie, Fischer Verlag Jena,
pp. 203-205.
D’Orbigny A (1835) Synopsis terrestrium et fluviatilium Molluscorum in suo per
American meridionalem itinere collectorum. Mag Zool 5: 1-44.
Felsenstein J (1981) A likelihood approach to character weighting and what it tells us
about parsimony and compatibility. Biol J Linn Soc 16: 183-196.
Felsenstein J (2004) Inferring phylogenies. Sinauer Associates, Sunderland, Massachusetts,
664 pp.
Folmer O, Black M, Hoeh WR, Lutz R, Vrijenhoek R (1994) DNA primers for
amplification of mitochondrial cytochrome c oxidase subunit I from diverse
metazoan invertebrates. Mol Mar Biol Biotech 3: 294-299.
Gilson HC (1939) Description of the Expedition. In: The Transactions of the Linnean
Society of London – The Percy Sladen Trust Expedition to Lake Titicaca in 1937.
Cambridge University Press, pp. 4-20
Glaubrecht M, Köhler F (2004) Radiating in ariver: systematics, molecular genetics and
morphological differation of vivparous freshwater gastropods endemic to the Kaek
49
5 Literature
River, Central Thailand (Cerithioidea, Pachychilidae). Biol J Linnean Soc 82: 275-
311.
Glöer P (2002) Süßwassergastropoden Nord- und Mitteleuropas: Bestimmungsschlüssel,
Lebensweise, Verbreitung. ConchBooks, Bad Kreuznach, 327 pp.
Gorthner A (1992) Bau, Funktion und Evolution komkplexer Gastropodenschalen in
Langzeit-Seen. Stutt Beit Naturk Ser.B, 190: 1-173.
Gorthner A (1994) What is an ancient lake?, in: Martens, K., Goddeeris, B., Coulter, G.
(Eds.), Speciation in Ancient Lakes. Arch. Hydrobiol. 44, 97–100.
Greenwood PH (1984) What is a species flock? In: Echelle AA, Kornfield I (Eds)
Evolution of fish species flocks. Orono Press, University of Maine, pp 13-19.
Fritz SC, Baker PA, Seltzer GO, Ballantyne A, Tapia P, Cheng H, Edwards RL (2007)
Quaternary glaciation and hydrologic variation in the South American tropics as
reconstructed from the Lake Titicaca drilling project. Quat Res 68: 410-420.
Haas F (1955) Mollusca: Gastropoda. In: The Transactions of the Linnean Society of
London – The Percy Sladen Trust Expedition to Lake Titicaca in 1937. Cambridge
University Press, pp. 275-308.
Haas F (1957) Eine neue endemische Schnecke aus dem Titikaka-See. Arch Moll 85: 137-
139.
Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis
program for Windows 95/98/NT. Nucl. Acids Symposium Series 41, 95-98.
Herbert P, Cywinska A, Ball S, de Waard J (2003) Biological identification through DNA
barcodes. In: Proceedings of the Royal Society B: Biological Sciences 270: 313-321
Hershler R, Ponder W (1998) Descripition of Shell Characters. In: A Review of
Morphological Characters of Hydroboiid Snails. Smith Cont Zool 600: 3-9.
Hershler R, Thompson F (1992) Genus Heleobia Stimpson 1865b. In: A Review of the
aquatic Gastropod subfamily Cochliopinae (Prosobranchia: Hydrobiidae). Malacol
Rev Suppl 5: 45-57.
Huang D, Meier R, Todd PA (2008) Loke Ming Chou Slow Mitochondrial COI Sequence
Evolution at the Base of the Metazoan Tree and Its Implications for DNA Barcoding.
J Mol Evol 66: 167-174.
Hubendick B (1955) The Anatomy of the Gastropoda. In: The Transactions of the Linnean
Society of London – The Percy Sladen Trust Expedition to Lake Titicaca in 1937.
Cambridge University Press, pp. 309-327.
50
5 Literature
Huelsenbeck JP, Ronquist F (2001) MRBAYES: Bayesian inference of phylogenetic trees.

Bioinformatics 17: 754-755.
Ilties A, Carmouze J-P, Lemoalle J (1992) Physico-Chemistry. In: Lake Titicaca – A
Synthesis of Limnological Knowledge (Ed. by C. Dejoux and A. Ilties). Kluwer
Academic Publishers, pp. 89-97.
IMARPE (2000) Descripcion Breve del Ambito del Lago Titicaca. Papele de Informe del
Instituto del Mar del Peru de Puno del Julio 2000, pp. 1-16.
Kabat A, Hershler H (1993) Diagnosis of the Familiy Hydrobiidae. In: The Prosobranchia
Snail Family Hydrobiidae (Gastropoda: Rissoodiea): Review of Classification and
Supraspecific Taxa. Smith Cont Zool 547: 5-6.
Kimura M (1968) Evolutionary Rate at the Molecular Level. Nature 217: 624-626.
Kimura M (1980) A simple method for estimating evolutionary rates of base substitutions
through comparative studies of nucleotide sequences. J Mol Evol 16: 111-120.
Kumar S, Tamura K, Nei M (2004) MEGA3: Integrated software for Molecular
Evolutionary Genetics Analysis and sequence alignment. Briefings in Bioinformatics
5: 150-163.
Knoop V, Müller K (2006) Vom Alignment zum Stammbaum – phylogenetische
Methoden in der Übersicht. In: Gene und Stammbäume. Ein Handbuch zur
Molekularen Phylogentik. Spektrum Akademischer Verlag, pp. 70-72.
Köhler F (2007) From DNA taxonomy to barcoding – how a vague idea evolved into a
biosystematic tool. Mitt Mus Nat Berl Zool 83 Suppl: 44-51.
Kolaczkowski B, Thorton JW (2004) Performance of maximum parsimony and likelihood
phylogenetics when evolution is heterogeneous. Nature 431: 980-984.
Korniushin AV (2004) The Bivalve Mollusca Fauna of Ancient Lakes In The Context Of
The Historical Biogeography Of The Balkan Region, pp. 219-241. In H. I. Griffiths,
B. Krystufek, and J. M. Reed (Eds): Balkan Biodiversity - Pattern and Process in the
European Hotspot, Kluwer Academic Publishers. Korniushin
Lauzanne L (1992) Native species The Orestias. In: Lake Titicaca – A Synthesis of
Limnological Knowledge (Ed. by C. Dejoux and A. Ilties). Kluwer Academic
Publishers, pp. 405-419.
Lavenu A (1992) Formation and geological evolution. In: Lake Titicaca – A Synthesis of
Limnological Knowledge (Ed. by C. Dejoux and A. Ilties). Kluwer Academic
51
5 Literature
Lüssen A, Falk TM, Villwock W (2003) Phylogenetic patterns in populations of Chilean

species of the genus Orestias (Teleostei: Cyprinodontidae): results of mitochondrial
DNA analysis. Mol Phyl Evol 29: 151-160.
Maddison WP & Knowles LL (2006) Infering Phylogeny Despit Incomplete Lineage
Sorting Syst Biol 55(1): 21-30
Martens K (1997) Speciation in ancient lakes (review). Trends Ecol Evol 12: 177-182.
Martens K, Coulter G, Goddeeris B (1994) Speciation in Ancient Lakes - 40 years after
Brooks. Arch Hydrobiol Beih Ergebn Limnol 44: 75-96.
McGuire JA et al. (2007) Mitochondrial Introgression and Incomplete Lineage sorting
Through Space And Time: Phylogenetics Of Crotaphytid Lizards. Evolution 61-12:
2879-2897
Moulton V (2003) SplitsTree: A network-based tool forexploring evolutionary
relationships in molecular data, pp. 312-328. In M. Salemi, and A.-M. Vandamme
(Eds): The Phylogenetic Handbook - A Practical Approach to DNA and Protein
Phylogeny, Cambridge University Press, Cambridge.
Mourguiart P (2000) Historical Changes in the Environment of the Lake Titicaca:
Evidence from Ostracod Ecology and Evolution. Adv Ecol Res 31: 497-520.
Nei M, Kumar S (2000) Molecular evolution and phylogenetics. Oxford University Press,
Oxford, 333 pp.
Newell ND (1949) Geology of the Lake Titicaca Region, Peru and Bolivia. The Geological
Society of America 36: 5-25.
Northcote TG (2000) Ecological Interactions Among an Orestiid (Pisces: Cyprinodontidae)
Species Flock in the Littoral Zone of Lake Titicaca. Adv Ecol Res 31: 399-420.
Pampuch Th, Echalar A (1998) Der Altiplano. In: Bolivien. Becksche Reihe Länder, pp.
17-18
Pilsbry HA (1924) South American land and freshwater molluscs. Notes and descriptions.
I. Molluscs of Lake Titicaca. Proc Acad Nat Sci Philad 76.
Posada D, Crandall KA (1998) MODELTEST: testing the model of DNA substitution.
Bioinformatics 14: 817-818.
R Development Core Team (2007) R: A Language and Environment for Statistical
Computing. Vienna: R Foundation for Statistical Computing.
Remigio EA, Lepitzki DAW, Lee JS, Hebert PDN (2001) Molecular systematic
relationships and evidence for a recent origin of the thermal spring endemic snails
52
5 Literature
Physella johnsoni and Physella wrighti (Pulmonata: Physidae). Can J Zool 79: 1941-
1950.
Rigsby CA, Bradbury JP, Baker PA, Rollins SM, Warren MR (2005) Late quaternary
paleolakes, rivers, and wetlands on the Bolivian Altiplano and their paleoclimatic
implications. J Quate Sc 20: 671-691.
Roche MA, Bourges J, Cortes J., Mattos R (1992) Climatology and Hydrology. In: Lake
Titicaca – A Synthesis of Limnological Knowledge (Ed. by C. Dejoux and A. Ilties).
Kluwer Academic Publishers, pp. 63-88.
Romero E (1973) Altiplano del Titicaca. In: Peru – Una nueva geografia, pp. 183-225.
Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under
mixed models. Bioinformatics 19: 1572-1574.
Rossiter A, Kawanabe H (2000) Ancient Lakes: Biodiversity, Ecology and Evolution. Adv
Ecol Res 31, Academic Press, London.
Rowe HD (2003) Late Quaterny lake-level changes constrained by radiocarbon and stable
isotope studies on sediment cores from Lake Titicaca, South America. Glob Plan
Change 38: 273-290.
Schäfer A (1997) Der Titicacasee. In: Biogeographie der Binnengewässer. B.G. Teubner
Stuttgart, pp.136-139.
Schön I, Martens K (2004) Adaptive, pre-adaptive and non-adaptive components of
radiations in ancient lakes: a review. Org Div Evol 4: 137-156.
Seehausen O (2004) Hybridization and adaptive radiation. Trends in Ecology and
Evolution 19 (4) 198-207
Seehausen O (2006) African cichlid fish: a model system in adaptive radiation research.
Proc Royal Soc B 273: 1987-1998.
Sterba G (1996) Unterfamilie Orestianinae – Titicacakärpflinge. In: Süßwasserfische der
Welt. Weltbildverlag, pp. 563-564.
Swofford DL (2002) PAUP. Phylogenetic analysis using parsimony (and other methods).
Sinauer Associates, Sunderland, Massachusetts.
Taylor D (1966) A remarkable snail fauna from Cuahuila, Mexico. The Veliger 9: 152-
228.
Theissen KM, Dunbar RB, Rowe HD, Mucciarone DA (2008) Multidecadal- to century-
scale arid episodes on the northernAltiplano during the middle Holocene.
Palaeogeography, Palaeoclimatology, Palaeoecology 257: 361-376.
53
5 Literature
Valdovinos C (2006) Estado de conocimiento de los gasterópodos dulceacuícolas de Chile.

Gayana 70: 88-95.
von Rintelen T, Wilson AB, Meyer M, Glaubrecht M (2004) Escalation and trophic
specialization drive adaptive radiation of freshwater gastropods in ancient lakes on
Sulawesi, Indonesia. Proc Roy Soc Lond 271: 2842-2850.
Wägele JW (2000) Monophyla. In: Grundlagen der Pyhlogenetischen Systematik . Verlag
Dr. Friedrich Pfeil, pp. 70-73
Wilke Th (2003) Salenthydrobia gen. nov. (Rissooidea: Hydrobiidae): a potential relict of
the Messinian salinity crisis
Wilke Th, Davis G, Falniowski A, Giusti F, Bodon M, Szarowska M (2001) Molecular
systematics of Hydrobiidae (Mollusca: gastropoda: Rissooidea): testing monophyly
and phylogenetic relationships. Proc Acad Nat Sci Phil 151: 1-21.
Wilke Th, Pfenniger M, Davis G (2002) Anatomical variacion in cryptic mudsnail species:
statistical discriminacion and evolutionary significance. Proc Acad Nat Sci Phil 152:
45-66.
Wilke Th, Davis G, Qiu D, Spear R (2006) Extreme mitochondrial sequence diversity in
the intermediate Schistosomiasis host Ocomelania hupensis robertsoni: Another case
of ancestral polymorphism? Malacologia 48: 143-157.
Wilson AB, Glaubrecht M, Meyer A (2004) Ancient lakes as evolutionary reservoirs.
evidence from the thalassoid gastropods of Lake Tanganyika. Proc Roy Soc Lond
271: 529-536.
Wirrmann D (1992) Geomorphology and Sedimentation. In: Lake Titicaca – A Synthesis
of Limnological Knowledge (Ed. by C. Dejoux and A. Ilties). Kluwer Academic
Xia X, Xie Z (2001) DAMBE: Software package for data analysis in molecular biology
and evolution. J Heredity 92: 371-373.
Zuckerhandl E, Pauling L (1965) Evolutionary divergence and convergence in proteins, pp.
97-166. In V. Bryson, and H. J. Vogel (Eds): Evolving Genes and Proteins,
Academic Press, New York.
54

Tesis Kroll

Uploaded by

Copyright:

Available Formats

Tesis Kroll

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tesis Kroll

Uploaded by

Copyright:

Available Formats

Evolutionary patterns of endemic gastropods in ancient Lake

Titicaca: the genus Heleobia (Rissooidea: Cochliopidae)

Rostock, August 2008

Submitted 4. August 2008 by

I want to express my gratitude to:

1.1 Ancient lakes in evolutionary- and biodiversity science

1.2 Lake Titicaca

1.2.1 Titicaca research

1.2.2 General information (Origin, morphology and hydrology)

Table 1. Summary of physical and limnological data (from Dejoux 1994).

1.3 The genus Heleobia (Stimpson, 1865)

1.4 Heleobia in Lake Titicaca

This problem becomes more severe due to a high degree of intraspecific

Figure 5. Set of seven recognizable intermediate morphotypes of Heleobia andicola

Dejoux (1992b) described massive problems of precise species identification for

1.5 Aims of this work

problems in interpreting the morphologic and anatomical characters of the Titicaca

2 Materials and Methods

2.1 Sampling design and specimens studied

Figure 6. Map of the study area indicating the sampling sites.

Specimens with ambiguous species determination were labeled as Heleobia sp.

2.2 Molecular analysis

2.2.1 DNA isolation

2.2.2 DNA amplification

2.2.3 DNA purification

2.2.4 DNA sequencing

2.2.5 Sequence saturation

2.3 Phylogenetic analyses of molecular data

2.3.1 Tree based methods

Bayesian inference analyses are based on the posterior probability distribution of

Maximum likelihood is a statistical method to find the probability (expressed as

2.3.2 Network analysis

2.3.3 Molecular clock analyses

2.4.1 Shell parameters

degree of homoplasy. Statistical character analysing is aggravated usually by the relatively

15. First whorl with character number 12: n

Principal component analyses (PCA)

3.1 Molecular analysis

3.1.1 DNA isolation and amplification

3.1.2 DNA purification

3.1.3 DNA sequencing

average nucleotide composition of G=19.9 %, A=23.0 %, T=17.4 % and C=39.6 %. The

3.2 Phylogenetic analyses of molecular data

3.2.1 Tree based methods

3.2.2 Network analysis

Results of the TCS statistical parsimony network combined with geographical

A combination of geographical with taxonomical information with the TCS

3.2.3 Molecular clock analysis

3.3 Morphological analyses

3.3.1 Shell characters

2) To the right (red circle): shell dimension.

The figure 18 shows that:

The figure 19 shows that:

8) Specimen without determination cluster with H. mirum.

4.2 Monophyly of Titicaca Heleobia spp.

b) many specimens were of uncertain taxonomical classification (classified as

4.3 Morphological characters in the view of molecular data

The example of combined character states might be an indication for independent

Constantini ML, Sabetta L, Mancinelli G, Rossi L (2004) Spatial variability of the