1. Introduction

1.1. The origin and evolution of mammals

1.1.1. The origin of mammals


The rise of mammals during the Mesozoic era was one of the most important events in vertebrate evolution (Kemp 2005). With over 5000 extant species and some 4000 fossil taxa, mammals are a highly diverse and successful animal group (Crompton and Sun 1985, Novacek 1992, Crompton and Luo 1993, Luo 2007). Today, mammals comprise taxa in all size ranges, from a 6 cm shrew to a 33 m blue whale (Kemp 2005, Luo 2007), and have developed an enormous number of ecological specializations such as scavenging, burrowing, gliding, arboreal, scansorial, and aquatic lifestyles (Luo and Wible 2005, Martin 2005, Ji et al. 2006, Meng et al. 2006, Luo 2007).

However, there is still a lively debate regarding the origin of this successful group. What mammalian character was the most important adaptation? So-called key innovations include lactation and preceding juvenile care, a manifold behaviour facilitated by an increase in brain size, as well as endothermy (Jerison 1971, Long 1972, Hopson 1973, Crompton et al. 1978, Koteja 2000, Kemp 2005).

An endothermic physiology, which maintains a constant body temperature, enables an animal to be active under a wider range of temperatures and allows for a more complex body plan (Kemp 2005). A high rate of sustainable aerobic activity allows for more sustained exercise and a higher maximum running speed in endotherms than in ectotherms and has advantages in e.g. predation, territory size, and predator avoidance (Koteja 2004, Kemp 2005). On the other hand, endothermy requires an immense increase in food intake (Bakker 1971, Kemp 2005, Kemp 2006).


But how did endothermy evolve? It is undoubted that such a complex character is unlikely to have evolved in a single step. Several hypotheses have been discussed, among which are the thermoregulation-first hypothesis via miniaturization, the aerobic capacity hypothesis, and the parental provision hypothesis (McNab 1978, Bennett and Ruben 1979, Ruben 1995, Farmer 2000, Koteja 2000). Also, invading a nocturnal niche is believed to be amongst the features that supported the evolution of endothermy in early mammals (Crompton et al. 1978). 

1.1.2. Nocturnality – a prerequisite of endothermy

Although a nocturnal life habit in early mammals had been suggested by Jerison in 1971, Crompton et al. (1978) were the first to propose that the acquisition of homeothermy enabled early mammals to invade a nocturnal niche without having to increase their resting metabolic rate. Only in a second step, the authors proposed, did mammals become diurnal, and a higher body temperature and resting metabolic rate were only secondarily acquired (Crompton et al. 1978). This perspective has been generally accepted and published in classical textbooks. Therein, the first mammals, such as Morganucodon and Megazostrodon, are pictured to have been small, highly active, nocturnal animals, and insulated by fur, living in the shadow of the dinosaurs, with life habits similar to modern hedgehogs, tenrecs, and shrews who feed on insects (Fig. 1) (Bakker 1971, Jerison 1971, Carroll 1988, Kemp 2005).

Figure 1. Reconstruction of Morganucodon, an early mammal from the Late Triassic of China, South Africa, India, and Europe (Kemp 2005).



Until the mid 1970s, Mesozoic mammals were known only from teeth, but in recent years, more and more mammaliaform fossils preserving more complete skeletons have been found (Luo et al. 2001, Ji et al. 2002, Meng et al. 2006, Luo et al. 2007). Presently, the nocturnality hypothesis is supported solely by features found in the fossil record and include huge orbits,

enlarged olfactory regions in the brain and improved hearing suggesting a nocturnal lifestyle, small body size, and a tooth shape similar to that of modern insectivorous animals (Carroll 1988, Kemp 2005).


In 1942, Walls described differences between the eyes of nocturnal and diurnal animals, such as eye size and shape, shape of the pupil, extent of curvature of cornea and lens, as well as visual cell shape and number, and their arrangement in the retina. Since the publication of this work, similar studies with additional information have followed (Ahnelt und Kolb 2000, Kaskan et al. 2005). However, soft-tissue morphological evidence for nocturnality as well as potential signals of endothermy in early mammals is lacking and, so far, cannot be provided by fossil findings (Ruben 1995). Ancient DNA studies have their challenges and limits, too, especially when it comes to molecules older than one million years (Hofreiter et al. 2001, Olson and Hassanin 2003, Schweitzer et al. 2009).

Hence, this thesis approaches the question of whether early mammals had indeed been nocturnal in a novel manner, i.e. by means of molecular techniques. Selection patterns acting on visual pigment genes, which support potential adaptation to changes in life habits, are investigated, and hypothetical ancestral visual pigments are inferred and resurrected, and their function is tested in vitro.

1.1.3. Evolution of therapsids and the acquisition of endothermy

After its erection by Linnaeus in 1758, Mammalia sensu lato (or Mammaliaformes according to Rowe 1988) is now recognized as a monophyletic group that includes the common ancestor of Sinoco no don, as well as living monotremes, and living therians (Crompton and Sun 1985, Luo et al. 2002, Kemp 2005). They form the sister group to Reptilia (Modesto and Anderson 2004) and are characterised by unique features such as insulation by fur, mammary glands, and a dentary-squamosal jaw articulation (Kemp 2005). Within synapsids, mammals belong to the clade Therapsida (Broom 1905). Therapsids originated in the Middle Permian and were one of the most successful amniote groups during the Permian, but were strongly affected by the P/T extinction event (Fig. 2) (Kemp 2005). Only anomodonts and cynodonts survived into the Triassic and the latter experienced a remarkable Middle Triassic diversification (Fig. 2) (Abdala and Ribeiro 2010). Cynodonts, more precisely Tritheledontidae and Tritylodontidae, are the closest relatives of mammals (Fig. 2) (Luo 1994). Tritylodontidae are small herbivorous forms that originated in the Early Triassic and which were abundant and remarkably diverse in the Early Jurassic (Sues 1986, Luo 1994, Abdala and Ribeiro 2010).


Tritheledontids are small insectivorous/carnivorous therapsids from the Late Triassic (Luo 1994, Kemp 2005, Abdala and Ribeiro 2010). During the Late Triassic and Early Jurassic early mammals and these cynodont tritylodontids and tritheledontids show a cosmopolitan distribution in the supercontinent Pangaea.

Figure 2. Simplified synapsid phylogeny based on accepted literature

(Reisz 1986, Luo 1994, Abdala 2007, Abdala et al. 2008, Fröbisch et al. in press).


Gow (1985) proposed a cynodont-mammal transition characterised by sudden and profound changes, such as small body size, determinate growth, the presence of a promontorium, and diphyodonty. However, it is now widely accepted that the transition happened stepwise in transitional clades such as tritheledontids, Sinoconodon, and Adelobasileus (Brink 1956, Luo 1994, Kemp 2005, Luo 2007). For example, the stapedial process, which is part of the mammalian middle ear, is present in tritylodontids and the mammaliaform Morganucodon, but not found in tritheledontids and the mammaliaform Sinoconodon (Luo 2007). Also, the quadratojugal, which allows for more mobility in the middle ear, is already lost in Sinoconodon and Morganucodon (Luo 2007). The evolution of sensitive hearing facilitated by the middle ear in earliest mammalian forms also argues for the exploitation of nocturnal habitats (Luo 2007).

Endothermy is considered one of the major features in mammalian evolutionary history as it enables an animal to be active independent from the temperature of its surroundings and because it is thought to have been a key adaptation that gave rise to the diverse and successful mammalian group. This aspect of mammalian physiology has been the subject of various studies, in birds as well as in both living synapsids and their close extinct relatives, i.e. non-mammalian therapsids (McNab 1978, Ruben 1995, Kemp 2005, Sánchez-Villagra 2010). In contrast to pelycosaur-grade synapsids, non-mammalian therapsids have evolved many modifications in their skull and postcranium as well as presumably in their physiology towards a mammalian organisation which possibly helped to overcome temperature fluctuations of the terrestrial environment (Kemp 2005).


More precisely, in a review Sánchez-Villagra (2010) observed that fibrolamellar bone, which is an indicator of rapid osteogenesis, overall rapid growth, and endothermy, is found in some therapsids, e.g. cynodonts.

Nasal turbinal bones are found to be present in the skull of all mammals, as well as in some therapsids (Hillenius 1992, Hillenius 1994, Laaß et al. 2010). These bones are associated with reduction of respiratory water loss, but are thought to have evolved in association with elevated ventilation rates and the evolution of endothermy (Hillenius 1992).

Crompton et al. (1978) proposed that endothermy initially arose as an adaptation that permitted the exploitation of a nocturnal niche, which was facilitated by an insulating fur that reduced the rate of heat loss, while maintaining a relatively low body temperature and metabolic rate, as is also the case in living monotremes. It was only in a subsequent shift to diurnal activity that early mammals acquired a higher metabolic rate and a higher body temperature in order to withstand temperature fluctuations (Crompton et al. 1978).

1.2. Enigmatic monotremes, the most basal mammals

1.2.1. Monotremes


Monotremes are the basalmost living mammals. They form the sister group to marsupials and placentals, i.e. Theria, though, this currently widely held view is sometimes still challenged by the so-called ‘Marsupionta hypothesis’, which states that marsupials and monotremes are sister to Theria (Janke et al. 2002, Grützner and Graves 2004, Bininda-Emonds et al. 2007, Rowe et al. 2008).

The name ‘Monotremata’ (Bonaparte 1837) means ‘single opening’ and refers to the common external opening for the urinary, defecatory, and reproductive systems; the cloaca (Warren et al. 2008). Today, monotremes, also called Pro(to)theria, consist of only five species: the semi-aquatic, duck-billed platypus (Ornith o rhychus anatinus Shaw 1799), the short-beaked echidna (Tachyglossus aculeatus Shaw 1792), and three species of long-beaked echidnas (Zaglossus attenboroughi Flannery and Groves 1998, Zaglossus bartoni Thomas 1907, Zaglossus brujini Peters and Doria 1876).

All living monotremes are nocturnal, homeothermic, endemic to the Australian continent, and show a low rate of reproduction (Dawson et al. 1979, Rissmiller 1999, Werneburg and Sánchez-Villagra 2010). They are insulated by fur, produce milk, have a single dentary and possess three middle ear bones, just like all other mammals (Campbell and Reece 2009). In sperm shape and chromosome arrangement, however, monotremes are unique among mammals (Watson et al. 1996). Unlike all other mammals, monotremes have cloacae, lay eggs, and have a reptile-like/sprawling gait (Campbell and Reece 2009). Females lack nipples, so the young suck milk directly from the abdominal skin (Warren et al. 2008, Campbell and Reece 2009). A genome analysis of the platypus revealed that the monotreme genome has many unique micro RNAs (miRNAs), but also shares some other miRNAs with either mammals or reptiles (Warren et al. 2008). Overall, monotremes exhibit an intriguing mosaic of reptilian and mammalian characters, in terms of anatomy, physiology, and reproduction (Griffiths 1989). Adult monotremes lack teeth (Warren et al. 2008), whereas fossil forms have "tribosphenic" teeth, which are one of the hallmarks of extant mammals (Li and Luo 2006, Rowe et al. 2008).


The origin of Monotremata presumably occurred sometime in the Late Triassic/Early Jurassic, a date supported by fossil as well as molecular data (Luo et al. 2002, Woodburne et al. 2003, Phillips et al. 2009). According to a recent study on Teinolophos, a new monotreme fossil from the Early Cretaceous of Australia, the divergence of the two residual living monotreme genera, platypus and echidna, occurred earlier than molecular estimates have suggested, i.e. in the Early Cretaceous (Rowe et al. 2008).

1.2.2.  Tachyglossus aculeatus, the short-beaked echidna

The echidna (Tachyglossidae Gill 1872), also known as the spiny anteater, is named after the monster of Greek myth, meaning ‘she viper’. Echidnas are covered in coarse hair and spines, and have elongate and slender snouts. With their short and strong limbs and large claws, they are powerful diggers (Griffiths 1989). Echidnas have a low body temperature, which is around 7°C below the usual range of placental mammas, a low metabolic rate, and are able to reduce their energy output by torpor and hibernation (Schmidt-Nielsen et al. 1966, Nicol and Andersen 2007). During the rainy season, it is inactive and shelters under shrubs and trees (Griffiths 1989).

The short-beaked echidna is the most widely distributed extant monotreme, and can be found both in Australia and southwestern New Guinea, where it occupies a diverse range of habitats from the coast to the highlands (Griffiths 1989, Nicol and Andersen 2006, Nicol and Andersen 2007). It has an adult body mass of about 3-4 kg and with a documented lifespan of approximately 50 years, which is 3.7 times that predicted from its body mass, Tachyglossus is exceptionally long-living (Hulbert et al. 2008). It feeds on insects, in particular termites and ants, which it catches with its distinctive snout and specialized tongue (Griffiths and Simpson 1966, Griffiths 1989). Although not threatened by extinction, the populations of short-beaked echidnas have been reduced due to hunting, habitat destruction, and exposure to invasive predatory species and diseases.


The short-beaked echidna,  Tachyglossus aculeatus, is nocturnal or crepuscular, depending on the temperature of its surroundings (Fig. 3) (Boisvert and Grisham 1988). The echidnan retina displays several features which are thought to result from adapting to dim-light vision, such as a circular pupil as well as the lack of oil droplets and a nictating membrane (Gresser and Noback 1935, Walls 1942, Young and Pettigrew 1991, Rowe 2000). The echidna has long been thougt to possess a pure rod retina (Bolk et al. 1934, O’Day 1952). However, Young and Pettigrew (1991) identified the presence of twin cones, which constitute 10-15% of the photoreceptors in the retina and have all the ultrastructural characteristics of the cones of placental mammals. The distribution of these cones is similar to that of cones in the retina of the nocturnal cat and their density is higher than that seen in some nocturnal primates and in the nocturnal rabbit (Young and Pettigrew 1991).

Figure 3. A short-beaked echidna, Tachyglossus aculeatus, in Australia

(Photo: Jasmina Hugi). 

1.3. Rhodopsin, a vertebrate visual pigment

1.3.1. The visual signaling cascade


Visual pigments, also called opsins, form the first crucial step in the visual transduction cascade (Yau 1994, Blumer 2004). In tetrapods, there are up to five different visual pigments, located within the rods and cones in the retina of the eye (Bowmaker and Hunt 2006). Cone opsins mediate colour (photopic) vision and include short-wavelength opsins (SWS) 1 and 2, a middle-wavelength pigment (MWS or Rh2), and a long-wavelength opsin (LWS) (Bowmaker and Hunt 2006, Yokoyama 2008, Wald 1968). Rhodopsin (Rh1) is the only visual pigment responsible for vision at night and/or dim-light (scotopic vision) (Menon et al. 2001, Yokoyama 2008). At intermediate light levels (mesopic vision), both rods and cones contribute to vision (Peichl 2005).

All opsins absorb light at different characteristic wavelengths, ranging from UV at about 350 nm to far red at about 630 nm (Fig. 4) (Yokoyama 2008). Coulour discrimination depends on

the presence of two or more types of cone photoreceptors containing opsins that show absorption maxima in different regions of the visible spectrum (Fig. 4) (Szél et al. 1996).


It has been suggested that rods have evolved from cones, with the M/LWS opsin class evolving first, followed by SWS1, SWS2, and finally the Rh class (Okano et al. 1992, Carleton et al. 2005).

Figure 4. Wavelength diagram. SWS 1 absorbs light at 355-440 nm, SWS2 at 410-490 nm, MWS at 480-535 nm, LWS at 490-570 nm, and Rhodopsin at about 500 nm.

http://www.energymedc.com/images/light spectrum.jpg.


Visual pigments are composed of a protein moiety (opsin), which is a member of the G-protein-coupled receptor family, and a light absorbing chromophore, namely 11-cis retinal, which is a derivative of vitamin A (Wald 1968, Sugawara et al. 2010); though some fish, reptiles, and aquatic mammals use a derivative of A2 (Menon et al. 2001). 11-cis retinal, which is covalently linked to the opsin via a protonated Schiff base at a highly conserved residue Lys296 in transmembrane helix 7, absorbs a single photon (Fig. 5) (Baylor et al. 1979, Heck et al. 2003, Park et al. 2008).

Figure 5. Structural formula of 11-cis retinal.



Photon absorption causes an isomerization from 11-cis to all-trans retinal and, further, to a conformational change of the protein moiety. This results in the dissociation of all-trans retinal from the opsin (Palczewski et al. 2000). The active rhodopsin, also called meta II state, activates transducin, a cytoplasmic membrane G-protein, by loading it with guanosine triphosphate (GTP), which in turn causes phosphodiesterase (PDE) to increase its activity, thereby lowering the concentration of cyclic guanosine monophphate (cGMP), an intracellular second-messenger molecule (Blumer 2004, Imai et al. 2005). A decrease in cGMP concentration leads to the closure of cGMP-regulated Na+ and Ca2+ ion-specific channels in the outer cell membrane and, further, to a hyperpolarized membrane potential (Blumer 2004). This light-induced hyperpolarization of the cell membrane influences second-order visual neurons by modulating the rate of neurotransmitter (glutamate) release from the synaptic terminal of the photoreceptor (Yau 1994). This chain of signaling events is also called "the vertebrate phototransduction cascade" (Fig. 6) (Blumer 2004).

Figure 6. The phototransduction cascade in the vertebrate eye

(Blumer 2004). 


One photoexcited rhodopsin molecule activates hundreds of transducin copies (Sagoo and Lagnado 1997, Menon et al. 2001). Thus, the amplitude of the photoreceptor response is dependent on how efficiently the phototransduction cascade is activated by the visual pigment (Sakurai et al. 2007).

Turn-off of photoreceptor cells is accomplished by a protein called RGS9 (regulator of G-protein signaling 9), which accelerates the transducin’s ability to hydrolyse GTP, which is the rate-limiting step in the photoresponse (Sagoo and Lagnado 1997, Blumer 2004).

Eventually, rhodopsin is restored by recombining enzymatically produced 11-cis retinal from isomerized all-trans retinal in the dark, which is delivered from adjacent retinal epithelial cells (Palczewski et al. 2000, Heck et al. 2003).

1.3.2. Rhodopsin, a G protein-coupled receptor


Rhodopsin is the visual pigment mediating vision at night and/or dim-light (Menon et al. 2001, Yokoyama 2008). It consists of five exons and four introns. Its protein-coding sequence is composed of approximately 1044 nucleotides, hence, 348 amino acids (Fig. 7) (Palczewski et al. 2000). There are seven transmembrane helices (TM), which are embedded in the membrane and encompass 194 amino acids in total (Fig. 7) (Menon et al. 2001, Sakmar et al. 2002).

Figure 7. Secondary structure of bovine rhodopsin

(Sakmar et al. 2002).


Packed in the crystal lattice to form an array of helical tubes, the extracellular surface domain comprises an amino-terminal tail and three interhelical loops; the cytoplasmic domain comprises a carboxyl-terminal tail and three cytoplasmic loops (Fig. 7) (Palczewski et al. 2000, Sakmar et al. 2002).

Rhodopsin is temperature-sensitive (McKibbin et al. 2007). With an isoelectric point at pH 5.43, rhodopsin is an acidic protein; it has more glutamic and aspartic acid than basic lysine and arginine residues (Radding and Wald 1956, Kito et al. 1968).

It is ascertained that the chromophore is covalently bound to the opsin at the highly conserved Lys296 (Heck et al. 2007, Park et al. 2008), but which residues participate in holding the 11-cis retinal inside the binding pocket before photoisomerization is still debated (Schädel et al. 2003, Park et al. 2008, Hildebrand et al. 2009).


Figure 8. Three-dimensional structure of bovine rhodopsin.

Downloaded from RCSB Protein Data Bank (www.pdb.org, ID: 1u19) and visualized in PyMOL (www.pymol.org - The PyMOL Molecular Graphics System, Version 1.3, Schrödinger, LLC.). 

To date, the bovine rhodopsin is the best studied of all visual pigments (Fig. 8) (Menon et al. 2001, Sakmar et al. 2002, Palczewski 2006). However, not only rod opsin sequence data but also biochemical and functional properties have now been analyzed in a variety of vertebrate taxa, including fish, amphibians, reptiles, and mammals (Wald and Brown 1958, Nathans and Hogness 1983, Nathans and Hogness 1984, Kawamura and Yokoyama 1998, Sakmar et al. 2002, Imai et al. 2005, Imai et al. 2007). Surprisingly, although they are the last survivors of the most basal clade of extant mammals, not much is known about the visual capacities of monotremes. So far only the rod opsin gene sequence and absorption maximum as well as single exons of two cone opsins, i.e. SWS2 and LWS, of the platypus have been published (Davies et al. 2007). Another study on the visual pigments of both platypus and echidna only addresses cone pigments (Wakefield et al. 2008). Thus, for a more reliable taxon sampling, incorporating the echidna rod opsin in this study was elementary. Furthermore, by studying the rhodopsin of the short-beaked echidna and its biochemical and functional properties in detail, this thesis also encompasses the first characterisation of a rhodopsin from a nocturnal animal, pinpointing differences to that of a diurnal animal; in this case the bovine rhodopsin.

1.4. Ancestral sequence reconstruction and selective constraint analyses

1.4.1. Resurrecting ancient genes


Ancestral sequence reconstruction (ASR) is nowadays widely used to test hypotheses about the functional evolution of ancient genes, to provide a glimpse into their evolutionary history, and, most importantly, to get a better understanding of the paleobiology of ancient organisms that presumably possessed these genes and proteins (Chang et al. 2002a, Thornton 2004, Chang et al. 2007).

In 1963, Pauling and Zuckerkandl were the first to introduce the idea of resurrecting ancient genes, after studying amino acid sequences of vertebrate hemoglobulin chains. Then in 1971, Fitch was the first to develop an algorithm to reconstruct ancestral character states, using the parsimony principle and, thus, also taking phylogeny into account. It was not until the 1980s that this algorithm was incorporated in computer programs such as PAUP (Swofford 1985), and that the first study using this method to infer ancestral sequences was published (Baba et al. 1984). Due to concurrent improvements in DNA synthesis, in 1990, Stackhouse et al. were the first to successfully resurrect a functional ancestral gene that had been inferred by parsimony. Since parsimony has some intrinsic limitations (Thornton 2004), it was a significant step in ancestral sequence reconstruction methods when Yang et al. developed PAML in 1995, a program which uses a maximum likelihood algorithm to infer ancestral sequences (Koshi and Goldstein 1996), thus allowing further knowledge about the process of molecular evolution to be included (Thornton 2004). Since then, resurrecting ancient proteins has become an increasingly popular tool for addressing evolutionary questions as it provides a great opportunity to study the mechanisms of functional change during evolution at a molecular level (Fig. 9) (Chang and Donoghue 2000, Chang et al. 2002a, Gaucher et al. 2003, Shi and Yokoyama 2003). Though it has its limitations, e.g. its hypothetical nature, the dubious accuracy of selecting the right algorithm to fit the data, as well as the limited interpretations based on recreating single molecules, ASR can provide data where paleontologists and the fossil record reach their limits (Chang et al. 2002a, Chang et al. 2007).

Ancestral sequences are usually inferred using a maximum likelihood phylogenetic algorithm, an alignment of extant sequences, a specific phylogeny, and a probabilistic model of sequence evolution. For each internal node in the phylogeny as well as each site in the sequence, an ancestral state with the highest likelihood is calculated. The confidence in any inferred ancestral state is described as its posterior probability, which is defined as the likelihood of the state divided by the sum of the prior-weighted likelihoods for all states. One uncertainty in the maximum likelihood approach is the assumption that the alignment, tree, model, and model parameters are a priori known to be correct. Another method, the Bayesian approaches addresses these sources of uncertainty by estimating likelihoods over several possible trees or parameter values, each weighted by its posterior probability (Smith et al. 2010). However, it has recently been suggested that maximum likelihood estimates are as reliable as Bayesian methods (Smith et al. 2010).


Figure 9. The ancestral gene resurrection strategy

(Thornton 2004). 

Ancestral sequence reconstruction methods allow researchers to address fascinating evolutionary questions and peer deep into the past. For example, Gaucher et al. (2003) were able to infer information about the lifestyle of Precambrian organisms. In order to understand in what environment the earliest life forms evolved, they investigated EF-Tu, a GDP-binding elongation factor which regulates the rate of protein synthesis and which is highly temperature-sensitive, in the common ancestor of all bacteria (Gaucher et al. 2003). Measuring the thermostability and GDP-binding affinity of E.coli bacteria containing resurrected EF-Tu genes indicate that bacterial ancestors had EF-Tus with an optimal GDP-binding temperature of 65°C, suggesting that bacteria originated in a thermophilic environment (Gaucher et al. 2003). Chang et al. (2002a) investigated the visual capacities of ancestral archosaurs living in the early Triassic, approximately 240 million years ago. Their data suggests that inferred ancestral archosaur rod opsins had been functional for vision at night and/or in dim-light (Chang et al. 2002a). Another study focused on steroid hormone receptors that evolved before the origin of Bilateria (Thornton et al. 2003). They found that these receptors were lost in invertebrates (Thornton et al. 2003). In chordates, they had experienced an increase in affinity for steroids after having first evolved oestrogen receptor-like functions (Thornton et al. 2003).


Since the vertebrate visual system is so adaptive, ancestral sequence reconstruction is a great tool for investigating rhodopsin, the visual pigment mediating scotopic vision, which then allows for fathoming visual capacities and life habits of early mammals. Hence, this approach was used to infer hypothetical ancestral mammalian rhodopsins, among others.

1.4.2.  In vitro expression systems in vision research

The visual system is one of the five senses that provide input for perception. This highly specialized and adaptive system is triggered by a large range of different light levels. It is instrumental in the survival of an animal, and changes can have profound consequences for the organism it inhabits.

In order to understand this crucial system and the proteins involved better, the evolutionary history of the different opsins involved in the visual signaling cascade and the differences they exhibit have been subject to various studies (Okano et al. 1992, Bowmaker and Hunt 2006); permitted by major improvements in in vitro expression systems in recent years. I n vitro expression systems allow not only for studying molecular properties of visual pigments


but also for synthesising hypothetical ancestral opsins (Oprian et al. 1987, Chang et al. 2002a, Sakmar et al. 2002, Chang 2003, Parry et al. 2005).

Furthermore, various biochemical assays have now been developed in order to characterise visual pigments and to identify differences between rod and cone opsins, including hydroxylamine stability, meta II decay and retinal regeneration, transducin activation, as well as acid bleaching (Kito et al. 1968, Shichida et al. 1994, Starace and Knox 1998, Imai et al. 2005, Imai et al. 2007, Sakurai et al. 2007).

Simultaneously, site-directed mutagenesis experiments have become a popular approach in vision research as they allow for the identification of key sites that are potentially responsible for changes in the different types of visual pigments (Sakmar et al.1989, Imai et al. 1997, Carvalho et al. 2006). Altering a specific amino acid can test if this exact amino acid has a significant impact on a protein’s function, eventually leading to a far-reaching adaptation and possibly to the origination of a newly adapted protein (Chang et al. 2007).


Thus, to date, much is known about biochemical and functional differences between cone and rod pigments, but detailed studies characterising and comparing differences in the rhodopsin of a nocturnal animal to that of a diurnal one are lacking. This thesis in part comprises the in vitro expression and the first detailed characterisation of a rod opsin from a nocturnal animal, the short-beaked echidna, potentially allowing the biochemical and functional properties of inferred and synthesised ancestral pigments to a nocturnal or a diurnal lifestyle to be determined.

1.4.3. Selective constraint analyses

In addition to ancestral gene reconstruction, the identification of selective constraint acting on genes of interest, has become a more popular approach in molecular evolutionary research in recent years (Yang and Bielawski 2000, Tan et al. 2005, Zhao et al. 2009a).

If mutations do not code for another amino acid, as is often the case if they occur in the third codon position, they are called synonymous (silent) substitutions (Page and Holmes 2006). Whereas those that lead to the translation of a different amino acid are referred to as non-synonymous (replacement) substitutions (Page and Holmes 2006).


In a highly adaptive system, adaptive changes, which can be a result of an accelerated rate of non-sysnonymous substitutions (dN) over synonymous substitutions (dS), can be traced using selective constraint analysis. The strength of selection acting on protein-coding genes is assessed by estimating ω, which is the ratio of non-synonymous (dN) to synonymous (dS)

substitutions (Yang 2002). Positive selection is identified whenever ω = dN/dS > 1 (Yang 2002, Pie 2006). For if ω = 1 and ω < 1 this would indicate neutral and purifying selection, respectively (Yang 2002, Pie 2006). Detected positive selection is a clear signal of adaptive evolution driven by selection (Yang 2002).

Selective constraint methods are now widely used. For example, Bakewell et al. (2007) investigated the degree of positive selection in human and chimpanzee genes and found more genes undergoing positive selection in chimp than in humans since their split; a finding which is in sharp contrast to the common belief that humans experienced more phenotypic adaptations than chimpanzees. Metzger and Thomas (2010) studied other G-protein coupled receptors, the CC chemokine receptor proteins, and found evidence for positive selection acting on residues in extracellular domains rather than in intracellular domains, which might be due to ligand-binding and pathogen interactions in the extracellular domains. Although selective constraint analyses provide a glimpse into the evolution of protein-coding genes, one must be aware that these analyses need to be carried out with care. For example, Tan et al. (2005) investigated the selective constraint acting on several opsins in primates and concluded that nocturnality could not have been the ancestral state. However, they did not take all exons of the short-, middle-, and long-wavelength opsins into account, and thus, disregarded important information (Tan et al. 2005).


With the visual system being a highly adaptive system, this method is nowadays often used in vision research, addressing not only paleobiological questions concerning e.g. vision capacities in ancestral primates and bats, but also ecological diversification in fish due to adaptations in their visual pigments (Sugawara et al. 2002, Spady et al. 2005, Tan et al. 2005, Zhao et al. 2009a, Shen et al. 2010).

Thus, this approach was used in this thesis to investigate the vertebrate rhodopsin and its single amino acids were inferred for mammalian and other branches, in order to make inferences of if and how the visual pigment responsible for dark and dim-light had experienced significant modifications in early mammals presumably due to changes in life habits.

1.5. Objectives of this thesis

This thesis represents the first study that investigates whether the first mammals had indeed been nocturnal, as indicated by the fossil record, by means of molecular evolution. Its focus lies on rhodopsin, the one visual pigment which is responsible for vision at night and/or dim-light.


1) The rhodopsin of the short-beaked echidna was expressed in vitro and investigated in detail. The echidna was interesting because, on the one hand, it represents one of the two last survivors of monotremes, the most basal mammals. On the other hand, it is a nocturnal animal and, so far, a detailed characterisation of a rhodopsin of a nocturnal animal is lacking.

2) Hypothetical ancestral rhodopsin amino acid sequences for the nodes Amniota, Mammalia, and Theria were inferred by maximum likelihood estimates and the proteins were expressed in vitro. Their biochemical and functional properties were examined and compared to rhodopsins of a nocturnal and a diurnal animal, i.e. echinda and bovine, respectively.

3) Selective constraint analyses were carried out in order to evaluate if the rhodopsin had experienced any dramatic changes in its function in tetrapods and along the branch leading to Mammalia in particular.



© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML-Version erstellt am: