The Evolution of Marsupial and Monotreme Chromosomes

Deep divisions in class Mammalia provide an opportunity to look back in time and reconstruct the events shaping their genomes. Mammals are divided into 3 major lineages that diverged from a common ancestor between 161 and 217 million years ago (mya) (fig. 1) [Phillips et al., 2009]. The egg-laying monotremes (subclass Prototheria) are the earliest offshoot of the class Mammalia and are represented by 1 extant species of platypus and 4 species of echidna. Marsupial (infraclass Metatheria) and eutherian mammals combined make up the subclass Theria. Marsupials and eutherians diverged 143–178 mya [Phillips et al., 2009; Luo et al., 2011]. This deep divergence of eutherian, marsupial and monotreme lineages is particularly valuable for gaining insight into the evolution of mammalian genomes and the reconstruction of ancestral chromosomes.

One of the major differences between these 3 mammalian lineages is their mode of reproduction. Marsupials give birth to altricial young, and evolution has resulted in more emphasis on a sophisticated lactation system than in prolonged in utero development as observed in eutherian species. It is important to stress that, despite laying eggs, monotremes possess features that are distinctively mammalian; they are fur-bearing animals and, even in the absence of nipples, feed their young milk that is produced by glands and oozes onto the skin of the mother’s abdomen to be licked by their young.

The unique features of marsupials and monotremes have intrigued biologists since they were first described. Among those features are their chromosomes, which differ markedly between these 2 groups. Marsupials are renowned for their low diploid numbers and characteristically large and well-conserved chromosomes, whereas monotremes have high diploid numbers, chromosome size differences reminiscent of reptiles and complicated sex chromosome systems.

Great insight into the evolution of mammalian chromosomes can be gained from including representative marsupials and monotremes in comparative genomic studies. Although many studies have reconstructed the ancestral eutherian karyotypes from comparisons of different eutherian species, an understanding of the homologies and chromosome rearrangements between all 3 mammalian lineages is required to accurately elucidate the rearrangements that have occurred throughout mammalian chromosome evolution and the reconstruction of the plesiomorphic karyotype of all mammals.

Here, we describe the current knowledge of marsupial and monotreme chromosomes and the insight gained into chromosome evolution within these lineages.

Marsupials are a diverse group of mammals consisting of over 300 species, divided between the Americas and Australasia. The American (Ameridelphia) and Australian (Australiadelphia) species diverged from a common ancestor about 80 mya and it is generally acknowledged that the ameridelphian family Didelphidae were the first offshoot of the marsupial lineage, diverging from the lineage that gave rise to the American families, then to the Australidelphia species around 80 mya [Beck, 2008; Meredith et al., 2008].

One feature of marsupial karyotypes that has fascinated cytogeneticists is their remarkable conservation. This is quite unlike eutherian mammals, in which extensive chromosome rearrangement has occurred between species. Marsupial genomes are typically packaged into relatively few, very large chromosomes, which have facilitated cytogenetic studies. Basic karyotype studies reporting chromosome number, size and morphology for many species in each marsupial family showed that marsupial chromosomes had changed little since divergence from a common ancestor [Hayman and Martin, 1969] and certainly much less than the chromosomes of their eutherian counterparts. This finding was later supported by G-banding studies [Rofe and Hayman, 1985]. The advent of molecular cytogenetic techniques has enabled us to more thoroughly investigate the level of karyotypic conservation and delve much deeper into the analysis of chromosome evolution in this group of mammals.

The limited range of diploid numbers was a telling sign of the limited amount of karyotypic variation amongst marsupials. Diploid numbers were bimodal in marsupials, with 2n = 14 and 2n = 22 complements common both in Australidelphia and Ameridelphia [Sharman, 1974; Hayman, 1990]. This led to 2 alternative hypotheses for the ancestral marsupial chromosome number [Sharman, 1973; Hayman and Martin, 1974; Reig et al., 1977; Rofe and Hayman, 1985]. Rofe and Hayman [1985] proposed a 2n = 14 ancestral marsupial karyotype, a chromosome complement common among several Australian families and observed to have changed little between American and Australian species, with fissions giving rise to higher diploid numbers that are seen in many families [Hayman and Martin, 1974]. The alternative hypothesis proposed that the 2n = 22 complement is ancestral and that lower diploid numbers are the result of fusion events [Sharman, 1973; Svartman and Vianna-Morgante, 1998]. The use of a well-supported phylogenetic tree provided further evidence for a 2n = 14 ancestor [Westerman et al., 2010], but the 2 hypotheses could not be distinguished without reference to an outgroup.

Two of the most speciose groups of marsupials have very different rates of karyotypic evolution. The family Dasyuridae is known for its extreme level of karyotypic conservation, with 40 of the 68 known dasyurid species examined to date possessing the conserved 2n = 14 karyotype [Hayman and Martin 1974; Young et al., 1982; Rofe and Hayman, 1985]. In contrast, the family Macropodidae (kangaroos, wallabies) has been subject to major chromosome rearrangements, with diploid numbers ranging from a species with the lowest marsupial diploid number of 2n = 10, 11 to 2n = 24, all of which can be derived from a 2n = 22 macropodid ancestor by simple (mostly Robertsonian) fusions and fissions [Rofe, 1978; Hayman, 1990]. For this reason, their chromosomes have been comprehensively studied, first by G-banding and more recently by new molecular cytogenetic techniques which provide better resolution of the chromosome restructuring events that have occurred to result in the karyotypes of extant macropod species. Chromosome painting has afforded an overview of large-scale rearrangements, and physical gene mapping has been used to resolve breakpoints and identify smaller inversions.

Cross-species chromosome painting on different marsupial species has proven exceptionally valuable for determining which chromosome regions are genetically homologous between closely and distantly related marsupials in order to determine how chromosomes have evolved in marsupials and to reconstruct the ancestral marsupial karyotype. The first such study used chromosome paints derived from 3 Australian species of the order Diprotodontia: the tammar wallaby (Macropus eugenii,2n = 16), the brushtail possum (Trichosurus vulpecula,2n = 22) and the long nosed potoroo (Potorous tridactylus, 2n = 12,XX female: 2n = 13,XY₁Y₂ male). These were compared in mostly reciprocal painting experiments to the fat tailed dunnart (Sminthopsis crassicaudata,2n = 14) of the order Dasyuromorphia, a group with extremely conserved 2n = 14 karyotypes [Rens et al., 1999]. These experiments detected 15 conserved chromosome segments, and the arrangement of these segments in the 2n = 14 ancestral karyotype was reconstructed from G-banding studies. The 2n = 14 karyotype of S. crassicaudata was thought to have remained very similar to that of the ancestral marsupial, separated by only 6 inversions.

To further test the hypothesis of the 2n = 14 ancestral marsupial karyotype, detailed chromosome painting was extended to a diprotodontid species, (the southern hairy-nosed wombat, Lasiorhinus latifrons) and a dasyurid (the striped-face dunnart, S. macroura), both of which have 2n = 14 karyotypes. Chromosome paints derived from the tammar wallaby, also a member of the order Diprotodontia, were used to determine the chromosome homologies between these species. This study provided confirmation of previous G-banding studies that the 2n = 14 karyotype present in different orders of Australian marsupials are virtually identical. Rearrangements detected in 2n = 14 species were the result of inversions. The wombat essentially retains the predicted ancestral arrangement, with the exception of 2 inversions on chromosome 1 [De Leo et al., 1999]. Several additional inversions have taken place on S. macroura chromosomes, a finding similar to that observed for the closely related S. crassicaudata[Rens et al., 1999]. Thus, even distantly related Australian marsupials share a conserved 2n = 14 karyotype, from which other karyotypes could easily be derived.

To determine whether this was also true for American species, and marsupials in general, Rens et al. [2001] used probes derived from the Brazilian gray short-tailed opossum (Monodelphis domestica) to paint onto chromosomes of 3 Australian species used in previous studies (S. crassicaudata, M. eugenii, T. vulpecula) and vice versa. All but one of the conserved segments identified in reciprocal painting experiments on Australian species [Rens et al., 1999] were also conserved in M. domestica. Indeed, 4 autosomes were homologous between M. domestica (1, 2, 5, and 8) and S. crassicaudata(2, 4, 6, and 5). Only 2 rearrangements differentiate the chromosomes of these species; M. domesticachromosomes 3 and 6 are homologous to S. crassicaudata chromosome 1, and chromosomes 4 and 7 are homologous to S. crassicaudata 3. Tellingly, this same rearrangement was observed by G-banding between a 2n = 14 American species Marmosops incanus(gray slender opossum) and M. domestica[Svartman and Vianna-Morgante, 1999], suggesting that the 2n = 14 American and Australian marsupial chromosome complements were homologous, as had been proposed from early G-banding studies of 2n = 14 karyotypes [Rofe and Hayman, 1985].

A better resolution of conserved segments was achieved by using chromosome paints derived from Aepyprymnus rufuscens (rufous bettong), the marsupial with the highest reported diploid chromosome number of 2n = 32. These paints detected 19 conserved chromosome segments when hybridized to 1 American (M.domestica) and 4 Australian species (S. crassicuadata, T. vulpecula, M. eugenii, and P. tridactylus) [Rens et al., 2003]. When these data were combined with G-banding data on many more species, it was possible to determine the chromosome rearrangements that have occurred during marsupial evolution and reconstruct the marsupial ancestral karyotype (fig. 2) [Rens and Ferguson-Smith, 2010].

Examination of the conserved segment organisation across 2n = 14 ameridelphian and austradelphian species showed that the 2n = 14 karyotypes observed are virtually identical, apart from a few intrachromosomal rearrangements [Rens and Ferguson Smith, 2010]. However, the arrangement of these 19 conserved segments differed markedly in 2n = 22 American and Australian species. Therefore, 2n = 22 karyotypes in the 2 superorders [Sharman, 1973] are not the same, so they do not represent an ancestral arrangement. However, this does not rule out the possibility that the marsupial ancestor had a 2n = 22 complement that underwent fusion events early in marsupial evolution to form the common 2n = 14 karyotype, from which unrelated 2n = 22 karyotypes later evolved. The observation of interstitial telomere signals (ITS) in the chromosomes of 2n = 14 American species supported the origin of the large chromosomes by fusion of smaller chromosomes [Svartman and Vianna-Morgante, 1998, discussed further below], although it did not reveal when this occurred.

To further test the alternative hypotheses of the ancestral marsupial karyotype, Westerman et al. [2010] used a well-resolved marsupial phylogenetic tree along with cytogenetic information to show that the conserved 2n = 14 karyotype appears almost unchanged in several American and Australian lineages, so that the most likely chromosome arrangement in the marsupial ancestor was 2n = 14, and higher diploid numbers are derived from fission events. However, this dichotomy cannot be resolved by any technique with out reference to an outgroup.

Chromosome painting has provided an excellent foundation for comparing the homology of marsupial chromosomes and reconstruction of a plesiomorphic karyotype, but to resolve the debate over the ancestral chromosome complement, comparisons to outgroups need to be made. With the exception of the X chromosome [Glas et al., 1999b], marsupial chromosome paints do not hybridize to eutherian chromosomes. Fortunately, the sequencing of several marsupial genomes and the availability of bacterial artificial chromosomes (BAC) clones for these species have enabled gene maps to be constructed from orthologous genes in divergent species such as chicken and human.

The opossum genome was deeply sequenced [Mikkelsen et al., 2007], and 97% of this sequence was anchored by mapping the ends of sequence scaffolds to opossum metaphase chromosomes using fluorescence in situ hybridization (FISH) [Duke et al., 2007]. The Tasmanian devil (Sarcophilus harrissi) has been the subject of 2 independent sequencing efforts, both of which employed next-generation sequencing technology [Miller et al., 2011; Murchison et al., 2012]. One of these projects sequenced flow-sorted devil chromosomes to assign sequences to chromosomes, but sequence arrangement on devil chromosomes was not determined [Murchison et al., 2012]. The tammar wallaby genome has only been lightly sequenced, and more than 300,000 sequence scaffolds remain unanchored [Renfree et al., 2011].

By extrapolating data from chromosome painting and the anchored opossum genome assembly, we can predict the genic content of chromosomes of other marsupials. Access to sequence data and BAC libraries for the tammar wallaby and devil has enabled cytogenetic gene maps to be efficiently generated, providing finer resolution of the homology and rearrangements between marsupial chromosomes as well as permitting comparisons to outgroups such as chicken and human to be made. Such comparisons are critical for a definitive reconstruction of the ancestral marsupial karyotype.

A map of tammar wallaby chromosome 5 was the first dense cytogenetic gene map constructed for any marsupial and a trial for an efficient approach to assigning genome sequence to chromosomes. A block of genes conserved between opossum chromosome 7 and the human X chromosome was identified, and 52 of the 58 genes occurring in this block were mapped to the short arm of tammar wallaby chromosome 5 [Deakin et al., 2008]. Rather than mapping every gene within an opossum-human conserved block, only genes at either end of all other opossum-human evolutionary conserved blocks were mapped, and the genic content of these blocks inferred from the opossum genome assembly. This resulted in a cytogenetic map of 141 genes across 15 opossum-human conserved blocks and the virtual assignment of 2,320 protein-coding genes [Deakin et al., 2008]. Three inversions that had not been detected by chromosome painting were revealed by comparison of the tammar wallaby map with the opossum genome assembly.

Chromosome painting determined that tammar wallaby chromosome 5, consisting of conserved segments C11 and C12, was homologous to most of opossum chromosome 4 and a small region on chromosome 7, respectively [Rens et al., 2003]. Gene mapping more accurately determined the boundaries of these conserved segments. In the tammar wallaby, C12 did not encompass the entire short arm and proximal region of the long arm of chromosome 5, but was instead restricted to the pericentric region, and part of C11 was located on the distal region of the short arm [Deakin et al., 2008]. The location of C12 on opossum chromosome 7 was inferred from known chromosome homologies rather than direct hybridization, due to difficulties experienced in hybridizing the C12 paint to opossum chromosomes [Rens et al., 2003]. Comparative mapping data has allowed this segment to be accurately assigned to the pericentric region of chromosome 7.

The same mapping strategy was used to assign blocks conserved between opossum and human to all tammar wallaby autosomes [Mohammadi et al., 2009; Deakin, 2010], and a virtual map of all chromosomes has been constructed [Wang et al., 2011]. A similar approach has been used to generate a map of the devil genome, where 105 genes corresponding to the ends of opossum-wallaby conserved blocks were mapped to devil chromosomes [Deakin et al., 2012]. Comparisons of gene arrangements between devil, tammar wallaby and opossum revealed substantial chromosome rearrangement between these species, making it necessary for more genes to be mapped before a virtual map of the genome can be constructed.

Of particular interest is the gene order on chromosome 1 of the Tasmanian devil, since it is this chromosome that has been extensively rearranged in the transmissible Devil Facial Tumour Disease [Deakin et al., 2012]. Confusingly, devil chromosome 1 has been mislabelled as chromosome 2 in some of the literature on this interesting tumour [e.g. Pearse and Swift, 2006], contradicting the long-established standard dasyurid karyotype [Martin, 1967]. The normal devil chromosome 1, like the ancestral chromosome 1 from which it evolved, consists of conserved segments C1 to C6, but these blocks are highly rearranged in devil. The most conserved chromosome corresponded to segments C11 and C12 on devil chromosome 3.

Acceptance of a 2n = 14 ancestral marsupial karyotype predicts that segments C10, C11 and C12 would have formed chromosome 3, as we observe in the devil. The alternative hypothesis of a 2n = 22 ancestor distributes these conserved segments across 2 chromosomes [Svartman and Vianna-Morgante, 1998]. We observe that genes spanning parts of all 3 segments in a representative outgroup species, the chicken (Gallus gallus), were all found together on chicken chromosome 1. This suggests that they were part of the same chromosome in the ancestor of all mammals, which was subsequently subject to various fission and fusion events in some lineages but remains as a single chromosomes in 2n = 14 species. This refutes the hypothesis that the marsupial ancestor had a 2n = 22 karyotype [Deakin et al., 2008]. Thus all evidence, from extensive G-banding studies to cross-species chromosome painting, phylogenetics and gene mapping, support the original hypothesis of Hayman and Martin [1974] for an marsupial ancestor with a 2n = 14 chromosome complement, which has been retained in 6 of the 7 marsupial orders.

The highly conserved karyotype among species of the family Dasyuridae, including the devil, means that by anchoring genome sequence to chromosomes in this species [Deakin et al., 2012], we have essentially determined the gene order on all dasyurid chromosomes. It is important to note that the devil genome sequence assembly, which assigned sequence to chromosomes by sequencing of flow-sorted chromosomes [Murchison et al., 2012], perpetuates the mislabelling of devil chromosomes 1 and 2 noted above. Of course, there could be small-scale rearrangements that have escaped detection by G-banding that could differentiate members of this family, but most genes would be predicted to be present in the same order. Once further gene mapping is completed for this species, enabling the generation of a virtual map of the genome, it will be possible to predict the gene order in the ancestral 2n = 14 plesiomorphic karyotype, taking into account the inversions differentiating the standard dasyurid karyotype from that of the marsupial ancestor.

Svartman and Vianna-Morgante [1998] explored the alternative hypothesis of a 2n = 22 marsupial ancestor. They argued that the 2n = 14 complement is derived from fusion of chromosomes from a 2n = 22 ancestor, using the presence of ITS in 2n = 14 and 2n = 18 species, and the absence of ITS in 2n = 22 species as evidence of such fusion events. They claim, for example, that ITS signals detected at the pericentric region of M. domestica chromosome 1 is evidence of a fusion between 2 and 10 of a putative 2n = 22 plesiomorphic karyotype. However, several other studies detecting ITS in American and Australian marsupial species demonstrated the co-localisation of these ITS with constitutive heterochromatin, suggesting that these ITS represent a component of satellite DNA and did not necessarily imply a past fusion event [Pagnozzi et al., 2000, 2002; Metcalfe et al., 2007]. However, Svartman [2009] claimed that, at least for M. incanus, the ITS signals are not within the boundaries of the pericentric heterochromatin, leaving it open for debate as to whether or not ITS can be used as evidence of past fusion events in marsupials.

Mapping ITS onto marsupial chromosome homology maps (fig. 2) makes clear that many ITS do not correspond to past fusion events. For instance, ITS detected in the pericentric region of 2 Sminthopsis species are unlikely to be evidence of fusion of chromosomes derived from a 2n = 22 complement. ITS signals are detected on chromosomes 1, 2, 3, and 6 in these species. If the marsupial ancestor was 2n = 22, chromosome 1 in these species would be the result of a fusion between a chromosome containing conserved segments C1, C2 and C3 with a chromosome consisting of segments C4–C6. However, chromosome painting indicates that this chromosome has been subject to at least 2 inversions [De Leo et al., 1999]. Thus, telomeric remnants of a fusion event would not be expected at the centromere on this chromosome in Sminthopsis species. Similarly, ITS on chromosome 6 are not evidence of a past fusion event, as this chromosome corresponds to a single chromosome in the predicted 2n = 22 ancestral karyotype [Metcalfe et al., 2007].

Furthermore, ITS were detected at pericentric locations on all chromosomes of tammar wallaby and the slender mouse opossum (Marmosops parvidens) [Pagnozzi et al., 2002], which were certainly not all formed by a fusion event. Indeed, tammar wallaby chromosome 4 consisting of conserved segments C4, C5 and C6, which form a single chromosome in the predicted ancestral marsupial 2n = 22 karyotype, has a strong ITS signal [Metcalfe et al., 2007]. ITS signals, therefore, do not provide conclusive evidence of past fusion events. Rather, some ITS on chromosomes not formed by centric fusions actually appear to correspond to remnants of telomeres from inversion events. This accounts for the ITS signals in 2n = 14 Sminthopsisspecies and the tammar wallaby chromosomes 4 and 5. Likewise, the ITS on M. domestica chromosome 1 are more likely to be the remnants of an inversion of the short arm containing segments C7 and C8 than a centric fusion.

Although marsupials are renowned for karyotypic conservation, some groups have been particularly prone to karyotypic change. This is exemplified by members of the suborder Macropodiformes, which consists of 3 extant families (Macropodidae, Potoroidae, Hypsiprymnodontidae). The Macropodiformes have an ancestral 2n = 22 complement, which differs from the 2n = 22 karyotype of ameridelphian species. This was initially proposed from G-banding studies [Rofe, 1978] and has since been verified using cross-species chromosome painting on members of the Macropodidae and Potoroidae families [Glas et al., 1999a; O’Neill et al., 1999; Rens et al., 1999, 2003]. This macropodiformes ancestral chromosome karyotype can be derived from the 2n = 14 ancestral marsupial complement by 5 fissions, 1 fusion and several inversions and centric shifts (fig. 3) [Rofe, 1978; Hayman, 1990; Eldridge and Close, 1993; Rens et al., 1999]. The macropodid ancestral 2n = 22 karyotype is observed to have been retained in some species in all 3 Macropodiformes families. Examples of various macropodiform karyotypes are shown in figure 4.

The family Macropodidae (kangaroos and wallabies) have experienced extensive karyotypic change since their divergence approximately 23 mya. As a result, their chromosome evolution has been intensively studied by G-banding [Rofe, 1978; Eldridge et al., 1989, 1990, 1991, 1992a, b; Eldridge and Pearson, 1997; Alsop et al., 2005], chromosome painting [Toder et al., 1997; Glas et al., 1999a; O’Neill et al., 1999] and FISH detection of ITS [Metcalfe et al., 1997, 1998, 2002, 2004, 2007]. It is clear that the 18 autosomal conserved segments identified by Rens et al. [2003] have undergone considerable reshuffling in this family. Even species within the same genus possessing the same diploid number and morphologically similar chromosomes do not have the same arrangement of the 18 conserved segments, and it appears they have undergone independent Robertsonian fusions of different ancestral chromosomes. This is illustrated by 2 species belonging the genus Macropus with 2n = 16 karyotypes, the tammar wallaby and the gray kangaroo (M. giganteus). Three of their chromosomes have been formed by the fusion of different conserved segments. For example, M. eugenii chromosome 1 was formed by the fusion of ancestral macropodiform (AnMac) chromosome 1 (C8, C9) to AnMac chromosome 10 (C1), whereas AnMac chromosome 1 fused to AnMac chromosome 8 (C2,C3) to form M. giganteus chromosome 1 (fig. 4).

The swamp wallaby (Macropus/Wallabia bicolor) possesses one of the most derived marsupial karyotypes, having only 4 very large autosomes and an XX female, XY₁Y₂ male sex chromosome system that resulted from an X-autosome fusion (see section on sex chromosomes). Chromosome painting demonstrated that chromosome 1 in this species has been formed by a centric fusion of AnMac chromosomes 6 and 10 and the tandem fusion of AnMac chromosome 4 and 10 [Toder et al., 1997]. The distribution of telomeric sequences on M. bicolor chromosome 1 indicate that in addition to these fusion events, an inversion event has probably taken place involving conserved segment C4, as there are 3 sets of ITS rather than the expected 2 [Metcalfe et al., 1998] (fig. 4). A centric fusion has given rise to M. bicolor chromosome 2 and the remaining 2 autosomes are equivalent to tammar wallaby chromosomes 3 and 5.

Macropodid karyotypes also afford an opportunity to assess the extent of convergence and reversal in marsupial karyotypic evolution. Comparisons of the arrangement of 18 autosomal conserved chromosome segments across macropodids reveal several examples of convergence (independent origin of the same segment associations in different species) and reversal (a secondarily derived arrangement resembling that of the ancestor). Six AnMac chromosomes (1, 5, 6, 8, 9, and 10) have been frequently involved in chromosome restructuring throughout macropodid evolution (fig. 4) [Hayman, 1990; Eldridge and Close, 1993; Eldridge and Johnston, 1993; Bulazel et al., 2007]. The fusion of AnMac chromosome 9 (C18) to AnMac chromosome 8 (C2 and C3) occurred independently in P. tridactylus and Dendrologous matschiei (Matschiei’s tree kangaroo). The fusion of AnMac chromosomes 3 and 6 in these species provide an example of reversal, as these chromosomes have reverted to the 2n = 14 ancestral association of segments C10, C11 and C12 on a single chromosome [Rens et al., 2003].

Comparisons made more widely across marsupials reveal examples of convergence. The fission of segments C8 /C9 and C13/14 in the ameridelphid species Didelphis marsupials, as well as in the australidelphid species T. vulpecula, are examples of convergence, as they occurred independently in these species [Rens et al., 2003]. Chromosome painting and gene mapping indicate that the fission of segments C10, C11 and C12 arose from different breakpoints in American and Australian species (fig. 2), providing another example of convergence [Rens et al., 2003; Deakin et al., 2008].

The X chromosome in placental mammals is unusually well conserved, perhaps because translocations and rearrangements interrupt an X inactivation mechanism that is coordinately controlled over the whole X chromosome. However, in contrast to the generally high conservation of marsupial autosomes, marsupial X chromosomes vary significantly in morphology, size and banding pattern, even between species with the conserved 2n = 14 karyotype [Rofe and Hayman, 1985]. Moreover, translocations or fusions between autosomes and the sex chromosomes are widespread amongst marsupials [Toder et al., 1997]. For instance, hybridization of tammar wallaby chromosome paints on the swamp wallaby (M. bicolor, 2n = 10,XX female, 2n = 11,XY₁Y₂ male) chromosomes revealed that the short arm of the swamp wallaby X chromosome was homologous to the tammar wallaby X, whereas the long arm shared homology with tammar wallaby chromosomes 2 and 7 (fig. 4), as does Y₂, representing the autosome to which the X was fused. The tammar wallaby Y chromosome hybridized predominantly to Y₁, corresponding to the original Y chromosome. The tammar wallaby Xp paint, corresponding to a region shared between the tammar wallaby X and Y chromosomes was also detected on both X and Y₂. P. tridactylus also has an X chromosome formed by the fusion of segments C5 and C6 to C19, resulting in 2n = 12,XX females and 2n = 13,XY₁Y₂ males (fig. 4). The X chromosome of all macropodids includes the nucleolus organizer (NOR) and nucleolus-associated heterochromatin. This region lies on chromosome 5p in the ancestral 2n = 14 karyotype, but was translocated to the X and Y in an ancestral macropodid; the NOR has since degenerated on the Y in all but one species and has amplified and elaborated on the X in others.

The X chromosome in most marsupials is represented by a single conserved chromosome segment (C19) [Rens et al., 2003]. Gene mapping and chromosome painting of the tammar wallaby X chromosome onto human chromosomes demonstrated homology between the wallaby X and two-thirds of the human X chromosome [Glas, 1999b]. The remainder of the human X chromosome maps to chromosome 5 in tammar wallaby and chromosome 3 in dasyurids, and to chromosomes 4 and 7 in the opossum (M. domestica) and also forms a separate evolutionary block in chicken. Thus, the marsupial X represents the original therian X chromosome to which an autosomal region was added early in the eutherian lineage.

The X chromosome of eutherian mammals is highly conserved in gene content, and usually in gene order, even among the most distantly related species [Delgado et al., 2009]. This has been ascribed to the dosage compensation mechanism that is coordinated over the whole X [Ohno, 1967] by the XIST gene that transcribes a key non-coding RNA involved in X chromosome inactivation in eutherians. Surprisingly, comparisons of the order of genes on the X chromosomes of tammar wallaby, devil and opossum revealed extensive rearrangement [Deakin et al., 2008, 2012], which would account for the variation observed in G-banding patterns. This may relate to differences in the control of X chromosome inactivation between marsupial and placental mammals, particularly the absence of XIST in marsupials. Genes that flank the XIST locus in eutherians map to opposite ends of the X chromosome in 2 opossum species, M. domestica [Davidow et al., 2007; Hore et al., 2007; Shevchenko et al., 2007] and Didelphis virginiana[Shevchenko et al., 2007], and tammar wallaby [Deakin et al., 2008]. This correlates with the locus-specific and partial X inactivation observed for genes on the marsupial X chromosome [Deakin et al., 2009; Al Nadaf et al., 2010]. Perhaps the acquisition of XIST and the evolution of a more tightly controlled inactivation mechanism than the partial inactivation observed in marsupials [Deakin et al., 2009; Al Nadaf et al., 2010] selected against rearrangements that might disrupt a whole X inactivation mechanism.

The basic marsupial Y chromosome is very small, in some species no more than 12 Mb [Toder et al., 2000 1997]. It represents a degraded relic of the original therian proto-Y chromosome, which was originally homologous to the original proto-X chromosome. To date, 12 genes have been mapped on the tammar wallaby Y chromosome, 11 of which have copies on the X chromosome from which they originated [Sankovic et al., 2006; Murtagh et al., 2012]. Orthologues of 10 of the tammar wallaby Y-borne gene sequences have been identified in the devil testis transcriptome [Murtagh et al., 2012]. This low tally of genes on the marsupial Y chromosome, nevertheless, far exceeds the 4 genes retained on the human Y from the original proto-XY, implying that the marsupial Y has been subject to a slower rate of degradation than the eutherian Y.

Sex determination (or at least testis determination) in marsupials is thought to be via a male-dominant gene because XXY animals have testes and XO animals lack them. However, some sexual dimorphisms are evidently a function of the dosage of X chromosomes (or the presence of a paternal X) since XXY animals possess a pouch and mammary glands, and XO animals have a scrotum. Marsupials were crucial in the identification of the mammal sex determining gene; the first candidate ZFY was first eliminated by the finding that this gene is autosomal in marsupials [Sinclair et al., 1988]. This led to the discovery, in humans and mice, of the sex determining gene SRY, which is ubiquitous in eutherian mammals and has an orthologue on the tiny marsupial Y; however, marsupials also have a Y-borne copy of the human sex reversing gene ATRY, which presented a rival candidate.

The application of molecular cytogenetics to understanding the evolution of marsupial chromosomes has greatly advanced this field of research. The ancestral marsupial karyotype has been confidently reconstructed.

There remain some important gaps, which can be filled in by a more thorough investigation of chromosome homologies and rearrangements in different taxa. For example, chromosome painting is yet to be performed on any members of the family Peramelidae (bandicoots), which predominantly have a 2n = 14 karyotype, and Petauroidea species with 2n = 22. Moreover, chromosome painting on at least one representative 2n = 14 and 2n = 22 American species will prove useful, as the conserved segment composition of chromosomes for all except M. domestica has been predicted based on homologies observed between G-banded chromosomes.

Although chromosome painting is valuable for detecting global regions of homology, small-scale rearrangements, particularly within a conserved segment, are undetectable. From the comparisons of gene order for just 3 species, it is evident that inversions, large and small, have featured heavily in the evolution of marsupial chromosomes. The generation of gene maps for additional marsupial species would be valuable. Fortunately, BAC libraries are currently available for 2 additional species, Isoodon macroura (northern brown bandicoot, 2n = 14) and D. virginiana(2n = 22), making it possible for gene maps to be constructed for these species. These 2 species belong to lineages that are important for accurate predictions of gene order on ancestral marsupial chromosomes and will provide higher resolution of the rearrangements that have shaped marsupial chromosomes.

Monotreme chromosomes have captivated cytogeneticists since their first controversial descriptions in the 1940s [reviewed in Murtagh and Sharman, 2009]. The presence of several unpaired chromosomes in mitosis and multivalent chains in male meiosis created an aura of mystery surrounding monotreme chromosomes. Attempts to characterize monotreme chromosomes by various banding patterns failed to conclusively distinguish the many small chromosomes [Wrigley and Graves, 1988a, b; Murtagh and Sharman, 2009]. It was only with the aid of molecular cytogenetics that the puzzle of monotreme chromosomes was finally solved.

In contrast to the few large and readily identifiable chromosomes characteristic of marsupial karyotypes, the monotreme genome consists of many smaller chromosomes. There are 6 large, easily distinguished chromosome pairs and many small chromosomes that are hard to differentiate. All platypus and echidna autosomes are either metacentric, submetacentric or subtelocentric; none of them are acrocentric.

Platypus (Ornithorhynchus anatinus) somatic cells contain 26 pairs of chromosomes (2n = 52) [Warren et al., 2008, supp data]; the short-beaked echidna (Tachyglossus aculeatus) somatic nuclei contain 63 chromosomes in the male and 64 chromosomes in the female. The platypus genome has been sized at only 1.92 Gb DNA (measured by flow cytometry), a size more typical of reptile than mammal genomes. The autosomes are ordered according to their size measured by flow cytometry, which has the advantage of being a non-arbitrary measuring method important for comparative genomics. Sex chromosomes are complex and are discussed in detail below.

In piecing together the puzzle of monotreme karyotypes, flow-sorted platypus chromosome paints were used to identify autosome pairs [Rens et al., 2004]. However, not all chromosomes could be physically separated into individual chromosomes, resulting in 4 paints that hybridized to more than one chromosome (e.g. chromosomes 9 and 12). To aid in the unambiguous identification of each platypus chromosome, a set of platypus anchor BAC clones (one for each chromosome) was developed [McMillan et al., 2007]. The combination of these resources has been a major step forward for monotreme cytogenetics, enabling the unambiguous localisation of genes to the smaller platypus chromosomes [Edwards et al., 2007; Grafodatskaya et al., 2007].

Unfortunately, it is difficult to establish the monotreme ancestral karyotype because there are only 2 extant families of monotremes, platypus and several closely related echidna species. Being basal to the mammalian radiation, there are no mammalian outgroup species. Chromosome rearrangements and homologies have been established between the platypus and short-beaked echidna [Rens et al., 2007]. Platypus chromosomes 1, 4, 5, 9, 11, 14, 16, and 19 are homologous to echidna chromosomes 1, 4, 3, 10, 11, 14, 19, and 22, respectively. All other chromosomes are derived from new combinations by centric fusion or fission in either species, showing that the monotreme karyotypes have not been stable after the platypus-echidna divergence around 30 mya [Phillips et al., 2009], but have continued to evolve, largely by Robertsonian fissions and fusions.

How do monotreme karyotypes relate to those of placental mammals? Ruiz-Herrera et al. [2012] detected 7 syntenic segmental associations characteristic for the placental ancestral karyotype in Ultracontigs of the platypus genome database, which suggests that these placental ancestral associations are ancestral for all mammalian species. The monotreme karyotypes are highly rearranged when compared to the placental ancestral karyotype, meaning that the monotreme karyotypes have their own dynamics of chromosome evolution. The platypus assembly is still too incomplete to define the chromosome rearrangements that occurred during the evolution of the monotreme karyotypes from the mammalian ancestral karyotype.

Sequencing of the platypus genome revealed a relatively high GC content of 45.5% compared to 41% for most eutherian species [Warren et al., 2008]. It is uncertain whether a high GC content is characteristic of all monotremes because the echidna genome is not fully sequenced yet. However, the platypus and echidna flow karyotypes provide information on relative GC contents of the individual chromosomes of both species [Rens et al., 2004, 2007]. In both species, most of the chromosome peaks do not deviate into AT or GC richer regions, implying that most monotreme chromosomes have similar GC content. In platypus, chromosomes 6, 11, 12, 15, X₂, X₄, and Y₄ have a somewhat higher GC content, whereas in echidna the larger chromosome 1, 2 and 4 have a relative lower GC content, and chromosome 3, 6 and X₅ have a relative higher GC content. The higher GC content for platypus chromosome 6 and echidna chromosome 3, 6 and X₅ can be explained by their inclusion of the NOR. (Note that although both platypus and echidna chromosome 6 are NOR-bearing, chromosome painting reveals that they are non-homologous).

The difference between the relative GC content of platypus and echidna chromosome 1, which are homologous, can be explained by an apparently AT-rich region that is identified on echidna chromosome 1 and 2 [Rens et al., 2007] and contains blocks of repetitive sequence [Kirby et al., 2007]. Similarly, the difference between platypus and echidna chromosome 4, which are also homologous, can be explained by another apparently AT-rich region on echidna chromosome 4 [Rens et al., 2007]. These examples show that homologous chromosomes, as determined by cross-species chromosome painting, may be different at a sequence level, depending on acquisition and amplification of repetitive sequences.

Centromeres are a crucial functional feature of chromosomes that appear as constrictions on metaphase chromosomes. In contrast to chromosomes of therian species, the centromeres of platypus chromosomes are not characterised by satellite DNA [Alkan et al., 2011].

Platypus chromosomes show nearly equal amounts of the retrotransposon LINE2 elements, but no LINE1 elements [Warren et al., 2008], in contrast to eutherian species in which the X chromosome contains a relatively large fraction of LINE1 elements. MON1 (SINE) elements are also abundant on platypus chromosomes [Kirby et al., 2007]. Young MON1 and LINE2 insertions are over-represented on the platypus Y chromosomes, a feature also seen in other species [Kortschak et al., 2009]. Platypus chromosomes do resemble reptilian chromosomes with respect to the low microsatellite content, with a high fraction of AT-rich microsatellites. The most common motifs in these microsatellites are also highly similar to the motifs used in lizard [Warren et al., 2008].

The evolution of the therian sex chromosome system has been studied in detail by researchers with interests ranging from sex determination and Y chromosome evolution to the mechanism of X inactivation. However, they are studying the complexity and functional implications of just 1 set of sex chromosomes among many different ones in vertebrates, and it is clear that the therian XY system evolved recently [Graves, 2008; Graves and Peichel, 2010]. The phylogenetic position of monotremes between mammals and reptiles means that sex chromosomes of monotremes can fill an important gap in our understanding of how the therian XY system evolved.

Monotremes have multiple sex chromosomes with different sizes and genetic contents. The raison d’être and evolutionary history of the monotreme sex chromosome complement is still very puzzling. A first step towards its understanding is the characterisation of the sex chromosomes and the comparison with genomes of other species.

The male platypus sex chromosome system consists of 5 different X and 5 different Y chromosomes. The female platypus has the 5 X chromosomes in pairs and contains no Y-specific sequences. The segregation of the 5 X chromosomes from the 5 Y chromosomes during meiosis into separate X- and Y-bearing sperm must occur via a polarised meiosis and be very accurate, since painting revealed no sperm or spermatogonia with mixtures of X and Y chromosomes [Grützner et al., 2004].

All 10 platypus sex chromosomes were unambiguously identified by chromosome sorting followed by chromosome painting. The sizes of these chromosomes differ considerably. X₁ is the largest sex chromosome with a DNA content of 106 Mb (measured by flow cytometry), which is between the DNA content of platypus chromosome 4 and 5 (123 and 92 Mb, respectively). Platypus X₄ is the smallest X chromosome (19 Mb). Y₁ is the largest Y chromosome with a DNA content of 48 Mb and Y₅ is the smallest Y-chromosome (DNA content too small to be measured by flow cytometry). The order of the sex chromosomes in the karyotype is not determined by size (as for autosomes), but according to their position in the male meiotic chain (fig. 5), which was ascertained by chromosome painting [Grützner et al., 2004]. The sex chromosome system evidently originated by serial translocation of a single XY pair with autosomes [Grützner et al., 2003, 2004; Rens et al., 2004].

Each sex chromosome has a region that pairs with its neighbour during meiosis. Thus a terminal region of X₁ pairs with one end of the Y₁, and the other end of Y₁ pairs with X₂; the X₂ pairs with Y₂ etc. The pairing regions occupy substantial pseudoautosomal regions at the tips of sex chromosomes, but the Y-specific regions are heterochromatic and much shorter than the comparable regions of the X, implying that the Y-chromosomes have degenerated during platypus evolution. For instance, Y₁ has lost around 50 Mb of DNA, and Y₅ has lost around 65 Mb of DNA. In comparison, chromosome size difference between human X and Y is around 100 Mb.

Because of the complexity of the platypus sex chromosome system, one may expect that the echidna has either a ‘normal’ one-X-one-Y system or exactly the same system as the platypus. A comparison performed by cross-species chromosome painting revealed that neither is the case [Rens et al., 2007]. The short-beaked echidna also has a multiple sex chromosome system, but the male has only 9 sex chromosomes: 5 X and 4 Y chromosomes. The missing Y chromosome is the tiny Y₅; however, chromosome painting with platypus Y₅ reveals that this chromosome has fused with Y₃. The female has the 5 X chromosomes in pairs. Surprisingly, not all of the sex chromosomes of platypus and echidna are homologous. Platypus and echidna have X₁, X₂ and X₃ in common, but platypus X₄ is homologous to an autosome (chromosome 27) in echidna, and echidna X₅ corresponds to a platypus autosome (chromosome 12p). The best explanation for this difference is that the original sex pair fused with an autosome in a common ancestor, producing a chain of 4. Two other fusions followed, producing a chain of 8, before platypus and echidna diverged 30 mya. Then an additional fusion occurred with a different autosome in the 2 lineages. The platypus-echidna difference shows that the translocation chain still evolved after platypus and echidna diverged around 30 mya and supports the hypothesis of consecutive translocations.

The homology of platypus X₄ to an echidna autosome is significant for the location of the MHC cluster in platypus. MHC class I and II genes are found on platypus X₃ and Y₃, and MHC class III is located on platypus Y₄ and X₅ (fig. 5) [Dohm et al., 2007]. Thus, these 2 clusters were separated by translocation with X₄. However, in echidna the MHC cluster is on neighbouring sex chromosomes because platypus X₄ is homologous to an echidna autosome.

It would be an advantage to know which chromosomes of the sauropsid ancestral karyotype became caught up in the sex chromosome system. The ancestral monotreme sex chromosomes could potentially be deduced by comparing homologous regions with an extant reptile outgroup.

The first important and surprising finding was that the orthologue of the chicken DMRT1 gene, a strong candidate for the bird sex determining gene, is located on X₅, the last X chromosome in the platypus sex chromosome chain [Grützner et al., 2004]. This prompted the search for other chicken Z orthologues. Many chicken Z genes can be found on the platypus sex chromosomes. Platypus X₅ has a large coverage of chicken Z genes, but X₁, X₂ and X₃ also contain chicken Z genes [Rens et al., 2007; Veyrunes et al., 2008]. No chicken Z genes have so far been mapped to X₄, perhaps because platypus X₄ is homologous to an autosome in echidna, so it was a late recruit to the chain.

Do the platypus sex chromosomes relate to the eutherian sex chromosomes? The answer is no. The platypus multiple XY system does not correspond to the therian XY- system. Early mapping using heterologous cDNA probes suggested that platypus X₁ is homologous to the eutherian X [Wrigley and Graves, 1988a]. However, by investigating the content of platypus X₁ and by comparative mapping of eutherian X orthologues, it was discovered that X₁ corresponds to eutherian autosome regions [Rens et al., 2007], and new gene mapping methods established that orthologues of genes within the conserved region of the X shared by eutherians and marsupials are all located on platypus chromosome 6 [Waters et al., 2005; Veyrunes et al., 2008]. The location of some chicken Z genes on platypus autosomes and the location of some chicken autosome genes on platypus Z chromosomes probably resulted from serial reciprocal translocations between Z-containing members of the chain and newly recruited autosomes [Rens et al., 2007; Veyrunes et al., 2008]. This is consistent with serial translocations of ancestral autosomes into the chain.

Thus, the sex chromosomes of platypus have no homology to the XY system of therian mammals, but very substantial homology with the ZW system of birds. This is a very significant finding for the study of eutherian sex chromosome evolution, implying that the eutherian X is much younger (166 my) than had been thought, and so must have evolved in a faster rate [Graves, 2006].

The non-homology of the platypus and eutherian sex chromosomes, and the position of SOX3(from which SRY evolved) on platypus chromosome 6, indicates that monotreme sex determination is not initiated by SRY, but by a different (unknown) sex determining gene(s). The most obvious candidate is DMRT1, now known to be sex determining in birds: the expectation was that DMRT1 became mutated or amplified on a Y-chromosome in order to initiate male development, in much the same way as in Medaka [Kondo et al., 2009]. However, DMRT1 lies on X₅, so it is present in 2 copies in females and a single copy in males, quite the reverse of its dosage in birds [El-Mogharbel et al., 2007]. Thus, it is thought that some other gene, probably located on Y₅, is responsible for testis determination.

It is tempting to propose that the chicken Z chromosome is the ancestral sauropsid sex chromosome, given the homology between bird and monotreme sex chromosomes, and the more recent finding that a gekko shares an almost identical Z chromosome [Kawai et al., 2009], although this seems not to be the ancestral gekko sex pair. In a large number of representative sauropsid species, the chromosome that became the Z in birds is conserved as one entity, although it is not a sex chromosome in all these studied species [Pokorna et al., 2011]. The conservation is a reflection of the high degree of general chromosome conservation in sauropsids [Pokorna et al., submitted]. Thus, the homology between the bird ZW and platypus sex chromosome chain might be an example of homoplasy that this autosome evolved into a sex chromosome in platypus, gekko and chicken. It is an interesting thought, however, that this autosome may have been particularly suitable to become a sex chromosome [O’Meally et al., 2012].

Vertebrate sex determination is regulated by a complex genetic network comprising several genes; the cascade is initiated by the SRY gene. There is no SRY in monotremes, and the gene from which it evolved (SOX3) is autosomal [Wallis et al., 2007]. Each of the genes in the therian testis determining pathway may be a candidate(s) for the primary sex determining gene(s) in monotremes. A first indication will be the location of this gene(s) on just one of the sex chromosomes (either a dose-dependent gene on the X, or a male-dominant gene on the Y). Several orthologues of genes active early in eutherian sex determination/differentiation have been mapped by FISH or PCR mapping using sorted chromosomes as templates.

Possible candidates are the genes WT1, SF1, LHX1, LHX2, FGF9, ATRX, and DAX1, which are active in eutherian male sex determination/ differentiation. SOX9 has a critical conserved role in vertebrate sex determination, since its activation and maintenance is needed for male development. However, SOX9 is located on platypus chromosome 15, so it is not the primary sex determining gene [Wallis et al., 2007]. SF1(NR5A1) has a crucial role early in eutherian sex determination by aiding SRY to bind to the TESCO enhancer of SOX9; however, its autosome location excludes such a role in platypus. SOX3 is also sex-reversing when expressed ectopically in somatic cells of testis [Sutton et al., 2011], but its location on platypus autosome 6 excludes it from this primary role [Wallis et al., 2007]. GATA4 is located on Y₁ and X₂, so it is unlikely to function as the primary sex determining gene [Grafodatskaya et al., 2007].

Only 2 genes active in female sex determination have been mapped so far, WNT4 and RSPO1. These 2 genes are located on platypus chromosome 5 and 16, positions that make a primary sex determining role unlikely [Grafodatskaya et al., 2007].

Recently, homologues of human CRSP7 were localised on platypus X₁ and Y₅. The X₁ homologue (CRSPX) and the Y₅(CRSPY) homologue are diverged with different expression levels in adult testis. However, neither activated mouse Sox9 in a reporter assay [Tsend-Ayush et al., 2012]. The hunt for the monotreme sex determining gene continues focusing on the tiny Y₅ of platypus, which is the smallest and most degraded Y, so it is probably the oldest and most original. This chromosome was first dismissed as a candidate because it is absent in echidna; however, it was discovered to have been translocated onto Y₃ in this species [Rens et al., 2007].

In eutherians, it is likely that the acquisition of SRY initiated the degeneration of the eutherian Y chromosome. Specialisation of adjacent genes to perform male-advantage functions selected for restriction of recombination at this site. The simplest assumption is that a primary sex determinant is localised on just one Y chromosome, rather than multiple factors being spread between all the Y chromosomes in the monotreme sex chromosome system. However, male-specific (e.g. spermatogenesis) genes other than the primary sex determinant may be located on the differential regions of the Y chromosomes and were also initiators of restriction of recombination at their sites. This led to degeneration of all Y chromosomes so that most genes on the X chromosomes (amounting to about 15% of the genome) lost their active partner and are therefore present in different dosages in XY males and XX females.

The above discourse shows that the monotreme genome has enriched our knowledge on mammalian genome evolution. Monotreme genomes will no doubt continue to be important for investigations into the appearance and evolution of specific mammalian traits, such as lactation, essential for an understanding of mammalian biology and evolution.

At present, the patchy assignment of the platypus genome assembly to chromosomes makes it difficult to determine the ancestral karyotype of all mammals. The genome of the short-beaked echidna (T. aculeatus) is currently being sequenced, and it would be advantageous for genome evolution studies to anchor this sequence to chromosomes, which will enable the demarcation of rearrangements between platypus and echidna to be determined. It is hoped that future research will make it possible to compare the platypus genome arrangement to that of outgroup species such as chicken and anole lizard.

Our understanding of marsupial and monotreme chromosomes and their evolution has greatly advanced due largely to the molecular cytogenetic techniques and the availability of genome sequence data for representative species. Comparisons of gene arrangement between all 3 major mammalian lineages and other amniotes have provided a new perspective on the evolution of mammalian chromosomes, particularly the sex chromosomes. The inclusion of these species in comparative studies has and will continue to greatly impact on our understanding of the evolutionary events that have shaped mammalian genomes.

J.E.D. is supported by an Australian Research Council Future Fellowship.