Evolution of placentation in cattle and antelopes

Abstract Bovids have enjoyed great evolutionary success as evidenced by the large number of extant species. Several important domestic animals are from this family. They derive from both subfamilies: cattle and their kin belong to Bovinae and sheep and goats to Antilopinae. The premise of this review, therefore, is that evolution of reproduction and placentation is best understood in a context that includes antelope-like bovines and antelopes. Many key features of placentation, including hormone secretion, had evolved before bovids emerged as a distinct group. Variation nevertheless occurs. Most striking is the difference in fusion of the binucleate trophoblast cell with uterine epithelium that yields a transient trinucleate cell in bovines and many antelopes, but a more persistent syncytium in wildebeest, sheep and goat. There is considerable variation in placentome number and villus branching within the placentome. Many antelopes have right-sided implantation in a bicornuate uterus whilst others have a uterus duplex. Finally, there has been continued evolution of placental hormones with tandem duplication of PAG genes in cattle, differences in glycosylation of placental lactogen and the emergence of placental growth hormone in sheep and goats. The selection pressures driving this evolution are unknown though maternal-fetal competition for nutrients is an attractive hypothesis.


Introduction
Cattle, buffaloes, sheep and goats account for around 80% of livestock ( Bar-On, Phillips, and Milo, 2018) and are of great cultural and economic importance (Clutton-Brock, 2012). The domesticated species derive from a much greater pool of bovids that includes antelopes and antelope-like bovines. It is a premise of this review that the evolution of bovids, and of ruminants in general, is best understood by considering wild as well as domestic species. Using this approach, we shall consider how the reproductive strategy of bovids has contributed to their evolutionary success.
The specific focus is placentation. The epitheliochorial placenta of ruminants is an advanced form of placentation that was derived from a more invasive one (Elliot and Crespi, 2009;Mess and Carter, 2006;Wildman et al., 2006;Carter and Mess, 2017). It is diffuse in the basal tragulids and cotyledonary in pecoran ruminants (Wooding and Burton, 2008). Placentation in Bovidae has not been reviewed since the work of Hradecký (Hradecky, 1986;Hradecky et al., 1988aHradecky et al., , 1988b. There are variations in the number of uterine caruncles and fetal cotyledons as well as in the shape and internal structure of the placentomes. Binucleate trophoblast cells and their hormones are of especial interest and have undergone evolution at the ordinal and family level.
The narrative will begin with the remarkable ascent of the bovids and their later domestication. Next it will deal with various aspects of placentation, especially the fate of the binucleate trophoblast cell and the role of syncytin, the product of an endogenous retroviral gene, in its fusion with uterine epithelium to form a fetal-maternal hybrid cell or syncytium. To search for evolutionary trends, the account will proceed to variations in placentation among the twelve tribes of bovids as well as the placental hormones. A concluding section will refer to the ruminant and bovid trees and attempt to define the branching points at which innovations occurred.

Evolution and domestication of bovids
Relations between eutherian mammals have been clarified by molecular phylogenetics and phylogenomics. Four orders share a common ancestor: carnivores (Carnivora), horses and tapirs (Perissodactyla), pangolins (Pholidota) and even-toed ungulates (Cetartiodactyla). Three of these orders are characterized by epitheliochorial placentation whereas the endotheliochorial placentation of carnivores reflects an evolutionary reversal (Carter and Mess, 2017). Whales and dolphins are now counted among the even-toed ungulates (Spaulding et al., 2009) together with 10 orders of land mammals, 6 of which are ruminants (Tab. 1). 4 Anim. Reprod., v.16, n.1, p.3-17, Jan./Mar. 2019 tragulids (present-day chevrotains). Tragulids are basal to pecoran ruminants and have a diffuse placenta without placentomes. Their relative abundance was greatly reduced by the Middle Miocene when there were roughly equal numbers of giraffids and bovids. The only giraffids living today are the okapi and four species of giraffe, but giraffids were more abundant and diverse in earlier epochs. What explains their fate?
One suggestion is that a longer gestation period put them at a disadvantage compared to bovids, which could better adapt to climate change by adopting seasonal breeding patterns (Clauss and Roessner, 2014). This might also explain why giraffids were never numerous in the Eurasian fauna (Fig. 1b), where they faced additional competition from deer and their kin (Cervidae). Table 1 Terrestrial mammals of Order Cetartiodactyla (Burgin et al., 2018). In addition, the order includes 11 families (40 genera

Domestication
Why were some ruminants domesticated and others not? Most domesticated animals (Tab. 2) are derived from species that are highly social and live in herds with a dominance hierarchy (cats are a notable exception). Another factor is flight distance: it is shorter in goats, which rely on agility to escape from predators, than in gazelles, which rely on speed. This could explain why goats were tamed and gazelles were not, although both had been a major food source for huntergatherers (Clutton-Brock, 2012). The transition from a hunter-gatherer economy to herding was regional and A trait shared by domesticates of many species is the retention of juvenile traits in the adult (neotony). This extends to behaviour as much as to anatomical features. It likely results from selection for tameness and against aggressive behaviour (Budiansky, 1994). The most convincing evidence comes from an experiment on the silver fox (Vulpes vulpes) where selection for tameness over many generations resulted in foxes that behaved much like domestic dogs (Belyaev, 1979). The experiment was recently revisited with a view to identifying genes associated with tame and aggressive behaviours (Kukekova et al., 2018).
The first bovids to be domesticated were goats and sheep. Some of the earliest evidence for both is from Jericho (Clutton-Brock, 2012). However, genomic evidence points to three centres of goat domestication within the Fertile Crescent (Daly et al., 2018). Domestication of taurine cattle occurred in the Near East and of humped or zebu cattle in present day Pakistan. They are thought to have been derived from two species of aurochs (Bos primigenius and B. namadicus) (Clutton-Brock, 2012;Pitt et al., 2018). Cattle were introduced to other regions as herding spread to Europe and North Africa and throughout Asia.
Thus, taurine and zebu cattle were introduced independently to southern Africa via the Sahara and the Horn of Africa and at least 120 breeds developed there. A separate domestication of taurine cattle in Egypt is not supported by current analyses of genetic data (Pitt et al., 2018). Elsewhere there was domestication of the yak in Tibet, Bali cattle in Indonesia and of water buffalo both in China (the swamp buffalo) and South Asia (the river buffalo). The gayal of South and Southeast Asia is a semi-domesticated gaur (Tab. 2).
There has, of course, been a great deal of crossbreeding and not just between taurine and zebu cattle. As an example, introgression has occurred between yak and Tibetan cattle (Wu et al., 2018). The latter have acquired haplotypes of two genes in the hypoxiainducible factor pathway: EGLN2 and HIF3a. One consequence of this is a reduced haematopoietic response to hypoxia, which confers an advantage to cattle living at high altitude. Introgression has also occurred between cattle and wild species, including the European bison or wisent (Wu et al., 2018).
Livestock biomass, 80% of it from cattle, buffaloes, sheep and goats, greatly exceeds the biomass of all wild mammals (0.1 versus 0.007 gigatons of carbon) ( Bar-On et al., 2018). Budiansky has argued, "domestication seems natural only because it happened, but it happened only because it was natural" (Budiansky, 1994). To the extent that animals were complicit in the domestication process, as he contends, this was a highly successful strategy. Anim. Reprod., v.16, n.1, p.3-17, Jan./Mar. 2019

Placentation
Most domesticated bovids have a bicornuate uterus with the placenta extending from the pregnant to the non-pregnant horn. There are four rows of caruncles in each horn and placentomes are 80-110 in number. This pattern was likely present in the common ancestor of pecoran ruminants (Klisch and Mess, 2007).
Large, grazing antelopes of the Tribe Hippotragini, such as the sable antelope (Hippotragus niger) and some members of Alcelaphini, such as black wildebeest (Connochaetes gnou), have a uterus duplex with a bifurcated cervical canal (Hradecky, 1982). This restricts the placenta to the gravid horn (Hradecky, 1983a). Interestingly, the fetal membranes are also restricted to one horn in the hartebeest (Alcelaphus buselaphus), an antelope with a bicornuate uterus (Hradecky, 1983a). In Hippotragini, perhaps as a compensation, there are 6-8 rows of caruncles and the placenta has a larger than average number of placentomes (Hradecky, 1983a(Hradecky, , 1983b.

Litter size
The reproductive strategy of ruminants entails a long gestation with a singleton fetus that is precocial at birth. This pattern was present in the common ancestor of Ferungulata and only pigs have reverted to a larger litter of three or more offspring (Klisch and Mess, 2007;Carter and Mess, 2017). Among the Bovidae, singleton pregnancy is the rule. Exceptions include twinning in the nilgai and four-horned antelope (Boselaphini) as well as some gazelles (Antilopini). Domestic sheep and goats (Caprini) bear twins and sometimes triplets, but many wild caprines have singleton pregnancies (Castello, 2016). In cattle and rarely in other ruminants, pigs and camels, vascular connections between twins of separate sex result in masculinization of the female reproductive tract. Freemartins are not known from other orders of mammal (Padula, 2005).

Binucleate trophoblast cells
Binucleate trophoblast cells (BNCs) were first described for the sheep placenta by Assheton (Assheton, 1906). Their critical role in ruminant placentation was not fully understood until the work of Wooding in sheep (Wooding, et al., 1980) and cattle (Wooding and Wathes, 1980;Wooding, 1982). This established that BNCs migrate towards and fuse with uterine epithelial cells to form a fetomaternal hybrid -either a syncytium or trinucleate cells. Some authors favour referring to BNCs as trophoblast giant cells (Klisch et al., 1999).
Because the maternal-fetal interface in ruminants contains hybrid elements, some authors recommend the term "synepitheliochorial" (Wooding and Burton, 2008), reserving epitheliochorial for species where only trophoblast and maternal epithelium occur at the interface (Carter and Enders, 2013).
BNCs occur in all ruminants, including chevrotains, where they contribute to a fetomaternal syncytium (Kimura et al., 2004;Wooding et al., 2007). A syncytium forms at early stages of implantation in deer and cattle and in sheep and goats it persists until later in gestation (Wooding et al., 1980;Wooding, 7 1982), whereas trinucleate cells are found in cattle (Wooding and Wathes, 1980) and deer (Wooding et al., 2018). Although this is the usual outcome in cattle, multinucleate cells can be observed that seemingly result from fusion of additional BNCs with trinucleate cells (Klisch et al., 1999). The trinucleate cells of cattle and deer atrophy and die after releasing their granules and are reabsorbed by the trophectoderm ( Wooding and Wathes, 1980;Wooding, 1982).
It has been difficult to discuss the alternative fates of BNCs in an evolutionary context, since comparative studies of bovid placenta tend to document the presence of binucleate cells, but not their subsequent fate (Hradecky, 1986;Hradecky et al., 1988b;Benirschke, 2012). Recently, a survey was made of a wide range of ruminants including a chevrotain (Tragulidae), eight bovids (Bovidae), eight deer (Cervidae), the pronghorn (Antilocapridae) and a giraffe (Giraffidae) (Wooding et al., 2018). This study used antibodies raised against pregnancy-associated glycoproteins and immunogold staining to localise granules in BNCs, trinucleate cells and syncytium. A fetomaternal syncytium was confirmed for the chevrotain, but most of the pecoran ruminants had trinucleate cells. Exceptions were the sheep and the blue wildebeest (Connochaetes taurinus) (Fig. 3).
The BNCs of ruminants are an evolutionary novelty. They do not occur in other even-toed ungulates nor are they related to the multinucleate giant cells of camelids (Klisch and Mess, 2007). Since they migrate to and fuse with the uterine epithelium, the population of BNCs requires continual renewal. In their early developmental stages, BNCs of cattle retain contact with the basal membrane of the trophoblast (Attiger et al., 2018). This raises the possibility that BNCs arise by mitosis of basally located stem cells rather than being derived from uninucleate trophoblast cells (Attiger et al., 2018). Table 3 Variations in placentation across the bovid family. Hradecký is the source for data on villous structure (Hradecky et al., 1988b

Syncytins
Fusion of cells from two individuals to form a viable unit seems problematic. The explanation seems to lie in expression by the binucleate cell of one or more syncytins. These are coded by genes of retroviral origin that have been incorporated into the genome and are expressed in the placenta where they promote cell fusion. Syncytin genes occur in primates, rodents, rabbits, carnivores and even in a marsupial and a lizard (Dupressoir et al., 2012;Cornelis et al., 2015;Cornelis et al., 2017). They are not orthologous genes as each represents an independent capture from a retrovirus (Lavialle et al., 2013). Syncytins are encoded by retroviral env genes. In a retrovirus, the envelope protein is responsible for fusion with a host cell (Dupressoir et al., 2012). Most syncytins function to promote fusion of trophoblast cells to form multinucleate syncytiotrophoblast. In pecoran ruminants, however, a syncytin (Syncytin-Rum1) was found that enables the binucleate cell to fuse with a uterine epithelial cell. The envelope protein has an immunosuppressive domain, and this is highly conserved in ruminant syncytin (Cornelis et al., 2013).
A second syncytin gene (Fematrin-1) is expressed by BNCs in taurine cattle, Bali cattle and water buffalo (Bovini) and sitatunga (Tragelaphus spekii) (Tragelaphini), although not in domestic sheep and goat (Caprini) (Nakaya et al., 2013). It was suggested that this gene could account for the formation of trinucleate cells in bovines as opposed to syncytial plaques in caprines (Nakaya et al., 2013). However, that interpretation is not compatible with recent work (Wooding et al., 2018), from which it is clear that trinucleate cells are basal and syncytium formation is the derived state.
It is not possible to pinpoint the time when a retrovirus became incorporated in the genome. A provisional estimate for Syncytin-Rum1 is more than 30 million years ago (mya) (Cornelis et al., 2013) and for Fematrin-1 18.3-25.4 mya (Nakaya et al., 2013).

Placentome shape
There are three basic shapes to placentomes (Andresen, 1927;Mossman, 1987). The convex type is typical of cattle and often has a narrow base giving it a mushroom shape. Flat placentomes are found mostly in deer, and the concave type is found in sheep and goats. The sheep placentome has a central concavity where extravasated maternal erythrocytes are taken up and processed by columnar trophoblast cells (Burton et al., 1976;Myagkaya and Vreeling-Sindelarova, 1976). Many antelopes have concave placentomes, examples being Kirk's dik-dik (Antilopini), the roan antelope (Hippotragini) and the topi (Damaliscus lunatus) (Alcelaphini) (Andresen, 1927). It is not known if these placentomes have a central haemophagous region like that of the sheep.
Branching patterns were much more varied among the antelopes. An interesting example was Reduncini where only 10-20 placentomes were formed. One might anticipate a complex internal structure to maximize the maternal-fetal interface, but the reverse was the case. In the kob, villi were up to 15 mm long and coursed through the entire depth of the placentome with minimal branching (Fig. 4C). Villi were evenly distributed and averaged 200 µm in diameter at half height (Fig. 4D). The same applied to the lechwe (Kobus leche), waterbuck and Bohor reedbuck. A similar pattern was found in the unrelated bush duiker (Cephalophini), where the villi were up to 10 mm long with only slight branching and 200-300 µm in diameter at half height. Findings were similar in bay duiker (Cephalophus dorsalis) and Maxwell's duiker (Philantomba maxwellii).
In contrast, the villi were extensively branched in the impala (Aepycerotini), where the villi were up to 10 mm long (Fig. 4 E-F) and in the steenbok (Raphicerus campestris) (Antilopini), where they were up to 2.5 mm long. Moderate branching occurred in the black wildebeest (Alcelaphini) and sable antelope (Hippotragini) with villi 8 and 4 mm in length, respectively.  (Hradecky et al., 1988b)). Note the branching of the villi that is typical of Subfamily Bovinae. B. Cross section of a placentome from the same animal. C. Longitudinal section of a placentome from the kob (Kobus kob) at 97 days gestation; fetal length 23.4 cm (M26812; kob 7 in (Hradecky et al., 1988b)). Note the straight villi with minimal branching known only from impala and duikers. D. Cross section of a placentome from the kob at 127 days gestation; fetal length 34.5 cm (M26814; kob 9 in (Hradecky et al., 1988b)). E. Longitudinal section of a placentome from the impala (Aepyceros melampus) in mid-gestation; fetal length 11 cm (M26463; impala 118 in (Hradecky et al., 1988b)). Extensive branching of the villi is representative of the pattern in most antelopes. F. Cross section of a placentome from the same animal. Labels: b, binucleate trophoblast cell; ca, base of caruncle; fc, fetal capillary; ms, maternal stroma; v, villus. Hradecký voucher specimens in the Harland W. Mossman Embryological Collection, University of Wisconsin Zoological Museum.

Placental hormones
Placental hormones act to adapt maternal physiology to pregnancy and lactation (Napso et al., 2018). In bovids, interferon-tau from the trophectoderm of the elongated blastocyst acts as a luteotrophic factor . Peptide hormones secreted by the placenta proper include pregnancy-associated glycoproteins (PAGs), placental lactogens, prolactinlike proteins, and placental growth hormone. The coding genes evolved from non-placental ones through one or more rounds of gene duplication.

Placental lactogens
Placental lactogens have arisen by convergent evolution in primates, rodents and ruminants (Soares, 2004). In ruminants, tandem duplication of the prolactin gene has given rise to genes for placental lactogen (PL) and a family of prolactin-like proteins.

taurinus)
(Alcelaphini), the roan antelope (Hippotragini), several sheep and goats and the muskox (Ovibos moschatus) (Caprini). The sequences of bovine and ovine PLs are very different, suggesting a high rate of molecular evolution (Wallis, 1993). Moreover, bovine PL is heavily glycosylated whereas ovine PL is not (Colosi et al., 1989). Within the placentome, ovine PL is located exclusively to the BNCs and the syncytium formed by fusion of BNCs with uterine epithelial cells (Wooding et al., 1992). Bovine PL is similarly restricted to BNCs and the hybrid trinucleate cells (Wooding and Beckers, 1987). Indeed, fusion of BNCs with maternal epithelium has been proposed as a mechanism for delivering PL and other placental hormones to the maternal tissues (Wooding and Beckers, 1987).
The best documented role of ovine prolactin is to stimulate the secretion of histotrophe ("uterine milk") by binding to prolactin receptors in the uterine glands (Noel et al., 2003). PL can be detected in the maternal plasma of sheep, but seems to have little effect on mammary gland development (Min et al., 1997). Bovine PL is barely detectable in maternal plasma (Byatt et al., 1987) so may act only in a paracrine manner. In contrast, the convergently evolved PLs of rodents are important for pregnancy maintenance (Galosy and Talamantes, 1995) and mammary gland development . Human and primate PLs (evolved by duplication of the growth hormone gene) act on maternal metabolism to improve nutrient availability to the fetus (Handwerger, 1991).
Ovine and bovine PLs do reach the fetal circulation, although by an unknown mechanism, and may promote fetal growth (Taylor et al., 1980;Byatt et al., 1987).

Placental prolactin-like proteins
Unlike PLs, the prolactin-like proteins (PRPs) do not bind to prolactin or growth hormone receptors and their function is unclear. No less than 12 PRP genes are expressed in bovine placenta (Ushizawa et al., 2005;Larson et al., 2006). Sheep and goat have two PRP genes each and they are homologous with genes in cattle (Ushizawa et al., 2007a;Ushizawa, et al., 2007b). Expression of the PRPs is limited to BNCs (Milosavljevic et al., 1989;Ushizawa et al., 2007a;Ushizawa et al., 2007b).

Placental growth hormone
A placental growth hormone has been described in domestic sheep and goat. The protein was localized to uninucleate and binucleate trophoblast cells and the syncytium of the ovine placentome (Lacroix et al., 1996). It is possible that gene duplication took place during the evolution of caprine ruminants (Wallis et al., 1998), but it could have a deeper origin as discussed below. Uterine gland hyperplasia and histotrophe secretion in sheep require the sequential action of interferon-tau, ovine PL and ovine placental growth hormone (Noel et al., 2003).

Pregnancy-associated glycoproteins
Pregnancy-associated glycoproteins (PAGs) are placental hormones that evolved in the common ancestor of Cetartiodactyla by duplication of the pepsin-F gene, which codes for an aspartic proteinase (Hughes et al., 2003;Wallace et al., 2015). There were two further rounds of gene duplication. The first gave rise to the "ancient PAGs," which retained the active site of the proteinase. The second occurred in the ruminant lineage and many of these "modern PAGs" lack the active site (Wallace et al., 2015). Bovids have a high number of PAG genes: 18 protein-coding genes and 14 pseudogenes have been identified in taurine cattle (Telugu et al., 2009). Multiple genes were revealed by Southern blotting of genomic DNA from a wide selection of bovids (Tab. 4) (Xie et al., 1997), but detailed studies have been confined to domestic species including sheep and goats (Xie et al., 1997;Tandiya et al., 2013). Antibodies raised against bovine PAG-1 and ovine PAG-2 have been used for immunolocalization studies across a range of species (Tab. 4;Wooding et al., 2018), but the full range of PAG proteins remains to be explored.
Nevertheless, there is an interesting distribution of gene expression between trophoblast of the cotyledons and intercotyledonary areas ( Wooding et al., 2005;Touzard et al., 2013). In bovine placenta, modern PAGs were transcribed mainly in the cotyledons and ancient PAGs in the intercotyledonary areas; the exception was ancient PAG-2, which was expressed in the cotyledons (Touzard et al., 2013). The cellular localization of representative PAG proteins also differed. Thus PAG-1 (ancient) was detected in the cytoplasm of BNCs from the intercotyledonary chorion. PAG-11 (modern) was detected in BNCs from both cotyledonary and intercotyledonary BNCs; however, no BNC stained for both PAG-1 and PAG-11. Finally, PAG-2 (ancient) was found in the uninucleate trophoblast cells of the cotyledons, but appeared to be excluded from BNCs (Touzard et al., 2013).
The function of PAGs remains unclear. In cattle, two of the ancient PAGs are more highly expressed in cotyledons from early gestation, perhaps implying a role in placentation (Wiedemann et al., 2018). Further, it has been suggested that ancient PAGs act as linking molecules at the fetomaternal interface and that modern PAGs have an immunomodulatory function (Wooding et al., 2005). These proteins are produced in large quantities and some are secreted to the maternal circulation, where they can be used to monitor pregnancy (Wallace et al., 2015). Table 4 Pregnancy-associated glycoproteins in bovids (Xie et al., 1997;Wooding et al., 2005;Wooding et al., 2018)

Conclusions
Many key features of placentation had evolved before Bovidae emerged as a separate family (Carter, 2014;Carter and Mess, 2017;Tab. 5). Even so, there are several interesting innovations in one or more of the twelve tribes. One can estimate when they emerged by reference to the current tree (Fig. 2), which follows the phylogeny of Hassinin et al. (Hassanin et al., 2012). There are some differences in branching order in the tree offered by Bibi (Bibi, 2013), such as a more basal position for Caprini, but these do not affect the arguments in this section.
One curious example is the appearance in Hippotragini and some members of Alcelaphini of a uterus duplex with a bifurcated cervical canal (Hradecky, 1982), restricting the placenta to the gravid horn (Hradecky, 1983a). This feature could have emerged in the common ancestor of the two tribes as even the hartebeest, with its bicornuate uterus, has the fetal membranes restricted to one horn (Hradecky, 1983a).
In many other species of Suborder Antilopinae, implantation usually occurs in the right horn, although ovulation alternates between right and left ovaries and the fetal membranes can extend to the left horn. Exceptions are not well documented. In the oribi, for example, implantation was clearly in the right horn of three females whereas a fourth carried twins, one confined to the right horn while the other fetus "occupied a portion of both horns" (Kellas, 1966). A proclivity to right-sided implantation may therefore have arisen after the common ancestor of antelopes diverged from that of bovines with reversal occurring mainly in species, such as domestic sheep, where twinning frequently occurs.
Another significant feature concerns the fate of BNCs. As recently documented (Wooding et al., 2018), the usual pattern is fusion with a uterine epithelial cell to form a trinucleate cell. In sheep and goats, however, a fetomaternal syncytium is formed that persists throughout gestation. The same happens in the blue wildebeest (Wooding et al., 2018). It is, therefore, notable that these species represent two tribes (Caprini and Alcelaphini) that share a common ancestor with a third (Hippotragini) ( Fig. 2; Hassanin Bibi, 2013). It seems likely that the most recent common ancestor of these three tribes had a persistent fetomaternal syncytium.
Most placental hormones predate the emergence of Bovidae (Tab. 5). An interesting exception is placental growth hormone. As noted above, it is likely that duplication of the growth hormone gene took place during the evolution of caprine ruminants (Wallis et al., 1998). However, as there is no data for antelopes, it could have a deeper origin. Therefore, it would be especially interesting to know if placental growth hormone is present in the wildebeest; as noted it shares a common ancestor with caprines (Hassanin et al., 2012) and like them forms a fetomaternal syncytium (Wooding et al., 2018). The sequences of placental lactogen and degree of glycosylation differ between domestic cattle and sheep, but there is insufficient data to show when the divergence occurred and whether it might be linked to the emergence of a placental growth hormone in the latter.
Fewer evolutionary novelties have been observed in Subfamily Bovinae. The capture of an env gene with the properties of a syncytin (Fematrin-1) is the most convincing example, as it is present in taurine cattle, Bali cattle and water buffalo (Bovini) and also in the sitatunga (Tragelaphini) (Nakaya et al., 2013). There are more PAGs in cattle than in sheep, but it is not clear when the tandem duplication of PAG genes took place.
As discussed elsewhere (Carter and Enders, 2013), it is difficult to identify the selection pressures behind placental evolution. One suggestion has been that evolution in cotyledon number, shape and interdigitation of fetal villi and maternal crypts has been driven by maternal-fetal competition for nutrients (Klisch and Mess, 2007). The assumption is that allocation of resources to the fetus will increase the fitness of the offspring, but decrease the mother's fitness for future reproduction (Haig, 2008). The convergent evolution of placental lactogens and growth 13 hormones in rodents, ruminants and primates can also be interpreted as attempts by the fetus to influence maternal acquisition and allocation of nutrients (Haig, 2008;Napso et al., 2018). Elsewhere it was argued that the interplay between the fetal semi-allograft and the maternal immune system is a significant force in placental evolution (Carter and Enders, 2013). An advantage of epitheliochorial over more invasive placentation is that it allows the immune system to be primed against uterine infection rather than suppressed to allow trophoblast invasion (Carter and Enders, 2013). Fusion of BNCs with uterine epithelium exposes trophoblast to the immune system, but the process is aided by syncytins with immunomodulatory properties (Cornelis et al., 2013) and the modern PAGs may also act in this fashion (Wooding et al., 2005). Unfortunately, little is known about the immune system of the uterus even in cattle and sheep. For the antelopes and antelope-like bovines that live in the wild this is one of many aspects that deserves further study. Table 5. Timeline of placental evolution, showing the appearance of distinct characters in various taxonomic clades (Carter, 2014). Approximate dates (million years ago, mya) above the family level are taken from Meredith et al. (Meredith et al., 2011) and for families and subfamilies from Bibi (Bibi, 2013