The nitrogenase metalloenzyme family, essential for supplying fixed nitrogen to the biosphere, is one of life's key biogeochemical innovations. The three forms of nitrogenase differ in their metal dependence, each binding either a FeMo-, FeV-, or FeFecofactor where the reduction of dinitrogen takes place. The history of nitrogenase metal dependence has been of particular interest due to the possible implication that ancient marine metal availabilities have significantly constrained nitrogenase evolution over geologic time. Here, we reconstructed the evolutionary history of nitrogenases, and combined phylogenetic reconstruction, ancestral sequence inference, and structural homology modeling to evaluate the potential metal dependence of ancient nitrogenases. We find that active-site sequence features can reliably distinguish extant Mo-nitrogenases from V-and Fe-nitrogenases and that inferred ancestral sequences at the deepest nodes of the phylogeny suggest these ancient proteins most resemble modern Mo-nitrogenases. Taxa representing early-branching nitrogenase lineages lack one or more biosynthetic nifE and nifN genes that both contribute to the assembly of the FeMo-cofactor in studied organisms, suggesting that early Monitrogenases may have utilized an alternate and/or simplified pathway for cofactor biosynthesis. Our results underscore the profound impacts that protein-level innovations likely had on shaping global biogeochemical cycles throughout the Precambrian, in contrast to organism-level innovations that characterize the Phanerozoic Eon. K E Y W O R D Sancestral sequence reconstruction, metal cofactor, metalloenzyme, nitrogen fixation, nitrogenase This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Amino acid substitution models are a key component in phylogenetic analyses of protein sequences. All commonly-used amino acid models available to date are time-reversible, an assumption designed for computational convenience but not for biological reality. Another significant downside to time-reversible models is that they do not allow inference of rooted trees without outgroups. In this paper, we introduce a maximum likelihood approach nQMaker, an extension of the recently published QMaker method, that allows the estimation of time non-reversible amino acid substitution models and rooted phylogenetic trees from a set of protein sequence alignments. We show that the non-reversible models estimated with nQMaker are a much better fit to empirical alignments than pre-existing reversible models, across a wide range of datasets including mammals, birds, plants, fungi, and other taxa, and that the improvements in model fit scale with the size of the dataset. Notably, for the recently published plant and bird trees, these non-reversible models correctly recovered the commonly estimated root placements with very high statistical support without the need to use an outgroup. We provide nQMaker as an easy-to-use feature in the IQ-TREE software (http://www.iqtree.org), allowing users to estimate non-reversible models and rooted phylogenies from their own protein datasets. The datasets and scripts used in this paper are available at https://doi.org/10.6084/m9.figshare.14516712.

The formation of a placenta is critical for successful mammalian pregnancy and requires remodelling of the uterine epithelium. In eutherian mammals, remodelling involves specific morphological changes that often correlate with the mode of embryonic attachment. Given the differences between marsupial and eutherian placentae, formation of a marsupial placenta may involve patterns of uterine remodelling that are different from those in eutherians. Here we present a detailed morphological study of the uterus of the brushtail possum (Trichosurus vulpecula; Phalangeridae) throughout pregnancy, using both scanning and transmission electron microscopy, to identify whether uterine changes in marsupials correlate with mode of embryonic attachment as they do in eutherian mammals. The uterine remodelling of T. vulpecula is similar to that of eutherian mammals with the same mode of embryonic attachment (non-invasive, epitheliochorial placentation). The morphological similarities include development of large apical projections, and a decrease in the diffusion distance for haemotrophes around the period of embryonic attachment. Importantly, remodelling of the uterus in T. vulpecula during pregnancy differs from that of a marsupial species with non-invasive attachment (Macropus eugenii; Macropodidae) but is similar to that of a marsupial with invasive attachment (Monodelphis domestica; Didelphidae). We conclude that modes of embryonic attachment may not be typified by a particular suite of uterine changes in marsupials, as is the case for eutherian mammals, and that uterine remodelling may instead reflect phylogenetic relationships between marsupial lineages.

Amino acid substitution models are a key component in phylogenetic analyses of protein sequences. All amino acid models available to date are time-reversible, an assumption designed for computational convenience but not for biological reality. Another significant downside to time-reversible models is that they do not allow inference of rooted trees without outgroups. In this paper, we introduce a maximum likelihood approach nQMaker, an extension of the recently published QMaker method, that allows the estimation of time non-reversible amino acid substitution models and rooted phylogenetic trees from a set of protein sequence alignments. We show that the non-reversible models estimated with nQMaker are a much better fit to empirical alignments than pre-existing reversible models, across a wide range of datasets including mammals, birds, plants, fungi, and other taxa, and that the improvements in model fit scale with the size of the dataset. Notably, for the recently published plant and bird trees, these non-reversible models correctly recovered the commonly known root placements with very high statistical support without the need to use an outgroup. We provide nQMaker as an easy-to-use feature in the IQ-TREE software (http://www.iqtree.org), allowing users to estimate non-reversible models and rooted phylogenies from their own protein datasets.

In mammalian pregnancy, the uterus is remodeled to become receptive to embryonic implantation. Since non-invasive placentation in marsupials is likely derived from invasive placentation, and is underpinned by intra-uterine conflict between mother and embryo, species with non-invasive placentation may employ a variety of molecular mechanisms to maintain an intact uterine epithelium and to prevent embryonic invasion. Identifying such modifications to the uterine epithelium of marsupial species with non-invasive placentation is key to understanding how conflict is mediated during pregnancy in different mammalian groups. Desmoglein-2, involved in maintaining lateral cell-cell adhesion of the uterine epithelium, is redistributed before implantation to facilitate embryo invasion in mammals with invasive placentation. We identified localization patterns of this cell adhesion molecule throughout pregnancy in two marsupial species with non-invasive placentation, the tammar wallaby (Macropus eugenii; Macropodidae), and the brushtail possum (Trichosurus vulpecula; Phalangeridae). Interestingly, Desmoglein-2 redistribution also occurs in both M. eugenii and T. vulpecula, suggesting that cell adhesion, and thus integrity of the uterine epithelium, is reduced during implantation regardless of placental type, and may be an important component of uterine remodeling. Desmoglein-2 also localizes to the mesenchymal stromal cells of M. eugenii and to epithelial cell nuclei in T. vulpecula, suggesting its involvement in cellular processes that are independent of adhesion and may compensate for reduced lateral adhesion in the uterine epithelium. We conclude that non-invasive placentation in marsupials involves diverse and complementary strategies to maintain an intact epithelial barrier.

15The nitrogenase metalloenzyme family, essential for supplying fixed nitrogen to the biosphere, is 16 one of life's key biogeochemical innovations. The three isozymes of nitrogenase differ in their 17 metal dependence, each binding either a FeMo-, FeV-, or FeFe-cofactor for the reduction of 18 nitrogen. The history of nitrogenase metal dependence has been of particular interest due to the 19 possible implication that ancient marine metal availabilities have significantly constrained 20 nitrogenase evolution over geologic time. Here, we combine phylogenetics and ancestral sequence 21 reconstruction -a method by which inferred, historical protein sequence information can be 22 linked to functional molecular properties -to reconstruct the metal dependence of ancient 23 nitrogenases. Inferred ancestral nitrogenase sequences at the deepest nodes of the phylogeny 24

Nearly neutral theory predicts that the effectiveness of natural selection will vary with species' neutral genetic diversity, or effective population size (Ne). This relationship allows comparisons between species with differentNeto reveal selective preferences for a variety of traits, including those where the effects of selection may be subtle. We investigate one such trait — amino acid composition — and find that the efficacy of selection, measured as the Codon Adaptation Index of Species (CAIS), can predict amino acid frequencies across vertebrates. In addition to amino acid frequency (an evolutionary outcome), we also measure amino acid flux (an evolutionary process) on the branches separating mouse and rat, and compare it to flux on the branches separating human and chimpanzee. We decompose flux into a shared component that reflects a variety of universal artifacts, and a species-specific component capturing any effects ofNe. The shared component of flux is correlated with disorder propensity, while the species component is correlated with selective preference as measured by CAIS, i.e., the two methods agree on which amino acids are preferred under more effective selection. Within highly exchangeable pairs of amino acids, we find strong preferences and flux imbalances in favor of arginine over lysine, and valine over isoleucine. Interestingly, these preferences are the same as those found in thermophilic lineages vs. mesophilic relatives.

No abstract