The Amt/Mep/Rh superfamily is a unique branch of the major facilitator superfamily with hundreds of members from a vast array of organisms spread in the three domains of life [1]. It is further divided into Amt/Mep subfamily as represented by E. coli Amt (ammonium transporter) and S. cerevisiae Mep (methylammonium permeases) and Rh subfamily by Rh antigen-related proteins of animal red cells and epithelial tissues. Model Amt/Mep members of prokaryotes, eukaryotic microbes and plants have been extensively studied owing to their importance for nitrogen uptake and assimilation. Amt/Mep were thought to act like active transporters for the charged NH4+ specie, but more recent studies yielded key physiological and structural evidence showing that they are gas channels for NH3 and not NH4+ transporters [2,3]. These studies have identified Amt/Mep as the first biological gas channels and established the structural principle for gas conductance function. While ammonium as the transport specificity of Amt/Mep has been confirmed, the substrate for the Rh subfamily remains uncertain. Rh proteins were proposed as Amt/Mep functional equivalents given their limited sequence similarity and apparent distant relationship. Subsequent studies of Rh homologues in different organisms yielded controversial results with regard to substrate identification. Physiological studies by the Berkeley group explored the substrate for Rh proteins in the unicellular green alga, indicating that the substrate for Rh proteins is the gas CO2 [2,4]. Other studies used mammalian Rh proteins for transient expression in yeast Δmep mutant or frog oocytes to assay ammonium or methylammonium uptake. Although the latter studies have not reached on the agreement of which species (gas vs cation) was involved and whether uptake was electrogenic or neutral, they all indicate that Rh proteins transport ammonium and/or methylammonium [2,5]. One caveat in these assays was that the concentrations of ammonium or methylammonium used were overwhelmingly higher than that under homeostatic physiological conditions. Here we assembled a large dataset of the Amt/Mep/Rh superfamily to explore what selective forces might govern the evolutionary relatedness of the two subfamilies and how their members would differ from one another in terms of their primary sequence relationship, transmembrane topologic organization and molecular phylogenetic patterns. We further analyzed and compared the sequence divergence and codon composition for potential occurrence of adaptive evolution in each subfamily separately. Our studies revealed the following observations that are consistent with the view that Amt/Mep and Rh subfamilies are sequence-related yet substrate-distinct biological gas channels. (1) Amt/Mep genes and Rh genes differ in organismal distribution. The former are ubiquitous in eubacteria, fungi and plants, are scattered in archaea and invertebrates, and are absent in vertebrates. In contrast, Rh genes are rare in prokaryotes and absent in plants, but show an increasing occurrence in microbial eukaryotes (except fungi) and become ubiquitous in metazoan with a major expansion in vertebrates. (2) Rh genes and Amt/Mep genes coexist in a diverse spectrum of species including microbial eukaryotes and certain invertebrates, signifying their paralogous relationship over a long period of evolutionary time. (3) Although Amt/Mep and Rh proteins are related to one another by marginal sequence relatedness, they are subject to independent evolution after their early isolation into two paralogous branches apparently in primitive eukaryotic microbes. (4) Members of the Amt/Mep subfamily are highly divergent internally, which could be shaped by environmental adaptation, and they are divided into three major clusters where archaeal Amt are dispersed with prokaryotic and eukaryotic members. (5) Amt/Mep proteins and Rh proteins share a limited set of conserved amino acids most of which (15/17 residues) are hydrophobic ones or the small glycine residues that can be assigned to the interval transmembrane segments. (6) Amt/Mep subfamily genes and Rh subfamily genes are highly different in the usage of the second codon positions, which identified a major driving force that discriminates the two subfamilies. (7) Amt/Mep proteins and Rh proteins strikingly differ in charge distribution: the former essentially lack membrane-embedded negative charged amino acids, in sharp contrast with the latter. (8) Along with the evolution of vertebrates, the erythrocyte Rh proteins have been subject to strong positive selection, particularly in the lineage from rodents to higher mammals. We discuss the implications of these findings in the light of gas channel function and organismal physiology.