17beta-hydroxysteroid dehydrogenases (17beta-HSDs) regulate androgen and estrogen concentrations in mammals. By 1995, four distinct enzymes with 17beta-HSD activity had been identified: 17beta-HSD-types 1 and 3, which in vivo are NADPH-dependent reductases; 17beta-HSD-types 2 and 4, which in vivo are NAD+-dependent oxidases. Since then six additional enzymes with 17beta-HSD activity have been isolated from mammals. With the exception of 17beta-HSD–type 5, which belongs to the aldoketo-reductase (AKR) family, these 17beta-HSDs belong to the short chain dehydrogenases/reductases (SDR) family. Several 17beta-HSDs appear to be examples of convergent evolution. That is, 17beta-HSD activity arose several times from different ancestors. Some 17beta-HSDs share a common ancestor with retinoid oxido-reductases and have retinol dehydrogenase activity. 17beta-HSD-types 2, 6 and 9 appear to have diverged from ancestral retinoid dehydrogenases early in the evolution of deuterostomes during the Cambrian, about 540 million years ago. This coincided with the origin of nuclear receptors for androgens and estrogens suggesting that expression of 17beta-HSDs had an important role in the early evolution of the physiological response to androgens and estrogens.

Various explanations have been proposed to account for complex differentiation and development in humans, despite the human genome containing only two to three times the number of genes in invertebrates. Ignored are the actions of adrenal and sex steroids—androgens, estrogens, glucocorticoids, mineralocorticoids, and progestins—which act through receptors that arose from an ancestral nuclear receptor in a protochordate. This ligand-based mechanism is unique to vertebrates and was integrated into the already robust network of transcription factors in invertebrates. Adrenal and sex steroids influence almost all aspects of vertebrate differentiation and development. I propose that evolution of this ligand-based mechanism in a primitive vertebrate was an important contribution to vertebrate complexity. Sequencing of genomes from a cephalochordate, such as amphioxus, and from hagfish and lamprey will establish early events in the evolution of steroid hormone signaling, and also allow genetic studies to elucidate how vertebrate complexity depends on steroid hormones.

Various lipophilic signals, including ecdysone, retinoic acid, estradiol, cortisol, testosterone and progesterone, act through nuclear receptors, a large group of transcription factors that regulate differentiation and development, which are central to metamorphosis. Here, we focus on environmental factors (e.g. climate, chemicals) in the evolution of nuclear receptors and other signal transduction proteins that interact with heat shock protein 90 (Hsp90), a chaperone that promotes the proper folding and trafficking in cells of proteins. Hsp90 also promotes functional folding of some mutant signal proteins, which would be otherwise destabilized. Stress diverts Hsp90 from stabilizing mutant signal transduction proteins and towards promoting proper folding of stress-damaged proteins and preventing the aggregation of denatured proteins. Reduced Hsp90 levels allow expression of cryptic mutations in signal transduction proteins and new developmental patterns. Thus, environmental stress in the form of extreme climate can influence the evolution of metamorphosis. We discuss how extreme cooling called “Snowball Earth”, which occurred in the late Proterozoic, diverted Hsp90 from chaperoning signal-transduction proteins. As a result, pre-existing mutant signal transduction-proteins were expressed in animals. Some mutations were selectively advantageous in animals that are seen in the Cambrian, when diverse pathways for metamorphosis in metazoans first appear in the fossil record. Other environmental factors, such as biological chemicals (e.g. the antibiotic geldanamycin) can reduce the levels of active HSP90 providing another mechanism for the emergence of mutant signaling pathways.

Vertebrates appear about 500 million years ago in the fossil record. This is only 25–50 million years after the great explosion of multicellular invertebrate body plans in the early Cambrian. On a geological time scale, this interval is a ‘blink of an eye’, suggesting that the evolution of regulatory genes is likely to be important in the origins of vertebrates. Here we present evidence for a role of steroid receptors in this process based on a phylogenetic analysis suggesting that receptors for androgens, glucocorticoids, mineralocorticoids, and progesterone evolved from an ancestral steroid receptor gene by two successive duplications over a brief time that could have coincided with the origins of vertebrates. Moreover, the duplications of these steroid receptors may be additional evidence for the two duplications on a genome-scale that have been proposed to be important in the evolution of vertebrates. The two successive duplications of steroid receptor genes and their subsequent sequence divergence leading to steroid-specific receptors that regulate growth, development, reproduction and homeostasis in vertebrates may have been one of the events important in vertebrate survival after the Cambrian during global extinctions that occurred about 440 and 370 million years ago.

Steroid dehydrogenases regulate the access of active steroids to their receptors. In particular, 11beta-hydroxysteroid dehydrogenase type-1 (11beta-HSD1) and 11beta-HSD2 regulate the levels of glucocorticoids, such as cortisol, and 17beta-HSD1 and 17beta-HSD2 regulate the levels of androgens and estrogens. Human 11beta-HSD1 and 11beta-HSD2 are distant homologs, with less than 25% amino acid sequence identity, as are human 17beta-HSD1 and 17beta-HSD2. In contrast, human 11beta-HSD2 and 17beta-HSD2 are close homologs, with about 43% sequence identity. Until recently, deciphering early events in the evolution 11beta-HSD2 and 17beta-HSD2 was difficult because only mammalian sequences were available. The completely sequenced Takifugu, Tetraodon and medaka genomes and the almost completed zebrafish genome provide an opportunity to investigate the evolution of 11beta-HSD2, 17beta-HSD2, and 11beta-HSD1. Unexpectedly, a search of the Takifugu, Tetraodon and medaka genomes only found an ortholog to 11beta-HSD2 and none to 17beta-HSD2, while the zebrafish genome contains orthologs of both enzymes. This suggests that 17beta-HSD2 was lost in teleosts after the divergence of zebrafish and medaka. Also unexpectedly, searches with 11beta-HSD1 only identified several fish 11beta-HSD3s, as well as an ortholog in Ciona, indicating that 11beta-HSD3 is the ancestor of 11beta-HSD1.

Receptors for the adrenal and sex steroids arose by a series of gene duplications from an ancestral nuclear receptor in a primitive vertebrate, at least 540 million years ago. Sequence analysis indicates many steroidogenic and steroid-inactivating enzymes, including cytochrome P450s and hydroxysteroid dehydrogenases, arose at the same time. The estrogen receptor (ER) appears to be the ancestral steroid receptor. Initially, the redundant duplicated ER had a low specificity for its new ligand. This raises the question: “How was specificity for responses to different steroids regulated early in the evolution of steroid receptors?” Selective expression of these steroid-metabolizing enzymes provided specificity for different steroid responses in primitive vertebrates. 17beta-hydroxysteroid dehydrogenase-type 1 (17beta-HSD-type 1) and 17beta-HSD-type 2, which preferentially catalyze the reduction and oxidation at C17 of androgens and estrogens, respectively, provide an example of this mechanism. Selective expression of either 17beta-HSD-type 1 or 17beta-HSD-type 2 can regulate synthesis or inactivation of androgens or estrogens in specific cells. Steroids also were important in the evolution of land animals, which began about 400 million years ago. Steroidogenic and steroid-inactivating enzymes were recruited to regulate steroid-mediated responses as organ function became more complex. For example, in the kidney 11beta-HSD-type 2 prevents binding of glucocorticoids to the mineralocorticoid receptor, which is crucial for aldosterone-mediated regulation of electrolyte transport in the distal tubule. We propose that Delta-5 steroids, such as dehydroepiandrosterone and its metabolites, were the ligands for the ancestral ER. Understanding the actions of Delta-5 steroids in amphioxus and lamprey may shed light on adrenarche and neurosteroid actions in humans.