MOLECULAR GENETICS OF THE STEROID RECEPTOR SUPERFAMILY 
Ronald M. Evans, Ph.D., Investigator 
It is axiomatic that the physiology of multicellu- 
lar organisms depends on hormonal signals between 
cells that act through receptors to elicit changes in 
both cell shape and function. In the case of steroid 
and thyroid hormones and active metabolites of vi- 
tamins A (retinoic acid) and D, the receptors are 
intracellular transcription factors that activate or re- 
press expression of target genes. Molecular studies 
by the Evans laboratory have revealed that these re- 
ceptors are evolutionarily related and share an over- 
all common structure that defines the steroid/ 
thyroid hormone receptor superfamily. The labora- 
tory has focused on the characterization of this hor- 
mone receptor family to provide insight into the 
origin and evolution of endocrine systems and 
to provide a simple means of characterizing the 
genetic networks that underlie cell fate and 
physiology. 
Retinoid X Receptor 
Derivatives of vitamin A, or retinoids, are a group 
of signaling molecules that are essential for several 
life processes, including growth, differentiation, 
and epithelial homeostasis. In the embryo, retinoic 
acid (RA) has been implicated in many roles, in- 
cluding the formation of the developing nervous 
system. RA exerts its effects on transcriptions for two 
classes of nuclear receptors, the retinoic acid recep- 
tors (RARs) and the retinoid X receptors (RXRs). 
The classification into the RAR and RXR subfamilies 
is mainly based on the differences in primary struc- 
ture, sensitivity to synthetic retinoid ligands, and 
the ability to regulate expression of target genes. 
A major question raised by the discovery of two 
retinoid-responsive systems is whether their func- 
tions are independent, interactive, or redundant. 
One approach to the answer is to determine whether 
these systems share common or discrete down- 
stream target genes. In previous studies, the Evans 
laboratory has provided evidence that expression of 
the gene for the cellular retinol-binding protein 
type II (CRBPII) is dramatically induced by the RXR 
but not by the RAR. This discovery supported the 
suggestion that RXRs may be functionally distinct 
from the RAR system. 
These data raise the question whether the two re- 
ceptor systems recognize common ligands. This is 
particularly relevant because they are both activated 
by all-trans RA, but in vitro, only the RARs are capa- 
ble of binding RA with high affinity. This discrep- 
ancy led the Evans group to propose that during a 
functional assay, cells convert all-trans RA to an 
RXR-specific ligand referred to as retinoid X. These 
observations also led them to develop a strategy to 
identify a metabolite of all-trans RA that might bind 
and activate RXR. It was not possible, in advance, to 
deduce how RA might be transformed, because the 
number of potential stereoisomers, conformers, and 
modifications could generate thousands of alterna- 
tive molecules. 
The Evans laboratory has recently described the 
identification of retinoid X as 9-cis RA, a naturally 
occurring RA stereoisomer. Although not previously 
seen in living organisms, 9-cis RA is evidently a prev- 
alent vertebrate hormone. It represents the first 
such hormone identified in 20 years and the ninth 
hormone of the steroid/thyroid family of molecules. 
RXR Heterodimers 
Accessory factors present in nuclear extracts ap- 
pear to be necessary for high-affinity binding of the 
RAR, thyroid hormone receptor (TR) , and vitamin D 
receptor (VDR) to their cognate hormone response 
elements. It was proposed that these accessory fac- 
tors might be members of the receptor family that 
stimulate receptor binding through heterodimer 
formation. Recently, in work partially funded by the 
National Institutes of Health, the Evans laboratory 
has found that RXR functions as a common heterodi- 
meric partner for the VDR, TR, and RAR. These het- 
erodimers display a high degree of cooperativity in 
binding to target DNA in vitro, as well as in aug- 
menting functional activity in vivo. 
Why the VDR, TR, and RAR interact with a com- 
mon partner is not yet clear. Presumably, in this 
complex, RXR does not require its unique ligand 
(9-cis RA) but rather serves as a silent partner to 
promote hormone response of the associated recep- 
tor. It is clear that characterization of the RXR gene 
family, its patterns of expression, and the nature of 
the RXR ligand in controlling this process will fur- 
ther the understanding of the complex molecular 
nature of hormonal signaling. 
Heterodimers in Drosophila 
A Drosophila homologue of the RXR has been 
identified that corresponds to the ultraspiracle 
(usp) locus and is associated with an embryonic 
segmentation defect and larval lethality. The homol- 
ogy between usp and RXR is sufficient to suggest 
398 
